AIMS: This study intended to assess the association of peripheral neuropathy with statins therapy amongst Type 2 diabetic patients.
METHODS: At Penang General Hospital, 757 cases were categorized into two groups (564 with statins therapy and 193 without statins therapy). The diagnosis of PN was investigated retrospectively for a period of 10 years (2006-2016). Confounding risk factors as age, diabetes period, hypertension, glycemic control, other co-morbidity, and prescriptions were matched.
RESULTS: About 129 (22.9%) cases from 564 statins users had PN. Only 30 (15.5%) subjects had PN from 193 statins non-users. Chi-square test showed a significant variance among statins treatment cohort and statin-free cohort in the occurrence of PN (P-value: 0.001). Spearman's investigation presented a positive correlation (r: 0.078, p-value: 0.031) among statins use and PN prevalence. Binary logistic regression was statistically significant for statins therapy as a predictor of peripheral neuropathy incidence (r2: 0.006, p-value: 0.027) amid diabetic patients. The relative risk of peripheral neuropathy connected with statins therapy is (RR: 1.47, 95% CI: 1.02-2.11). The excess relative risk is 47.1%. While the absolute risk (AR) is 7.3% and the number needed to harm (NNH) is 14.
CONCLUSIONS: The study indicated a positive association between peripheral neuropathy and statins utilization. Peripheral neuropathy was higher amongst statins users than the statins-free group.
METHOD: In this study, the MNSI data were collected from the Epidemiology of Diabetes Interventions and Complications (EDIC) clinical trials. Two different datasets with different MNSI variable combinations based on the results from the eXtreme Gradient Boosting feature ranking technique were used to analyze the performance of eight different conventional ML algorithms.
RESULTS: The random forest (RF) classifier outperformed other ML models for both datasets. However, all ML models showed almost perfect reliability based on Kappa statistics and a high correlation between the predicted output and actual class of the EDIC patients when all six MNSI variables were considered as inputs.
CONCLUSIONS: This study suggests that the RF algorithm-based classifier using all MNSI variables can help to predict the DSPN severity which will help to enhance the medical facilities for diabetic patients.