gallbladder polyps: systematic review reward your...

SPONSORS

Canadian Association of General SurgeonsCanadian Society of Surgical Oncology

Canadian Association of Thoracic SurgeonsDepartment of Surgery, Dalhousie University

James IV Association of SurgeonsDepartment of Surgery, University of Toronto

Department of Surgery, University of CalgaryDépartement de chirurgie, Université de SherbrookeDepartment of Surgery, McMaster UniversityDépartement de chirurgie, Université de MontréalDepartment of Surgery, Western University

Vol. 61, No. 3, June/juin 2018canjsurg.ca

Risk factors for infection, revision, death, blood transfusion and longer hospital stay 3 months and 1 year after primary THA or TKA

Tranexamic acid administration to older patients undergoing primary THA conserves hemoglobin and reduces blood loss

Clinical and operative outcomes of patients with acute cholecystitis who are treated initially with image-guided cholecystostomy

Diagnostic accuracy of transabdominal ultrasonography for gallbladder polyps: systematic review

R7

Ad Number: MBZ_FLE_P26541_CMAJ_E_4Publication(s): CMAJ Magazine

This ad prepared by: SGL Communications • 2 Bloor St. West, Toronto, Ontario • phone 416.413.7495 • fax 416.944.7883 File Location: SGL_MercedesBenzCanada:Magazine Ads:2017:O79685_CMAJ November Corporate Sales:MBZ_FLE_P26541_CMAJ_E_4.indd

JOB SPECIFICS

Client: Mercedes-Benz CanadaCreative Name: C-Wagon Fleet MagazineAgency Docket #: MBZ FLE O79685Main Docket #: SMZ COR O79685Art Director: Paul WiersmaCopy Writer: NonePrint Production: Tom BurtonRetoucher: Jano KirijianLive: 7” x 9.5”Trim: 8.125” x 10.875”Bleed: 8.375” x 11.125”Artwork Scale: 1:1Print Scale: 100%

FILE SPECIFICATIONS:

File Name: MBZ_FLE_P26541_CMAJ_E_4.inddCreation Date: 10-3-2017 10:50 AMLast Modified: 10-13-2017 2:58 PMWorkstation: T16-0268InDesign Version: CC 2015 App. Version: 11.4.1Round #: 1 Page Count: 1GRAPHIC PRODUCTION:

Operator: Dennis SousaCorrection: None

SIGNOFFS:

Creative:

Production:

Premedia:

Proofreading:

Account:

Client:

PREMEDIA OPERATOR:

Operator: RJ

INKS:

Cyan

MAGENTA

YELLOW

BLACK

FONTS & PLACED IMAGES

Family Style

Times RegularCorporate S Light, DemiCorporate A Condensed Regular

File Name Colour Space Eff. Res (PPI)

2016_CARS_PICTURES01287_v2_Eng_mag.tif CMYK 300 ppiMB-star_M_n_4C.epsVorab_MB_WM-claimEN_LB_ML_n_1C.ai

This proof was produced by the following department:

PREPRESS

© 2017 Mercedes-Benz Canada Inc.

Reward your success.Mercedes-Benz Corporate Sales is proud to continue a strong partnership withall Canadian Medical Association members. Reward yourself with a special incentive in addition to existing special offers towards new Mercedes-Benz,Mercedes-AMG and smart vehicles.

See additional details at mercedes-benz.ca/cma

S:7”

S:9.5”

T:8.125”

T:10.875”

B:8.375”

B:11.125”

cover-june18.indd 2 2018-05-22 9:15 AM

For more information, please visit

joule.cma.ca/inhouse

98%course

satisfactionrate

Bringing courses from the Physician Leadership Institute (PLI) in-house to your organization helps build the skills of individuals while strengthening the capacity and impact of your entirehealth care team.

Reduce travel related costs and time

Provide a tailored learning experience based on the needs of your organizatio

Are available to the entire health care team—physicians and non-physicians

Accredited by CFPC and CCHL

Delivered by physicians and other subject matter experts joule.cma.ca/inhouse

Bring the Physician Leadership Institute in-house

BARIATRIC SURGERY – ACADEMIC GENERAL SURGEON

The Department of Surgery (www.surgery.queensu.ca) at Queen’s University is seeking an Academic GeneralSurgeon for a geographically full-time position. The candidate will have Fellowship training in bariatric surgery andhave advanced minimally invasive surgical skills. The preferred candidate will be an established academic surgeonand must have an excellent record of success working in a multi-disciplinary clinical environment. They must alsodemonstrate the strong potential for outstanding educational contributions and exhibit significant academic andscholarly research.

Qualified applicants will hold an MD degree (or equivalent), have completed postgraduate/fellowship qualificationsprior to appointment, and have or be eligible for licensure with the College of Physicians and Surgeons of Ontarioand certification from the Royal College of Physicians and Surgeons of Canada. Preference will be given tocandidates with an advanced degree.

The successful candidate will be required to collaborate with General Surgeons at the Kingston Health SciencesCentre and will also participate in the general surgery call schedule. As a full-time member of the Division ofGeneral Surgery, Department of Surgery, the successful candidate will also be expected to contribute to theteaching, supervision and mentorship of undergraduate, graduate and postgraduate students.

Queen’s University is recognized nationally for the quality of its undergraduate and graduate programs, whichattracts outstanding students. Queen’s University is an integral part of the vibrant Kingston community in the heartof the Thousand Islands region of southeastern Ontario. It has a community spirit and amenities unmatched by anyother Canadian university. The University and the region offer an outstanding academic and communityenvironment (www.queensu.ca).

The University invites applications from all qualified individuals. Queen’s is committed to employment equity anddiversity in the workplace and welcomes applications from women, visible minorities, Aboriginal peoples, personswith disabilities, and LGBTQ persons. All qualified candidates are encouraged to apply; however, in accordancewith Canadian immigration requirements, Canadian citizens and permanent residents of Canada will be givenpriority. To comply with Federal laws, the University is obliged to gather statistical information about how manyapplicants for each job vacancy are Canadian citizens/permanent residents of Canada. Applicants need not identifytheir country of origin or citizenship, however, all applications must include one of the following statements: “I am aCanadian citizen/permanent resident of Canada”; OR, “I am not a Canadian citizen/permanent resident of Canada”.Applications that do not include this information will be deemed incomplete.

The University will provide support in its recruitment processes to applicants with disabilities, includingaccommodation that takes into account an applicant’s accessibility needs. If you require accommodation during theinterview process, please contact the Department of Surgery at 613 533-2660.

The review of applications will begin on June 1, 2018 and will continue until the position is filled. Applications are toinclude a letter summarizing expertise, qualifications, and accomplishments relevant to the position, a CV, and thenames and full contact information of three referees. Complete applications are to be directed to:

Dr. John Rudan, Head, Department of SurgeryQueen’s University, Victory 3Kingston General Hospital

Kingston, ON K7L 2V7Fax 613 544-9174

Email [email protected]

S-8

03

classifiedcjs-jun18.indd 217 2018-05-17 2:46 PMcover-june18.indd 3 2018-05-22 9:16 AM

© 2018 Joule Inc. or its licensors Can J Surg, Vol. 61, No. 3, June 2018 145

EDITORIAL • ÉDITORIAL

148 Patient outcomes versus �nancial outcomes: Which should we listen to? E.J. Harvey

149 Résultats chez les patients ou résultats �nanciers : Que faut-il prioriser? E.J. Harvey

COMMENTARY • COMMENTAIRE

150 One thousand consecutive in-hospital deaths following severe injury: Has the etiology of traumatic inpatient death changed in Canada? D.J. Roberts, C. Harzan, A.W. Kirkpatrick, E. Dixon, S.C. Grondin, P.B. McBeth, G.G. Kaplan, C.G. Ball

153 The current state of resident trauma training: Are we losing a generation? P.T. Engels, N.L. Bradley, C.G. Ball

HISTORY OF SURGERY: FIRST WORLD WAR HISTOIRE DE LA CHIRURGIE : PREMIÈRE GUERRE MONDIALE

155 Massacre of Canadian Army Medical Corps personnel after the sinking of HMHS Llandovery Castle and the evolution of modern war crime jurisprudence J. Doucet, G. Haley, V. McAlister

RESEARCH • RECHERCHE

158 Management of intra-abdominal vascular injury in trauma laparotomy: a South African experience R. Weale, V. Kong, V. Manchev, W. Bekker, G. Oosthuizen, P. Brysiewicz, G. Laing, J. Bruce, D. Clarke

165 Risk factors for infection, revision, death, blood transfusion and longer hospital stay 3 months and 1 year after primary total hip or knee arthroplasty C. Rhee, L. Lethbridge, G. Richardson, M. Dunbar

177 Tranexamic acid administration to older patients undergoing primary total hip arthroplasty conserves hemoglobin and reduces blood loss H. El Beheiry, A. Lubberdink, N. Clements, K. Dihllon, V. Sharma

185 Troponin T monitoring to detect myocardial injury after noncardiac surgery: a cost– consequence analysis G. Lurati Buse, B. Manns, A. Lamy, G. Guyatt, C.A. Polanczyk, M.T.V. Chan, C.Y. Wang, J.C. Villar, A. Sigamani, D.I. Sessler, O. Berwanger, B.M. Biccard, R. Pearse, G. Urrútia, R.W. Szczeklik, I. Garutti, S. Srinathan, G. Malaga, V. Abraham, C.K. Chow, M.J. Jacka, M. Tiboni, G. Ackland, D. Macneil, R. Sapsford, M. Leuwer, Y. Le Manach, P.J. Devereaux

195 Clinical and operative outcomes of patients with acute cholecystitis who are treated initially with image-guided cholecystostomy I. Molavi, A. Schellenberg, F. Christian

REVIEW • REVUE

200 Diagnostic accuracy of transabdominal ultrasonography for gallbladder polyps: systematic review E. Martin, R. Gill, E. Debru

Vol. 61, No. 3, June/juin 2018canjsurg.ca

canjsurg.ca

contents-june18.indd 145 2018-05-18 12:10 PM

146 J can chir, Vol. 61, No 3, juin 2018 © 2018 Joule Inc. or its licensors

DISCUSSIONS IN SURGERY DISCUSSIONS EN CHIRURGIE

208 Users’ guide to the surgical literature: how to assess a qualitative study L. Gallo, J. Murphy, L.H. Braga, F. Farrokhyar, A. Thoma

215 CAREER/CLASSIFIED ADVERTISING CARRIÈRES ET ANNONCES CLASSÉS

© 2018 Joule Inc. ISSN 0008-428X. For information on permission to reproduce material from the Canadian Journal of Surgery (CJS) see canjsurg.ca.

All editorial matter in CJS represents the opinions of the authors and not necessarily those of the publisher. We assume no responsibility or liability for damages arising from any error or omission or from the use of any information or advice contained in CJS, including articles, editorials, reviews, letters and advertisements. All reproduction rights are reserved.

Printed by The Lowe-Martin Group, Ottawa. Appears in February, April, June, August, October and December.

Return undeliverable Canadian copies to the CMA Member Service Centre, 1870 Alta Vista Dr, Ottawa ON K1G 6R7 (email [email protected]).

© 2018 Joule Inc. ISSN 0008-428X. Pour obtenir des renseignements au sujet des permissions à obtenir afin de reproduire des extraits du Journal canadien de chirurgie (JCC), consulter canjsurg.ca.

Tous les articles à caractère éditorial dans le JCC représentent les opi nions de leurs auteurs, qui ne sont pas nécessairement celles de l’éditeur. Nous n’assumons aucune respon sabilité pour les dommages résultant de toute erreur ou omission, ou de l’utilisation de renseignements ou de conseils contenus dans le JCC, y compris les articles, éditoriaux, revues, lettres et annonces. Tous droits de reproduction réservés.

La revue est imprimée par The Lowe-Martin Group, Ottawa. Elle paraît en février, avril, juin, août, octobre et décembre.

Retournez toutes copies canadiennes non livrées au Centre des ser vices aux membres, Association médicale canadienne, 1870, prom. Alta Vista, Ottawa (Ontario) K1G 6R7 (courriel : [email protected]).

We believe in open access to researchTo ensure continued worldwide free access to all CJS content, articles submitted for publication as of Jan. 1, 2014, are subject to a submission fee of $100 (Canadian funds). Submission fees will be waived for corresponding authors af�liated with CJS sponsors. Accepted Research, Review and Continuing Medical Education

articles are subject to a publication fee of $700, and Commentaries and Discussions are subject to a publication fee of $500, payable on acceptance in Canadian funds.

Benefits of open access• For researchers and institutions: increased visibility, usage and impact for their work• For government: a better return on investment for funding research• For society: ef�cient, effective patient care resulting in better outcomes

CJS articles are available free of charge on the journal website (canjsurg.ca) and in PubMed Central.

contents-june18.indd 146 2018-05-18 12:10 PM


COEDITORS CORÉDACTEURS

Edward J. Harvey, MD, Montreal [email protected]

LCol Vivian C. McAlister, MB, London [email protected]

ASSOCIATE EDITORS RÉDACTEURS ASSOCIÉS

BASIC SCIENCE AND SURGICAL BIOLOGYIan McGilvray, MD, PhD, Toronto Timothy Daniels, MD, Toronto

BREAST SURGERYMuriel Brackstone, MD, London

CARDIOVASCULAR SURGERYMichel Carrier, MD, Montreal Michael Chu, MD, London

CRITICAL CARERaymond Kao, MD, London

ENDOCRINE SYSTEMSam Wiseman, MD, Vancouver

EVIDENCE-BASED MEDICINEMichelle Ghert, MD, Hamilton Kelly Vogt, MD, London

GASTROINTESTINAL AND COLORECTAL SURGERYMarcus Burnstein, MD, Toronto Jason Park, MD, Winnipeg

GLOBAL SURGERYDan Deckelbaum, MD, Montreal Vanessa Fawcett, MD, Edmonton

GENERAL SURGERYDaniel Birch, MD, Edmonton Andrew Beckett, MD, Montreal

HEPATOBILIARY AND PANCREATIC SURGERYShiva Jayaraman, MD, Toronto

MEDICAL EDUCATIONCarol Hutchison, MD, Calgary

MILITARY MEDICINEHomer Tien, MD, Toronto Carlos J. Rodriguez, MD, Bethesda

ORTHOPEDIC FOOT AND ANKLE SURGERYKarl-André Lalonde, MD, Ottawa

ORTHOPEDIC SURGERYGraham Elder, MD, Sault Ste. Marie

PEDIATRIC SURGERY, GENERALVacant

PEDIATRIC SURGERY, ORTHOPEDICJames G. Wright, MD, MPh, Toronto

SPINAL SURGERYRaja Rampersaud, MD, Toronto

SPORTS MEDICINEPaul Martineau, MD, Montreal

SURGICAL ONCOLOGYGeoff Porter, MD, Halifax

SURGICAL ONCOLOGY, MUSCULOSKELETALFrank O’Dea, MD, St. John’s

SURGICAL ONCOLOGY, SOFT TISSUEVacant

THORACIC SURGERYRichard Malthaner, MD, London

TRAUMA SURGERY, ORTHOPEDICWilliam Dust, MD, Saskatoon

TRAUMA SURGERY, SOFT TISSUEMary vanWijngaarden-Stephens, MD, Edmonton Chad Ball, MD, Calgary

VASCULAR SURGERYKent Mackenzie, MD, Montreal Tom Forbes, MD, Toronto

EDITORIAL STAFF ÉQUIPE DE RÉDACTION

MANAGING EDITOR DIRECTRICE DE LA RÉDACTIONWendy Carroll ([email protected]) tel 800 663-7336 (Canada, USA) or 613 731-8610 x8419

EDITORIAL COORDINATOR COORDONNATRICE DE LA RÉDACTIONHolly Clark ([email protected]) tel 800 663-7336 (Canada, USA) 613 731-8610 x8424

COPYEDITOR • RÉVISEUREGloria Baker ([email protected])

TRANSLATION • TRADUCTIONCMA French language services

PRODUCTION STAFF ÉQUIPE DE PRODUCTION

MANAGER, PRODUCTION & GRAPHIC DESIGN GESTIONNAIRE DE LA PRODUCTION ET CONCEP-TION GRAPHIQUEJennifer Pershick

PRODUCTION DESIGNERS INFOGRAPHISTES DE LA PRODUCTIONCarole Lalonde, Sarah O’Neill, Clara Walker

WEB PUBLISHER ÉDITION EN LIGNEJames Manship

ADVERTISING PUBLICITÉ

ADVERTISING ANNONCES PUBLICITAIRESLaurie McLeod, Susan Ritchie 613 731-8610 x8460/8475 [email protected]

ADVERTISING SALES VENTES PUBLICITAIRESTrish Sullivan ([email protected]) cell 905 330-8770

CMAJ GROUP STAFF ÉQUIPE CMAJ GROUP

EDITOR IN CHIEF, INTERIM ÉDITEUR EN CHEF INTÉRIMAIREDiane Kelsall

PUBLISHER, INTERIM ÉDITRICE INTÉRIMAIREHolly Bodger

The Canadian Journal of Surgery aims to contribute to the effective continuing medical education of Canadian surgical specialists and to provide surgeons with an effective vehicle for the dissemination of observations in the areas of clinical, basic science and education research.Readers can find CJS online at canjsurg.ca.Submission of new manuscripts can be made at http://mc.manuscriptcentral.com/cjs.

Le Journal canadian de chirurgie vise à dispenser une éducation médicale continue efficace aux spécialistes en chirurgie au Canada, et fournir aux chirurgiens un mécanisme efficace pour diffuser les constatations de la recherche clinique, fondamentale et éducative.Les lecteurs trouveront en direct le JCC à l’adresse canjsurg.ca.Nous favorisons l’envoi électronique de manuscrits. Veuillez visiter le http://mc.manuscriptcentral.com/cjs.

Owned and published by Joule Inc., a wholly

owned subsidiary of the Canadian Medical Association.

All prescription drug advertisements have been cleared by the Pharmaceutical Advertising Advisory Board.Toutes les annonces de médicaments prescrits ont été approuvés par le Conseil consultatif de publicité pharmaceutique.

masthead-june18.indd 147 2018-05-18 12:14 PM



The views expressed in this editorial are those of the author and do not necessarily reflect the position of the Canadian Medical Association or its subsidiaries.

Patient outcomes versus financial outcomes: Which should we listen to?

O utcomes are important in medicine. Tradition-ally, we relied on our own internal scoring to determine how well we were doing with our

care delivery. But we have spent the last two decades making a concerted effort to determine what patient-reported outcomes (PROs) can mean for validation of our procedures and protocols. We are still stuck when it comes to interpretation of these scores.

A PRO is the information reported by a patient, but without interpretation of that response by the sur-rounding health care team. The strength of a PRO is that it pertains directly to the patient’s quality of life and functional status and is a patient-based reflection of how well we did with our interventions. Unfortu-nately a lot of internal patient information may go into an evaluation. Their perceived outcome in a global evaluation after appendectomy may have as much to do with parking and hospital food as it does with the skill of their surgeon. Researchers have accounted for this by using directed questionnaires to interpret patient responses.

The tools or instruments used to measure PROs are patient-reported outcome measures (PROMs) and patient-reported experience measures (PREMs). A PROM is a questionnaire administered to determine the patient’s health status. A PREM gauges a patient’s perception of the health care they have received. These tools give a more objective value to subjective patient input. We use many objective scores for treatment evaluation, includ-ing physiologic parameters, markers and radiographs among others, but PROMs are now often being used as adjuncts to these scores.

Hospital administrators have other outcomes in mind. They are more interested in cash �ow, and PREMs can be used for economic viability scoring. Some general sys-tems (e.g., EuroQol, 12-Item Short Form Health Sur-vey) measure health-related quality of life that can be used to estimate quality-adjusted life-years (QALYs), which can be used as discussion points for economic evaluation. As surrogate measurements of outcomes over time, QALYs help us to justify expensive interventions or equipment.

Much as described in the article by Rhee and col-leagues in this month’s issue,1 we have increasingly

replaced PROs with economic outcomes. It is much easier to examine these data using data-mining tech-niques and large patient care databases with financial information embedded. This is a double-edged sword. We would like to know if measures are economically viable. In today’s cash-strapped hospital environment, it behooves us to promote �nancially responsible treatment options. Unfortunately, the disconnect between the �nance department and patient outcomes will only be accentuated if we do not marry the two in some way. All new technology is going to be more expensive up front. There is no way, for example, that a modern fixation plate with minimal contact points and locking screw tech-nology will be able to compete with a cold milled �at plate on a per cost analysis. New minimally invasive lapa-roscopy equipment is always more expensive. In the United States, reimbursement is being based on patient satisfaction scores.2 But relying only on patient satisfac-tion scores without considering the overall care picture may be dangerous. Patient satisfaction does not depend on hospital protocols, on compliance with surgical care measures for accreditation or even on whether the pro-cedure is accepted practice.

I fear if we do not get a handle on what to measure and at what time point, and on whether the results are in context, then we will be hard pressed to determine if we actually treat patients adequately.

Edward J. Harvey, MD

Coeditor, Canadian Journal of Surgery

Competing interests: E.J. Harvey is the Chief Medical Of�cer of Greybox Healthcare (Montreal) and Chairman of the Board of NXT-Sens Inc. (Montreal).

DOI: 10.1503/cjs.006818

References

1. Rhee C, Lethbridge L, Richardson G, et al. Risk factors for infec-tion, revision, death, blood transfusion and longer hospital stay 3 months and 1 year after primary total hip or knee arthroplasty. Can J Surg 2018;61:165-76.

2. Lyu H, Wick EC, Housman M, et al. Patient satisfaction as a possi-ble indicator of quality surgical care. JAMA Surg 2013;148:362-7.

edit-june.indd 148 2018-05-18 12:04 PM



Les opinions exprimées dans cet éditorial sont celles de l’auteur et ne représentent pas nécessairement celles de l’Association médicale canadienne ou ses filiales.

Résultats chez les patients ou résultats financiers : Que faut-il prioriser?

En médecine, les résultats comptent. De tous temps, nous nous sommes �és à notre jugement pour évaluer l’ef�cacité des soins que nous prodiguons. Mais nous

avons passé les deux dernières décennies à déployer des efforts concertés pour déterminer l’impact des résultats déclarés par les patients (RDP) sur la validation de nos inter-ventions et de nos protocoles. Et nous sommes toujours dans une impasse quand vient le temps d’interpréter ces résultats.

Les RDP sont des données transmises par nos patients, mais sans que l’équipe soignante concernée puisse les inter-préter. L’atout des RDP est qu’ils sont le re�et direct de la qualité de vie et du statut fonctionnel des patients, et en ce sens, ils témoignent du degré de réussite de nos interven-tions de l’avis des patients eux-mêmes. Malheureusement, beaucoup de données internes propres aux patients peuvent être incluses dans une telle évaluation : les résultats qu’ils perçoivent dans une évaluation globale suivant une appendi-cectomie peuvent avoir autant à voir avec le stationnement ou la nourriture de l’hôpital qu’avec la compétence de leur chirurgien. Les chercheurs ont tenu compte de ce type de variables en utilisant des questionnaires dirigés pour inter-préter les réponses des patients.

Les outils ou les instruments utilisés pour évaluer les RDP sont des mesures des résultats déclarés par les patients (MRDP) et des mesures de l’expérience déclarée par les patients (MEDP). Les MRDP sont des questionnaires administrés pour déterminer l’état de santé des patients. Les MEDP évaluent la perception des patients quant aux soins de santé qu’ils ont reçus. Ces outils donnent une valeur plus objective aux réponses subjectives des patients. Nous utilisons plusieurs paramètres objectifs pour évaluer le traite-ment, notamment des paramètres physiologiques, des mar-queurs et des radiographies, mais les MEDP sont désormais souvent utilisés comme mesures d’appoint.

Les administrateurs hospitaliers ont d’autres résultats en tête. Ils se préoccupent davantage de questions monétaires, et les MEDP peuvent servir à établir des scores de viabilité économique. Certains questionnaires généraux (p. ex., Euro-Qol, SF-12) mesurent la qualité de vie en lien avec la santé, ce qui permet d’estimer les années de vie pondérées en fonc-tion de la qualité (AVAQ), et peut servir de point de départ à une discussion sur les coûts. En tant que marqueurs substi-tuts au �l du temps, les AVAQ nous aident à justi�er des interventions ou des appareils coûteux.

Comme le décrivent bien Rhee et ses collaborateurs dans leur article publié dans ce numéro1, nous avons progressive-

ment remplacé les RDP par des résultats économiques. Ces données sont beaucoup plus faciles à analyser grâce aux tech-niques d’exploration des données et aux volumineuses bases de données sur les soins aux patients qui incluent des ren-seignements de nature économique. C’est une arme à double tranchant. Nous voulons savoir si les mesures sont écono-miquement viables. Compte tenu des contraintes budgétaires auxquelles le milieu hospitalier est soumis actuellement, il nous incombe de promouvoir des options thérapeutiques économiquement responsables. Malheureusement, le clivage entre la dimension économique et les résultats chez les patients ne pourra que s’accentuer si nous n’arrivons pas à arrimer les deux. Toute forme de technologie nouvelle sera coûteuse au départ. En effet, dans une analyse des coûts, com-ment, une plaque d’ostéosynthèse moderne avec un minimum de points de contact peut-elle rivaliser avec les systèmes à base de plaques à vis verrouillées? De même, les nouveaux appa reils de laparoscopie minimalement effractifs sont toujours plus coûteux. Aux États-Unis, le remboursement est établi à partir des scores de satisfaction des patients2, mais il pourrait être dangereux de se �er uniquement aux scores de satisfaction des patients sans tenir compte du tableau d’ensemble. La satisfac-tion des patients ne dépend ni des protocoles hospitaliers, ni de la conformité aux normes chirurgicales agréées, ni même du fait que l’intervention soit une pratique acceptée ou non.

Ma crainte est que si nous ne faisons pas le point sur les paramètres à mesurer et sur le moment opportun de le faire, et si nous ne replaçons pas les résultats dans leur contexte, nous aurons beaucoup de dif�culté à déterminer si les traite-ments prodigués à nos patients sont adéquats.

Edward J. Harvey, MD

Corédacteur, Journal canadien de chirurgie

Intérêts concurrents: E.J. Harvey est médecin hygiéniste en chef de Greybox Healthcare (Montréal) et président du Conseil d’adminis-tration de NXT-Sens Inc. (Montréal).

DOI: 10.1503/cjs.007318

Références

1. Rhee C, Lethbridge L, Richardson G, et al. Risk factors for infec-tion, revision, death, blood transfusion and longer hospital stay 3 months and 1 year after primary total hip or knee arthroplasty. Can J Surg 2018;61:165-76.

2. Lyu H, Wick EC, Housman M, et al. Patient satisfaction as a possible indicator of quality surgical care. JAMA Surg 2013;148:362-7.

edit-june-fr.indd 149 2018-05-18 12:08 PM



One thousand consecutive in-hospital deaths following severe injury: Has the etiology of traumatic inpatient death changed in Canada?

A wide range of factors have traditionally led to in-hospital death fol-lowing severe injury. Although the most dramatic is ongoing hemor-rhage and subsequent physiological exhaustion, other factors include

severe neurological injuries, sepsis, multiorgan failure and the progression of preceding medical comorbidities.1 It is also evident that the relative incidence of these causes has changed over time. More speci�cally, innovation and improvements in care have led to a reported decrease in deaths from acute respiratory distress syndrome (ARDS), multiorgan failure and sepsis. Advance-ments that have contributed to this improvement include lung-protective ven-tilation strategies, a reduction in crystalloid fluid usage, aggressive deep venous thrombosis prophylaxis, early antimicrobial therapy and source control for infection, and damage control resuscitation.2,3

The most relevant and recent literature discussing the etiology of posttrau-matic inpatient death is a review of 813 patients by Kahl and colleagues.4 Although this American group clearly outlines an increase in deaths related to pre-existing patient comorbidities, it remains unclear if these data are also rel-evant in Canada. More speci�cally, given differences in helmet laws, gun con-trol, organized trauma systems and even health care itself, we conducted an ecological, ethically approved review of Canadian inpatient deaths after severe injury (1000 consecutive inpatient deaths [injury severity score (ISS) ≥ 12] at the Foothills Medical Centre in Calgary, Alta.). The dominant goal of this review was to elucidate the potential change in early mortality among severely injured patients. Detailed analyses of patient records, including the medical death certi�cates and coroner’s reports, were completed by two experienced reviewers (C.H. and C.G.B.).

Between Feb. 21, 2005, and Dec. 31, 2013, a total of 9941 consecutive severely injured (ISS ≥ 12) patients were admitted to the Foothills Medical Centre. This cohort was typical of the admission pro�le for critically injured patients to our trauma service (Table 1). The overall early inpatient mortality was 10.1%. The primary causes of death included severe neurological trauma (traumatic brain injury [TBI], spinal cord injury); acute hemorrhagic shock/exsanguination; sepsis; sudden, unexpected inpatient events; and a multitude of other uncommon causes (e.g., medical events that immediately preceded the injury) (Table 2). Over the nearly 8-year study period, deaths due to ARDS and sepsis nearly disappeared (no patient died of ARDS or sepsis after

Derek J. Roberts, MD, PhD Christina Harzan, MD Andrew W. Kirkpatrick, MD, MSc Elijah Dixon, MD, MSc Sean C. Grondin, MD, MPH Paul B. McBeth, MD Gilaad G. Kaplan, MD Chad G. Ball, MD, MSc

This manuscript was presented at the 2016 Trauma Association of Canada (TAC) confer-ence, Halifax, NS, in May, 2016.

Accepted Sept. 6, 2017

Correspondence to: C.G. Ball Associate Professor Department of Surgery University of Calgary Foothills Medical Centre 1403-29th St Northwest Calgary AB T2N 2T9 [email protected]

DOI: 10.1503/cjs.014116

A wide range of factors have traditionally led to early in-hospital death follow-ing severe injury. The primary goal of this commentary was to evaluate the causes of early posttraumatic inpatient deaths over an extended period. Although early posttraumatic in-hospital death remains multifactorial, severe traumatic brain injuries are the dominant cause and have increased in propor-tion over time. Other traditional causes of death have also decreased owing to improved clinical care.

SUMMARY

death-ball.indd 150 2018-05-18 12:22 PM

COMMENTARY

Can J Surg, Vol. 61, No. 3, June 2018 151

2010). The overall mortality decreased signi�cantly among all patients, from 14.4% in 2005 to 8.7% in 2013 (p = 0.001). The mean patient age at the time of death increased over the study period from 56.6 years in 2005 to 61.4 years in 2013 (p = 0.041).

The precise time of death following admission varied signi�cantly and was associated with the primary cause of record (Table 3). Overall, 212 patients (21.2%) also under-went withdrawal of care (WOC) irrespective of their pri-mary injury. Although most (83%) of these patients had

severe neurological injuries (Table 4) as the documented reason for WOC, there was no change in the rate of with-drawal over the study interval (p = 0.24). Also, WOC was more common when patients survived longer in hospital (> 48 hours) (p = 0.021). Overall, 90% of the WOC mech-anisms of death were related to neurological and/or spine injuries. It should be noted that the potential impact of a donation after cardiac death (DCD) program on early mortality and the WOC process was not a practical con-cern as our institution did not use DCD until 2016.

Despite signi�cant differences in traumatic mechanisms and age, patients who died following a severe injury in southern Alberta showed similar patterns to those reported by Kahl and colleagues.4 More speci�cally, overall mortal-ity decreased over time (mean 10%). Although the mean age of death increased, the occurrence of death caused by ARDS, sepsis, and therefore the speci�c sequelae of injury, decreased over time. This is a direct result of advances in

Table 1. Demographic characteristics of 1000 patients admitted to the Foothills Medical Centre trauma service who experienced early inpatient death

Characteristic Mean (range) No. of patients

Age, yr 59.7 (21–81) —

Sex

Male — 720

Female — 280

Injury characteristics

ISS 29 (15–59) —

Mechanism

Blunt — 943

Penetrating — 57

Admission vital signs

Heart rate, bpm 109 (66–150) —

Systolic blood pressure, mm Hg 95 (64–151) —

Respiratory rate, bpm 24 (10–36) —

Glasgow Coma Scale score 11 (3–15) —

bpm = beats/breaths per minute; ISS = injury severity score.

Table 3. Primary cause of death, as related to the timing of admission

Time of death after admission No. of patients

< 6 hours 130

Neurological 21

Exsanguination 104

Sepsis/ARDS 0

Unexpected event 5

6–24 hours 153

Neurological 62

Exsanguination 81

Sepsis/ARDS 4

Unexpected event 6

24–48 hours 163

Neurological 85

Exsanguination 66

Sepsis/ARDS 6

Unexpected event 6

> 48 hours 554

Neurological 461

Exsanguination 22

Sepsis/ARDS 50

Unexpected event 21

ARDS = acute respiratory distress syndrome.

Table 2. Primary causes of inpatient death following critical injury (n = 1000)

Cause of death No. of patients

Neurological (severe traumatic brain or spinal cord injuries)

770

Acute traumatic shock/exsanguination 183

Pelvic fracture 123

Abdominal source 123

Thoracic source 22

Extremity or other source 3

Sepsis/ARDS 24

Gastrointestinal 12

Pulmonary 11

Other 1

Sudden, unexpected inpatient events 21

Pulmonary embolus 12

Myocardial infarction 4

Cerebrovascular accident 3

Tension pneumothorax 1

Cardiac tamponade 1

Progression of preceding event 2

Myocardial infarction 1

Seizure/brain injury 1

Withdrawal of care 212


Table 4. Underlying diagnosis in patients who underwent withdrawal of care (n = 212)

Injury-related cause of death No. of patients

Severe neurological injury

Traumatic brain injury 179

Severe spinal cord injury 13

Unexpected event 9

Progression of preceding event 8

Physiologic exhaustion/exsanguination 2

Sepsis/ARDS 1



COMMENTAIRE

152 J can chir, Vol. 61, No 3, juin 2018

care. The benefits of pattern recognition in caring for elderly, injured patients on a more frequent basis owing to increased volumes also cannot be understated. This improvement applies not only to the clinicians, but also to the nursing staff and rehabilitation specialists.

It is clear that although the primary causes of death were varied, the most common causes were severe neuro-logical injuries (TBI and spinal cord injury; 77.0%) and acute hemorrhagic shock/exsanguination within the torso (18.3%). Although our ability to maintain life in patients with severe neurological injury has improved signi�cantly, we still lack treatments that improve functional recovery for all patients. This concept overlaps with the willingness to engage in the WOC process in many cultures.5 Exsan-guination secondary to ongoing hemorrhage also remains a signi�cant challenge despite advancements such as damage control resuscitation, hybrid operating theatres and recog-nition of the urgency to arrest bleeding. Although some patients died of sudden, unexpected inpatient events (2.1%) and others died from medical events preceding their injuries (0.2%), these uncommon occurrences are considered much more difficult to prevent and predict from a public health point of view.

In summary, although the mean age of patients and the mean age at time of death are increasing over time, the causes of death that occured as a direct result of the in juries are decreasing. Severe neurological trauma and exsanguination remain the dominant causes of inpatient

death and require enhanced efforts at injury prevention. These patterns echo those reported by Kahl and col-leagues4 in the United States. The impact of WOC in scen arios of perceived medical futility is substantial in Cana da and must now be better elucidated in Canadian studies on posttraumatic inpatient mortality.

Af�liations: From the Department of Surgery, University of Calgary, Calgary, Alta. (Roberts, Harzan, Kirkpatrick, Dixon, Grondin, McBeth, Ball); and the Department of Medicine, University of Calgary, Calgary, Alta. (Kaplan).

Competing interests: None declared.

Contributors: All authors contributed substantially to the conception, writing and revision of this article and approved the �nal version for publication.

References

1. Sauaia A, Moore F, Moore EE, et al. Epidemiology of trauma deaths: a reassessment. J Trauma 1995;38:185-93.

2. Ball CG. Damage control resuscitation: history, theory and tech-nique. Can J Surg 2014;57:55-60.

3. Ball CG, Das D, Roberts DJ, et al. The evolution of trauma surgery at a high-volume Canadian centre: implications for public health, prevention, clinical care, education and recruitment. Can J Surg 2015;58:19-23.

4. Kahl JE, Calvo RY, Sise M, et al. The changing nature of death on the trauma service. J Trauma 2013;75:195-201.

5. Ball CG, Navsaria P, Kirkpatrick AW, et al. The impact of country and culture on end-of-life care for injured patients: results from an international survey. J Trauma 2010;69:1323-33.




The current state of resident trauma training: Are we losing a generation?

Paul T. Engels, MD Nori L. Bradley, MD, MSc Chad G. Ball, MD, MSc

This manuscript was presented in part, in oral form, at the Canadian Surgery Forum, Sept. 16, 2017, Victoria, BC, Canada.

Accepted Oct. 20, 2017; Published online Apr. 1, 2018

Correspondence to: P.T. Engels Hamilton General Hospital 6 North Wing, Rm 616 237 Barton St East Hamilton ON L8L 2X2 [email protected]

DOI: 10.1503/cjs.014417

General surgeons provide life-saving trauma care to a large portion of Canad ians. Although trauma care has evolved signi�cantly over the last few decades and now requires fewer operations, when a life-saving operation is required the expectation of competence to perform this operation has not been reduced. A recent study from the United States found decreased resident case-log volumes of trauma operations. Such �ndings raise the alarm of declining exposure of residents to trauma opera-tions and beg the question of whether graduating residents are competent to care for trauma patients. Examination of the Canadian setting reveals a dearth of pub-lished information about the actual exposure of Canadian general surgery residents to trauma care. With the forthcoming evolution of general surgery education into competency-based medical education, we sound a call to action to ensure that all graduating general surgeons are able to provide the care that both the Royal College of Physicians and Surgeons of Canada and the Canadian public demand.

SUMMARY

F or decades, general surgeons have provided clinical care to many of the nearly 250 000 injured Canadians annually. Given both our geography (more than 7 million Canadians reside farther than a one-hour drive from

a level-1 trauma centre) and weather, we continue to rely on local hospitals to provide life- and limb-saving trauma surgeries before transferring patients to a major centre for de�nitive care.1 At the heart of these hospitals and operations are general surgeons.

Our approach to trauma care has evolved in recent decades. Augmented involvement of subspecialties, provision of care within a multidisciplinary team-based paradigm and increasing responsibility for trauma education have required that general surgeons develop their nontechnical skills to be the “quar-terbacks” of trauma care. The wider availability of high-quality diagnostic im aging, evidence-based strategies for nonoperative management and increased use of interventional and endovascular technology have reduced the frequency of emergency operative interventions for trauma. Thankfully, the required com-petency to perform a life-saving operation has not changed.

The impact of modern trauma care on surgical training was recently evaluated: Strummwasser and colleagues2 reviewed the operative case logs of general surgery residents in the United States over the past 15 years. The number of trauma and nontrauma operations as well as the impact of fellowship training on trauma case logs were assessed. Not surprisingly, results revealed both decreased operative trauma and nontrauma exposure among residents over time. The authors con-cluded that “even with fellowship training, the graduating trauma/critical care fel-low is still only as experienced in open trauma surgery as a general surgery resi-dent who graduated 15 years ago.” The authors acknowledge the substantial evolution in trauma care that has occurred over this timeframe, but also raise dif-�cult questions: What is the acceptable operative exposure in residency training? What should a graduating general surgery resident be expected to be able to do?

