at SciVerse ScienceDirect

International Journal of Surgery 11 (2013) 795e800

REVIEW

Contents lists available

International Journal of Surgery

journal homepage: www.journal-surgery.net

Review

Elevated serum bilirubin in assessing the likelihood of perforationin acute appendicitis: A diagnostic meta-analysis

Salvatore Giordano*, Markus Pääkkönen, Paulina Salminen, Juha M. GrönroosDepartment of Surgery, Turku University Hospital and the University of Turku, Turku, Finland

a r t i c l e i n f o

Article history:Received 11 October 2012Received in revised form17 April 2013Accepted 17 May 2013Available online 31 May 2013

Keywords:AppendicitisBilirubinHyperbilirubinemiaDiagnostic meta-analysisJaundice

* Corresponding author. Department of Surgery, Tthe University of Turku, OS 299, PL 52, 20521443346119, þ358 02 3130521; fax: þ358 02 3132284

E-mail addresses: salvatoregiordano@yahoo.it, s(S. Giordano).

1743-9191/$ e see front matter � 2013 Surgical Assohttp://dx.doi.org/10.1016/j.ijsu.2013.05.029

a b s t r a c t

Hyperbilirubinemia associates with perforated acute appendicitis, but the precise test characteristicshave not been determined by sufficiently powered studies.

A systematic literature search of reports on hyperbilirubinemia in acute appendicitis was performed.After review and quality assessment of eight studies encompassing a total of 4974 patients the sensitivity,specificity and other measures of accuracy of hyperbilirubinemia as a predictor of perforation in acuteappendicitis were pooled using a random-effects model.

Summary estimates for hyperbilirubinemia (cutoff 1 mg/dl) as a predictor of perforation in acuteappendicitis were as follows: sensitivity, 0.49 (95% confidence interval [CI], 0.45e0.52); specificity, 0.82(95% CI, 0.80e0.83); positive likelihood ratio, 2.51 (95% CI, 1.58e4.00); negative likelihood ratio, 0.58(95% CI, 0.44e0.76); diagnostic odds ratio was 4.42 (95% CI, 2.21e8.83). In summary receiving operatingcharacteristic (SROC) analysis, the area under curve was 0.73.

The presence of hyperbilirubinemia does not alone distinguish a perforation in acute appendicitis.Serum bilirubin level is of some value as a predictor of appendiceal perforation. Patients with hyper-bilirubinemia combined with symptoms and signs consistent with severe acute appendicitis should beconsidered for early appendectomy.

� 2013 Surgical Associates Ltd. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Acute appendicitis remains a diagnostic challenge.1 Aside fromsymptoms and signs, ultrasonography andmultidetector computedtomography (MDCT) have become the most useful tools with arespective sensitivity and specificity of 98.5% and 98.0% in expertcentres.2 Routine use of MDCT increases radiation exposure and canbe very expensive and time-consuming. Thus methods to diagnoseand estimate the extent of acute appendicitis without radiologicalexaminations are of interest. Observing serum bilirubin level meritsconsideration as it seems to elevate in perforated appendicitis, itsdetermination is fast and can be performed in panel with othercommonly performed blood tests. Blood white blood cell count andespecially C-reactive protein have already been found somewhat topredict the severity of acute appendicis, but without definitive andconvincing results.3e5

urku University Hospital andTurku, Finland. Tel.: þ358.alvatore.giordano@gmail.com

ciates Ltd. Published by Elsevier Lt

There are several diagnostic scoring systems such as the Alvar-ado score,6 the modified Alvarado score for use in paediatric pa-tients,7 PAS (Paediatric Appendicitis Score),8 RIPASA (RajaIsteriPengiran Anak Saleha Appendicitis) score for use in Asian patients,9

and a newer Appendicitis Inflammatory Response (AIR) score.10

Most often these scoring systems combine symptoms (duration ofpain, migration of pain, nausea, vomiting), signs (tenderness, fever)and/or laboratory measurements (leucocytosis, CRP). All of thesescores have been proven useful in predicting acute appendicitis inpatients presenting with pain in the lower right fossa, but none ofthem evaluates the risk of appendiceal perforation nor useshyperbilirubinemia as a predictor.

