Muscle strength and episodic memory are linked in later life, but factors associated with their first transitions to low reserve remain uncertain.

We analyzed four longitudinal cohorts of adults aged 50–100 years with at least three synchronous strength and memory assessments. Participants entered follow-up when both cohort-standardized residual scores exceeded −1 SD. Continuous-time panel multistate models estimated muscle-first and memory-first transition intensities. Cohort-specific effects were pooled using random-effects meta-analysis with Hartung–Knapp inference. The primary contrast was the ratio of covariate-specific intensity ratios (IRRs) for memory-first versus muscle-first transitions.

The primary analysis included 46,356 participants, with 6,036 first observed muscle-first, 8,539 memory-first and 897 both-low transitions. Older age was associated with a relative memory-first pattern (IRR ratio per 5 years 1.110, 95% CI 1.082–1.139); activities of daily living (ADL) limitation was associated with a relative muscle-first pattern (0.664, 95% CI 0.559–0.788). Diabetes was associated with higher intensities in both domains, but its pathway contrast remained uncertain (0.751, 95% CI 0.554–1.019; P = 0.058). Age and ADL associations persisted after adjustment for baseline reserve and multimorbidity and when participants with two assessments were included. Age-specific SD thresholds attenuated the age contrast to 1.048 (95% CI 1.016–1.081). Diabetes findings varied with the low-reserve threshold.

Age and ADL limitation were associated with different relative patterns of first low-reserve transitions. Diabetes was associated with higher transition intensities in both domains without a clear pathway difference at the primary threshold. These observational associations do not establish a biological sequence or clinical screening thresholds.