You are on page 1of 6

American Journal of Emergency Medicine (2011) 29, 972–977

www.elsevier.com/locate/ajem

Original Contribution

Using appendicitis scores in the pediatric ED☆
Katherine Mandeville MD a , Tamara Pottker MD a , Blake Bulloch MD a,⁎, Jiexin Liu PhD, MBA, MS b
a b

Department of Emergency Medicine, Phoenix Children's Hospital, Phoenix, AZ 85016, USA Phoenix Children’s Hospital, Phoenix, AZ 85016, USA

Received 19 April 2010; revised 26 April 2010; accepted 29 April 2010

Abstract Study Objective: The aims of the study were to prospectively evaluate the Alvarado and Samuel (pediatric appendicitis score [PAS]) appendicitis scoring systems in children and determine performance based on sex. Methods: Children with abdominal pain concerning for appendicitis were recruited. Nine parameters evaluated by the scores were documented before imaging/surgery consultation. Test characteristics were calculated on all patients and by sex. Results: Two hundred eighty-seven patients enrolled; median age was 9.8 years; and 155 (54%) were diagnosed with pathologic examination-confirmed appendicitis. Patients with appendicitis had mean PAS of 7.6, and those without had mean of 5.6 (P b .001). Patients with appendicitis had a mean Alvarado of 7.2, and those without had a mean of 5.2 (P b .001). In appendicitis patients, PAS cutoff of 6 or greater would give 137 correct diagnoses; sensitivity, 88%; specificity, 50%; and positive predictive value (PPV), 67%. An Alvarado cutoff of 7 or greater would give 118 correct diagnoses; sensitivity, 76%; specificity, 72%; and PPV, 76%. Both performed better in males than females. Conclusion: Regardless of sex, neither PAS nor Alvarado has adequate predictive values for sole use to diagnose appendicitis. © 2011 Elsevier Inc. All rights reserved.

1. Introduction
Appendicitis is one of the most common causes of acute abdominal pain in pediatrics and is the most common indication for emergency abdominal surgery in childhood [1]. For the past decade, there has been greater use of abdominal computed tomographic (CT) scans in the ED attempting to improve the diagnostic accuracy of appendicitis and to decrease the negative laparotomy rate [2]. The can use, however, results in exposure to ionizing radiation,

☆ Financial disclosure: Study received no funding support. ⁎ Corresponding author. Tel.: +1 602 546 1910. E-mail address: bbulloch@phoenixchildrens.com (B. Bulloch).

which in pediatric patients may lead to increased cancer risks [3-5]. Ultrasound imaging has become more popular and negates the risk of radiation. However, ultrasound use is operator dependent, and different institutions have varying expertise in its use and availability [6,7]. More recently, different clinical scoring systems aiding in the diagnosis of appendicitis have been developed. The 2 most used scores (Alvarado and the pediatric appendicitis score [PAS] by Samuel) are based on the presence or absence of symptoms with each feature being scored as a 1 or 2 depending on the feature and each given different levels of significance (Table 1). For each score, the author identified a cut point at which surgery was recommended vs observation. Most studies examining these scores have had varying results [8-14].

