PEDIATRICS/ORIGINAL RESEARCH

What Are the Most Clinically Useful Cutoffs for the Alvarado
and Pediatric Appendicitis Scores? A Systematic Review
Mark H. Ebell, MD, MS*; JoAnna Shinholser, BSHP
*Corresponding Author. E-mail: ebell@uga.edu, Twitter: @markebell.

Study objective: The objective of this study is to systematically review the accuracy of the Alvarado score and Pediatric
Appendicitis Score and to identify optimal cutoffs for low- and high-risk populations.

Methods: We performed a systematic review of the literature and identified 26 studies of the accuracy of the Alvarado
score and Pediatric Appendicitis Score. Data were abstracted in parallel, and only prospective, cohort studies that
avoided verification bias were included. We calculated summary likelihood ratios for low-, moderate-, and high-risk
groups, using all possible cutoffs based on available data, even if not reported in the original study.

Results: The pretest probability of appendicitis was approximately 33% in studies of children and approximately 66% in
studies of adults. Likelihood ratios at different cutoffs for the Alvarado score in adults were as follows: 0.03
(<4 points), 0.42 (4 to 6 points), and 3.4 (
7 points); and 0.01 (<5 points), 0.98 (5 to 8 points), and 6.7 (
9 points).
Likelihood ratios for the Alvarado score in children were as follows: 0.02 (<4 points), 0.27 (4 to 6 points), and
4.2 (
7 points); and 0.04 (<5 points), 1.2 (5 to 8 points), and 8.5 (
9 points). For the Pediatric Appendicitis
Score, likelihood ratios were 0.13 (<4 points), 0.70 (4 to 7 points), and 8.1 (
8 points).

Conclusion: For children with a pretest probability of acute appendicitis of 60% or less, an Alvarado score below
4 rules out the diagnosis; this is also true for a score less than 5 if the pretest probability is up to approximately 40%. In
adults with a pretest probability greater than or equal to 60%, an Alvarado score of 8 or higher rules in the diagnosis,
whereas one of 9 or higher rules in the diagnosis at pretest probabilities greater than or equal to 40%. The Pediatric
Appendicitis Score did not identify clinically useful low- or high-risk groups at typical pretest probabilities. [Ann Emerg
Med. 2014;64:365-372.]

Please see page 366 for the Editor’s Capsule Summary of this article.

A feedback survey is available with each research article published on the Web at www.annemergmed.com.
A podcast for this article is available at www.annemergmed.com.
0196-0644/$-see front matter
Copyright © 2014 by the American College of Emergency Physicians.
http://dx.doi.org/10.1016/j.annemergmed.2014.02.025

SEE EDITORIAL, P. 373.
INTRODUCTION
Background
Clinical decision rules integrate several findings from the
medical history, physical examination, and simple laboratory tests
to predict the likelihood of a disease. Several clinical decision rules
have been developed and prospectively evaluated for their accuracy
in diagnosis of appendicitis in both adults and children. Among
the most widely studied are the Alvarado score and the Pediatric
Appendicitis Score.1,2 The Alvarado score can be used in adults
and children, whereas the Pediatric Appendicitis Score is used only
in children and adolescents. The scores are summarized in Table 1.
Pauker and Kassirer3 proposed the threshold model of
diagnosis, which identifies test and treatment thresholds for
clinical decisionmaking. For example, according to the initial
clinical examination results, patients with acute abdominal pain
and a probability of appendicitis that is below the test threshold
may be discharged home without additional diagnostic tests,
whereas those with a high probability of disease that is above the
treatment threshold may be treated with immediate
appendectomy. Patients who have an intermediate risk of
appendicitis that is between the test and treatment thresholds
might undergo imaging or observation for further data gathering.
A truly useful clinical decision rule would classify patients into
low-, moderate-, and high-risk groups that correspond to the
zones below the test threshold, between the test and treatment
thresholds, and above the treatment threshold, respectively.4 On
the other hand, if a clinical decision rule for appendicitis creates
low- and high-risk groups, but the low-risk group is not low risk
enough to rule out appendicitis and the high-risk group is not
high risk enough to rule it in, then the clinical decision rule does
not have good clinical relevance.

Volume 64, no. 4 : October 2014 Annals of Emergency Medicine 365

Downloaded for Mahasiswa 1 FK UNPAD (mhs.clinicalkey1@fk.unpad.ac.id) at Universitas Padjadjaran from ClinicalKey.com by Elsevier on October 04, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
Most Clinically Useful Cutoffs for Appendicitis Scores Ebell & Shinholser

