carbon dioxide insufflation deflects airborne particles ... insufflation deflects airborn...

ww.sciencedirect.com

Journal of Hospital Infection xxx (2016) 1e6

Available online at w

Journal of Hospital Infection

journal homepage: www.elsevierhealth.com/journals / jhin

Carbon dioxide insufflation deflects airborne particlesfrom an open surgical wound model

P. Kokhanenko a, G. Papotti a, J.E. Cater b, A.C. Lynch c, J.A. van der Linden d,C.J.T. Spence a,*

a Fisher & Paykel Healthcare Ltd, Auckland, New ZealandbDepartment of Engineering Science, University of Auckland, Auckland, New ZealandcDepartment of Colorectal Surgery, Peter MacCallum Cancer Centre, Melbourne, AustraliadDepartment of Cardiothoracic Surgery and Anesthesiology, Karolinska University Hospital, Karolinska Institute, Stockholm,Sweden

A R T I C L E I N F O

Article history:Received 12 October 2016Accepted 14 November 2016Available online xxx

Keywords:InfectionFlow visualizationMedical deviceSurgical site infection

* Corresponding author. Address: 15 MEast Tamaki, Auckland 2013, New Zealand. T

E-mail address: [email protected]

http://dx.doi.org/10.1016/j.jhin.2016.11.0060195-6701/ª 2016 The Author(s). Published bunder the CC BY-NC-ND license (http://creat

Please cite this article in press as: Kokhanemodel, Journal of Hospital Infection (2016),

S U M M A R Y

Background: Surgical site infections remain a significant burden on healthcare systemsand may benefit from new countermeasures.Aim: To assess the merits of open surgical wound CO2 insufflation via a gas diffuser toreduce airborne contamination, and to determine the distribution of CO2 in and over awound.Methods: An experimental approach with engineers and clinical researchers wasemployed to measure the gas flow pattern and motion of airborne particles in a model ofan open surgical wound in a simulated theatre setting. Laser-illuminated flow visualiza-tions were performed and the degree of protection was quantified by collecting andcharacterizing particles deposited in and outside the wound cavity.Findings: The average number of particles entering the wound with a diameter of <5 mmwas reduced 1000-fold with 10 L/min CO2 insufflation. Larger and heavier particles had agreater penetration potential and were reduced by a factor of 20. The degree of pro-tection was found to be unaffected by exaggerated movements of hands in and out of thewound cavity. The steady-state CO2 concentration within the majority of the wound cavitywas >95% and diminished rapidly above the wound to an atmospheric level (w0%) at aheight of 25mm.Conclusion: Airborne particles were deflected from entering the wound by the CO2 in thecavity akin to a protective barrier. Insufflation of CO2 may be an effective means ofreducing intraoperative infection rates in open surgeries.

ª 2016 The Author(s). Published by Elsevier Ltdon behalf of The Healthcare Infection Society. This is an open access article

under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

aurice Paykel Place,el.: þ64 9 574 0123.o.nz (C.J.T. Spence).

y Elsevier Ltd on behalf of Tivecommons.org/licenses/by-

nko P, et al., Carbon dioxidehttp://dx.doi.org/10.1016/j.

Introduction

Surgical site infection (SSI) remains a significant cause ofmorbidity, prolonged hospitalization, increased cost of treat-ment, and mortality. SSIs range in severity from relatively

he Healthcare Infection Society. This is an open access articlenc-nd/4.0/).

insufflation deflects airborne particles from an open surgical woundjhin.2016.11.006

http://creativecommons.org/licenses/by-nc-nd/4.0/

mailto:[email protected]

www.sciencedirect.com/science/journal/01956701

http://www.elsevierhealth.com/journals/jhin

http://dx.doi.org/10.1016/j.jhin.2016.11.006

http://creativecommons.org/licenses/by-nc-nd/4.0/



P. Kokhanenko et al. / Journal of Hospital Infection xxx (2016) 1e62

trivial superficial wound exudate to causing, at least in part,one-third of postoperative deaths.1

