Use of changes over time in seruminflammatory parameters in patientswith equivocal appendicitisHan-Ping Wu, MD,a,e Chin-Yi Huang, MS,b Yu-Jun Chang, BSB,c Chu-Chung Chou, MD,a andChing-Yuang Lin, MD, PhD,d,e Changhua, Taiwan

Background. The aim of this prospective study was to determine whether the change between the initialand subsequent analysis of serum inflammatory markers measured 6 to 10 hours later could improvediagnostic accuracy in appendicitis.Methods. The study group comprised 225 patients who presented to hospital with equivocal signs ofappendicitis from 2003 to 2004. Based on the period from the onset of symptoms to admission, receiveroperating characteristic (ROC) curves were used to determine cutoff values of the changes betweenprimary and repeated laboratory examinations in diagnosing appendicitis.Results. ROC analysis showed that the cutoff values for the change in total percentage neutrophilcount on the first day after onset of symptoms (3.2%), and a change of in CRP concentration abovethe baseline on the second day (9.5 mg/L) and the third day (17.0 mg/L) after onset of symptoms weresignificant parameters for diagnosing acute appendicitis.Conclusion. The cutoff values of the change in total neutrophil count on the first day after onset ofsymptoms and the change in CRP on the second and third days after onset of symptoms during in-hospital observation may serve as useful parameters to surgeons in differentiating acute appendicitisfrom other acute abdominal diseases. (Surgery 2006;139:789-96.)

From the Department of Emergency Medicine,a Evidence Base Medicine Center,b Laboratory of Epidemiology andBiostatistics,c and Department of Pediatrics Changhua Christian Hospital,d Institute of Medical Research,e Chang

Jung Christian University, Changhua, Taiwan

To rule out appendicitis a great number of pa-tients with right lower quadrant pain are evaluatedby primary clinicians every day. As evidenced byhigh rates of misdiagnosis and perforation, the di-agnosis of appendicitis can still be difficult and isprobably the most common problem in clinicaloperation.1–3 Diagnostic difficulties lead surgeonsto carry out unnecessary appendectomies, whereasmisdiagnosis can lead to perforation and abscessformation.3–5 In general, a normal appendix foundduring appendectomy represents a misdiagnosis,and a delayed diagnosis of acute appendicitis maylead to perforation and peritonitis. Thus, improv-ing accuracy is desirable both for earlier diagnosis

Accepted for publication December 15, 2005.

Reprint requests: Ching-Yuang Lin, MD, PhD, Professor andDirector, Children’s Hospital, Changhua Christian Hospital,135 Nanhsiao Street, Changhua 500, Taiwan. E-mail: 100966@cch.org.tw

0039-6060/$ - see front matter

© 2006 Mosby, Inc. All rights reserved.

doi:10.1016/j.surg.2005.10.021

and for avoiding unnecessary appendectomies. Al-though technologic advancements, such as spiralcomputed tomography (CT), nuclear medicinescans, and ultrasonography, are valuable aids,1,3–6

these procedures are not available in all healthcaresettings and the cost remains high. In contrast,some preoperative laboratory tests, such as totalwhite blood cell (WBC) count, total neutrophilcount, manual band count and C-reactive protein(CRP) count, are fast, economical, and universallyavailable, and often aid primary clinicians with de-cision making about patients with suspected appen-dicitis.4–8

Classically, the diagnosis of acute appendicitis isoften based on a brief history, physical examina-tion, and laboratory findings. Furthermore, oncethe diagnosis of appendicitis is made, consequentoperative intervention will be indicated. To ourknowledge, the diagnostic accuracy based on theseclinical findings ranges from 70% to 80%.1,4,9 Insome situations, the clinical presentations are atyp-ical, and diagnostic errors are made easily. Duringobservation of patients with an equivocal diagnosis

of appendicitis, repeated laboratory examination

SURGERY 789

790 Wu et al SurgeryJune 2006

may provide primary clinicians further informationabout decision making. However, the role of re-peated laboratory examinations is not clear, andthe diagnostic value remains controversial. Thepurpose of this study was to investigate whethersome of these repeated serum inflammatory mark-ers, including WBC count, percentage neutrophilcount, manual band count, and CRP concentra-tion, could be used to predict acute appendicitisafter 6 to 10 hours of observation. In addition, wealso aimed to determine the change between pri-mary and repeated analysis of serum inflammatorymarkers over time to improve diagnostic accuracyin patients with an equivocal diagnosis of appendi-citis.

