Predictors of critical acute pancreatitis: a prospective cohort study.

Critical acute pancreatitis (CAP) has recently emerged as the most ominous severity category of acute pancreatitis (AP). As such there have been no studies specifically designed to evaluate predictors of CAP. In this study, we aimed to evaluate the accuracy of 4 parameters (Acute Physiology and Chronic Health Evaluation [APACHE] II score, C-reactive protein [CRP], D-dimer, and intra-abdominal pressure [IAP]) for predicting CAP early after hospital admission. During the study period, data on patients with AP were prospectively collected and D-dimer, CRP, and IAP levels were measured using standard methods at admission whereas the APACHE II score was calculated within 24 hours of hospital admission. The receiver-operating characteristic (ROC) curve analysis was applied and the likelihood ratios were calculated to evaluate the predictive accuracy. A total of 173 consecutive patients were included in the analysis and 47 (27%) of them developed CAP. The overall hospital mortality was 11% (19 of 173). APACHE II score ≥11 and IAP ≥13 mm Hg showed significantly better overall predictive accuracy than D-dimer and CRP (area under the ROC curve-0.94 and 0.92 vs. 0.815 and 0.667, correspondingly). The positive likelihood ratio of APACHE II score is excellent (9.9) but of IAP is moderate (4.2). The latter can be improved by adding CRP (5.8). In conclusion, of the parameters studied, APACHE II score and IAP are the best available predictors of CAP within 24 hours of hospital admission. Given that APACHE II score is rather cumbersome, the combination of IAP and CRP appears to be the most practical way to predict critical course of AP early after hospital admission.