[Establishment and validation of a risk predictive model of preoperative drug-induced limitation of pupil dilation in type 2 diabetes mellitus patients with concomitant cataract].

ABSTRACT

Objective: To establish and validate a risk predictive model of preoperative drug-induced limitation of pupil dilation (PD) in type 2 diabetes mellitus (T2DM) patients with concomitant cataract. Methods: A cross-sectional study was performed, in which 376 T2DM patients with concomitant cataract who received cataract operation in the Second Affiliated Hospital of Zhejiang University School of Medicine from October 2022 to March 2023 were randomly selected as the study subjects. Of the 376 patients, 268 who were admitted to the hospital from October to December 2022 served as the modeling group, and were divided into PD limited group (n=187) and PD unlimited group (n=81) based on whether they had drug-induced limitation of PD. Logistic regression was used to establish a risk predictive model, R software was used to draw the nomogram, Hosmer-Lemeshow test was utilized to judge the model's goodness of fit, and receiver operating characteristic (ROC) curve was adopted to validate the predicting efficacy of the model. Another 108 T2DM patients who received cataract operation in the same hospital from January to March 2023 served as the validation group, and Hosmer-Lemeshow test and ROC curve were used for the external validation of the model. Results: In the modeling group (n=268), there were 124 males and 144 females, with the mean age of (66.6±6.8) years, while in the validation group (n=108), there were 51 males and 57 females, with the mean age of (64.9±9.1) years. The incidence of preoperative drug-induced limitation of PD was 69.8% (187/268) in T2DM patients with concomitant cataract. T2DM disease course (OR=1.134, 95%CI 1.074-1.198, P<0.001), body mass index (BMI) (OR=0.863, 95%CI 0.767-0.972, P=0.015), glycohemoglobin (HbA1c) level (OR=1.397, 95%CI 1.055-1.849, P=0.019) and baseline pupil dimeter (OR=0.089, 95%CI 0.045-0.179, P<0.001) were the risk factors of drug-induced limitation of PD. Hosmer-Lemeshow test showed χ2=6.231 and P=0.621, the area under curve (AUC) of ROC curve was 0.897 (95%CI 0.857-0.937, P<0.001), and when the Youden index was the maximum (0.655), the model's sensitivity and specificity was 0.877 and 0.778, respectively. The external validation results demonstrated that the AUC of ROC curve was 0.928 (95%CI 0.875-0.981, P<0.001), the maximum Youden index was 0.761, the sensitivity was 0.932, the specificity was 0.829, and the overall accuracy was 89.8%. Conclusion: The risk predictive model established in the current study can provide reference for the clinical assessment of the risk of preoperative drug-induced limitation of PD in T2DM patients with concomitant cataract.