PATIENTS AND METHODSPatient population. From 2003 to 2004, we ana-

lyzed prospectively 645 patients admitted for clini-cally suspected acute appendicitis at the ChanghuaChristian Hospital, Taiwan. The duration of symp-toms comprised the period from the time the pa-tients first felt ill until the time of admission. Weidentified the duration within 24 hours as day 1, 24to 48 hours as day 2, and 48 to 72 hours as day 3 inour study. The term “clinically suspected acute ap-pendicitis” was reserved for patients with clinicalsymptoms and signs that included migration ofpain, anorexia, nausea, rebound tenderness in theright lower quadrant of the abdomen, pyrexia, andtenderness over the right iliac fossa. Patients under4 years of age were excluded. Also, patients whosesymptoms and signs lasted more than three dayswere excluded from analysis. Appendicitis was di-agnosed by the faculty physician treating the pa-tients. This study was approved by the institution’sHuman Subjects Review Committee.

In total, 420 patients were treated by operativemethods without further investigation, whereas 225patients were admitted for observation. The deci-sion to intervene surgically was made at the discre-tion of senior surgeons. A repeated laboratory ex-amination was carried out in these patients underobservation after a median duration of 8 hours(range, 6 to 10 hours). The ultimate diagnosis ofthe patients who were treated by operative methodswas based on histologic examination of the excisedappendix. In addition, a patient was defined ashaving a normal appendix when symptoms andsigns in a nonoperative patient subsided and thepatient was discharged, and when follow-up at theoutpatient clinic 2 weeks later confirmed that thediagnosis of appendicitis could be ruled out. Anormal appendix was also defined when an unin-

flamed appendix was found in a patient who had

undergone appendectomy with intent to treatacute appendicitis (normal appendectomy).

Methods. This was a prospective observationalstudy and was not intended to influence the indi-cations and timing of operative approach. Facultyphysicians completed standardized patient datasheets on patient presentation to the hospital. Thefollowing data were recorded on the data sheet:age, gender, body temperature, time of onset ofsymptoms, and time of admission. The faculty phy-sicians’ evaluation of the possibility of appendicitiswas made based on the Alvarado score.10 On ad-mission, blood samples were obtained from all se-lected participant, and analyzed for WBC count,percentage neutrophil count, manual band count,and CRP concentration. Repeated laboratory ex-amination was carried out in patients with an equiv-ocal diagnosis of appendicitis after 6 to 10 hours ofobservation. The change between primary and re-peated laboratory examinations was recorded onthe sheet at the time the primary laboratory test wascarried out. The correlation between the changeand the period recorded from the onset of symp-toms to admission was analyzed statistically in pa-tients with equivocal appendicitis.

The total WBC count and percentage neutrophilcounts were measured by an automated five-partleukocyte differential count hematology analyzer(Cell-Dyn 4000R System, Abbot Laboratories, Ab-bot Park, Ill). A WBC count above 10,000/mm3,and a neutrophil count greater than 80% wereconsidered abnormal. The nonsegmented bandcounts were carried out manually by counting 100consecutive cells in a peripheral blood film. A man-ual band count greater than 5% was consideredabnormal. The concentration of CRP in serum wasmeasured by immunoturbidimetry (BeckmanCoulter, Fullerton, Calif). A CRP concentrationabove the detection level of 8.0 mg/L was taken tobe elevated.

In addition, we further analyzed the results ofimage studies, including abdominopelvic CTs andabdominal ultrasonographics in these 225 patients.Reports of CTs were all made by senior radiologists,and abdominal ultrasonographies were carried outby gastrointestinal subspecialists. We compared thediagnostic accuracy of repeated blood tests versusimaging studies for appendicitis.

Statistical analysis. The methods of statisticalanalysis were the t test, multivariate logistic regres-sion analysis, and the receiver operating character-istic (ROC) curve. Possibility levels of less than 0.05were taken as statistically significant. The differ-ence between groups was presented as 95% confi-

dence intervals (CIs). Our logistic regression anal-

Surgery Wu et al 791Volume 139, Number 6

ysis used an automatic stepwise procedure. Webegan by selecting the strongest candidate predic-tor, then testing additional candidate predictors,one at a time, for inclusion in the model. At eachstep, we checked to see whether a new candidatepredictor would significantly improve the model. Ifthe new predictor was included in the model, wealso checked to see whether any other predictorsalready in the model should remain or be removed.If a newly entered predictor did a better job ofexplaining the default behavior, then it was possi-ble for a predictor already in the model to beremoved from the model, because it no longeruniquely explained enough. This stepwise proce-dure was continued until all of the candidate pre-dictors had been tested thoroughly for inclusionand removal. The variables selected in our logisticregression analysis were: gender, WBC count, per-centage neutrophil count, manual band count andCRP concentration.

