To the Editor: The ALLHAT-LLT authors concluded
that "pravastatin did not reduce either all-cause mortality or CHD significantly."1 This conclusion is misleading because it suggests
that, despite considerable prior evidence that statins reduce CHD events,
statins are not effective in this population, with this drug, or in this trial.
In fact, the data show that the effect of pravastatin on CHD events
was exactly as expected from previous data.
Figure 1 presents data from Table 5 of the ALLHAT-LLT article. The
regression line is plotted from the 8 previous statin trials without the ALLHAT-LLT
data. I have plotted percentage change in cholesterol as the independent variable
and the odds of cardiovascular events as the dependent variable. The ALLHAT-LLT
investigators found that pravastatin produced a 0.91 reduction in the risk
of CHD events and a difference in total cholesterol of 9.6%. The reduction
in CHD events expected from the regression equation, for that difference in
cholesterol, is almost identical at 0.906. Of course, unless there is a threshold
effect, the "real" relationship would pass through the origin (0,1) where
no change in cholesterol would produce no change in events. If so, either
the effect of cholesterol lowering on CHD events is nonlinear (greater with
greater changes in cholesterol) or the relationship calculated from the available
data does not reflect the truth.
Lipid-Lowering Therapy and Risk of Coronary Events. JAMA. 2003;289(16):2070. doi:10.1001/jama.289.16.2070-b