Many surgical centres attempt to reduce surgical proceduresperformed during the night, and delaying appendectomy for 12e24 h does not seem to increase complications.11 Conservativetreatment with antibiotics has also been suggested for uncompli-cated acute appendicitis in selected patients.12e14 Appendicealperforation combined with peritonitis always mandates an imme-diate emergency operation or percutaneous drainage.12,13 Thus,there is an increasing demand for fast and safe markers to specif-ically separate patients with a complicated perforated appendicitisfrom those with uncomplicated acute appendicitis.

d. All rights reserved.

Table 1Summarized quality assessment using the modified standards for reporting ofdiagnostic Accuracy checklist.

Quality item Number of studies complyingwith item (of 8 included)

Representative spectrum 8Selection biasa avoided 6Acceptable reference standard 8Partial verificationb avoided 8Differential verificationc avoided 3Incorporationd avoided 3Reference standard results blinded 0Information biase avoided 0Index test results blinded 0Relevant clinical information provided 8Uninterpretable/intermediate test

results reported8

Withdrawals explained 6

Reference standard e occurrence of perforation in acute appendicitis, index test e

S. Giordano et al. / International Journal of Surgery 11 (2013) 795e800796

REVIEW

Elevated serum bilirubin level and jaundice are commonlyobserved in patients suffering from a septic condition.15e17 Acuteappendicitis is the most common intra-abdominal infectious focusin a surgical patient and Escherichia coli and Bacteroides fragilis arethemost frequent bacterial isolates in this condition. Bacteraemia isknown to cause endotoxemia leading to impaired excretion ofbilirubin from the bile canaliculi.18,19 A few case reports havedescribed hyperbilirubinemia in patients with a severe post-operative infection after appendectomy or with a complicatedacute appendicitis.17,20e26

More recently, several retrospective and some prospectivestudies have analysed hyperbilirubinemia as a marker of acuteappendicitis and perforation, but none of these studies has beensufficiently powered to analyse the precise sensitivity and speci-ficity in diagnostics.27e35 We performed a meta-analysis to deter-mine the value of serum bilirubin measurement in diagnosingacute appendiceal perforation.

Hyperbilirubin.a When eligible patients are not enrolled consecutively/randomly.b When a nonrandom set of patients does not undergo the reference standard.c When a set of patients is verified with a second or third reference standard,

especially when this selection depends on the index test result.d When the index test is incorporated in the reference standard.e When the index test results are interpreted with knowledge of the results of the

reference standard.

2. Materials and Methods

An English-language literature search was performed through Medline,Cochrane Library, Embase, Scopus and Google Scholar for any study on appendicitisand hyperbilirubinemia from January 1950 to March 2013 aiming to identify studiesevaluating an association between appendiceal perforation and hyperbilirubinemia.

The search was run independently by two authors (S.G. and M.P.) according tothe validated methods of the Preferred Reporting Items for Systematic Reviews andMeta-Analyses (PRISMA) statement.36 The terms employed in the search were:‘appendix’, ‘appendicitis’, ‘bilirubin’, ‘hyperbilirubin’, ‘hyperbilirubinemia’, ‘hyper-bilirubinaemia’, ‘jaundice’. In addition, the reference lists of all relevant articles werescrutinised. In order to form a highly sensitive search strategy, there were no re-strictions on publication status or study design. However, case reports wereexcluded.

Although no language restrictionswere imposed initially, for the full-text reviewand final analysis, our resources permitted the review of articles published in En-glish, Italian, Spanish, German, Swedish or Finnish. All studies designed with at least100 patients, reporting all-comers populations with acute appendicitis and bilirubinmeasurements, were qualified for subsequent analyses. Letters, conference abstractsand grey literature to the journal editors were excluded because of the limited datapresented.

Two reviewers (S.G. and M.P.) read the titles and/or abstracts of the identifiedreferences and eliminated irrelevant studies. Studies that were considered eligiblefor inclusion were read fully in duplicate, and both reviewers independentlydetermined their suitability for inclusion in the study. Discrepancies were managedby consensus.

Data extraction was performed independently by two reviewers (S.G. and M.P.)from all eligible studies using a standardized Excel file. The two reviewers wereblinded to publication details, and all extracted data had to be agreed upon. Dataretrieved from the reports included study design, participant characteristics,outcome measurements, publication year, and methodological quality.

We assessed the methodological quality of the studies by using the qualityassessment of diagnostic accuracy studies (QUADAS) tool.37,38 Two independentreviewers (S.G. and M.P.) applied the QUADAS tool to the articles included in thestudy, and differences in assessments were resolved by consensus in the presence ofa third investigator (J.G.). The details of the methodological assessment are shown inTable 1. The score was not used as an inclusion criterion or a weighting factor.

