Living alone, long-term cardiometabolic stability, and medication management in adults with type 2 diabetes: a secondary analysis of the ACCORD trial.

Abstract

Living alone is an easily ascertained household characteristic that may shape diabetes self-management. Its relation to long-term cardiometabolic stability and medication-management patterns remains unclear. We conducted a secondary analysis of the ACCORD trial. After excluding two participants with missing baseline adult co-residence information, living alone was defined as not living with another adult at baseline. Outcomes were repeated HbA1c, fasting plasma glucose (FPG), systolic blood pressure (SBP), diastolic blood pressure (DBP), low-density lipoprotein cholesterol (LDL), and high-density lipoprotein cholesterol (HDL). Mixed-effects models estimated differences in repeated levels, linear models estimated within-person variability (standard deviation and average real variability), and exploratory Cox models examined microvascular and neuropathy outcomes. Medication adherence and treatment-management patterns were assessed from longitudinal ACCORD records. Among 10,249 participants, 2,078 (20.3%) were classified as living alone. Participants living alone were older, more often female, and had more baseline depression. In fully adjusted mixed-effects models, living alone was associated with slightly higher HbA1c (beta 0.06%, 95% CI 0.03 to 0.10), FPG (beta 2.28 mg/dL, 95% CI 0.72 to 3.83), DBP (beta 0.40 mmHg, 95% CI 0.05 to 0.75), and LDL cholesterol (beta 1.77 mg/dL, 95% CI 0.45 to 3.09). Living alone was also associated with greater variability, most consistently for SBP and DBP; HbA1c SD and HDL SD/ARV remained below family-specific false-discovery-rate thresholds, whereas HbA1c ARV was borderline. The DBP mean-level association was attenuated when analyses began at month 12, and neuropathy outcomes were retained as exploratory. Suboptimal oral glucose-lowering medication adherence was descriptively more frequent among participants living alone, but the adjusted association was borderline (adjusted OR 1.08, 95% CI 1.00 to 1.17; P = 0.052). In ACCORD, living alone was associated with small differences in glycemic and lipid levels, greater variability in selected cardiometabolic indicators, and medication-management patterns linked to subsequent metabolic control. Findings suggest self-management instability rather than a uniform increase in vascular outcomes.