The informative value of each biologic markerwas determined, and the optimal operating pointwas defined by ROC analysis. The decision tochoose the optimal operating point was madewith full understanding of the implication of thechoices.11,12 The optimal operating point was de-termined as the predictive probability such thatboth sensitivity and specificity are maximized. TheROC is a monotonic curve such that as sensitivitygoes up, specificity goes down, and as sensitivitygoes down, specificity goes up. The optimal oper-ating point corresponds to a point that minimizesthe probability of both types of misclassifications.To achieve this effect, the optimal operating pointis determined by the following steps: 1) for each i,mi � min(xi,yi), where xi is the sensitivity and yi isthe specificity for the ith predicted probability; 2)let M � max(m1, m2,..., m); 3) the optimal oper-ating point corresponds to the predicted probabil-ity that produces M.12 Once the optimal operatingpoint is chosen, the optimal operating point pro-vides a graphic interpretation of predicting re-sponse. The test characteristics of the cutoff pointsof these serum markers, including sensitivity (Sn),specificity (Sp), positive predictive value (PPV),negative predictive value (NPV), area under theROC curve (AUC), positive likelihood ratio (LR�),and negative likelihood ratio (LR�) were exam-ined. The AUC, calculated using the trapezoidalrule, was considered a global measure of the diag-nostic value of the parameter. An optimal test re-sult gives a value of 1.0, and a useless test resultgives a value of 0.5. LR� and LR� were calculated

for the best cutoff values. Statistical analyses were

carried out with SPSS software (version 11.0, SPSS,Inc., Chicago, Ill).

RESULTSDuring the study period, 225 patients whose

symptoms and signs lasted less than 3 days wereenrolled in the study. There were 93 patients onday 1, 71 patients on day 2, and 61 patients on day3. The patients comprised 121 males (53.8%) and104 females (46.2%) (mean age, 29.9 � 19.3 years;range, 4 to 87 years). Of those, 107 had histologi-cally proven acute appendicitis, and 118 had nor-mal appendices. Among those with normal appen-dices, 85 were treated by nonoperative methodsand finally discharged from our hospital, and 33underwent a normal appendectomy. Of the 420patients who underwent immediate appendectomy,342 had appendicitis, whereas 78 had normal ap-pendices. Patients with normal appendices werelater diagnosed as having had infectious enteritis,functional gastrointestinal disorders, diverticulitis,Meckel’s diverticulitis, mesenteric lymphadenopa-thy, lymphoma, pelvic inflammatory disease, orruptured tubo-ovarian abscess. The most commondiagnosis was infectious enteritis followed by func-tional gastrointestinal disorders. Baseline charac-teristics of both the 420 who underwent immediateappendectomy, and the 225 in the study are pre-sented in Table I.

The change between primary and repeated se-rum markers in patients with appendicitis and nor-mal appendices were analyzed on the first 3 daysafter onset of symptoms (Table II). A change intotal neutrophil count indicated a change in thepercentage of white blood cells that were neutro-phils. On day 1, the change in total neutrophilcount (delta percentage neutrophils) was greaterin patients with appendicitis than in those withnormal appendices (P � .05). On day 2, thechanges in percentage neutrophil count, manualband count, and CRP concentration were allgreater in patients with appendicitis (P � .05). OnDay 3, only the change in CRP concentration (deltaCRP) was statistically significant (P � .005). Theresults of forward stepwise logistic regression anal-ysis indicated that the significant changes in dis-cerning acute appendicitis from other causes ofabdominal pain were delta percentage neutrophilson day 1, and delta CRP on days 2 and 3 (P � .05)(Table III).