In Canada, we do not have a national database of resident operative logs. Each Royal College residency program individually assesses its residents and determines whether they have met the clinical objectives of training (i.e., whether they are

current-engels.indd 153 2018-05-18 12:16 PM

COMMENTAIRE


eligible to sit their graduating examination). There is no man-datory minimum for the number of operations performed, type or complexity of managed injuries, or number of call nights completed. The inherent ambiguity of this approach has now fostered competency-based medical education.

So, how do we know that our residents are actually com-petent to provide trauma care? The existing literature reports a disconnect between graduating residents’ perception of their competence (80% state being comfortable on call at a level-1 trauma centre3) and that of the fellowship program directors who supervise them (who deemed two-thirds of incoming fellows unable to operate for 30 unsupervised min-utes in a major procedure4).

In Canada, the provision of trauma care is still embedded within the identity of the Royal College–trained general sur-geon. The objectives of training for general surgery5 outline medical expert competencies in the principles of initial man-agement and resuscitation (i.e., Advanced Trauma Life Sup-port); nonoperative management of traumatic injuries; knowledge of trauma surgical anatomy; management of in jur-ies to the head, neck, chest, abdomen, pelvis, extremities and soft tissues; and special populations, such as trauma in preg-nancy, pediatrics and geriatrics. The current operative “A list,” for which “the graduate must be competent to independ-ently perform these procedures” includes multiple trauma operations: surgical airway; surgical exploration of penetrat-ing neck injuries; resuscitative thoracotomy; trauma laparot-omy, including exploration of retroperitoneal hematomas; and damage control surgery for massive intra-abdominal hemorrhage or multiple intra-abdominal injuries. Further-more, the collaborator competency mandates the graduate must be able to “assume the role of trauma team leader (TTL).” The bar appears to be set, and it is appropriately high. After all, these graduating surgeons will be expected to provide life- and limb-saving trauma care to the 20% of Canadians who do not have local access to a level-1 trauma centre staffed with dedicated expert trauma surgeons.1

Although mounting evidence supports that performance of a specific number of operations does not necessarily equate to competency, we can agree that surgical trainees should receive appropriate exposure to trauma patients and conditions. So, what is the typical trauma experience for Canadian general surgery residents? The answer is not clear. We do not have mandatory national operative case logs to monitor resident volume. A search of Canadian journals uncovered a dearth of literature on this topic. Even more concerning is that conversations with colleagues, chief resi-dents and recent graduates entering fellowship programs reveal an emerging narrative of graduates not ever having fulfilled the role of TTL successfully or having been an independent operator for a trauma laparotomy. Even the nonoperative management of severely injured polytrauma patients on a trauma service remains suspect.

Can we assume that our residents are obtaining exposure to trauma care on a daily basis given their regular on-call

duties? Perhaps. However, there is impressive nationwide variation on the types of clinical rotations residents com-plete, the type of hospital in which they rotate, and the level of training on a given rotation. For example, some training programs possess a single trauma centre, but �ve or more mandatory teaching hospitals. As a result, a resident’s expos ure to trauma care is almost exclusively limited to their three- to six-month rotations through the trauma cen-tre. In contrast, other training programs may centralize their residents to core hospitals (including the trauma cen-tres), possibly providing more consistent, longitudinal and comprehensive trauma care exposure. But even if the resi-dent is on call at a trauma hospital, does (s)he respond to all trauma activations or just the operative cases? Does (s)he function as the TTL? And if (s)he is doing home call, how does (s)he get to the hospital fast enough to gain the experi-ence of doing a resuscitative thoracotomy? Although most programs rotate senior residents through the trauma ser-vice, outliers to this tradition remain.

In the modern era of medical education and competency-based medical education with entrustable professional activities, we argue clear answers to these questions are sorely needed.

As with most problems, the �rst step is recognition and acknowledgement. We hereby challenge all Canadian general surgeons to start a realistic and accurate measure-ment of resident trauma education. We suspect it will be the only method for establishing a clear plan to ensure that all graduating general surgeons are able to provide the care that both the Royal College and the Canadian public demand.Af�liations: From the Department of Surgery, McMaster University, Hamilton, Ont. (Engels); the Department of Surgery, University of British Columbia, Vancouver, BC (Bradley); and the Department of Surgery, University of Calgary, Calgary, Alta. (Ball).



References 1. Hameed SM, Schuurman N, Razek T, et al. Access to trauma systems

in Canada. J Trauma 2010;69:1350-61.

2. Strumwasser A, Grabe D, Inaba K, et al. Is your graduating general surgery resident quali�ed to take trauma call? A 15-year appraisal of the changes in general surgery education for trauma. J Trauma Acute Care Surg 2017;82:470-80.

3. Friedell ML, VanderMeer TJ, Cheatham ML, et al. Perceptions of graduating general surgery chief residents: Are they con�dent in their training? J Am Coll Surg 2014;218:695-706.

4. Mattar SG, Alseidi AA, Jones DB, et al. General surgery residency inadequately prepares trainees for fellowship: results of a survey of fellowhsip program directors. Ann Surg 2013;258:440-9.

5. Objectives of training in the specialty of general surgery. Royal College of Physicians and Surgeons of Canada. Available: www.royalcollege.ca/cs/groups/public/documents/document/ltaw/mtyx/~edisp/rcp-00161418.pdf ( accessed 2017 Sept. 19).

current-engels.indd 154 2018-05-18 12:16 PM


HISTORY OF SURGERY: FIRST WORLD WAR HISTOIRE DE LA CHIRURGIE : PREMIÈRE

GUERRE MONDIALE

Massacre of Canadian Army Medical Corps personnel after the sinking of HMHS Llandovery Castle and the evolution of modern war crime jurisprudence

D uring the First World War, the Allies used two kinds of shipborne transport for casualties. In the English Channel, which the belliger-ents recognized as a Danger Zone, ambulance transport ships wore

no special markings, travelled at night unilluminated and were escorted by armed “P” boats. In the North Atlantic, passenger liners converted to hospi-tal ships followed the rules of the Geneva Convention; they travelled without escort vessels, wearing fully illuminated, huge Red Cross markings. The Canadian Army by 1918 had a well-organized system to repatriate its soldiers to Canada; among the assets available was the Llandovery Castle, one of the �nest “Castle” ships of the Union Castle line, named after a Welsh castle in honour of the line’s Welsh chair. The ship was built for the long East Africa route and was considered well appointed, boasting such features as the �rst elevators aboard a passenger ship. It was a natural choice for conversion to a hospital ship.1,2

At 9:30 pm on June 27, 1918, His Majesty’s Hospital Ship (HMHS) Llandovery Castle was sunk in the Atlantic Ocean by a torpedo from the German submarine U-86, about 116 miles southwest of Fastnet, Ireland. The hospital ship had 980 lifeboat spots for crew and patients, but there were no patients on board as it made its way from Halifax, Nova Scotia, to Liverpool, England. The torpedo struck the starboard side of the ship, killing the engin-eers, wrecking the wireless and extinguishing all lights. Despite signals from the bridge, the ship could not be stopped, and lifeboats had to be lowered from the listing ship into fast-moving water. Two lifeboats were swamped and wrecked along the side, including one that went through the propellers, killing at least 11 Canadian Army Medical Corps (CAMC) nursing sisters — only an orderly survived. The captain of HMHS Llandovery Castle, R.A. Sylvester, was observed to be the last of the initial survivors to leave the stricken ship. The ship’s boilers exploded after it went down about 50 feet. U-86 surfaced as survivors of the torpedo attack tried to make their way to one of the hospital ship’s 19 lifeboats.1,2

Jay Doucet, MD Gregory Haley, MD Vivian McAlister, MB

Accepted May 3, 2018

Correspondence to: A. Gregory Haley Commander Senior Staff Officer Surgeon General Canadian Forces Health Services Group Headquarters c/o National Defence Headquarters 101 Colonel By Dr Ottawa ON K1A 0K2 [email protected]

DOI: 10.1503/cjs.006518

Events after the sinking of the hospital ship Llandovery Castle on June 27, 1918, by the German submarine U-86 outraged Canadians. Survivors aboard a single life raft gave evidence that many of the 234 souls lost had made it to lifeboats but were rammed and shot by the submarine. Many of those who died were nurses. Three German of�cers were charged with war crimes after the war. The submarine’s captain evaded capture. The remain-ing two of�cers’ defence that they were following the captain’s orders failed and they were convicted. This ruling was used as a precedent to dismiss similar claims at the war crime trials after the Second World War. It is also the basis of the order given to members of modern militaries, including the Canadian Armed Forces, that it is illegal to carry out an illegal order.

SUMMARY

wwi-june-18.indd 155 2018-05-18 12:02 PM

HISTOIRE DE LA CHIRURGIE : PREMIÈRE GUERRE MONDIALE


The description of what happened next is taken from the judgment rendered at the German Imperial Court of Jus-tice on July 16, 1921.3 Commander of U-86, First Lieuten-ant Helmut Patzig, ordered that the wreckage, the lifeboats and the survivors be searched for evidence of combatants, or armaments. Captain Sylvester was taken aboard the sub-marine for interrogation, where he was accused of trans-porting eight American airmen. A doctor, Major Thomas Lyon of the CAMC, was also interrogated and accused of being one of the airmen. Dr. Lyon was roughly treated and suffered a broken ankle before he and the others were returned to their lifeboats.4 U-86 came alongside Captain Sylvester’s lifeboat, and he was questioned a second time about explosions heard from the Llandovery Castle as it sank, indicating that it was carrying munitions. After each interrogation, U-86 left the scene. On a third occasion, it returned and appeared to be intent on ramming the life-boats. Some witnesses claimed that the submarine veered off a collision course at the last moment. A short time later,

U-86 �red 14 rounds from the 8.8 cm gun on its stern. On deck were boatswain Meissner, Patzig and Lieutenants Ludwig Dithmar and John Boldt. The court noted that U-86 stayed on the surface for some time, demonstrating no concern of any threat. Witnesses claimed that the aim of those on deck of U-86 was to kill all survivors. Six CAMC members and 18 of the Llandovery Castle crew ultimately survived.5 Among the dead were 14 CAMC nursing of�-cers, some of whom had been seen in stable lifeboats. The sight of the �oating bodies of the dead nurses in their bil-lowing dresses shocked the captain of HMS Morea when that ship sailed past the wreckage the next day.

The Treaty of Versailles forced Germany to hold the �rst war crimes trials, which were deeply unpopular in Germany. Initially the Allies submitted a list of more than 900 Germans they wanted extradited for Allied military trials, but diplo-macy reduced the number of cases to 12, including the Llandovery Castle, to be held at the German Supreme Court at Leipzig.6 The German court found that Patzig, who had absconded to the free city of Danzig by the time of the trial, illegally ordered the sinking of HMHS Llandovery Castle. The judges attributed his subsequent actions to a desire to cover up the crime and to co-opt fellow of�cers. Dithmar and Boldt were initially uncooperative. Dithmar claimed to have manned the forward gun, which was not �red, in an effort to blame Boldt. Then he claimed that he acted under Patzig’s orders. The court rejected the claim of acting on orders, because Dithmar should have known that “the killing of shipwrecked people, who have taken refuge in lifeboats, is forbidden” by convention and by international law. Patzig was found guilty in absentia of homicide. Dithmar and Boldt were convicted of assisting him.

The episode shocked Canadians in the �nal phase of the First World War. Propaganda posters were made, and “Llandovery Castle” was used as the Canadian rallying cry during the Hundred Days Offensive, which ended the war. Despite the prominence the Llandovery Castle massacre held during the war, memory of the tragic event quickly faded, in contrast to the memory of the Halifax Explosion.7 Although it was the largest ever single loss of Canadian medical providers’ lives, there is no speci�c memorial. The medical personnel are commemorated on the Halifax Memorial with other Canadians “lost at sea” in wartime, while the merchant navy crewmen are listed on the Tower Hill monument in London, England. In 1924, unlabelled photographic portraits of HMHS Llandovery Castle’s senior medical of�cer, Lieutenant Colonel Thomas Howard MacDonald, and Matron Margaret Marjory (Pearl) Fraser were placed in the entrance hall of Dalhousie University’s new medical research building, which had been funded by the Rockefeller and Carnegie foundations. Two of us (J.D., V.M.) remember when these portraits were removed after 70 years, during the building’s refurbishment for adminis-tration. It was only when we searched out the identity of the portrait sitters that we became aware of the tragedy.

Contemporary fundraising poster designed to harness popular outrage against the massacre of survivors of the sinking of HMHS Llandovery Castle as it sailed from Halifax, Nova Scotia to Liverpool, England, to collect wounded soldiers. Reproduced with permission from the Canadian War Museum. Appendix 1, available at canjsurg.ca/006518-a1, includes additional images.

wwi-june-18.indd 156 2018-05-22 10:34 AM

HISTORY OF SURGERY: FIRST WORLD WAR


Despite loss of the Llandovery Castle massacre in popu-lar memory, the event did establish a very signi�cant leg-acy. The Leipzig trials were considered deeply offensive by German nationalists and a total failure by the Allies owing to the few convictions and short sentences given. However, the failure in a German court of Lieutenant Dithmar’s defence that he was only following orders was used as precedent to reject similar claims made at the Nuremberg Trials after the Second World War. Canada is proud of its pivotal role in the establishment of the International Crim-inal Court (ICC), which grew out of our experience at the Leipzig and Nuremberg Trials.8 War crimes, for which there is no statute of time limitation, remain relevant today. For example, antipersonnel improvised explosive devices have been shown to cause super�uous injury and unnecessary suffering, ful�lling the medical component of a war crime allegation.9 The ICC also rejects the defence of following orders. Modern militaries, including Canada’s military, observe the Law of Armed Con�ict (LOAC) dur-ing the application of legal force.10 The Canadian Armed Forces teaches and requires all its members to be familiar with LOAC. Prior to each mission, rules of engagement are promulgated, and this brie�ng always highlights that it is wrong for a CAF member to carry out an illegal order. This humane legacy may bring some solace to the relatives and descendants of the crew and the medical staff of HMHS Llandovery Castle, who routinely braved great peril while carrying out their noncombatant duties in the First World War.Affiliations: From the Division of Trauma, Surgical Critical Care, Burns & Acute Care Surgery, University of California San Diego Health (Doucet); and the Canadian Forces Health Services, Ottawa, Ont. (Haley, McAlister).



References

1. Hurd A. The Sinking of Hospital Ships. In: History of the Great War – The Merchant Navy, Volume III. London (UK): John Murray Co. 1929.

2. McGreal S. The War on the Hospital Ships 1914-1918. Barnsley (UK): Pen & Sword Books Ltd.; 2008.

3. German war trials: judgment in case of Lieutenants Dithmar and Boldt. Am J Int Law 1922;16:708-24.

4. Royal Canadian Medical Service Association. Major Thomas Lyon, Canadian Army Medical Corps. Victoria (BC): the association. Avail-able: www.royalcdnmedicalsvc.ca/wp-content/uploads/2015/10/Thomas-Lyon.pdf (accessed 2018 Apr. 27).

5. Hankel G. The Leipzig Trials — German war crimes and their legal consequences after World War One. Volume 4, The Human Rights Series. Dordrecht (NL): Republic of Letters Publishing; 2014.

6. Canadian Great War Project. The sinking of the Canadian Hospi-tal Ship. Available: www.canadiangreatwarproject.com/writing/llandoveryCastle.asp (accessed 2018 Apr. 27).

7. McAlister CN, Marble AE, Murray TJ. The 1917 Halifax Explosion: the �rst coordinated local civilian medical response to disaster in Canada. Can J Surg 2017;60:372-4.

8. Global Affairs Canada. Canada and the International Criminal Court. Ottawa (ON): Government of Canada. Available: www.international.gc.ca/court-cour/index.aspx?lang=eng (accessed 2018 Apr. 27).

9. Smith S, Devine M, Taddeo J, et al. Injury pro�le suffered by targets of antipersonnel improvised explosive devices: prospective cohort study. BMJ Open 2017;7:e014697.

10. Rouillard LP. Ethics, human right and the law of armed con�ict. Can Mil J 2011;12:6-12.

wwi-june-18.indd 157 2018-05-18 12:02 PM



Management of intra-abdominal vascular injury in trauma laparotomy: a South African experience

Background: Intra-abdominal vascular injury (IAVI) is uncommon but continues to be associated with high mortality rates despite technological advances in the past decades. In light of these ongoing developments, we reviewed our contemporary experience with IAVI in an attempt to clarify and re�ne our management strategies and the outcome of these patients.

Methods: We retrospectively reviewed the charts of all patients admitted between January 2011 and December 2014 at a major trauma centre in South Africa who were found to have an IAVI during laparotomy for trauma. We collected demographic and clinical data including mechanism of injury, location and severity of the injury, con-current injuries, physiologic parameters and clinical outcome.

Results: We identi�ed 110 patients with IAVIs, of whom 98 had sustained penetrat-ing injuries (55 gunshot wounds and 43 stab wounds). There were 84 arterial injuries (including 21 renal and 17 mesenteric) and 74 venous injuries (including 21 renal and 17 inferior vena caval). Combined venous and arterial injuries were found in almost one-third of patients (34 [30.9%]). Fifty-seven patients (51.8%) required intensive care admission. The overall mortality rate was 28.2% (31 patients); the rate was 62% for aortic injuries and 47% for inferior vena cava injuries. Liver injury, large bowel injury, splenic injury and elevated lactate level were all associated with a statistically signi�cantly higher mortality rate.

Conclusion: The mortality rate for IAVI remains high despite decades of operative experience in high-volume centres. Open operative techniques alone are unlikely to achieve further reduction in mortality rates. Integration of endovascular techniques may provide an alternative strategy to improve outcomes.

Contexte : Les lésions vasculaires intraabdominales (LVIA) sont rares, mais elles sont toujours associées à un taux de mortalité élevé, malgré les progrès technologiques des dernières décennies. À la lumière de ces renseignements, nous avons passé en revue l’expérience récente en matière de LVIA a�n de clari�er et de parfaire nos stra-tégies de prise en charge et d’améliorer les résultats des patients.

Méthodes : Nous avons examiné de manière rétrospective les dossiers de tous les patients admis entre janvier 2011 et décembre 2014 dans un grand centre de trauma-tologie d’Afrique du Sud chez qui une laparotomie a révélé la présence d’une LVIA. Nous avons recueilli des données démographiques et cliniques portant notamment sur le mécanisme lésionnel, la localisation et la gravité de la lésion, les blessures concomi-tantes, les paramètres physiologiques et l’issue clinique.

Résultats : Nous avons recensé 110 patients atteints de LVIA, dont 98 avaient subi des blessures par pénétration (55 causées par un projectile d’arme à feu et 43 par une arme blanche). Nous avons dénombré 84 lésions artérielles (dont 21 rénales et 17 mésenté-riques) et 74 lésions veineuses (dont 21 rénales et 17 touchant la veine cave inférieure). Dans l’ensemble, nous avons constaté des lésions veineuses et artérielles chez près du tiers des patients (34 patients, soit 30,9 %). Cinquante-sept patients (51,8 %) ont dû être admis à l’unité des soins intensifs. Le taux de mortalité global était de 28,2 % (31 patients); il était de 62 % pour les cas de lésions aortiques et de 47 % pour les lésions touchant la veine cave inférieure. Les lésions au foie, au gros intestin et à la rate ainsi que les taux élevés de lactate ont tous été associés à une hausse statistiquement signi�cative du taux de mortalité.

Conclusion : Le taux de mortalité associé aux LVIA reste élevé malgré des décennies d’expérience chirurgicale dans des centres de traumatologie traitant un grand nombre de patients. Les techniques opératoires ouvertes seules sont peu susceptibles de don-ner lieu à une baisse de ce taux. L’intégration des techniques endovasculaires pourrait constituer une solution de rechange pour améliorer les résultats.

Ross Weale, MBBS, BSc Victor Kong, MSc, PhD, MRCS Vassil Manchev, FCS(SA) Wanda Bekker, FCS(SA) George Oosthuizen, FCS(SA) Petra Brysiewicz, PhD Grant Laing, PhD, FCS(SA) John Bruce, FCS(SA) Damian Clarke, PhD, FCS(SA)

Accepted Sept. 5, 2017; Published online Apr. 1, 2018

Correspondence to: R. Weale Wessex Deanery Winchester SO22 5DH United Kingdom of Great Britain and Northern Ireland [email protected]

DOI: 10.1503/cjs.009717

manage-weale.indd 158 2018-05-18 12:00 PM

RESEARCH


M uch has been published about intra-abdominal vascular injury (IAVI) over the last 60 years. Most of these publications were retrospective;

very few were documented large case series. The persis-tent theme throughout is that the management of IAVI is challenging and that mortality rates are high. Access-ing the injured vessels may be dif�cult, and patients are often in a physiologically compromised state. In addi-tion, competing priorities and concomitant intra-abdominal contamination restrict the options for vascu-lar repair.1

The surgical response to these taxing injuries has been multifaceted. The initial approach was to perfect the sur-gical exposure and to debate the optimal operative tech-niques to manage these injuries.2,3 Although standardizing these surgical lessons was important, it has become appar-ent that there is a natural limit to the extent to which improved operative techniques alone can further improve outcome. There has been much recent focus on imaging techniques, perioperative resuscitation, aggressive replace-ment of blood products and damage-control strategies. However, the mortality rate has not been substantially reduced over the last 2 decades (37% in 1975–1980 v. 33% in 2004–2009).4–6 It is accepted that exsanguination eclipses coagulopathy as the primary cause of death in IAVI, with 1 study showing that only 24% of the deaths from uncon-trollable hemorrhage were attributable to some form of coagulopathy.6 Controlling exsanguination is therefore the single most important objective if one hopes to avoid death in these patients, yet controlling bleeding in these cases remains a challenge. The most recent military reports sug-gest that the next evolution in the management of these exsanguinating injuries will be a combination of surgical and endovascular-based modalities.7,8

In light of these ongoing developments, we reviewed our contemporary experience with IAVI to attempt to clar-ify and re�ne our management strategies and the outcome of these patients.

METHODS

We retrospectively reviewed the charts of all patients admitted between January 2011 and December 2014 through the Pietermaritzburg Metropolitan Trauma Ser-vice, Pietermaritzburg, South Africa, who were found to have an IAVI during laparotomy for trauma. The data col-lected included sex, age, mechanism of injury, location and severity of the injury, and admission physiologic parameters, including lactate level. All abdominal vascular injuries were graded with the American Association for the Surgery of Trauma Organ Injury Scale (AAST-OIS) for abdominal vascular injury9 (Table 1). Other data col-lected included any concomitant solid-organ injury, dura-tion of hospital stay, admission to the intensive care unit (ICU) and end mortality. Ethics approval for this study

and for maintenance of the register was obtained from the Biomedical Research Ethics Committee of the University of KwaZulu-Natal.

Clinical setting

The Pietermaritzburg Metropolitan Trauma Service pro-vides de�nitive trauma care to the city of Pietermaritz-burg, the capital of KwaZulu-Natal province. It also serves as the trauma referral centre for 19 other provincial hospitals within the province. The service manages a high volume of trauma cases. It is pertinent to note the logis-tical challenges of prehospital medicine in this setting. Much of the catchment areas is rural, and the transfer time to hospital is often much greater than in the Euro-pean or North American setting. Many patients whose condition is unstable die even before admission to hospital or access to prehospital medicine.

Injury management

The potential for an IAVI exists in all cases of penetrating torso trauma, and a systematic approach is required in this situation. Important clues to the presence of such an injury include the physiologic state of the patient on presentation and the path of the projectile. Patients with penetrating torso trauma whose condition is unstable are expedited to the operating room. In patients whose condition is stable, imaging may be used selectively, and treatment can be individualized according to the �ndings. At laparotomy a systematic approach is essential. Four-quadrant packing is used to control active bleeding, soiling is mopped up and the bowel eviscerated, and enteric leakage is controlled with packs. Damage-control techniques are implemented,

Table 1. American Association for the Surgery of Trauma Organ Injury Scale for abdominal vascular injury9*

Grade I Unnamed superior mesenteric artery or superior mesenteric vein branches. Unnamed inferior mesenteric artery or inferior mesenteric vein branches. Phrenic artery/vein. Lumbar artery/vein. Gonadal artery/vein. Ovarian artery/vein. Other nonnamed small arterial or venous structures requiring ligation.

Grade II Right, left or common hepatic artery. Splenic artery/vein. Right or left gastric arteries. Gastroduodenal artery. Inferior mesenteric artery, trunk, or inferior mesenteric vein, trunk. Primary named branches of the mesenteric artery (such as ileocolic artery) or mesenteric vein. Other named abdominal vessels requiring ligation/repair.

Grade III Superior mesenteric vein, trunk. Renal artery/vein. Iliac artery/vein. Hypogastric artery/vein. Vena cava, infrarenal.

Grade IV Superior mesenteric artery, trunk. Celiac axis proper. Vena cava, suprarenal and infrahepatic. Aorta, infrarenal.

Grade V Portal vein. Extraparenchymal hepatic vein. Vena cava, retrohepatic or suprahepatic, aorta, suprarenal, subdiaphragmatic.

*Applicable for extraparenchymal vascular injuries. If the vessel injury is within 2 cm of the organ parenchyma, refer to the specific organ injury scale. Increase 1 grade for multiple grade III or IV injuries involving more than 50% vessel circumference. Downgrade 1 grade if less than 25% vessel circumference laceration for grade IV or V.


RECHERCHE


and there is no place in this cohort of patients for pro-longed de�nitive management of enteric injuries.

The management of the IAVI follows standard guide-lines. If a large central hematoma is recognized, proxi-mal control of the aorta at the hiatus is recommended before the hematoma is explored. If a suprarenal injury is suspected, a combined left and right medial visceral rotation is required to expose this part of the aorta. If the injury is believed to be infrarenal, a right rotation usually suffices. Lateral hematomas are not explored unless there is active bleeding or rapid expansion. Pelvic hematomas are explored selectively. The path of the projectile is delineated to make sure it is away from any major vessels in the pelvis. All nonessential vessels are ligated. Simple venous injuries are repaired and complex ones ligated. Essential arterial injuries are managed with primary repair if possible; if this is not possible, more complex individual solutions are used, including tempo-rary shunting and the use of interposition grafting or extra-anatomic bypass. The primary concern is always the physiologic status of the patient, and this dictates management.

Statistical analysis

Data are reported as mean or median values for continu-ous variables and proportions for categorical variables. We used the nonparametric Kruskal–Wallis test to assess differences in presenting heart rate, systolic blood pres-sure and lactate level dependent on AAST-OIS grade. We assessed categorical variables using the χ2 test. We used a post hoc Dunn test to assess differences between groups.

We assessed concurrent organ injuries and examined their relation with mortality and ICU admission using χ2 analysis. The 4 most common organ injuries and the single physiologic parameter most associated with death were included in a multiple logistic regression analysis. The dependent variable was death. The statistical signi�cance level was accepted as 0.05 for all analyses. All statistical analysis was performed with the use of R 3.3.3 software (R Foundation).

RESULTS

During the study period, 1283 patients underwent lapa-rotomy for trauma, of whom 110 (8.6%) were found to have an IAVI. Ninety patients (81.8%) were male, and the mean age was 29 years (Table 2). Of the 110 injuries, 98 (89.1%) were penetrating trauma, and 12 (10.9%) were blunt. Of the 98 penetrating trauma cases, 55 were gun-shot wounds and 43 were stab wounds. The mean admis-sion physiologic parameters were heart rate 105 beats/min, systolic blood pressure 102 mm Hg and serum lac-tate level 5 mmol/L.

Arterial injuries

There were 84 arterial injuries in total: renal (21), mesen-teric (17), aortic (8), external iliac (7), superior mesenteric (6), inferior mesenteric (6), common iliac (5), splenic (5), hepatic (2), internal iliac (2), sigmoid (2) and, in 1 case each, gonadal, omental and pancreaticoduodenal. Table 3 summarizes the patients’ clinical characteristics by AAST-OIS grade. There was a signi�cant difference in heart rate and systolic blood pressure across grade I–IV injuries. The post hoc Dunn analysis revealed this differ-ence to be signi�cant between grade I versus grade III (p < 0.001) and between grade III versus grade IV (p = 0.006) for systolic blood pressure, and between grade I versus grade IV (p = 0.02) for heart rate (Table 4).

The management strategies used in the patients with arterial injuries are summarized in Table 5. Aortic injury was associated with the highest mortality rate, 62%.

Venous injuries

Seventy-four venous injuries were identi�ed: renal (21), inferior vena cava (17), common iliac (11), external iliac (6), internal iliac (4), superior mesenteric (4), inferior mesenteric (3), gonadal (3), portal (2), pelvic (2) and hepatic (1). Of the 17 inferior vena cava injuries, 9 were infrarenal. Table 6 summarizes the patients’ outcome by AAST-OIS grade. There was no signi�cant difference in rates of death or ICU admission across AAST-OIS grade for all venous injuries.

Table 2. Demographic and clinical characteristics of patients with intra-abdominal vascular injury

CharacteristicNo. (%) of patients*

n = 110

Age at presentation, yr; mean ± SD 29 ± 10

Male sex 90 (81.8)

Affected vessel†

Artery 84 (76.4)

Vein 74 (67.3)

Wound type

Penetrating 98 (89.1)

Gunshot wound 55 (50.0)

Stab wound 43 (39.1)

Blunt 12 (10.9)

Physiologic parameters on admission

Heart rate, beats/min; mean ± SD 105 ± 22

Systolic blood pressure, mm Hg; mean ± SD 102 ± 27

Lactate level, mmol/L; mean ± SD 4.7 ± 3.6

Outcome

Admitted to ICU 57 (51.8)

Died 31 (28.2)

Recovery without ICU admission 22 (20.0)

ICU = intensive care unit; SD = standard deviation.

*Except where specified otherwise.

†In 34 patients, both types of vessel were affected.


RESEARCH


Table 5. Management of arterial injuries

Artery

Procedure; no. of injuries

Death (%)Admission to

ICU (%)Repaired Ligated PTFE graft RSVG Embolized Packed Total

Renal 0 18 0 0 3 0 21 4 (19) 11 (52)

Mesenteric 0 17 0 0 0 0 17 4 (24) 12 (70)

Aorta 7 1 0 0 0 0 8 5 (62) 3 (38)

External iliac 5 1 0 1 0 0 7 2 (29) 3 (43)

Superior mesenteric 2 4 0 0 0 0 6 3 (50) 3 (50)

Inferior mesenteric 0 6 0 0 0 0 6 2 (33) 3 (50)

Common iliac 2 1 1 1 0 0 5 0 (0) 4 (80)

Splenic 0 5 0 0 0 0 5 2 (40) 3 (60)

Hepatic artery 0 1 0 0 0 1 2 1 (50) 1 (50)

Internal iliac 0 0 1 0 1 0 2 0 (0) 0 (0)

Sigmoid 0 2 0 0 0 0 2 0 (0) 1 (50)

Gonadal 0 1 0 0 0 0 1 0 (0) 0 (0)

Omental 0 1 0 0 0 0 1 0 (0) 0 (0)

Pancreaticoduodenal 0 1 0 0 0 0 1 0 (0) 0 (0)

Total 16 59 2 2 4 1 84 — —

ICU = intensive care unit; PTFE = polytetrafluoroethylene; RSVG = reverse saphenous vein graft.

Table 4. Post hoc Dunn analysis between all AAST-OIS groups for arterial IAVIs

Variable; AAST-OIS grade

AAST-OIS grade; p value

I II III

Systolic blood pressure

II 0.2 — —

III < 0.001 0.05 —

IV 0.12 0.9 0.006

Heart rate

II 0.1 — —

III 0.1 0.4 —

IV 0.02 0.9 0.2

AAST-OIS = American Association for the Surgery of Trauma Organ Injury Scale; IAVI = intra-abdominal vascular injury.

Table 3. Clinical characteristics of patients with arterial IAVIs, by AAST-OIS grade

Characteristic

AAST-OIS grade

p value*

Combined arterial injury

n = 8Totaln = 68

In = 19

IIn = 9

IIIn = 21

IVn = 19

Heart rate, beats/min; median ± SD

98 ± 27 112 ± 22 106 ± 14 115 ± 20 0.04† 93 ± 24 68

Systolic blood pressure, mm Hg; median ± SD

118 ± 25 96 ± 28 70 ± 22 99 ± 26 < 0.001† 113 ± 31 —

Lactate level, mmol/L; median ± SD

4.7 ± 3.1 3.5 ± 3.8 3.0 ± 1.7 4.4 ± 4.2 0.1† 5.1 ± 5.2 —

No. (%) died 4 (21) 3 (33) 4 (19) 4 (21) 0.8‡ 4 (50) —

No. (%) admitted to ICU 13 (68) 6 (67) 11 (52) 11 (58) 0.7‡ 2 (25) —

Total no. of arteries 19 18 21 26 — 16 84

AAST-OIS = American Association for the Surgery of Trauma Organ Injury Scale; IAVI = intra-abdominal vascular injury; ICU = intensive care unit; SD = standard deviation.

*Statistical tests were performed excluding combined injuries so as to fit the test assumption of nonoverlapping data.

†Kruskal–Wallis test.

‡χ2 test.


RECHERCHE


The management strategies used in the patients with venous injuries are summarized in Table 7. Inferior vena cava injuries had the highest mortality rate, 47%; in most cases (13) of inferior vena cava injury, the patient under-went primary repair.

Arterial and venous injury combined

Combined venous and arterial injuries were found in 34 patients, of whom 11 (32%) died and 21 (62%) were admitted to the ICU.

Concurrent injuries

Concurrent nonvascular injuries were as follows: small bowel (63), large bowel (33), liver (33), kidney (32), stom-ach (22), pancreas (19), diaphragm (12), duodenum (10), spleen (10), bladder (7) and ureter (4) (Table 8). The in juries associated with the highest mortality rates were of the spleen (60%, p = 0.02) stomach (50%, p = 0.01) duo-denum (50%, p = 0.1) and liver (45%, p = 0.008); the lowest mortality rates were associated with ureteric (25%) and renal (25%) injuries. The 4 organ injuries most strongly suggestive of death on the χ2 test and the physiologic param-eter most associated with death (lactate level, Table 9) were

included in a multiple logistic regression analysis. Liver injury, large bowel injury, splenic injury and elevated lac-tate level were all associated with a statistically signi�cantly higher mortality rate (Table 10).

Clinical outcome

Fifty-seven patients (51.8%) required admission to the ICU. Thirty-one patients (28.2%) died; causes included renal dys-function (25 patients), respiratory (6), intra-abdominal sepsis (4), wound sepsis (4), cardiac (3) and neurologic (3). The remaining patients did not require an ICU stay and had an uneventful postoperative recovery. All 4 patients who had vascular grafts had associated enteric injury. One of these patients, who had a destructive external iliac artery injury, had a prosthetic graft to restore continuity. Graft sepsis and graft failure developed, and the patient ultimately died due to hemorrhage from a septic false aneurysm.

DISCUSSION

Numerous reports over the past half-century have reiter-ated that IAVIs are associated with a high mortality rate. Our �ndings suggest that this remains the case. Controlling exsanguination is key to achieving a reduction in death rates

Table 6. Outcome of patients with venous IAVIs, by AAST-OIS grade

Characteristic

AAST-OIS grade

p value*

Combined venous injury

n = 5Totaln = 69

In = 3

IIn = 1

IIIn = 26

IVn = 31

Vn = 3

No. (%) died 0 (0) 0 (0) 9 (35) 11 (35) 2 (67) 0.5 2 (40) —

No. (%) admitted to ICU 2 (67) 1 (100) 15 (58) 4 (14) 1 (33) 0.6 3 (60)

Total no. of veins 5 6 27 33 3 — 10 74

AAST-OIS = American Association for the Surgery of Trauma Organ Injury Scale; IAVI = intra-abdominal vascular injury; ICU = intensive care unit.

*Fisher exact test; performed excluding combined injuries so as to fit the test assumption of independent samples.

Table 7. Management of venous injuries

Vein

Procedure; no. of veins

Death (%)Admission to

ICU (%)Repaired Ligated PTFE graft Packed Total

Renal 2 19 0 0 21 4 (19) 11 (52)

Inferior vena cava.

13 4 0 0 17 8 (47) 9 (53)

Common iliac 6 4 1 0 11 3 (27) 3 (27)

External iliac 2 4 0 0 6 1 (17) 5 (83)

Internal iliac 4 0 0 0 4 3 (75) 1 (25)

Superior mesenteric

0 4 0 0 4 0 (0) 2 (50)

Inferior mesenteric

0 3 0 0 3 0 (0) 2 (67)

Gonadal 0 3 0 0 3 0 (0) 2 (67)

Portal 1 0 0 1 2 2 (100) 0 (0)

Pelvic 0 2 0 0 2 1 (50) 1 (50)

Hepatic 0 0 0 1 1 0 (0) 0 (0)

Total 28 43 1 2 74 — —

ICU = intensive care unit; PTFE = polytetrafluoroethylene.


RESEARCH


following IAVIs, and this represents an ongoing surgical challenge. The methods of surgical ligation and repair, which were pioneered in the military con�icts of the mid to late 20th century, have remained largely unchanged.

Repair remains the only feasible option in aortic injuries; aortic ligation is almost universally fatal.10 In this cohort of patients, after aortic injury, the highest mortality rate was found for the hepatic (50%), superior mesenteric (50%) and splenic (40%) arteries. All were managed by ligation apart from 1 injury to the mesenteric artery, which was repaired. Almost all cases of actively bleeding renal artery injuries were managed by nephrectomy. If there is a nonexpanding perirenal hematoma, this should be managed conservatively, as opening Gerota’s fascia generally results in uncontrolled bleeding, which can be dealt with only by nephrectomy. If

conservative management is embarked on, endovascular and endo-urological techniques can be used to salvage the situa-tion once the initial operative management is complete.

Inferior vena cava injuries remain highly lethal, as evi-denced by our mortality rate of 47%. Although primary repair was the most commonly used approach in our study, ligation is increasingly applied, especially for infrarenal caval injuries.11,12 There were 17 inferior vena cava injuries in the current series, 4 of which (3 infrarenal and 1 supra-renal) were ligated. All 4 patients had associated enteric injuries, but all survived. Among the 13 cases of inferior vena cava repair, 8 patients (62%) died, all but 1 of whom had associated enteric injury.

