Association between the estimated 10‐year cardiovascular mortality risk and cognitive function in older adults: results from the FINGER trial

Background Dementia risk scores could be instrumental in the implementation of multidomain lifestyle interventions for dementia prevention. However, as relatively new tools, their performance requires further validation. Given the extensive overlap between risk factors for dementia and cardiovascular disease (CVD), multidomain interventions usually have an important vascular component and better‐established cardiovascular risk scores could become useful in this context. More evidence is needed on their ability to predict the risk of cognitive decline and dementia. SCORE/SCORE‐Older People (OP) are the official tools of the European Society of Cardiology to estimate the 10‐year CVD‐mortality risk. In this study, we investigated the association between SCORE/SCORE‐OP and cognitive function within FINGER, the first large trial showing the efficacy of a multidomain lifestyle‐ and vascular‐based intervention in preventing cognitive decline. Method SCORE (age<65) and SCORE‐OP (age≥65) were calculated in the population of the 2‐year FINGER trial (N = 1236, baseline and 24 months). The cognitive outcomes (baseline, 12 and 24 months) were the overall z‐score from a Neuropsychological Test Battery (NTB, 14 tests, primary outcome) and the z‐scores of the NTB cognitive domains (Memory, Executive Function, and Processing Speed, secondary outcomes). The associations between (i) SCORE/SCORE‐OP and cognition, (ii) SCORE/SCORE‐OP and changes in cognition, and (iii) changes SCORE‐OP and changes in cognition were assessed using linear mixed‐models repeated‐measures with maximum likelihood estimation. Result Higher baseline risk of CVD‐mortality, estimated with SCORE/SCORE‐OP, was significantly and consistently associated with worse cognition, (NTB total composite: Estimate = ‐0.036; CI: ‐0.042 to ‐0.030; P‐value>0.0001; individual NTB cognitive domains: P‐values<0.0001) and yearly lower improvement in cognitive performance (NTB total composite: Estimate = ‐0.009, CI: ‐0.011 to ‐0.007, P‐value>0.0001; individual NTB cognitive domains: P‐values<0.0001). 2‐year change in SCORE‐OP was not significantly associated with changes in cognition during the FINGER study period. Conclusion SCORE/SCORE‐OP could be useful in predicting the risk of cognitive decline in a trial aimed at improving lifestyle and vascular‐related risk factors of dementia. However, more evidence is needed on their efficacy in predicting dementia risk and the response to multidomain lifestyle‐ and vascular‐ based interventions. Future studies on other well‐established and easy‐access cardiovascular risk scores could also provide relevant findings.