Endocrinology · Medical Literature
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.
Clinical relevance
Among adults with type 2 diabetes enrolled in the ACCORD trial, is living alone associated with differences in long-term cardiometabolic levels, visit-to-visit variability of cardiometabolic indicators, medication-management patterns, and microvascular or neuropathy outcomes compared with living with another adult? __SNAPSHOT__ {"studyDesign":"Secondary analysis (post-hoc observational analysis) of a randomized controlled trial (ACCORD trial); mixed-effects models for repeated measures, linear models for within-person variability, and exploratory Cox proportional-hazards models","setting":"77 clinical sites in the United States and Canada (ACCORD trial infrastructure); institutional affiliations of the secondary analysis authors: Second People's Hospital of Hunan Province and Second Xiangya Hospital, Central South University, Changsha, China","sampleSize":"n=10,249 (living alone: n=2,078 [20.3%]; living with ≥1 adult: n=8,171 [79.7%]); 2 participants excluded for missing co-residence data from the original 10,251 ACCORD enrollees","population":"Adults aged 40–79 years with type 2 diabetes, baseline HbA1c ≥7.5%, and either prior clinical cardiovascular disease or additional cardiovascular risk factors; participants living alone were older, more often female, and had higher rates of baseline depression","intervention":"Living alone at baseline (defined as not co-residing with another adult; baseline household characteristic, not a randomized intervention)","comparator":"Living with one or more other adults at baseline","primaryOutcome":"No single pre-specified primary outcome; co-primary analytic targets were: (1) longitudinal repeated levels of HbA1c, FPG, SBP, DBP, LDL, and HDL via mixed-effects models; (2) within-person variability (SD and average real variability [ARV]) of the same six indicators; (3) medication adherence and treatment-management patterns; (4) exploratory time-to-event microvascular and neuropathy outcomes via Cox models","followUp":"Aligned with the ACCORD trial duration; the original ACCORD trial follow-up was approximately 3.5–5 years; exact analytic follow-up for this secondary analysis not separately reported","fundingSource":"Not reported in the available abstract and full-text excerpt; the underlying ACCORD trial was funded by the National Heart, Lung, and Blood Institute (NHLBI), U.S. government"}
Key findings
In fully adjusted mixed-effects models, living alone was associated with statistically significant but small absolute differences in four of six cardiometabolic indicators: HbA1c (β = +0.06%, 95% CI 0.03–0.10), fasting plasma glucose (β = +2.28 mg/dL, 95% CI 0.72–3.83), diastolic blood pressure (β = +0.40 mmHg, 95% CI 0.05–0.75), and LDL cholesterol (β = +1.77 mg/dL, 95% CI 0.45–3.09). No statistically significant mean-level differences were reported for SBP or HDL (values not reported in available text). Living alone was also associated with greater within-person variability, most consistently for SBP and DBP; HbA1c SD and HDL SD/ARV did not survive family-specific false-discovery-rate (FDR) correction, and HbA1c ARV was described as borderline. The DBP mean-level association was attenuated when analyses began at month 12, raising the possibility of early measurement confounding. 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 1.00–1.17; P = 0.052). Suboptimal oral glucose-lowering medication adherence was independently associated with higher subsequent HbA1c (β = +0.08%, 95% CI 0.07–0.10; P 2.0 (61.8% vs. 56.0%; adjusted HR 1.14, 95% CI 1.04–1.24; P = 0.004) and ankle jerk loss (50.9% vs. 46.6%; adjusted HR 1.11, 95% CI 1.02–1.21; P = 0.020); however, neither survived FDR correction (MNSI q = 0.068; ankle jerk q = 0.162) and were therefore retained as exploratory signals only. NNT/NNH: not calculable from available data. Adverse events: not applicable (observational exposure, not a therapeutic intervention).
Practice implications
This secondary analysis of the ACCORD trial established that living alone is associated with small but statistically significant increases in mean HbA1c, FPG, DBP, and LDL cholesterol, and with greater visit-to-visit variability in blood pressure indicators. Crucially, the absolute magnitudes of mean-level differences are small and unlikely to be clinically meaningful in isolation (e.g., 0.06% HbA1c, 1.77 mg/dL LDL), suggesting that mean risk-factor levels do not adequately capture the self-management disadvantage associated with living alone. The more clinically relevant signal may lie in cardiometabolic instability (variability), particularly for blood pressure, and in medication-management patterns, with suboptimal adherence demonstrating meaningful associations with subsequent glycemic and lipid control. The study did not establish a causal effect of living alone on vascular outcomes; hard microvascular endpoints were not significantly increased after multiple-comparison correction. Results apply to adults with T2DM at high cardiovascular risk recruited in a protocolized trial with intensive follow-up—a population that may differ substantially from patients seen in routine primary care, particularly those with fewer resources or less structured follow-up. Participants who live alone due to recent bereavement, social isolation, or functional limitation may represent a higher-risk subgroup not distinguishable in this dataset. Living alone is also not equivalent to loneliness, social isolation, or lack of social support. Because only the abstract and selected full-text excerpts were available to inform this summary, additional nuance in the statistical analyses, subgroup findings, and sensitivity analyses may not be fully reflected. No change in clinical practice protocols is indicated based on this study alone, but the findings suggest that household structure is a clinically ascertainable variable that may inform individualized patient education and adherence support strategies. __LIMITATIONS__ Secondary (post-hoc) observational analysis of a randomized trial: living alone was not randomized, so confounding by unmeasured social, behavioral, and clinical variables cannot be excluded.,Living alone was defined only at baseline; changes in household composition during follow-up were not captured, introducing exposure misclassification.,The baseline variable did not distinguish between adults living with children/dependents only versus truly alone, potentially misclassifying some participants.,The ACCORD cohort was selected for high cardiovascular risk and intensive clinical follow-up; results may not generalize to lower-risk community-dwelling populations with less structured care.,Multiple comparisons across six cardiometabolic indicators and multiple variability metrics required FDR correction; several findings of interest did not survive correction and must be treated as exploratory.,Living alone is a crude household proxy that does not capture loneliness, perceived social support, functional capacity, or reasons for living alone, limiting mechanistic interpretation.,The borderline adjusted OR for oral glucose-medication adherence (1.08, 95% CI 1.00–1.17; P = 0.052) does not meet conventional significance thresholds.,The attenuation of the DBP mean-level association when analyses began at month 12 raises the possibility of early measurement or ascertainment bias.,Medication adherence was participant-reported at protocol visits and may be subject to social-desirability bias; pharmacy records or pill counts were not available.,Funding source for this secondary analysis was not reported; industry sponsorship of the primary ACCORD trial components should be considered in contextualizing findings.,Hard cardiovascular outcomes (myocardial infarction, stroke, cardiovascular death) were not assessed in this analysis. __BOTTOM_LINE_CATEGORY__ potentially_practice_informing __BOTTOM_LINE__ In this large ACCORD secondary analysis, living alone was associated with small but statistically significant elevations in HbA1c, FPG, DBP, and LDL, and with greater blood pressure variability over time, but not with significantly increased microvascular or hard vascular outcomes after multiple-comparison correction. The more actionable signal is the link between suboptimal medication adherence and subsequent glycemic and lipid deterioration, independent of living situation. Clinicians may consider routinely ascertaining household structure as a prompt to assess medication adherence and social support needs, particularly in patients with blood pressure variability or recurrent glycemic fluctuations, while recognizing that the absolute cardiometabolic differences associated with living alone are modest.