Department of Endocrinology, Diabetology and Internal Medicine, Medical University of Bialystok, Poland M.C. Skłodowskiej-Curie 24A, 15-276 Bialystok, Poland.
Clinical Research Centre, Medical University of Bialystok, Poland; M.C. Skłodowskiej-Curie 24A, 15-276 Bialystok, Poland.
Nutrients. 2019 Apr 12;11(4):834. doi: 10.3390/nu11040834.
The major risk factors of T2DM (type 2 diabetes mellitus) development are still under investigation. We evaluate the possible risk factors associated with type 2 diabetes (T2DM) in adult subjects during a five-year prospective cohort study. We recruited 1160 subjects who underwent oral glucose tolerance test, anthropometric measurements, and body composition and body fat distribution analysis at a baseline visit and again at follow-up after approximately five years. The conclusions of this study are based on observation of 219 subjects who attended both the first and follow-up visits. The fasting serum insulin was measured, and HOMA-IR (homeostatic model assessment of insulin resistance) was calculated. During the follow-up period, T2DM was diagnosed in 7.4% of participants, impaired fasting glucose in 37.7%, and impaired glucose tolerance in 9.3%. Logistic regression models, adjusted for age, were constructed. The changes in glucose concentration, visceral fat tissue content, insulin resistance, and %loss of muscle mass were chosen as the potential predictors for T2DM development. A set of independent variables was extracted. The constructed feature set comprised change in HOMA-IR (OR (odds ratio) = 1.01, < 0.01) and change in %loss of muscle mass (OR = 0.84, < 0.03). With an aim to validate the prediction capability using the selected attributes, a support vector machine classifier and leave-one-out cross-validation procedure was applied, yielding 92.78% classification accuracy. Our results show the correlation between the %loss of muscle mass and T2DM development in adults, independent of changes in insulin resistance.
2 型糖尿病(T2DM)发展的主要危险因素仍在研究中。我们在一项为期五年的前瞻性队列研究中评估了与成人 2 型糖尿病(T2DM)相关的可能危险因素。我们招募了 1160 名受试者,他们在基线访视时接受了口服葡萄糖耐量试验、人体测量学测量以及身体成分和体脂分布分析,大约五年后再次进行随访。本研究的结论是基于 219 名参加了第一次和随访的受试者的观察得出的。测量了空腹血清胰岛素,并计算了 HOMA-IR(胰岛素抵抗的稳态模型评估)。在随访期间,7.4%的参与者被诊断为 T2DM,37.7%的参与者为空腹血糖受损,9.3%的参与者为糖耐量受损。构建了调整年龄的逻辑回归模型。选择血糖浓度、内脏脂肪组织含量、胰岛素抵抗和肌肉量损失百分比的变化作为 T2DM 发展的潜在预测指标。提取了一组独立变量。构建的特征集包括 HOMA-IR 的变化(OR(比值比)= 1.01,<0.01)和肌肉量损失百分比的变化(OR = 0.84,<0.03)。为了使用所选属性验证预测能力,应用了支持向量机分类器和留一法交叉验证程序,得到了 92.78%的分类准确率。我们的结果表明,在成年人中,肌肉量损失与 T2DM 发展之间存在相关性,与胰岛素抵抗的变化无关。