Unsupervised cluster analysis of clinical and metabolite characteristics in patients with chronic complications of T2DM: an observational study of real data
Cuicui Wang,Yan Li,Jun Wang,Kunjie Dong,Chenxiang Li,Guiyan Wang,Xiaohui Lin,Hui Zhao
DOI: https://doi.org/10.3389/fendo.2023.1230921
IF: 6.055
2023-10-20
Frontiers in Endocrinology
Abstract:Introduction: The aim of this study was to cluster patients with chronic complications of type 2 diabetes mellitus (T2DM) by cluster analysis in Dalian, China, and examine the variance in risk of different chronic complications and metabolic levels among the various subclusters. Methods: 2267 hospitalized patients were included in the K-means cluster analysis based on 11 variables [Body Mass Index (BMI), Systolic Blood Pressure (SBP), Diastolic Blood Pressure (DBP), Glucose, Triglycerides (TG), Total Cholesterol (TC), Uric Acid (UA), microalbuminuria (mAlb), Insulin, Insulin Sensitivity Index (ISI) and Homa Insulin-Resistance (Homa-IR)]. The risk of various chronic complications of T2DM in different subclusters was analyzed by multivariate logistic regression, and the Kruskal-Wallis H test and the Nemenyi test examined the differences in metabolites among different subclusters. Results: Four subclusters were identified by clustering analysis, and each subcluster had significant features and was labeled with a different level of risk. Cluster 1 contained 1112 inpatients (49.05%), labeled as "Low-Risk"; cluster 2 included 859 (37.89%) inpatients, the label characteristics as "Medium-Low-Risk"; cluster 3 included 134 (5.91%) inpatients, labeled "Medium-Risk"; cluster 4 included 162 (7.15%) inpatients, and the label feature was "High-Risk". Additionally, in different subclusters, the proportion of patients with multiple chronic complications was different, and the risk of the same chronic complication also had significant differences. Compared to the "Low-Risk" cluster, the other three clusters exhibit a higher risk of microangiopathy. After additional adjustment for 20 covariates, the odds ratios (ORs) and 95% confidence intervals (95%CI) of the "Medium-Low-Risk" cluster, the "Medium-Risk" cluster, and the"High-Risk" cluster are 1.369 (1.042, 1.799), 2.188 (1.496, 3.201), and 9.644 (5.851, 15.896) (all p <0.05). Representatively, the "High-Risk" cluster had the highest risk of DN [OR (95%CI): 11.510(7.139,18.557), ( p <0.05)] and DR [OR (95%CI): 3.917(2.526,6.075), ( p <0.05)] after 20 variables adjusted. Four metabolites with statistically significant distribution differences when compared with other subclusters [Threonine (Thr), Tyrosine (Tyr), Glutaryl carnitine (C5DC), and Butyryl carnitine (C4)]. Conclusion: Patients with chronic complications of T2DM had significant clustering characteristics, and the risk of target organ damage in different subclusters was significantly different, as were the levels of metabolites. Which may become a new idea for the prevention and treatment of chronic complications of T2DM.
endocrinology & metabolism