0735-6757/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2010.04.018

informed consent was obtained from parents or legal guardians. the PAS's performance has not previously been stratified by sex [14-16]. The questionnaire consisted of historical. Assessors were blinded to each other's evaluation. were nonverbal. we had 2 independent assessments. The interobserver assessments were performed within 15 minutes of each other in an attempt to eliminate changes in the patient's condition. appendicitis with recommendation of surgery The Alvarado score has been studied more extensively than the PAS with several studies assessing the Alvarado score finding statistically significant improved performance in male vs female patients. probable appendicitis with recommendation of surgery Score of 9-10. Emergency department physicians and pediatric emergency medicine fellows were educated about the data collection form before initiation of the study. level 1. except the primary investigators. a review of the pathology database. Enrollment occurred mainly between the hours of 2 PM and midnight when research staff was available that correlated with the ED peak census times. The primary objective of this study was to prospectively assess the performance of the Alvarado and PAS scoring systems in a pediatric ED. If patients were taken to the operating room and determined to have no evidence of appendicitis. pediatric ED with approxiAll data analyses were conducted with SPSS 17. Questionnaires were completed before any diagnostic imaging and before laboratory work was obtained. Ill).3°C) Leukocytes (≥10 000) Differential white blood cell count with 75% polymorphonuclear cells Point value 1 1 1 2 2 1 1 1 Score system ≤5. 3.” ≥37. If the caregiver was not able to be reached. Chicago.3°C or 99°F) Leukocytes (L) (≥10 000) Differential white blood cell count with 75% polymorphonuclear cells Score system Score of 5-6. The study was approved by the hospital's institutional review board. All followup. To determine interobserver reliability of clinical findings used by the scoring systems. Final diagnosis was determined by pathology report if the patient had surgery or by a follow-up telephone call at 2 weeks if the patient did not have surgery. To our knowledge. Both the PAS and Alvarado scores have the potential value of eliminating excessive use of abdominal CT scans and permitting more rapid surgical consultation in patients with suspected appendicitis. Pediatric emergency medicine faculty or a pediatric emergency medicine fellow providing care for the patient completed the data collection forms. laboratory. The secondary objective was to determine if these scoring systems performed better based on sex. Alvarado (MANTRELS) score Migration of pain (M) Anorexia/acetone (A) Nausea/vomiting (N) Right lower quadrant tenderness (T) Rebound pain (R) Elevation in temperature (E) (≥37. telephone calls were performed by the principle investigator (PI) to a caregiver of the patient. however.Using appendicitis scores in the pediatric ED Table 1 PAS and Alvarado scoring systems Point value 1 1 1 2 1 1 2 1 973 mately 60 000 visits per year. repeat telephone calls were made 1 week later. and review of the radiology database for all patients who underwent abdominal CT scans. Data analysis 2. and medical records were reviewed to determine the final diagnosis. not appendicitis with recommendation of observation ≥6. had undergone previous abdominal surgery. Written. Patients were excluded if they were pregnant. or had radiologic imaging of the abdomen within the previous 2 weeks. It was assumed that resolution of symptoms with no diagnosis of appendicitis 2 weeks after the initial visit would indicate that the patient did not have appendicitis. Methods This study was a prospective observational study conducted at an urban. compatible with appendicitis with recommendation of observation Score of 7-8. and physical examination variables included in the 2 scoring systems. none of the participating physicians. Enrollment was audited by review of an ED log of all patients presenting with a complaint of abdominal pain. and informed assent was obtained from children 7 years and older. Only those patients with follow-up were included in the data analysis.0 for Windows (SPSS Inc. Verbal children between 4 and 17 years of age with suspected appendicitis were enrolled from June 2008 to May 2009. whenever possible. they were coded as not having appendicitis. Bivariate analyses was used to assess the differences in patients' demographic . very probable appendicitis with recommendation of surgery Samuel (PAS) score Migration of pain Anorexia Nausea/vomiting Right lower quadrant tenderness Cough/hopping/percussion/tenderness in the right lower quadrant Elevation in temperature (“Pyrexia. were introduced to the actual scoring systems. had a chronic medical condition.

