To the Editor.—
I am interested in the fall in cholesterol level over time and its positive association with overall and cardiovascular mortality as reported by Anderson et al.1 As I understand it, individual least squares regression lines were computed for each individual having at least five cholesterol measurements. Positive and negative slopes as binary variables relative to no change were entered into a model that included age and sex. Baseline cholesterol values of individuals with positive as opposed to those with negative slopes were not reported.I am concerned that individuals with negative slopes would have higher baseline cholesterol values due to the phenomenon of regression to the mean. The positive association with mortality would then be attributable to the higher baseline cholesterol values rather than the fall in cholesterol level.2,3Why the converse finding—ie, an inverse relationship between an increase in cholesterol level and mortality—is not
Hainer JW. Cholesterol and Mortality. JAMA. 1987;258(12):1608–1609. doi:10.1001/jama.1987.03400120058022
Customize your JAMA Network experience by selecting one or more topics from the list below.