Editor’s Capsule Summary MATERIALS AND METHODS

Study Design
What is already known on this topic We searched PubMed, using the following initial search
It is unclear whether appendicitis scoring instruments strategy: (Clinical Prediction Guides/Broad[filter]) AND
are of any value. (appendicitis) AND (sensitivity[tiab] OR specificity[tiab] OR
receiver[tiab] OR “likelihood ratio”[tiab] OR “predictive
What question this study addressed value”[tiab]) OR “alvarado score”[tiab].
This systematic review compared the 2 published and We also searched the reference lists of previous systematic
validated appendicitis scores and their ability to rule reviews to identify any additional studies not identified by our
in or rule out appendicitis at various pretest clinical initial search.
suspicion levels.
What this study adds to our knowledge Selection of Participants
We included only original research studies that gathered
In contrast to work recently published in this journal, data prospectively for a series of adults or children with acute
this study suggests that scoring instruments may have abdominal pain or clinically suspected appendicitis and also
a role in specific situations. For example, in children, reported sufficient data to calculate the sensitivity and specificity
if the clinical suspicion for appendicitis is less than or likelihood ratio for a clinical decision rule. We excluded the
60% and the Alvarado score is less than 4, derivation study used to create each clinical decision rule, studies
observation without imaging may be preferred. In gathering data retrospectively or using chart review, studies that
adults, if the clinical suspicion for appendicitis is failed to report how they followed up patients who did not
greater than 50% and the Alvarado score is greater undergo surgery, studies of mixed populations of adults and
than 9, surgery without imaging may be best. children, and any studies using a case-control design. We also
excluded studies that calculated any modification of the Alvarado
How this is relevant to clinical practice score. We included studies performed in any country and
These systematic review data suggest optimal ways in published in any language as long as they met our inclusion
which the Alvarado score might be applied. criteria described above.
These inclusion criteria ensured selection of high-quality
diagnostic test studies that used a prospective cohort design and
avoided verification and spectrum bias. Because we included only
studies that enrolled patients and gathered data prospectively,
Goals of This Investigation the persons performing the index test were blinded to the final
In this study, we will perform a diagnostic meta-analysis of diagnosis. To assess quality, we used criteria recommended by
high-quality studies of the Alvarado score and the Pediatric the Cochrane Handbook for Diagnostic Test Accuracy Reviews
Appendicitis Score and evaluate the clinical utility of the full range (http://srdta.cochrane.org/handbook-dta-reviews, chapter 9,
of calculable pairs of cutoffs to define low-, moderate-, and high- Table 9.1) and adapted for this topic.
risk groups. We will also determine the accuracy of decision
thresholds that were not explicitly reported by the original studies.
Data Collection and Processing
Each abstract was initially reviewed by both of the authors
Table 1. The Alvarado score and the Pediatric Appendicitis to determine whether it met our inclusion criteria, and
Score.1,2 differences were resolved by discussion until consensus was
Clinical Variable Alvarado Score PAS achieved. We then abstracted study design characteristics in
Migration of pain 1 1 parallel and again met to resolve any discrepancies. Finally, data
Anorexia 1 1 about the accuracy of each clinical decision rule were abstracted
Nausea or vomiting 1 1 in parallel and discrepancies resolved by consensus discussion.
Right lower quadrant tenderness 2 2 Studies enrolling only persons aged 14 years and older were
Rebound pain 1
Elevated temperature* 1 1
classified as involving adults, and those enrolling only persons
Leukocytosis (
10,000/mL) 2 1 aged 18 years and younger were classified as involving children.
Shift of WBC count to the left 1 1 We analyzed studies enrolling both adults and children separately
(
75% polymorphonucleocytes) from those enrolling only children or only adults. For studies
Cough/percussion/hopping cause pain in the RLQ 2 published in languages other than English, we relied on assistance
Total 10 10
from native speakers or used Google Translate to translate
PAS, Pediatric Appendicitis Score; WBC, white blood count; RLQ, right lower quadrant. key parts.
*Fever generally defined as greater than or equal to 37.3 C (91.2 F) for the Alvarado
score and greater than or equal to 37.3 C (99.2 F) or 38.0 C (100.4 F) for PAS. In addition to the risk categories originally described by the
authors, where possible we also abstracted data for other risk

366 Annals of Emergency Medicine Volume 64, no. 4 : October 2014

Downloaded for Mahasiswa 1 FK UNPAD (mhs.clinicalkey1@fk.unpad.ac.id) at Universitas Padjadjaran from ClinicalKey.com by Elsevier on October 04, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
 Volume 64, no. 4 : October 2014

Ebell & Shinholser

Table 2. Characteristics of included studies.
Percentage
Percentage Operated
Downloaded for Mahasiswa 1 FK UNPAD (mhs.clinicalkey1@fk.unpad.ac.id) at Universitas Padjadjaran from ClinicalKey.com by Elsevier on October 04, 2018.

With With
Reference Mean Age (SD), Appendicitis Appendicitis
Study Population Standard n Range, Years Male (%) (0%-100%) (0%-100%) Country
Adults
Al Qahtani, 200411 Consecutive patients with suspected Surgeryþclinical f/u by telephone 2–3 211 32, 13–72 59.2 56.9 87.7 Saudi Arabia
appendicitis days after d/c
Baidya, 200712 Clinically suspected appendicitis Surgeryþclinical f/u at hospital d/c 231 26.3, 16–72 61 51.5 93.7 India
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.

Kim, 200813 Clinically suspected appendicitis Surgeryþclinical f/u at 3 mo 157 37.1 (16.5), 15–84 40.1 57.3 96.8 Korea
Limpawattanasiri, 201114 Clinically suspected appendicitis Surgeryþinpatient f/u at 24þ h 1,000 NR, 15–72 40.7 71.5 85.3 Thailand
Pouget-Baudry, 201015 RLQ abdominal pain Surgeryþfollow-up clinical f/u at 1 wk 233 31.5, 15–88 48.1 73.4 98.3 France
Pruekprasert, 200416 Clinically suspected appendicitis Surgeryþfinal f/u at hospital d/c 231 27, 14–75 58 80.5 92.5 Thailand
Sanabria, 200717 Patients >14 y with pain in the right Surgeryþfollow-up telephone 374 29.5 (10.8), 15–71 47.6 55.2 76.60 Columbia
lower quadrant call at 30 days
Memon, 200918 Clinically suspected appendicitis Surgical findings 100 24.8 (9), 13–55 65 91 91 Pakistan
undergoing surgery
Inci, 201119 Clinically suspected appendicitis Surgical findings 66 26.5 (11.3), 14–72 52.9 86.4 86.4 Turkey
undergoing surgery
Canavosso, 200820 Clinically suspected appendicitis Surgical findings 207 26.6, 13–82 52.2 91.3 91.3 Argentina
undergoing surgery
Denizbasi, 200321 Clinically suspected appendicitis Surgical findings 221 26.6, 14þ 53.6 79.2 79.2 Turkey
undergoing surgery
Kang, 198940 Clinically suspected appendicitis Surgeryþclinical follow-up 62 45.8, 18–78 66.1 67.7 85.7 China
Sigdel, 201022 Clinically suspected appendicitis Surgical findings 100 27.5 (9.8), 15–68 72 94 94 Nepal
undergoing surgery
Children
Shreef, 201023 Suspected acute appendicitis Surgeryþ24-h f/u or telephone call 350 9.3, 8–14 56.1 37.7 72.5 Egypt,
Saudi Arabia

