Objectives: Understanding how the importance of modifiable risk factors for dementia varies by cognitive status and sex is vital for the development of effective approaches to dementia prevention. We aimed to calculate population attributable fractions (PAFs) for incident dementia associated with sets of risk factors while exploring sex differences in individuals who are cognitively normal (CN) or has mild cognitive impairment (MCI).

Methods: Longitudinal data from the Rush University Memory and Aging Project (MAP) were analysed. Included participants were aged over 50 years and were CN or with a diagnosis of MCI at their baseline assessment. Analyses considered fifteen potential dementia risk factors covering cardiometabolic, lifestyle, psychosocial and sensory domains. We used Cox proportional hazard models to estimate the hazard ratios for incident dementia associated with dementia risk factors and calculated weighted PAFs. All analyses were repeated stratified by sex.

Results: The analytical sample comprised 754 cognitively normal participants (77.2% female) and 242 participants with a diagnosis of MCI (71.9% female), of whom 214 (28.4%) and 120 (49.6%) were diagnosed with dementia across the follow-up, respectively. Although the weighted overall PAF was similar for CN (24.7%) and MCI (25.2%) subgroups, sex differences were present in both. Compared to in females, PAFs were higher in males in both CN (42.5% vs. 25.1%) and MCI (51.6% vs 12.3%) subgroups. The profiles of contributing risk factors also varied by sex. In males, the highest PAFs were smoking (11.1%), vision impairment (6.2%) and stroke (6.0%) in CN and smoking (13.3%), physical inactivity (12.9%) and heart attack (7.9%) in MCI. In females, the highest PAFs were unmarried marital status (4.9%), depression (4.1%) and social isolation (3.8%) in CN and vision impairment (4.4%), increased alcohol intake (3.5%) and depression (2.6%) in MCI.

Conclusions: These findings support the notion that dementia risk is modifiable after the onset of MCI. They also highlight the potential benefits of considering an individual’s cognitive status and sex when formulating dementia prevention strategies.