There is much debate over sources of SSIs; however, a largeproportion of SSIs from low-infection-risk operations are thoughttobecausedbycontaminatedairborneparticles.2e5Whyteet al.suggested that 98% of bacteria in orthopaedic wounds camedirectly or indirectly from the air.6 Particles may settle directlyin the wound or be transferred to the wound after settling oninstruments or the surgeon’s hands. Infectious airborne particlesmay be (i) single bacterial cells or spores, fungal spores, or vi-ruses; (ii) aggregates of several cells, spores, or viruses; or (iii)pathogen-laden carrier particles such as skin scales, lint, andrespiratory droplets. These micro-organisms and particles rangein size from 0.02e0.3 mm viruses to 0.3e10 mm bacteria andbacterial clumps, and skin scales from <5 mm to 500 mm indiameter, with a thickness of about 2e5 mm.2,7e9

Airborne particles are ubiquitous and may carry viable mi-cro-organisms.10 Seal and Clark measured as many as5.6 million particles/m3 in a ventilated operating theatre dur-ing surgery.11 Gosden et al. found that operating theatreairborne contamination was dependent almost exclusively onthe number of people present and their activity within theroom.12 Skin scales are shed from the human body at a rate ofabout 7000 particles/min and may carry Staphylococcusaureus, which forms part of typical skin flora and is frequentlyresponsible for SSIs.13 Skin scales spread into the operatingroom from both patient and staff passing freely through thepores of tightly woven fabrics.8e10,14,15

Attempts to reduce dispersion of airborne particles andwound contamination include the use of low-turbulence flow(LTF) ventilation. Conventional ventilation systems create ahigh-efficiency particulate air (HEPA)-filtered unidirectionaldownward flow over the sterile field at a flow rate achieving�20 air changes per hour.4 Persson and van der Linden intro-duced a gas diffuser for CO2 insufflation of cardiothoracicwounds to prevent arterial air embolism in open heart sur-gery.16 The authors proposed that the heavier-than-air CO2

remained inside and covered the wound cavity like a protectivebarrier reducing airborne contamination. They showed a 7.1-fold reduction in contamination rate with a CO2 flow of 10 L/min, although they did not confirm the cushion mechanism orshow the spatial distribution of CO2 and whether elevatedlevels of CO2 posed a risk to theatre staff.

This study visualizes the effect of CO2 insufflation on themotion of airborne particles over an open wound and quantifiesany influence on particle deposition rates. Flow visualizationexperiments were performed with and without CO2 insufflationusing aerosolized glycol and a laser sheet in a simulated theatreenvironment and wound model. The effect of the motion ofsurgeons’ hands was also assessed. A quantitative evaluation ofthe degree of protection provided by CO2 insufflation againstthe entry of particles into the wound was performed atdifferent locations in the wound cavity. The presence andconcentration of CO2 was also measured directly to confirm therole of CO2 in the protection mechanism and to elucidate theextent of the CO2 coverage.

Figure 1. Schematic of modelled wound geometry.

Methods

A theatre environment with LTF ventilation was simulatedwith a positive-pressure clean room enclosure (bioBUBBLE,

Please cite this article in press as: Kokhanenko P, et al., Carbon dioxidemodel, Journal of Hospital Infection (2016), http://dx.doi.org/10.1016/j.

Fort Collins, CO, USA). The ventilation flow rate was controlledto achieve a mean vertical velocity of 0.25 m/s at a height of1.2 m above fixed floor level in accordance with operatingtheatre design standard DIN 1946-4:2008-12. Aerosolized glycolfrom a QTFX-1500 fog generator was introduced to the downflow with particle diameters ranging from 0.4 to 15 mm (89%<5 mm, 10% 5e10 mm, 1% >10 mm).

A simplified abdominal surgical incision was modelled as anoptically transparent elliptical cavity (Figure 1) with a length,width, and depth of 180, 100, and 50mm, respectively.17 AVita-diffuser� (Cardia Innovation AB, Stockholm, Sweden) wasused for insufflation of 10 L/min CO2 and positioned at thenarrow end of the wound at a depth of 25mm below the woundedge. For flow patterns over and in the wound, with andwithout CO2, insufflation was visualized by observing aero-solized particles in a 2-mm-thick vertical laser (1 W, 532 nm,LT-301). Video recordings (1280�720 pixels) of 60 s durationwere taken for each experimental set-up. The distribution andmotion of particles over each minute-long recording werevisualized in time-lapse-like images created by selecting themaximum pixel intensity over all video frames.

