Consideration of exposure data markedly increased the significance of the associations with ND, but AD. All of this was to be expected. Since our exposure criterion “have you ever had a drink?” excluded only five subjects, a notable effect on AD should not be expected. However, with respect to ND, more non-exposed subjects were eliminated, the effects were comparatively stronger. Interestingly, roughly equally significant associations were obtained using either model despite the fact Model 1 excluded the data from nearly half the subjects, while Model 2 excluded data from only 10% of the subjects. This may suggest that, with respect to already existing biobank datasets, that either ND model can be used depending on what information is available in the data set.