The box plots show the distribution of percent-age neutrophil count on day 1, and CRP values ondays 2 and 3 between patients with normal appen-dices and patients with appendicitis (Fig 1). Serum

levels of both primary and repeated tests in these

C, white

792 Wu et al SurgeryJune 2006

markers were greater in patients with appendicitisthan in patients with normal appendices. Whetheror not the changes in biomarkers between the 2examinations on the first three days after onset of

Table I. Baseline characteristics of the subjects en

Patients with immediate app(n � 420)

Normalappendices

Simpleappendicitis*

No. of subjects 78 270Age (mean � SD) 22.1 � 17.9 21.2 � 18.1Gender (female:male) 38:40 97:173Imaging studiesCT (n) 17 48Ultrasonography (n) 18 21Laboratory testPrimary (n) 73 263Repeated (n) 0 0

*Non-perforated appendicitis.

Table II. Change between primary and repeated snormal appendices on the first 3 days after onset o

Day*Normal appendice

(mean � SD)

Day 1WBC (mm3) �764 � 3675Neutrophils (%) �2.46 � 12.60Manual bands (%) �0.33 � 2.72CRP (mg/L) 3.42 � 26.26

Day 2WBC (mm3) �461 � 10227Neutrophils (%) �6.31 � 13.67Manual bands (%) �1.46 � 5.21CRP (mg/L) �0.25 � 7.64

Day 3WBC (mm3) �1855 � 5499Neutrophils (%) �22.44 � 106.96Manual bands (%) �1.19 � 6.32CRP (mg/L) �4.69 � 13.99

CRP, C-reactive protein; NS, not significant; SD, standard deviation; WB*The period from the onset of symptoms to admission.

Table III. Stepwise logistic regression analysis ofthe change in significant biomarkers between the2 examinations in predicting appendicitis

Day Difference Coefficient SE OR 95% CI P value

1 Neutrophils 0.12 0.04 1.12 1.04–1.22 .0052 CRP 0.54 0.24 1.72 1.07–2.75 .0243 CRP 0.12 0.06 1.13 1.01–1.26 .032

CI, confidence interval; SE, standard error; OR, odds ratio.

symptoms could predict appendicitis was analyzed

by ROC curves (Fig 2). The AUCs were all greaterthan 0.5. The cutoff value of delta percentage neu-trophils on day 1 was 3.2% (AUC, 0.71; 95% CI,0.60 to 0.80; Sn, 0.60; Sp, 0.82; PPV, 0.82; NPV,0.60; LR�, 3.30; LR�, 0.50), and the cutoff value indelta CRP on day 2 was 9.5 mg/L (AUC, 0.97; 95%CI, 0.84 to 0.99; Sn, 0.93; Sp, 0.90; PPV, 0.95; NPV,0.88; LR�, 13.42; LR�, 0.11) and on day 3 was 17.0mg/L (AUC, 0.91; 95% CI, 0.74 to 0.98; Sn, 0.82;Sp, 1.00; PPV, 1.00; NPV, 0.90; LR�, —; LR�, 0.18)in distinguishing acute appendicitis from otheracute abdominal diseases.

The use of these serum parameters as the soleindicator for diagnosing appendicitis on days 1 to 3in the 225 patients with equivocal presentations forappendicitis gave: Sn, 0.75; Sp, 0.92; PPV, 0.92;

in our study

omy Patients with a short-term observation(n � 225)

rforatedendicitis

Normalappendices

Simpleappendicitis*

Perforatedappendicitis

72 118 89 18� 24.4 27.8 � 19.3 30.9 � 17.9 39.1 � 13.17:45 61:57 36:53 7:11

10 23 32 422 39 24 7

72 118 90 170 118 90 17

markers in patients with appendicitis andptoms

Appendicitis(mean � SD) P value

4 � 4129 NS4.11 � 12.07 .0131.48 � 6.07 NS

14.49 � 17.31 NS

736 � 4877 NS2.07 � 9.16 .0034.40 � 12.60 .014

52.26 � 55.33 �.001

�844 � 3855 NS�1.26 � 11.80 NS

4.39 � 12.39 NS65.94 � 59.91 .003

blood cell.

rolled

endect

Peapp

26.32

erumf sym

s

NPV, 0.77. In addition, of these 225 patients, 59

patients with appendicitis and normal appendices.

Surgery Wu et al 793Volume 139, Number 6

underwent CTs, and 70 underwent ultrasonogra-phies, the use of abdominopelvic CT scan for diag-nosing appendicitis showed: Sn, 0.89; Sp, 0.65;PPV, 0.80; and NPV, 0.79, and abdominal ultra-sound showed: Sn, 0.42; Sp, 0.72; PPV, 0.52; andNPV, 0.60.