Time-averaged CO2 concentrations within the CO2-filledwound cavity weremeasured with a gas analyser (CheckMate II,PBI Dansensor, Ringsted, Denmark) and probe traversed verti-cally from the wound floor (�50mm) to a height of 50mmabove the wound edge. A 5mL/min sample flow rate wasconsidered to be small enough (0.05% of the insufflation flowrate) to have negligible effect on the overall concentrations.

The degree of protection offered by CO2 insufflation atdifferent locations in the cavity was quantified as a protectionfactor given by (ISO 14644-1):

PFx ¼ log�Cx

�Cref

�; (Eq. 1)

where PFx is the degree of protection at position x, Cx is theconcentration of particles at position x, and Cref is the refer-ence concentration of particles measured in the background.A 100-fold reduction of particles in the ‘protected’ areacompared with the background therefore corresponded with aprotection factor of two, and 1000 times a factor of three, andso on. Particle concentrations were measured by collectingand counting deposited particles on glass slides held


P. Kokhanenko et al. / Journal of Hospital Infection xxx (2016) 1e6 3

horizontally at the centre of the wound at depths of0, �10, �30, and �50 mm (wound floor), and on the woundedge. Experiments were also performed simulating the motionof a surgeon’s hands within the wound in a worst-case scenarioby intentionally attempting to disrupt CO2. Repeated move-ments with two hands were made in and out and sidewayswithin the wound cavity for the respective tests’ full 60 s du-rations. The background particle concentration was simulta-neously sampled 200 mm above the top of the wound. Repeatswere performed to provide N ¼ 3 for each experimental set-up. A programme to detect and count the particles was writ-ten using Matlab� (MathWorks, Natick, MA, USA) and validatedmanually with <1% variation.

Results

Particle trace images with and without CO2 insufflation areshown in Figure 2. The border of the wound and location of the

Figure 2. Flow and particle distribution in an open surgical wound msimulated theatre environment with 0.2 m/s downward flow ventilatio


diffuser are highlighted with white dotted lines. Time-averaged CO2 concentrations in and over the wound cavityduring insufflation are plotted and overlaid in Figure 3. Errorbars show the range over which the CO2 concentrations fluc-tuated in steady-state conditions. Average protection factorscalculated for each measurement condition and locationinvestigated are given in Table I.

Discussion

Aerosolized glycol was selected for simulating airborneparticles for two reasons. First, the size range roughly matchedthe most frequent particle sizes found in theatres. Two studiesfound that >99% of particles sampled in ventilated operatingtheatre were<5 mm.11,18 Second, the particles were calculatedto be small enough to follow the motion of the gas flowsfaithfully and to give an accurate picture of the flow pattern.The Stokes number (Stk) is a dimensionless number in fluid

odel (a) without (control), and (b) with insufflation of CO2 in an.


Figure 3. Carbon dioxide concentrations measured vertically above the centre of the wound cavity.

Table I

Mean (N¼ 3) protection factors for different wound depths and conditions investigated (standard deviations are shown in parentheses)

Depth No insufflation (control) CO2 insufflation CO2 insufflation with hand

movement

<5 mm >5 mm <5 mm >5 mm <5 mm >5 mm

Wound edge e e >3.4 (0.4) >1.4 (0.1) 3.5 (0.4) 1.3 (0.0)0 mm centre (top) e e 2.9 (1.0) 1.1 (0.6) e e

�10 mm centre e e 3.4 (1.7) >1.2 (0.2) e e

�30 mm centre e e 2.4 (0.3) 1.6 (0.0) e e

�50 mm centre (floor) �0.2 (0.1) �0.1 (0.0) 2.9 (0.3) 1.3 (0.2) 3.9 (1.0) >1.4 (0.2)


mechanics that indicates whether a particle will continue in astraight line under its own inertia as a fluid turns (Stk>>1) or ifviscous forces dominate and the particle will follow the fluidflow faithfully (Stk >>1). The Stokes number for the largestparticles generated (15 mm) was 0.001 in air and 0.002 in CO2.Experiments therefore not only traced the motion of airborneparticles directly but also accurately modelled the gas flows.

