The predictive value of proteinuria in acute pancreatitis.

Acute pancreatitis has a highly variable clinical course. Early and reliable predictors for the severity of acute pancreatitis are lacking. Proteinuria appears to be a useful predictor of disease severity and outcome in a variety of clinical conditions. This study aims to investigate the predictive value of proteinuria on admission for the severity of acute pancreatitis compared with other commonly used predictors; the APACHE II score, Modified Glasgow score and C-reactive protein (CRP). This is a post-hoc analysis of 64 patients admitted with acute pancreatitis treated in one teaching hospital, who participated in a previous randomized trial. Proteinuria was defined as a Protein/Creatinine (P/C) ratio >23 mg/mmol. The primary endpoint was severe acute pancreatitis. Secondary endpoints included infectious complications, need for invasive intervention, ICU stay and in-hospital mortality. Proteinuria was present in 30/64 patients (47%). Eleven patients (17%) had severe acute pancreatitis. There was no difference in incidence of severe acute pancreatitis between patients with and without proteinuria: 6/30 patients (20%) versus 5/34 patients (15%) respectively (p = 0.58). Likewise, the occurrence of infectious complications, need for intervention and ICU stay and mortality did not differ significantly (p = 0.58, p = 0.99, p = 0.33 and p = 0.60 respectively). The diagnostic performance of the P/C ratio for the prediction of severe pancreatitis was inferior to the Modified Glasgow score (p = 0.04) and CRP (p = 0.03). Proteinuria on admission does not seem to be a reliable predictor for disease severity in acute pancreatitis. The diagnostic performance of the P/C ratio is inferior to the Modified Glasgow score and CRP.