Abstract
Background
We aimed to examine the multimorbidity patterns within a representative sample of UK older adults and their association with concurrent and subsequent memory.
Methods
Our sample consisted of 11,449 respondents (mean age at baseline was 65.02) from the English Longitudinal Study of Ageing (ELSA). We used fourteen health conditions and immediate and delayed recall scores (IMRC and DLRC) over 7 waves (14 years of follow up). Latent class analyses were performed to identify the multimorbidity patterns and linear mixed models were estimated to explore their association with their memory trajectories. Models were adjusted by socio-demographics, BMI and health behaviors.
Results
Results showed 8 classes: Class 1:Heart Disease/Stroke (26%), Class 2:Asthma/Lung Disease (16%), Class 3:Arthritis/Hypertension (13%), Class 4:Depression/Arthritis (12%), Class 5:Hypertension/Cataracts/Diabetes (10%), Class 6:Psychiatric Problems/Depression (10%), Class 7:Cancer (7%) and Class 8:Arthritis/Cataracts (6%). At baseline, Class 4 was found to have lower IMRC and DLRC scores and Class 5 in DLRC, compared to the no multimorbidity group (n=6380, 55.72% of total cohort). For both tasks, in unadjusted models, we found an accelerated decline in Classes 1, 3 and 8; and, for DLRC, also in Classes 2 and 5. However, it was fully attenuated after adjustments.
Conclusions
These findings suggest that individuals with certain combinations of health conditions are more likely to have lower levels of memory compared those with no multimorbidity and their memory scores tend to differ between combinations. Socio-demographics and health behaviours have a key role to understand who is more likely to be at risk of an accelerated decline.