Prospective Comparison of tand CT Scan in the EvaluatiAppendicitis: A Proposed Al

Winson Jianhong Tan, MBBS (Hons), MRCS, Sanchalika AWeng Hoong Chan, MBBS, FRCS, Wai Keong Wong,Hock Soo Ong, MBBS, FRCS, FAMS, FACS

BACKGROUND: Although computed tomography (CT)tion risk remains a concern. We compa

aluabenctioscay a

mpli-y are

estinalakages

as thepected

anagementhe Alvarado

218 2015 by the American College of Surgeons. Published by Elsevier Inc.This is an open access article under the CC BY-NC-ND license

http://dx.doi.org/10.1016/j.jamcollsurg.2014.10.010

ISSN 1072-7515/14ing CT use are therefore essential to minimsary CT scans.9 We previously proposed a malgorithm for suspected appendicitis with t

icine, DUKE NUS Graduate Medical School (Acharyya), Singapore.Correspondence address: Dr Winson Jianhong Tan, MBBS (Hons),MRCS, Department of General Surgery, Singapore General Hospital, Out-ram Road, Singapore 169608. email: tanwinson156@hotmail.comappendicitis in adults. It has decreased negative appen-dectomy rates to less than 10%.4-6 However, the radiationexposure with CT poses a concern, particularly in appen-dicitis, which occurs predominantly in young patientsmost susceptible to the adverse effects of radiation.7,8

Available literature has estimated that at least 25% ofCT scans are not clinically warranted and may posemore harm than benefit.9 Rules for clinical decisions guid-

ize unneces-

CME questions for this article available athttp://jacscme.facs.org

Disclosure Information: Authors have nothing to disclose. Timothy J Eberlein,Editor-in-Chief, has nothing to disclose.

Received September 11, 2014; Revised October 17, 2014; AcceptedOctober 21, 2014.From the Department of General Surgery, Singapore General Hospital(Tan, Goh, Chan, Wong, Ooi, Ong) and the Centre for Qualitative Med-ratios of CT scan with each of the AS cutoff values.RESULTS: The study included 134 males (38.3%) and 216 females (61.7%). The overall prevalence of

appendicitis was 44.3% in the total study population; 37.5% in females and 55.2% in males.There were 168 patients (48%) who underwent surgery, with a negative appendectomy rateof 7.7%. Positive likelihood ratio of disease was significantly greater than 1 only in patientswith an AS of 4 and above. An AS of 7 and above in males and 9 and above in females has apositive likelihood ratio comparable to that of CT scan.

CONCLUSIONS: Evaluation by CT is beneficial mainly in patients with AS of 6 and below in males and 8 andbelow in females. We propose an objective management algorithm with the AS guiding subse-quent evaluation. (J Am Coll Surg 2015;220:218e224. 2015 by the American College ofSurgeons. Published by Elsevier Inc. This is an open access article under the CC BY-NC-NDlicense [http://creativecommons.org/licenses/by-nc-nd/3.0/].)

Acute appendicitis is one of the most common causes ofacute abdominal pain requiring surgical intervention,with a lifetime risk of 8.6% for males and 6.7% for fe-males.1,2 Historically, negative appendectomy rates ofmore than 20% were considered the norm. However,

this is no longer acceptable because even though cocation rates in the setting of negative appendectomlow, conditions such as incisional hernias, intobstruction secondary to adhesions, and stump lecan result in significant morbidity.Computed tomography (CT) scan has emerged

dominant imaging modality for evaluation of sus3itive likelihood ratios were determined for various AS and for CT scan. The AS ranges thatbenefitted most from CT evaluation were determined by comparing the positive likelihood(AS) with those of CT scan in the evtifying a subset of patients who will

STUDY DESIGN: We performed prospective data colledicitis who were evaluated with CTand correlated with eventual histolog(http://creativecommons.org/licenses/by-nc-nd/3.0/).he Alvarado Scoreon of Suspectedgorithm to Guide CT Use

charyya, PhD, MPH, Yaw Chong Goh, MBBS, FRCS,MBBS, FRCS, London Lucien Ooi, MBBS, FRCS, MD,

has reduced negative appendectomy rates, its radia-red the performance statistics of the Alvarado Scoretion of suspected appendicitis, with the aim of iden-efit from CT evaluation.n on 350 consecutive patients with suspected appen-ns. The AS for each patient was scored at admissionnd CT findings. The sensitivity, specificity, and pos-