Most Clinically Useful Cutoffs for Appendicitis Scores

Bond, 199024 Consecutive patients with Surgeryþclinical f/u at 2 wk 189 NR, 2–17 NR 61.4 94.3 US
abdominal pain <1 wk
Escribá, 201125 Suspected appendicitis who Surgeryþtelephone call 10 days 99 11.2 (3.7), 4–17.8 62.6 42.4 95.5 Spain
had blood drawn in ED after discharge from ED
Mandeville, 201126 Verbal children 4 to 17 y with Surgeryþf/u telephone call 2 wk later 287 9.8 (3.1), 4–16 52.6 54 NR USA
clinical suspicion of appendicitis
Borges, 200327 Clinically suspected appendicitis Surgeryþclinical at 1 wk 76 NR, <18 52.6 71.1 96.4 Brazil
(visit or telephone)
Annals of Emergency Medicine 367

Schneider, 200728 Clinically suspected appendicitis Surgeryþf/u telephone call 2 wk later 588 11.9, 3–21 NR 33.5 91.7 USA
Bhatt, 200929 Abdominal pain <3 days and Surgeryþf/u telephone call 1 mo later 246 10.9 (3.4), 4–18 59.8 33.7 87.4 Canada
appendicitis suspected
Goldman, 200830 Abdominal pain <7 days Surgeryþtelephone call 5–7 days later 849 NR, 1–17 NR 14.5 NR Canada
Zuniga, 201231 Clinically suspected Surgeryþclinical f/u 7 days later 101 9.5 (2.8) 54.5 27.7 93.3 Spain
appendicitis <7 days
Wu, 201238 Clinically suspected appendicitis Surgeryþtelephone call 2 wk after 1,395 11.1 (4.2), 3–18 46.2 63.2 NR Taiwan
discharge from ED
Wu, 201239 Clinically suspected appendicitis Surgeryþtelephone call 2 wk after 594 11.1 (4.2), 4–18 60.3 51.5 74.8 Taiwan
discharge from ED
f/u, follow-up; d/c, discharge; h, hours; wk, week; mo, month.
Most Clinically Useful Cutoffs for Appendicitis Scores
categories. Many studies reported the number of patients with
and without a final diagnosis of appendicitis for each value of the
Alvarado score or Pediatric Appendicitis Score. Even if the
original study reported only the accuracy for a single
dichotomous cutoff or pair of cutoffs, this information allowed us
to abstract data for a series of dichotomous cutoffs and for
commonly reported pairs of cutoffs such as less than 4, 4 to 6,
and greater than or equal to 7 points, and less than 5, 5 to 8, and
greater than or equal to 8 points for the Alvarado score.
Data were recorded in parallel in a Google Docs spreadsheet.
After discrepancies were reconciled as described above, the final
data set was imported into Stata (version 12.1; StataCorp,
College Station, TX) for analysis.
Primary Data Analysis
We used the MIDAS library (Ben Dwamena, 2007. MIDAS:
A program for Meta-analytical Integration of Diagnostic
Accuracy Studies in Stata. Division of Nuclear Medicine,
Department of Radiology, University of Michigan Medical
School, Ann Arbor, Michigan) to calculate summary likelihood
ratios for the final diagnosis of appendicitis for patients who
scored above and below each cutoff. MIDAS uses a bivariate
mixed-effects regression model5,6 that has been modified for
meta-analysis of diagnostic accuracy studies.7
We took the approach of Ohle et al6 to calculate likelihood
ratios for low-, moderate-, and high-risk groups. For example, if
low-, moderate-, and high-risk groups were defined as less than
4, 4 to 6, and greater than or equal to 7 points, respectively, we
calculated the likelihood ratios for less than 4 versus greater than
or equal to 4, less than 7 versus greater than or equal to 7, and 4
to 6 versus all other scores. In some cases, studies contributed
data for the upper or lower cutoff, but not both. We also report
95% confidence intervals (CIs) for each summary likelihood
ratio. Bivariate summary receiver operating characteristic curves
were created for the key lower and upper clinical cutoff points
with the MIDAS procedure where possible. Because they use a
bivariate random-effects regression model, they can be used to
provide a valid estimate of the summary likelihood ratios.
RESULTS
Our initial PubMed search yielded 526 studies, and a search
for “‘pediatric appendicitis score’[tiab]” yielded 14 studies. The
reference lists for previously published systematic reviews8-10
were reviewed and identified 10 additional studies not found by
the initial PubMed searches. The final total was 544 unique
original research studies, of which a total of 29 met our inclusion
criteria. The remaining studies generally did not study accuracy
or did not report sufficient data to calculate accuracy, were
retrospective, or used a case-control design.
The characteristics of included studies are summarized in
Table 2. Of the 29 studies, 13 reported on use of the Alvarado
score in adults11-22,40 and 11 on use of the Alvarado score or
Pediatric Appendicitis Score in children.23-31,38,39 Five additional
studies reported data for a mixed population of adults and
children32-36 and were excluded from the analysis.
368 Annals of Emergency Medicine
Downloaded for Mahasiswa 1 FK UNPAD (mhs.clinicalkey1@fk.unpad.ac.id) at Universitas Padjadjaran from ClinicalKey.com by Elsevier on October 04, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
Ebell & Shinholser
Excluding studies that enrolled only patients who underwent
surgery, the pooled percentage of patients with appendicitis was
66.0% in the adult studies (range 51.5% to 80.5%), 38.8% in
the mixed-population studies (range 35.0% to 58.3%), and
33.4% in the studies of children (range 14.5% to 71.1%). The
mean age ranged from 25 to 37 years in studies of adults, 23 to
30 years in studies with a mixed population, and 9.3 to
11.9 years in studies of children.
Because we used methodologically rigorous inclusion criteria,
all included studies were prospective cohort studies of unselected
patients with abdominal pain or suspected appendicitis, had
adequate follow-up, and reported sufficient data to calculate
Alvarado or Pediatric Appendicitis Score scores. None of the
studies blinded outcome assessors, and both scores were always
calculated before surgery or on admission. Five studies enrolled
only patients undergoing surgery,18-22 whereas the remainder
were all judged to have adequate clinical follow-up in
nonoperated patients to avoid verification bias. Most studies
specified that the decision to perform surgery was independent of
the clinical decision rule(s). In 3 studies, the decision to operate
was guided by the score result,12,14,20 and for 5 studies it was not
clear whether the surgery decision was independent.11,19,24,30,31
A summary of the quality assessment is shown in Figure 1.
The accuracy of the Alvarado score in adults for the most
commonly used sets of decision thresholds is summarized in
Table 3. The likelihood ratio for the low-risk groups in adults
ranged from 0.01 to 0.38 and was less than 0.05 for 6 of 10 study
populations. Two studies had significantly higher likelihood
ratios for the low-risk group.13,15 One of these 2 studies found
appendicitis in 23 of 55 patients with a score less than 4 points,
but the score was calculated by an intern on call, so lack of
clinical experience may have contributed to a failure to detect
clinical signs and symptoms of appendicitis.15 The other was a
Korean study that found appendicitis in 11 of 34 patients with a
score less than 4 points; however, we could find no explanation
for the higher rate of appendicitis compared with that in other
studies.13 In regard to the high-risk groups in adults, the
likelihood ratio ranged from 1.76 to 13.7. The summary
likelihood ratio was 3.4 (95% CI 2.5 to 4.6) for a cutoff of
greater than or equal to 7 points and 6.7 (95% CI 3.5 to 12.7)
for a cutoff of greater than or equal to 9 points. Bivariate
summary receiver operating characteristic curves for cutoffs of
less than 4, greater than or equal to 7, and greater than or equal
to 9 in adults are shown in Figure E1A through C (available
online at http://www.annemergmed.com); receiver operating
characteristic curves for cutoffs of less than 4, less than 5, greater
than or equal to 7, and greater than or equal to 9 in children are
shown in Figure E2A through D (available online at http://www.
annemergmed.com).
The accuracy of the Alvarado score and Pediatric Appendicitis
Score in children is summarized in Table 4. The likelihood of
appendicitis was significantly higher in the low-risk group for the
Pediatric Appendicitis Score (0.13; 95% CI 0.04 to 0.4) than for
the Alvarado score (0.02 for a cutoff of <4 points; 0.04 for a
cutoff of <5 points). On the other hand, the high-risk group for
Volume 64, no. 4 : October 2014
Ebell & Shinholser Most Clinically Useful Cutoffs for Appendicitis Scores

