Cognition is multidimensional, and each domain plays a unique and crucial part in successful daily life engagement. However, less attention has been paid to multi-domain cognitive health for the elderly, and the role of lifestyle factors in each domain remains unclear.

We conducted a cross-sectional study of 3,230 older adults aged 60+ years in Xiamen, China, in 2016. The Montreal Cognitive Assessment (MoCA) was used to measure general cognition and six specific sub-domains. To account for educational effects, we adjusted the MoCA score and divided respondents into three education-specific groups (low, moderate, and high education groups with ≤5, 6~8, and ≥9 years of education, respectively). A series of proportional odds models were used to detect the associations between two categories of lifestyle factors – substance abuse (cigarette and alcohol) and leisure activity (TV watching, reading, smartphone use, social activity, and exercise) – and general cognition and the six sub-domains in those three groups.

Among the 3,230 respondents, 2,617 eligible participants were included with a mean age of 69.05 ± 7.07 years. Previous or current smoking/drinking was not associated with MoCA scores in the whole population, but unexpectedly, the ex-smokers in the low education group performed better in general cognition (OR = 2.22) and attention (OR = 2.05) than their never-smoking counterparts. Modest TV watching, reading, and smartphone use also contributed to better cognition among elderly participants in the low education group. For the highly educated elderly, comparatively longer reading (>3.5 hours/week) was inversely associated with general cognition (OR = 0.53), memory (OR = 0.59), and language (OR = 0.54), while adequate exercise (5~7 days/week) was positively related to these factors with OR = 1.48, OR = 1.49, and OR = 1.53, respectively. For the moderately educated elderly, only modest reading was significantly beneficial.

Lifestyle factors play different roles in multidimensional cognitive health in different educational groups, indicating that individual intervention strategies should be designed according to specific educational groups and different cognitive sub-domains.