Vol. 220, No. 2, February 2015 Tan et al Using Alvarado Score to Guide CT Use 219score (AS) (Table 1) guiding CT use.10 This algorithmwas, however, developed based on retrospective datawith its antecedent limitations.This study aimed to compare the performance statistics

of the AS with CT scan in the evaluation of suspectedappendicitis. Thereafter, we attempt to use the AS tostratify patients with suspected appendicitis into sub-groups that might benefit from CT evaluation. An objec-tive algorithm for the management of suspectedappendicitis guided by the AS is then proposed.

METHODSWe performed an analysis of prospectively collected datafrom 450 consecutive patients with suspected appendi-citis, admitted to the General Surgery Department atSingapore General Hospital. The study ran from August2013 to March 2014, and only patients who underwentCT evaluation were included in the final analysis. Deci-sion for CT evaluation was left to the discretion of theattending surgeon during the initial assessment. Patientdemographics, presenting signs and symptoms, and rele-

Table 1. The Alvarado Scoring System

Mnemonic (MANTRELS) Value

Symptom

Migration 1

Anorexia-acetone 1

Nausea-vomiting 1

Signs

Tenderness in right lower quadrant 2

Rebound pain 1

Elevation of temperature >37.3C 1Laboratory

Leukocytosis 2

Shift to the left 1

Total score 10vant laboratory values were prospectively collected andrecorded in a standardized data collection sheet. The ASof each patient was scored by the attending surgeon atthe point of admission, before the decision was madefor CT evaluation, and it was recorded in the clinicalchart. Computed tomography findings, surgical findings,and histologic results were recorded for each patient whenapplicable. Study data were collected and managed usingthe REDCap electronic data capture tools hosted atSingapore General Hospital. REDCap (Research Elec-tronic Data Capture) is a secure, web-based applicationdesigned to support data capture for research studies.11

In order to ensure short-term follow-up, all patientswere reviewed in person by a clinician outpatient at leastonce within 2 weeks from discharge. Subsequent follow-up visits were determined based on clinical indication.Patients discharged without surgery were treated with an-tibiotics only if they were diagnosed with conditions thatwarranted therapy. Empirical treatment with antibioticswas not practiced. Repeat admissions for patients dis-charged without surgery were identified by a search ofthe National Electronic Health Record database inSingapore, a database that captures the admission infor-mation of every person in Singapore who has visited thepublic health care system. A case of missed diagnosiswas defined as readmission within 2 weeks from initialdischarge, with eventual surgery showing acute appendi-citis on histology.Appendicitis was considered present when patients who

had undergone surgery had a final histology showingacute appendicitis. A case was considered to be a negativeappendectomy when a patient had undergone surgerywith the clinical impression of acute appendicitis buthad no features of appendicitis in histology. Patientswho did not undergo surgery were considered not tohave appendicitis if they did not re-present within 2 weeksfrom initial discharge with acute appendicitis. Computedtomography scans were read by the radiologist on dutywhen the scans were ordered, and findings were catego-rized into 3 groups: positive for acute appendicitis, nega-tive for acute appendicitis, and equivocal findings.Sensitivity, specificity, positive and negative predictive

values, and likelihood ratios were estimated for each ofthe cut off AS scores ranging from 2 to 10, using histol-ogy results as the gold standard. Scores of zero and 1 wereomitted because there were no patients with such scores.The same diagnostic performance measures were calcu-lated for CT scan using the same gold standard. Equiv-ocal CT scans were considered positive for acuteappendicitis in the calculations above. This method ofclassifying equivocal scans was chosen because in ourinstitution, most surgeons would offer a diagnostic lapa-roscopy for patients who present with suspected appendi-citis and an equivocal CT scan. Because we wereconcerned that classification in this manner may influ-ence the eventual findings, we repeated the above statisti-cal analysis first by excluding and thereafter by classifyingthese equivocal cases as negative for acute appendicitisseparately. It is our institutions practice to remove all ap-pendixes even if there were no macroscopic features ofacute appendicitis intraoperatively. This is guided byexisting data, which revealed that up to 33% of macro-scopically normal appendixes have features of inflamma-tion on histology.12

The range of AS for which patients were least likely tobenefit from CT evaluation was determined by identi-fying AS ranges that had positive likelihood ratios not

e and outcomes of study cohort.

