The interaction of cognitive and brain reserve with frailty in the association with mortality: an observational cohort study

Abstract

Summary: Background: A higher cognitive reserve and brain reserve could decrease mortality risk, but the interaction of these factors with general age-related loss of physical fitness (eg, frailty) remains unclear with regards to mortality. We investigated the associations of cognitive and brain reserve with mortality and the interaction of cognitive and brain reserve with frailty within these associations. Methods: Within the observational population-based cohort of the Rotterdam Study, we included participants who visited the research centre for a cognitive assessment between March 2, 2009, and March 1, 2012. Participants with an incomplete assessment of cognition, no data on education attainment, no MRI or an MRI of insufficient quality, three or more missing frailty criteria, or a dementia diagnosis were excluded. Participants were followed up until their death or May 1, 2019. Cognitive reserve was defined as a latent variable that captures variance across five cognitive tests. Brain reserve was defined as the proportion of healthy-appearing brain volume relative to total intracranial volume measured with 1·5 Tesla MRI. Frailty was defined according to Fried's frailty phenotype; participants meeting at least one of the five criteria were considered frail. Hazard ratios (HRs) for associations of cognitive reserve, brain reserve, frailty, and reserve–frailty interactions with the risk of mortality were estimated using Cox regression models. Findings: 2878 individuals in the Rotterdam Study who visited the research centre for a cognitive assessment were considered eligible. 1388 individuals were excluded due to incomplete or missing data or a dementia diagnosis. 1490 participants with valid information on cognitive reserve, brain reserve, and frailty were included (mean age 74·3 years [SD 5·5]; 815 [55%] female participants). 810 (54%) participants were classified as frail. A higher cognitive reserve (HR 0·87 per SD, 95% CI 0·76–0·99, p=0·036) and a higher brain reserve (0·85 per SD, 0·72–1·00, p=0·048) were associated with a lower risk of mortality, after adjusting for sex, age, educational level, body-mass index, smoking status, and number of comorbidities. The association between cognitive reserve and mortality was more pronounced (0·77 per SD, 0·66–0·90, p=0·0012) when the cognitive reserve–frailty interaction (p=0·0078) was included, indicating that higher cognitive reserve is related to lower mortality in individuals with frailty. The brain reserve–frailty interaction was non-significant. Interpretation: Higher cognitive reserve and higher brain reserve were associated with a lower mortality risk. Additionally, cognitive reserve and frailty interact in the association with mortality, such that higher cognitive reserve is particularly associated with lower mortality in frail participants. Funding: Netherlands Organization for Health Research and Development and EU Horizon 2020 research programme.