The management of iliac artery injuries is extremely challenging, with most cases in the current series under-going primary repair. One patient with an external iliac artery injury required a prosthetic interposition graft, and in 1 case a bleeding internal iliac artery injury was embo l-ized. Interposition grafting is used in large-calibre vessels where �ow must be preserved and has been reported in aortic, superior mesenteric artery, renal and iliac artery injuries.13–15 Although complex injuries may require an interposition graft, the presence of intra-abdominal sepsis means that these repairs are at high risk for the develop-ment of graft sepsis. The consequences of graft failure in this setting are usually catastrophic. The 2 alternatives to an interposition graft are extra-anatomic bypass and a tem-porary intravascular shunt. However, creating an extra-anatomic bypass is usually not feasible in an injured patient whose condition is unstable. Although the use of tempo-rary intravascular shunts as part of a damage-control strat-egy has been reported,16 it has not found widespread use at our institution.

Although exsanguination eclipses coagulopathy as the primary cause of death in IAVI, failure to implement ade-quate resuscitative and damage-control strategies results in dismal outcomes. This was shown by our parent institution, King Edward VIII Hospital, in Durban, where the mortal-ity rate for caval injuries increased from 35.7% to 88% over a 15-year period 2 decades ago. The increase was attributed to the lack of implementation of damage-control tech-niques in response to a massive increase in devastating gun-shot wounds during a period of great political instability.17 Since that period, damage-control approaches to IAVIs have gained widespread acceptance in our environment. Our current mortality rate is in keeping with the interna-tional and national literature, no doubt largely due to adop-tion of damage-control approaches.6

It is unlikely that we are going to achieve further reduc-tions in mortality rates using exclusively open operative tech-niques. The most recent military reports suggest that the next evolution in the management of IAVIs will likely be endovascular-based modalities.18 The advent of these tech-niques has expanded the scope for nonoperative approaches to IAVIs. Endovascular approaches can be used to arrest

Table 10. Multiple logistic regression analysis of factors associated with death

Variable OR (95% CI)

Elevated lactate level on admission 1.64 (1.34–2.12)

Large bowel injury 8.00 (2.32–32.54)

Liver injury 7.56 (1.98–36.06)

Spleen injury 12.04 (1.92–87.80)

Stomach injury 1.91 (0.51–7.10)

CI = confidence interval; OR = odds ratio.

Table 9. χ2 analysis of physiologic parameters as predictors of death

Parameter

Outcome; median ± SD

p valueDiedn = 31

Survivedn = 79

Lactate level, mmol/L 7.5 ± 4.2 3.5 ± 2.6 < 0.001

Systolic blood pressure, mm Hg

89 ± 28 106 ± 25 0.002

Heart rate, beats/min 107 ± 23 104 ± 22 0.4

SD = standard deviation.

Table 8. Distribution and outcome of nonvascular injuries

InjuryNo. of

patients

Rate of admission to ICU, % p value*

Death rate, % p value*

Small bowel 63 58 0.01 30 0.3

Large bowel 33 70 0.01 45 0.008

Liver 33 52 1.0 45 0.008

Kidney 32 60 0.4 25 0.6

Stomach 22 59 0.4 50 0.01

Pancreas 19 53 0.9 47 0.04

Diaphragm 12 58 0.6 33 0.7

Duodenum 10 70 0.2 50 0.1

Spleen 10 60 0.6 60 0.02

Bladder 7 43 0.6 29 1.0

Ureter 4 25 0.3 25 0.9

ICU = intensive care unit.

*χ2 test.


RECHERCHE


hemorrhage through balloon occlusion or embolization, or to repair an injured vessel with an endovascular graft. Embol ization is an endovascular option for injuries to small and medium-sized nonessential vessels.19 In our series, 4 patients underwent embolization (3 of the renal artery and 1 of the inferior internal iliac artery); all survived. Emboliza-tion of the internal iliac artery or its branches is generally well tolerated.20 Although not performed in our analysis, suc-cessful embolization of superior mesenteric artery injuries has been reported.21 The superior mesenteric artery territory is a particularly attractive area for endovascular intervention as it is dif�cult to surgically access this area. The development of hybrid operating rooms such as the so-called RAPTOR suite (resuscitation with angiography, percutaneous techniques and operative repair) may allow for more seamless integra-tion of open and endovascular approaches in the manage-ment of IAVIs. Currently, there is great interest in resuscita-tive endovascular balloon occlusion of the aorta.22 This innovation was prompted by the recent military experience, and reports on its use are limited.23 This technique could be instrumental in facilitating proximal control while allowing endovascular intervention in vessels that are dif�cult to access without recourse to massive retroperitoneal dissections. The exact place of these techniques is yet to be de�ned.

Limitations

This was a single-centre experience. For a more thorough analysis, a larger study across multiple centres is required.

CONCLUSION

Despite the standardization of operative approaches and the implementation of damage-control surgery and resus-citation over the last 50 years, the mortality rate for IAVI remains high. Exsanguination remains the most common cause of death. It is hoped that the ongoing development of endovascular techniques and approaches in the manage-ment of these injuries may improve outcomes in the future.Af�liations: From the Department of General Surgery, Wessex Deanery, Wessex, United Kingdom (Weale); Pietermaritzburg Metropolitan Trauma Service, Department of Surgery, University of KwaZulu-Natal, Durban, South Africa (Kong, Manchev, Bekker, Oosthuizen, Laing, Bruce, Clarke); the School of Nursing and Public Health, University of KwaZulu-Natal, Durban, South Africa (Brysiewicz); and the Department of Surgery, University of the Witwatersrand, Johannesburg, South Africa (Clarke).


Contributors: All authors designed the study. R. Weale, V. Kong, V. Manchev, G Laing and J. Bruce acquired the data, which R. Weale, V. Kong, G. Oosthuizen, G. Laing and D. Clarke analyzed. R. Weale and D. Clarke wrote the article, which all authors reviewed and approved for publication.

References

1. Ksycki M, Ruiz G, Perez-Alonso AJ, et al. Iliac vessel injuries: dif�cult injuries and dif�cult management problems. Eur J Trauma Emerg Surg 2012;38:347-57.

2. Buscaglia LC, Blaisdell FW, Lim RC. Penetrating abdominal vascular injuries. Arch Surg 1969;99:764-9.

3. Mattox KL, McCollum WB, Jordan GL, et al. Management of upper abdominal vascular trauma. Am J Surg 1974;128:823-8.

4. Eachempati SR, Robb T, Ivatury RR, et al. Factors associated with mortality in patients with penetrating abdominal vascular trauma. J Surg Res 2002;108:222-6.

5. Kashuk JL, Moore EE, Johnson JL, et al. Postinjury life threatening coagulopathy: Is 1:1 fresh frozen plasma:packed red blood cells the answer? J Trauma 2008;65:261-70, discussion 270-1.

6. Sorrentino TA, Moore EE, Wohlauer MV, et al. Effect of damage control surgery on major abdominal vascular trauma. J Surg Res 2012; 177:320-5.

7. Morrison JJ, Rasmussen TE. Noncompressible torso hemorrhage: a review with contemporary de�nitions and management strategies. Surg Clin North Am 2012;92:843-58.

8. Stannard A, Morrison JJ, Scott DJ, et al. The epidemiology of non-compressible torso hemorrhage in the wars in Iraq and Afghanistan. J Trauma Acute Care Surg 2013;74:830-4.

9. Moore EE, Cogbill TH, Jurkovich GJ, et al. Organ injury scaling III: chest wall, abdominal vascular, ureter, bladder and urethra. J Trauma 1992;33:337-9.

10. Asensio JA, Chahwan S, Hanpeter D, et al. Operative management and outcome of 302 abdominal vascular injuries. Am J Surg 2000; 180:528-33, discussion 533-4.

11. Van Rooyen PL, Karusseit VO, Mokoena K. Inferior vena cava injuries: a case series and review of the South African experience. Injury 2015;46:71-5.

12. Sullivan PS, Dente CJ, Patel S, et al. Outcome of ligation of the inferior vena cava in the modern era. Am J Surg 2010;199:500-6.

13. Morrison JJ, Stannard A, Rasmussen TE, et al. Injury pattern and mortality of noncompressible torso hemorrhage in UK combat casu-alties. J Trauma Acute Care Surg 2013;75:S263-8.

14. Murthy R, Hastings G, Richard H, et al. Angiography and endovas-cular intervention for vascular trauma to the extremities. Semin Interv Radiol 2003;20:81-8.

15. Velmahos GC, Demetriades D, Chahwan S, et al. Angiographic embolization for arrest of bleeding after penetrating trauma to the abdomen. Am J Surg 1999;178:367-73.

16. Ding W, Wu X, Li J. Temporary intravascular shunts used as a damage control surgery adjunct in complex vascular injury: collective review. Injury 2008;39:970-7.

17. Clarke DL, Madiba TE, Muckart DJ. Inferior vena caval injury in the �rearm era. S Afr J Surg 1999;37:107-9.

18. Whigham CJ Jr, Bodenhamer JR, Miller JK. Use of the Palmaz stent in primary treatment of renal artery intimal injury secondary to blunt trauma. J Vasc Interv Radiol 1995;6:175-8.

19. Villas PA, Cohen G, Putnam SG 3rd, et al. Wall stent placement in a renal artery after blunt abdominal trauma. J Trauma 1999; 46:1137-9.

20. Lyden SP, Srivastava SD, Waldman DL, et al. Common iliac artery dissection after blunt trauma: case report of endovascular repair and literature review. J Trauma 2001;50:339-42.

21. Kakizawa H, Toyota N, Hieda M, et al. Traumatic mesenteric bleeding managed solely with transcatheter embolization. Radiat Med 2007; 25:295-8.

22. Sadeghi M, Nilsson KF, Larzon T, et al. The use of aortic balloon occlusion in traumatic shock: �rst report from the ABO trauma registry. Eur J Trauma Emerg Surg 2017;Aug 11 [Epub ahead of print]. doi:10.1007/s00068-017-0813-7.

23. Smith SA, Hilsden R, Beckett A, et al. The future of resuscitative endovascular balloon occlusion in combat operations. J R Army Med Corps 2017;Aug 9 [Epub ahead of print]. doi:10.1136/jramc-2017-000788.




Risk factors for infection, revision, death, blood transfusion and longer hospital stay 3 months and 1 year after primary total hip or knee arthroplasty

Background: Total joint replacement (TJR) is increasingly performed in older patients with more comorbidities, who are considered at higher risk for postoperative complications. We aimed to identify and calculate the odds ratio of the risk factors for infection, revision and death 3 months and 1 year after TJR as well as for postopera-tive blood transfusion and longer hospital stay.

Methods: We analyzed all primary total hip arthroplasty (THA) and total knee arthroplasty (TKA) cases in Nova Scotia between Apr. 1, 2000, and Mar. 31, 2014, as identi�ed from the Discharge Abstract Database. We used the Charlson Comorbidity Index as a surrogate measure of comorbidities. We used hospital and physician bill-ings data and Nova Scotia Vital Statistics data to identify the postoperative events in this cohort.

Results: A total of 10 123 primary THA and 17 243 primary TKA procedures were performed during the study period. The mean patient age was 66.1 (standard devia-tion 11.7) years and 67.1 (standard deviation 9.3) years, respectively. With THA, the risk of infection was higher in patients with heart failure and those with diabetes. For TKA, liver disease and blood transfusion were associated with a higher risk of infec-tion. Revision rates were higher among patients with hypertension and those with paraparesis/hemiparesis for THA, and among patients with metastatic disease for TKA. Signi�cant risk factors for death included metastatic disease, older age, heart failure, myocardial infarction, dementia, rheumatologic disease, renal disease, blood transfusion and cancer. Multiple medical comorbidities and older age were associated with higher rates of blood transfusion and longer hospital stay.

Conclusion: We have identi�ed the risk factors associated with higher rates of post-operative complications and longer hospital stay after TJR. The results enable indi-vidualized risk strati�cation during the preoperative consultation.

Contexte : Les arthroplasties totales (AT) sont de plus en plus pratiquées chez les patients âgés présentant de plus nombreuses comorbidités et considérés de ce fait exposés à un risque accru de complications postopératoires. Nous avons voulu déter-miner et calculer le rapport des cotes pour les facteurs de risque d’infection, de révi-sion chirurgicale et de décès 3 mois et 1 an après l’AT, de même que de transfusions sanguines postopératoires et de prolongation du séjour hospitalier.

Méthodes : Nous avons analysé toutes les interventions primaires pour prothèse totale de la hanche (PTH) et prothèse totale du genou (PTG) en Nouvelle-Écosse entre le 1er avril 2000 et le 31 mars 2014, répertoriées dans la base de données sur les congés des patients. Nous avons utilisé le score de comorbidité de Charlson comme marqueur de substitution des comorbidités. Nous avons utilisé les données de factura-tion des hôpitaux et des médecins et les données de l’état civil de la Nouvelle-Écosse pour recenser les événements postopératoires dans cette cohorte.

Résultats : En tout, 10 123 PTH primaires et 17 243 PTG primaires ont été effec-tuées pendant la période de l’étude. L’âge moyen des patients était de 66,1 ans (écart-type 11,7) et de 67,1 ans (écart-type 9,3), respectivement. Avec la PTH, le risque d’infection a été plus élevé chez les patients atteints d’insuf�sances cardiaques et les patients diabétiques, tandis qu’avec la PTG, il a été plus élevé chez les patients atteints de maladie hépatique et traités par transfusions sanguines. Les taux de révision chirur-gicale ont été plus élevés chez les patients hypertendus et ceux qui souffraient de

Chanseok Rhee, MD Lynn Lethbridge, MA Glen Richardson, MD Michael Dunbar, MD, PhD

Accepted Oct. 20, 2017

Correspondence: C. Rhee Halifax Infirmary Site Capital Health 1799 Robie St Halifax NS B3H 3A7 [email protected]

DOI: 10.1503/cjs.007117

risk-rhee.indd 165 2018-05-18 12:18 PM

RECHERCHE


G iven the aging population, the incidence and prev-alence of osteoarthritis is expected to increase, and consequently the demand for total joint replace-

ment (TJR) continues to rise.1 In addition, more TJR pro-cedures are being performed in older patients and those with multiple medical comorbidities.2,3 Such patients are inevitably at a higher risk for serious postoperative compli-cations and may require more rigorous optimization of their medical condition preoperatively and more in-depth discussion when obtaining informed consent.4,5

Many researchers attempting to identify risk factors for complications have focused on a singular adverse event, and only a few investigators have considered multiple com-plications within the same cohort.6–10 In addition, it is often dif�cult to generalize the results of a study to different population groups owing to region-speci�c factors such as ethnic composition and socioeconomic differences among the populations.11,12

The purpose of this study was to determine the risk fac-tors for postoperative infection, revision and death 3 months and 1 year after primary total hip arthroplasty (THA) and total knee arthroplasty (TKA) through the use of large administrative databases. In addition, we sought to identify the risk factors for postoperative blood transfusion and longer hospital stay.

METHODS

The study population included all patients who underwent primary hip or knee arthroplasty in Nova Scotia between Apr. 1, 2000, and Mar. 31, 2014, as identi�ed from the Canadian Institute for Health Information Hospital Dis-charge Abstract Database. Follow-up data were available until the end of 2014/15, which allowed calculation of complication rates for up to 1 full year for all patients. We used Canadian Classi�cation of Health Intervention codes to select the arthroplasty procedures included in the study (Appendix 1, available at canjsurg.ca/007117-a1). Revision procedures were excluded, as were cases identi�ed as emer-gency cases, which largely include arthroplasty surgery sec-ondary to injury.

We linked data on use of health care services to the study cohort to identify patient characteristics, risks and outcomes. Hospital and physician billings data for the 5 years before and after the index procedure were incorpo-

rated to capture information on use from hospitals as well as outpatient observations, which included of�ce visits, home visits and visits to long-term care facilities by phys-icians. We used Nova Scotia Vital Statistics data to identify deaths that occurred in or out of hospital within the prov-ince. To account for socioeconomic factors, we included median neighbourhood income. Statistics Canada pro-duces summary statistics for various geographic levels, the smallest being a dissemination area. We converted patients’ residential postal code to dissemination area to link neighbourhood income to the database.

We focused on patient risk factors including age, sex and comorbidities from the Charlson Comorbidity Index, which were included individually rather than as aggregate score. We also analyzed blood transfusion as a risk factor, as it has been shown to be associated with complications after TJR.13,14 Using International Classi�cation of Dis-eases codings, we employed algorithms developed for the Canadian Chronic Disease Surveillance System to specify the presence of individual health conditions for which both hospital and physician billing data sources are used. Neigh-bourhood income, which has been included as a control in previous studies,15,16 was included as a measure of socioeco-nomic status. We included dummy variables for each health care facility in all models to control for institutional-level factors that may affect outcomes.

The study received ethics approval from the Nova Sco-tia Health Authority.

Statistical analysis

We used multivariate regression analysis to examine indi-vidual associations between risk factors and patient out-comes. We modelled binary outcomes using the logistic procedure. Associations are presented as odds ratios (ORs). Centre A, which is a tertiary care teaching hospital, served as a reference for calculation of ORs compared to 4 regional hospitals within the province. Similarly, the age group of 59 years or less and male sex were used as refer-ences in comparison to other age groups and female sex, respectively. We regressed length of stay against covariates using a log-linear specification. We employed the log transformation to account for outliers (i.e., very long length of stay) to reduce skewness in the data and improve the interpretability of the results.

paraparésie ou d’hémiparésie dans les cas de PTH, et chez les patients atteints de maladies métastatiques dans les cas de PTG. Les facteurs de risque de décès signi�ca-tifs incluaient maladie métastatique, âge avancé, insuf�sance cardiaque, infarctus du myocarde, démence, maladie rhumatismale, maladie rénale, transfusions sanguines et cancer. La présence de comorbidités multiples et l’âge avancé ont été associés à des taux plus élevés de transfusions sanguines et à des séjours hospitaliers plus longs.

Conclusion : Nous avons déterminé les facteurs de risque associés aux taux plus élevés de complications postopératoires et aux séjours hospitaliers prolongés après une AT. Les résultats permettent d’établir une strati�cation individualisée des risques dès la consultation préopératoire.

risk-rhee.indd 166 2018-05-18 12:18 PM

RESEARCH


We calculated diagnostic statistics that measure both calibration (how well the models predict actual out-comes) and discrimination (how well the models distin-guish between cases and noncases) to examine the over-all performance of the models. Since calibration and discrimination measures do not necessarily move in the same direction,17,18 we included estimates for both aspects, namely, the adjusted R-squared (max-rescaled R-squared), the Brier score (which includes information on both calibration and discrimination) and the c-statistic (the area under the receiver operating charac-teristic curve for binary outcomes). The R-squared adjusted for the number of covariates is reported for the length of stay model. We used SAS 9.4 (SAS Institute) for all analysis.

RESULTS

A total of 27 366 patients who underwent primary joint replacement during the study period were included in the �nal statistical analysis. Of the 27 366, 10 123 (mean age 66.1 [standard deviation 11.7] yr) underwent pri-mary THA, and 17 243 (mean age 67.1 [standard devia-tion 9.3 yr) underwent primary TKA. The patients’ demographic characteristics and the overall incidence of major postoperative complications are summarized in Table 1.

Infection

The overall infection rate was 0.9% at 3 months and 1.2% at 1 year following primary THA. Infection rates at 3 months and 1 year increased over the study period (p = 0.005 and p < 0.001, respectively) (Fig. 1 and Fig. 2). Con-gestive heart failure was associated with increased risk of infection at both 3 months (OR 6.1, 95% CI 2.8 to 13.5) and 1 year (OR 4.8, 95% CI 2.2 to 10.2) (Table 2 and Table 3). Diabetes with complications (de�ned as diabetes with associated end-organ damage) was a signi�cant risk factor for infection at 3 months (OR 2.0, 95% CI 1.2 to 3.3). Patients with osteoporosis had a decreased infection rate at 1 year (OR 0.4, 95% CI 0.2 to 0.9).

The overall infection rate after TKA was 1.1% at 3 months and 1.6% at 1 year (Fig. 3 and Fig. 4). Mild liver disease was a statistically signi�cant risk factor for prosthetic knee joint infection at both 3 months (OR 3.9, 95% CI 1.6 to 9.6) and 1 year (OR 4.0, 95% CI 1.8 to 8.5) (Table 4 and Table 5). Blood transfusion was shown to be a statistically signi�cant risk factor for infection at 1 year (OR 1.6, 95% CI 1.1 to 2.4). In addition, patients aged 70–79 years (OR 0.6, 95% CI 0.4 to 0.8), those aged 80 or more (OR 0.6, 95% CI 0.3 to 0.9) and women (OR 0.7, 95% CI 0.6 to 0.9) were less likely to have prosthetic joint infection at 1 year compared to their respective ref-erence group.

Revision

The overall revision rate after THA was 0.9% at 3 months and 1.6% at 1 year. Revision rates at 3 months (p = 0.001) and 1 year (p = 0.02) increased over the study period (Fig. 1 and Fig. 2). Patients with paraparesis/hemiparesis had an increased revision rate at 3 months (OR 7.1, 95% CI 1.5 to 32.7) (Table 2).

The revision rate after TKA was 0.5% at 3 months and 1.4% at 1 year (Fig. 3 and Fig. 4). Patients with metastatic disease had a significantly increased revision rate at 3 months (OR 5.5, 95% CI 1.4 to 21.6) (Table 4). In addi-tion, patients aged 70–79 (OR 0.6, 95% CI 0.4 to 0.8), those aged 80 or more (OR 0.6, 95% CI 0.3 to 0.9) and women (OR 0.7, 95% CI 0.6 to 0.9) were less likely to undergo revision surgery at 1 year compared to their respective reference group (Table 5).

Death

The overall mortality rate after THA was 0.4% at 3 months and 1.2% at 1 year (Fig. 1 and Fig. 2). At both 3 months and 1 year, congestive heart failure (OR 4.1, 95% CI 1.7 to 10.0, and OR 3.8, 95% CI 2.1 to 6.8, respectively), metastatic disease (OR 34.4, 95% CI 9.4 to

Table 1. Demographic characteristics and overall complication rates for patients who underwent primary hip or knee arthroplasty in Nova Scotia, Apr. 1, 2000, to Mar. 31, 2014

VariablePrimary THAn = 10 123

Primary TKAn = 17 243

Age, mean ± SD; yr 66.1 ± 11.7 67.1 ± 9.3

Age group, yr; no. (%) of patients

< 60 2722 (26.9 ) 3697 (21.4)

60–69 3208 (31.7) 6455 (37.4)

70–79 2987 (29.5) 5487 (31.8)

≥ 80 1206 (11.9) 1604 (9.3)

Sex, no. (%) of patients

Male 4487 (44.3) 6940 (40.2)

Female 5636 (55.7) 10 303 (59.8)

Neighbourhood household income, median ± SD; $

57 290 ± 22 410 56 240 ± 21 890

Length of stay, mean ± SD; d

5.7 ± 5.6 5.1 ± 4.7

Transfusion rate, % 14.1 8.1

Complication rate, %

3 mo

Infection 0.9 1.1

Revision 0.9 0.5

Death 0.4 0.3

1 yr

Infection 1.2 1.6

Revision 1.6 1.4

Death 1.2 0.9

SD = standard deviation; THA = total hip arthroplasty; TKA = total knee arthroplasty.

risk-rhee.indd 167 2018-05-18 12:18 PM

RECHERCHE


126.1, and OR 15.3, 95% CI 7.2 to 32.7, respectively), blood transfusion (OR 2.6, 95% CI 1.3 to 5.3, and OR 2.1, 95% CI 1.3 to 3.2, respectively) and age 80 or more (OR 9.3, 95% CI 2.0 to 43.3, and OR 3.0, 95% CI 1.6 to 5.6, respectively) were associated with increased mortality rates

(Table 2 and Table 3). Patients with previous myocardial infarction (OR 7.9, 95% CI 2.6 to 24.3) and those with dementia (OR 4.4, 95% CI 1.1 to 17.5) were at increased risk for death at 3 months. Diabetes with complications (OR 1.7, 95% CI 1.1 to 2.9) was a statistically signi�cant

Fig. 2. Complication rates at 1 year after total hip arthroplasty in Nova Scotia, Apr. 1, 2000, to Mar. 31, 2014.

2.5

2.0

1.5

1.0

0.5

0

2000/01

Infection Revision Death

2001/02 2003/04 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/152004/05 2005/06 2006/07

% o

f pa

tient

s

Year

Fig. 1. Complication rates at 3 months after total hip arthroplasty in Nova Scotia, Apr. 1, 2000, to Mar. 31, 2014.

2.5

2.0

1.5

1.0

0.5

0

2000/01


2001/02 2003/04 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/152004/05 2005/06 2006/07

% o

f pa

tient

s

Year

risk-rhee.indd 168 2018-05-18 12:18 PM

RESEARCH


risk factor for death at 1 year. Women showed lower mor-tality rates than men at 3 months (OR 0.4, 95% CI 0.2 to 0.8) and 1 year (OR 0.7, 95% CI 0.4 to 1.0).

The mortality rate after TKA was 0.3% at 3 months and 0.9% at 1 year (Fig. 3 and Fig. 4). Identi�ed risk fac-tors for an increased mortality rate at 3 months and 1 year included previous myocardial infarction (OR 22.5, 95% CI 6.9 to 73.9, and OR 6.8, 95% CI 3.3 to 14.0,

respectively), congestive heart failure (OR 5.3, 95% CI 2.3 to 11.9, and OR 2.9, 95% CI 1.7 to 5.1, respectively), cerebrovascular accident (OR 4.0, 95% CI 1.6 to 9.8, and OR 2.2, 95% CI 1.2 to 4.2, respectively), rheumato-logic disease (OR 2.7, 95% CI 1.2 to 6.2, and OR 2.0, 95% CI 1.1 to 3.4, respectively), renal disease (OR 3.0, 95% CI 1.0 to 8.4, and OR 2.5, 95% CI 1.3 to 4.8, respectively) and age 80 or more (OR 10.6, 95% CI 2.3

Table 2. Association of risk factors with hip infection, revision and death 3 months after total hip arthroplasty

Risk factor

OR (95% CI)


Year of surgery 1.1 (1.0 to 1.2) 1.1 (1.0 to 1.2) 1.0 (0.9 to 1.1)

Blood transfusion (n = 1425 [14.1%])

1.1 (0.6 to 2.1) 0.9 (0.5 to 1.8) 2.6 (1.3 to 5.3)

Diabetes with complications (n = 1504 [14.8%])

2.0 (1.2 to 3.3) 1.1( 0.6 to 2.1) 1.5 (0.6 to 3.5)

Diabetes without complica-tions (n = 341 [3.4%])

0.3 (0.1 to 1.1) 1.0 (0.3 to 2.7) 0.5 (0.1 to 2.0)

Hypertension (n = 4766 [47.0%])

0.8 (0.5 to 1.3) 1.2 (0.8 to 2.0) 1.3 (0.7 to 2.6)

Myocardial infarction (n = 105 [1.0%])

0.8 (0.1 to 6.2) 1.8 (0.4 to 8.1) 7.9 (2.6 to 24.3)

Congestive heart failure (n = 246 [2.4%])

6.1 (2.8 to 13.5) 1.3 (0.5 to 3.7) 4.1 (1.7 to 10.0)

Ischemic heart disease (n = 1230 [12.1%])

0.8 (0.4 to 1.6) 1.2 (0.7 to 2.3) 1.2 (0.5 to 2.9)

Osteoporosis (n = 1139 [11.3%])

0.4 (0.2 to 1.0) 1.0 (0.5 to 1.9) 0.9 (0.3 to 2.4)

Peripheral vascular disease (n = 211 [2.1%])

1.9 (0.6 to 5.5) 1.3 (0.5 to 3.9) 0*

Cerebrovascular accident (n = 232 [2.3%])

0.9 (0.2 to 3.6) 1.0 (0.3 to 3.2) 2.3 (0.7 to 8.1)

Dementia (n = 84 [0.8%]) 0.7 (0.1 to 5.8) 1.4 (0.3 to 6.8) 4.4 (1.1 to 17.5)

Chronic obstructive pulmonary disease (n = 1121 [11.0%])

1.1 (0.6 to 2.0) 1.5 (0.9 to 2.7) 1.8 (0.8 to 3.7)

Rheumatologic disease (n = 507 [5.0%])

1.4 (0.6 to 3.3) 1.9 (0.9 to 4.0) 1.5 (0.4 to 5.3)

Peptic ulcer (n = 113 [1.1%]) 1.2 (0.2 to 8.6) 0* 1.2 (0.2 to 9.8)

Mild liver disease (n = 81 [0.8%])

1.3 (0.2 to 9.8) 1.4 (0.2 to 10.8) 2.7 (0.3 to 27.1)

Paraparesis/hemiparesis (n = 32 [0.3%])

2.6 (0.2 to 28.5) 7.1 (1.5 to 32.7) 3.4 (0.3 to 40.6)

Renal disease (n = 217 [2.1%]) 2.4 (0.9 to 6.2) 0.9 (0.3 to 3.0) 1.4 (0.4 to 5.2)

Cancer (n = 828 [8.2%]) 0.7 (0.3 to 1.7) 0.9 (0.4 to 2.0) 0.9 (0.3 to 2.5)

Moderate to severe liver disease (n = 12 [0.1%])

0* 0* 0*

Metastatic disease (n = 95 [0.9%])

1.2 (0.1 to 11.5) 0.8 (0.1 to 7.3) 34.4 (9.4 to 126.1)

Median household income 1.0 (1.0 to 1.0) 1.0 (1.0 to 1.0) 1.0 (1.0 to 1.0)

Age 60–69 yr (n = 3208) 1.5 (0.8 to 2.8) 1.1 (0.6 to 2.2) 4.0 (0.9 to 18.2)

Age 70–79 yr (n = 2987) 1.4 (0.7 to 2.7) 1.5 (0.8 to 2.9) 3.7 (0.8 to 17.1)

Age ≥ 80 yr (n = 1206) 1.7 (0.8 to 3.6) 1.7 (0.8 to 3.6) 9.3 (2.0 to 43.3)

Female sex (n = 5639) 1.2 (0.8 to 1.9) 0.9 (0.6 to 1.3) 0.4 (0.2 to 0.8)

Centre B (n = 1116) 1.4 (0.7 to 2.8) 1.9 (0.9 to 3.7) 2.4 (0.9 to 5.9)

Centre C (n = 1750) 2.1 (1.2 to 3.6) 2.6 (1.4 to 4.5) 0.9 (0.3 to 2.4)

Centre D (n = 892) 0.9 (0.4 to 2.3) 1.2 (0.5 to 3.1) 2.5 (0.8 to 7.6)

Centre E (n = 2019) 1.1 (0.5 to 2.0) 1.5 (0.8 to 2.8) 1.5 (0.6 to 3.7)


*The number of incidents was insufficient to calculate the odds ratio of corresponding complications.

risk-rhee.indd 169 2018-05-18 12:18 PM

RECHERCHE


to 48.6, and OR 6.6, 95% CI 3.4 to 12.9, respectively) (Table 4 and Table 5).

Blood transfusion

A total of 14.1% and 8.1% of patients received a blood transfusion during their hospital stay for primary THA and TKA, respectively. The overall transfusion rates decreased over the study period for both THA and TKA (p < 0.001 for both). Risk factors for blood transfusion for both THA and TKA included diabetes without complications (OR 1.6, 95% CI 1.1 to 2.2, and OR 1.6, 95% CI 1.2 to 2.1, respec-tively), previous myocardial infarction (OR 2.6, 95% CI 1.6 to 4.1, and OR 6.9, 95% CI 4.6 to 10.4, respectively), con-gestive heart failure (OR 1.7, 95% CI 1.2 to 2.3, and OR 2.4, 95% CI 1.8 to 3.1, respectively), osteoporosis (OR 1.5, 95% CI 1.3 to 1.8, and OR 1.4, 95% CI 1.2 to 1.7, respec-tively), rheumatologic disease (OR 1.6, 95% CI 1.2 to 2.0, and OR 1.5, 95% CI 1.2 to 1.9, respectively), peptic ulcer

(OR 1.7, 95% CI 1.1 to 2.8, and OR 1.7, 95% CI 1.2 to 2.4, respectively), renal disease (OR 2.0, 95% CI 1.4 to 2.8, and OR 2.0, 95% CI 1.4 to 2.7, respectively), cancer (OR 1.3, 95% CI 1.1 to 1.6, and OR 1.2, 95% CI 1.0 to 1.5, respectively), age 70–79 (OR 1.7, 95% CI 1.4 to 2.0, and OR 3.0, 95% CI 2.4 to 3.6, respectively), age 80 or more (OR 3.2, 95% CI 2.6 to 4.0, and OR 6.1, 95% CI 4.9 to 7.6, respectively) and female sex (OR 2.4, 95% CI 2.1 to 2.8, and OR 2.3, 95% CI 2.0 to 2.6, respectively) (Table 6).

Length of hospital stay

The average length of hospital stay was 5.7 days and 5.1 days for primary THA and TKA, respectively. The average length of stay decreased over the study period for both procedures (p < 0.001 for both). Multiple medical comorbidities, blood transfusion, more advanced age and female sex were all risk factors for longer hospital stay after TJR (Table 7).

Table 3. Association of risk factors with hip infection, revision and death 1 year after total hip arthroplasty

Risk factor

OR (95% CI)



Blood transfusion 1.2 (0.7 to 2.0) 0.9 (0.6 to 1.4) 2.1 (1.3 to 3.2)

Diabetes with complications 1.5 (1.0 to 2.4) 0.7 (0.4 to 1.2) 1.7 (1.1 to 2.9)

Diabetes without complications 0.3 (0.1 to 1.0) 1.0 (0.4 to 2.4) 0.9 (0.4 to 1.9)

Hypertension 0.8 (0.6 to 1.2) 1.4 (1.0 to 2.0) 0.9 (0.6 to 1.4)

Myocardial infarction 0.7 (0.1 to 5.7) 1.5 (0.4 to 5.1) 2.1 (0.9 to 5.2)

Congestive heart failure 4.8 (2.2 to 10.2) 0.8 (0.3 to 2.2) 3.8 (2.1 to 6.8)

Ischemic heart disease 0.6 (0.3 to 1.2) 1.2 (0.8 to 1.9) 1.4 (0.9 to 2.4)

Osteoporosis 0.4 (0.2 to 0.9) 0.9 (0.5 to 1.5) 0.9 (0.5 to 1.6)

Peripheral vascular disease 1.9 (0.7 to 4.9) 1.4 (0.6 to 3.1) 0.6 (0.2 to 2.0)

Cerebrovascular accident 1.1 (0.3 to 3.3) 1.3 (0.6 to 3.0) 1.3 (0.5 to 3.3)

Dementia 0.6 (0.1 to 4.7) 1.0 (0.2 to 4.6) 2.6 (0.9 to 7.5)

Chronic obstructive pulmonary disease

1.1 (0.6 to 1.8) 1.3 (0.8 to 2.0) 1.3 (0.8 to 2.1)

Rheumatologic disease 1.6 (0.8 to 3.1) 1.6 (0.9 to 2.9) 2.0 (1.1 to 3.9)

Peptic ulcer 0.8 (0.1 to 5.9) 0* 0.9 (0.2 to 3.9)

Mild liver disease 0.9 (0.1 to 6.8) 2.3 (0.7 to 7.6) 1.4 (0.3 to 7.2)

Paraparesis/hemiparesis 1.7 (0.2 to 17.3) 3.5 (0.9 to 14.1) 0.5 (0.0 to 7.4)

Renal disease 2.1 (0.9 to 4.9) 1.2 (0.5 to 2.8) 1.8 (0.8 to 4.0)

Cancer 0.8 (0.4 to 1.7) 1.1 (0.6 to 1.9) 1.7 (1.0 to 2.9)

Moderate to severe liver disease 0* 0* 0*

Metastatic disease 1.7 (0.4 to 8.5) 1.5 (0.4 to 5.3) 15.3 (7.2 to 32.7)


Age 60–69 yr 1.7 (1.0 to 2.8) 1.2 (0.8 to 1.8) 0.9 (0.5 to 1.8)

Age 70–79 yr 1.4 (0.8 to 2.4) 1.3 (0.9 to 2.1) 1.4 (0.8 to 2.5)

Age ≥ 80 yr 1.7 (0.8 to 3.2) 1.2 (0.7 to 2.2) 3.0 (1.6 to 5.6)

Female sex 1.0 (0.7 to 1.4) 0.8 (0.6 to 1.1) 0.7 (0.4 to 1.0)

Centre B 1.3 (0.7 to 2.3) 1.3 (0.7 to 2.1) 1.1 (0.6 to 1.9)

Centre C 2.3 (1.4 to 3.7) 2.3 (1.5 to 3.4) 0.8 (0.4 to 1.3)

Centre D 0.8 (0.4 to 1.8) 0.9 (0.4 to 1.8) 1.0 (0.5 to 2.1)

Centre E 1.1 (0.6 to 1.9) 1.3 (0.8 to 2.1) 0.9 (0.5 to 1.5)



risk-rhee.indd 170 2018-05-18 12:18 PM

RESEARCH


Model diagnostics

Table 8 shows the summary diagnostic statistics for each model. Across outcomes, the mortality models per-formed the best in terms of both calibration and discrim-

ination measures. Most of the c-statistic values were close to or greater than 0.80 for each of the mortality outcomes, which suggests that each is a strong model in terms of discrimination. Lix and colleagues19 sug-gested that a Brier score less than 0.25 is an acceptable

Fig. 3. Complication rates at 3 months after total knee arthroplasty in Nova Scotia, Apr. 1, 2000, to Mar. 31, 2014.

2.5

2.0

1.5

1.0

0.5

0

2000/01


2001/02 2003/04 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/152004/05 2005/06 2006/07

% o

f pa

tient

s

Year

Fig. 4. Complication rates at 1 year after total knee arthroplasty in Nova Scotia, Apr. 1, 2000, to Mar. 31, 2014.

2000/01


2001/02 2003/04 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/152004/05 2005/06 2006/07

% o

f pa

tient

s

Year

3.0

2.5

2.0

1.5

1.0

0.5

0

risk-rhee.indd 171 2018-05-18 12:18 PM

RECHERCHE


prediction error where the smaller the value, the less the error. The Brier score for all models was well below this threshold. The 3-month models generally performed better than the 1-year models for all 3 measures. The R-squared value was highest for length of stay, which is not surprising given the large number of statistically sig-ni�cant covariates.

DISCUSSION

The aim of the current study was to identify the risk factors for major complications after THA and TKA.