4% 29.001 .417 . interobserver reliability was tested with κ test.8% No appendicitis (n = 132) 10.3% (23%-37%) 10. 962 patients presented to the ED with a complaint of abdominal pain.5% (0%-4%) 8.8) 61.001 b.8% 36.2% 74.1% (78%-91%) 95.7% 7.2% 49.3% (5%-15%) Specificity 0.2% (90%-97%) 94.5% (89%-97%) 91.4) 42.5% (58%-74%) 43. 487 patients were diagnosed with appendicitis.6 (2.5% (89%-97%) 93.051 .6 (2.4% 6.2% (98%-100%) PPV 54% 53% 53% 55% 58% 64% 69% 76% 81% 89% 94% NPV 10% 18% 52% 69% 74% 74% 72% 64% 54% 49% 0 1 2 3 4 5 6 7 8 9 10 100% 94.6% 5.2% (90%-97%) 93.3°C or 99°F) Leukocytes (≥10 000) Differential white blood cell count with 75% polymorphonuclear cells 9. The χ2 test was used for comparison of categorical variables (eg.5) 6.3% 47.2% 75. Overall 9.4% 89.117 . and 287 patients meeting the inclusion and exclusion criteria gave given consent and enrolled.001 characteristics (such as age and sex).1% 36. and κ coefficients were calculated.7% 83.5% 53.2% 76.0% 58. and Student t test was applied for comparisons of continuous variables.1% (69%-82%) 59.6% 62.6% (76%-88%) 76.002 b. A receiver operating characteristic (ROC) curve was constructed to assess sensitivity and specificity for both PAS and Alvarado scores at diagnosing appendicitis.4) 7.9% (36%-51%) 12. Table 3 Score PAS and Alvarado score performance at each cutoff value PAS Sensitivity Specificity 0.8 (3.0 (3.001 .2 (2. Mandeville et al. Results During the study period. who had a history and/ or physical examination concerning appendicitis.4% (51%-67%) 30.002 b. All patients who gave consent had complete data for both the PAS and Alvarado scores. Subgroup analysis on ROC curves was constructed for males and females separately.4% (83%-93%) 80% (73%-86%) 66.8% (0%-3%) 1. and PAS and Alvarado scores between patients with appendicitis and patients without appendicitis.2 (2.5% 92.0% 1% (0%-3%) 2% (0%-4%) 8% (3%-13%) 15% (9%-22%) 30% (22%-38%) 50% (42%-59%) 67% (60%-76%) 80% (74%-88%) 92% (87%-97%) 98% (96%-100%) PPV 54% 53% 53% 54% 56% 60% 67% 74% 80% 86% 87% NPV 10% 18% 50% 63% 71% 79% 74% 67% 58% 49% Alvarado Sensitivity 100% 94.031 . symptoms.6% 62.2% (90%-97%) 94. sex.4% 37.1% 64. .6 (2.4% (34%-51%) 56. as originally intended by the scoring system developers. Computed scores were used to assess the accuracy of the Alvarado system with a score of 7 or greater.6% 70.1) 5.6% (87%-96%) 87.3 (2.5) 40.024 b.6% 47.9% (7%-18%) NPV indicates negative predictive value.4% K.6 (2.001 b.7% (84%-94%) 88.2% (90%-97%) 93.1) 52.1% (81%-92%) 82.974 Table 2 Patient characteristics grouped by presence or absence of appendicitis Appendicitis (n = 155) Age (y) Sex Male Female Mean PAS (SD) Mean Alvarado score (SD) Migration of pain Anorexia Nausea/vomiting Right lower quadrant tenderness Rebound pain (R) Cough/hopping/percussion/tenderness in the right lower quadrant Elevation in temperature (≥37.8% 68.5% (92%-100%) 99.8% (49%-66%) 72% (65%-80%) 84.7% 43.1% 83.3% (4%-14%) 22% (15%-30%) 42.202 .0% 69.4) 45. Of those.2% P .0) 35.4% 57.3% 38. 4. symptoms).9% 39. Of those patients with interobserver scores.0% 0.3% (88%-96%) 89. and a PAS score of 6 or greater.4% 90.