We re-calculated the numbers of true-positive, true-negative, false-positive andfalse-negative, and results. Only patients with histological findings of acute orperforated appendicitis were included in order to standardize the data between thestudies. Therefore, we excluded patients with no histological signs of acuteinflammation. Retrieved information is displayed for each study in Table 2.

We followed standard methods recommended for meta-analyses of diagnostictest evaluations.39 The retrieved studies were included if they reported outcomesabout acute appendicitis, appendiceal perforation and serum bilirubin value.Hyperbilirubinemia was defined as reported from the considered studies, a bilirubinlevel greater than 1mg/dl or>20.5 mmol/l.27,29e35 Datawere retrieved only from thearticles and no attempt to get missing data from the authors was made.

2.1. Statistical analysis

Statistical analyses were performed using Meta-DiSc software (for Windows; XICochrane Colloquium; Barcelona, Spain).40

The following measures of test accuracy were computed for each study: sensi-tivity (Se), specificity (Sp), positive likelihood ratio (PLR), negative likelihood ratio(NLR) and diagnostic odds ratio (DOR).

Se of a test is the proportion of participants with disease and a positive finding ofperforated appendicitis and hyperbilirubinemia. Sp measures the proportion ofnegatives subjects correctly identified (patients with normal bilirubin and withoutappendiceal perforation). PLR is the probability of a positive finding in participantswith disease divided by the probability of the same finding in participants withoutdisease. Similarly, NLR is the probability of a negative finding in a participant withdisease divided by the probability of the same finding in participants without dis-ease. DOR is the ratio of the odds of a positive finding in a participant with diseaserelative to a positive finding in a participant without disease.39

The analysis was based on a summary receiver operating characteristic (SROC)curve,41,42 with a Q-value defined to be the false discovery rate analogue of the p-value. Sensitivity and specificity for the single test threshold identified for eachstudywere used to plot the SROC curve. Random-effects model was used to calculatethe average sensitivity, specificity, and the other measures across the studies.43

Heterogeneity has been assessed by using I2 statistic, which describes the per-centage of total variation across studies that is due to heterogeneity rather thanchance.44 Usually, values of the I2 statistic <25% are indicative of low heterogeneity,those ranging between 25% and 75% of moderate heterogeneity, and those >75% ofhigh heterogeneity. I2 <75% was considered as non-important heterogeneity. Weperformed a random-effect analysis in all cases, which considers both within- andbetween-study variation,45 because of the observational nature of most studiesincluded in this analysis. A p < 0.05 was considered statistically significant.

Finally, we conducted sensitivity analyses omitting each study in turn todetermine whether the results were influenced excessively by a single study. Pub-lication bias was assessed using the funnel plot.

3. Results

Literature search yielded 29 articles, eight27,29e35 of which werepertinent to this issue and sources of information on appendixperforation and hyperbilirubinemia. The literature search flowchartis shown in Fig. 1.

The eight studies included27,29e35 involved a total of 4974 pa-tients, and two of them were carried out in the United States,27,30

four in Europe31e33,35 and two in Asia.29,34 All publications are inEnglish. Surgical and histological diagnosis is the reference stan-dard in all studies included. The summary of the quality assessmentis illustrated in Table 1 and the methodological quality of thestudies included is very similar as none of the studies fulfilled allitems of the QUADAS checklist.37 The primary concern withmethodological quality relates to the time delay between the bili-rubin measurement and clinical tests. Furthermore, it is unclearwhether the assessors were blinded to the outcome of index andreference tests. Table 2 shows data extracted from the studies,seven of them are retrospective cohorts.

Table 2Summary of data retrieved from studies, sensitivity and specificity of hyperbilirubinemia as marker of appendix perforation.