We found 1-year infection rates of 1.2% and 1.6% after THA and TKA, respectively, which is in accor-dance with the literature.20–23 Patients younder than 60 years of age had a signi�cantly higher infection rate than did those aged 70 years or older after TKA but not after THA. This outcome was shown in some pre-vious studies24,25 but not in others.22,26,27 Mild liver dis-ease was also a signi�cant risk factor for infection at 3 months and 1 year after TKA in our study. This is in keeping with the literature: the presence of hepatitis and the markers of active hepatitis such as thrombo-cytopenia and liver �brosis have been reported to be

Table 4. Association of risk factors with knee infection, revision and death 3 months after total knee arthroplasty

Risk factor

OR (95% CI)



Blood transfusion (n = 1402 [8.1%]) 1.6 (1.0 to 2.5) 2.0 (0.9 to 4.2) 1.3 (0.6 to 2.8)

Diabetes with complications (n = 3810 [22.1%])

1.0 (0.7 to 1.5) 1.1 (0.7 to 1.9) 0.3 (0.1 to 1.0)

Diabetes without complications (n = 823 [4.8%])

0.9 (0.4 to 1.9) 0.7 (0.2 to 2.3) 2.5 (0.6 to 10.2)

Hypertension (n = 9300 [53.9%]) 1.2 (0.9 to 1.6) 1.3 (0.9 to 2.1) 1.2 (0.7 to 2.2)

Myocardial infarction (n = 142 [0.8%]) 1.1 (0.2 to 4.7) 0* 22.5 (6.9 to 73.9)

Congestive heart failure (n = 405 [2.4%])

0.5 (0.2 to 1.7) 1.8 (0.5 to 6.3) 5.3 (2.3 to 11.9)

Ischemic heart disease (n = 2262 [13.1%])

1.0 (0.6 to 1.6) 0.7 (0.3 to 1.6) 0.4 (0.1 to 1.1)

Osteoporosis (n = 1460 [8.5%]) 1.1 (0.6 to 1.8) 1.2 (0.5 to 2.7) 1.3 (0.6 to 2.8)

Peripheral vascular disease (n = 305 [1.8%])

1.4 (0.6 to 3.6) 2.4 (0.7 to 7.9) 0.8 (0.2 to 4.2)

Cerebrovascular accident (n = 422 [2.5%])

1.4 (0.6 to 3.2) 0.5 (0.1 to 4.0) 4.0 (1.6 to 9.8)

Dementia (n = 83 [0.5%]) 1.0 (0.1 to 7.6) 0.0 (0.0 to 0.0) 0*

Chronic obstructive pulmonary disease (n = 2424 [14.1%])

1.4 (0.9 to 2.0) 0.9 (0.5 to 1.8) 1.8 (0.9 to 3.5)

Rheumatologic disease (n = 1009 [5.9%])

1.2 (0.7 to 2.1) 0.8 (0.3 to 2.1) 2.7 (1.2 to 6.2)

Peptic ulcer (n = 255 [1.5%]) 1.5 (0.6 to 3.8) 0.9 (0.1 to 6.5) 1.5 (0.3 to 6.7)

Mild liver disease (n = 136 [0.8%]) 3.9 (1.6 to 9.6) 3.1 (0.7 to 13.1) 0*

Paraparesis/hemiparesis (n = 44 [0.3%])

1.4 (0.2 to 11.1) 0* 0*

Renal disease (n = 370 [2.2%]) 1.3 (0.5 to 3.0) 2.0 (0.7 to 5.6) 3.0 (1.0 to 8.4)

Cancer (n = 1429 [8.3%]) 1.0 (0.6 to 1.8) 1.0 (0.4 to 2.2) 1.7 (0.8 to 3.9)

Moderate to severe liver disease (n = 33 [0.2%])

0.8 (0.1 to 6.6) 0* 0*

Metastatic disease (n = 134 [0.8%]) 2.8 (0.9 to 8.5) 5.5 (1.4 to 21.6) 2.0 (0.2 to 16.8)


Age 60–69 (n = 6455) 0.8 (0.6 to 1.2) 0.8 (0.5 to 1.2) 4.8 (1.1 to 21.1)

Age 70–79 yr (n = 5487) 0.6 (0.4 to 0.9) 0.4 (0.2 to 0.7) 4.1 (0.9 to 18.4)

Age ≥ 80 yr (n = 1604) 0.7 (0.4 to 1.2) 0.4 (0.2 to 1.1) 10.6 (2.3 to 48.6)

Female sex (n = 10 303) 0.8 (0.6 to 1.0) 0.7 (0.4 to 1.0) 0.9 (0.5 to 1.7)

Centre B (n = 1806) 0.7 (0.4 to 1.2) 0.5 (0.2 to 1.1) 0.9 (0.3 to 2.6)

Centre C (n = 3443) 0.9 (0.6 to 1.3) 0.6 (0.4 to 1.1) 0.9 (0.4 to 2.0)

Centre D (n = 1867) 0.2 (0.1 to 0.6) 0.2 (0.1 to 0.7) 2.0 (0.9 to 4.5)

Centre E (n = 3676) 0.9 (0.6 to 1.3) 0.6 (0.3 to 1.0) 0.9 (0.4 to 2.1)



risk-rhee.indd 172 2018-05-18 12:18 PM

RESEARCH


closely associated with increased infection rates after TJR.28,29 Another well-documented risk factor for infection is postoperative blood transfusion,13,30 and our study showed its association with infection for TKA but not THA. In a recent meta-analysis, Kim and colleagues31 summarized 6 papers to show the associa-tion between allogeneic blood transfusion and surgical site infection. However, only 1 study was conducted solely in patients who had undergone THA, and that study did not show a statistically signi�cant association between transfusion and infection.32 Four studies included in the meta-analysis were conducted in both THA and TKA populations but did not analyze the data for these populations separately.13,33–35 Therefore, although allogeneic blood transfusion has its inherent risks and must be administered judiciously, it does not appear to increase the rate of surgical site infection after THA as it does after TKA. The results also may

vary depending on the exact definition of infection (e.g., deep-seated, super�cial).32

The rate of revision at 1 year was 1.6% and 1.4% after THA and TKA, respectively. The paraplegia/hemiplegia category includes patients with �accid or spastic paralysis secondary to problems in the central or peripheral nervous system. Paraplegia/hemiplegia was associated with early revision after THA but not after TKA. In patients with paraplegia or hemiplegia, THA has been shown to result in a higher revision rate owing to the more complex nature of the surgery, which is linked to increased rates of disloca-tion, aseptic loosening and periprosthetic fracture.36,37 The lower revision rates in women and in patients aged 70 or more in our study may have been due to the lower infec-tion rates in these groups.

Mortality rates after TJR were closely associated with various medical comorbidities and age. This correlation is well established in the literature.38–40 Despite the increasing

Table 5. Association of risk factors with knee infection, revision and death 1 year after total knee arthroplasty

Risk factor

OR (95% CI)



Blood transfusion 1.6 (1.1 to 2.4) 1.5 (1.0 to 2.5) 1.1 (0.7 to 1.8)

Diabetes with complications 1.0 (0.8 to 1.4) 1.1 (0.8 to 1.5) 0.8 (0.5 to 1.3)

Diabetes without complications 1.0 (0.5 to 1.8) 1.0 (0.5 to 2.0) 1.1 (0.5 to 2.2)

Hypertension 1.3 (1.0 to 1.7) 1.1 (0.8 to 1.4) 1.3 (0.9 to 1.9)

Myocardial infarction 1.4 (0.5 to 4.0) 1.0 (0.2 to 4.3) 6.8 (3.3 to 14.0)

Congestive heart failure 1.2 (0.6 to 2.4) 1.9 (0.9 to 3.8) 2.9 (1.7 to 5.1)

Ischemic heart disease 0.9 (0.6 to 1.3) 0.7 (0.5 to 1.2) 0.9 (0.5 to 1.4)

Osteoporosis 1.1 (0.7 to 1.7) 1.1 (0.7 to 1.8) 0.8 (0.5 to 1.5)

Peripheral vascular disease 1.1 (0.5 to 2.4) 1.7 (0.8 to 3.7) 1.5 (0.7 to 3.3)

Cerebrovascular accident 1.6 (0.9 to 3.0) 1.3 (0.6 to 2.8) 2.2 (1.2 to 4.2)

Dementia 0.7 (0.1 to 4.9) 0* 0.5 (0.1 to 4.1)


1.3 (1.0 to 1.8) 0.9 (0.6 to 1.3) 1.3 (0.9 to 2.0)

Rheumatologic disease 1.3 (0.9 to 2.1) 0.9 (0.5 to 1.6) 2.0 (1.1 to 3.4)

Peptic ulcer 1.2 (0.5 to 2.8) 0.6 (0.1 to 2.3) 1.5 (0.6 to 3.6)

Mild liver disease 4.0 (1.8 to 8.5) 2.0 (0.7 to 5.8) 5.5 (2.2 to 13.8)

Paraparesis/hemiparesis 1.9 (0.4 to 8.3) 1.1 (0.1 to 8.9) 0.6 (0.1 to 5.9)

Renal disease 1.1 (0.5 to 2.3) 1.3 (0.6 to 2.7) 2.5 (1.3 to 4.8)

Cancer 1.0 (0.6 to 1.6) 1.1 (0.7 to 1.8) 1.7 (1.0 to 2.7)

Moderate to severe liver disease 1.0 (0.2 to 5.1) 2.1 (0.4 to 11.4) 1.5 (0.2 to 8.9)

Metastatic disease 2.0 (0.7 to 5.8) 2.1 (0.7 to 6.4) 4.5 (1.8 to 11.0)


Age 60–69 yr 0.8 (0.6 to 1.1) 0.8 (0.6 to 1.0) 1.7 (0.9 to 3.2)

Age 70–79 yr 0.6 (0.4 to 0.8) 0.4 (0.3 to 0.6) 2.4 (1.3 to 4.6)

Age ≥ 80 yr 0.6 (0.3 to 0.9) 0.3 (0.1 to 0.5) 6.6 (3.4 to 12.9)

Female sex 0.7 (0.6 to 0.9) 0.7 (0.5 to 0.9) 0.8 (0.6 to 1.1)

Centre B 0.8 (0.5 to 1.2) 0.9 (0.5 to 1.4) 0.8 (0.4 to 1.5)

Centre C 1.0 (0.7 to 1.3) 1.1 (0.8 to 1.6) 1.1 (0.7 to 1.8)

Centre D 0.3 (0.2 to 0.6) 0.8 (0.5 to 1.3) 2.0 (1.2 to 3.3)

Centre E 0.9 (0.7 to 1.3) 0.9 (0.6 to 1.3) 0.9 (0.6 to 1.5)



risk-rhee.indd 173 2018-05-18 12:18 PM

RECHERCHE


number of older patients with more medical comorbidi-ties undergoing primary TJR, overall mortality rates are generally decreasing.41–46 However, the number of major postoperative complications after TJR has been increas-ing, resulting in increased morbidity and higher eco-nomic burden.46

Strengths and limitations

The main strength of the current study is the use of a large, representative population. Consequently, these �nd-ings are particularly helpful for local physicians and phys-icians in areas where the population shares characteristics similar to those in Nova Scotia. In addition, we used the same cohort of the population to analyze multiple postop-erative outcomes, which is relatively uncommon, as most studies focus on a single outcome. The challenges of using administrative data for research purposes have been well

documented.47,48 Along with limited information, inaccura-cies of diagnostic coding to measure outcomes have been highlighted.49,50 As administrative data have become more ubiquitous in health research, however, data entry person-nel have developed expertise in coding, and researchers have learned to use techniques such as data linkages across multiple sources to enhance accuracy and increase the functionality of these readily available data sources.

CONCLUSION

Our study identi�ed multiple risk factors for complications and a longer hospital stay after TJR. The results enable individualized risk strati�cation during the preoperative consultation, which will, in turn, allow involvement of the appropriate consulting services to optimize the patient’s preoperative medical condition.

Table 7. Factors influencing length of hospital stay after total joint replacement

Factor

Coefficient (95% CI)

Total hip arthroplasty Total knee arthroplasty

Year of surgery −0.06 (−0.06 to −0.05) −0.05 (−0.06 to −0.05)

Blood transfusion 0.28 (0.25 to 0.30) 0.32 (0.30 to 0.34)

Diabetes with complications

0.05 (0.02 to 0.08) 0.05 (0.04 to 0.07)

Diabetes without complications

0.09 (0.04 to 0.14) 0.09 (0.06 to 0.12)

Hypertension 0.01 (−0.01 to 0.03) 0.02 (0.00 to 0.03)

Myocardial infarction 0.25 (0.17 to 0.34) 0.35 (0.28 to 0.41)

Congestive heart failure 0.13 (0.07 to 0.19) 0.12 (0.08 to 0.16)

Ischemic heart disease 0.05 (0.02 to 0.07) 0.03 (0.01 to 0.05)

Osteoporosis 0.02 (0.00 to 0.05) 0.03 (0.01 to 0.05)

Peripheral vascular disease 0.05 (−0.01 to 0.11) 0.01 (−0.03 to 0.06)

Cerebrovascular accident 0.07 (0.02 to 0.13) 0.07 (0.03 to 0.11)

Dementia 0.32 (0.23 to 0.41) 0.33 (0.25 to 0.41)


0.08 (0.05 to 0.11) 0.07 (0.06 to 0.09)

Rheumatologic disease 0.05 (0.01 to 0.08) 0.01 (−0.01 to 0.04)

Peptic ulcer 0.12 (0.04 to 0.20) 0.09 (0.05 to 0.14)

Mild liver disease 0.05 (−0.05 to 0.15) 0.05 (−0.02 to 0.12)

Paraparesis/hemiparesis 0.39 (0.24 to 0.54) 0.43 (0.32 to 0.55)

Renal disease 0.06 (0.00 to 0.12) 0.07 (0.03 to 0.11)

Cancer 0.04 (0.01 to 0.07) 0.01 (−0.01 to 0.03)

Moderate to severe liver disease

0.12 (−0.13 to 0.36) 0.08 (−0.06 to 0.22)

Metastatic disease 0.08 (−0.01 to 0.17) 0.03 (−0.04 to 0.10)

Median household income 0.00 (0.00 to 0.00) 0.00 (0.00 to 0.00)

Age 60–69 yr 0.07 (0.05 to 0.09) 0.03 (0.01 to 0.04)

Age 70–79 yr 0.18 (0.16 to 0.21) 0.10 (0.09 to 0.12)

Age ≥ 80 yr 0.37 (0.34 to 0.40) 0.23 (0.21 to 0.25)

Female sex 0.09 (0.07 to 0.11) 0.08 (0.07 to 0.09)

Centre B −0.10 (−0.13 to −0.07) −0.09 (−0.11 to −0.06)

Centre C 0.09 (0.07 to 0.12) 0.00 (−0.02 to 0.02)

Centre D 0.11 (0.08 to 0.14) −0.20 (−0.22 to −0.18)

Centre E −0.03 (−0.05 to −0.01) −0.07 (−0.09 to −0.05)

CI = confidence interval.

Table 6. Association of risk factors with blood transfusion after total hip arthroplasty and total knee arthroplasty

Risk factor

OR (95% CI)

Total hip arthroplasty

Total knee arthroplasty

Year of surgery 0.9 (0.9 to 0.9) 0.9 (0.9 to 0.9)

Diabetes with complications 1.1 (0.9 to 1.3) 1.1 (1.0 to 1.3)

Diabetes without complications

1.6 (1.1 to 2.2) 1.6 (1.2 to 2.1)

Hypertension 0.9 (0.8 to 1.1) 1.0 (0.9 to 1.1)

Myocardial infarction 2.6 (1.6 to 4.1) 6.9 (4.6 to 10.4)

Congestive heart failure 1.7 (1.2 to 2.3) 2.4 (1.8 to 3.1)

Ischemic heart disease 1.3 (1.1 to 1.6) 1.1 (0.9 to 1.3)

Osteoporosis 1.5 (1.3 to 1.8) 1.4 (1.2 to 1.7)

Peripheral vascular disease 0.9 (0.6 to 1.4) 0.9 (0.6 to 1.4)

Cerebrovascular accident 1.0 (0.7 to 1.5) 1.3 (0.9 to 1.8)

Dementia 1.9 (1.2 to 3.2) 1.7 (1.0 to 3.1)


1.2 (1.0 to 1.5) 1.0 (0.9 to 1.2)

Rheumatologic disease 1.6 (1.2 to 2.0) 1.5 (1.2 to 1.9)

Peptic ulcer 1.7 (1.1 to 2.8) 1.7 (1.2 to 2.4)

Mild liver disease 1.3 (0.7 to 2.5) 1.6 (0.9 to 2.9)

Paraparesis/hemiparesis 1.3 (0.5 to 3.4) 1.9 (0.8 to 4.6)

Renal disease 2.0 (1.4 to 2.8) 2.0 (1.4 to 2.7)

Cancer 1.3 (1.1 to 1.6) 1.2 (1.0 to 1.5)

Moderate to severe liver disease

1.7 (0.4 to 6.8) 4.0 (1.6 to 9.8)

Metastatic disease 1.3 (0.7 to 2.2) 1.7 (1.0 to 3.0)

Median household income 1.0 (1.0 to 1.0) 1.0 (1.0 to 1.0)

Age 60–69 yr 1.0 (0.9 to 1.3) 1.5 (1.2 to 1.8)

Age 70–79 yr 1.7 (1.4 to 2.0) 3.0 (2.4 to 3.6)

Age ≥ 80 yr 3.2 (2.6 to 4.0) 6.1 (4.9 to 7.6)

Female sex 2.4 (2.1 to 2.8) 2.3 (2.0 to 2.6)

Centre B 0.7 (0.6 to 0.9) 1.0 (0.8 to 1.2)

Centre C 1.6 (1.3 to 1.9) 2.3 (2.0 to 2.7)

Centre D 1.1 (0.9 to 1.4) 1.2 (0.9 to 1.5)

Centre E 2.4 (2.0 to 2.8) 2.8 (2.4 to 3.3)


risk-rhee.indd 174 2018-05-18 12:18 PM

RESEARCH


Af�liations: From the Department of Surgery, Dalhousie University, Halifax, NS (Rhee, Lethbridge, Richardson, Dunbar).

Acknowledgements: The authors thank the Nova Scotia Department of Health and Wellness for granting access to the data and extracting the required variables in a timely manner.


Contributors: C. Rhee, L. Lethbridge and M. Dunbar designed the study. L. Lethbridge acquired the data, which all authors analyzed. C. Rhee and L. Lethbridge wrote the article, which all authors reviewed and approved for publication.

References

1. Nemes S, Rolfson O, W-Dahl A, et al. Historical view and future demand for knee arthroplasty in Sweden. Acta Orthop 2015;86:426-31.

2. Oviedo Baena AM, Moeschler SM, Smith HM, et al. Perioperative comorbidities and complications among patients undergoing primary total knee arthroplasty: a retrospective analysis and prospective sur-vey. J Clin Anesth 2015;27:558-65.

3. Carr AJ, Robertsson O, Graves S, et al. Knee replacement. Lancet 2012;379:1331-40.

4. Yu S, Garvin KL, Healy WL, et al. Preventing hospital readmissions and limiting the complications associated with total joint arthro-plasty. J Am Acad Orthop Surg 2015;23:e60-71.

5. Belmont PJ Jr, Goodman GP, Kusnezov NA, et al. Postoperative myocardial infarction and cardiac arrest following primary total knee and hip arthroplasty: rates, risk factors, and time of occurrence. J Bone Joint Surg Am 2014;96:2025-31.

6. Triantafyllopoulos GK, Soranoglou SG, Memtsoudis SG, et al. Rate and risk factors for periprosthetic joint infection among 36,494 pri-mary total hip arthroplasties. J Arthroplasty 2018;33:1166-70.

7. Pruzansky JS, Bronsom MJ, Grelsamer RP, et al. Prevalence of mod-i�able surgical site infection risk factors in hip and knee joint arthro-plasty patients at an urban academic hospital. J Arthroplasty 2014;29: 272-6.

8. Ali AM, Loef�er MD, Aylin P, et al. Factors associated with 30-day readmission after primary total hip arthroplasty: analysis of 514,455 procedures in the UK National Health Service. JAMA Surg 2017; 152:e173949.

9. Hunt LP, Blom A, Wilkinson JM. An analysis of 30-day mortality after weekend versus weekday elective joint arthroplasty in England and Wales: a cohort study using the National Joint Registry Dataset. Bone Joint J 2017;99-B:1618-28.

10. Messina G, Forni S, Rosadini D, et al. Risk adjusted mortality after hip replacement surgery: a retrospective study. Ann Ist Super Sanita 2017;53:40-5.

11. Allen Butler R, Rosenzweig S, Myers L, et al. The Frank Stinch�eld Award: the impact of socioeconomic factors on outcome after THA: a prospective, randomized study. Clin Orthop Relat Res 2011;469: 339-47.

12. Zhang W, Lyman S, Boutin-Foster C, et al. Racial and ethnic dispar-ities in utilization rate, hospital volume, and perioperative outcomes after total knee arthroplasty. J Bone Joint Surg Am 2016;98:1243-52.

13. Friedman R, Homering M, Holberg G, et al. Allogeneic blood trans-fusions and postoperative infections after total hip or knee arthro-plasty. J Bone Joint Surg Am 2014;96:272-8.

14. Hart A, Khalil JA, Carli A, et al. Blood transfusion in primary total hip and knee arthroplasty. Incidence, risk factors, and thirty-day complication rates. J Bone Joint Surg Am 2014;96:1945-51.

15. Clement ND, Muzammil A, Macdonald D, et al. Socioeconomic sta-tus affects the early outcome of total hip replacement. J Bone Joint Surg Br 2011;93:464-9.

16. Hollowell J, Grocott MP, Hardy R, et al. Major elective joint replace-ment surgery: socioeconomic variations in surgical risk, postoperative morbidity and length of stay. J Eval Clin Pract 2010;16:529-38.

17. Cook NR. Use and misuse of the receiver operating characteristic curve in risk prediction. Circulation 2007;115:928-35.

18. Steyerberg EW, Vickers AJ, Cook NR, et al. Assessing the perfor-mance of prediction models: a framework for some traditional and novel measures. Epidemiology 2010;21:128-38.

19. Lix LM, Quail J, Fadahunsi O, et al. Predictive performance of comorbidity measures in administrative databases for diabetes cohorts. BMC Health Serv Res 2013;13:340.

20. Tayton ER, Frampton C, Hooper GJ, et al. The impact of patient and surgical factors on the rate of infection after primary total knee arthroplasty: an analysis of 64,566 joints from the New Zealand Joint Registry. Bone Joint J 2016;98-B:334-40.

21. Cancienne JM, Werner BC, Browne JA. Complications of primary total knee arthroplasty among patients with rheumatoid arthritis, psoriatic arthritis, ankylosing spondylitis, and osteoarthritis. J Am Acad Orthop Surg 2016;24:567-74.

22. Jamsen E, Huhtala H, Puolakka T, et al. Risk factors for infection after knee arthroplasty. A register-based analysis of 43,149 cases. J Bone Joint Surg Am 2009;91:38-47.

23. Bohl DD, Sershon RA, Fillingham YA, et al. Incidence, risk factors, and sources of sepsis following total joint arthroplasty. J Arthroplasty 2016;31:2875-9.

24. Meehan JP, Danielsen B, Kim SH, et al. Younger age is associated with a higher risk of early periprosthetic joint infection and aseptic mechanical failure after total knee arthroplasty. J Bone Joint Surg Am 2014;96:529-35.

25. Malinzak RA, Ritter MA, Berend ME, et al. Morbidly obese, dia-betic, younger, and unilateral joint arthroplasty patients have ele-vated total joint arthroplasty infection rates. J Arthroplasty 2009; 24(Suppl):84-8.

26. Namba RS, Inacio MC, Paxton EW. Risk factors associated with sur-gical site infection in 30,491 primary total hip replacements. J Bone Joint Surg Br 2012;94:1330-8.

27. Wu C, Qu X, Liu F, et al. Risk factors for periprosthetic joint infection after total hip arthroplasty and total knee arthroplasty in Chinese patients. PLoS One 2014;9:e95300.

Table 8. Overall diagnostic statistics for study outcomes

Time point; outcomeAdjusted R-squared Bier score c-statistic

Total hip arthroplasty

3 mo

Infection 0.056 0.008 0.691

Revision 0.055 0.009 0.685

Death 0.253 0.004 0.864

Length of hospital stay 0.375 — —

1 yr

Infection 0.051 0.012 0.686

Revision 0.041 0.016 0.663

Death 0.166 0.011 0.796

Total knee arthroplasty

3 mo

Infection 0.025 0.011 0.636

Revision 0.052 0.005 0.703

Death 0.194 0.003 0.775

Length of hospital stay 0.361 — —

1 yr

Infection 0.025 0.016 0.626

Revision 0.026 0.014 0.639

Death 0.128 0.009 0.771

risk-rhee.indd 175 2018-05-18 12:18 PM

RECHERCHE


28. Issa K, Boylan MR, Naziri Q, et al. The impact of hepatitis C on short-term outcomes of total joint arthroplasty. J Bone Joint Surg Am 2015;97:1952-7.

29. Orozco F, Post ZD, Baxi O, et al. Fibrosis in hepatitis C patients predicts complications after elective total joint arthroplasty. J Arthro-plasty 2014;29:7-10.

30. Klika AK, Small TJ, Saleh A, et al. Primary total knee arthroplasty allogenic transfusion trends, length of stay, and complications: nation-wide inpatient sample 2000–2009. J Arthroplasty 2014;29:2070-7.

31. Kim JL, Park JH, Han SB, et al. Allogeneic blood transfusion is a sig-ni�cant risk factor for surgical-site infection following total hip and knee arthroplasty: a meta-analysis. J Arthroplasty 2017;32:320-5.

32. Dowsey MM, Choong PF. Obesity is a major risk factor for pros-thetic infection after primary hip arthroplasty. Clin Orthop Relat Res 2008;466:153-8.

33. Newman ET, Watters TS, Lewis JS, et al. Impact of perioperative allogeneic and autologous blood transfusion on acute wound infec-tion following total knee and total hip arthroplasty. J Bone Joint Surg Am 2014;96:279-84.

34. Rosencher N, Kerkkamp HE, Macheras G, et al. Orthopedic Sur-gery Transfusion Hemoglobin European Overview (OSTHEO) study: blood management in elective knee and hip arthroplasty in Europe. Transfusion 2003;43:459-69.

35. Innerhofer P, Klingler A, Klimmer C, et al. Risk for postoperative infection after transfusion of white blood cell-�ltered allogeneic or autologous blood components in orthopedic patients undergoing pri-mary arthroplasty. Transfusion 2005;45:103-10.

36. King G, Hunt LP, Wilkinson JM, et al. Good outcome of total hip replacement in patients with cerebral palsy: a comparison of 389 patients and 425,813 controls from the National Joint Registry for England and Wales. Acta Orthop 2016;87:93-9.

37. Yoshino S, Shiga H, Nakamura H, et al. Longterm outcome of total joint arthroplasty in nonambulatory patients with rheumatoid arthri-tis. J Rheumatol 1999;26:1076-9.

38. Hunt LP, Ben-Shlomo Y, Clark EM, et al. 45-day mortality after 467,779 knee replacements for osteoarthritis from the National Joint Registry for England and Wales: an observational study. Lancet 2014; 384:1429-36.

39. Belmont PJ Jr, Goodman GP, Hamilton W, et al. Morbidity and mortality in the thirty-day period following total hip arthroplasty: risk factors and incidence. J Arthroplasty 2014;29:2025-30.

40. Belmont PJ Jr, Goodman GP, Waterman BR, et al. Thirty-day post-operative complications and mortality following total knee arthro-plasty: incidence and risk factors among a national sample of 15,321 patients. J Bone Joint Surg Am 2014;96:20-6.

41. Cram P, Lu X, Kaboli PJ, et al. Clinical characteristics and outcomes of Medicare patients undergoing total hip arthroplasty, 1991–2008. JAMA 2011;305:1560-7.

42. Ong KL, Kurtz SM, Lau E, et al. Prosthetic joint infection risk after total hip arthroplasty in the Medicare population. J Arthroplasty 2009;24(Suppl):105-9.

43. Hip and knee replacements in Canada: Canadian Joint Replacement Registry 2015 annual report. Toronto: Canadian Institute for Health Information; 2015.

44. Hunt LP, Ben-Shlomo Y, Clark EM, et al. 90-day mortality after 409,096 total hip replacements for osteoarthritis, from the National Joint Registry for England and Wales: a retrospective analysis. Lancet 2013;382:1097-104.

45. Lalmohamed A, Vestergaard P, de Boer A, et al. Changes in mortality patterns following total hip or knee arthroplasty over the past two decades: a nationwide cohort study. Arthritis Rheumatol 2014;66:311-8.

46. Kirksey M, Chiu YL, Ma Y, et al. Trends in in-hospital major mor-bidity and mortality after total joint arthroplasty: United States 1998–2008. Anesth Analg 2012;115:321-7.

47. Katz A, Soodeen RA, Bogdanovic B, et al. Can the quality of care in family practice be measured using administrative data? Health Serv Res 2006;41:2238-54.

48. Zhan C, Miller MR. Administrative data based patient safety research: a critical review. Qual Saf Health Care 2003;12(Suppl 2): ii58-63.

49. Upadhyaya CD, Mumaneni PV. Comparison of ICD-9-based, retro-spective, and prospective assessments of perioperative complications: assessment of accuracy in reporting [editorial]. J Neurosurg Spine 2011;14:14-5, discussion 5.

50. O’Malley KJ, Cook KF, Price MD, et al. Measuring diagnoses: ICD code accuracy. Health Serv Res 2005;40:1620-39.

risk-rhee.indd 176 2018-05-18 12:18 PM



Tranexamic acid administration to older patients undergoing primary total hip arthroplasty conserves hemoglobin and reduces blood loss

Background: Tranexamic acid effects in older people are dif�cult to predict. This study investigated the following research questions: 1) Is tranexamic acid effective in older patients undergoing primary total hip arthroplasty (THA)? and 2) Is there a dif-ference in the effect of tranexamic acid between younger and older patients?

Methods: This was a 2-phase retrospective matched-pair study of patients who underwent THA in 2007–2013. All procedures were performed by surgeons with at least 10 years’ experience as senior consultant. In the �rst phase, 58 patients aged 65 years or more who received tranexamic acid were matched 1:1 with patients who did not receive tranexamic acid for age, sex, American Society of Anesthesiologists (ASA) classi�cation and body mass index. In the second phase, 58 patients aged 65 years or more who received tranexamic acid were matched 1:1 with patients less than 65 years of age who received tranexamic acid for sex, ASA classi�cation and body mass index. The primary outcome measures were percent maximum decrease in hemoglobin level and estimated blood loss after surgery.

Results: In the �rst phase, patients who received tranexamic acid conserved postop-erative hemoglobin by a mean of 10.26 g/L (standard deviation [SD] 9.89 g/L) com-pared to the control group (p < 0.001). The mean difference in the estimated periop-erative blood loss between the 2 groups was 410 mL (SD 376 mL) (p < 0.001), which indicated less bleeding in the treatment group. In the second phase, there was no dif-ference between the younger (mean age 55.1 [SD 7.28] yr) and older (mean age 75.6 [SD 6.35] yr) groups in mean lowest postoperative hemoglobin level or percent decrease in hemoglobin level.

Conclusion: Tranexamic acid reduced the postoperative decrease in hemoglobin level and blood loss in older patients. Moreover, the signi�cant hemoglobin-sparing effect of tranexamic acid in older patients was similar to that observed in younger patients.

Contexte : Les effets de l’acide tranexamique sont dif�ciles à prévoir chez les per-sonnes âgées. Avec cette étude, nous avons voulu répondre aux 2 questions suivantes : 1) L’acide tranexamique est-il ef�cace chez les patients âgés soumis à une intervention chirurgicale primaire pour prothèse totale de la hanche (PTH)?, et 2) L’acide tranexa-mique produit-il un effet différent selon que les patients sont jeunes ou âgés?

Méthodes : Cette étude rétrospective en 2 phases sur des paires appariées a regroupé des patients soumis à une intervention pour PTH entre 2007 et 2013. Toutes les interventions ont été effectuées par des chirurgiens détenant au moins 10 ans d’expé-rience à titre de consultants principaux. Au cours de la première phase, 58 patients de 65 ans ou plus ayant reçu de l’acide tranexamique ont été assortis (rapport 1:1), selon l’âge, le sexe, la classi�cation ASA (American Society of Anesthesiologists) et l’indice de masse corporelle, à des patients n’en ayant pas reçu. Au cours de la deuxième phase, 58 patients de 65 ans ou plus ayant reçu de l’acide tranexamique ont été assortis (rapport 1:1), selon le sexe, la classi�cation ASA et l’indice de masse corporelle, à des patients de moins de 65 ans ayant aussi reçu de l’acide tranexamique. Les paramètres principaux étaient la diminution maximale en pourcentage du taux d’hémoglobine et la perte sanguine estimée après l’intervention chirurgicale.

Résultats : Pour la première phase, les patients qui ont reçu l’acide tranexamique ont maintenu une hémoglobine postopératoire moyenne à 10,26 g/L (écart-type [É.-T.] 9,89 g/L) comparativement au groupe témoin (p < 0,001). La différence moyenne entre les 2 groupes pour ce qui est des pertes sanguines periopératoires a été de 410 mL (É.-T. 376 mL) (p < 0,001), indiquant de ce fait une perte sanguine moindre dans le groupe traité. Pour la deuxième phase, on n’a noté aucune différence entre le

Hossam El Beheiry, MD, PhD Ashley Lubberdink, MD Nigel Clements, MD Kiran Dihllon, MScN Vicky Sharma, MScN

Accepted for publication Nov. 17, 2017

Correspondence to: H. El Beheiry Mississauga Hospital 100 Queensway W Mississauga ON L5B 1B8 [email protected]

DOI: 10.1503/cjs.012817

acid-elbeheriy.indd 177 2018-05-22 9:10 AM

RECHERCHE


T he anti�brinolytic action of tranexamic acid stabilizes the formed clot and consequently enhances micro-vascular hemostasis.1,2 Thus, recent systematic

reviews and meta-analyses of published randomized con-trolled trials concluded that administration of tranexamic acid reduces blood loss and the need for transfusion in patients undergoing total hip arthroplasty (THA).3–5 How-ever, there has been a persistent lack of knowledge on the effects of tranexamic acid in older patients undergoing THA because this population has been severely underrepresented in most randomized controlled trials.6–12 In addition, the effect of tranexamic acid in older people is dif�cult to pre-dict. This is because of the presence of 2 opposing factors. First, geriatric patients in general are at higher risk for peri-operative surgical bleeding, which emanates from acquired coagulation disorders and use of anticoagulation and anti-platelet medications as well as osteoporosis and osteope-nia.13,14 Second, pharmacokinetic studies suggest that older patients who receive tranexamic acid may have less perioper-ative bleeding because of increased blood tranexamic acid levels resulting from an aging-induced decrease in glomeru-lar �ltration rate and volume of distribution.15,16

Therefore, the objective of the current study was to investigate the effect of tranexamic acid on blood loss and reduction in the incidence and volume of allogeneic blood transfusion in older patients undergoing THA. To achieve the study objective, we investigated 2 primary hypotheses: 1) tranexamic acid administered at the time of skin incision to older patients undergoing THA reduces the decrease in postoperative hemoglobin concentration and 2) tranexamic acid reduces the decrease in postoperative hemoglobin level more in younger patients than in older patients undergoing THA.

METHODS

This was a retrospective single-centre multisurgeon matched-pair study including patients who underwent THA in 2007–2013. The study was conducted at a tertiary health care facility after appropriate research ethics board approval.

We retrospectively reviewed the charts of 382 patients who had primary total hip replacement during the study period. Patients included in the retrospective study had primary THA, were aged 21 years or more at the time of surgery, were classi�ed as American Society of Anesthesi-ologists (ASA) level I, II or III, and received tranexamic acid (given as a single bolus intravenously just before skin

incision) or did not receive tranexamic acid. Patients who had revision THA, were classi�ed as ASA level IV, had a risk factor for thromboembolism, received an anti�brinol-ytic or coagulant agent preoperatively, had a blood transfu-sion intraoperatively or received postoperative anticoagu-lant therapy that differed from the institutional protocol for joint replacements were excluded. The charts of eligi-ble patients were reviewed in detail. Data were extracted from a database kept in the orthopedic department as well as the clinic charts kept with the orthopedic surgeons. The chart review was performed by a single person (A.L.) who was not involved in the analysis of the results.

To examine the study hypotheses, this investigation was conducted in 2 phases. The first phase determined the effects of tranexamic acid administration in older patients (age ≥ 65 yr) by comparing each older patient to a matched control patient who did not receive the anti�brinolytic (negative control). Patients were matched 1:1 on age (5-yr intervals), sex, ASA level and body mass index (3-kg/m2 intervals). In the second phase, we compared the ef�cacy of tranexamic acid in older and younger patients (positive control). The patients in the first phase who received tranexamic acid were matched with a younger patient (< 65 yr) who received tranexamic acid on sex, ASA level and body mass index (3-kg/m2 intervals). Every younger patient should have received tranexamic acid similarly to her/his corresponding older patient.

The primary outcome measures were percent decrease in patient’s hemoglobin level and estimated blood loss after surgery. We calculated these from chart data consisting of preoperative hemoglobin level (PreHB), and lowest post-operative hemoglobin level before discharge and before any postoperative blood transfusion (PostHB). The lowest postoperative hemoglobin level usually occurred on the third or fourth postoperative day. We calculated the pri-mary outcomes as follows: percent decrease in patient’s hemoglobin level = [PreHB – PostHB/PreHB]*100, and estimated blood loss = Estimated patient’s blood volume*[ln(PreHB/PostHB)].17 We determined estimated patient’s blood volume as follows:18

PBV (male) = (0.3669 × Ht3) + (0.03219 × Wt ) + 0.6041

PBV (female) = (0.3561 × Ht3) + (0.03308 × Wt) + 0.1833

where PBV = patient’s blood volume in millilitres, Ht = height in metres and Wt = weight in kilograms.

groupe plus jeune (âge moyen 55,1 ans [É.-T. 7,28 ans]) et le groupe plus âgé (âge moyen 75,6 ans [É.-T. 6,35 ans]) pour ce qui est du taux d’hémoglobine postopératoire moyen le plus bas ou le pourcentage de baisse du taux d’hémoglobine.

Conclusion : L’acide tranexamique a permis d’atténuer la baisse postopératoire de l’hémoglobine et les pertes sanguines chez les patients âgés. De plus, l’effet signi�catif de l’acide tranexamique sur le maintien de l’hémoglobine chez les patients âgés a été similaire à ce qui s’observe chez les patients plus jeunes.


RESEARCH


The secondary outcome measures were occurrence of allogenic blood transfusion perioperatively, number of allo-genic blood units transfused perioperatively, length of hos-pital stay, occurrence of surgical infection postoperatively, occurrence of deep vein thrombosis (assessed clinically) and occurrence of pulmonary embolism (assessed clinically and by imaging studies). The following confounders and co interventions were recorded: patient demographic char-acteristics, medications, comorbidities, name of surgeon, duration of surgery, type of anesthesia (general v. regional), �uid administration, preoperative iron therapy, preopera-tive erythropoietin therapy, preoperative coagulation pro-�le and details of the surgical technique.

Statistical analysis and sample size calculation

We compared means of categorical variables using the paired Student t test. Nonparametric data and data that deviated signi�cantly from normal distribution were com-pared with the use of the Wilcoxon matched pairs signed-rank test. We performed the statistical analysis using Stata 10 (StataCorp).

