Risk tool can predict heart failure in patients with diabetes

A novel machine learning-derived risk score can predict the risk for heart failure among outpatients with type 2 diabetes mellitus (T2DM), according to a study published online Sept. 13 in Diabetes Care to coincide with the annual meeting of the Heart Failure Society of America, held from Sept. 13 to 16 in Philadelphia.

Matthew W. Segar, M.D., from the University of Texas Southwestern Medical Center in Dallas, and colleagues used data from 8,756 patients participating in the Action to Control Cardiovascular Risk in Diabetes trial to develop a machine learning-derived model to predict the risk for heart failure among patients with T2DM. The random survival forest (RSF) methods model was validated in a cohort of individuals with T2DM using the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial.

The researchers created an integer-based risk score for five-year heart failure incidence called the WATCH-DM (weight [BMI], age, hypertension, creatinine, high-density lipoprotein cholesterol, diabetes control [fasting plasma glucose], QRS duration, myocardial infarction, and coronary artery bypass graft). For each one-unit increment in the risk score, there was a 24 percent higher relative risk for heart failure within five years. There was a graded increase in the cumulative five-year incidence of heart failure, from 1.1 percent in quintile 1 (WATCH-DM score ≤7) to 17.4 percent in quintile 5 (WATCH-DM score ≥14). The RSF-based risk prediction model and the WATCH-DM risk score performed well with good discrimination in the external validation cohort (C-index, 0.74 and 0.70, respectively). Furthermore, the investigators noted acceptable calibration and broad risk stratification (five-year heart failure risk range from 2.5 to 18.7 percent across quintiles 1 to 5).

"We hope that this risk score can be useful to clinicians on the ground -- primary care physicians, endocrinologists, nephrologists, and cardiologists -- who are caring for patients with diabetes and thinking about what strategies can be used to help them," a coauthor said in a statement.

Several authors disclosed financial ties to the pharmaceutical and medical device industries.

 

Risk tool can predict heart failure in patients with diabetes

A novel machine learning-derived risk score can predict the risk for heart failure among outpatients with type 2 diabetes mellitus (T2DM), according to a study published online Sept. 13 in Diabetes Care to coincide with the annual meeting of the Heart Failure Society of America, held from Sept. 13 to 16 in Philadelphia. Matthew W. Segar, M.D., from the University of Texas Southwestern Medical Center in Dallas, and colleagues used data from 8,756 patients participating in the Action to Control Cardiovascular Risk in Diabetes trial to develop a machine learning-derived model to predict the risk for heart failure among patients with T2DM. The random survival forest (RSF) methods model was validated in a cohort of individuals with T2DM using the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial. The researchers created an integer-based risk score for five-year heart failure incidence called the WATCH-DM (weight [BMI], age, hypertension, creatinine, high-density lipoprotein cholesterol, diabetes control [fasting plasma glucose], QRS duration, myocardial infarction, and coronary artery bypass graft). For each one-unit increment in the risk score, there was a 24 percent higher relative risk for heart failure within five years. There was a graded increase in the cumulative five-year incidence of heart failure, from 1.1 percent in quintile 1 (WATCH-DM score 7) to 17.4 percent in quintile 5 (WATCH-DM score 14). The RSF-based risk prediction model and the WATCH-DM risk score performed well with good discrimination in the external validation cohort (C-index, 0.74 and 0.70, respectively). Furthermore, the investigators noted acceptable calibration and broad risk stratification (five-year heart failure risk range from 2.5 to 18.7 percent across quintiles 1 to 5). We hope that this risk score can be useful to clinicians on the ground -- primary care physicians, endocrinologists, nephrologists, and cardiologists -- who are caring for patients with diabetes and thinking about what strategies can be used to help them, a coauthor said in a statement. Several authors disclosed financial ties to the pharmaceutical and medical device industries.