Authors Study type Number ofpatients

Age (years,(mean, range)

Hystologicallyconfirmedappendicitis (n)

Perforatedappendix (n)

Sensitivity Specificity Positivelikelihood ratio

Negativelikelihood ratio

Estrada 2007 Retrospective non-randomized 170 33 y (5e66) 157 41 0.56 0.69 1.81 0.64Khan 2008 Prospective non-randomized 122 29 y (8e73) 118 18 0.72 0.18 0.88 1.54Sand 2009 Retrospective non-randomized 538 36 y (6e91) 376 97 0.70 0.86 5.06 0.35Käser 2010 Retrospective non-randomized 1073 22 y (5e92) 725 155 0.38 0.78 1.71 0.8Atahan 2011 Retrospective non-randomized 351 31 y (18e83) 302 45 0.80 0.84 4.9 0.24Emmanuel 2011 Retrospective non-randomized 472 27 y (5e82) 386 45 0.60 0.70 1.99 0.57Hong 2012 Retrospective non-randomized 977 31 y 732 245 0.32 0.84 2.01 0.81McGowan 2013 Retrospective non-randomized 1271 e 1053 154 0.55 0.90 5.76 0.50

S. Giordano et al. / International Journal of Surgery 11 (2013) 795e800 797

REVIEW

Fig. 2(A) and (B) shows the forest plot of sensitivity and speci-ficity for hyperbilirubinemia in predicting appendix perforation.The sensitivity ranged from 0.32 to 0.80 (mean, 0.49; 95% confi-dence interval [CI], 0.45e0.52, Fig. 2(A)), and specificity from 0.18 to0.90 (mean, 0.82; 95% CI, 0.80e0.83, Fig. 2(B)). PLR was 2.51 (95% CI,1.58e4.00, Fig. 2(C)), NLR was 0.58 (95% CI, 0.44e0.76, Fig. 2(D)),and DOR was 4.42 (95% CI, 2.21e8.83, Fig. 2(E)).

X2 values of sensitivity, specificity, PLR, NLR and DORwere 82.65(p< 0.001), 313.14 (p< 0.001), 158.02 (p< 0.001), 80.78 (p< 0.001)and 100.59 (p< 0.001), respectively, with all indicating a significantheterogeneity between the studies (91.3e97.8%).

The SROC curve summarizes the overall diagnostic accuracy,showing the tradeoff between sensitivity and specificity. The SROC

Fig. 1. Flow-chart summarizing

curve for hyperbilirubinemia in appendiceal perforation demon-strating true-positive rates versus false-positive rates from indi-vidual studies is shown in Fig. 3. The SROC curve is positionedtowards the desirable upper left corner of the SROC curve, and themaximum joint sensitivity and specificity, i.e. the Q-value, corre-sponding to the point closest to the ideal top-left corner of the SROCspace in symmetric curves, was 0.68. The area under the curve(AUC) was 0.73 indicating an acceptable level of overall accuracyand discrimination (Fig. 3).

Finally, the exclusion of any study from the analysis did notmaterially change the summary estimates. The funnel plot of all theincluded studies showed an absence of significant asymmetry(Fig. 2(F)).

literature search results.

Fig. 2. Forest plots of estimates of sensitivity (A), specificity (B), positive likelihood ratio (C), negative likelihood ratio (D) and diagnostic odds ratio (E) for hyperbilirubinemia andappendiceal perforation and funnel plot of all included studies (F). The size of the solid circle indicates the effect size of each study.

S. Giordano et al. / International Journal of Surgery 11 (2013) 795e800798

REVIEW

4. Discussion

This diagnostic meta-analysis of 4974 patients provides acompelling insight into the diagnostic value of serum bilirubinmeasurements for appendiceal perforation in an emergency roomsetting Appendicitis is most often uncomplicated. Whereas non-perforated acute appendicitis can be treated by appendectomyfollowed by a short recovery period, a perforated appendicitis maybe a life-threatening condition.46 An early diagnosis of perforationimproves outcome, and allows the surgeon to operate early andprepare for a potentially difficult procedure.

Although meta-analysis is not yet a widely approved method tosummarize evidence from studies on diagnostic tests, we believethat pooling the diagnostic accuracy from eligible studies providesvaluable information for surgeons and researchers until largerstudies are available.