Sample size calculation for the �rst phase was based on the difference in the primary outcome (decrease in hemoglo-bin level from preoperative level) between the older patients who received tranexamic acid and those who did not.19 Avail-able data showed that tranexamic acid can produce a saving of about 25% of the preoperative mean hemoglobin value. Based on this information, the required sample size was 47 matched pairs of patients (2-tailed paired t test: effect size f = 0.36, α = 0.05, power = 0.8). We increased the sample size by about 20% to compensate for expected incomplete patient charts, resulting in 58 pairs; i.e., 116 patients. Simi-larly, the sample size estimate for the second phase was based on the difference in the primary outcome between the older and younger patients who received tranexamic acid.19 Based on previous experience, we expected that tranexamic acid would produce a lesser decrease in preoperative hemoglobin level, by about 20%, in younger patients than older patients. Accordingly, the required sample size was 48 matched pairs of patients (2-tailed paired t test: effect size f = 0.41, α = 0.05, power = 0.8). Again, we increased the sample size by about 20% to compensate for expected incomplete patient charts, resulting in 58 patients per group; i.e., 116 patients.

RESULTS

In patients included in both phases of the study, the deci-sion to administer tranexamic acid was based on the absence of risk factors for postoperative thromboembolism and surgeon’s preference. All patients had uncemented hip replacement through the lateral surgical approach. The primary THA procedures were performed by 5 orthopedic surgeons, all of whom had at least 10 years’ experience as senior consultant.

Effect of tranexamic acid in older patients

The treatment and control groups had similar demographic characteristics, which indicated adequate matching (Table 1). The mean tranexamic acid dosage was 18.07 (standard devia-tion [SD] 3.59) mg/kg. Postoperatively, patients in the treat-ment group conserved hemoglobin by a mean of 10.26 g/L (SD 9.89 g/L) compared to control (p < 0.001) (Table 2). The mean difference in the estimated blood loss between the 2 groups was 410 mL (SD 376 mL) (p < 0.001), which indi-cated less total perioperative bleeding in the treatment group.

Table 1. Preoperative characteristics of matched pairs of older patients (age ≥ 65 yr)

Characteristic

No. (%) of patients*

p value†Control group

n = 58

Tranexamic acid groupn = 58

Demographic

Age, mean ± SD; yr 75.12 ± 6.20 75.57 ± 6.35 0.08

Sex 0.5

Female 42 (72) 42 (72)

Male 16 (28) 16 (28)

Body mass index, mean ± SD

28.3 ± 4.52 27.8 ± 4.73 0.2

ASA classification 0.5

II 11 (19) 11 (19)

III 47 (81) 47 (81)

Preoperative laboratory values, mean ± SD

Hemoglobin level, g/L 131.07 ± 11.58 132.67 ± 11.31 0.4

Platelet count, × 109/L 245 ± 63 250 ± 82 0.5

International normalized ratio

1.02 ± 0.05 0.98 ± 0.06 0.7

Surgery

Surgical site 0.5

Right 31 (53) 30 (52)

Left 27 (47) 28 (48)

Surgical duration, mean ± SD; min

84.12 ± 23.61 76.60 ± 17.67 0.08

Prosthesis 0.5

DePuy Synthes, Johnson & Johnson

50 (86) 40 (69)

Zimmer Biomet 8 (14) 18 (31)

Anesthesia 0.5

General 19 (33) 14 (24)

Spinal 39 (67) 44 (76)

Patients per surgeon

Surgeon 1 0 (0) 38 (66)

Surgeon 2 5 (9) 20 (34)

Surgeon 3 27 (47) 0 (0)

Surgeon 4 17 (29) 0 (0)

Surgeon 5 9 (16) 0 (0)

Tranexamic acid dosage, mean ± SD; mg/kg

— 18.07 ± 3.59

ASA = American Society of Anesthesiologists; SD = standard deviation.

*Except where noted otherwise.

†Paired t test.


RECHERCHE


Table 2. Perioperative outcomes recorded in matched pairs of older patients (age ≥ 65 yr)

Outcome


p value†Control group

n = 58Tranexamic acid group

n = 58

Hematologic

Preoperative hemoglobin level, mean ± SD; g/L

131.07 ± 11.58 132.67 ± 11.31 0.4

Lowest postoperative hemoglobin level, mean ± SD; g/L

89.53 ± 10.20 99.79 ± 10.43 < 0.001

% decrease in hemoglobin level, mean ± SD

31.56 ± 6.14 24.65 ± 6.27 < 0.001

Estimated total blood loss, mean ± SD; mL

1618 ± 448 1208 ± 407 < 0.001

Packed erythrocytes transfusion 3 (5) 1 (2) 0.6

Total volume of packed erythrocytes given, mL

2100 (7 units) 600 (2 units) 0.3

Postoperative recovery

Length of hospital stay, mean ± SD; d

6.24 ± 3.62 5.31 ± 2.82 0.05

Composite complication rate, % 8.6 6.9 0.4

Deep vein thrombosis 2 (3) 0 (0) 0.1

Pulmonary embolism 2 (3) 2 (3) 0.5

Infection 1 (2) 2 (3) 0.3

Discharged to rehabilitation 11 (19) 12 (21) 0.4



†Paired t test.

Fig. 1. Mean maximum decrease in postoperative hemoglobin level in control (no tranexamic acid) and treatment (tranexamic acid) groups, by age group. Error bars represent standard deviation. *Signifi-cantly different from control (p < 0.05).

15

20

25

30

35

40

65–69

Mea

n m

axim

um d

ecre

ase

in h

emog

lobi

n le

vel,

%

Control

Tranexamic acid

*

n = 15,15

70–74

n = 12,12

75–79

n = 14,14

80–84

n = 12,12

≥ 85

n = 5,5 n = 58,58

≥ 65

**

*

*

*

Age, yr


RESEARCH


The composite rate of complications (deep vein throm-bosis, pulmonary embolism and infection) was not statistic-ally signi�cantly different between the treatment (8.6%) and control (6.9%) groups (Table 2). The hospital stay was longer in the control group than in the treatment group (6.2 d [SD 3.6 d] v. 5.3 d [SD 2.8 d], p = 0.05).

When we examined the percent maximum decrease in postoperative hemoglobin concentration by age group, in each age group, the value was statistically significantly higher in the control group than in the treatment group (Fig. 1) (p < 0.05).

Efficacy of tranexamic acid in younger and older patients

The mean age of the younger and older patients was 55.1 (SD 7.28) years and 75.6 (SD 6.35) years, respectively (Table 3). The mean tranexamic acid dosage was similar in the 2 groups (17.77 [SD 4.45] mg/kg and 18.07 [SD 3.59] mg/kg, respectively). There was no difference between the 2 groups in mean lowest postoperative hemo-globin level or percent maximum decrease in postoperative hemoglobin level. The estimated perioperative blood loss was similar in the younger (mean 1281 mL [SD 382 mL]) and older (mean 1208 mL [SD 407 mL]) patients. Postop-eratively, the daily decrease in hemoglobin level was almost identical in the 2 groups (Fig. 2).

The composite complication rate was signi�cantly dif-ferent between the younger and older patients (0% v. 6.9%, p = 0.02) (Table 4). The hospital length of stay in the younger (5.3 d [SD 2.48 d) and older (5.3 d [SD 2.82 d) groups was similar.

DISCUSSION

We found that a single bolus of tranexamic acid adminis-tered intravenously at the initial skin incision during pri-mary THA surgery reduced the postoperative decrease in hemoglobin level in older patients (≥ 65 yr). The signi�-cant hemoglobin-sparing effect of tranexamic acid in older patients was similar to that in younger patients. The com-posite complication rate was lower in younger patients than in older patients.

Other investigators have reported different dosage regi-mens for the administration of tranexamic acid during THA, including single and multiple injections, and contin-uous infusion following a loading dose.21 A single injection of nearly 20 mg/kg at the time of the initial skin incision was used in our cohort. This dosage showed signi�cant ef�cacy and is simple in application and clinical utility. In a pharmacokinetic study, Benoni and colleagues20 found that such single doses maintain therapeutic serum and joint �uid levels for 8 hours after injection. A study in which multiple tranexamic acid injections were used21 did not show results superior to ours.

Our results validated our primary hypothesis that tran-examic acid administered at the time of skin incision to older patients undergoing hip replacement surgery reduces the decrease in postoperative hemoglobin con-centration. Based on recent large studies that showed aging as an independent risk factor associated with allo-genic blood transfusion,22–24 we also hypothesized that older patients experience more perioperative bleeding than do younger patients and that tranexamic acid is more effective in younger patients. However, our results were at odds with this proposition: our results in older patients

Table 3. Preoperative characteristics of matched pairs of younger and older patients who received tranexamic acid

Characteristic


p value†Age < 65 yr

n = 58 Age ≥ 65 yr

n = 58

Demographic

Age, mean ± SD; yr 55.10 ± 7.28 75.57 ± 6.35 < 0.001

Sex 0.5

Female 42 (72) 42 (72)

Male 16 (28) 16 (28)

Body mass index, mean ± SD

29.3 ± 3.66 27.8 ± 4.73 0.4

ASA classification 0.5

II 11 (19) 11 (19)

III 47 (81) 47 (81)

Preoperative laboratory values, mean ± SD

Hemoglobin level, g/L 135.58 ± 10.03 132.67 ± 11.31 0.5

Platelet count, × 109/L- 230 ± 66 250 ± 82 0.7

International normalized ratio

1.10 ± 0.07 0.98 ± 0.06 0.6

Surgery

Surgical site < 0.001

Right 20 (34) 30 (52)

Left 38 (66) 28 (48)

Surgical duration, mean ± SD; min

83.38 ± 21.71 76.60 ± 17.67 0.2

Prosthesis < 0.001

DePuy Synthes, Johnson & Johnson

30 (52) 40 (69)

Zimmer Biomet 28 (48) 18 (31)

Anesthesia 0.1

General 10 (17) 14 (24)

Spinal 48 (83) 44 (76)

Patients per surgeon

Surgeon 1 28 (48) 38 (66)

Surgeon 2 20 (34) 20 (34)

Surgeon 3 10 (17) 0 (0)

Surgeon 4 0 (0) 0 (0)

Surgeon 5 0 (0) 0 (0)

Tranexamic acid dosage, mean ± SD; mg/kg

17.77 ± 4.45 18.07 ± 3.59 0.7

ASA = American Society of Anesthesiologists; SD = standard deviation.


†Paired t test.


RECHERCHE


were unexpectedly similar to those of previously published reports in younger patients, including prospective11,25,26 and retrospective27–29 studies.

Although older patients who did not receive tranexamic acid had more perioperative bleeding and hemoglobin loss than older patients who received the anti�brinolytic in our study, the transfusion rates in the 2 groups were similar.

This was expected because most patients did not reach the trigger hemoglobin levels for allogenic transfusion according to standard protocols. A large sample will be required to explore whether tranexamic acid can reduce transfusion rates in older patients undergoing primary THA. In addition, the incidence of thromboembolism in our older group was higher than that previously

Fig. 2. Mean hemoglobin level before and after surgery in younger and older patients who received tranexamic acid. Error bars represent standard deviation.

Preoperative Postoperative day 1

Time point

Postoperative day 2 Postoperative day 3

150

140

130

120

110

100

90

Mea

n he

mog

lobi

n le

vel,

g/L

< 65 yr

≥ 65 yr

Table 4. Perioperative outcomes recorded in matched pairs of younger and older patients who received tranexamic acid

Outcome


p value†Age < 65 yrn = 58

Age ≥ 65 yrn = 58

Hematologic

Preoperative hemoglobin level, mean ± SD; g/L

135.58 ± 10.03 132.67 ± 11.31 0.5

Lowest postoperative hemoglobin level, mean ± SD; g/L

101.48 ± 10.41 99.79 ± 10.43 0.6

% decrease in hemoglobin level, mean ± SD

25.06 ± 6.50 24.65 ± 6.27 0.4

Estimated total blood loss, mean ± SD; mL

1281 ± 382 1208 ± 407 0.3

Packed erythrocytes transfusion 0 (0) 1 (2) 0.2

Total volume of packed erythrocytes given, mL

— 600 (2 units) 0.2

Postoperative recovery

Length of hospital stay, mean ± SD; d

5.30 ± 2.48 5.31 ± 2.82 0.3

Composite complication rate, % 0 6.9 0.02

Pulmonary embolism 0 (0) 2 (3) 0.08

Infection 0 (0) 2 (3) 0.08

Discharged to rehabilitation 2 (4) 12 (21) 0.004



†Paired t test.


RESEARCH


reported.30–32 This may have been due to the fact that our study was not powered to determine the prevalence of thromboembolism in older people undergoing total hip replacement.

Limitations

The current study has limitations. First, the cohorts were not matched for year of the procedure. However, there was no association between the extent of the primary out-comes of the study and the date of surgery in the entire cohort. Hence, there is no apparent sampling bias due to the lack of matching for surgical date. Second, the lowest postoperative hemoglobin concentration was that meas-ured during the hospital stay. The hemoglobin level may have drifted to lower values after discharge from hospital; hence, the reported decrease in hemoglobin level may not precisely re�ect the actual outcomes. However, the hemo-globin concentration would probably have reached a nadir before patients were discharged from the hospital, as is evi-dent from Figure 2. Finally, the decision to give tranexamic acid was partly based on the surgeon’s prefer-ence. This might have led to sampling bias, especially if surgeon expertise and technique are considered. However, all the surgeons who performed the procedures in this study had more than 10 years of experience, and their sur-gical technique was not signi�cantly different within the groups studied. Therefore, sampling bias was reduced to a minimum.

CONCLUSION

Tranexamic acid decreased perioperative blood loss and had a significant perioperative hemoglobin-sparing effect in older patients undergoing primary THA, simi-lar to that seen in younger patients. Hence, there is no need for adjustment of the tranexamic acid dosage in older patients. Tranex amic acid administration should be an important component of blood management pro-grams in geriatric patients undergoing hip replacement surgery.33

Acknowledgements: The authors are grateful to the nursing staff at the Department of Orthopedic Surgery, Trillium Health Partners, Mississauga Hospital for their invaluable help during data collection for the study.

Af�liations: From the Department of Anesthesia, Trillium Health Part-ners, Mississauga Hospital, Mississauga, Ont. (El Beheiry, Lubberdink); the Department of Orthopedic Surgery, Trillium Health Partners, Mis-sissauga Hospital, Mississauga, Ont. (Clements, Dihllon, Sharma); the Department of Anesthesia, University of Toronto, Toronto, Ont. (El Beheiry, Lubberdink); and the Department of Surgery, Division of Orthopedics, University of Toronto, Toronto, Ont. (Clements).


Contributors: H. El Beheiry and N. Clements designed the study. All authors acquired the data, which H. El Beheiry and A. Lubberdink ana-lyzed. H. El Beheiry wrote the article, which all authors reviewed and approved for publication.

References

1. Mahdy AM, Webster NR. Perioperative systemic haemostatic agents. Br J Anaesth 2004;93:842-58.

2. Katsumata S, Nagashima M, Kato K, et al. Changes in coagulation–�brinolysis marker and neutrophil elastase following the use of tour-niquet during total knee arthroplasty and the in�uence of neutrophil elastase on thromboembolism. Acta Anaesthesiol Scand 2005;49:510-6.

3. Sukeik M, Alshryda S, Haddad FS, et al. Systematic review and meta-analysis of the use of tranexamic acid in total hip replacement. J Bone Joint Surg Br 2011;93:39-46.

4. Gandhi R, Evans HMK, Mahomed SR, et al. Tranexamic acid and the reduction of blood loss in total knee and hip arthroplasty: a meta-analysis. BMC Res Notes 2013;6:184.

5. Xie J, Jun M, Yue C, et al. Combined use of intravenous and topical tranexamic acid following cementless total hip arthroplasty: a ran-domised clinical trial. Hip Int 2016;26:36-42.

6. Husted H, Blønd L, Sonne-Holme S, et al. Tranexamic acid reduces blood loss and blood transfusions in primary total hip arthroplasty: a prospective randomized double-blind study in 40 patients. Acta Orthop Scand 2003;74:665-9.

7. Lemay E, Guay J, Côté C, et al. Tranexamic acid reduces the need for allogenic red blood cell transfusions in patients undergoing total hip replacement. Can J Anaesth 2004;51:31-7.

8. Yamasaki S, Masuhara K, Fuji T. Tranexamic acid reduces blood loss after cementless total hip arthroplasty — prospective randomized study in 40 cases. Int Orthop 2004;28:69-73.

9. Kazemi SM, Mosaffa F, Eajazi A, et al. The effect of tranexamic acid on reducing blood loss in cementless total hip arthroplasty under epi-dural anesthesia. Orthopedics 2010;33:17.

10. Malhotra R, Kumar V, Garg B. The use of tranexamic acid to reduce blood loss in primary cementless total hip arthroplasty. Eur J Orthop Surg Traumatol 2011;21:101-4.

11. Imai N, Dohmae Y, Suda K, et al. Tranexamic acid for reduction of blood loss during total hip arthroplasty. J Arthroplasty 2012;27:1838-43.

12. Niskanen RO, Korkala OL. Tranexamic acid reduces blood loss in cemented hip arthroplasty: a randomized, double-blind study of 39 patients with osteoarthritis. Acta Orthop 2005;76:829-32.

13. Nicolle AL, Talks KL, Hanley JP. Congenital and acquired bleeding problems in elderly patients. Rev Clin Gerontol 2005;15:9-26.

14. Van PY, Schreiber MA. Hematologic issues in the geriatric surgical patient. Surg Clin North Am 2015;95:129-38.

15. Yang QJ, Jerath A, Bies RR. Pharmacokinetic modeling of tranex-amic acid for patients undergoing cardiac surgery with normal renal function and model simulations for patients with renal impairment. Biopharm Drug Dispos 2015;36:294-307.

16. Aymanns C, Keller F, Maus S, et al. Review on pharmacokinetics and pharmacodynamics and the aging kidney. Clin J Am Soc Nephrol 2010; 5:314-27.

17. Good L, Peterson E, Lisander B. Tranexamic acid decreases external blood loss but not hidden blood loss in total knee replacement. Br J Anaesth 2003;90:596-9.

18. Nadler SB. Prediction of blood volume in normal human adults. Surgery 1972;51:224-32.

19. Faul F, Erdfelder E, Lan AG, et al. G*Power 3: a �exible statistical power analysis program for the social, behavioral, and biomedical sci-ences. Behav Res Methods 2007;39:175-91.

20. Benoni G, BjÖrkman S, Fredin H. Application of pharmacokinetic data from healthy volunteers for the prediction of plasma concentrations of tranexamic acid in surgical patients. Clin Drug Invest 1995;10:280-7.

21. Niskanen RO, Korkala OL. Tranexamic acid reduces blood loss in cemented hip arthroplasty: a randomized, double-blind study of 39 patients with osteoarthritis. Acta Orthop 2005;76:829-32.

22. Hart A, Khalil J, Carli A, et al. Blood transfusion in primary total hip and knee arthroplasty. Incidence, risk factors and thirty-day compli-cation rates. J Bone Joint Surg Am 2014;96:1945-51.


RECHERCHE


23. Saleh A, Small TDO, Pillai ALPC, et al. Allogenic blood transfusion following total hip arthroplasty: results from the Nationwide Inpa-tient Sample, 2000 to 2009. J Bone Joint Surg Am 2014;96:e155.

24. Browne JA, Adib F, Brown TE, et al. Transfusion rates are increas-ing following total hip arthroplasty: risk factors and outcomes. J Arthroplasty 2013;28(Suppl 8):34-7.

25. Lee YC, Park SJ, Kim JS, et al. Effect of tranexamic acid on reducing postoperative blood loss in combined hypotensive epidural anesthesia and general anesthesia for total hip replacement. J Clin Anesth 2013;25:393-8.

26. Hourlier H, Fennema P. Single tranexamic acid dose to reduce peri-operative morbidity in primary total hip replacement: a randomised clinical trial. Hip Int 2014;24:63-8.

27. George DA, Sarraf KM, Nwaboku H. Single perioperative dose of tranexamic acid in primary hip and knee arthroplasty. Eur J Orthop Surg Traumatol 2015;25:129-33.

28. Wei W, Wei B. Comparison of topical and intravenous tranexamic acid on blood loss and transfusion rates in total hip arthroplasty. J Arthroplasty 2014;29:2113-6.

29. March GM, Elfatori S, Beaulé PE. Clinical experience with tranex-amic acid during primary total hip arthroplasty. Hip Int 2013;23: 72-9.

30. Poeran J, Rasul R, Suzuki S, et al. Tranexamic acid use and postoper-ative outcomes in patients undergoing total hip or knee arthroplasty in the United States: retrospective analysis of effectiveness and safety. BMJ 2014;349:g4829.

31. Whiting DR, Gillette BP, Duncan C, et al. Preliminary results suggest tranexamic acid is safe and effective in arthroplasty patients with severe comorbidities. Clin Orthop Relat Res 2014;472: 66-72.

32. Duncan CM, Gillette BP, Jacob AK, et al. Venous thromboembo-lism and mortality associated with tranexamic acid use during total hip and knee arthroplasty. J Arthroplasty 2015;30:272-6.

33. Hare GM, Freedman J, David Mazer C. Review article: risks of anemia and related management strategies: Can perioperative blood management improve patient safety? Can J Anaesth 2013;60: 168-75.

REPRINTSfrom the CMAJ GroupGoing to a conference? Teaching a course? Need extra copies?

Contact Sheridan Press1-800-635-7181 ext. [email protected]/cma/eoc

HOMELESS PEOPLE WITH MENTAL ILLNESSCost of services for homeless people with mental illness in 5 Canadian cities

2017

0006

Reprint-half-horizontal.pdf 1 2018-02-14 3:30 PM




Troponin T monitoring to detect myocardial injury after noncardiac surgery: a cost–consequence analysis

Background: Myocardial injury after noncardiac surgery (MINS) is a mostly asymptomatic condition that is strongly associated with 30-day mortality; however, it remains mostly undetected without systematic troponin T monitoring. We evalu-ated the cost and consequences of postoperative troponin T monitoring to detect MINS.

Methods: We conducted a model-based cost–consequence analysis to compare the impact of routine troponin T monitoring versus standard care (troponin T measure-ment triggered by ischemic symptoms) on the incidence of MINS detection. Model inputs were based on Canadian patients enrolled in the Vascular Events in Noncar-diac Surgery Patients Cohort Evaluation (VISION) study, which enrolled patients aged 45 years or older undergoing inpatient noncardiac surgery. We conducted prob-ability analyses with 10 000 iterations and extensive sensitivity analyses.

Results: The data were based on 6021 patients (48% men, mean age 65 [standard deviation 12] yr). The 30-day mortality rate for MINS was 9.6%. We determined the incremental cost to avoid missing a MINS event as $1632 (2015 Canadian dollars). The cost-effectiveness of troponin monitoring was higher in patient subgroups at higher risk for MINS, e.g., those aged 65 years or more, or with a history of athero-sclerosis or diabetes ($1309).

Conclusion: The costs associated with a troponin T monitoring program to detect MINS were moderate. Based on the estimated incremental cost per health gain, implementation of postoperative troponin T monitoring seems appealing, particularly in patients at high risk for MINS.

Contexte : Les lésions myocardiques après chirurgie non cardiaque (CNC) sont majoritairement asymptomatiques et fortement associées au risque de mortalité dans les 30 jours; toutefois, dans la plupart des cas, elles ne sont pas détectées en l’absence d’une surveillance systématique de la troponine T. Nous avons évalué les coûts et les conséquences d’une telle surveillance pour détecter les lésions myocardiques après CNC.

Méthodes : Nous avons mené une analyse coût–conséquence modélisée pour com-parer la surveillance systématique de la troponine T aux soins habituels seuls (mesure de la troponine T seulement s’il y a présence de symptômes d’ischémie) sur la fréquence de détection de lésions myocardiques après CNC. Les données ayant servi à l’analyse provenaient des patients canadiens ayant participé à l’étude de cohorte VISION, qui visait à évaluer les complications vasculaires chez les patients de 45 ans et plus ayant subi une CNC. Nous avons mené des analyses de probabilité avec 10 000 itérations et des analyses de sensibilité approfondies.

Résultats : Les données portaient sur 6021 patients (48 % du sexe masculin; âge moyen de 65 ans [écart-type de 12 ans]). Le taux de mortalité dans les 30 jours associé à une lésion myocardique après CNC était de 9,6 %. Nous avons déterminé que le coût marginal de la détection de la présence d’une lésion par surveillance de la tropo-nine T était de 1632 $ (dollars canadiens en 2015). Le rapport coût–ef�cacité était plus bas pour les sous-groupes de patients à risque élevé de lésion myocardique après CNC, comme les patients de 65 ans et plus ou ceux ayant des antécédents d’athéro-sclérose ou de diabète (1309 $), que pour leurs pairs.

Conclusion : Les coûts associés à un programme de surveillance de la troponine T pour détecter les lésions myocardiques après CNC étaient modérés. Le coût marginal estimé par gain de santé indique que la mise en œuvre de ce type de programme pour-rait être une option intéressante, surtout pour les patients à risque élevé de lésion myocardique après CNC.

Giovanna Lurati Buse, MD, MSc Braden Manns, MD, MSc Andre Lamy, MD Gordon Guyatt, MD, MSc Carisi A. Polanczyk, MD Matthew T.V. Chan, MD Chew Yin Wang, MBChB Juan Carlos Villar, MD, PhD Alben Sigamani, MD Daniel I. Sessler, MD Otavio Berwanger, MD Bruce M. Biccard, MBChB Rupert Pearse, MD Gerard Urrútia, MD R. Wojciech Szczeklik, MD, PhD Ignacio Garutti, MD, PhD Sadeesh Srinathan, MD, MSc German Malaga, MD, MSc Valsa Abraham, MD Clara K. Chow, MBBS, PhD Michael J. Jacka, MD, MSc Maria Tiboni, MD Gareth Ackland, MD, PhD Danielle Macneil, MD Robert Sapsford, MD Martin Leuwer, MD, PhD Yannick Le Manach, MD Philip J. Devereaux, MD, PhD

Accepted Nov. 17, 2017

Correspondence to: G. Lurati Buse Anesthesiology Department University Hospital of Düsseldorf Moorenstrasse 5 40225 Düsseldorf Germany [email protected]

DOI: 10.1503/cjs.010217

monitor-lurati.indd 185 2018-05-18 12:25 PM

RECHERCHE


A bout 500 000 noncardiac surgical procedures take place in Canada annually.1 In the Vascular Events in Noncardiac Surgery Patients Cohort Evaluation

(VISION) study, myocardial injury after noncardiac surgery (MINS) (defined as a peak troponin T level ≥ 0.03 ng/mL due to myocardial ischemia) was the most frequent vascular complication (8%).2 The crude 30-day mortality rate of MINS was 9.6%, and MINS was strongly associated with death. It may account for 34% of deaths within this period.2

Because only a small minority of patients who experi-ence MINS have symptoms,2 most cases go undetected without systematic postoperative troponin T monitoring. Monitoring for MINS by means of perioperative troponin T levels may offer an opportunity to intervene and poten-tially reduce subsequent adverse events. Although the optimal treatment for MINS remains unclear, promising observational data suggest mortality advantages in patients given treatments after MINS and, thus, that MINS is likely modifiable.3,4 Moreover, in the VISION study, patients with MINS who died did so a mean of 9 days after their initial troponin T level elevation, which indi-cates that there is time to initiate treatment after MINS is detected. Perhaps as a consequence, Canadian5 and inter-national6 guidelines recommend troponin monitoring after noncardiac surgery in patients at high cardiovascular risk.

The need to be judicial with resources requires con-sideration of both bene�t and cost of any intervention.7,8 Information on the resource and health implications of routine troponin T monitoring after noncardiac surgery is limited. The goal of this initial, basic model was to esti-mate, using different postoperative troponin T monitoring strategies, the cost and health consequences resulting from routine troponin T monitoring in patients with various levels of preoperative MINS risk undergoing noncardiac surgery, with a focus on the detection of MINS (that is, without burdening the model with assumptions with regard to treatment effect).

METHODS

The basis for these cost–consequence analyses was the VISION Study (clinicaltrials.gov, identifier NCT00512109). A previous report presents the details of enrolment and follow-up.2 Since the VISION study did not measure resource use, this cost–consequence analysis was model-based.

Population

The VISION study enrolled patients aged 45 years or more who underwent noncardiac surgery that required an overnight hospital stay and who had general or regional anesthesia. This analysis includes all Canadian patients

enrolled in the VISION study between September 2007 and October 2010 who had their troponin T level mea-sured with the fourth-generation (non–high-sensitive) assay (Fig. 1). We excluded patients who had an elevated troponin T level in the 7 days before surgery. The research ethics board at each site approved the protocol, and writ-ten informed consent was obtained from all patients.

Definition of myocardial injury after noncardiac surgery

We previously established the diagnostic criteria for MINS based on its prognostic impact on 30-day mortality.2 Details on the adjudication procedure have been previously published;2 in short, physicians evaluated extensive in-hospital documentation of all patients with troponin eleva-tion for ischemic features ful�lling the universal de�nition of myocardial infarction6 and for alternative nonischemic causes for increased troponin levels (i.e., sepsis, pulmonary embolism and cardioversion). Using Cox regression, we analyzed the association between alternative diagnostic cri-teria for MINS and 30-day mortality after adjustment for preoperative characteristics and perioperative complica-tions. Irrespective of the presence of ischemic symptoms or electrocardiographic changes, a peak troponin T level of 0.03 ng/mL or greater was independently associated with 30-day mortality (adjusted hazard ratio 3.87, 95% con�-dence interval 2.96–5.08). Therefore, MINS was de�ned as a troponin T value of 0.03 ng/mL or greater resulting from ischemia that occurs within 30 days after noncardiac surgery.2 Differentiation between type 1 and type 2 infarc-tion6 was not attempted.

For these analyses, we considered MINS events detected during the �rst 3 postoperative days (i.e., during the pro-posed troponin T monitoring period).

Model structure and computer simulation

We conducted a cost–consequence analysis from the per-spective of the Canadian health care system. We measured the health consequences as the number of detected MINS events during the monitoring period. We expressed costs in 2015 Canadian dollars.

In our base-case model, we compared fourth-generation troponin T monitoring 6–12 hours after surgery and on postoperative days 1, 2 and 3 with standard care (i.e., reliance on suggestive myocardial ischemic symptoms to trigger evaluation for potential MINS). The troponin level was not systematically measured preoperatively.

The model was structured as a decision tree (Fig. 2). It included the following health states: true-positive (detected MINS), true-negative (no MINS), false-negative (missed MINS) and false-positive. In the patients screened for ele-vation of the troponin T level, a false-positive health state was de�ned as troponin T elevation that was not due to


RESEARCH


myocardial ischemia (i.e., sepsis, pulmonary embolism or cardioversion). In the standard care alternative, false-positive referred to patients primarily assessed because of symptoms suggestive of myocardial ischemia (e.g., chest pain) but deemed to be of noncardiac origin after further investigation.

The reference analysis was a probabilistic sensitivity analysis that calculated the mean cost and the mean num-ber of detected events over 10 000 iterations generated by a second-order Monte Carlo simulation. We ran the model in Microsoft Excel spreadsheets with corresponding macros. We validated the model by extreme values and by the number of MINS cases estimated by the model against the primary data. Calibration of cost estimates was not possible because primary cost data were not available.

Data inputs

Data generated for 6021 Canadian VISION study patients informed the probabilities of the monitoring results. Since the VISION study did not include nonscreened patients,

we estimated the number of detected and missed cases of MINS after standard care using the number of symptom-atic MINS cases (assumed detected in standard care) and the number of asymptomatic MINS cases (assumed unde-tected in standard care). The VISION study did not collect information on clinical symptoms in patients without ele-vation of the troponin T level of 0.04 ng/mL or greater. We assumed the incidence of noncardiac chest pain (i.e., false-positive in the standard-care group) to be 1% and explored the impact of this assumption in sensitivity analysis. We opted for this very conservative estimate of false-positive health state to avoid any overestimation of the cost in the standard-care group. Table 1 summarizes the model parameters and their distributions.

The VISION study did not collect data on resource use except for coronary angiography. We estimated the cost of the 2 alternatives based on predefined diagnostic algo-rithms to con�rm or exclude MINS in the case of elevated troponin T levels in screened patients or suggestive symp-toms in the standard-care group. The algorithms included fourth-generation troponin T measurements as monitoring

Fig. 1. Flow chart showing selection of study population. Note: VISION = Vascular Events in Noncardiac Surgery Patients Cohort Evaluation.

Canadian patients enrolled in VISION studyn = 6314

Canadian patients who fulfilled VISION eligibility criteria

n = 8981

Patients included in economic analysis of postoperative troponin T screening

n = 6021

2667 patients (29.7%) not enrolled:

• 1801 (67.5%) declined participation• 122 (4.6%) not identified before or within 24 h of surgery• 5 physicians (0.2%) declined participation• 739 (27.7%) other reasons

293 (4.6%) excluded from economic analysis of postoperative troponin T screening:

• 114 (1.8%) troponin T level not quantified below 0.04 ng/mL• 25 (0.4%) preoperative troponin T elevation• 154 (2.4%) all scheduled troponin T measurements missing


RECHERCHE


and as serial follow-up in the case of elevated troponin T or of clinical symptoms (triggered troponin T measure-ments), a cardiology consultation and follow-up visits, and serial electrocardiography and echocardiography for all patients with elevated troponin T levels. Canadian VISION data were the source for the probability of cor-onary angiography and for the probability that the tropo-nin T level exceeded the upper limit of normal as late as on the last day of scheduled measurement, thus resulting in additional, triggered troponin T measurements in screened patients. The model did not include costs related to the noncardiac surgical procedure itself because these resource items were identical in the 2 alternatives.12 Costs were inflated by the Canadian health care Consumer Price Index as necessary.

Sensitivity and scenario analyses

We explored the impact of the assumptions on resource use and costs resulting from the predefined diagnostic algorithms using sensitivity analyses. The worst-case scen-ario maximized the incremental costs of the monitoring program by assuming simultaneously a 25% cost increase in the screened patients and a 25% cost reduction for stan-dard care. The best-case scenario reduced the incremental cost of the monitoring program by assuming the inverse.

We also assessed the impact of the assumption for all patients with elevated troponin T levels to undergo echo-cardiography by assuming that 50% and 75% of such patients would have echocardiography.

We assessed the impact of varying the false-positive rate in the standard-care patients (imputing 0% for sensi-tivity analyses and assuming 1% for the reference case). In the overall VISION population (i.e., not limited to Canadian centres), nonischemic causes for troponin ele-vation were adjudicated in 0.59% of patients (95/16 087),2 as opposed to the 0.37% assumed in the current analysis (based on 22 cases of nonischemic tropo-nin elevation in 6021 Can adian patients); therefore, we also ran a sensitivity analysis imputing 0.59% false- positive in the monitoring alternative.

To assess the cost-effectiveness of troponin T monitor-ing in populations at various levels of risk for MINS, we analyzed the following subgroups: 1) age 65 years or more, 2) urgent/emergent surgery, 3) history of coronary artery disease and 4) age 65 years or more, or history of coronary artery disease, peripheral vascular disease, cerebrovascular event or diabetes.

Patients experience MINS at different times after the procedure. Therefore, varying timelines after the proced-ure and increasing numbers of troponin T measurements will detect events that had not yet occurred at the time of

Fig. 2. Decision tree representing the alternatives and the health states at the end of the monitoring period. Note: MINS = myocardial injury after noncardiac surgery.

Troponin T screening,4 measurements

Standard care

True-negative: no MINS

True-negative: no MINS

True-positive: MINS

True-positive: MINS with ischemic symptoms

False-negative: MINS with negativescheduled troponin T measurements

False-negative: MINS without ischemic symptoms

False-positive: symptoms without MINS

False-positive: nonischemic troponin T elevation


RESEARCH


the previous measurement. This represents a major differ-ence from screening programs for static disease (e.g., cancer), where a marginal approach is used for increasing the number of tests. To assess the impact of using different monitoring strategies, we estimated the cost and number of cases of MINS detected using monitoring based on 1) a single troponin T measurement 6–12 hours

postoperatively, 2) measurements 6–12 hours postopera-tively and on postoperative day 1, 3) measurements 6–12 hours postoperatively and on postoperative days 1 and 2, and 4) measurement only on postoperative days 1 and 2. In the base case, the troponin T level was measured 6–12 hours postoperatively and daily up to the third post-operative day (VISION study protocol).

Table 1. Parameters and their distributions of the model of troponin T monitoring on the detection of myocardial injury after noncardiac surgery

Parameter Point estimate Distribution α/β or SE Source

Health state

True-negative troponin T monitoring 91.18% b 5490/531 VISION study2

True-positive troponin T monitoring 8.39% b 505/5516 VISION study2

False-negative troponin T monitoring 0.07% b 4/6017 VISION study2

False-positive troponin T monitoring 0.37% b 22/5999 VISION study2

True-negative standard care 90.63% b 5457/564 Residual

True-positive standard care 1.52% b 91/5920 VISION study2†

False-negative standard care 6.94% b 418/5603 VISION study2†

False-positive standard care 0.92% b 55/5966 Expert-based

Cost parameters*

Electrocardiography $11.05 NA — Ontario Health Insurance Plan Schedule of Benefits9

Echocardiography $208.80 NA — Ontario Health Insurance Plan Schedule of Benefits9

Troponin T measurement (per measurement) $18.00 NA — Laboratory Reference Centre affiliated with Hamilton Health Sciences

Cardiologist consultation $157 NA — Ontario Health Insurance Plan Schedule of Benefits10

Cardiologist partial assessment $31 NA — Ontario Health Insurance Plan Schedule of Benefits10

Coronary angiography $2903.99 NA — Clement et al.,11 CanadianConsumer Price Index

Probability that troponin T level would exceed upper limit of normal in last scheduled measurement

15.23% b 46/256 VISION study2 (patients with all 4 measurements)

Probability of angiography in patients who experience MINS

0.99% b 5/500 VISION study2

Probability of angiography in patients with symptomatic MINS


Probability of angiography in patients with false-positive troponin T monitoring


Probability of angiography in patients with false-negative troponin T monitoring

25% b 1/3 VISION study2

Cost of troponin monitoring true-negative $72.00 Log normal 3.60 Based on predefined diagnostic algorithms (see Methods)

Cost of troponin monitoring true-positive $597.24 Log normal 27.39 Based on predefined diagnostic algorithms (see Methods)

Cost of troponin monitoring false-negative $1276.32 Log normal 41.73 Based on predefined diagnostic algorithms (see Methods)

Cost of troponin monitoring false-positive $605.75 Log normal 18.44 Based on predefined diagnostic algorithms (see Methods)

Cost of standard care true-negative $0.00 NA — Based on predefined diagnostic algorithms (see Methods)

Cost of standard care true-positive $734.27 Log normal 29.27 Based on predefined diagnostic algorithms (see Methods)

Cost of standard care false-negative $0.00 NA — Based on predefined diagnostic algorithms (see Methods)

Cost of standard care false-positive $58.10 Log normal 3.47 Based on predefined diagnostic algorithms (see Methods)

MINS = myocardial injury after noncardiac surgery; NA = not applicable; SE = standard error; VISION = Vascular Events in Noncardiac Surgery Patients Cohort Evaluation.