Figure 1. Assessment of study quality. Green indicates low risk of bias, yellow unclear risk of bias, and red high risk of bias for each
aspect of study quality.

the Pediatric Appendicitis Score had a higher likelihood ratio et al30 enrolled all children with abdominal pain, not just
than for the Alvarado scores, but with considerable heterogeneity. children with right lower quadrant pain or suspected
Two studies were outliers, with higher likelihood ratios for the appendicitis. We could find no reason why the other study26
low-risk group than other studies.26,30 The study by Goldman might have been an outlier. Seven studies reported data sufficient

Table 3. Performance of Alvarado score in adults.

Likelihood Ratio (95% CI)
Study n Reference Standard Low Risk Moderate Risk High Risk
<4 4–6
7
Denizbasi, 200321 221 Surgery 1.76 (1.34–2.29)
Canavosso, 200820 207 Surgery 0.10 (0–4.90) 0.23 (0.13–0.40) 1.96 (1.17–3.30)
Sigdel, 201022 100 Surgery 2.46 (0.79–7.65)
Memon, 200918 100 Surgery 0.04 (0–0.83) 0.54 (0.35–0.82) 5.24 (0.82–33.5)
Inci, 201119 66 Surgery 2.53 (1.00–6.41)
Al Qahtani, 200411 211 Surgeryþf/u 0.01 (0–0.22) 0.09 (0.04–0.20) 6.04 (3.73–9.79)
Pruekprasert, 200416 231 Surgeryþf/u 2.65 (1.63–3.95)
Sanabria, 200717 374 Surgeryþf/u 0.11 (0.04–0.29) 0.45 (0.33–0.60) 2.40 (1.90–3.04)
Baidya, 200712 231 Surgeryþf/u 6.64 (4.05–10.9)
Pouget-Baudry, 201015 233 Surgeryþf/u 0.26 (0.17–0.41) 1.07 (0.75–1.54) 4.65 (2.14–10.1)
Limpawattanasiri, 201114 1000 Surgeryþf/u 3.41 (2.79–4.17)
Kang, 198940 62 Surgeryþf/u 3.57 (1.46–8.76)
Summary (all studies) 3,568 0.03 (0.01–0.33) 0.42 (0.20–0.89) 3.43 (2.53–4.65)
I2¼71 I2¼84 I2¼61
Summary (excluding surgery only studies)* 4.29 (2.98–6.17)
I2¼59
<5 5–8
9
Canavosso, 200820 207 Surgery 0.01 (0–0.27) 0.78 (0.59–1.02) 7.14 (1.05–48.4)
Memon, 200918 100 Surgery 0.01 (0–0.21) 1.31 (0.72–2.37) 5.54 (0.36–84.3)
Al Qahtani, 200411 211 Surgeryþf/u 0.01 (0–0.13) 0.84 (0.61–1.15) 13.7 (5.2–36.0)
Sanabria, 200717 374 Surgeryþf/u 0.12 (0.06–0.25) 1.00 (0.85–1.18) 4.96 (2.79–8.82)
Kim, 200813 157 Surgeryþf/u 0.38 (0.20–0.72)
Limpawattanasiri, 201114 1,000 Surgeryþf/u 0.01 (0–0.21)
Summary 2,049 0.01 (0–0.41) 0.98 (0.84–1.13) 6.69 (3.51–12.7)
I2¼94 I2¼29 I2¼0
*Meta-analysis of likelihood ratios for low- and moderate-risk groups is not possible because of the small number of studies.

Volume 64, no. 4 : October 2014 Annals of Emergency Medicine 369

Downloaded for Mahasiswa 1 FK UNPAD (mhs.clinicalkey1@fk.unpad.ac.id) at Universitas Padjadjaran from ClinicalKey.com by Elsevier on October 04, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
Most Clinically Useful Cutoffs for Appendicitis Scores Ebell & Shinholser

Table 4. Performance of the Alvarado score and Pediatric Appendicitis Score in children.*
Likelihood Ratio (95% CI)
Study n Low Risk Moderate Risk High Risk
Alvarado score <4 4–6
7
Bond, 199024 189 0.01 (0–0.24) 0.24 (0.13–0.44) 3.12 (2.16–4.50)
Schneider, 200728 588 0.06 (0.02–0.18) 0.49 (0.38–0.63) 3.76 (3.01–4.69)
Shreef, 201023 350 0.03 (0–0.47) 0.19 (0.13–0.30) 5.09 (3.76–6.88)
Escriba, 201125 99 0.03 (0–0.53) 0.17 (0.06–0.44) 10.3 (4.44–24.0)
Wu, 201239 594 4.38 (3.33–5.76)
Mandeville, 201126 287 0.38 (0.21–0.70) 0.31 (0.21–0.46) 2.72 (2.04–3.62)
Summary 1,513 0.02 (0–0.36) 0.27 (0.19–0.40) 4.21 (3.33–5.32)
I2¼83 I2¼67 I2¼49
Alvarado score <5 5–8
9
Bond, 199024 189 0.01 (0–0.19) 1.03 (0.80–1.33) 6.17 (3.58–14.7)
Borges, 200327 76 0.12 (0.04–0.31)
Schneider, 200728 588 0.18 (0.11–0.29) 1.25 (1.09–1.43) 6.75 (3.89–11.7)
Shreef, 201023 350 0.01 (0–0.19) 1.15 (1.01–1.32) 10.6 (4.22–26.5)
Escriba, 201125 99 0.02 (0–0.33) 1.25 (0.84–1.87) 52.6 (3.27–847)
Mandeville, 201126 287 0.30 (0.19–0.48) 1.07 (0.87–1.32) 6.67 (2.95–15.1)
Summary 1,589 0.04 (0–0.36) 1.16 (1.06–1.27) 8.47 (5.61–12.8)
I2¼80 I2¼0 I2¼0
PAS <4 4–7
8
Schneider, 200728 588 0.16 (0.08–0.32) 0.71 (0.60–0.84) 5.04 (3.63–7.00)
Goldman, 200830 849 0.67 (0.58–0.77) 4.20 (3.05–5.80) 1.69 (0.35–8.02)
Bhatt, 200929 246 0.04 (0.01–0.26) 0.70 (0.53–0.92) 11.3 (5.59–22.8)
Escriba, 201125 99 0.03 (0–0.48) 0.50 (0.31–0.83) 79.6 (5.00–1266)
Mandeville, 201126 287 0.43 (0.21–0.88) 0.42 (0.31–0.55) 3.37 (2.35–4.85)
Zuniga, 201231 101 0.09 (0.01–1.43) 0.58 (0.37–0.91) 8.34 (3.38–20.6)
Summary 2,170 0.13 (0.04–0.40) 0.70 (0.45–1.11) 8.10 (4.13–15.9)
I2¼96 I2¼95 I2¼91
*All studies used surgeryþclinical follow-up as the reference standard.