The trace images in Figure 2 illustrate a distinct reduction ofairborne particles entering the wound cavity with insufflationof CO2. In the control case without insufflation (Figure 2a)streamlines of particles can be seen entering and in some casesexiting the wound. With insufflation (Figure 2b) the airflow andparticles are deflected from entering the cavity at the top ofthe wound and interface between air and CO2. Figure 3 showsthe cavity is filled with CO2 and for the majority is above aconcentration of 95%. Above the wound a covering of CO2

w5mm thick forms as the CO2 spills over the wound edge eakin to continuing to overfill a cup with water and a protectivebarrier.16 In this region the CO2 concentration varied by up to�10% due to mixing, whereas elsewhere in the domain theconcentration was stable to within�1%. The CO2 concentrationrapidly diminished to 0%, 25mm above the wound, demon-strating negligible risk to theatre staff breathing in elevatedlevels of CO2.

The protection factors (PFs) in Table I confirm the deflectionmechanism shown qualitatively in Figure 2. The case withoutinsufflation served as the experimental control and wasnecessary to assess any protection offered by the cavity


geometry itself. The negative PFs showed that the cavity ge-ometry assisted particle deposition likely due to reduced flowvelocities within the cavity and shorter relaxation times. ThePF for particles <5 mm was on average 3.0 (P < 0.001) acrossthe wound, which corresponds to a 1000-fold reduction inparticle deposition rate. PFs were higher for particles <5 mmthan larger particles (P < 0.001), which was expected due tothe higher Stokes number of the larger particles. The averagePF for particles >5 mm was 1.3 (P < 0.001). We surmised thatthe movement of a surgeon’s hands would disrupt the cushionof CO2. However, there was no significant difference (P > 0.2)in PFs with exaggerated hand movements.

Persson and van der Linden in an operating room with ul-traclean airflow showed that CO2 insufflation reduced thenumber of particles <5 mm and >5 mm diameter entering anexperimental open surgical wound by 18- and 15.5-fold,respectively, when the surgeons leaned forward over thewound during sham surgery. No airborne particles weredetected in the cavity when the surgeons were inactive,standing upright next to the open wound cavity.19 Debate overthe benefit or perhaps harm of different ventilation systemswith different performance characteristics is ongoing.20,21 Asearly as 1976, Clark et al. found a reduction in both the con-centration of airborne particles and SSI rates after installing ahigh-flow HEPA-filtered vertical ventilation system.22 Later,however, Taylor and Bannister found a 27-fold increase inwound contamination when the surgeon leaned over the woundand into vertically directed ‘laminar’ downflow.21


P. Kokhanenko et al. / Journal of Hospital Infection xxx (2016) 1e6 5

It is not clear in the literature whether airborne particleconcentrations correlate with surgical site infection rates.23

Attempts to show correlations appear to be confounded by anumber of other factors that influence SSI rates such as thetype of operation, surgical technique, antibiotic prophylaxis,patient risk factors, and the degree of wound contamination.Whyte et al. demonstrated an effect of airborne contaminationand estimated that a 13-fold reduction in airborne bacteriawould reduce wound contamination by w50% during biliarysurgery where there was little contamination from non-airborne routes.5 Reduction of infection rates with reducedairborne contamination have also been shown for orthopaedic,spinal, and sham operations.24e28 If the regression model ofLidwell et al. of orthopaedic joint sepsis rates (Eq. 2) is appliedto the lowest CO2 insufflation PF factor on a wound surface of1.3 from the present study, with an air contamination of 177bacteria/m3 averaged across 15 hospitals, then a reduction ofjoint sepsis rate from 3.2% to 1.4% could be expected.25

Joint sepsis rate ð%Þ ¼ 0:84þ0:18ðAir contaminationÞ0:5 (Eq. 2)

The ability to visualize the motion of small or individualparticles was limited both by a magnification of 0.2 mm/pixeland the intensity of light scattered. It was therefore notpossible to detect isolated particles and it was necessary tocollect and count deposited particles on a microscope toquantify the qualitatively observed deflection mechanism. Infour experiments the maximum number of particles wascollected on the background slide before any particles wereobserved at the location of interest. For these cases the PFs arereported in Table I as ‘greater than’ based on an assumedparticle count of unity since an infinite PF is not verifiable.