*In 2015 Canadian dollars.

†Proportion of cases of ischemic troponin T elevation; i.e., including false-negative troponin T monitoring.


RECHERCHE


RESULTS

Just less than half of the 6021 Canadian VISION study patients were men (2886 [47.9%]) and were aged 65 years or more (2975 [49.4%]). Table 2 shows the patients’ demographic characteristics, preoperative risk factors, type of surgery.

Reference model

Table 3 summarizes the resources used in the various health states according to the diagnostic algorithms. The unit costs and their sources are shown in Table 2.

The incremental cost of a monitoring program consisting of 4 troponin T measurements was $112.18 per screened patient, including follow-up diagnostic tests and consulta-tions in patients with elevated troponin T levels (Table 4). The incremental cost per additional case of MINS detected was $1632.51. The distribution of the simulations in the incremental cost-effectiveness plan is shown in supplemen-

tary Fig. S1, Appendix 1 (available at canjsurg.ca/010217-a1). Under the assumption of 500 000 noncardiac surgical pro-cedures per year in Canada,1 the annual cost of a MINS monitoring program would amount to $56.1 million, with the incremental detection of 34 354 MINS events. The absolute 30-day mortality rate among patients with MINS was 9.6% (95% con�dence interval 8.0–11.4).2

Sensitivity and scenario analyses

The incremental cost to detect an additional case of MINS rose to $2138 under the worst-case scenario and was $1134 in the best-case scenario. Table 5 shows the results of other sensitivity analyses.

The incremental cost per additional case detected was lower in patients at higher risk for MINS (Table 6; Sup-plementary Fig. 2, Appendix 1).

Myocardial injury after noncardiac surgery occurred in 34.6% of patients within the �rst 6–12 hours after surgery, in 22.9% on the �rst postoperative day, in 26% on the

Table 2. Baseline characteristics overall and by health state at the end of the monitoring period of Canadian VISION patients2

Characteristic

Health state; no. (%) of patients*

All n = 6021

True-negative troponin T monitoring n = 5490

True-positive troponin

monitoring n = 505

False-negative troponin T monitoring n = 4

False-positive troponin T monitoring n = 22

Age, mean ± SD; yr 65 ± 12 64 ± 11 72 ± 12 74 ± 7 73 ± 10

Age ≥ 65 yr 2975 (49.4) 2593 (47.2) 359 (71.1) 4 (100) 19 (86)

Age ≥ 75 yr 1415 (23.5) 1154 (21.0) 248 (49.1) 2 (50) 11 (50)

Male sex 2886 (47.9) 2583 (47.0) 288 (57.0) 3 (75) 12 (54)

History of congestive heart failure 191 (3.2) 135 (2.4) 52 (10.3) 0 (0) 4 (18)

History of coronary artery disease 1005 (16.7) 814 (14.8) 184 (36.4) 2 (50) 5 (23)

Current atrial fibrillation 192 (3.2) 147 (2.7) 39 (7.7) 0 (0) 6 (27)

History of cerebrovascular event 428 (7.1) 345 (6.3) 77 (15.2) 0 (0) 6 (27)

History of peripheral vascular disease 292 (4.8) 211 (3.8) 75 (14.8) 1 (25) 5 (23)

History of hypertension 3241 (53.8) 2851 (51.9) 368 (72.9) 4 (100) 18 (82)

History of diabetes 1119 (18.6) 951 (17.3) 158 (31.3) 1 (25) 9 (41)

Urgent/emergent surgery† 607 (10.1) 514 (9.4) 87 (17.2) 1 (25) 5 (23)

Vascular surgery‡ 240 (4.0) 196 (3.6) 43 (8.5) 0 (0) 1 (4)

General surgery§ 1006 (16.7) 897 (16.3) 100 (19.8) 2 (50) 7 (32)

Major urogynecologic surgery¶ 745 (12.4) 677 (12.3) 66 (13.1) 0 (0) 2 (9)

Major orthopedic surgery** 1649 (27.4) 1481 (27.0) 161 (31.9) 2 (50) 5 (23)

Neurosurgery†† 386 (6.4) 362 (6.6) 24 (4.8) 0 (0) 0 (0)

Low-risk surgery‡‡ 1924 (32.0) 1809 (33.0) 109 (21.6) 0 (0) 6 (27)

SD = standard deviation; VISION = Vascular Events in Noncardiac Surgery Patients Cohort Evaluation.


†Procedure performed within 72 hours of the surgery-triggering acute event.

‡Included thoracic aorta or aortoiliac reconstructive procedures, peripheral vascular reconstruction without aortic cross-clamping, extracranial cerebrovascular surgery and endovascular abdominal aortic aneurysm repair.

§Included complex visceral resection, partial or total colectomy or stomach surgery, other intra-abdominal surgery, and major head and neck resection for tumour.

¶Included visceral resection (e.g., nephrectomy, ureterectomy, bladder resection, retroperitoneal tumour resection, radical procedure for cancer [i.e., exenteration], hysterectomy).

**Included major hip or pelvis surgery (hemi or total hip arthroplasty, internal fixation of hip, pelvic arthroplasty).

††Included craniotomy and spine surgery involving multiple levels of the spine.

‡‡Included parathyroid, thyroid, breast, hernia or local anorectal procedure, radical prostatectomy, transurethral prostatectomy, oopherectomy, salpingectomy, endometrial ablation, peripheral nerve surgery, ear/nose/throat surgery, vertebral disc surgery, spinal fusion, knee arthroplasty, hand surgery, cosmetic surgery, arteriovenous access surgery for dialysis, other surgery not fulfilling the major criteria as above.


RESEARCH


Table 3. Resource use according to predefined diagnostic algorithms by health state

Health state

Resource; no. of uses

Cardiologist consultation

Cardiologist partial

assessment Electrocardiography Echocardiography

Scheduled troponin T

measurement

Triggered troponin T

measurement Angiography

True-negative troponin T monitoring

0 0 0 0 4 0 0

True-positive troponin T monitoring

1 3 4 1 4 1 × probability of first elevation of troponin T level

occurring on postoperative

day 3*

1 × probability of angiography after

MINS

False-negative troponin T monitoring

1 3 2 1 4 1 1 × probability of angiography in false-negative

troponin T monitoring

False-positive troponin T monitoring

1 0 4 1 4 1 × probability of first elevation of troponin T level

occurring on postoperative

day 3*

1 × probability of angiography in false-positive

troponin T monitoring

True-negative standard care

0 0 0 0 0 0 0

True-positive standard care

1 3 4 1 0 2 1 × probability of angiography after

symptomatic MINS

False-negative standard care

0 0 0 0 0 0 0

False-positive standard care

0 0 2 0 0 2 0

MINS = myocardial injury after noncardiac surgery.

*Last day of scheduled troponin T measurement.

Table 4. Cost and number of detected cases of myocardial injury after noncardiac surgery with troponin T monitoring and with standard care

VariableTroponin T monitoring Standard care

Incremental cost, $*

Incremental cost per additional case of MINS detected*

Cost, $ 123.87 11.70 112.18 1632.51

No. of cases of MINS detected

0.084 0.015 0.069 —


*2015 Canadian dollars.

Table 5. Scenario analyses assessing the impact of varying costs and false-positive rates in the standard care alternative on the cost-effectiveness estimates

VariableIncremental cost per patient screened, $*

Incremental cost per additional case of MINS detected, $*

Reference case 112.18 1632.51

25% cost increase in monitoring and 25% cost reduction in standard care alternative (worst case)

146.81 2137.70

25% cost reduction in monitoring and 25% cost increase in standard care alternative (best case)

77.88 1134.28

50% of patients with elevated troponin T levels undergo echocardiography

102.74 1495.28

75% of patients with elevated troponin T levels undergo echocardiography

107.38 1564.85

0% false-positive in standard care 112.58 1639.39

0.59% false-positive with monitoring 113.58 1650.16




RECHERCHE


second postoperative day and in 16.5% on the third post-operative day. A monitoring protocol that measured the troponin T level 6–12 hours after surgery and daily on postoperative days 1, 2 and 3 resulted in the lowest incre-mental costs for detected MINS (Table 7).

DISCUSSION

This analysis suggests that the incremental cost to avoid missing a MINS event through troponin T monitoring after noncardiac surgery in unselected patients aged 45 years or more would be less than $1650. The estimated incremental cost to detect an additional case of MINS was less than $1350 in selected populations (e.g., patients aged ≥ 65 yr, those undergoing urgent/emergent surgery and those with a history of cardiovascular disease).

Among the screening protocols tested, monitoring con-sisting of troponin T measurements 6–12 hours after sur-gery and on postoperative days 1, 2, and 3 resulted in the lowest incremental costs per case of MINS detected. Under the assumption of an annual surgical volume of 500 000 inpatient noncardiac surgical procedures in Can-ada,1 a budget of around $56 million would allow physi-cians to identify over 34 000 additional MINS cases.

Mantha and colleagues13 evaluated the cost-effectiveness of a postoperative troponin T monitoring strategy to initi-

ate heart rate control and surveillance in a coronary care unit after abdominal aortic aneurysm repair. They popu-lated their model based on data from the literature and assumed a hypothetical relative risk of 0.55 for myocardial ischemia using the strategies mentioned. They concluded that troponin T monitoring after abdominal aortic aneur-ysm repair was cost-effective (US$12 641 per quality-adjusted life-year [QALY]).

Torborg and colleagues14 conducted an economic analysis of perioperative troponin monitoring after non-cardiac surgery in South Africa. They assumed a 25% reduction in 30-day rates of cardiovascular mortality and myocardial infarction after initiation of treatment with acetylsalicylic acid and statins in patients with positive screening results. The monitoring alternative dominated standard care, i.e., it prevented 30-day adverse events at lower cost.

Our model was populated by a large cohort of patients undergoing a broad spectrum of noncardiac surgical pro-cedures, and it was limited neither by the use of QALY not speci�c for postoperative events nor by assumptions of a hypothetical treatment effect. In spite of these differ-ences in methods, all evaluations suggest that there may be health gains achievable by troponin T monitoring after noncardiac surgery within commonly applied ceiling ratios.

Table 6. Cost-effectiveness ratio, cost and incremental number of myocardial injury after noncardiac surgery events detected in populations at various risk

Population

Incremental cost per case of MINS

detected, $* Annual volume Cost, $ millions*

No. of incremental cases of MINS

detected

Age ≥ 45 yr 1633 500 000 56.1 34 354

Age ≥ 65 yr 1337 247 000 31.7 23 753

History of coronary artery disease 1084 83 500 12.3 11 359

Urgent/emergent surgery 1192 50 500 7.3 6087

Age ≥ 65 yr or history of coronary artery disease, peripheral vascular disease, cerebrovascular event or diabetes

1309 280 500 36.9 28 171



Table 7. Cost-effectiveness ratio, cost and incremental number of myocardial injury after noncardiac surgery events detected with various troponin T monitoring alternatives

Timing of troponin T measurement

Incremental cost per case of MINS detected,

$* Cost, $ millions*No. of incremental cases of

MINS detected

6–12 h postoperatively and postoperative days 1, 2 and 3

1633.00 56.1 34 354

6–12 h postoperatively 14 248.00 46.8 3283

6–12 h postoperatively and postoperative day 1

4429.00 51.0 11 518

6–12 h postoperatively and postoperative days 1 and 2

2599.55 52.8 20 313




RESEARCH


Strengths and limitations

Our study’s strengths include the large, representative, contemporary sample of patients (> 6000, broad inclusion criteria) undergoing noncardiac surgery in Canada. In addition to assessing the cost per additional case of MINS detected, we also assessed affordability (i.e., the cost associ-ated with a troponin T monitoring program).15 Finally, the results are not limited by extensive assumptions regarding monitoring and treatment effect or by long-term extrapo-lations; rather, they rely closely on the observed data.

This cost–consequence analysis has limitations. The VISION study did not collect data on resource use. To estimate resource use in the presence of elevated troponin T levels, we applied diagnostic algorithms both for tropo-nin T monitoring and for standard care, based on expert opinion that was not generated in a Delphi panel. We addressed this limitation by conducting scenario analyses that varied the cost of the monitoring and standard-care alternatives: even in the worst-case scenario, the incremen-tal cost to detect an additional case of MINS remained moderate (< $2200).

Furthermore, given that the VISION study did not have a standard-care alternative, the model relied on the following assumptions with regard to the health states. First, we assumed that the proportion of MINS cases detected and missed by clinical assessment corresponded to the symptomatic and asymptomatic cases, respectively, of MINS. Second, because ischemic symptoms in patients with normal troponin T levels were not collected in the VISION study, we estimated the proportion of false-positive �ndings (i.e., noncardiac chest pain) with standard care based on expert opinion. However, our sensitivity analyses suggest a limited impact of this parameter, and we opted for a conservative estimate of false-positive �ndings in the standard care alternative. As such, the results pre-sented here underestimate the actual cost of standard care. In other words, the model overestimated the cost of tropo-nin T monitoring after noncardiac surgery.

The source for estimation of coronary angiogram cost included both in-hospital and outpatient angiography. The model did not take into account whether coronary angiog-raphy in patients with MINS was performed during the hospital stay for noncardiac surgery or after discharge (plausible in the case of asymptomatic patients). However, coronary angiography was done in less than 1% of patients with MINS; as such, the impact on the overall cost of tro-ponin T monitoring was minor. Furthermore, because more cases of MINS were detected with the troponin T monitoring alternative than with standard care, any poten-tial impact of uncertainty with regard to angiography cost would have led to overestimation of the cost of troponin T monitoring.

Finally, the VISION study measured troponin T in all patients; thus, it did not provide estimates of the effect of

troponin T monitoring on outcomes. There is no evidence from randomized controlled trials of the effectiveness of troponin T monitoring after noncardiac surgery on out-comes important to patients, and the optimal treatment of MINS has not been established. Therefore, we opted for a cost–consequence analysis with detected MINS as the health consequence of interest rather than assuming a treatment effect to estimate gain in survival or QALYs. However, this fact can also be considered as a strength because the model did not require extended assumptions or extrapolations. Furthermore, modelling guidelines12,16 agree that lack of data — in this case, evidence from ran-domized controlled trials to establish treatment for MINS — should not prevent economic evaluations.16 A large randomized trial examining MINS treatment has recently been completed.17

CONCLUSION

More than 500 000 Canadians undergo inpatient noncar-diac surgery annually, and around 40 000 patients will experience MINS, a condition strongly associated with 30-day mortality. The condition remains undetected in more than 4 in 5 patients (i.e., about 34 000 patients yearly in Canada) because of the lack of ischemic symptoms. Based on the estimated incremental cost per health gain, the implementation of a troponin T monitoring program after noncardiac surgery seems appealing.Af�liations: From the University Hospital of Düsseldorf, Düsseldorf, Germany (Lurati Buse); the University Hospital of Basel, Basel, Switzer-land (Lurati Buse); the University of Calgary, Calgary, Alta. (Manns); McMaster University, Hamilton, Ont. (Lamy, Guyatt, Tiboni, Devereaux); the Hospital de Clínicas de Porto Alegre, Universidade Federal de Rio Grande do Sul, Brazil (Polanczyk); the Chinese Univer-sity of Hong Kong, Sha Tin, NT, Hong Kong (Chan); the University of Malaya, Kuala Lampur, Malaysia (Wang); the Fundación Cardioinfan-til – Instituto de Cardiología, Bogotá and Universidad Autónoma de Buca ramanga, Bucaramanga, Colombia (Villar); St. John’s Medical Col-lege and Research Institute, Bangalore, India (Sigamani); the Depart-ment of Outcomes Research, Cleveland Clinic, Cleveland, Ohio (Sessler); the HCor Research Institute (Hospital do CoraÇão), São Paulo, Brazil (Berwanger); the Groote Schuur Hospital and University of Cape Town, Cape Town, South Africa (Biccard); Barts and The London School of Medicine and Dentistry, London, UK (Pearse); the Hospital de Sant Pau, Barcelona, Spain (Urrútia); the Jagiellonian University Medical College, Krakow, Poland (Szczeklik); the Hospital General Uni-versitario Gregorio Marañón, Madrid, Spain (Garutti); the University of Manitoba, Winnipeg, Man. (Srinathan); the Universidad Peruana Cayet-ano Heredia, Lima, Peru (Malaga); the Christian Medical College, Lud-hiana, India (Abraham); the George Institute for Global Health, Univer-sity of Sydney, Sydney, Australia (Chow); the University of Alberta Hospital, Edmonton, Alta. (Jacka); the University College London, Lon-don, UK (Ackland); the London Health Sciences Centre, London, Ont. (Macneil); the University of Leeds, Leeds, UK (Sapsford); the Royal Liv-erpool Broadgreen University Hospital Trust, Liverpool, UK (Leuwer); and the HÔpital Pitié-Salpêtrière, Paris, France (Le Manach).

Funding: Canadian Institutes of Health Research (6 grants) (Ottawa, Ontario, Canada); Heart and Stroke Foundation of Ontario (2 grants) (Toronto, Ontario, Canada); Academic Health Science Centre Alterna-tive Funding Plan Innovation Fund grant (Toronto, Ontario, Canada); Population Health Research Institute grant (Hamilton, Ontario, Canada); Clarity Research Group grant; Surgical Associates Research Grant, Department of Surgery, McMaster University (Hamilton, Ontario,


RECHERCHE


Canada); Hamilton Health Sciences New Investigator Fund grant (Ham-ilton, Ontario, Canada); Hamilton Health Sciences grant (Hamilton, Ontario, Canada); Ontario Ministry of Resource and Innovation grant (Toronto, Ontario, Canada); Stryker Canada (Waterdown, Ontario, Can-ada); Department of Anesthesia, McMaster University (2 grants) (Hamil-ton, Ontario, Canada); Department of Medicine, Saint Joseph’s Health-care (2 grants) (Hamilton, Ontario, Canada); Father Sean O’Sullivan Research Centre (2 grants) (Hamilton, Ontario, Canada); Department of Medicine, McMaster University (2 grants) (Hamilton, Ontario, Canada); Roche Diagnostics Global Of�ce (3 grants) (Basel, Switzerland); Hamil-ton Health Sciences Summer Studentships (6 grants) (Hamilton, Ontario, Canada); Department of Clinical Epidemiology and Biostatistics (now Health Research Methods, Evidence, and Impact) grant, McMaster Uni-versity (Hamilton, Ontario, Canada); Division of Cardiology grant, McMaster University (Hamilton, Ontario, Canada); Canadian Network and Centre for Trials Internationally grant (Hamilton, Ontario, Canada); Winnipeg Health Sciences Foundation Operating Grant (Winnipeg, Manitoba, Canada); Department of Surgery Research Grant, University of Manitoba (2 grants) (Winnipeg, Manitoba, Canada); Diagnostic Services of Manitoba Research Grant (2 grants) (Winnipeg, Manitoba, Canada); Manitoba Medical Services Foundation grant (Winnipeg, Manitoba, Canada); Manitoba Health Research Council grant (Winnipeg, Manitoba, Canada); Faculty of Dentistry Operational Fund grant, University of Manitoba (Winnipeg, Manitoba, Canada); Department of Anesthesia (now Department of Anesthesia and Perioperative Medicine) grant, Uni-versity of Manitoba (Winnipeg, Manitoba, Canada); University Medical Group Start-up Fund, Department of Surgery, University of Manitoba (Winnipeg, Manitoba, Canada), Fellowship for prospective researchers, Swiss National Science Foundation (Bern, Switzerland)

Competing interests: Roche Diagnostics provided the troponin T assays and �nancial support for the VISION study. P. Devereaux has received other funding from Roche Diagnostics and Abbott Diagnostics for investigator-initiated studies. No other competing interests declared.

Contributors: G. Lurati Buse, A. Lamy, G. Guyatt, C. Polanczyk, O. Berwanger, V. Abraham, Y. Le Manach and P. Devereaux designed the study. G. Lurati Buse, C. Polanczyk, M. Chan, C.Y. Wang, J. Villar, A. Sigamani, O. Berwanger, B. Biccard, R. Pearse, G. Urrútia, I. Garutti, S. Srinathan, G. Malaga, V. Abraham, M. Jacka, M. Tiboni, G. Ackland, D. Macneil, R. Sapsford and M. Leuwer acquired the data, which G. Lurati Buse, B. Manns, A. Lamy, G. Guyatt, C. Polanczyk, C.Y. Wang, D. Sessler, R.W. Szczeklik, S. Srinathan, G. Malaga, C. Chow, Y. Le Manach and P. Devereaux analyzed. G. Lurati Buse, C. Polanczyk and V. Abraham wrote the article, which all authors reviewed and approved for publication.

References

1. Weiser TG, Regenbogen SE, Thompson KD, et al. An estimation of the global volume of surgery: a modelling strategy based on available data. Lancet 2008;372:139-44.

2. Botto F, Alonso-Coello P, Chan MT, et al.; Vascular events In non-cardiac Surgery patIents cOhort evaluatioN (VISION) Writing Group, on behalf of The Vascular events In noncardiac Surgery patIents cOhort evaluatioN (VISION) Investigators. Myocardial injury after noncardiac surgery: a large, international, prospective cohort study establishing diagnostic criteria, characteristics, predictors, and 30-day outcomes. Anesthesiology 2014;120:564-78.

3. Devereaux PJ, Xavier D, Pogue J, et al. Characteristics and short-term prognosis of perioperative myocardial infarction in patients undergoing noncardiac surgery: a cohort study. Ann Intern Med 2011; 154:523-8.

4. Foucrier A, Rodseth R, Aissaoui M, et al. The long-term impact of early cardiovascular therapy intensi�cation for postoperative tropo-nin elevation after major vascular surgery. Anesth Analg 2014;119: 1053-63.

5. Duceppe E, Parlow J, MacDonald P, et al. Canadian Cardiovascular Society Guidelines on Perioperative Cardiac Risk Assessment and Management for Patients Who Undergo Noncardiac Surgery. Can J Cardiol 2017;33:17-32.

6. Thygesen K, Alpert JS, Jaffe AS, et al. Third universal de�nition of myocardial infarction. Eur Heart J 2012;33:2551-67.

7. Eddy DM. Comparing bene�ts and harms: the balance sheet. JAMA 1990;263:2493-8.

8. Wilson JMJ, Jungner G. Principles and practice of screening for disease. Public Health Papers 34. Geneva: World Health Organization; 1968.

9. Schedule of bene�ts: physician services under the Health Insurance Act. Section J. Diagnostic and therapeutic procedures. Ontario Minis-try of Health and Long-Term Care. 2015. Available: www.health.gov.on.ca/en/pro/programs/ohip/sob/physserv/physserv_mn.html (ac -cessed 2018 May 2).

10. Schedule of bene�ts: physician services under the Health Insurance Act. Section A. Consultations and visits. Ontario Ministry of Health and Long-Term Care. 2015. Available: www.health.gov.on.ca/en/pro/programs/ohip/sob/physserv/physserv_mn.html (accessed 2018 May 2).

11. Clement FM, Ghali WA, Rinfret S, et al.; APPROACH Investiga-tors. Economic evaluation of increasing population rates of cardiac catheterization. BMC Health Serv Res 2011;11:324.

12. Guidelines for the economic evaluation of health technologies. 3rd ed. Ottawa: Canadian Agency for Drugs and Technologies in Health (CADTH); 2006.

13. Mantha S, Foss J, Ellis JE, et al. Intense cardiac troponin surveillance for long-term bene�ts is cost-effective in patients undergoing open abdominal aortic surgery: a decision analysis model. Anesth Analg 2007;105:1346-56.

14. Torborg A, Ryan L, Kantor G, et al. The pharmacoeconomics of routine postoperative troponin surveillance to prevent and treat myo-cardial infarction after non-cardiac surgery. S Afr Med J 2014;104: 619-23.

15. Drummond M, Brown R, Fendrick AM, et al.; ISPOR Task Force. Use of pharmacoeconomics information — report of the ISPOR Task Force on use of pharmacoeconomic/health economic information in health-care decision making. Value Health 2003;6:407-16.

16. Weinstein MC, O’Brien B, Hornberger J, et al.; ISPOR Task Force on Good Research Practices — Modeling Studies. Principles of good practice for decision analytic modeling in health-care evaluation: report of the ISPOR Task Force on Good Research Practices — Modeling Studies. Value Health 2003;6:9-17.

17. Duceppe E, Yusuf S, Tandon V, et al. Design of a randomized placebo-controlled trial to assess dabigatran and omeprazole in patients with myocardial injury after noncardiac surgery (MANAGE). Can J Cardiol 2018;34:295-302.




Clinical and operative outcomes of patients with acute cholecystitis who are treated initially with image-guided cholecystostomy

Background: Percutaneous cholecystostomy (PC) tube placement followed by delayed cholecystectomy has been shown to be an effective treatment option in high-risk populations such as older and critically ill patients. The goal of this study was to review the short -and long-term clinical and operative outcomes of patients with acute cholecystitis initially treated with PC tube placement.

Methods: We conducted a retrospective review of patients who underwent image-guided PC tube insertion between 2001 and 2011 at the Royal University Hospital or St. Paul’s Hospital, Saskatoon. Clinical outcomes, complications and elective chole-cystectomy follow-up were noted.

Results: A total of 140 patients underwent PC tube insertion, 76 men and 64 women with a mean age of 68.4 (standard deviation 17.7) years. Of the 140, 94 (67.1%) had an American Society of Anesthesiologists classi�cation score of III or IV. Percutaneous cholecystostomy tubes remained in place for a median of 21.0 days, and the median hospital stay was 7.0 days. Readmission owing to complications from PC tubes occurred in 21 patients (15.0%), and 10 (7.1%) were readmitted with recurrent chole-cystitis after tube removal. Forty-four patients (31.4%) returned for subsequent elec-tive cholecystectomy, of whom 32 (73%) underwent laparoscopic cholecystectomy, 4 (9%) underwent open cholecystectomy, and 8 (18%) underwent laparoscopic con-verted to open cholecystectomy.

Conclusion: Percutaneous cholecystostomy is a safe procedure that can be performed in patients who are older or have numerous comorbidities. However, less than one-third of such patients in our cohort subsequently had the de�nitive intervention of elective chole-cystectomy, with a high rate of conversion from laparoscopic to open cholecystectomy.

Contexte : Il a été démontré que la pose d’un drain de cholécystostomie percutanée suivie d’une cholécystectomie tardive serait une option thérapeutique ef�cace chez les populations à risque élevé, comme les patients âgés et gravement malades. L’objectif de cette étude était de revoir l’issue clinique et chirurgicale à court et à long terme chez les patients ayant présenté une cholécystite aiguë traitée par cholécystostomie percutanée.

Méthodes : Nous avons procédé à une revue rétrospective des patients ayant subi une cholécystostomie percutanée guidée à l’aide de l’imagerie entre 2001 et 2011 à l’Hôpital royal universitaire ou à l’Hôpital St. Paul de Saskatoon. Nous avons ensuite pris note de l’issue clinique, des complications et des cholécystectomies non urgentes subséquentes.

Résultats : En tout, 140 patients ont subi une cholécystostomie percutanée, 76 hommes et 64 femmes âgés en moyenne de 68,4 ans (écart-type 17,7 ans). Sur les 140 patients, 94 (67,1 %) présentaient un score ASA (American Society of Anesthesiolo-gists) de III ou IV. Les drains de cholécystostomie percutanée sont restés en place pen-dant une période médiane de 21,0 jours et la durée médiane des séjours hospita liers a été de 7,0 jours. Vingt-et-un patients (15,0 %) ont dû être réadmis en raison de complica-tions liées aux drains de cholécystostomie, et 10 patients (7,1 %), en raison d’une récur-rence de la cholécystite après le retrait du drain. Quarante-quatre patients (31,4 %) sont revenus pour une cholécystectomie non urgente, dont 32 (73 %) ont subi une cholécys-tectomie laparoscopique, 4 (9 %), une cholécystectomie laparoto mique, et 8 (18 %) une cholécystectomie laparoscopique convertie en cholécystectomie laparotomique.

Conclusion : La cholécystostomie percutanée est une approche sécuritaire envisage-able chez les patients plus âgés présentant plusieurs comorbidités. Toutefois, dans notre cohorte, moins du tiers de ces patients ont par la suite subi la cholécystectomie non urgente dé�nitive, et le taux de conversion de cholécystectomie laparoscopique en cholécystectomie laparotomique a été élevé.

Ida Molavi, MD Angela Schellenberg, MD, PhD Francis Christian, MBBS

Accepted Nov. 17, 2017

Correspondence to: A. Schellenberg Department of Surgical Oncology University of Toronto Princess Margaret Cancer Centre 6th floor OPG Building 700 University Ave Toronto ON M5G 1Z5 [email protected]

DOI: 10.1503/cjs.003517

clinnical-schellenberg.indd 195 2018-05-18 12:28 PM

RECHERCHE


T he accepted standard for treatment of acute chole-cystitis is laparoscopic cholecystectomy.1,2 How-ever, laparoscopic cholecystectomy in patients at

high risk with multiple comorbidities, such as older people and those who are critically ill, has been associated with a postoperative morbidity rate of over 20% in some series and a mortality rate reaching 5%.3–7 The presence of comorbidities has been shown to be a predictor of conver-sion to an open procedure8 and to result in increased peri-operative complications.9 Older patients have been found to have an increase in complicated biliary tract disease, with conversion rates of laparoscopic to open cholecystec-tomy ranging from 5% to 37%.4,5,10

Image-guided percutaneous cholecystostomy (PC) tube placement has been found to be a safe and effective alternative to cholecystectomy in patients at high risk with serious comorbidities.11–16 This procedure provides an immediate treatment option in these sick patients but is essentially a temporizing measure until the immediate emergency is treated and the patient’s medical condition is optimized for an elective cholecystectomy procedure several weeks after placement of the PC tube.17 Percuta-neous cholecystostomy tube placement followed by delayed cholecystectomy has been shown to be an effec-tive treatment option in this high-risk population.18 There have been inconsistent results regarding the supe-riority of early versus delayed laparoscopic cholecystec-tomy. Randomized and retrospective studies favoured early cholecystectomy for acute cholecystitis owing to a lower morbidity rate, shorter hospital stay and lower cost,19–21 whereas a systematic review did not show any significant difference in the rate of complications.22 Improved outcomes have been shown with delayed chole-cystectomy following percutaneous draining in this high-risk cohort.23

The resolution of the acute episode of cholecystitis after PC tube placement may itself be followed by com-plications, including a 1-year cholecystitis recurrence rate of up to 35% and 30-day overall mortality rate of up to 15.4%.24,25 In most studies published after 1995, the over-all mortality rate was 13.3%.25 However, the selection bias toward patients with comorbidities and older patients in this patient population results in a mortality rate related more to underlying overall health than to PC tube placement.

We aimed to determine short- and long-term clinical and operative outcomes of patients with acute cholecystitis treated initially with PC tube placement.

METHODS

We conducted a retrospective study of patients with a diagnosis of acute cholecystitis who underwent image-guided PC tube insertion from 2001 to 2011 at the Royal University Hospital or St. Paul’s Hospital, Saskatoon.

Patients with cholecystitis caused by malignant disease, ascending infection from common bile duct stones or pan-creatic disease were excluded from the study. The study was approved by the research ethics board of the Univer-sity of Saskatchewan.

We de�ned acute cholecystitis based on the Tokyo Cri-teria.26 The general surgeon on call for the acute care ser-vice at our institution was responsible for the decision to place a PC tube. All PC tubes were placed under ultra-sonography or computed tomography guidance by a con-sultant staff radiologist.

The chart review included a 2-year follow-up period. Patient demographic characteristics, comorbidities, labora-tory findings, image findings and follow-up including readmission due to PC tube complications and elective cholecystectomy were noted. Complications after PC tube insertion included readmission to hospital secondary to tube dislodgement, site leakage and need for reinsertion. The primary outcome was the proportion of patients who subsequently underwent elective cholecystectomy. Second-ary outcomes included hospital length of stay and the length of time the tube was in place.

We reviewed operative reports for patients who returned for elective cholecystectomy to identify the type of chole cystectomy performed (laparoscopic v. open), con-version to open surgery, the reason for the conversion and the American Society of Anesthesiologists (ASA) classi�ca-tion score.

Data are presented as median and quartile range as appropriate. We tested signi�cance using the Pearson χ2 test.

RESULTS

We reviewed the charts of 140 patients who underwent image-guided PC tube placement for acute cholecystitis. Of the 113 radiology reports that described the presence or absence of gallstones, 21 (18.6%) showing acalculous cholecystitis. The study cohort included patients who were admitted to the intensive care unit. There were 76 men (54.3%) and 64 women (45.7%) with a mean of 68.4 (stan-dard deviation 17.7) years. Patients had elevated leukocyte counts and multiple comorbidities (Table 1). Of the 140 patients, 19 were lost to follow-up.

Outcomes

In most cases, a PC tube was placed because the patient was a poor surgical candidate owing to multiple comorbid-ities, critical illness or advanced age. The timing of tube placement was immediately on presentation or after on going symptoms and elevated leukocyte count following initial intravenous antibiotic treatment in patients at high risk. Eight cases were due to gallbladder factors including gallbladder perforation and empyema.


RESEARCH


The PC tube was eventually removed in all patients for whom follow-up data were available. The tubes remained in place for a median of 21.0 (quartile range 31.0) days, and patients remained in hospital for a median of 7.0 (quartile range 10.0) days after tube insertion (Table 2). In 11 cases, the tube remained in place until elective cholecystectomy. Home care was arranged for all patients who were sent home with PC tubes.

After discharge from hospital, 21 patients (15.0%) were readmitted to hospital because of PC tube compli-cations (Table 2). The most common complications resulting in readmission were tube displacement from the gallbladder (9 patients), clogged tube (7) and pain at the PC site (6). Less common reasons were leaking around the tube, tube breakage and removal of the tube by the patient. After the tubes had been removed, 10 patients (7.1%) were readmitted with cholecystitis; most were managed nonoperatively, and only 3 subsequently under-went cholecystectomy.

Elective cholecystectomy

A total of 44 patients (31.4%) underwent subsequent elec-tive cholecystectomy following initial PC tube insertion (Table 3). Three had subtotal cholecystectomy. No deaths were documented in the 30-day postoperative period.

DISCUSSION

This cohort of 140 patients who underwent initial PC tube insertion for acute cholecystitis at our institution over a 10-year period were found to have serious comor-bidities, in most cases cardiac. About one-third of patients returned for subsequent delayed elective chole-cystectomy, of which 18% were converted from a laparo-scopic to an open pro cedure. Percutaneous cholecystos-tomy tube insertion had minimal associated morbidity, and patients were typically in hospital for 1 week follow-ing the procedure. The tube typically remained in place for 3 weeks, during which time 15.0% of patients required readmission to hospital for tube-related compli-cations. Following removal of the tube, 7.1% of patients were readmitted with recurrent cholecystitis.

Several investigators have reported outcomes after treatment of cholecystitis with PC.16,27,28 These studies, however, are limited to relatively small cohorts and lack consistency regarding a delayed versus early laparoscopic cholecystectomy approach after initial PC tube insertion. Furthermore, a 2013 Cochrane review showed that these trials also were not adequately powered, with a high risk of bias and differences in patient inclusion criteria.29 There-fore, there are no clear guidelines on the role of PC in the management of acute cholecystitis.

Table 1. Baseline demographic characteristics of patients with acute cholecystitis who underwent percutaneous cholecystostomy tube placement

CharacteristicNo. (%) of patients*

n = 140

Age, mean ± SD; yr 68.4 ± 17.7

Male sex 76 (54.3)

Weight, mean ± SD; kg 84.4 ± 20.2

Laboratory values, mean ± SD

Leukocyte count, × 109/L 14.4 ± 6.3

Hemoglobin level, g/L 117 ± 19.7

Comorbidities

Respiratory 48 (34.3)

Cardiac 98 (70.0)

Metabolic 64 (45.7)

Gastrointestinal 34 (24.3)

Renal 26 (18.6)

Hepatic 10 (7.1)

Malignant disease 27 (19.3)

Steroid use 5 (3.6)

Received anticoagulant treatment 30 (21.4)

Previous abdominal surgery 53 (37.8)

Radiological findings

Dilated gallbladder (n = 111) 94 (84.7)

Gallbladder wall thickness ≥ 4 mm (n = 122) 110 (90.2)

Pericholecystic fat stranding (n = 97) 79 (81.4)

Cholelithiasis (n = 113) 92 (81.4)



Table 2. Outcomes

Outcome No. (%) of patients*

Hospital length of stay after tube insertion, median (quartile range); d

7.0 (10.0)

No. of days tube in place, median (quartile range) 21.0 (31.0)

Readmission due to tube complications 21 (15.0)

1 readmission 14 (10.0)

2 readmissions 5 (3.6)

3 readmissions 2 (1.4)

Readmission for cholecystitis after tube removal 10 (7.1)


Table 3. Outcome of delayed elective cholecystectomy

OutcomeNo. (%) of patients

n = 44

Procedure

Laparoscopic 32 (73)

Open 4 (9)

Laparopscopic converted to open 8 (18)

ASA classification score

I or II 24 (54)

III or IV 20 (45)

Reason for conversion to open

Extensive adhesions 4 (50)

Severe inflammation 1 (12)

Difficult anatomy/dissection 3 (38)

ASA = American Society of Anesthesiologists.


RECHERCHE


In a recent study, Mirzahi and colleagues30 compared the outcome of delayed cholecystectomy in patients with and without initial PC tube insertion. They reported that delayed, elective laparoscopic cholecystectomy was a rou-tine practice at their institution, therefore giving them a larger cohort and hence possible bias. Patients who under-went PC had a longer hospital stay and a higher rate of conversion to open cholecystectomy than the group with-out PC (11% v. 4%). The PC group also presented a greater rate of overall biliary-related complications and surgical site infections. Percutaneous cholecystostomy was an independent predictor of conversion to open surgery, along with presence of cirrhosis and choledocholithiasis.

Less than one-third (31.4%) of patients in our cohort returned for delayed, elective cholecystectomy; in 18%, a laparoscopic procedure was converted to an open proced-ure. This is in keeping with reported rates of conversion from laparoscopic to open cholecystectomy of 14%–67% in patients who have had PC tube placement.3,10,14,28,30–33

In our study, 15.0% of patients required 1 or more readmissions because of complications related to the PC tube, most commonly tube dislodgement or clogging, and pain at the insertion site. In a systematic review, Winbladh and colleagues25 also reported slippage of the PC tube, in 8.6% of patients. The rate of recurrent cholecystitis after tube removal was 7.1% in our patient cohort, lower than that reported by Sanjay and colleagues,28 22%.

Just over half (54%) of our patients who underwent elective cholecystectomy had an ASA score of I or II. Pre-vious studies have also con�rmed the intuitive conclusion that predictive factors of eventual cholecystectomy include younger age and fewer comorbidities.34,35 However, the ASA score calculated by anesthesia at the time of elective cholecystectomy is not always independent of observer bias and also does not necessarily re�ect the clinical condition at the time of initial presentation.

