Longitudinal Trajectories of Frailty and Functional Decline and Risk of Incident Stroke in Older Adults with Dysglycemia

Technical Summary: Longitudinal Trajectories of Frailty and Functional Decline and Risk of Incident Stroke in Older Adults with Dysglycemia

Problem Statement
Stroke remains a leading cause of mortality and disability, particularly among older adults with dysglycemia (diabetes or prediabetes), who face nearly double the risk compared to normoglycemic individuals. While frailty and limitations in instrumental activities of daily living (IADL) are established independent predictors of adverse cardiovascular outcomes, existing research largely treats these as static, single-time-point exposures. There is a lack of understanding regarding the longitudinal co-development of frailty and functional decline, and how their joint trajectories influence the risk of incident stroke in high-risk geriatric populations. This study addresses the gap in identifying distinct, dynamic trajectories of the frailty index (FI) and IADL limitation and evaluating their combined association with stroke risk.

Methodology
The study utilized data from the China Health and Retirement Longitudinal Study (CHARLS), a nationally representative cohort. The analysis included 2,372 community-dwelling adults aged ≥60 years with diabetes or prediabetes, free of baseline stroke, with follow-up data spanning 2011 to 2020 (median follow-up of 7 years).

• Variable Construction:
  • Frailty Index (FI): Calculated using a deficit accumulation approach based on 39 health deficits across physical, cognitive, psychological, and sensory domains.
  • IADL Limitation: Assessed via a binary indicator (difficulty in one or more of five tasks: managing money, phone calls, medications, shopping, cooking).
  • Outcome: Incident stroke was defined by self-reported physician diagnosis at follow-up waves.
• Statistical Modeling:
  • Trajectory Identification: Growth Mixture Modeling (GMM) was applied to identify latent classes of FI trajectories. Latent Class Growth Analysis (LCGA) with a logit link was used for IADL limitation trajectories. Model selection relied on fit indices (AIC, BIC, aBIC, LMR, BLRT, entropy, and posterior probabilities).
  • Joint Modeling: Joint Latent Class Growth Analysis was performed to derive combined trajectory patterns of FI and IADL limitation simultaneously.
  • Association Analysis: Multivariable logistic regression estimated Odds Ratios (ORs) and 95% Confidence Intervals (CIs) for the association between trajectory classes and incident stroke, adjusting for demographics, lifestyle factors, and comorbidities (Model 3).
  • Sensitivity: Robustness was tested by excluding participants with baseline cardiovascular disease (CVD) or hypertension, and by stratifying by sex.

Key Contributions

1. Dynamic Trajectory Identification: The study identified five distinct longitudinal trajectories for both the FI (e.g., low-stable, low-increasing, high-increasing, moderate-fluctuating, moderate-increasing) and IADL limitation (e.g., no limitation, late-onset, early-onset recovering, fluctuating, persistent).
2. Joint Pattern Characterization: By combining these domains, the study established five joint trajectory patterns, capturing the heterogeneity in the co-development of physiological vulnerability and functional decline.
3. Risk Stratification Framework: The research moves beyond static assessments to demonstrate that the pattern of decline (e.g., progressive vs. stable, persistent vs. recovering) provides a more granular risk profile for cerebrovascular events.

Results

• Trajectory Distribution: Five joint patterns were identified. The largest group (45.3%) exhibited low FI with no IADL limitation. The highest-risk group (9.3%) was characterized by high FI with persistent IADL limitation.
• Stroke Incidence: Over the follow-up period, 292 participants (12.3%) developed incident stroke.
• Association with Stroke:
  • In the fully adjusted model, the high-FI/persistent-IADL-limitation group exhibited the highest risk of incident stroke (OR 3.37, 95% CI 2.18–5.21) compared to the low-FI/no-limitation reference group.
  • Other high-risk joint patterns included increasing-FI/late-IADL-limitation (OR 2.67), moderate-FI/recovering-IADL (OR 2.20), and increasing-FI/early-IADL-limitation (OR 2.05).
  • Notably, the moderate-fluctuating FI trajectory alone was not significantly associated with stroke (OR 1.12), suggesting that transient frailty fluctuations may not confer lasting cerebrovascular risk.
• Sensitivity: The association between the high-risk joint trajectory and stroke remained robust across all sensitivity analyses, including sex-stratified models and exclusions of baseline comorbidities.

Significance and Claims
The authors claim this is the first study to examine the joint longitudinal trajectories of FI and IADL limitation and their association with incident stroke specifically within a dysglycemic population. The primary significance lies in the finding that the co-occurrence of high frailty and persistent functional decline confers a greater risk (more than a threefold increase) than either factor assessed in isolation or at a single time point.

The paper posits that these findings support the integration of multidimensional geriatric assessments into cardiovascular risk stratification. It suggests that older adults with diabetes or prediabetes exhibiting concurrent progressive frailty and persistent IADL limitation represent a high-priority subgroup for intensive stroke prevention strategies. Conversely, the study highlights that maintaining physical and functional integrity is protective, as the low-risk group comprised nearly half the sample. The authors conclude that routine longitudinal monitoring of both frailty and IADL status in primary care could facilitate early identification of trajectory shifts and timely intervention.