220 Tan et al Using Alvarado Score to Guide CT Use J Am Coll Surgsignificantly different from those of CT scans. Likelihoodratios were selected as the parameter for comparisonbecause they were independent of disease prevalence anddepended only on the intrinsic ability of the diagnostictest to distinguish between diseased and nondiseased indi-viduals. The pairwise comparisons of predictive valuesand likelihood ratios are based on the methods describedby Moskowitz and Pepe (2006)13 and Nofuentes and Cas-tillon (2007),14 respectively. The above statistics were sub-analyzed by sex because the performance of the AS hasbeen shown to vary according to sex.15 Statistical analyseswere performed using Statistical Package for the SocialSciences (SPSS) Version 17. Performance measures,including sensitivity, specificity, positive and negative pre-dictive values, and diagnostic likelihood ratios were calcu-lated and compared using the BDT comparatorprogram.16 A p value of less than 0.05 was consideredto indicate statistical significance.The study was carried out under the approval of the

Centralized Institutional Review Board of the SingaporeHealth Services.Figure 1. Management coursRESULTSThere were 450 patients admitted for suspected appendi-citis from August 2013 to March 2014. One hundred pa-tients were not evaluated with CT scans and were excludedfrom the study. Altogether, 350 patients underwent CTevaluation. There were no cases of missed diagnosis inthese patients, who were all evaluated with CT scans.There were 134 males (38.3%) and 216 females

(61.7%). The overall median age of the patients was 33years (range 15 to 82 years): 32 years for males and 33years for females. Among the 350 patients who presentedwith suspected appendicitis and were evaluated with CTscans, the overall prevalence of appendicitis was 44.3%in the total study population; 37.5% in females and55.2% in males (Fig. 1).Nineteen (5.4%) of the CT scans were deemed equiv-

ocal, 11 in females and 8 in males. Surgery was performedfor 168 patients (48%), of whom 40, 126, and 2 under-went open appendectomy, laparoscopic appendectomy,and laparotomy, respectively. The overall negative appen-dectomy rate was 7.7%.The number of patients within each AS cut off category

is illustrated in Table 2. The sensitivity, specificity, posi-tive and negative predictive values, and positive likelihoodratio of the various AS cut-off values compared with CTscan are illustrated in Table 3. Sub-analysis of the positivelikelihood ratios of the various AS values stratified by sexand compared with CT scan are illustrated in Table 4.Alvarado Scores of 7 and above in males (AS 7, p

0.513; AS 8, p 0.442; AS 9, p 0.398; AS 10,p 0.896) and 9 and above in females (AS 9, p 0.513; AS 10, p 0.638) have positive likelihood ra-tios comparable to those of CT scan. Analysis afterexcluding equivocal scans or after classifying equivocalTable 2. Number of Patients within Each Alvarado ScoreCategory Cut Off with Sex Stratification

Alvarado Score

Sex

Male, n Female, n

2 134 2163 127 2034 120 1875 101 1566 72 1237 37 738 17 409 5 1110 1 3

are subjected to CT evaluation. Within these score ranges,

Table

3.

Perform

ance

Statistics

ofAlvaradoScore

CutOffValuesandResultsfrom

PairwiseComparisonoftheseStatistics

withthatofCTScan

Sensitivity,

%pValue*

Specificity,

%pValue*

Positivepredictive

value,%

pValue*

Negative

predictive

value,%

pValue*

Positivelikelihoodratio

pValue*

CTscan

98.7

(95.4e99.8)

88.5

(83.2e92.7)

87.4

(81.6e92.0)

98.8

(95.8e99.9)

8.570(5.782e12.701)

Alvarado

Score

2100.0

(97.6e100.0)

0.479

0(0.0e1.9)