Interleukin-6 significantly improves predictive value of systemic inflammatory response syndrome for predicting severe acute pancreatitis

Abstract

Background: Predicting severe acute pancreatitis (AP) is important for triage, prognosis, and designing therapeutic trials. Persistent systemic inflammatory response syndrome (SIRS) predicts severe AP but its diagnostic accuracy is suboptimal. Our objective was to study if cytokine levels could improve the predictive value of clinical variables for the development of severe AP. Methods: Consecutive patients with AP were included in a prospective cohort study at a tertiary care center. Serum levels of IL-6, TNF-α, IL-10, MCP-1, GM-CSF and IL-1β were measured at day 3 of onset of AP. Variables such as age, co-morbidity, etiology, SIRS, and cytokines were modeled to predict severe AP by multivariable regression analysis. Genotyping was done to correlate IL-6, TNF-α and MCP-1 gene polymorphisms with cytokine levels. Results: Of 236 patients with AP, 115 patients admitted within 7 days of onset formed the study group. 37 of the 115 (32%) patients developed organ failure. Independent predictors of organ failure were persistent SIRS (OR 34; 95% CI: 7.2-159) and day 3 serum IL-6 of >160 pg/ml (OR 16.1; 95% CI:1.8-142). IL-6 gene (-174 G/C) GG genotype was associated with significantly higher levels of IL-6 compared to CC/CG genotype. Serum IL-6 >160 pg/ml increased the positive predictive value of persistent SIRS from 56% to 85% and specificity from 64% to 95% for predicting OF without compromising its sensitivity and negative predictive value. Conclusion: Serum IL-6 of >160 ng/ml added significantly to the predictive value of SIRS for severe AP.