Our analysis has some strengths and limitations. The pooledspecificity and SROC values are relatively high (0.82, Fig. 2(B) and0.73, Fig. 3). This demonstrates that a normal serum bilirubin valuecombined with symptoms and signs of acute appendicitis supportsthe presence of an uncomplicated appendicitis. A major limitationderives from the weaknesses observed in the included studies.Eight studies were used for this pooled analysis: seven of themwere retrospective27,30e35 and only one prospective cohort29 withclear heterogeneity in methods and settings (e.g. country, operative

indications, time since diagnosis Table 2). We extracted data onlyfrom patients with histological findings of acute or perforatedappendicitis, and excluded patients with no histological signs ofacute inflammation. Indeed, no patients presenting organ failurerequiring resuscitation or a laparotomy were present in theincluded articles. Possible confounders are present in all studies,and if a surgeon is conscious of the predictive value he or she mightbe more prone to operate e and therefore diagnose appendicitis ein the presence of hyperbilirubinemia and therefore be more likelyto diagnose appendicitis. Our meta-analysis showed a pooledsensitivity of 0.49, a pooled specificity of 0.82, positive likelihoodratio of 2.51 and negative likelihood ratio of 0.58. A positive like-lihood ratio >5 or negative likelihood ratio <0.2 is indicative ofstrong diagnostic evidence.47 For reference, a positive likelihoodratio of 2 increases probability of disease by 15%,48 and a negativelikelihood ratio of 0.5 decreases the probability of disease by 15%.48

Our relatively low PLR and high NLR, combined with poorsensitivity and specificity, do not give a sustained evidence ofelevated serum bilirubin value in acute appendicitis perforation.However, AUC value in our SROC curve was 0.73 indicating amoderate accuracy of the value of hyperbilirubinemia in disclosingappendix perforation, as it is >0.7.49 Important statistical hetero-geneity (I2 > 75%) was found in all of the analysis (Fig. 2), whichmight be a confounding factor for the results. All included studiesdiffered in design and conduct as well as in participants and

Fig. 3. SROC curve for hyperbilirubinemia and appendiceal perforation. Each study isrepresented by each solid circle in the meta-analysis. The size of the solid circle in-dicates the size of each study. SROC curve summarizes the overall diagnostic accuracy.

S. Giordano et al. / International Journal of Surgery 11 (2013) 795e800 799

REVIEW

interventions. Substantial heterogeneity was identified in terms ofboth size and direction of effect (I2 ¼ 91.3e97.8%, Fig. 2(A)e(E)). Forthis reason a sensitivity analyses was conducted excluding in turneach study. However, the exclusion of any study from the analysisdid not materially change the summary estimates, which indicatesthat publication bias is unlikely to have appreciably influenced thepooled results. The number of patients considered for this meta-analysis was relatively small (a total of 4974 patients), diagnosticmethods used were not standardised with lack of elucidation in thepapers and there was not enough specific data about timing ofoperation since diagnosis.

In summary, this study presents a cumulative analysis of almost5000 patients from studies in different settings and demonstratedthe value of hyperbilirubinemia as a predictor of appendicealperforation. A patient presenting with an acute appendicitis un-dergoes a primary evaluation with a clinical examination, bloodtests and possibly imaging studies in the emergency room anddecision is made between observation, with or without antimi-crobial treatment, planned surgery during the following 12e24 h orimmediate emergency operation.3e10,13,14 If symptoms and signssuggest a perforated appendicitis serum bilirubin measurement isone additional tool in the primary evaluation. Hyperbilirubinemiaalone is not a strong enough predictor, but might be more usefulwhen integrated into a scoring system. In this regard, furtherstudies would be useful especially in resource-poor settings withlimited availability of a CT or other imaging studies.

Ethical approvalNone.

FundingNone.

Author contributionSalvatore Giordano: Developed the idea, performed the statis-

tical analysis and wrote the manuscript.Markus Pääkkönen: Revised the manuscript.Paulina Salminen: Revised the manuscript.Juha Grönroos: Revised the manuscript and supervised the work.

Conflict of interestNone.

Acknowledgements and potential conflict of interest statement

There is nothing to disclose in this manuscript.No commercial associations or disclosures may pose, or create

any conflict of interest with information presented in thismanuscript.

References

1. Petroianu A. Diagnosis of acute appendicitis. Int J Surg 2012;10:115e9.2. Pickhardt PJ, Lawrence EM, Pooler BD, Bruce RJ. Diagnostic performance of

multidetector computed tomography for suspected acute appendicitis. AnnIntern Med 2011;154:789e96.

3. Shindoh J, Niwa H, Kawai K, Shindoh J, Ohata K, Ishihara Y, Takabayashi N, et al.Diagnostic power of inflammatory markers in predicting severity of appendi-citis. Hepatogastroenterology 2011;58:2003e6.

4. Grönroos JM, Forsström JJ, Irjala K, Nevalainen TJ. Phospholipase A2, C-reactiveprotein, and white blood cell count in the diagnosis of acute appendicitis. ClinChem 1994;40:1757e60.