to calculate a simple dichotomous cutoff of 7 or more points for
the Pediatric Appendicitis Score.25,26,28-31,38 This yielded a
positive likelihood ratio of 5.2 (95% CI 3.1 to 8.6) and negative
likelihood ratio of 0.38 (95% CI 0.20 to 0.74).
LIMITATIONS
A limitation of the current study is threats to validity in the
included studies. There were also differences in the level of
training of the physicians gathering the clinical data. However,
we limited our analysis to high-quality, prospective studies that
avoided verification bias. Another limitation is that pretest
probabilities are not well known to most physicians in their
practice setting and that the test and treatment thresholds are
generally determined intuitively by each physician rather than by
a formal assessment of the benefits or harms of testing, treating,
or neither. Finally, CIs were fairly large for some estimates,
notably, the high-risk group for a cutoff of 9 for the Alvarado
score in adults and most of the low-risk groups. The latter was
due primarily to 2 studies that were outliers.26,30
DISCUSSION
Previous systematic reviews on this topic have had limitations.
For example, some analyzed results only for either a single
cutoff,5 a single pair of cutoffs (ie, <5, 5 to 6, or
7 points),6 or
combined results from studies using different cutoffs.7 Two
previous systematic reviews included studies that gathered data
retrospectively or had incomplete follow-up of nonoperated
patients, which might lead to partial verification bias.5,6 Finally,
we identified high-quality, primary studies that met our inclusion
criteria but were not identified by previous studies because of
language exclusion criteria or publication date.5-7
In the current study, we limited our analysis to studies that
avoided these important biases: they used prospective data
collection, performed follow-up to determine the outcome for
nonoperated patients, and did not use a case-control design. We
also examined the accuracy for cutoffs not reported by the
original study to identify all high-quality studies that examined a
particular cutoff for the test or treatment threshold.
We were interested in identifying optimal cutoffs for the test
and treatment thresholds because this reflects modern clinical
decisionmaking about appendicitis. Patients with a high
probability of appendicitis are generally taken directly to the
operating theater, those with a very low probability are either
observed or discharged, and those with an intermediate
probability generally undergo imaging. The lower decision
threshold corresponds to the “test threshold,” whereas the upper
cutoff corresponds to the “treatment threshold.”
Figure 2 summarizes the clinical implications of our study for
the Alvarado score. We used likelihood ratios for test and

