Investigating Weight Variability and Cognition

In addition to these overall analyses, we performed sensitivity analyses to investigate the robustness of the associations considering different subgroups of the population. First, we performed separate analyses for placebo and pravastatin treatment groups. Next, it has been shown that high blood pressure variability was associated with worse cognitive function [10 (link)]. As variability in weight and blood pressure may have a common cause, we additionally adjusted our models for systolic blood pressure variability and mean systolic blood pressure from baseline to month 30. This allows for the separation of effects of weight variability from those originating from blood pressure variability. Systolic blood pressure variability was defined as the intraindividual SD from baseline to month 30, where blood pressure was measured every 3 months, as previously done [10 (link), 17 (link)]. Furthermore, we included both weight loss (slope) and visit-to-visit body weight variables in the same multivariable-adjusted linear regression model to test independence of the two phenotypes. Last, we performed analyses excluding individuals who developed any of the following diseases during follow-up, to ensure weight loss did not follow as a result: incident diabetes, non-fatal cancer, non-fatal stroke or TIA, hospitalisation because of heart failure and non-fatal coronary or cardiovascular events.