and 155 patients (54%) were diagnosed with pathologic examination-confirmed appendicitis.587 for PAS.9% and 83. 136 patients (47. Percentages of agreement between the 2 pairs of assessors were 87. (B) Receiver operating characteristic curves for the PAS and Alvarado scores for female patients.6 (SD ± 2. 0. respectively. yielding a positive predictive value (PPV) of 67. Fig. a negative predictive value of 78.715-0.4%) were female.4).7%. When the Alvarado scoring system was used for diagnosing appendicitis. a negative predictive value of 72%. a sensitivity of 88%. 1 shows both of the ROC curves. When PAS scores were stratified by sex. With a cutoff score of 6. Interobserver reliability was obtained on 182 (63.721-0. 1 Receiver operating characteristic curves for the PAS and Alvarado scores.6%. Of the patients with appendicitis. In female patients.718 and the negative likelihood ratio was 0. and the negative likelihood ratio was 0. a sensitivity of 86%. a PAS cutoff score of 6 or greater would have led to 137 correct appendicitis diagnoses among 155 appendicitis patients. the same cutoff score of 6 and greater would have led to 82 correct appendicitis diagnoses among the 95 male patients with appendicitis with a PPV of 78.827). whereas those without appendicitis had a mean PAS score of 5.1%. .1 (range. 97 (53%) had appendicitis. and a specificity of 72%. the same cutoff score would yield 40 correct appendicitis diagnoses among the 60 female patients with appendicitis with a PPV of 66.232. and a specificity of 50%.2 (SD ± 2.001).332.6 (SD ± 2. Patients with appendicitis had a mean Alvarado score of 7.5%. Patients with appendicitis had a mean PAS score of 7. Overall κ coefficients were 0.4%) of the 287 patients. and a specificity of 74% (P b . the PAS positive likelihood ratio was 1. Of those patients with interobserver scores.001).7%) were less than 10 years of age. and a specificity of 59%.8 years with an SD of 3.2 (SD ± 2.1) (P b . Table 3 shows the sensitivity and specificity for all cutoff values for both scoring systems with the score of 7 highlighted for the Alvarado score and the score of 6 highlighted for the PAS score as these are previous published cutoff values for determining the diagnosis of appendicitis. Table 2 shows the characteristics of patients with and without appendicitis. and the area under the ROC curve for the Alvarado scores was 0.0) (P b . 125 patients (43. 416). a sensitivity of 76%.5% for Alvarado scale and PAS scale. With a cutoff score of 7.001).768. and patients without appendicitis had a mean Alvarado score of 5.4). Fig.Using appendicitis scores in the pediatric ED 975 The median age was 9.833). a cutoff value of 7 or greater would have led to 118 correct diagnoses among the 155 patients with appendicitis yielding a PPV of 76.1% (put in 95% CIs for all of these).777 (95% confidence interval [CI]. The area under the ROC curve for the PAS scores was 0.771 (95% CI. 0.660 for Alvarado and 0. (A) Receiver operating characteristic curves for the PAS and Alvarado scores for male patients. a sensitivity of 67%. the Alvarado score positive likelihood ratio was 2.