DISCUSSIONAcute appendicitis is the most common problem

requiring emergency abdominal operation.1–6

Early diagnosis of appendicitis can prevent perfo-ration, abscess formation and postoperative com-plications, and can decrease cost by decreasing hos-pitalizations time. However, definitive diagnosis ofappendicitis is made in only 50% to 70% of patientsat the time of initial assessment of acute abdominalpain in the emergency department.9,10,13 Patientswith an equivocal diagnosis of appendicitis are ad-mitted to the hospital usually for a short period ofobservation and require further investigation. Clin-ically, although the utility of repeated laboratoryexaminations may seem helpful, these serial analy-ses have not been studied thoroughly enough andapplied universally. Also, it is unclear which seruminflammatory markers should be focused on byprimary clinicians according to the period of timesymptoms were present. Moreover, it is unknownhow well changes between primary and repeatedexaminations can function as a strong discrimina-tor in patients with an equivocal diagnosis of ap-pendicitis after short-term observation. In this pro-spective study, we analyzed 225 patients with aclinically equivocal diagnosis of appendicitis anddetermined the effective discriminators for equivo-cal appendicitis based on the first 3 days after onsetof symptoms. We also established cutoff values ofthe changes between the 2 analyses to discern ap-pendicitis from other acute abdominal diseases.

Numerous studies have concluded that an in-creased percentage of the neutrophil count aids inthe diagnosis of acute appendicitis on admis-sion.4,5,14 However, previous studies did not men-tion how long the patients’ symptoms and signspersisted from onset to admission. The durationmay influence the true discriminating power of thepercentage neutrophil count and limit its clinicalapplication. In our study, based on the result offorward stepwise logistic regression analysis, thepercentage neutrophil count was found to be auseful inflammatory parameter on day 1. We alsofound that the change between primary and re-peated examinations increased the accuracy of di-agnosing appendicitis. The cutoff point of thechange in the percentage neutrophil count on day

Fig 1. Box plots of percentage neutrophil counts on day1 (A) and CRP levels on day 2 (B) and day 3 (C) in

1 was determined based on the result of ROC anal-

794 Wu et al SurgeryJune 2006

ysis. We found that once the change in percentageneutrophil count (delta percentage neutrophilcount) was greater than 3.2% on day 1, the proba-bility of acute appendicitis increased. Although itssensitivity was not high enough (0.60), the specific-ity was acceptable (0.82). On the first day afteronset of symptoms, we think that ruling out appen-dicitis seems more important than to undergo anunnecessary appendectomy in patients with anequivocal presentation for appendicitis. Therefore,we chose the cutoff value of delta percentage neu-trophil count to maintain a higher specificity thansensitivity on day 1. We believe that this cutoff valueof delta percentage neutrophil count can help sur-geons to lower the negative appendectomy rate onday 1, and may serve as an effective discriminatorfor patients with equivocal appendicitis within thefirst 6 to 10 hours of observation. Severe inflamma-tion and infection has been shown to prime andactivate neutrophils within 3 to 6 hours afterinjury.15–17 Based on these reports, we believe thatthe activation of neutrophils primed by acute in-flammation occurs on the first day after onset ofthe symptoms of appendicitis.

We also found that increases in CRP concentra-tion on day 2 and day 3 proved to be an importantpredictor in diagnosing appendicitis. Clinically,once the change in CRP concentration (delta CRP)was greater than 9.5 mg/L on day 2, and greaterthan 17.0 mg/L on day 3, the probability of acuteappendicitis increased. Based on large AUCs,changes in CRP concentration served as a strongdiscriminator in patients with an equivocal diagno-sis of appendicitis on day 2 and day 3 observation.CRP has been used to screen for occult inflamma-tion, as a marker of disease activity, and as a diag-nostic tool.7 However, several studies have reportedthat serial analyses of serum CRP show an increaseafter 12 to 24 hours of onset of symptoms of in-flammation.4,5,7,14,18–20 When symptoms of acuteappendicitis proceed rapidly, an increase in CRPconcentration may not be observed on admissionand may explain why the change in CRP concen-tration was significant on day 2 and day 3, but noton day 1.

