Analysis of risk factors for hypoglycemia in patients with gastrointestinal bleeding induced by somatostatin for injection and establishment of the prediction model.

OBJECTIVES: This study aimed to explore the risk factors of hypoglycemia in patients with gastrointestinal bleeding caused by somatostatin for injection and to establish a prediction model based on logistic regression combined with receiver operating characteristic (ROC) curve.

MATERIALS AND METHODS: This retrospective study analyzed patients diagnosed with gastrointestinal bleeding and treated with somatostatin from January 2022 to May 2023 and collected hypoglycemic events. Univariate and multivariate logistic regression analysis were used to determine the independent influencing factors of somatostatin-induced hypoglycemia, and a prediction model was established. ROC analysis was used to evaluate the prediction model.

RESULTS: A total of 331 patients were enrolled in this study, and 42 patients developed hypoglycemic events. Age and co-infection were found to be significant risk factors for hypoglycemia in patients with gastrointestinal bleeding induced by somatostatin. Binary logistic regression fitting established the hypoglycemia prediction model Logit (P) = -4.125+0.053Yage +1.366Yco-infection (co-infection: Xco-infection = 1, non-co-infection: Xno co-infection = 0), Hosmer-Lemeshow test results showed that the model had a good fit (χ2 = 10.552, df = 8, p = 0.228), and the AUC of the prediction model to predict the risk of hypoglycemia caused by somatostatin in patients with gastrointestinal bleeding was 0.744 (95% CI: 0.653 - 0.835, p < 0.001), the sensitivity was 57.14%, and the specificity was 93.77%.

CONCLUSION: Among adult patients with gastrointestinal bleeding treated with somatostatin for injection, our study found that age and co-infection were significant risk factors for somatostatin-induced hypoglycemia in this patient population, and the fitted models had high predictive value in predicting the occurrence of hypoglycemia.