Thursday, July 16, 2009

New Study Shows that At Least Two-Thirds of Patients Receiving Placebo in "Double-Blind" NRT Trials Know they are Receiving Placebo

Finding Casts Doubt on Conclusions Regarding Effectiveness of Nicotine Replacement Therapy

NOTE: This is a revised version of the original post. Corrections made based on feedback from the lead study author are listed here.

A study published online ahead of print in the journal Nicotine and Tobacco Research reports that two-thirds of placebo subjects in a double-blind, placebo-controlled trial of precessation nicotine patch therapy were aware that they were receiving placebo, rather than nicotine. Of the 165 patients receiving placebo patches, 112 believed they had not received an active patch. Overall, participants guessed whether they had received nicotine or placebo at a rate significantly higher than what would have been expected by chance alone (p<0.0001).

The Rest of the Story

Keep in mind, first of all, that this was a trial of precessation nicotine replacement therapy. Thus, subjects were still smoking when they received the patch. If the proportion of subjects receiving placebo who knew they were receiving a dud patch was 67% when they were still allowed to smoke, imagine the rate when subjects were not smoking at all.

It is clear that the overwhelming majority of patients who receive placebo in "blinded" NRT trials are not blinded as to their status.

The significance of this finding is that it casts serious doubt upon the conclusions of many previous studies which have purported to show that NRT products are more effective than placebo in "double-blinded" trials.

One of the key facets of a clinical trial of the effectiveness of nicotine replacement therapy compared to placebo is that the subjects need to be blinded as to their treatment status. The finding I am reporting today demonstrates that - most likely - all of the "double-blinded" nicotine replacement therapy trials have actually not been blinded studies and therefore, their conclusions are cast into doubt.

A review of NRT trials revealed that only 17 of 73 published articles assessed whether subjects were successfully blinded as to their treatment status and that in 12 of the 17 studies, subjects guessed their status with greater than chance accuracy. The article concluded that we presently do not even know enough about the assessment of blinding success, much less about whether this problem of unsuccessful blinding negates the conclusions of the existing studies.

This blinding problem is particularly problematic given the financial conflicts of interest of the investigators drawing conclusions about the effectiveness of nicotine replacement therapy. That the expert panel which concluded that NRT is effective and recommended its use in all individuals trying to quit ignored the blinding problem is perhaps not surprising given the fact that the chair of the panel and seven of its members had significant financial conflicts of interest with Big Pharma.

In the present article, despite the finding - reported in the article's results section - that blinding was dismally unsuccessful - the paper does not even mention this problem in its discussion section. The paper concludes that NRT was effective even though it has demonstrated that the study was not blinded!

According to the paper, the lead author does not have any present financial interest in nicotine replacement therapy, but has in the past received royalties from nicotine patch sales. The conflict of interest statement notes that the lead author: "has received royalties from sales of certain nicotine patches and is named as inventor on nicotine skin patch patents that expired in 2008."

While there is no conflict of interest at present because the patent has expired, the lead author of the study has had a substantial past financial interest.

If you look at the data in the paper, you'll see that the entire effect of the nicotine patch occurred within the first week of attempted cessation. There were no differences in subjects' success for the rest of the study (in fact, placebo patients appear to have done a little bit better).

Also of note, the point prevalence of abstinence showed no effect of the nicotine patch, other than at 1 week. There was no effect of the patch on point abstinence at 10 weeks postquit and no effect of the patch on point abstinence at 6 months (although there was an effect for patients whose withdrawal symptoms were relieved by the precessation patch).

This suggests that a large proportion of the subjects receiving placebo immediately realized they were not receiving nicotine, were discouraged, and gave up right away. Some of them later overcome their initial frustration, but almost the entire observed effect of the study appears to be attributable to this initial frustration on the part of a large percentage of placebo-receiving subjects who realized that they had been given a dud patch.

The rest of the story is that the entire literature on the effectiveness of nicotine replacement therapy has been thrown into serious question. And the effects of financial conflicts of interest with Big Pharma are readily apparent in the biased conclusions of the most important review of the literature (the NIH expert panel and its clinical practice guideline for smoking cessation) and its failure to address the blinding problem.

(Thanks to John Polito for the tip. To see one of John's excellent commentaries on this issue, please click here for his recent letter published in the Canadian Medical Association Journal).

No comments: