To the Editor:

Fitchett et al have presented supporting evidence for the ‘lower LDL-C is better’ concept.

Much of the proposed benefit is in events, however defined, and that may include silent heart attacks, hospital interventions and angina. For example, event benefit in ASCOT (1) included a 41% reduction in stable angina, a nonfatal pain but a pain that may well cause fewer hospital visits and catheterization laboratory-generated interventions. Such event reductions could therefore simply result from the nitroglycerin-mimicking action (promoting the NO-synthase pathway) of statins – and when nitroglycerin is the treating agent of choice.

Moreover, ‘event’ benefit may only be found in some groups, such as men before their sixties, while actual medical benefit, such as fewer myocardial infractions (MI) and fewer deaths, is never found in women and often not in others, for example, in the post age 70 patients who comprised PROSPER (references in 2), and when mortality was not improved and new cancers significantly increased (P=0.02).

The two main problems with the lower-is-better concept and the suggested 12% lowered mortality (in any patient) for 1 mmol/L LDL-C reduction after five years of statin use (3) are illustrated by the following:

• When comparing atorvastatin with placebo in ASCOT, we find that up to the median study end of 3.3 years the mortality curves simply do not separate, while the recently published SPARCL trial found five fewer deaths after 4.9 years in patients on a placebo than in those on top-dose atorvastatin (4). Ominously, the study discussed, TNT, ended with two more deaths on the top-dose versus the lowest-dose atorvastatin, the final nail in the coffin of any hope that atorvastatin, or increasing doses thereof, may possibly extend lives. Incidentally, ASCOT also ended with two fewer MIs, one fatal, in women taking a placebo.

• Patient outcomes also call to question the lower-is-better concept, when possibly the largest follow-up study in hypercholesterolemic patients (n=41,801, six years, taking simvastatin) finds doubled mortality in patients with the lowest LDL-C and quadrupled mortality in those with lowest total cholesterol (Figure 1 [5]). The drug cost would have been approximately 1/4 billion dollars. The authors suggest caution in ‘hyper-responders’ (to statin), a caveat hard to square with what Fitchett et al propose.

Figure 1).

All-cause deaths while on simvastatin for 6 years. Primary prevention in 41,801 hypercholesterolemic Japanese patients. Mean age at baseline 58 years, mean cholesterol at baseline 270 mg/dL (black scale) (7 mmol/L [white scale]). Sixty-eight per cent of study participants were women. From Table 5 in reference 5

In summary, when the top-selling statin, atorvastatin, proves not to save lives in high-risk patients, when two meta-analyses conclude that there is no mortality and/or MI benefit in women (reference in 2) and when the J-LIT patient follow-up study suggests harm, the lower-is-better concept must be re-examined, including all nonsupportive studies.

From my evaluation of the science, ‘lower’ may well turn out to be worse for overall health and survival in westernized populations, including Canadians, especially because the Quebec Cardiovascular Study (6) even failed to find cholesterol a risk factor in men.

One may wonder how the CTT meta-analysis (3) derived reported mortality benefit. Clearly, this is largely by including the massive HPS (7), but where the male, female, nondiabetic and diabetic patient Kaplan-Meier mortality curves still have not been released. Despite the lack of such data, the CTT authors imply a benefit in all, through statistical maneuvering and effect assumptions, even when not supported by the individual trials.

It is time to re-evaluate event-based medicine and the lower-is-better concept, when all-cause mortality is not improved in treating vascular diseases, and when statins have become a major strain on patients and systems.