Accuracy in Diagnosing Pediatric Appendicitis: Discordance in Diagnosis and Masqueraders
MetadataShow full item record
BACKGROUND: The classification of perforated versus nonperforated appendicitis remains poorly defined despite being clinically significant and requiring dramatically different therapeutic approaches. We hypothesized that there is substantial variability in the determination of perforated and nonperforated appendicitis across specialties due to the lack of standardized criteria. To identify and quantify the degree of variability in the diagnosis of acute versus perforated appendicitis, a comparative analysis of radiologic results, operative findings, and pathologic review was undertaken to identify the degree of discordance across specialties as the primary objective. Given the new suggested paradigm of non-operative management of acute appendicitis, a secondary objective was to identify the incidence of atypical diagnoses detected among appendectomy specimens to better elucidate those potentially missed in non-operative management. METHODS: An IRB-approved retrospective review of 1311 appendectomies over a 16-month period at an independent children's hospital allowed a comparison of the diagnostic classification of appendicitis as perforated (PA) or nonperforated (NPA) based on radiology (R), operative (O), and pathology (P) reports. Three groups, P+O (N=1241), P+R (N=516), O+R (N=512), were compared to identify inter-specialty discordance in classification. The length of stay (LOS) was analyzed as a proxy for clinical behavior to confirm that the diagnostic classification was consistent with the clinical behavior of being perforated or nonperforated (PA with LOS >48hours and NPA with LOS ≤48 hours). Given the numerous pathologic descriptives of appendicitis, a more extensive pathologic review of all appendectomies was done to identify an intra-specialty variance in diagnostic classification and to identify any other diagnosis in the setting of presumed appendicitis. This arm of the study involved an expanded IRB-approved retrospective review of pediatric patients (n=6816) who underwent appendectomies at an independent children's hospital over an 11-year period from January 2000 to December 2010. Inclusion criteria were age <17 and surgery for presumed appendicitis, thus excluding incidental appendectomies (n=269) with a final review of 6547 specimens. RESULTS: In the assessment of the degree of discordance, the subsets P+O, P+R, O+R revealed a discordance of 11%, 15.7% and 16.6% within the classification of appendicitis respectively. In the O+R group, 35% of the cases that were operatively designated as 'perforated' contradicted the radiologic diagnosis of being 'nonperforated' appendicitis. Cases designated as PA in all subsets (P+O, P+R, O+R) clinically behaved as perforated with a mean LOS>48 hours (97, 95, 95, respectively), whereas the cases designated as nonperforated appendicitis (NPA) exhibited greater variation from the expected LOS≤48 hours, with means 35, 83, and 62, respectively. The more exhaustive pathologic review identified 7 classifications for appendicitis in 91.6% patients, 5% false positives, and ~3% coincident other diagnoses. 5998 (91.6%) subjects showed true appendicitis including acute non-perforated, perforated, chronic, suppurative, gangrenous, necrotizing, and catarrhal appendicitis. In 224 subjects (3.4%), diagnoses other than appendicitis were identified: non-inflammatory obstruction (n=71), other infectious etiologies (n=58), non-specific inflammatory changes (n=58), extra-appendiceal pathology (n=31), tumors (n=10), and foreign body (n=2). Of the tumors, 6 patients with true appendicitis had co-existing carcinoid tumors. 325 specimens (5.0%) were documented as negative appendicitis. CONCLUSION: Variability in the classification of appendicitis between specialties suggests an error rate inherent in diagnosis. This error rate may explain why identifying best clinical guidelines for length of antibiotic therapy and treatment has been elusive. Standardizing classification criteria across specialties may help identify best practices for optimal use of hospital resources and improve diagnostic accuracy for meaningful clinical trials. This study represents the largest analysis of the incidence of pathologies that mimic appendicitis in the pediatric population, conveying a broad overlap of diagnoses that present in a similar fashion. Given how commonly appendicitis is diagnosed, follow-up for routine appendectomies has been streamlined and expedited in such a way that the review of pathology by the practitioner may be overlooked. The number of infectious etiologies and tumors detected reinforces the increasing importance of pathology review in post-operative follow-up to appropriately diagnose uncommon conditions that may necessitate further work-up and treatment. Additionally, the possibility of missing an alternative or co-incidental diagnosis such as a carcinoid tumor in the non-operative management of appendicitis merits some reflection in planning operative versus non-operative management. Given the recent popularity of non-operative therapy for appendicitis, this research gives the clinician a reliable incidence of other potential diagnoses if symptoms recur. Having percentages for conditions that present similarly also allows for proper counseling of patients who opt for non-operative management.