5. Grönroos JM, Grönroos P. Leucocyte count and C-reactive protein in the diag-nosis of acute appendicitis. Br J Surg 1999;86:501e4.

6. Alvarado A. A practical score for the early diagnosis of acute appendicitis. AnnEmerg Med 1986;15:557e64.

7. Kalan M, Talbot D, Cunliffe WJ, Rich AJ. Evaluation of the modified Alvaradoscore in the diagnosis of acute appendicitis: a prospective study. Ann R Coll SurgEngl 1994;76:418e9.

8. Escriba A, Gamell AM, Fernandez Y, Quintillá JM, Cubells CL. Prospective vali-dation of two systems of classification for the diagnosis of acute appendicitis.Pediatr Emerg Care 2011;27:165e9.

9. Chong CF, Thien A, Mackie AJ, Tin AS, Tripathi S, Ahmad MA. Comparison ofRIPASA and Alvarado scores for the diagnosis of acute appendicitis. SingaporeMed J 2011;52:340e5.

10. de Castro SM, Unlü C, Steller EP, van Wagensveld BA, Vrouenraets BC. Evalu-ation of the appendicitis inflammatory response score for patients with acuteappendicitis. World J Surg 2012;36:1540e5.

11. Abou-Nukta F, Bakhos C, Arroyo K, Koo Y, Martin J, Reinhold R, et al. Effects ofdelaying appendectomy for acute appendicitis for 12 to 24 hours. Arch Surg2006;141:504e6.

12. Vons C, Barry C, Maitre S, Pautrat K, Leconte M, Costaglioli B, et al. Amoxicillinplus clavulanic acid versus appendicectomy for treatment of acute uncompli-cated appendicitis: on open-label, non-inferiority, randomized controlled trial.Lancet 2011;377:1573e9.

13. Wilms IM, de Hoog DE, de Visser DC, Janzing HM. Appendectomy versusantibiotic treatment for acute appendicitis. Cochrane Database Syst Rev2011;11. CD008359. PMID: 22071846.

14. Grönroos JM. Clinical suspicion of acute appendicitis e is the time ripe formore conservative treatment? Minim Invasive Ther Allied Technol 2011;20:42e5.

15. Miller DJ, Keeton DG, Webber BL, Pathol FF, Saunders SJ. Jaundice in severebacterial infection. Gastroenterology 1976;71:94e7.

16. Johnson AM. Sepsis and jaundice. Pediatrics 1993;91:1018.17. Whitehead MW, Hainsworth I, Kingham JG. The causes of obvious jaundice in

South West Wales: perceptions versus reality. Gut 2001;48:409e13.18. Franson TR, Hierholzer Jr WJ, LaBrecque DR. Frequency and characteristics

of hyperbilirubinemia associated with bacteremia. Rev Infect Dis 1985;7:1e9.

19. Cheng CW, Lin HS, Ye JJ, Yang CC, Chiang PC, Wu TS, et al. Clinical significanceof and outcomes for Bacteroides fragilis bacteremia. J Microbiol Immunol Infect2009;42:243e50.

20. Ogilvie TA. Acholuric jaundice and appendicitis. Br J Surg 1951;38:529e30.21. Eley A, Hargreaves T, Lambert HP. Jaundice in Severe Infections. Br Med J

1965;5453:75e7.22. Miller DF, Irvine RW. Jaundice in acute appendicitis. Lancet 1969;1:321e3.23. Seller RA. Jaundice in acute appendicitis. Lancet 1969;1:838.24. Ermich S, Kantor Z. Jaundice in the course of acute appendicitis. Pol Tyg Lek

1970;25:1583e4.25. Meyer G, Landi JP, Pitchumont CS. Acute appendicitis presenting as cholestatic

jaundice. N Y State J Med 1984;84:131e2.26. Lotveit T. Acute appendicitis in patients with Gilbert’s syndrome. Acta Chir

Scand 1985;151:701e2.27. Estrada JJ, Petrosyan M, Barnhart J, Tao M, Sohn H, Towfigh S, et al. Hyper-

bilirubinemia in appendicitis: a new predictor of perforation. J Gastrointest Surg2007;11:714e8.

28. Khan S. Evaluation of hyperbilirubinemia in acute inflammation of ap-pendix: a prospective study of 45 cases. Kathmandu Univ Med J 2006;4:281e9.

