Analysis of risk factors and establishment of risk prediction model for linezolid-associated thrombocytopenia

Objective　To explore the risk factors of linezolid associated thrombocytopenia and establish a risk prediction model. Methods　A total of 796 patients with infection treated by linezolid from January 2010 to December 2020 in 7 intensive care units (ICU) of the First Medical Center of PLA General Hospital were selected. The basic clinical data of the patients were extracted by case retrieval system, and the patients were divided into thrombocytopenia group (n=228) and non-thrombocytopenia group (n=568) according to the occurrence of thrombocytopenia. The general data of the two groups were compared, and the risk factors of linezolid-associated thrombocytopenia were analyzed by multivariate logistic regression, and a risk prediction model was established. Then, the receiver operating characteristic curve (ROC) was used to evaluate the effectiveness of the model. Results　Multivariate analysis showed age ≥80 years, duration of linezolid >7 d, WBC count ≥16×109/L, platelet count <200×109/L, and creatinine clearance (Ccr) <50 ml/min were independent risk factors for linezolid associated thrombocytopenia. The area under the curve (AUC) of the risk prediction model was 0.782(95%CI 0.747-0.816, P<0.001), The cut-off value was 6 points, the sensitivity was 82.1%, the specificity was 59.1%, the positive predictive value was 44.7%, and the negative predictive value was 89.1%. Conclusion　The risk prediction model based on age, duration of linezolid use, WBC count, platelet count and, Ccr can effectively predict the occurrence of linezolid associated thrombocytopenia. DOI: 10.11855/j.issn.0577-7402.2021.08.06