370 Annals of Emergency Medicine Volume 64, no. 4 : October 2014

Downloaded for Mahasiswa 1 FK UNPAD (mhs.clinicalkey1@fk.unpad.ac.id) at Universitas Padjadjaran from ClinicalKey.com by Elsevier on October 04, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
Ebell & Shinholser
Figure 2. Clinical application of optimal test and treatment
thresholds for the Alvarado score in adults and children and the
Pediatric Appendicitis Score in children. Green indicates
probability of appendicitis below 3% and red a probability of
85% or higher. Bold columns indicates typical pretest
probability of a final diagnosis of appendicitis for children (33%)
and adults (66%) presenting with clinically suspected
appendicitis from included studies. LR, Likelihood ratio.
treatment thresholds calculated in Tables 3 and 4 and calculated
posttest probabilities of appendicitis over a plausible range of
pretest probabilities. Ohmann et al37 proposed that clinical rules
not miss appendicitis in more than 5% of patients, which
corresponds to the test threshold to define a low-risk group. They
also proposed that the negative appendectomy rate be no more
than 15%, which corresponds to a treatment threshold of
85%. We argue that a lower missed appendicitis rate of perhaps
3% is more appropriate, particularly in children.
In children, when the pretest probability of acute appendicitis is
60% or less (true for 8 of 11 studies in our analysis), an Alvarado
score below 4 is associated with a probability of appendicitis of less
than 3%. For a range of pretest probabilities greater than 40%,
typical for studies of consecutive adults presenting with clinically
suspected appendicitis, a score of 9 is associated with an 85% or
higher probability of appendicitis, a point at which most surgeons
would be comfortable ruling in the diagnosis. A score of 7 or higher
rules in appendicitis at pretest probabilities of at least 60%.
However, at these pretest probabilities, the Alvarado score is not
useful for ruling in appendicitis in children or ruling it out in adults.
Nevertheless, thoughtful application of the Alvarado score has
the potential to significantly reduce use of computed tomography
and other imaging in children who have a low likelihood of
appendicitis. The positive likelihood ratio of the Pediatric
Appendicitis Score for ruling in appendicitis when above 8 points
was similar to that of the Alvarado score, and the negative likelihood
ratio was higher, making it worse at ruling out appendicitis. Thus,
it has less clinical value than the Alvarado score, according to our
analysis (Figure 2).
Kharbanda et al41 recently published a multicenter validation
of the Low Risk Appendicitis Score. Their original rule was
95.5% sensitive and 36.3% specific, with a negative likelihood
ratio of 0.12. They changed the rule post hoc and found that
Volume 64, no. 4 : October 2014
Downloaded for Mahasiswa 1 FK UNPAD (mhs.clinicalkey1@fk.unpad.ac.id) at Universitas Padjadjaran from ClinicalKey.com by Elsevier on October 04, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
Most Clinically Useful Cutoffs for Appendicitis Scores
they could achieve somewhat better sensitivity (98.1%) and
negative likelihood ratio (0.08) with the new rule. Of course, this
new rule would require prospective validation before it is applied.
Clearly, knowledge of the pretest probability is a critical
element to intelligently making use of the Alvarado score and
similar clinical decision rules. However, it is challenging for the
typical clinician in most settings to know what this pretest
probability is for common conditions in their practice or
emergency department (ED). It is hoped that increased use of
electronic health records will allow emergency physicians and
surgeons to have some idea of the percentage of children and
adults who present with acute abdominal pain and ultimately
receive a diagnosis of appendicitis.
It is also important to identify test and treatment thresholds to
guide clinical practice and the development and validation of
clinical decision rules. However, few data are available in regard
to these thresholds. There are currently several approaches to
determining thresholds (eg, cost-utility analysis, decision analysis,
use of clinical vignettes), and more work is needed to determine
the best approach.
Finally, developers of clinical decision rules for the diagnosis
of appendicitis and other conditions should explicitly consider
the typical range of pretest probabilities in different settings, as
well as the test and treatment thresholds, when deriving and
validating clinical decision rules. We argue that the most useful
rules identify patients whose scores are below the test threshold,
above the treatment threshold, or both.4
In summary, in children with a pretest probability of acute
appendicitis of 60% or less, an Alvarado score below 4 points
rules out the diagnosis satisfactorily. When the pretest probability
is greater than or equal to 60% in adult populations, an Alvarado
score of 7 or higher rules in the diagnosis, whereas a score of 9 or
higher rules in the diagnosis at pretest probabilities greater than
or equal to 40%.
Supervising editor: Kathy N. Shaw, MD, MSCE
Author affiliations: From the Department of Epidemiology and
Biostatistics, College of Public Health, University of Georgia,
Athens, GA.
Author contributions: MHE conceived and designed the study,
supervised the conduct of the systematic review and data
collection, performed the literature searches, conducted the
analysis, and wrote the article. MHE and JS performed selection of
studies and data abstraction. JS approved the article. MHE takes
responsibility for the paper as a whole.
Funding and support: By Annals policy, all authors are required to
disclose any and all commercial, financial, and other relationships
in any way related to the subject of this article as per ICMJE conflict
of interest guidelines (see www.icmje.org). The authors have stated
that no such relationships exist.
Publication dates: Received for publication July 14, 2013.
Revision received January 25, 2014. Accepted for publication
February 28, 2014. Available online April 14, 2014.
Annals of Emergency Medicine 371
Most Clinically Useful Cutoffs for Appendicitis Scores
REFERENCES
1. Alvarado A. A practical score for the early diagnosis of acute
appendicitis. Ann Emerg Med. 1986;15:557-564.
2. Samuel M. Pediatric appendicitis score. J Pediatr Surg.
2002;37:877-881.
3. Pauker SG, Kassirer JP. The threshold approach to clinical decision-
making. N Engl J Med. 1980;302:1109-1117.
4. Ebell MH. AHRQ white paper: use of clinical decision rules for point-of-
