Friday, April 21, 2006

Silliness in the New England Journal of Medicine or something worse?

I'm rather confused as to how this study in the New England Journal of Medicine got published.

The title is Feasibility of Treating Prehypertension with an Angiotensin-Receptor Blocker. Its purpose was to determine whether it is "feasible" to use a blood pressure medicine (candesartan, brand name Atacand, manufactured by AstraZeneca) to treat prehypertension. This is a condition designated by the JNC VII (Joint Committee on prevention, detection, evaluation, and treatment of high blood pressure) describing adults with a systolic blood pressure (the high number) from 120 to 139 and a diastolic blood pressure (the low number) from 80 to 89.

Before explaining the silliness of this study, a little background is in order.

The JNC VII described this entity because of data suggesting that patients with prehypertension are at increased risk of cardiovascular disease compared to those with completely normal blood pressure (less than 120/80).

This new "disease" was itself controversial in 2003 because it involved labeling a massive number of patients as having an abnormal blood pressure that were previously felt to be normal. According to one survey, 27% of the adult U.S. population has true hypertension. The prehypertension designation labeled another 31% of previously healthy individuals as being abnormal. Imagine the uneasiness that these patients, previously thinking that they were perfectly healthy, might now be experiencing.

Now imagine that you're the CEO of a pharmaceutical firm producing blood pressure medications looking for new opportunities to expand your market. Overnight, you are presented with a brand new patient population consisting of 31% of the nation's adult population.


The JNC VII, not finding evidence that treating such prehypertension with medications lowered the risk of cardiovascular events simply recommended lifestyle changes such as diet and exercise. However, just because such evidence is not currently available doesn't mean that it's not true. It would therefore be of tremendous economic advantage to attempt to gather such evidence.

Did the new article attempt to do that? The study was set up as follows: A group of patients with prehypertension were given placebo and a similar group was given candesartan. 16% fewer patients in the candesartan group went on to develop true hypertension than in the control group during the four year study period.

My feeling is -- who cares? All this proves is the rather unprovocative conclusion that such therapy delays a percentage of patients from getting to a point where medical therapy is shown to be useful. (I'm actually surprised that the medication didn't prevent almost all of the study group from going on to develop prehypertension.) It still doesn't answer the truly relevant question of whether actually treating prehypertension actually benefits anyone.

The fact that along with the study's results, the candesartan group had a similar incidence of side-effects as the placebo group led the investigators to optimistically conclude that

"Treatment of prehypertension with candesartan appeared to be well tolerated and reduced the risk of incident hypertension during the study period. Thus, treatment of prehypertension appears to be feasible."

Nowhere did they look at clinical outcomes such as death, heart attack, stroke, etc. Nowhere did they establish the clinical relevance of such treatment or of their results.

What is even more intriguing to me is the complete lack of a reason to perform this study in the first place. While not insignificant, this essentially amounted to a pilot study. As such, it doesn't seem appropriate for publication in a journal with the reputation of the NEJM.

Something else that I find unfathomable is that if all you're trying to establish is the feasibility of treating prehypertension (perhaps with an eye towards later proving that it improves outcomes as well), why use an expensive proprietary drug such as candesartan? Wouldn't it make more sense to perform the study using a cheap generic drug that was proven to be as efficacious in preventing adverse outcomes as more expensive drugs?

If I were going to design such a study, I'd look at the very large ALLHAT trial which found that one of the cheapest antihypertensive drugs, a diuretic called chlorthalidone, was just as effective as more expensive drugs at preventing such bad outcomes (and in some subsets of patients more effective).

It is puzzling that such this "feasibility" study selected a drug that costs almost two dollars a pill instead of one that costs about 12 cents a pill. Perhaps an answer to this question lies in the authors' potential conflict of interest declaration at the end of the article:

Dr. Julius reports having served as a consultant to Novartis and Servier, and having received lecture fees from Novartis and Merck and grant support from AstraZeneca and Novartis. Dr. Nesbitt reports having served as a consultant to AstraZeneca, Novartis, and Pfizer, and having received lecture fees from Boehringer Ingelheim, Pfizer, and Novartis and grant support from AstraZeneca. Dr. Egan reports having served as a consultant to AstraZeneca, Novartis, Pfizer, and Merck; having received lecture fees from Boehringer Ingelheim, Novartis, and Pfizer and grant support from Pfizer, Novartis, and AstraZeneca; and having received royalties from Elsevier. Dr. Weber reports having served as a consultant to Novartis, Pfizer, Merck, and Sankyo, and having received lecture fees from Novartis, Sankyo, Bristol-Myers Squibb, Pfizer, Merck, and Sanofi-Aventis. Dr. Grimm reports having served as a consultant to Pfizer and Merck, and having received lecture fees from Merck, Pfizer, and Novartis. Dr. Michelson is an employee of AstraZeneca. Dr. Black reports having served as a consultant to MSD, Myogen, Pfizer, Novartis, and BMS/Sanofi, and having received lecture fees from Pfizer, BMS/Sanofi, Boehringer Ingelheim, and Novartis and grant support from Pfizer. Dr. Oparil reports having served as a consultant to Merck, Novartis, Pfizer, Salt Institute, and Sankyo Pharma; having received lecture fees from Bristol-Myers Squibb, Merck, Pfizer, and Sankyo Pharma; having received grant support from Abbott Laboratories, AstraZeneca, Aventis, Biovail, Boehringer Ingelheim, Bristol-Myers Squibb, Forest Laboratories, GlaxoSmithKline, Novartis, Merck, Pfizer, Sankyo Pharma, Sanofi-Synthelabo, and Schering-Plough; and reports being a member of the board of directors of Encysive Pharmaceuticals.

Kind of long huh?

So why did the NEJM even publish this dog? Could it be that some people in high places are acquaintances with some other people in high places?

Don't get me wrong. I'm basically a free market kind of guy. But I would still like to see my discipline's most prestigious journal practice a higher level of oversight than may have been exercised here.


Blogger james gaulte said...

I also was also surprised that this article appeared in the NEJM.I would not have been had it appeared in one of the so called "throw a way" journals.You would think after the seemingly high road editorial approach the editors took regarding the Vioxx issue they would be more cautious in publishing anything that even hinted at drug promotion.

April 21, 2006 4:50 AM  
Blogger Margaret Polaneczky, MD (aka TBTAM) said...

Great post. I was truly shocked at this study when I read it. I think big pharma is getting itself into places we never thought they could before. Or maybe it's always been this way, and I just never saw it before.

April 25, 2006 7:29 PM  
Anonymous Anonymous said...

while it is true that disease dismongering to further drug companies' economic interests is bad, but look at the merits of the idea. if treating people at risk of developing hypertension reduces that risk, i do not see a problem with that. potentially, it can be life saving for thousands of people, because as we know hypertension is a risk factor for heart disease and stroke. however, it will have to be a personal choice whether or not to take candersartan on the basis of this study.

Also, they did not look at 'real' outcomes such as death, heart attacks, etc, because it's just a 2year study, and in a population of young and relatively healthy people, to get a statistically significant result in those outcomes would require many more times the number of participants.

April 29, 2006 9:29 AM  
Anonymous Anonymous said...

hello m edwards ballinger texas

May 17, 2006 3:11 PM  
Anonymous Anonymous said...


JP Kassirer, M.D.
Tufts University School of Medicine

December 18, 2006 10:10 AM  

Post a Comment

<< Home