A total of 18.6% of patients in our cohort had docu-mented acalculous cholecystitis. Sicker patients with multi-ple comorbidities are more likely to have acalculous chole-cystitis and, in turn, are more likely to be treated with PC. This is re�ected in the recommendation for laparoscopic cholecystectomy for acalculous cholecystitis in more �t patients only.36

Limitations

Our research was limited as a retrospective study, and hence there is the possibility of selection bias. This could be remedied in future studies by assigning speci�c criteria for PC over laparoscopic cholecystectomy or vice versa in a prospective manner and measuring outcomes using pro-spective evaluation models. Another limitation of this study was the inability to accurately assign ASA scores, since it is well known that these are not always accurate, unless in the context of a prospective study.

CONCLUSION

There are no current guidelines based on de�nitive data to guide the decision as to which patients should undergo PC tube insertion on initial presentation, and, hence, this is at the surgeon’s discretion. Better de�nitive criteria are needed for choosing one treatment (PC) over another (laparoscopic cholecystectomy) in the context of the initial presentation of patients with acute cholecystitis in order to optimize treat-ment selection and outcomes. Percutaneous cholecystos-tomy tube insertion can be performed in patients at high risk with numerous comorbidities without signi�cant mor-bidity and mortality. However, less than one-third of such patients in our cohort subsequently had the de�nitive inter-vention of elective cholecystectomy. The rates of initial open cholecystectomy and conversion from laparoscopic to open cholecystectomy were quite high in this group.

Af�liations: From the Department of General Surgery, University of Saskatchewan, Saskatoon, Sask. (Molavi, Schellenberg, Christian); and the Department of Surgical Oncology, University of Toronto, Toronto, Ont. (Schellenberg).


Contributors: All authors designed the study. A. Schellenberg and I. Molavi acquired the data, which all authors analyzed. All authors wrote and reviewed the article and approved the �nal version for publication.

References

1. Miller RE, Kimmelstiel FM. Laparoscopic cholecystectomy for acute cholecystitis. Surg Endosc 1993;7:296-9.

2. Lujan JA, Parrilla P, Robles R, et al. Laparoscopic cholecystectomy in the treatment of acute cholecystitis. J Am Coll Surg 1995;181:75-7.

3. Pessaux P, Regenet N, Tuech JJ, et al. Laparoscopic versus open cholecystectomy: a prospective comparative study in the elderly with acute cholecystitis. Surg Laparosc Endosc Percutan Tech 2001;11:252-5.

4. Bingener J, Richards ML, Schwesinger WH, et al. Laparoscopic cholecystectomy for elderly patients: Gold standard for golden years? Arch Surg 2003;138:531-6.

5. Tambyraja AL, Kumar S, Nixon SJ. Outcome of laparoscopic chole-cystectomy in patients 80 years and older. World J Surg 2004;28:745-8.

6. Maxwell JG, Tyler BA, Rutledge R, et al. Cholecystectomy in patients aged 80 and older. Am J Surg 1998;176:627-31.

7. Brunt LM, Quasebarth MA, Dunnegan DL. Outcome analysis of laparoscopic cholecystectomy in the extremely elderly. Surg Endosc 2001;15:700-5.

8. Goh JC, Tan JK, Lim JW, et al. Laparoscopic cholecystectomy for acute cholecystitis: an analysis of early versus delayed cholecystec-tomy and predictive factors for conversion. Minerva Chir 2017;72: 455-63.

9. Giger UF, Michel JM, Opitz I, et al. Risk factors for perioperative complications in patients undergoing laparoscopic cholecystectomy: analysis of 22,953 consecutive cases from the Swiss Association of Laparoscopic and Thoracoscopic Surgery database. J Am Coll Surg 2006;203:723-8.

10. Uecker J, Adams M, Skipper K, et al. Cholecystitis in the octogenar-ian: Is laparoscopic cholecystectomy the best approach? Am Surg 2001;67:637-40.

11. Silberfein EJ, Zhou W, Kougias P, et al. Percutaneous cholecystostomy for acute cholecystitis in high-risk patients: experience of a surgeon-initiated interventional program. Am J Surg 2007;194:672-7.


RESEARCH


12. Nasim S, Khan S, Alvi R, et al. Emerging indications for percutane-ous cholecystostomy for the management of acute cholecystitis — a retrospective review. Int J Surg 2011;9:456-9.

13. Joseph T, Unver K, Hwang GL, et al. Percutaneous cholecystostomy for acute cholecystitis: ten-year experience. J Vasc Interv Radiol 2012; 23:83-8.

14. Spira RM, Nissan A, Zamir O, et al. Percutaneous transhepatic chole-cystostomy and delayed laparoscopic cholecystectomy in critically ill patients with acute calculus cholecystitis. Am J Surg 2002;183:62-6.

15. Welschbillig-Meunier K, Pessaux P, Lebigot J, et al. Percutaneous cholecystostomy for high-risk patients with acute cholecystitis. Surg Endosc 2005;19:1256-9.

16. Cherng N, Witkowski ET, Sneider EB, et al. Use of cholecystos-tomy tubes in the management of patients with primary diagnosis of acute cholecystitis. J Am Coll Surg 2012;214:196-201.

17. Akyürek N, Salman B, Yüksel O, et al. Management of acute calcu-lous cholecystitis in high-risk patients: percutaneous cholecystotomy followed by early laparoscopic cholecystectomy. Surg Laparosc Endosc Percutan Tech 2005;15:315-20.

18. Berber E, Engle KL, String A, et al. Selective use of tube chole-cystostomy with interval laparoscopic cholecystectomy in acute cho-lecystitis. Arch Surg 2000;135:341-6.

19. Gutt CN, Encke J, Koninger J, et al. Acute cholecystitis: early versus delayed cholecystectomy, a multicenter randomized trial (ACDC study, NCT00447304). Ann Surg 2013;258:385-93.

20. Chang TC, Lin MT, Wu MH, et al. Evaluation of early versus delayed laparoscopic cholecystectomy in the treatment of acute cho-lecystitis. Hepatogastroenterology 2009;56:26-8.

21. Casillas RA, Yegiyants S, Collins JC. Early laparoscopic cholecystec-tomy is the preferred management of acute cholecystitis. Arch Surg 2008;143:533-7.

22. Gurusamy KS, Davidson C, Gluud C, et al. Early versus delayed laparoscopic cholecystectomy for people with acute cholecystitis. Coch rane Database Syst Rev 2013;(6):CD005440.

23. Karakayali FY, Akdur A, Kirnap M, et al. Emergency cholecystec-tomy vs percutaneous cholecystostomy plus delayed cholecystectomy for patients with acute cholecystitis. Hepatobiliary Pancreat Dis Int 2014;13:316-22.

24. Ha JP, Tsui KK, Tang CN, et al. Cholecystectomy or not after per-cutaneous cholecystostomy for acute calculous cholecystitis in high-risk patients. Hepatogastroenterology 2008;55:1497-502.

25. Winbladh A, Gullstrand P, Svanvik J, et al. Systematic review of cholecystostomy as a treatment option in acute cholecystitis. HPB (Oxford) 2009;11:183-93.

26. Hirota M, Takada T, Kawarada Y, et al. Diagnostic criteria and severity assessment of acute cholecystitis: Tokyo Guidelines. J Hepa-tobiliary Pancreat Surg 2007;14:78-82.

27. Rodriguez-Sanjuan JC, Arruabarrena A, Sanchez-Moreno L, et al. Acute cholecystitis in high surgical risk patients: Percutaneous chole-cystostomy or emergency cholecystectomy? Am J Surg 2012;204:54-9.

28. Sanjay P, Mittapalli D, Marioud A, et al. Clinical outcomes of a per-cutaneous cholecystostomy for acute cholecystitis: a multicentre analysis. HPB (Oxford) 2013;15:511-6.

29. Gurusamy KS, Rossi M, Davidson BR. Percutaneous cholecystec-tomy for high-risk surgical patients with acute calculous cholecystitis. Cochrane Database Syst Rev 2013;(8):CD007088.

30. Mizrahi I, Mazeh H, Yuval JB, et al. Perioperative outcomes of delayed laparoscopic cholecystectomy for acute calculous cholecystitis with and without percutaneous cholecystostomy. Surgery 2015;158:728-35.

31. Dolan JP, Diggs BS, Sheppard BC, et al. The national mortality bur-den and signi�cant factors associated with open and laparoscopic cholecystectomy: 1997–2006. J Gastrointest Surg 2009;13:2292-301.

32. Abi-Haidar Y, Sanchez V, Williams SA, et al. Revisiting percutane-ous cholecystostomy for acute cholecystitis based on a 10-year expe-rience. Arch Surg 2012;147:416-22.

33. Cheruvu CV, Eyre-Brook IA. Consequences of prolonged wait before gallbladder surgery. Ann R Coll Surg Engl 2002;84:20-2.

34. De Mestral C, Gomez D, Haas B, et al. Cholecystostomy: a bridge to hospital discharge but not delayed cholecystectomy. J Trauma Acute Care Surg 2013;74:175-80.

35. Kortram K, van Ramshorst B, Bollen TL, et al. Acute cholecystitis in high risk surgical patients: percutaneous cholecystostomy versus lap-aroscopic cholecystectomy (CHOCOLATE trial): study protocol for a randomized controlled trial. Trials 2012;13:7.

36. Schuld J, Glanemann M. Acute cholecystitis. Viszeralmedizin 2015; 31:163-5.



REVIEW • REVUE

Diagnostic accuracy of transabdominal ultrasonography for gallbladder polyps: systematic review

Background: Previous research has shown variable but generally poor accuracy of transabdominal ultrasonography in the diagnosis of gallbladder polyps. We per-formed a systematic review of the literature with the aim of helping surgeons interpret and apply these �ndings in the preoperative assessment and counselling of their patients.

Methods: We searched PubMed, MEDLINE and the Cochrane database using the keywords “gallbladder,” “polyp,” “ultrasound,” “pathology” and “diagnosis” for English-language articles published after 1990 with the full-text article available through our institutional subscriptions. Polyps were de�ned as immobile features that on transabdominal ultrasonography appear to arise from the mucosa and that lack an acoustic shadow, and pseudopolyps were de�ned as features such as in�am-mation, hyperplasia, cholesterolosis and adenomyomatosis that convey no risk of malignant transformation.

Results: The search returned 1816 articles, which were narrowed down to 14 pri-mary sources involving 15 497 (range 23–13 703) patients who had preoperative transabdominal ultrasonography, underwent cholecystectomy and had postoperative pathology results available. Among the 1259 patients in whom a gallbladder polyp was diagnosed on ultrasonography, 188 polyps were con�rmed as true polyps on patho-logic examination, and 81 of these were found to be malignant. Of the 14 238 patients for whom a polyp was not seen on ultrasonography, 38 had a true polyp on pathologic examination, none of which were malignant. For true gallbladder polyps, transabdom-inal ultrasonography had a sensitivity of 83.1%, speci�city of 96.3%, positive predic-tive value of 14.9% (7.0% for malignant polyps) and negative predictive value of 99.7%.

Conclusion: Transabdominal ultrasonography has a high false-positive rate (85.1%) for the diagnosis of gallbladder polyps. Further study of alternative imaging modali-ties and reevaluation of existing management guidelines are warranted.

Contexte : Des recherches antérieures ont montré la précision variable, mais géné-ralement médiocre, de l’échographie transabdominale pour le diagnostic des polypes de la vésicule biliaire. Nous avons procédé à une revue systématique de la littérature scienti�que a�n d’aider les chirurgiens à interpréter et à appliquer ces résultats lors de l’évaluation préopératoire, et à conseiller leurs patients.

Méthodes : Nous avons interrogé les réseaux PubMed, MEDLINE et la base de données Cochrane à partir des mots clés « gallbladder », « polyp », « ultrasound », « pathology » et « diagnosis » (vésicule biliaire, polype, échographie, pathologie et diagnostic) pour recenser les articles en langue anglaise publiés après 1990, pour lesquels le texte intégral était accessible par abonnement institutionnel. À l’échogra-phie, les polypes étaient dé�nis comme des structures �xes semblant émaner de la muqueuse et dépourvues d’ombre acoustique, et les pseudopolypes étaient dé�nis par des caractéristiques telles que l’in�ammation, l’hyperplasie, la cholestérolose et l’adé-nomyomatose ne comportant pas de risque de transformation maligne.

Résultats : La recherche a généré 1816 articles qui ont été ramenés à 14 sources principales regroupant 15 497 (éventail, 23–13 703) patients ayant subi une écho-graphie transabdominale préopératoire et une cholécystectomie, et pour lesquels on disposait des résultats de l’examen anatomopathologique postopératoire. Sur les 1259 patients chez qui des polypes intravésiculaires ont été diagnostiqués à l’écho graphie, 188 polypes ont été jugés vrais à l’examen anatomopathologique, et 81 d’entre eux se sont révélés malins. Parmi les 14 238 patients chez lesquels

Erin Martin, MD, MSc Richdeep Gill, MD Estifanos Debru, MD

Accepted for publication Nov. 22, 2017

Correspondence to: E. Martin Division of General Surgery University of Calgary North Tower, Room 1006 1403 29th Ave NW Calgary AB T2N 2T9 [email protected]

DOI: 10.1503/cjs.011617

review-martin.indd 200 2018-05-18 12:43 PM

REVIEW


G allbladder cancer is a rare but serious disease. There were 500 new cases of gallbladder cancer and 265 deaths from this disease in Canada in

2015.1 Gallbladder polyps are malignant or potentially malignant exophytic mucosal lesions such as adenomas, adenocarcinomas or carcinomas.2 Pseudopolyps are fea-tures such as in�ammation, hyperplasia, cholesterolosis and adenomyomatosis that convey no risk of malignant transformation.2 On transabdominal ultrasonography, gall-bladder polyps are defined as immobile features that appear to arise from the mucosa and that lack an acoustic shadow (which is typically seen with gallstones).3 Previous pathologic review showed that all malignant gallbladder tumours are found in polyps larger than 6 mm, that polyps 10–20 mm in size have a 43%–77% incidence of gallblad-der cancer, and that polyps larger than 20 mm are “pathognomonic” for gallbladder cancer.2 As such, current guidelines for management of gallbladder polyps are heavi ly focused on their size. Polyps measuring 10 mm or less should be managed with watchful waiting (with uncer-tainty regarding the number or frequency of visits or ultra-sonography examinations required), those measuring 11–20 mm should be managed with close follow-up or cholecystectomy, and those larger than 20 mm require cholecystectomy.2,4

Transabdominal ultrasonography has been shown to be inaccurate in measuring the size of polypoid gallbladder lesions. Choi and colleagues5 found a mean deviation of only 0.89 mm for true polyps, but Guo and colleagues6 reported that transabdominal ultrasonography universally overestimates polyp size by 4.24 mm (standard deviation [SD] 0.19 mm). These variations are potentially important when one considers that the suggested management of polyps changes substantially over a 10-mm range, and they could translate into patients’ being classi�ed into a higher risk category and being offered more ultrasonography examinations and surgical procedures than is appropriate. Guo and colleagues6 also found that transabdominal ultra-sonography overestimates the size of cholesterolosis by 5.12 mm (SD 0.21 mm), much of which may be inaccu-rately labelled as true polyps.

Transabdominal ultrasonography has also been shown to be a �awed modality for serial monitoring of gallbladder polyps, as longitudinal studies have demonstrated minimal to no progression in size. Kratzer and colleagues7 followed

2415 patients with a mean polyp size of 5.0 mm (SD 2.1 mm) and found that there was no statistically sig-nificant change in mean size at 30 months (5 mm [SD 2.8 mm]) or 84 months (4.0 mm [SD 2.3 mm]). Ito and colleagues8 followed 417 patients with polyps measur-ing less than 10 mm for a median of 17 months (range 1–81 mo) and found growth greater than 3 mm in only 6% of patients. High-risk features and malignant disease did not develop in any patient. Although these studies are small, they indicate that gallbladder polyps may not follow the standard progression of adenoma to adenocarcinoma.

Previous studies have shown poor correlation between ultrasonography diagnosis and �nal pathology,2–4 and an attempted meta-analysis was aborted owing to a lack of sat-isfactory research.2 This diagnostic uncertainty introduces the potential for unnecessary imaging, unnecessary surgery and all the associated risks to patients and costs to the health care system.

We performed a systemic review with the aim of help-ing surgeons interpret previous research and apply the �ndings in the preoperative assessment and counselling of their patients.

METHODS

We studied transabdominal ultrasonography as it is the most widely available and most commonly used imaging modality for biliary disease. We searched MEDLINE, PubMed and the Cochrane database using the keywords “gallbladder,” “polyp,” “ultrasound,” “pathology” and “diagnosis.” The articles were screened for relevance and duplicates. We included articles for review if they contained data for patients who had had preoperative transabdominal ultrasonography and had undergone cholecystectomy, and for whom postoperative pathology results were available. The search was limited to English-language articles pub-lished after 1990 with the full-text article available through our institutional subscriptions. We used the year 1990 as a cut-off given the substantial improvement in transabdom-inal ultrasonography technique after that year.9,10 Articles were excluded from review if they did not focus on gall-bladder polyps, if transabdominal ultrasonography was not used, if patients did not undergo cholecystectomy or if the article did not contain complete pathologic data. We also reviewed the references of screened articles.

aucun polype n’avait été détecté à l’échographie, 38 étaient porteurs d’un vrai polype à l’examen anatomopathologique et aucun ne s’est révélé malin. En ce qui concerne les vrais polypes intravésiculaires, l’échographie transabdominale a une sensibilité de 83,1 %, une spécificité de 96,3 %, une valeur prédictive positive de 14,9 % (7,0 % dans le cas des polypes malins) et une valeur prédictive négative de 99,7 %.

Conclusion : L’échographie transabdominale présente un taux de résultats faux positifs élevé (85,1 %) pour le diagnostic des polypes de la vésicule biliaire. Il faudra approfondir la recherche sur d’autres techniques d’imagerie et réévaluer les lignes directrices actuelles de prise en charge.


REVUE


We examined the de�nitions of true polyps and pseudo-polyps used by the authors of each article for consistency with the standardized definitions used in this review (polyps: immobile features that on transabdominal ultra-sonography appear to arise from the mucosa and that lack an acoustic shadow;3 pseudopolyps: features such as in�am-mation, hyperplasia, cholesterolosis and adenomyomatosis that convey no risk of malignant transformation2). We then retrieved the data directly from the articles in concor-dance with the standardized de�nition.

RESULTS

The literature search produced 1816 results. A total of 1789 abstracts were excluded because they were off topic (e.g., ultrasonography of thyroid nodules) or were dupli-cates (n = 1775), or because they did not have a full-text English-language version available (n = 14). Of the 27 arti-cles screened, 4 were excluded because they involved the incorrect imaging modality (e.g., a study of 3-dimensional ultrasonography only), and 9 were excluded because of absent or inadequate pathology data. No new articles were identi�ed on screening of the references of the 27 full-text articles. This resulted in 14 primary articles for inclusion in our review (Fig. 1).

The sample size in the 14 studies ranged from 23 to 13 703 patients (Table 1). Among the 1259 patients in whom a gallbladder polyp was diagnosed on preoperative transabdominal ultrasonography, 188 polyps were con-�rmed as true polyps on pathologic examination, and 81 of

these were found to be malignant. This corresponds to 188 true-positive ultrasonography diagnoses and 1071 false-positive ultrasonography diagnoses (Fig. 2). Among the 1071 patients with a positive ultrasonography report and a negative pathology report, 735 pathology reports were sig-ni�cant for pseudopolyps, and 401 reports were signi�cant for cholelithiasis, with some of the false-positive reports being positive for both pseudopolyps and cholelithiasis (Fig. 3). Of the 14 238 patients for whom a polyp was not seen on ultrasonography, 38 had a true polyp on pathologic examination, none of which were malignant.

For the diagnosis of true gallbladder polyps by trans-abdominal ultrasonography, the sensitivity was 83.1%, the speci�city was 96.3%, the positive predictive value was 14.9% (7.0% for malignant polyps), and the negative predictive value was 99.7% (Table 2). Stated another way, 85.1% of the patients in whom a gallbladder polyp was diagnosed on ultrasonography did not actually have a true polyp.

Data on the size of the true polyps were not consistently reported, but the available data show that true polyps (especially malignant polyps) tended to be larger than 10 mm but with a minimum SD of 4.2 mm (Table 3).

DISCUSSION

The proportions of true- and false-positive results in the included studies show generalized overestimation of the presence of true gallbladder polyps. Among the patients with a false-positive ultrasonography report,

Fig. 1. Flow diagram showing article selection.

MEDLINEn = 1608

Abstracts identifiedn = 1816

Articles screenedn = 27

Articles included in reviewn = 14

Excluded n = 1789• Off topic or duplicate n = 1775• Full text not available n = 14

Excluded n = 13• Incorrect imaging modality n = 4• No pathology data n = 9

PubMedn = 170

Cochrane databasen = 38


REVIEW


pathologic examination showed a pseudopolyp in 59% and cholelithiasis in 38%, both of which are known to increase the rate of false-positive results.2 Generally, pseudopolyps are more likely to be found on final

pathologic examination than gallstones, which indicates that the former are more dif�cult to distinguish from true polyps and are more likely to confound ultraso-nography diagnosis.

Table 1. Ultrasonography and pathology data from the 14 primary articles reviewed

StudyNo. of

patients

Negative ultrasonography report Positive ultrasonography report

No. of patients

No polyp

True polyp

True polyp, malignant

No. of patients

No polyp

No polyp, pseudopolyp

No polyp, gallstones

True polyp

True polyp, malignant

Akyurek et al.11 853 797 787 10 0 56 46 17 12 10 0

Ali Channa et al.12 28 0 — — — 28 23 15 8 5 1

Chattopadhyay et al.13

23 0 — — — 23 20 8 12 3 2

Choi et al.5 59 0 — — — 59 46 46 5 13 3

Damore et al.4 41 0 — — — 41 40 27 15 1 0

French et al.3 13 703 13 441 13 413 28 0 262 256 139 129 6 3

Huang et al.14 143 0 — — — 143 121 121 23 22 6

Ito et al.8 80 0 — — — 80 72 54 9 8 1

Khan et al.15 26 0 — — — 26 22 21 12 4 1

Mainprize et al.16 34 0 — — — 34 30 29 22 4 2

Park et al.17 180 0 — — — 180 114 82 57 66 25

Sarkut et al.18 138 0 — — — 138 116 77 39 22 21

Xu et al.19 59 0 — — — 59 50 20 30 9 7

Zielinski et al.20 130 0 — — — 130 115 79 28 15 9

All 15 497 14 238 14 200 38 0 1259 1071 735 401 188 81

All excluding French et al.3

1794 797 787 10 0 997 815 596 272 182 78

Fig. 2. Ratio of true-positive to false-positive diagnoses of gallbladder polyp on ultrasonography.

46

No.

of

diag

nose

s

10 5 4 4 98

23

3

20

13

46

1 6

40 256

12172

2230 50

15 188

115 1071

66

22 22

116

114

Akyur

ek et

al.11

Ali Cha

nna e

t al.1

2

Chatto

padh

yay e

t al.1

3

Choi e

t al.5

Damor

e et a

l.4

Fren

ch et

al.3

Huang

et al

.14

Ito et

al.8

Khan e

t al.1

5

Main

prize

et al

.16

Park et

al.17

Sarkut

et al

.18

Xu et a

l.19

Zielin

ski e

t al.2

0All

False positive True positive


REVUE


Other imaging modalities have also been shown to be inaccurate in measuring polyps, and the modalities are not well correlated with one another. Choi and col-leagues5 showed that there was a mean difference of

5.66 mm in measurements of polyps for cholesterolosis between ultrasonography and computed tomography and a mean difference of 2.17 mm in measurements of noncholesterol polyps. Those authors also found that a

Fig. 3. Ratio of gallstones to pseudopolyps on ultrasonography with false-positive diagnosis.

12

1715 21

29

20

54

8

8

12

5

46

27

139

15

129

121

9

1222

30

79

735

28401

57

82

23

Gallstones Pseudopolyps

77

39

No.

of

galls

tone

s/ps

eudo

poly

ps

Ali Cha

nna e

t al.1

2

Chatto

padh

yay e

t al.1

3

Choi e

t al.5

Damor

e et a

l.4

Fren

ch et

al.3

Huang

et al

.14

Ito et

al.8

Khan e

t al.1

5

Main

prize

et al

.16

Park et

al.17

Sarkut

et al

.18

Xu et a

l.19

Zielin

ski e

t al.2

0All

Akyur

ek et

al.11

Table 2. Statistical analysis of data from primary sources

Study Sensitivity Specificity

Positive predictive value Negative predictive

valueTrue polypMalignant

polyp

Akyurek et al.11 0.500 0.945 0.179 0.000 0.988

Ali Channa et al.12 — — 0.179 0.042 —


— — 0.130 0.091 —

Choi et al.5 — — 0.220 0.061 —

Damore et al.4 — — 0.024 0.000 —

French et al.3 0.176 0.981 0.023 0.012 0.998

Huang et al.14 — — 0.154 0.047 —

Ito et al.8 — — 0.100 0.014 —

Khan et al.15 — — 0.154 0.043 —

Mainprize et al.16 — — 0.118 0.063 —

Park et al.17 — — 0.367 0.180 —

Sarkut et al.18 — — 0.159 0.153 —

Xu et al.19 — — 0.153 0.123 —

Zielinski et al.20 — — 0.115 0.073 —

All 0.831 0.963 0.149 0.070 0.997

All excluding French et al.3

0.948 0.945 0.183 0.087 0.988


REVIEW


polypoid lesion that is smaller or is not present on com-puted tomography, as compared to a lesion seen on ultrasonography, is more likely to be cholesterolosis than a true polyp, which was supported by Song and colleagues.21 Song and colleagues21 also found that polyps diagnosed on ultrasonography were more likely to be cholesterolosis in young patients with a high body mass index. Other modalities such as 3-dimensional ultrasonography and endoscopic ultrasonography have shown promise in the diagnosis of gallbladder polyps, but further study of their accuracy and cost–bene�t ratio is required.

It is important to note that our results are heavily weighted by the extensive review conducted by French and colleagues3 of 13 703 patients. Their data represent 20.8% of the patients with a positive preoperative ultra-sonography report in the 14 articles reviewed. Their large data set results in lower sensitivity and speci�city values than those seen in the smaller study by Akyurek and colleagues,11 which may be a centre-speci�c trend but may also be closer to the true values for gallbladder polyps on ultrasonography. The data are summarized with and without the values from French and col-leagues,3 and, although the differences are minor, they must be considered.

Recommendations

Despite the �nding that transabdominal ultrasonography is a limited modality for the diagnosis and monitoring of gallbladder polyps, its limitations must be weighed against the severity of gallbladder cancer. Therefore, we recommend that patients with polyps smaller than 10 mm be educated as to the limitations of ultrasonogra-phy and the low likelihood of current or future malig-nant disease (Fig. 4). They can be offered elective chole-

cystectomy. If patients with polyps smaller than 10 mm decline cholecystectomy, they should be followed with transabdominal ultrasonography at 6 and 12 months to check for rapid polyp growth and other high-risk fea-tures. Patients with polyps 10–20 mm in size should be offered cholecystectomy, and, if surgery is contraindi-cated, they should be followed with transabdominal ultrasonography at 6 and 12 months. Patients with polyps greater than 20 mm should proceed to cholecys-tectomy with the same threshold for surgical contraindi-cations as other oncologic procedures. Patients with polyps of any size should be offered cholecystectomy if they have any of the following high-risk features: rapid growth of 3 mm or more in a 6-month period, sessile polyps, solitary polyps, thickened gallbladder wall, cho-ledocholithiasis, biliary colic, atypical right upper quad-rant pain, history of choledochal cyst, comorbid primary sclerosing cholangitis or age more than 50 years.2,22,23

Limitations

The data in this review are biased by the fact that most patients who undergo ultrasonography have abdominal problems, that most patients who undergo cholecystec-tomy have symptomatic disease, and that 20.8% of the positive ultrasonography results in the review came from 1 study.3 Our review is further limited by the lack of large studies, speci�cally of randomized and blinded studies, and by the relative homogeneity of the included studies. In addition, a meaningful likelihood ratio for the presence of true polyps based on an ultrasonography diagnosis cannot be calculated given that the true incidence of gallbladder polyps is not known, incidence data gathered by ultra-sonography reviews are inherently �awed, and incidence data gathered by pathologic review are biased because the patients required cholecystectomy.

Table 3. Size of true polyps on final pathologic report

Study

Size, mm; no of cases Minimum size, mm

Maximum size, mm

Median size, mm

Mean size ± SD, mm Comment < 5 5–10 > 10

Ali Channa et al.12 0 27 1 — 15 — — Incomplete data


0 0 3 12 20 18 16.7 ± 4.2 Benign and malignant polyps

Damore et al.4 0 1 0 — 6 — — Incomplete data

Khan et al.15 0 0 4 — — — — Incomplete data

Mainprize et al.16 0 0 2 — — — — Incomplete data

Park et al.17 0 1 24 8 30 15 15.2 ± 5.6 Malignant polyps only

Xu et al.19 — — — 3 12 — — Incomplete data

Zielinski et al.20 2 (< 6)

3 (6–9)

8 ( ≥ 10)

4 53 18 18.5 ± 15.4 Benign and malignant polys

Akyurek et al.11 1 (< 6)

1 (6–10)

8 — — — — Incomplete data

Sarkut et al.18 0 1 21 — — — — Incomplete data

Huang et al.14 — Benign — — — — 10.8 ± 4.7 Incomplete data

— Malignant — — — — 13.5 ± 4.2 Incomplete data



REVUE


CONCLUSION

In this systematic review involving 1259 patients who had positive �ndings on preoperative transabdominal ultra-sonography and then underwent cholecystectomy, ultrasonography had a sensitivity of 83.1% for true gall-bladder polyps, a speci�city of 96.3%, a positive predic-tive value of 14.9% (7.0% for malignant true polyps) and a negative predictive value of 99.7%. Further studies are needed to clarify the optimal use and cost–bene�t ratio of transabdominal ultrasonography in the diagnosis of gall-bladder polyps.Af�liation: From the Department of Surgery, University of Calgary, Calgary, Alta.


Contributors: E. Martin and E. Debru designed the study. E. Martin acquired and analyzed the data, which R. Gill also analyzed. E. Martin wrote the article, which all authors reviewed and approved for publication.

References

1. Canadian Cancer Society’s Advisory Committee on Cancer Statistics. Canadian cancer statistics 2015. Toronto: Canadian Cancer Society; 2015.

2. Gurusamy KS, Abu-Amara M, Farouk M, et al. Cholecystectomy for gallbladder polyp. Cochrane Database Syst Rev 2009;(1):CD007052.

3. French DG, Allen PD, Ellsmere JC. The diagnostic accuracy of transabdominal ultrasonography needs to be considered when man-aging gallbladder polyps. Surg Endosc 2013;27:4021-5.

4. Damore LJ, Cook CH, Fernandez KL, et al. Ultrasonography incor-rectly diagnosis gallbladder polyps. Surg Laparosc Endosc Percutan Tech 2001;11:88-91.

5. Choi JH, Yun JW, Kim YS, et al. Pre-operative predictive factors for gallbladder cholesterol polyps using conventional diagnostic imaging. World J Gastroenterol 2008;14:6831-4.

6. Guo J, Wu G, Zhou Z. Polypoid lesions of the gallbladder: report of 160 cases with special reference to diagnosis and treatment in China. Int J Clin Exp Pathol 2015;8:11569-78.

7. Kratzer W, Haenle MM, Voegtle A, et al. Ultrasonographically detected gallbladder polyps: A reason for concern? A seven-year follow-up study. BMC Gastroenterol 2008;8:41.

Fig. 4. Suggested management of gallbladder polyps detected on transabdominal ultrasonography.

Polyp < 10 mm

Polyp < 10 mm: consider cholecystectomyPolyp 10–20 mm: offer cholecystectomy

Educate aboutlimitations of

ultrasonography diagnosis

Considercholecystectomy

Patient declines

Follow with clinical evaluation and transabdominal ultrasonography

at 6 and 12 mo

Patient declines

Polyp unchangedSize increase or

any high-risk feature

Offercholecystectomy

Educate aboutlimitations of

ultrasonography diagnosis

Educate about limitations ofultrasonography diagnosis

Recommend cholecystectomy

Polyp 10–20 mm Polyp > 20 mm High-risk features*

*Rapid growth of ≥ 3 mm in 6 mo, sessile polyps, solitary polyps, thickened gallbladder wall,

choledocholithiasis, biliary colic, atypical right upper quadrant pain, comorbid primary sclerosing

cholangitis, patient age > 50 yr


REVIEW


8. Ito H, Hann LE, D’Angelica M, et al. Polypoid lesions of the gall-bladder: diagnosis and follow-up. J Am Coll Surg 2009;208:570-5.

9. Shea JA, Berlin J, Escarce J, et al. Revised estimates of diagnostic test sensitivity and speci�city in suspected biliary tract disease. Arch Intern Med 1994;154:2573-81.

10. History of ultrasound. 2016. UltrasoundSchoolsInfo. Available: www.ultrasoundschoolsinfo.com/history/ (accessed 2016 Nov. 30).

11. Akyurek N, Salman B, Irkorucu O, et al. Ultrasonography in the diagnosis of true gallbladder polyps: the contradiction in the litera-ture. HPB (Oxford) 2005;7:155-8.

12. Ali Channa M, Zubair M, Mumtaz TA, et al. Management of polyp-oid lesions of the gallbladder. J Surg Pak Int 2009;14:77-9.

13. Chattopadhyay D, Lochan R, Balupuri S, et al. Outcome of gall blad-der polypoidal lesions detected by transabdominal ultrasound scan-ning: a nine year experience. World J Gastroenterol 2005;11:2171-3.

14. Huang CS, Lien HH, Jeng JY, et al. Role of laparoscopic cholecys-tectomy in the management of polypoid lesions of the gallbladder. Surg Laparosc Endosc Percutan Tech 2001;11:242-7.

15. Khan MR, Al Ghamdi S, Nasser MFM. Management of polypoid lesions of gallbladder: a retrospective study at King Abdullah Hospi-tal, Bisha, Kingdom of Saudi Arabia. Pak J Surg 2012;28:182-5.

16. Mainprize KS, Gould SW, Gilbert JM. Surgical management of pol-ypoid lesions of the gallbladder. Br J Surg 2000;87:414-7.

17. Park JK, Yoon YB, Kim YT, et al. Management strategies for gall-bladder polyps: Is it possible to predict malignant gallbladder polyps? Gut Liver 2008;2:88-94.

18. Sarkut P, Kilicturgay S, Ozer A, et al. Gallbladder polyps: factors affecting surgical decision. World J Gastroenterol 2013;19:4526-30.

19. Xu HX, Yin XY, Lu MD, et al. Comparison of three- and two-dimensional sonography in diagnosis of gallbladder disease: prelimi-nary experience. J Ultrasound Med 2003;22:181-91.

20. Zielinski MD, Atwell TD, Davis PW, et al. Comparison of surgically resected polypoid lesions of the gallbladder to their pre-operative ultrasound characteristics. J Gastrointest Surg 2009;13:19-25.

21. Song HL, Shin JH, Kim H, et al. Clinical and radiologic preopera-tive predicting factors for GB cholesterol polyp. J Korean Surg Soc 2012;82:232-7.

22. Baltayiannis N, Gavressea T, Rizos S. Gallbladder polyps: diagnosis and treatment. Hell J Surg 2010;82:233-8.

23. Buckles DC, Lindor DK, LaRusso NF, et al. In primary sclerosing cholangitis, gallbladder polyps are frequently malignant. Am J Gas-troenterol 2002;97:1138-42.

CJS’s top viewed articles*

1. Research questions, hypotheses and objectives Farrugia et al. Can J Surg 2010;53:278–81

2. Blinding: Who, what, when, why, how? Karanicolas et al. Can J Surg 2010;53:345–8 3. Clinical practice guideline: management of acute pancreatitis Greenberg et al. Can J Surg 2016;59:128–40 4. Hardware removal after tibial fracture has healed Sidky and Buckley Can J Surg 2008;51:263–8

5. Surgical approach in primary total hip arthroplasty: anatomy, technique and clinical outcomes Petis et al. Can J Surg 2015;58:128–39 6. Complications associated with laparoscopic sleeve gastrectomy for morbid obesity: a surgeons’ guide Sarkosh et al. Can J Surg 2013;56:347–52 7. Treatment of an infected total hip replacement with the PROSTALAC system Scharfenberger et al. Can J Surg 2007;50:24–8 8. Pharmacological management of postoperative ileus Zeinali et al. Can J Surg 2009;52:153–7 9. De�ning medical error Grober and Bohnen Can J Surg 2005;48:39–4410. Tracheostomy: from insertion to decannulation Engels et al. Can J Surg 2009;52:427–33 *Based on page views on PubMed Central of research, reviews, commentaries and discussions in surgery. Updated May 16, 2018.



DISCUSSIONS IN SURGERY • DISCUSSIONS EN CHIRURGIE

Users’ guide to the surgical literature: how to assess a qualitative study

T he randomized controlled trial (RCT) and the systematic review of RCTs, both forms of quantitative research, represent the gold standard in clinical research.1 While quantitative research seeks to establish con-

clusions through causal determination, predictions and statistical analysis, it is limited in its examination of perspectives, attitudes and beliefs of individual participants in favour of objective, numerical data.1,2

Unlike quantitative research, qualitative research contributes to the literature through the observation, description and interpretation of theories about social interactions and individual experiences as they occur in their natural setting.3 This type of research has the potential to enhance the understanding of sur-geons’ and patients’ preferences, attitudes and beliefs, as well as assess how these may change with time.

While qualitative research methods have been well documented since 1985, there is growing recognition that this study design is well suited for the surgical literature.4 In 2016, Maragh-Bass and colleagues5 performed an analysis on publi-cation trends in surgical literature that showed that qualitative surgical research has gained in popularity, representing more than half of all articles published in 32 surgical journals since 2011. Recent examples of qualitative studies within the surgical literature include the role of salespeople in the surgical suite,6 the per-spectives of orthopedic surgeons on patient candidacy for total joint arthroplasty7 and the postsurgical barriers to exercise in the bariatric patient.8 Despite the popu larity of qualitative research, such studies are often considered low-level evi-dence as they are routinely likened to case reports, expert opinions or anecdotal �ndings owing to a lack of familiarity with its methods.1,5,9 This thinking is ulti-mately misleading, as there is ample evidence within the literature to suggest that qualitative research has a useful role to play in the surgical domain.5

To date, there is no widely accepted standard for the methodological assess-ment of qualitative research.10 Despite ongoing debate, this article seeks to familiarize surgeons with the basic techniques for the critical appraisal of quali-tative studies in the surgical literature.

