Rossi and colleagues report no association between elevated lipoprotein(a) [Lp(a)] and incident dementia in a large propensity-score-matched TriNetX cohort [1]. The study is valuable because Lp(a) is a causal cardiovascular risk factor and vascular prevention is central to dementia care [2,3]. We suggest, however, that the matched analysis may estimate a conditional direct association rather than the total clinical effect of lifelong Lp(a) exposure, and that limited at-risk time may further shape the null finding.