29. Khan S. Elevated serum bilirubin in acute appendicitis: a new diagnostic tool.Kathmandu Univ Med J 2008;6:161e5.

30. Sand M, Bechara FG, Holland-Letz T, Sand D, Mehnert G, Mann B. Diagnosticvalue of hyperbilirubinemia as a predictive factor for appendiceal perforationin acute appendicitis. Am J Surg 2009;198:193e8.

S. Giordano et al. / International Journal of Surgery 11 (2013) 795e800800

REVIEW

31. Käser SA, Fankhauser G, Willi N, Maurer CA. C-reactive protein is superior tobilirubin for anticipation of perforation in acute appendicitis. Scand J Gastro-enterol 2010;45:885e92.

32. Atahan K, Üreyen O, Aslan E, Deniz M, Çökmez A, Gür S, et al. Preoperativediagnostic role of hyperbilirubinaemia as a marker of appendix perforation. JInt Med Res 2011;39:609e18.

33. Emmanuel A, Murchan P, Wilson I, Balfe P. The value of hyperbilirubinaemia inthe diagnosis of acute appendicitis. Ann R Coll Surg Engl 2011;93:213e7.

34. Hong YR, Chung CW, Kim JW, Kwon CI, Ahn DH, Kwon SW, et al. Hyper-bilirubinemia is a significant indicator for the severity of acute appendicitis. JKorean Soc Coloproctol 2012;28:247e52.

35. McGowan DR, Sims HM, Zia K, Uheba M, Shaikh IA. The value of biochemicalmarkers in predicting a perforation in acute appendicitis. ANZ J Surg 2013;83:79e83.

36. Moher D, Liberati A, Tetzlaff J, Altman DG. PRISMA Group. Preferred reportingitems for systematic reviews and meta-analyses: the PRISMA statement. Int JSurg 2010;8:336e41.

37. Whiting PF, Weswood ME, Rutjes AW, Reitsma JB, Bossuyt PN, Kleijnen J.Evaluation of QUADAS, a tool for the quality assessment of diagnostic accuracystudies. BMC Med Res Methodol 2006;6:9.

38. Leeflang MM, Deeks JJ, Gatsonis C, Bossuyt PM. Cochrane Diagnostic Test Ac-curacy Working Group. Systematic reviews of diagnostic test accuracy. AnnIntern Med 2008;149:889e97.

39. Deville WL, Buntinx F, Bouter LM, Montori VM, de Vet HC, van der Windt DA.Conducting systematic reviews of diagnostic studies: didactic guidelines. BMCMed Res Methodol 2002;2:9.

40. Zamora J, Abraira V, Muriel A, Khan K, Coomarasamy A. Meta-DiSc: a softwarefor meta-analysis of test accuracy data. BMC Med Res Methodol 2006;6:31.

41. Moses LE, Shapiro D, Littenberg B. Combining independent studies of a diag-nostic test into a summary ROC curve: data-analytic approaches and someadditional considerations. Stat Med 1993;12:1293e316.

42. Lau J, Ioannidis JP, Balk EM, Milch C, Terrin N, Chew PW, et al. Diagnosing acutecardiac ischemia in the emergency department: a systematic review of theaccuracy and clinical effect of current technologies. Ann Emerg Med 2001;37:453e60.

43. Irwig L, Tosteson AN, Gatsonis C, Lau J, Colditz G, Chalmers TC, et al. Guidelinesfor meta-analyses evaluating diagnostic tests. Ann Intern Med 1994;120:667e76.

44. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency inmeta-analyses. BMJ 2003;327:557e60.

45. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials1986;7:177e88.

46. Fike FB, Mortellaro VE, Juang D, Sharp SW, Ostlie DJ, St Peter SD. The impact ofpostoperative abscess formation in perforated appendicitis. J Surg Res2011;170:24e6.

47. Jaeschke R, Guyatt GH, Sackett DL. Users’ guides to the medical literature. III.How to use an article about a diagnostic test. B. What are the results and willthey help me in caring for my patients? The Evidence-Based Medicine WorkingGroup. JAMA 1994;2(271):703e7.

48. McGee S. Simplifying Likelihood Ratios. Gen Intern Med 2002;17:647e50.49. Swets JA. Measuring the accuracy of diagnostic systems. Science 1988;3(240):

1285e93.