Other studies have counted the number of bacterial coloniesin agar plates to assess airborne contamination; however, theyhave included the added variable of the viability of infectingmaterial.16,28,29 In the present study we counted all particlesdeposited, whether they contained viable bacteria or not. Thismay account for the larger PFs measured in the present studycompared with those of Persson and van der Linden.16,19

Whyte et al. noted that when bacteria from bile were pre-sent, they were so numerous in the wound that they maskedthe presence of airborne bacteria.5 For contaminated surgerieswhere endogenous infection sources may dominate airbornesources and neglect any benefits of reduced airborne deposi-tion, CO2 insufflation may still have bacteriostatic benefitsagainst aerobic bacteria, and improve oxygenation.30 Also, aCO2-covered wound surface is insulated from sub-body-temperature theatre ventilation, so if CO2 insufflation iswarmed and humidified, it may assist with wound temperaturemaintenance and reduce the incidence of hypothermia.31

We have demonstrated the essential physics that insuffla-tion of an open wound cavity with CO2 affords protection fromairborne contamination even when disturbed. CO2 insufflationmay be an effective supplementary means of reducing intra-operative infection rates in open surgeries.

Conflict of interest statementC.J.T Spence, K. Kokhanenko and G. Papotti are employeesof Fisher & Paykel Healthcare Ltd (Auckland, New Zealand)which markets and sells the Vita-diffuser. J.A. van derLinden is a shareholder of Cardia Innovation AB (Stockholm,Sweden) which manufactures the Vita-diffuser.


Funding sourceThis project was funded by Fisher & Paykel Healthcare.

References

1. National Collaborating Centre for Women’s Children’s Health (UK).Surgical site infection: prevention and treatment of surgical siteinfection. 2008. Contract No.: 74.

2. Chauveaux D. Preventing surgical-site infections: measures otherthan antibiotics. Orthop Traumatol Surg Res 2015;101(1,Supplement):S77eS83.

3. Bagnall NM, Vig S, Trivedi P. Surgical-site infection. Surgery (Oxf)2009;27:426e430.

4. Dharan S, Pittet D. Environmental controls in operating theatres.J Hosp Infect 2002;51:79e84.

5. Whyte W, Hambraeus A, Laurell G, Hoborn J. The relative impor-tance of the routes and sources of wound contamination duringgeneral surgery. II. Airborne. J Hosp Infect 1992;22:41e54.

6. Whyte W, Hodgson R, Tinkler J. The importance of airborne bac-terial contamination of wounds. J Hosp Infect 1982;3:123e135.

7. Cole EC, Cook CE. Characterization of infectious aerosols in healthcare facilities: an aid to effective engineering controls and pre-ventive strategies. Am J Infect Control 1998;26:453e464.

8. Mackintosh CA, Lidwell OM, Towers AG, Marples RR. The di-mensions of skin fragments dispersed into the air during activity.J Hyg (Lond) 1978;81:471e479.

9. Lidwell OM, Mackintosh CA, Towers AG. The evaluation of fabricsin relation to their use as protective garments in nursing andsurgery. II. Dispersal of skin organisms in a test chamber. J Hyg(Lond) 1978;81:453e469.

10. Tham KW, Zuraimi MS. Size relationship between airborne viablebacteria and particles in a controlled indoor environment study.Indoor Air 2005;15:48e57.

11. Seal DV, Clark RP. Electronic particle counting for evaluating thequality of air in operating theatres: a potential basis for standards?J Appl Bacteriol 1990;68:225e230.

12. Gosden PE, MacGowan AP, Bannister GC. Importance of air qualityand related factors in the prevention of infection in orthopaedicimplant surgery. J Hosp Infect 1998;39:173e180.