our high percentage of patients with appendicitis may not accurately reflect patient population of other EDs. Goldman et al [17] found that a PAS score of 7 or greater (rather than the 6 originally proposed by Samuel) gave a sensitivity of 94%. Our higher proportion of appendicitis was similar to the original Alvarado and PAS studies. even among experienced physicians. Some earlier studies reported that the sex of the patient may affect the effectiveness of the scoring systems [13. used a cutoff value of 6 or greater for recommending surgical evaluation of patients for probable appendicitis and reported a sensitivity of 100%. the same cutoff score would yield 33 correct appendicitis diagnoses among the 60 female patients with appendicitis. This variation of interobserver scores highlights the inherent difficulty and subjectivity of the scoring system itself. some potentially significant variation in our interobservers with 88% agreement for the Alvarado score and 83. In contrast. we found with male patients that the same cutoff score of 7 and greater would correctly diagnose 80 among the 95 male patients with appendicitis. and a specificity of 66%. the same cutoff score of 7 and greater would have yielded 80 correct appendicitis diagnoses among the 95 male patients with appendicitis with a PPV of 80. our results were still similar to those of Schneider et al [13] and Bhatt et al [18]. a specificity of 98%.5% had appendicitis. The scores were designed to predict and recommend the need for surgery without the input of imaging. and specificity of 81%. We only enrolled patients suspected of having appendicitis. resulting in 20 (33.5% for the PAS scale. and interobserver scores were obtained on 63% of patients. There was. In addition. and 54% had pathologic examination-proven appendicitis. The pediatric appendicitis score.9%) with appendicitis. Schneider et al [13] found that the same cutoff score of 6 or greater had a PPV of 54%. If we look at the data to try and identify cutoffs where no resource use would be needed. Discussion The Alvarado scoring system was originally derived retrospectively and included both pediatric and adult patients. and that if this cutoff score was used to take patients to the operating room. Goldman et al [17] applied the scoring system to all children who presented with abdominal pain. Schneider et al [13] observed 588 patients aged 3 to 21 years. CT. Several studies looking at CT scans have found sensitivities of 95% to 97% with 5. In their sample only 14. We found that both the PAS and Alvarado scoring systems performed overall fair with somewhat improved sensitivities in male patients but not clinically useful enough to use this test solely to diagnose appendicitis.4%) with appendicitis. Schneider et al [13] and Bhatt et al [18] enrolled only patients with suspected appendicitis and their populations both had a rate of appendicitis of approximately 34%.8%) with appendicitis and incorrectly diagnosing 19 patients (33. the same score would correctly diagnose 33 patients among the 60 with appendicitis. a sensitivity of 55%. K. a sensitivity of 84%. However. and a specificity of 65% [12]. Bhatt et al [18] found a sensitivity of 92. sensitivity of 72%. When the PAS score was stratified by sex. or operating room depending on local resources. however.6%) incorrect diagnoses of appendicitis. the same cutoff score of 6 and greater would have correctly diagnosed 82 among the 95 male patients with appendicitis. with 23 (41. they found a PPV of 65%. For female patients. the same cutoff score would have detected 40 of the 60 female patients with appendicitis. Mandeville et al.8% and a specificity of 69. The .18]. with a specificity of 92% [11]. a sensitivity of 82%. and using the Alvarado recommended score of 7 as a cutoff value for having appendicitis. Three prospective studies in children to date have come up with differing results.001). The difference in these studies may be the population of patients enrolled.8%. In female patients. others would then proceed to ultrasound. In female patients. The CT scan use has increased for the past several years because of its rapid diagnosis and improved sensitivity and specificity for diagnosing appendicitis in children.14. only 4% would have been taken without having appendicitis. there is no clinically reasonable level where the number of negative appendectomies and number of missed appendicitis cases would be considered acceptable (Table 3). This selection thus led us to have a high percentage of enrolled patients with appendicitis (54%). A more useful exercise would be to develop an algorithm where these scores are used to identify a subset of patients that can be sent home from the ED without imaging or surgical consultation. Our results were similar to theirs and confirm that the scoring system did not perform as well as in the original study.976 When the Alvarado scores were stratified by sex.7%) with appendicitis.3%) missed cases of appendicitis and 20 (26. When we looked at the Alvarado score and stratified it by sex. there are several different modalities that are currently used. and a specificity of 82% (P b . In addition. with a PPV of 70. developed by Samuel. It was developed primarily to identify patients with suspected appendicitis that could be taken directly to the operating room. This would have resulted in missing 15 patients (15. Several studies have assessed the Alvarado scoring system with mixed results [9-14]. We only enrolled patients with suspected appendicitis because we felt this subgroup was the one that would benefit the most from a clinical scoring system to reduce the necessity of further imaging. When evaluating appendicitis. In contrast. in male patients.2%. resulting in missing 27 patients (45%) with appendicitis and incorrectly diagnosing 14 patients (18. we did have full scoring information on all patients. This would have resulted in missing 13 patients (13. This percentage may potentially be greater than other EDs because our hospital acts as a local referral center for many possible pediatric appendicitis patients due to a shortage of local pediatric surgeons in the area.1%) of patients without appendicitis being diagnosed with appendicitis.3%.

