Well, it could still be that in elderly people, higher cholesterol levels are associated with lower mortality, but in younger people, higher levels are associated with higher mortality.

There is precedent for this: BMI. In younger populations, higher BMI is associated with higher mortality, especially in recent analyses. Yet, higher BMI is associated with lower mortality in the elderly.

This may be because the sorts of things that kill elderly people are different than those that kill younger people, so the cost-benefit balance for any given biological attribute shifts as one ages.

Regardless of the reason why, the fact that cholesterol does not seem to be correlated with higher mortality in the elderly does not rule out cholesterol as an important entity in the causal chain that results in CVD or other chronic diseases. All it means is that, in the elderly, the negative effects of cholesterol, if any, may be counterbalanced by positive effects — or at least were in the population studied.

The data say what the data say…Moving beyond that point is a matter of interpretation, not certainty.