In our study, patients with “clinically suspectedacute appendicitis” were enrolled, and their clini-cal diagnoses were made based on the Alvarado

Fig 2. The ROC curves of the change in total neutrophilcount on day 1 (A) and in CRP concentration on day 2(B) and day 3 (C) for discriminating between appendi-

citis and normal appendices.

Surgery Wu et al 795Volume 139, Number 6

score. However, a patient with a normal appendixsuffering from right lower quadrant pain may havean inflammatory condition in the abdomen relatedto something else. In our study, the patients whounderwent appendectomy but had normal appen-dices were diagnosed finally as having infectiousenteritis, functional gastrointestinal disorders, di-verticulitis, mesenteric lymphadenopathy, lym-phoma, pelvic inflammatory disease, and rupturedtubo-ovarian abscess. The percentage of neutrophilcounts and CRP levels in other abdominal inflam-matory conditions may be increased at the time ofinitial assessment, but we found that the changesbetween primary and repeated examinations didnot show a significant increase after short-term ob-servation. Therefore, we believe that the changesafter short-term observation in percentage neutro-phil counts on day 1, and CRP on day 2 and day 3are relatively specific for acute appendicitis and canbe used to exclude other inflammatory conditionsin the abdomen.

CT and ultrasonography have become impor-tant tools for diagnosing appendicitis. However,these procedures are not readily available in allhealthcare settings in the United States, Europe,and Taiwan. In addition, their cost remains high.In Taiwan, the cost of a CT is 2 times higher thanadmission to the ward overnight, and 10 timeshigher than admission to the observation unit inthe emergency department. The cost of carryingout an abdominal ultrasonography is 3 timeshigher than admission to the observation unit. Re-peated laboratory tests in patients under short-termobservation cost less than a CT or ultrasonographyin our medical center. Moreover, in some patients,CT or ultrasonography may show equivocal find-ings.21–23 The radiologist may be uncertain how tointerpret such a study and not be able to give animmediate diagnosis of appendicitis.21–23 Patientswith equivocal CT findings can not be dischargedhome and need to be admitted for observation.Therefore, we believe that repeated laboratory testsin patients under short-term observation are notonly a relatively reliable method of diagnosingequivocal appendicitis but also an economic diag-nostic tool.

Patient recall error in our study needs to beaddressed. We classified the duration of symptomsinto days 1 to 3 according to the statements ofpatients and their family. Although it is possiblethat patient recall errors may have occurred andcarry misclassification, such misclassification maynot lead to severe bias in group difference, becausemisclassified patients would be divided into an ap-

pendicitis group and a normal appendix group.

Therefore, we assume any patient recall errorswould not bias our study.

During the period of in-hospital observation,once the diagnosis of appendicitis was made, ap-pendectomy was indicated.24–26 Accurate preoper-ative diagnosis of appendicitis is important andessential. Patients with equivocal signs of appendi-citis require short-term observation and can un-dergo repeated laboratory examination. We main-tain that analyses of serum inflammatory markerscan aid primary clinicians in integrating the indi-cations of operative intervention during observa-tion as a potential alternative to abdominopelvicCT or ultrasonography.

REFERENCES1. Andersson RE. Meta-analysis of the clinical and laboratory

diagnosis of appendicitis. Br J Surg 2004;91:28-37.2. Wu HP, Lin CY, Chang CF, Chang YJ, Huang CY. Predictive

value of C-reactive protein at different cutoff levels in acuteappendicitis. Am J Emerg Med 2005;23:449-53.

3. Kokoska ER, Silen ML, Tracy TF Jr, Dillon PA, Cradock TV,Weber TR. Perforated appendicitis in children: risk factorsfor the development of complications. Surgery 1998;124:619-25.

4. Wu HP, Chang CF, Lin CY. Predictive inflammatory param-eters in the diagnosis of acute appendicitis in children. ActaPaediatr Taiwan 2003;44:227-31.

5. Tompson MM, Underwood MJ, Dookeran KA, Lloyd DM,Bell PR. Role of sequential leucocyte counts and C-reactiveprotein measurement in acute appendicitis. Br J Surg1992;70:822-4.

6. Barron B, Hanna C, Passalaqua AM, Lamki L, Wegener WA,Goldenberg DM. Rapid diagnostic imaging of acute, non-classic appendicitis by leukoscintigraphy with sulesomab, atechnetium 99m-labeled antigranulocyte antibody Fab= frag-ment. LeukoScan Appendicitis Clinical Trial Group. Sur-gery 1999;125:288-96.