13. Clark RP, Cox RN, Lewis HE. Deposition and dispersion of particlesfrom the human micro-environment. J Physiol 1971;216:19pe20p.

14. Anderson DJ. Surgical site infections. Infect Dis Clin North Am2011;25:135e153.

15. Bitkover CY, Marcusson E, Ransjo U. Spread of coagulase-negativestaphylococci during cardiac operations in a modern operatingroom. Ann Thorac Surg 2000;69:1110e1115.

16. Persson M, van der Linden J. Wound ventilation with carbon di-oxide: a simple method to prevent direct airborne contaminationduring cardiac surgery? J Hosp Infect 2004;56:131e136.

17. Cater JE, van der Linden J. Simulation of carbon dioxide insuf-flation via a diffuser in an open surgical wound model. Med EngPhys 2015;37:121e125.

18. Scaltriti S, Cencetti S, Rovesti S, Marchesi I, Bargellini A, Borella P.Risk factors for particulate and microbial contamination of air inoperating theatres. J Hosp Infect 2007;66:320e326.

19. Persson M, van der Linden J. Wound ventilation with ultraclean airfor prevention of direct airborne contamination during surgery.Infect Control Hosp Epidemiol 2004;25:297e301.

20. McHugh SM, Hill ADK, Humphreys H. Laminar airflow and theprevention of surgical site infection. More harm than good? Sur-geon 2015;13:52e58.

21. Taylor G, Bannister G. Infection and interposition between ultra-clean air source and wound. Bone Joint J 1993;75:503e504.

22. Clark RE, Amos WC, Higgins V, Bemberg KF, Weldon CS. Infectioncontrol in cardiac surgery. Surgery 1976;79:89e96.


http://refhub.elsevier.com/S0195-6701(16)30532-1/sref1


















































































23. Hansen D, Krabs C, Benner D, Brauksiepe A, Popp W. Laminar airflow provides high air quality in the operating field even duringreal operating conditions, but personal protection seems to benecessary in operations with tissue combustion. Int J Hyg EnvHealth 2005;208:455e460.

24. Lidwell O, Lowbury E, Whyte W, Blowers R, Stanley S, Lowe D.Effect of ultraclean air in operating rooms on deep sepsis in thejoint after total hip or knee replacement: a randomised study. BrMed J (Clin Res Ed) 1982;285:10e14.

25. Lidwell O, Lowbury E, Whyte W, Blowers R, Stanley S, Lowe D.Airborne contamination of wounds in joint replacement opera-tions: the relationship to sepsis rates. J Hosp Infect1983;4:111e131.

26. Charnley J. Postoperative infection after total hip replacementwith special reference to air contamination in the operating room.Clin Orthop Relat Res 1972;87:167e187.

27. Gruenberg MF, Campaner GL, Sola CA, Ortolan EG. Ultraclean airfor prevention of postoperative infection after posterior spinalfusion with instrumentation: a comparison between surgeries


performed with and without a vertical exponential filtered air-flow system. Spine 2004;29:2330e2334.

28. Friberg B, Friberg S, Burman LG. Correlation between surface andair counts of particles carrying aerobic bacteria in operating roomswith turbulent ventilation: an experimental study. J Hosp Infect1999;42:61e68.

29. Pasquarella C, Pitzurra O, Herren T, Poletti L, Savino A. Lack ofinfluence of body exhaust gowns on aerobic bacterial surfacecounts in a mixed-ventilation operating theatre. A study of 62 hiparthroplasties. J Hosp Infect 2003;54:2e9.

30. Marshall JK, Lindner P, Tait N, Maddocks T, Riepsamen A, van derLinden J. Intra-operative tissue oxygen tension is increased bylocal insufflation of humidified-warm CO2 during open abdominalsurgery in a rat model. PLoS One 2015;10:e0122838.

31. Frey JM, Janson M, Svanfeldt M, Svenarud PK, van der Linden J.Local insufflation of warm humidified CO2 increases open woundand core temperature during open colon surgery: a randomizedclinical trial. Anesth Analg 2012;115:1204e1211.

















































carbon dioxide insufflation deflects airborne particles ... insufflation deflects airborn...

Documents