Development and validation of a nomogram for predicting cefoperazone/sulbactam-induced hypoprothrombinaemia in Hospitalized adult patients.

Cefoperazone/sulbactam-induced hypoprothrombinaemia is associated with longer hospital stays and increased risk of death. The aim of this study was to develop and validate a nomogram for predicting the occurrence of cefoperazone/sulbactam-induced hypoprothrombinaemia in hospitalized adult patients. This retrospective cohort study involved hospitalized adult patients at Xi'an Central Hospital from January 2020 to December 2022 based on the Chinese pharmacovigilance system developed and established by the Adverse Drug Reaction Monitoring Center in China. Independent predictors of cefoperazone/sulbactam-induced hypoprothrombinaemia were obtained using multivariate logistic regression and were used to develop and establish the nomogram. According to the same standard, the clinical data of hospitalized patients using cefoperazone/sulbactam at the Third Affiliated Hospital of Xi'an Medical University from January 1, 2023 to June 30, 2023 were collected as the external validation group. The 893 hospitalized patients included 95 who were diagnosed with cefoperazone/sulbactam-induced hypoprothrombinaemia. Our study enrolled 610 patients: 427 in the training group and 183 in the internal validation group. The independent predictors of cefoperazone/sulbactam-induced hypoprothrombinaemia were surgery (odds ratio [OR] = 5.279, 95% confidence interval [CI] = 2.597-10.729), baseline platelet count ≤50×109/L (OR = 2.492, 95% CI = 1.110-5.593), baseline hepatic dysfunction (OR = 12.362, 95% CI = 3.277-46.635), cumulative defined daily doses (OR = 1.162, 95% CI = 1.162-1.221) and nutritional risk (OR = 16.973, 95% CI = 7.339-39.254). The areas under the curve (AUC) of the receiver operating characteristic for the training and internal validation groups were 0.909 (95% CI = 0.875-0.943) and 0.888 (95% CI = 0.832-0.944), respectively. The Hosmer-Lemeshow tests yielded p = 0.475 and p = 0.742 for the training and internal validation groups, respectively, confirming the goodness of fit of the nomogram model. In the external validation group (n = 221), the nomogram was equally robust in cefoperazone/sulbactam-induced hypoprothrombinaemia (AUC = 0.837, 95%CI = 0.736-0.938). The nomogram model constructed in this study had good predictive performance and extrapolation, which can help clinicians to identify patients at high risk of cefoperazone/sulbactam-induced hypoprothrombinaemia early. This will be useful in preventing the occurrence of cefoperazone/sulbactam-induced hypoprothrombinaemia and allowing timely intervention measures to be performed.