7. Clyne B, Olshaker JS. The C-reactive protein. J Emerg Med1999;17:1019-25.

8. Paajanen H, Mansikka A, Laato M Ristamaki R, Pulkki K,Kostiainen S. Novel serum inflammatory markers in acuteappendicitis. Scand J Clin Lab Invest 2002;62:579-84.

9. Andersson RE, Hugander A, Ravn H, Offenbartl K, GhaziSH, Nystrom PO, et al. Repeated clinical and laboratoryexaminations in patients with an equivocal diagnosis ofappendicitis. World J Surg 2000;4:479-85.

10. Saidi RF, Ghasemi M. Role of Alvarado score in diagnosisand treatment of suspected acute appendicitis. Am J EmergMed 2000;2:230-1.

11. Ackerman DL, Greenland S, Bystritsky A. Use of receiver-operator characteristic (ROC) curve analysis to evaluatepredictors of response to clomipramine therapy. Psycho-pharmacol Bull 1996;32:157-65.

12. Riddle DL, Stratford PW. Interpreting validity indexes fordiagnostic tests: an illustration using Berg balance test. PhysTher 1999;79:939-48.

13. Samuwl M. Pediatric appendicitis score. J Pediatr Surg2002;37:877-81.

14. Paajanen H, Mansikka A, Laato M, Kettunen J, Kostiainen S.Are serum inflammatory markers age dependent in acuteappendicitis? J Am Coll Surg 1997;164:303-8.

15. Botha AJ, Moore FA, Moore EE, Kim FJ, Banerjee A, Peter-

796 Wu et al SurgeryJune 2006

son VM. Postinjury neutrophil priming and activation: anearly vulnerable window. Surgery 1995;118:358-64.

16. Van Leeuwen HJ, Van Der Tol M, Van Strijp JA, Verhoef J,van Kessel KP. The role of tumour necrosis factor in thekinetics of lipopolysaccharide-mediated neutrophil primingin whole blood. Clin Exp Immunol 2005;140:65-72.

17. Brazil TJ, Dagleish MP, McGorum BC, Dixon PM, Haslett C,Chilvers ER. Kinetics of pulmonary neutrophil recruitmentand clearance in a natural and spontaneously resolvingmodel of airway inflammation. Clin Exp Allergy 2005;35:854-65.

18. Gurleyik E, Gurleyik G, Unalmiser S. Accuracy of serumC-reactive protein measurements in diagnosis of acute ap-pendicitis compared with surgeon’s clinical impression. DisColon Rectum 1995;38:1270-4.

19. Oosterhuis WP, Zwinderman AH, Teeuwen M. C-reactiveprotein in the diagnosis of acute appendicitis. Eur J Surg1993;159:115-9.

20. Gronroos JM, Forsstron JJ, Irjala K, Nevalainen TJ. Phospho-lipase A2, C-reactive protein, and white blood cell count inthe diagnosis of acute appendicitis. Clin Chem 1994;40:

21. Daly CP, Cohan RH, Francis IR, Caoili EM, Ellis JH, Nan B.Incidence of acute appendicitis in patients with equivocalCT findings. Am J Roentgenol 2005;184:1813-20.

22. Funaki B, Grosskreutz SR, Funaki CN. Using unenhancedhelical CT with enteric contrast material for suspected ap-pendicitis in patients treated at a community hospital. Am JRoentgenol 1998;171:997-1001.

23. Weyant MJ, Eachempati SR, Maluccio MA, Rivadeneira DE,Grobmyer SR, Hydo LJ, et al. Interpretation of computedtomography does not correlate with laboratory or patho-logic findings in surgically confirmed acute appendicitis.Surgery 2000;128:145-52.

24. Guller U, Jain N, Peterson ED, Muhlbaier LH, Eubanks S,Pietrobon R. Laparoscopic appendectomy in the elderly.Surgery 2004;135:479-88.

25. Long KH, Bannon MP, Zietlow SP, Helgeson ER, HarmsenWS, Smith CD, et al. A prospective randomized comparisonof laparoscopic appendectomy with open appendectomy:clinical and economic analyses. Surgery 2001;129:390-400.

26. Meter DE, Guzzetta PC, Barber RG. Perforated appendicitisin children: is there a best treatment? J Pediatr Surg

1757-60. 2003;38:1520-4.