CLINICAL SCENARIO

A 57-year-old construction worker, who had a total knee replacement (TKR) by another surgeon at a peripheral hospital, has not been able to return to work

Lucas Gallo, BHSc, MD Jessica Murphy, MSc Luis H. Braga, MD, PhD Forough Farrokhyar, PhD Achilleas Thoma, MD, MSc

Accepted Oct. 20, 2017

Correspondence to: A. Thoma Department of Surgery McMaster University 206 James St. South Suite 101 Hamilton ON L8P 3A9 [email protected]

DOI: 10.1503/cjs.013117

Qualitative research contributes to the medical literature through the obser-vation, description and interpretation of theories about social interactions and individual experiences as they occur in their natural setting. This type of research has the potential to enhance the understanding of surgeons’ and patients’ preferences, attitudes and beliefs, as well as assess how these may change with time. To date, there is no widely accepted standard for the methodological assessment of qualitative research. Despite ongoing debate, this article seeks to familiarize surgeons with the basic techniques for the critical appraisal of qualitative studies in the surgical literature.

SUMMARY

userguie-thoma.indd 208 2018-05-18 12:34 PM

DISCUSSIONS IN SURGERY


after 6 months. You, an orthopedic surgeon, and your senior orthopedic resident review the operative record as well as the pre- and postsurgery radiographs; you cannot �nd any-thing wrong with the previously performed surgery. Your resident informs you that the patient was upset that he was on the waiting list for 2 years and was in constant pain while he continued working the year before surgery. His surgeon sent him for 7 physiotherapy sessions, after which the patient was told he was �t to return to work 6 weeks after surgery. You and your resident decide to investigate if there is any research that delves into the issue why some patients are unable to return to work after seemingly well-performed TKRs; you plan to present this case and your �ndings at the next orthopedic surgery grand rounds.

LITERATURE SEARCH

As the research question explores an underlying social phenomenon — the factors in�uencing a patient’s deci-sion to return to work following TKR — the ideal article type would be a qualitative study. You use MEDLINE to perform a literature search.11

Deriving keywords from the research question, you use the medical subject heading (MeSH) “qualitative research,” along with the search terms “total knee replacement” AND “return to work”; this search yields 2 qualitative studies.12,13

One of these articles is a systematic review assessing the in�uence of patient factors on employment following hip and knee replacement.12 Although this article references the topic of interest, it is not speci�c to TKR and does not cite the qualitative literature. As the data are presented from quantitative sources only, the study does not provide infor-mation regarding the patient perspective, experiences and social interactions associated with return to work following TKR. The second article is a qualitative study by Bardgett and colleagues13 published in 2016. You determine this is the only article to address the question posed in the clinical scenario. You decide to critically appraise this article.

A synopsis and the characteristics of the study by Bardgett and colleagues13 can be found in Table 1 and Table 2, respectively.

Are the results valid?

Creswell14 described 9 common characteristics of qualitative research: • Natural setting — data are collected directly from partici-

pants in the setting where they experience the phenomenon.• Researcher as key instrument — researchers collect data

through direct observation, interviews and the examination of documents as opposed to questionnaires or alternative instrumentation.

Table 1. Key features from the study by Bardgett and colleagues13*

Objective To gain insight in patients’ perspectives of the factors influencing return to work following knee replacement.

Setting A single secondary care setting in a large teaching hospital in northern England.

Methods Ten semistructured interviews. The interviews were transcribed and analyzed using a qualitative thematic approach.

Results Three themes identified: delays in surgical intervention, limited and often inconsistent advice from health care profession-als regarding return to work, and absence of rehabilitation to optimize recovery and facilitate return to work.

Conclusion The identified themes all contribute to potential delays in successful return to work. There is a need to tailor health care intervention to this cohort of patients to optimize outcomes.

*Similar to the appraisal of the quantitative studies, a specific framework is needed to evaluate a qualitative study. Box 1 outlines a series of questions that can be used to appraise a qualitative article. This framework assesses the quality of the study methods, examines the credibility of the results, and determines the applicability of the study’s conclusions to your patient.

Table 2. Demographic characteristics of participants in the study by Bardgett and colleagues13

Patient SexAge at time

of surgery, yrTime lapse between

surgery and interview, mo Type of employment Sector

Time taken to return to work,

wk

1 Male 58 9 Information technology Private 12

2 Female 49 21 Supermarket assistant Private 13

3 Female 55 14 General practitioner receptionist

Public 12

4 Male 59 21 Project engineer Private 6 from home10 at workplace

5 Male 59 11 Self-employed manager

Private 6

6 Female 57 25 Teacher Public 6

7 Female 47 23 Self-employed farmer Private 8

8 Male 58 35 Estates officer Public 12

9 Female 40 20 Administration Public 2 from home5 at workplace

10 Male 55 8 HGV driver Private 10


DISCUSSIONS EN CHIRURGIE


• Multiple sources of data — interviews, observations, and available documentation are used to collect data.

• Inductive data analysis — data are organized from simple to more abstract concepts to develop conclusions and themes.

• Participants’ meaning — meaning is derived from partici-pants directly without involvement of study researchers.

• Emergent design — the research process may be subject to change once data extraction begins and new information is obtained.

• Theoretical lens — the study and its conclusions are viewed through the appropriate social, political, or histori-cal context.

• Interpretive inquiry — researchers and participants inter-pret the data based on their own personal background and prior understanding.

• Holistic account — researchers use the data to identify and describe the “big picture” of the topic under inves-tigation.2,14

For qualitative research to be relevant to the practising surgeon, it must provide an understanding of a social phe-nomenon that has previously gone unrecognized or offer new insight into an already familiar area of social interaction.10

Although personal opinions and hypotheses have their own place in the literature, qualitative research represents a structured methodological approach to further one’s under-standing of a social interaction or personal experience as well as develop theories regarding its cause and overall relevance.1

In qualitative research, validity (as it refers to the idea of discovering truth) is not a single, �xed, or universal con-cept.15 As a result, qualitative researchers have replaced the word “validity” in favour of more appropriate terminology, such as credibility, rigour and trustworthiness.15,16

Was the formulation of the research question clearly described?Just as in quantitative research, the research question in qualitative research tells the reader what is going to be discovered, generated, or explored.17 Often, this research question is divided into two types: the central question, which is the most general question to ask, and the sub-questions, which subdivide the central question into spe-ci�c questions.17 Often in qualitative research the question will look at the “how” or “what” in order to describe or understand a situation or process.17

Bardgett and colleagues13 provide a brief background into previous research in the area, stating that they focused only on the quantitative side of returning to work.13 The authors explain that a qualitative approach to the research question is needed to obtain information directly from the patients based on their attitudes and perspectives.

The central research question for the appraised article is why certain factors may influence a patient’s return to work.13 The subquestion was to look at the potential de�-ciencies in the delivery of care directly pertinent to return

to work.13 The central and subquestions indicate that the aim of study was to gain a greater insight into the factors in�uencing return to work from the patient’s perspective.

Was the rationale for participant selection and observation sufficiently explained?Qualitative studies do not have a predetermined sample size or a statistical method to assess the appropriateness of the chosen sample size.10,18 To address this, participants are selected to meet particular criteria — a process referred to as “purposive sampling.”10,18 The criteria for participant selection may evolve during the course of data extraction and analysis to include typical, unusual or important cases to explore new themes and perspectives as they emerge. To account for this, relevant information, such as religion, socioeconomic status and profession, should be acknow-ledged by the researcher in order to establish the appropri-ate context for the sample’s perspectives.19

Bardgett and colleagues13 explain that participants were taken from a cohort of 50 patients, all under the age of 60 years at the time of surgery; the patients were originally recruited from a postal questionnaire study. From those 50 patients, 37 who were employed before surgery con-sented to participate. Purposive sampling was then used to select 10 patients who represented “a range of characteristics known to in�uence rates of return to work,” including age, sex and type of employment.13 Although the authors explain why the chosen patients were used for this study, they do not indicate why a sample size of 10 patients was used.

Were data collection methods and instruments adequately explained?As previously stated, in order to present a thorough assessment, researchers should use multiple sources; typ-ically, qualitative studies use 3 basic techniques for data collection: • Observation — the direct or indirect surveillance of

study participants10,20

• Participant interviews — structured, semistructured, or unstructured discussions to enable participants to describe a phenomenon as they experience it21

• Document analysis — the direct examination of relevant information linked to the topic under investigation (i.e., medical charts, operative reports).10,22 It has been suggested that in the case of an interview, the

characteristics of the interviewer be given to the reader, along with any relationship that the interviewer may have with the participants.23 This information allows the reader to conclude how the interviewer could possibly in�uence the responses of the participants and, furthermore, how they may lead the interview.23

Bardgett and colleagues13 used a semistructured patient interview focusing on the preoperative and early postopera-tive phases of the patients’ journey (see Box 1 for interview questions). The authors justify this data collection method,




stating that patients were able to discuss topics that were directly relevant to them, even those not previously con-sidered by the research team. The interviews had the fol-lowing limitations. First, the time at which the interview took place following surgery, which ranged from 8–35 months, introduced the possibility of recall bias; therefore, responses and recollections from the participants may have been affected. Second, the authors did not de�ne the criteria for “early postoperative phase”; therefore, the reader is unaware of speci�c timelines. Finally, the reader is not given adequate information regarding the interviewer. Bardgett and colleagues13 explain that it was a research-trained physiotherapist; however, it is unknown whether this physiotherapist was a third-party individual or if he or she was treating the patients.

In summary, Bardgett and colleagues13 used only one form of data collection, did not disclose the characteristics of the interviewer, and did not state if there was a prior rela-tionship between the interviewer and the participants. This could introduce some inconsistency into the data and ulti-mately affect how the reader interprets the results.

Was the data extraction of sufficient detail and scope?Tong and colleagues23 explain that the following informa-tion should be disclosed to readers regarding data extraction: the number of coders, a description of how the themes were derived, a mention of any software that was used for the extraction of data and a mention of whether the interviewees were able to review their transcripts.

The study by Bardgett and colleagues13 states that one author took the audio recordings and “transcribed at ver-batim.” It was stated that after the first coder listened repeatedly to the recordings, a second coder revisited them in order to verify the initial codes and themes found by the �rst reviewer. Therefore, the article did list how many coders were involved in data extraction. In regard to how the themes were derived, Bardgett and colleagues13 stated that a thematic approach was used to identify common pat-terns among the interviews. The authors state that this type of analysis allowed for the identi�cation, analysis and reporting of patterns found across the data set, while work-ing with themes that were identi�ed a priori.13 Although not explicitly stated, the authors did say that each tran-

script was coded by hand; therefore, one could assume that no software was used during the process. Finally, as there was no mention, one can assume that the interviewees did not review their transcripts.

Were the method and credibility of data synthesis, interpretation and presentation sufficiently explained?A variety of analytical approaches may be used to interpret qualitative studies, including ethnography, grounded theory, phenomenological analysis and content analysis. Each method, although unique, ultimately attempts to develop a theory or narrative from the qualitative data.10 Although the details of each analytical method are beyond the scope of this article, in this section we review the gen-eral features of data analysis that are relevant to most methods — triangulation and data saturation.

In the context of qualitative analysis, triangulation refers to the act of verifying outcomes through multiple sources of information. Denzin24 and Patton25 identi�ed 4 types of triangulation:• Method — the use of multiple forms of data collection

(i.e., interviews, direct observation and �eld notes) to gather information about the same phenomenon26

• Investigator — the use of two or more researchers to com-pare several different observations and conclusions24,26

• Theory — the use of existing social science theories to substantiate or refute �ndings of a qualitative study10,26

• Data source — using data from different participants (i.e., individuals, groups, or communities) in order to compare and contrast multiple perspectives and validate data.26

The theory behind data saturation is fundamental to qualitative research. Speci�cally, it refers to the point in which new data �t into an identi�ed theory without the need for revision.10 Although there is no single approach to reaching data saturation, researchers have agreed on the following principles to identify data saturation: no new data, no new themes, no new coding, and the ability to replicate the study.27,28 Failure to reach data saturation suggests that not all relevant themes or perspectives may be represented in study outcomes, ultimately jeopardizing the credibility of the study’s results.

Although Bardgett and colleagues13 do not cite triangula-tion directly, they do use a form of investigator triangulation: they used multiple researchers to compare observations and conclusions, thereby strengthening the reader’s con�dence that the data were interpreted and coded appropriately and instilling con�dence in the results.

In regard to data saturation, Bardgett and colleagues13 state that the aim of the study was not to reach data satura-tion, but to “identify important key themes.” Although the authors give an explanation, the failure to use data satura-tion as an end point suggests that all relevant themes may not be adequately accounted for in the data and, therefore, that the credibility of these �ndings may be questioned.

Box 1. Semistructured interview guide from the study by Bardgett and colleagues13

Introduction1. Could you start by explaining to me what your job involves?2. How did your arthritis affect you/your work?3. What has happened since your operation?4. Was the experience after the operation what you were expecting?5. What was the involvement of your employer?6. Is there anything that has helped you or would have helped you to return to work more easily?7. What influenced the decision to return to work at that time?Close




What are the results?

Qualitative research relies on suf�cient detail on the part of the study authors to elicit a clear picture of the perspec-tives and experiences of study participants.10 To do this, researchers may opt to supplement their analysis with excerpts from interview transcripts, observation notes and relevant documentation. Use of excerpts allows the reader to judge for themselves whether the results accurately re�ect the data and assess whether the data clearly supports the study’s conclusions.10 Conversely, if the study conclu-sions are not represented by the data excerpts, the reader may doubt the interpretation skills of the study author or the methods for analysis.

Bardgett and colleagues13 identi�ed three themes that influenced a patient’s return to work following TKR (Table 1). Each theme references direct excerpts from interview transcripts to substantiate the authors’ interpre-tations and conclusions.

Were participants’ characteristics clearly presented?In qualitative research, a description of the basic demo-graphic characteristics of study participants is necessary in order to properly interpret the �ndings of the research.23 As the act of being observed and/or reordered will undoubtedly affect study participants, it is important to know the characteristics of both the participants and the interviewer so that the reader may understand how the two groups may in�uence each other, ultimately affecting the study results.23

With regard to demographic information, Bardgett and colleagues13 supply the reader with information on sex, age at the time of surgery and type of employment for each participant (Table 2). This information allows the reader to interpret the results in the context of the patients’ age and employment background, all of which have been rec-ognized as important factors in�uencing return to work.13

Were study-derived constructs and their credibility adequately explained?“Constructs” refer to mental abstractions that attempt to convey meaning about a particular topic in only a few words. They provide a shared meaning that allows the study authors to communicate ideas clearly and precisely to their audiences.29 For example, the term “ageism” is a construct for prejudice or discrimination on the basis of a person’s age. As constructs are not directly observable —we cannot directly observe ageism even though we may associate ageism with different signs or actions — clear de�nitions are required to ensure conceptual clarity and good qualitative research.29

It has been suggested that direct quotations may be used to de�ne and support the use of constructs in the qualita-tive literature.23 The incorporation of direct quotations from participants is said to add “transparency and trust-

worthiness” to both the raw data and to the interpretation of the �ndings.23 Quotations allow the reader to assess how the data (in the form of quotations) relate to the conclu-sions made by the study authors.

In the article assessed in our clinical scenario,13 direct quotations from participants are used to de�ne and support each construct claimed by the authors. The themes they identi�ed are also clearly stated, and the quotations used are well integrated with additional information supplied to the reader.13 This article also lists the study number of the respective participant after each quotation, as suggested by Tong and colleagues.23 The reader is able to see a clear association between the author’s interpretation and the interview transcript excerpt, ultimately strengthening the study �ndings and supporting the author’s analysis.

Were study constructs differentiated from pre-existing constructs?As previously stated, for qualitative research to be relevant to the practising surgeon, it must provide an understand-ing of a social phenomenon that has previously gone unrecognized or offer new insight into an already familiar area of social interaction.10 Although empirically developed constructs do not need to agree with existing theories or beliefs, it is helpful to the reader if the authors relate these constructs to the prevailing knowledge and the existing literature.10

Bardgett and colleagues13 effectively relate their derived themes to pre-existing constructs de�ned in the literature. For example, when referencing “delays in surgical inter-vention” as a construct affecting return to work following TKR, Bardgett and colleagues13 express how this theme relates to the “physician–patient relationship” previously established in the literature. By differentiating from pre-existing constructs, authors are able to expand on theories already established in the literature.

Was the transferability/generalizability of the results discussed?Within the qualitative literature, the term “transferability” is synonymous with external validity and refers to the application of the study �ndings beyond the setting in which the study was conducted.9 Given the contextual nature of qualitative research insofar that it records social interactions and personal experiences within a given set-ting, the reader must carefully assess the transferability of study results to other sociocultural settings.19

There were many limitations identi�ed that may affect the transferability of the study results in the article by Bardgett and colleagues.13 First, as their study was based out of a large teaching hospital, the �ndings may be appli-cable only to severe cases requiring interventions at an aca-demic as opposed to a community hospital or clinic. In addition, all study participants were white British citizens younger than 60 years who were working at the time of




surgery. This raises the question as to whether the results could be generalized to less severe cases requiring TKR and to patients of different ethnic backgrounds, ages and employment status, which may have associated factors in�uencing return to work.

Were the data reported according to accepted guidelines?The Standards for Reporting Qualitative Research (SRQR) developed by O’Brien and colleagues make up a 21-item checklist endorsed by the EQUATOR Network.3,30 The SRQR functions to improve the transparency of qualitative research by establishing a list of recommendations that should be represented in all qualitative studies.3 These standards were developed with the intent to assist readers, editors and reviewers to assess qualitative study methods and the applica-tion of its results.

Applying the SRQR to the study by Bardgett and col-leagues13 shows that 20 of the 21 items are represented; “researcher characteristics and re�exivity” is the only item not reported. The article by Bardgett and colleagues has good overall adherence to the SRQR reporting standards and therefore shows suf�cient transparency.

Will the results help me care for my patient?

Are the results of this study applicable to my situation?As previously stated, there were limitations identi�ed in the study by Bardgett and colleagues13 that may affect the trans-ferability of data to real-world patients. It is the responsibility of individual physicians to consider how their patients may be similar to the patients referenced in the study population.

Does the study present a compelling theory?For qualitative research to be useful, it must be believable. As a result, the utility of a qualitative study depends on the nar-rative and arguments it presents. To assess this, Elder and Miller10,31 proposed the following characteristics to be evalu-ated by the reader: parsimony (the use of minimal assump-tions to explain the data), consistency (whether the study conforms with the existing literature or presents reasons for its disagreement), fertility (whether the study presents areas for future research) and clarity (whether the study’s narrative is free of redundancy, ambiguity, or contradiction).31 Simply, the reader must ask themselves whether the study makes sense and whether the account is compelling.31

The article by Bardgett and colleagues13 demonstrates consistency, fertility and clarity. Speci�cally, it describes its role within the existing literature, suggests areas for future research and presents a clear and concise narrative that is substantiated with patient transcriptions. Conversely, the study uses multiple unsupported assumptions to justify patient selection and therefore does not demonstrate appropriate parsimony. For example, patients younger

than 60 years and working at the time of surgery may not be representative of the population seeking to return to work following TKR.

RESOLUTION OF THE SCENARIO

At the next orthopedic surgery grand rounds, you present this case and your �ndings to your colleagues. You suggest that although the qualitative study by Bardgett and col-leagues13 identi�es three key themes that may in�uence this patient’s experience of return to work following TKR, including delays in surgical intervention, limited and incon-sistent advice among health care providers to optimise return to work and the provision of rehabilitation to opti-mise recovery and return to work, you admit that further analysis may be required to refine and/or revise these theories, given that the study authors did not reach the point of data saturation and investigated only a small sam-ple at a single site. You suggest that you and your col-leagues initiate a national qualitative study to assess patient factors in�uencing return to work following TKR.

CONCLUSION

The purpose of qualitative research is to better understand the social interactions and perspectives of individuals within a given setting. As no widely accepted standard for appraisal exists, readers should use this guideline (Box 2) not as a �nite checklist, but rather as a guide to familiarize themselves with qualitative studies. As the critical appraisal of qualitative research may differ according to the speci�c methodology used, readers should reference a text speci�c to the qualitative study method, just as Bardgett and col-leagues13 cited the article by Braun and Clarke32 as the basis for their thematic analysis, when seeking a more thorough appraisal of the literature.

Box 2. Guidelines for how to assess a qualitative research articleA. Are the results valid?

• Was the research question formulation clearly described?• Was the rationale for participant selection and observation sufficiently

explained?• Were data collection methods and instruments adequately explained?• Was the data extraction of sufficient detail and scope?• Were the method and credibility of data synthesis, interpretation and

presentation sufficiently explained?B. What are the results?

• Were participants’ characteristics clearly presented?• Were study-derived constructs and their credibility adequately explained?• Were study constructs differentiated from pre-existing constructs?• Was the transferability/generalizability of the results discussed?• Were the data reported according to accepted guidelines? C. Will the results help me in caring for my patient?

• Are the results of this study applicable to my setting?• Does the study present a compelling theory?




Af�liations: From the Faculty of Health Sciences, McMaster University, Hamilton, ON (Gallo); the Division of Plastic surgery, McMaster University, Hamilton, ON (Thoma, Murphy); the Division of Urology, McMaster Uni-versity, Hamilton, ON (Braga); the Dept of Surgery, McMaster University, McMaster University, Hamilton, ON (Braga, Farrokhyar, Thoma); and the Dept of Health Research Methods, Evidence, and Impact (Formerly Dept of CE&B), McMaster University, Hamilton, ON (Farrokhyar, Thoma).



References

1. Shuval K, Harker K, Roudsari B, et al. Is Qualitative research second class science? A quantitative longitudinal examination of qualitative research in medical journals. PLoS ONE 2011;6:e16937.

2. Cypress B. Qualitative research: the “what,”“why,”“who,” and “how”! Dimens Crit Care Nurs 2015;34:356-61.

3. O’Brien B, Harris I, Beckman T, et al. Standards for reporting quali-tative research. Acad Med 2014;89:1245-51.

4. Morse J. Qualitative health research. Qual Health Res 2015;25:3. 5. Maragh-Bass A, Appelson J, Changoor N, et al. Prioritizing qualitative

research in surgery: a synthesis and analysis of publication trends. Surgery 2016;160:1447-55.

6. O’Connor B, Pollner F, Fugh-Berman A. Salespeople in the surgical suite: relationships between surgeons and medical device representa-tives. PLoS ONE 2016;11:e0158510.

7. Frankel L, Sanmartin C, Hawker G, et al. Perspectives of orthopaedic surgeons on patients’ appropriateness for total joint arthroplasty: a qualitative study. J Eval Clin Pract 2015;22:164-70.

8. Peacock J, Sloan S, Cripps B. A qualitative analysis of bariatric patients’ post-surgical barriers to exercise. Obes Surg 2013;24:292-8.

9. Malterud K. Qualitative research: standards, challenges, and guidelines. Lancet 2001;358:483-8.

10. Guyatt G, Rennie D, Meade M, et al. Users’ guides to the medical liter-ature: A manual for evidence-based clinical practice. 3rd ed. New York (NY):McGraw-Hill; 2015.

11. Waltho D, Kaur M, Haynes B, et al. Users’ guide to the surgical litera-ture: how to perform a high-quality literature search. Can J Surg 2015;58:349-58.

12. Malviya A, Wilson G, Kleim B, et al. Factors in�uencing return to work after hip and knee replacement. Occup Med (Chic Ill) 2014;64:402-9.

13. Bardgett M, Lally J, Malviya A, et al. Return to work after knee replacement: a qualitative study of patient experiences. BMJ Open 2016;6:e007912.

14. Creswell JW. Qualitative Inquiry and Research Design: Choosing Among Five Approaches. 3rd ed. Thousand Oaks (CA): Sage; 2012.

15. Golafshani N. Understanding reliability and validity in qualitative research. Qual Rep 2003;8:597-607.

16. Creswell J, Miller D. Determining validity in qualitative inquiry. Theory Pract 2000;39:124-30.

17. Creswell JW, Clark VLP. Principles of qualitative research: designing a qualitative study. [Powerpoint Slides] Lincoln (NE): Of�ce of Qual-itative & Mixed Methods Research, University of Nebraska; 2004.

18. Sandelowski M. Sample size in qualitative research. Res Nurs Health 1995;18:179-83.

19. Kuper A, Lingard L, Levinson W. Critically appraising qualitative research. BMJ 2008;337:a1035-a1035.

20. Laitinen H, Kaunonen M, Åstedt-Kurki P. Methodological tools for the collection and analysis of participant observation data using grounded theory. Nurse Res 2014;22:10-5.

21. Gill P, Stewart K, Treasure E, et al. Methods of data collection in qualitative research: interviews and focus groups. BDJ 2008; 204:291-5.

22. Bowen G. Document analysis as a qualitative research method. Qual-itative Research Journal. 2009;9:27-40.

23. Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups. Int J Qual Health Care 2007;19:349-57.

24. Denzin N. Sociological methods: a sourcebook. New York (NY): McGraw-Hill; 1978.

25. Patton M. Enhancing the quality and credibility of qualitative analy-sis. Health Serv Res 1999;34:1189-1208.

26. Carter N, Bryant-Lukosius D, DiCenso A, et al. The use of triangula-tion in qualitative research. Oncol Nurs Forum 2014;41:545-7.

27. Fusch P, Ness L. Are we there yet? Data saturation in qualitative research. Qual Rep 2015;20:1408-16.

28. Guest G, Bunce A, Johnson L. How many interviews are enough? An experiment with data saturation and variability. Field Methods 2006;18:59-82.

29. Laerd Dissertation. Constructs in quantitative research [Internet]. Available: http://dissertation.laerd.com/constructs-in-quantitative-research.php (accessed 2017 Aug. 25).

30. The EQUATOR Network. Enhancing the QUAlity and Trans-parency Of Health Research [Internet]. Available: www.equator-network.org/ (accessed 2017 July 22).

31. Elder N, Miller W. Reading and evaluating qualitative research studies. J Fam Pract 1995;41:279.

32. Braun V, Clarke V. Using thematic analysis in psychology. Qual Res Psychol 2006;3:77-101.



The Department of Surgery at The University of British Columbia (UBC) invites applications for an academic surgeon to be appointed at the rank of Assistant Professor, Tenure Track. A clinical appointment at a Vancouver based hospital in the relevant surgical specialty is anticipated.

The UBC Department of Surgery has training programs at the undergraduate, graduate and postgraduate levels, and pursues research to make innovative advancements in knowledge and practice to improve health. The Department consists of more than 400 physicians and scientists as well as over 70 administrative, research and technical staff. Specialty training programs are offered in Cardiac Surgery, Colorectal Surgery, General Surgery, Neurosurgery, Otolaryngology, Pediatric Surgery, Plastic Surgery, Radiation Oncology, Surgical Oncology, Thoracic Surgery and Vascular Surgery. The successful candidate will join a dynamic clinical and research enterprise in the domains of diabetes, transplantation, and biomedical engineering, and will have access to research partnerships at Vancouver Coastal Health Research Institute and the BC Children’s Hospital Research Institute.

Reporting to the Head of the UBC Department of Surgery, the successful candidate will develop a research program in translational transplantation and regenerative medicine and will be expected to compete successfully for external grant funding. Areas of particular interest could include stem cell biology, immune tolerance, and cell engineering. The successful candidate will be expected to participate in the teaching activities of the Department, as well as provide mentorship and training to undergraduate, graduate and post-graduate trainees.

The successful candidate will hold an MD or MD/ Ph.D. and should have or be eligible for certification from the Royal College of Physicians and Surgeons of Canada in a surgical discipline. The successful candidate will have training and experience in a field of translational transplantation research, regenerative medicine, or cell therapy, have demonstrated ability to achieve excellence in research and teaching, and have a commitment to academic service.

Salary will be commensurate with qualifications and experience. A letter of application outlining the applicant’s research and teaching interests, accompanied by a detailed curriculum vitae and names of three references should be directed to:

Karen Larsen Human Resources Manager, UBC Department of Surgery

Email [email protected] with subject line: Assistant Professor position

Review of applications will begin on June 1, 2018 and continue until the position is filled. The anticipated start date for this position is January 1, 2019 or upon a date to be mutually agreed.

Equity and diversity are essential to academic excellence. An open and diverse community fosters the inclusion of voices that have been underrepresented or discouraged. We encourage applications from members of groups that have been marginalized on any grounds enumerated under the B.C. Human Rights Code, including sex, sexual orientation, gender identity or expression, racialization, disability, political belief, religion, marital or family status, age, and/or status as a First Nation, Metis, Inuit, or Indigenous person. All qualified candidates are encouraged to apply; however Canadians and permanent residents will be given priority.

S-8

01

THE UNIVERSITY OF BRITISH COLUMBIAFaculty of Medicine

Assistant Professor (Tenure Track) | Department of Surgery

CAREER OPPORTUNITY

med.ubc.ca | surgery.ubc.ca

The Division of Cardiac Surgery at the London Health Sciences Centre, University Campus, seeks a full-time surgical assistant to join us in providing expert preoperative, intraoperative, and postoperative care. Surgical procedures comprise a wide variety of cases, including off pump coronary artery bypass grafting, minimally invasive and robotic cardiac procedures. Candidates must have some previous surgical experience, be eligible for licensure in the province of Ontario and be prepared to work as part of a team. Mentorship will be provided by existing group of surgical assistants, as well as by cardiac surgery residents and consultants.

Please send c.v. and references to:

Marsha Goldie Assistant to the Division of Cardiac Surgery

London Health Sciences Centre University Campus

339 Windermere Road, London, ON N6A 5A5 E-mail: [email protected]

S-8

04

SURGICAL ASSISTANT CARDIAC SURGERY

The Canadian Journal of Surgery is pleased to accept career/ classified advertisements. The deadline is 1 month before issue date.

Rates: Display ads: 1 page $1200; 2/3 page $900; 1/2 page vert/horiz $800; 1/3 page $650; 1/4 page $500. Word ads: $120 for the first 40 words or less, additional words $1.20 each (additional $25 for frame). Special Display under 100 words, 55 × 55 mm, $205.

VISA, MASTERCARD AND AMERICAN EXPRESS ACCEPTED.

Advertisements should be sent to: email [email protected]; tel 800 663-7336 or 613 731-8610 x8460/8475.

The Ontario Human Rights Code prohibits discriminatory employ ment advertising.

CAREER/CLASSIFIED ADVERTISING

classifiedcjs-jun18.indd 215 2018-05-18 11:19 AM


The Department of Surgery at Sunnybrook Health Sciences Centre (SHSC) is inviting applications for one full-time surgeon with advanced fellowship training in upper extremity surgery. The effective date of this appointment is as early as January 2019, or beyond.

Clinically, the successful candidate will provide advanced surgical care with a predominant focus on upper extremity surgery within Sunnybrook’s Holland Musculoskeletal Program, with expected contributions to the hospital’s trauma program as well as Working Condition Program (WCP).

The ideal candidate will possess a graduate level degree relevant to academic orthopaedic surgery, preferably with training in health services research, bioengineering, or imaging technologies, or a significant track record of research in an area related to orthopedic surgery. Evidence of excellence in teaching and research is required.

Applicants must have certification in Orthopedic surgery from the Royal College of Physicians and Surgeons of Canada (or equivalent). Candidates must be eligible for appointment at the rank of Assistant Professor or higher at the University of Toronto and be eligible for certification with the Royal College of Physicians and Surgeons of Canada, as well as licensure with the College of Physicians and Surgeons of Ontario.

SHSC is one of Canada’s largest leading clinical and research hospitals. SHSC is a fully affiliated academic health sciences facility of the University of Toronto and the home to one of the University’s three Academies for undergraduate education. SHSC has an international reputation for clinical excellence and research and is an acknowledged academic leader. The Sunnybrook Research Institute hosts well established programs in basic, applied, and clinical research occupying over 250,000 sq.ft. of state-of-the-art infrastructure with annual peer review grant support of $75M.

The Department of Surgery at SHSC performs over 15,000 surgical cases annually, with Orthopaedic Surgery comprising ~ 5000 annual surgical cases. As one of the largest trauma centres in Canada, over 1700 ‘trauma code’ patients are managed through our Emergency Department annually. The hospital’s Department of Surgery remains integral to Sunnybrook’s ecosystem supporting its Trauma activity, Cancer Program, and provides comprehensive care for patients with upper extremity surgical disorders.

Applications will be accepted until July 1, 2018. Please submit a letter of intent, curriculum vitae and the name of three referees to:

Attention: Ms. Carolyn Gimera Dr. Albert Yee, MD, FRCSC

Holland MSK Program Chief, and Marvin Tile Chair, Division Head Orthopaedic Surgery Sunnybrook Health Sciences Centre

2075 Bayview Avenue, Room MG 371-B Toronto, Ontario, Canada M4N 3M5

Tel 416-480-6815 • Fax 416-480-4395 Email [email protected]

Sunnybrook is strongly committed to diversity within its community and especially welcomes applications from visible minority group members, women, Aboriginal persons, persons with disabilities, members of sexual minority groups, and others who may contribute to the further diversification of ideas. All qualified candidates are encouraged to apply; however, Canadians and permanent residents will be given priority.

For more information about the Faculty of Medicine/Department of Surgery, please visit our home page at http://surgery.utoronto.ca

ACADEMIC UPPER EXTREMITY/TRAUMA SURGERYDepartment of Surgery, Sunnybrook Health Sciences Centre & University of Toronto

S-8

02

classifiedcjs-jun18.indd 216 2018-05-18 11:19 AM

For more information, please visit

joule.cma.ca/inhouse

98%course

satisfactionrate

Bringing courses from the Physician Leadership Institute (PLI) in-house to your organization helps build the skills of individuals while strengthening the capacity and impact of your entirehealth care team.

Reduce travel related costs and time

Provide a tailored learning experience based on the needs of your organizatio

Are available to the entire health care team—physicians and non-physicians

Accredited by CFPC and CCHL

Delivered by physicians and other subject matter experts joule.cma.ca/inhouse

Bring the Physician Leadership Institute in-house

BARIATRIC SURGERY – ACADEMIC GENERAL SURGEON

The Department of Surgery (www.surgery.queensu.ca) at Queen’s University is seeking an Academic General Surgeon for a geographically full-time position. The candidate will have Fellowship training in bariatric surgery and have advanced minimally invasive surgical skills. The preferred candidate will be an established academic surgeon and must have an excellent record of success working in a multi-disciplinary clinical environment. They must also demonstrate the strong potential for outstanding educational contributions and exhibit significant academic and scholarly research.

Qualified applicants will hold an MD degree (or equivalent), have completed postgraduate/fellowship qualifications prior to appointment, and have or be eligible for licensure with the College of Physicians and Surgeons of Ontario and certification from the Royal College of Physicians and Surgeons of Canada. Preference will be given to candidates with an advanced degree.

The successful candidate will be required to collaborate with General Surgeons at the Kingston Health Sciences Centre and will also participate in the general surgery call schedule. As a full-time member of the Division of General Surgery, Department of Surgery, the successful candidate will also be expected to contribute to the teaching, supervision and mentorship of undergraduate, graduate and postgraduate students.

Queen’s University is recognized nationally for the quality of its undergraduate and graduate programs, which attracts outstanding students. Queen’s University is an integral part of the vibrant Kingston community in the heart of the Thousand Islands region of southeastern Ontario. It has a community spirit and amenities unmatched by any other Canadian university. The University and the region offer an outstanding academic and community environment (www.queensu.ca).

The University invites applications from all qualified individuals. Queen’s is committed to employment equity and diversity in the workplace and welcomes applications from women, visible minorities, Aboriginal peoples, persons with disabilities, and LGBTQ persons. All qualified candidates are encouraged to apply; however, in accordance with Canadian immigration requirements, Canadian citizens and permanent residents of Canada will be given priority. To comply with Federal laws, the University is obliged to gather statistical information about how many applicants for each job vacancy are Canadian citizens/permanent residents of Canada. Applicants need not identify their country of origin or citizenship, however, all applications must include one of the following statements: “I am a Canadian citizen/permanent resident of Canada”; OR, “I am not a Canadian citizen/permanent resident of Canada”. Applications that do not include this information will be deemed incomplete.

The University will provide support in its recruitment processes to applicants with disabilities, including accommodation that takes into account an applicant’s accessibility needs. If you require accommodation during the interview process, please contact the Department of Surgery at 613 533-2660.

The review of applications will begin on June 1, 2018 and will continue until the position is filled. Applications are to include a letter summarizing expertise, qualifications, and accomplishments relevant to the position, a CV, and the names and full contact information of three referees. Complete applications are to be directed to:

Dr. John Rudan, Head, Department of Surgery Queen’s University, Victory 3 Kingston General Hospital

Kingston, ON K7L 2V7 Fax 613 544-9174

Email [email protected]

S-8

03

classifiedcjs-jun18.indd 217 2018-05-17 2:46 PMcover-june18.indd 4 2018-05-22 9:17 AM

R7

Ad Number: MBZ_FLE_P27673_CMAJ_4Publication(s): Canadian Medical Association Journal

This ad prepared by: SGL Communications • 2 Bloor St. West, Toronto, Ontario • phone 416.413.7495 • fax 416.944.7883 File Location: SGL_MercedesBenzCanada:Magazine Ads:2018:O80824_Fleet E-Class Wagon:MBZ_FLE_P27673_CMAJ_4.indd

JOB SPECIFICS

Client: Mercedes-Benz CanadaCreative Name: CMAJ April issueAgency Docket #: MBZ FLE O80824Main Docket #: SMZ COR O80824Art Director: Paul WiersmaCopy Writer: NonePrint Production: Tom BurtonRetoucher: Jano KirijianLive: 7” x 9.5”Trim: 8.125” x 10.875”Bleed: 8.375” x 11.125”Artwork Scale: 1:1Print Scale: 100%

FILE SPECIFICATIONS:

File Name: MBZ_FLE_P27673_CMAJ_4.inddCreation Date: 4-4-2018 4:55 PMLast Modified: 4-11-2018 3:04 PMWorkstation: T16-0254InDesign Version: CC 2018 App. Version: 13.0.1Round #: 1 Page Count: 1GRAPHIC PRODUCTION:

Operator: Dennis SousaCorrection: None

SIGNOFFS:

Creative:

Production:

Premedia:

Proofreading:

Account:

Client:

PREMEDIA OPERATOR:

Operator: DS

INKS:

Cyan

MAGENTA

YELLOW

BLACK

FONTS & PLACED IMAGES

Family Style

Corporate A Condensed RegularCorporate S Light

File Name Colour Space Eff. Res (PPI)

S205_7_8_Front_v16_EN_39L_v2_Eng_mag.tif CMYK 326 ppiMB-star_M_n_4C.epsVorab_MB_WM-claimEN_LB_ML_n_1C.ai

This proof was produced by the following department:

PREPRESS

Just what the doctors order.Mercedes-Benz Corporate Sales is proud to continue a strong partnership with all Canadian MedicalAssociation members. Save up to $1,500 in addition to existing special offers on new Mercedes-Benz, AMG and smart vehicles. This competitive incentive refl ects our commitment to providing the best ownership experience to CMA members. See additional details at mercedes-benz.ca/cma

S:7”S:9.5”

T:8.125”T:10.875”

B:8.375”B

:11.125”

cover-june18.indd 1 2018-05-22 1:50 PM

gallbladder polyps: systematic review reward your...

Documents