care decision support. Med Decis Mak. 2010;30:712-721.
5. van Houwelingen HC, Arends LR, Stijnen T. Advanced methods in
meta-analysis: multivariate approach and meta-regression. Stat Med.
2002;21:589-624.
6. van Houwelingen HC, Zwinderman KH, Stijnen T. A bivariate approach
to meta-analysis. Stat Med. 1993;12:22730-22784.
7. Reitsma JB, Glas AS, Rutjes AWS, et al. Bivariate analysis of sensitivity
and specificity produces informative summary measures in diagnostic
reviews. J Clin Epidemiol. 2005;58:982-990.
8. Kulik DM. Does this child have appendicitis? a systematic review of
clinical prediction rules for children with acute abdominal pain. J Clin
Epidemiol. 2013;66:95-104.
9. Ohle R, O’Reilly F, O’Brien KK, et al. The Alvarado score for predicting
acute appendicitis: a systematic review. BMC Med. 2011;9:139.
10. Liu JLY, Wyatt JC, Deeks JJ, et al. Systematic reviews of clinical decision
tools for acute abdominal pain. Health Technol Assess.
2006;10:1-167.
11. Al Qahtani HH, Muhammad AA. Alvarado score as an admission
criterion for suspected appendicitis in adults. Saudi J Gastroenterol.
2004;10:8691.
12. Baidya N, Rodrigues G, Rao A, et al. Evaluation of Alvarado score in
acute appendicitis: a prospective study. Internet J Surg. 2007;9:1.
13. Kim K, Rhee JE, Lee CC, et al. Impact of helical computed tomography
in clinically evident appendicitis. Emerg Med J. 2008;25:477-481.
14. Limpawattanasiri C. Alvarado score for the acute appendicitis in a
provincial hospital. J Med Assoc Thai. 2011;94:441-449.
15. Pouget-Baudry Y, Mucci S, Eyssartier E, et al. The use of the Alvarado
score in the management of right lower quadrant abdominal pain in
the adult. J Visc Surg. 2010;147:e40-e44.
16. Pruekprasert P, Maipang T, Geater A, et al. Accuracy in diagnosis of
acute appendicitis by comparing serum C-reactive protein
measurements, Alvarado score and clinical impression of surgeons.
J Med Assoc Thai. 2004;87:296-303.
17. Sanabria A, Domínguez LC, Bermúdez C, et al. [Evaluation of diagnostic
scales for appendicitis in patients with lower abdominal pain].
Biomedica. 2007;27:419-428.
18. Memon AA, Vohra LM, Khaliq T, et al. Diagnostic accuracy of Alvarado
score in the diagnosis of acute appendicitis. Pak J Med Sci.
2009;25:118-121.
19. Inci E, Hocaoglu E, Aydin SP, et al. Efficiency of unenhanced MRI in
the diagnosis of acute appendicitis: comparison with Alvarado
scoring system and histopathological results. Eur J Radiol. 2011;80:
253-258.
20. Canavosso L, Carena P, Carbonell JM, et al. [Right iliac fossa pain and
Alvarado score]. Cir Esp. 2008;83:247-251.
21. Denizbasi A, Unluer EE. The role of the emergency medicine resident
using the Alvarado score in the diagnosis of acute appendicitis
compared with the general surgery resident. Eur J Emerg Med.
2003;10:296-301.
Did you know?
Annals accepts audio and video files as ancillaries to the main article.
Visit http://www.annemergmed.com/content/instauth/ for more details!
372 Annals of Emergency Medicine
Downloaded for Mahasiswa 1 FK UNPAD (mhs.clinicalkey1@fk.unpad.ac.id) at Universitas Padjadjaran from ClinicalKey.com by Elsevier on October 04, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
Ebell & Shinholser
22. Sigdel GS, Lakhey PJ, Mishra PR. Tzanakis score vs. Alvarado score in
acute appendicitis. JNMA J Nepal Med Assoc. 2010;49:96-99.
23. Shreef KS, Waly AH, Abd-Elrahman S, et al. Alvarado score as an
admission criterion in children with pain in right iliac fossa. Afr J
Paediatr Surg. 2010;7:163-165.
24. Bond GR, Tully SB, Chan LS, et al. Use of the MANTRELS score in
childhood appendicitis: a prospective study of 187 children with
abdominal pain. Ann Emerg Med. 1990;19:1014-1018.
25. Escriba A, Gamell AM, Fernandez Y, et al. Prospective validation of two
systems of classification for the diagnosis of acute appendicitis.
Pediatr Emerg Care. 2011;27:165-169.
26. Mandeville K, Pottker T, Bulloch B, et al. Using appendicitis scores in
the pediatric ED. Am J Emerg Med. 2011;29:972-977.
27. Borges PS, Lima M, Neto GH. The Alvarado score validation in
diagnosing acute appendicitis in children and teenagers at the
Instituto Materno Infantil de Pernambuco, IMIP. Res Bras Saude
Matern Infant. 2003;3:439-445.
28. Schneider C, Kharbanda A, Bachur R. Evaluating appendicitis scoring
systems using a prospective pediatric cohort. Ann Emerg Med.
2007;49:778-784, 784.e1.
29. Bhatt M, Joseph L, Ducharme FM, et al. Prospective validation of the
Pediatric Appendicitis Score in a Canadian pediatric emergency
department. Acad Emerg Med. 2009;16:591-596.
30. Goldman RD, Carter S, Stephens D, et al. Prospective validation of the
Pediatric Appendicitis Score. J Pediatr. 2008;153:278-282.
31. Zúñiga RV, Arribas JL, Montes SP, et al. Application of Pediatric
Appendicitis Score on the emergency department of a secondary level
hospital. Pediatr Emerg Care. 2012;28:489-492.
32. Andersson M, Andersson RE. The Appendicitis Inflammatory Response
Score: a tool for the diagnosis of acute appendicitis that outperforms
the Alvarado score. World J Surg. 2008;32:1843-1849.
33. de Castro SM, Unlu C, Steller EP, et al. Evaluation of the Appendicitis
Inflammatory Response Score for patients with acute appendicitis.
World J Surg. 2012;36:1540-1545.
34. Saidi RF, Ghasemi M. Role of Alvarado score in diagnosis and
treatment of suspected acute appendicitis. Am J Emerg Med.
2000;18:230-231.
35. Chan MY, Tan C, Chiu MT, et al. Alvarado score: an admission criterion
in patients with right iliac fossa pain. Surgeon. 2003;1:39-41.
36. Jang SO, Kim BS, Moon DJ. [Application of Alvarado score in
patients with suspected appendicitis]. Korean J Gastroenterol.
2008;52:27-31.
37. Ohmann C, Yang Q, Franke C. Diagnostic scores for acute appendicitis.
Eur J Surg. 1995;161:273-281.
38. Wu HP, Yang WC, Wu KH, et al. Diagnosing appendicitis at different
time points in children with right lower quadrant pain: comparison
between Pediatric Appendicitis Score and the Alvarado score. World J
Surg. 2012;36:216-221.
39. Wu HP, Chen CY, Kuo IT, et al. Diagnostic values of a single serum
biomarker at different time points compared with Alvarado score and
imaging examinations in pediatric appendicitis. J Surg Res.
2012;174:272-277.
40. Kang W-M, Lee C-H, Chou Y-H, et al. A clinical evaluation of ultrasonography
in the diagnosis of acute appendicitis. Surgery. 1989;105:154-159.
41. Kharbanda AB, Dudley NC, Bajaj L, et al. Validation and refinement of a
prediction rule to identify children at low risk for acute appendicitis.
Arch Pediatr Adolesc Med. 2012;166:738-744.
Volume 64, no. 4 : October 2014
Ebell & Shinholser Most Clinically Useful Cutoffs for Appendicitis Scores

