Can the Combination of White Blood Cell Count and C-Reactive Protein Rule Out Pediatric Appendicitis, a Retrospective Analysis?

Background: Pediatric appendicitis (PA) is the most frequent cause of pediatric emergency surgery. Standard PA scoring systems incorporate inflammatory biomarkers (white blood cell count [WBC] or C-reactive protein [CRP]); however, the individual biomarker sensitivities have not been reported beyond single-center sites. We aimed to measure the individual and combined sensitivities of WBC and CRP to rule out PA in a large multicenter hospital system. Methods: We did a retrospective study of pediatric emergency department patients (age <18) with abdominal pain, using a deidentified electronic health database from a 175-hospital system in the United States, comparing PA versus non-PA patients. Pediatric appendicitis patients were identified by ICD-10 diagnosis codes and had advanced imaging, whereas non-PA patients had advanced imaging and were discharged home with the ICD-10 code for abdominal pain. Using matched propensity matching, we calculated receiver operator characteristics and Youden's index to find the optimal cutpoints. Sensitivity was calculated at the optimal and traditional cutpoints for each biomarker, also combined in parallel fashion. Results: We identified 7414 subjects (3707 PA matched to 3707 controls). The WBC in PA patients was 14.99 ± 5.00 k/mm3 versus 9.63 ± 3.83 k/mm3 in non-PA patients (P < .001). The CRP in PA patients was 5.13 ± 6.43 mg/dL versus 1.22 ± 2.17 mg/dL in controls (P < .001). The optimal cutpoints (Youden's index) were 11.5 k/mm3 for WBC and 1.11 mg/dL for CRP. The sensitivity of WBC to rule out PA varied between 75.6%-83.9%, depending on the cutpoint (10.0-11.5 K/mm3). Similarly, the sensitivity of CRP to rule out PA varied from 64.9%-66.7%, depending on the cutpoint (1.0-1.11 mg/dL). Combined analysis showed that low WBC and CRP had a sensitivity of 90.0%-93.0%, yielding a negative predictive value of 99.2%-99.4% to rule out PA. Conclusion: The PA patients in our study had significantly higher inflammatory biomarkers than the non-PA patients. Combining WBC less than 10.0 k/mm3 and CRP less than 1.0 mg/dL was 93.0% sensitive to rule out PA. Researchers should consider this combination of biomarkers to rule out PA in future prospective studies.