Objectives: To evaluate the utility of C-reactive protein (CRP) in the detection of necrosis in acute pancreatitis and to determine the best cut-off point for CRP used for this purpose.
Material and methods: We performed a retrospective study of 157 patients with acute pancreatitis who underwent computed tomography (CT) with intravenous contrast material between 72 h and 8 days after the onset of symptoms and whose serum CRP values were determined by nephelometry 24 h or later after the onset of pain.
Results: Ninety-four patients were men and 63 were women, with a mean age of 61 years (range, 15-96 years). The cause of pancreatitis was biliary lithiasis in 53.5%, alcohol in 20.4%, and idiopathic in 10.8%. Other causes were found in 15.3%. The mean (standard deviation) time elapsed between symptom onset and extraction to evaluate CRP was 3.21 (1.7) days. The patients were divided into two groups according to the results of CT: 132 patients with acute intersitial edematous pancreatitis and 25 patients with acute necrotizing pancreatitis. The mean CRP concentrations were: 322 mg/l (range, 10.7-538) in patients with acute necrotizing pancreatitis and 133 mg/l (range, 3-442) in those with acute interstitial pancreatitis; this difference was statistically significant (p < 0.001). The area under the ROC curve of CRP vs. the occurrence of necrosis was 0.862 (95% CI, 0.778, 0.946). To evaluate the presence of pancreatic necrosis a cut-off level of 200 mg/l showed a sensitivity of 88% and a specificity of 75% while a cut-off level of 279 mg/l presented a sensitivity of 72% and a specificity of 88%.
Conclusions: In acute pancreatitis, a CRP value of less than or equal to 200 mg/l obtained at 72 h of symptom onset is useful for ruling out, with a high degree of probability, the presence of necrosis. With higher values, additional investigations should be performed to determine the presence of pancreatic necrosis; nevertheless, with values higher than 279 mg/l the risk of necrosis markedly increases.