Figure E1. Bivariate receiver operating characteristic curves for different low and high cutoffs of the Alvarado score in adults.
A, A cutoff of less than 4 points. B, A cutoff of 7 or more points. C, A cutoff of 9 or more points. Because of the small number of
studies and events within studies, the MIDAS procedure in Stata was unable to create a receiver operating characteristic curve for
the cutoff of less than 5 points.

Volume 64, no. 4 : October 2014 Annals of Emergency Medicine 372.e1

Downloaded for Mahasiswa 1 FK UNPAD (mhs.clinicalkey1@fk.unpad.ac.id) at Universitas Padjadjaran from ClinicalKey.com by Elsevier on October 04, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
Most Clinically Useful Cutoffs for Appendicitis Scores Ebell & Shinholser

Figure E2. Bivariate receiver operating characteristic curves for different low and high cutoffs of the Alvarado score in children.
A, A cutoff of less than 4 points. B, A cutoff of less than 5 points. C, A cutoff of 7 or more points. D, A cutoff of 9 or more points.

372.e2 Annals of Emergency Medicine Volume 64, no. 4 : October 2014

Downloaded for Mahasiswa 1 FK UNPAD (mhs.clinicalkey1@fk.unpad.ac.id) at Universitas Padjadjaran from ClinicalKey.com by Elsevier on October 04, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.