Sunday, November 09, 2008

A Massive Experiment Doesn't Dent the Conventional Wisdom

The conventional wisdom is that powerful pharmaceuticals (which can kill) are not to be used too much, meaning without extensive diagnostic work-ups. Certainly, the idea is that powerful pharmaceuticals must not be given to normal healthy people even in an experimenta trial.

Massive overwhelming evidence from a huge experiment shows
dramatic benefits of statins for people who do not have high blood cholesterol but do have "high levels of a protein called high-sensitivity C-reactive protein, or CRP, which indicates inflammation in the body."

The study, presented Sunday at an American Heart Association convention in New Orleans and published online in The New England Journal of Medicine, found that the risk of heart attack was more than cut in half for people who took statins.

Those people were also almost 50 percent less likely to suffer a stroke or need angioplasty or bypass surgery, and they were 20 percent less likely to die. The statin was considered so beneficial that an independent safety monitoring board stopped what was supposed to be a five-year trial last March after less than two years.

However, the consensus of the profession appears to be pretty much to ignore the study

The study is sparking debate over who should take a blood test to check CRP and under what circumstances someone with high CRP should be given a statin. Because heart disease is a complex illness affected by many risk factors — including smoking, hypertension, being overweight and having a family history of heart disease — most researchers said high CRP alone should not justify prescribing statins to people who have never had heart problems.

Some experts cautioned against testing people for CRP unless they had other indications of being at risk for heart disease, and they said more research was needed to pinpoint the patients for whom the benefit of statins outweighs the risks. Others recommended testing more frequently and using statins for people with low cholesterol if they have high CRP and some other risk factors.

Now lets understand. The study was stopped early, because it was considered un-ethical to give placebos to some of the patients.

Now it might make sense to attempt further study to find if there are people with elevated CRP would benefit from not getting statins. Given the results to date, it seems to be at least arguable that a study to determine if placebo works better than statin for some such patients would be absolutely immoral and unethical even if they gave informed consent. It seems to me insane to assume the results of such a study and give the recommendation for treatment "no statins if the patients just fit the criteria of the study which showed a 20% decline in mortality with statins but don't have other characteristics."

An experiment with a new alternative drug called "placebo" would almost certainly not be approved even if patients were told of the results of the study and signed a form swearing that they wanted to give placebo a trial even though they fit the characteristics of a group in a trial where placebo treatment implied a 20% increase in death rates.

Just doing it without explaining anything to the patients is fine however.

Some decades from now I want a calculation of how many lives were saved or lost, because people with high CRP but not the other characteristics were not prescibed statins and I want the un-named advocates of that approach to be named.

In exchange I would be glad to promise that if their advice doesn't cause a positive number of human deaths I will surrender my left testicle.

The attitude of critics of the alleged overuse of pharmaceuticals (Sidney Wolfe allowed his name to be used) is to me indistinguishable from that of global warming deniers. They recommend not doing something which the evidence, such as it is, supports because there should be more studies.

The article by PAM BELLUCK stresses the risks of statins and the fact that subgroups within the study might have benefits less than those risks. That would principally be a risk of death of one in one million. How excellent would the other risk factors have to be to identify a group with high CRP who would have longer life expectancy without Statins ? Has any such differential in the benefits of statins ever been found (including in animal studies)? Ask an expert.

Oh and what about giving statins to people who are perfectly normal and healthy in every way. Are their benefits ? Do they outweigh the risks ? We have no 0 (zero) evidence on this, because no such experiment has been performed or would ever be allowed. I don't know about animal studies. I certainly think that a study of the effect of statins on life expectancy in normal mammals would be interesting.


pj said...

You might want to read this blog post about this latest statin trial.

Although the relative risk reduction sounds high, the absolute risk reduction is low because the underlying risk of MI, stroke, or particularly death, is quite low for those with isolated elevation in CRP.

So you need to treat hundreds of people with statins to prevent one person getting an MI or stroke. So then you must trade off the cost and side-effects of treating all these people against preventing that one event.

Robert said...

Low and lower than the costs of side effects are different. The principal cost (as far as I know) is rhabdomiolosis (sp?) with a one in a million risk of death. One in hundreds (or thousands) is very good compared to one in a million.

The social cost of statins is the cost of manufacturing them, which is low. If we don't want to make pharmaceutical companies even more profitable than they are, that can be managed.

pj said...

Rhabdomyolysis is the very rare but scary side effect, myositis is more common but not as severe, but very many people get other side effects like muscle pains. There's still a question over the relationship with cancer rates.

Costs include testing people for CRP levels (not cheap if you're proposing a screening programme) or alternatively the massive cost of prescribing for an entire nation and the deaths and side-effects that will develop in people who weren't going to benefit from the statins in the first place (we don't know the prevalence of raised CRP in the population).

These sorts of questions are faced all the time in, for example, questions of the effect of the oral contraceptive pill on (low) rates of blood clots or breast cancer in young women.

I know the US is different, but in the UK we have systems in place to assess whether treatments and screening programmes are cost-effective - and it isn't necessarily an easy question.