68(7):504-5.28(3):147-51. Ann Emerg Med 2007.49 (6):778-84.9]. and its pooled sensitivity is 88% making it more useful than these scoring systems in limiting CT use [16].32(3):386-91.Using appendicitis scores in the pediatric ED specificities between 93% and 99% [2. Acta Chir Iugos 2008.25(4): 184-6. Ronckers C.25:489-93. the predictive values are not sufficient to be used solely for making the diagnosis of appendicitis. Continuing diagnostic challenge of acute appendicitis: evaluation through modified Alvarado score. [17] Goldman RD. [13] Schneider C. Harker M. Trop Gastroenterol 2004. Kalle T. [9] Malik AA. Although the scores do perform better in males. Ann Emerg Med 1986.30(2):94-8. et al. [7] Karakas SP.55(1):55-61. [2] Sivit CJ. 977 [6] Hahn HB. Sharma D. [4] Chodick G. A practical score for the early diagnosis of acute appendicitis. Henretig FM. Philadelphia (Pa): Williams & Wilkins. Evaluation of the Alvarado score in the diagnosis of acute appendicitis. p. The use of the clinical scoring system by Alvarado in the decision to perform computed tomography for acute appendicitis in the ED. Lin LH. Prospective validation of the pediatric appendicitis score. Pediatric emergency medicine. Carter S. Wani NA. [10] McKay R. Recent concern about CT scan use has been its ionizing radiation exposure and the potential for increasing cancer risk. Acute appendicitis in children: comparison of clinical diagnosis with ultrasound and CT imaging. Receiver operating characteristic analysis of the diagnostic performance of a computed tomographic examination and the Alvarado score for diagnosing acute appendicitis: emphasis on age and sex of the patients. Bachur R. Applegate KE. . Review of radiation risks from computed tomography: essentials for the pediatric surgeon. J Pediatr Surg 2002. Chen DF. Dugalic VD. Isr Med Assoc J 2007. Radiology 2000. et al. Evaluation of the Alvarado score in the diagnosis of acute appendicitis. Dudgeon DL. 1607. Ultrasound has been studied. et al.216:430-3. et al. in addition to its expense and availability. Cancer risks following diagnostic and therapeutic radiation exposure in children. Am J Emerg Med 2007. Stephens D.42(4):603-7. J Pediatr Surg 2007. [12] Samuel M. Leonidas JC. Application of the MANTRELS scoring system in the diagnosis of acute appendicitis in children. Shalev V.16:591-6. Sijacki AD. [5] Kleinerman RA. [16] Shrivastava UK. Prospective validation of the pediatric appendicitis score in a Canadian Pediatric Emergency Department. Min YG. et al. et al. et al. Acad Emerg Med 2009. [15] Subotic AM. Fleisher GR.37(6): 872-81. J Comput Assist Tomogr 2008. et al. [18] Bhatt M.153:278-82.9(8): 284-7. Kharbanda A. Aust N Z J Surg 1998. Ducharme FM. et al. Sonography of acute appendicitis in children: 7 years experience. References [1] Ludwig S. et al. Shepherd J. [3] Rice H. J Pediatr 2008. Joseph L.36(Suppl 14):121-5. Conclusion Neither the PAS nor Alvarado scores have adequate predictive values in the diagnosis of appendicitis. Pediatr Radiol 2006. editors.46(3):128-31. [14] Sun JS. Pediatr Radiol 1998. Gupta A.15(5):557-64. Pediat Radiol 2000. [8] Alvarado A. Evaluating appendicitis scoring systems using a prospective pediatric cohort. [11] Hsiao KH. Excessive lifetime cancer mortality risk attributable to radiation exposure from computed tomography examinations in children. Noh HW. Frush D. Evaluation of suspected appendicitis in children and young adults: helical CT. Acta Paediatr Tw 2005. Guelfguat M. 2006. Pediatric appendicitis score. Hoepner FU. 6.