http://www.cnn.com/2004/HEALTH/conditions/03/08/cholesterol.lowering.ap/index.html

The problem with this study is that all the effects of these drugs are
attributed to reduction in LDL.  What if the reduction in heart attack
levels (or whatever metric you'd like to use) is caused by the drug and
not the reduction in LDL?  Or what if it's caused by non-LDL effects (such
as reduction of inflamation)?  I guess we won't know because:

"the dose of Pravachol [the drug that didn't lower LDL as much] used in
the study costs about $900 a year, while Lipitor costs $1,400 [the drug
that lowered LDL quite a bit]"

Moreover, consider this:

"4,162 patients with new heart attacks or severe chest pain were randomly
assigned to the two drugs. After two years of follow up, 26 percent
getting the weaker Pravachol had died or experienced a variety of other
ill events, including new heart attacks, bypass surgery, rehospitalization
for chest pain or strokes. The same happened in 22 percent on Lipitor."

That's 166 people, a 4% improvement by using Lipitor.  But there's no
control group (at least none reported in the above-referenced
documentation).  What good is a study without a control group?