Prediction and evaluation of a nomogram model for recurrent acute pancreatitis

The purpose of this study was to investigate the influencing factors for recurrent acute pancreatitis and construct the nomogram model to predict the risk of recurrent acute pancreatitis.

Patients diagnosed with acute pancreatitis in the Affiliated Hospital of Southwest Medical University were enrolled. We collected these patients' basic information, laboratory data, imaging information. Using Logistic regression and least absolute shrinkage and selection operator regression to select risk factor for Cross-Validation Criterion. To create nomogram and validated by receiver operator characteristic curve, calibration curves and decision curve analysis.

A total of 533 patients with acute pancreatitis were included, including 99 recurrent acute pancreatitis patients. The average age of recurrent acute pancreatitis patients was 49.69 years old, and 67.7% of them were male. At the same time, in all recurrent acute pancreatitis patients, hypertriglyceridemic pancreatitis is the most important reason (54.5%). Regression analysis and least absolute shrinkage and selection operator regression showed that smoking history, acute necrotic collection, triglyceride, and alcohol etiology for acute pancreatitis were identified and entered into the nomogram. The area under the receiver operator characteristic curve of the training set was 0.747. The calibration curve showed the consistency between the nomogram model and the actual probability.

In conclusion, some high-risk factors like smoking history, acute necrotic collection, triglyceride, and alcohol etiology for acute pancreatitis may predict recurrent pancreatitis and their incorporation into a nomogram has high accuracy in predicting recurrence.