Objective: To assess the BISAP and APACHE II scores in predicting severity according to the 2012 Atlanta classification and whether the obesity factor added to these scores improves prediction.
Material and methods: A prospective study between January 2013 and April 2014 including all patients with acute pancreatitis was performed according to the new Atlanta 2012 classification. ROC curves were fabricated for BISAP, BISAP-O, APACHE-II scores and Apache O and appropriate cutoffs were selected to the sensitivity, specificity, PPV, NPV, RPP and RPN.
Results: We studied 334 patients. 65.27% were overweighted or obese. The biliar etiology was 86.53%. Only 8.38% had severe pancreatitis and 1.5% died. Areas under the ROC curve and cut points selected were: BISAP: 0.8725, 2; BISAP-O: 0.8246, 3; APACHE-II: 0.8547, 5; APACHE-O: 0.8531, 6. Using these cutoffs the sensitivity, specificity, PPV, NPV, RPP and RPN were BISAP: 60.71%, 91.83%, 40.48%, 96.23 %, 7.43, 0.43; BISAP-O: 60.71%, 86.93%, 29.82%, 96.03%, 4.76, 0.45; APACHE-II: 85.71%, 76.14%, 24.74%, 98.31%, 3.6, 0.19; APACHE-O: 82.14%, 79.41%, 26.74%, 97.98%, 4, 0.22.
Conclusions: BISAP, BISAP-O, APACHE-II and APACHE-O systems can be used to identify patients at low risk of severity because of its high NPV, however their use should be cautious considering that the RPP and RPN do not reach optimal levels indicating that their value in predicting severity is limited. On the other hand adding the obesity factor did not improve their predictive ability.