Updated– Last year, in what may have been an unprecedented action, a paper on the effects of Transcendental Mediation (TM) in African Americans was withdrawn by the editors only 12 minutes before the paper’s scheduled publication in Archives of Internal Medicine. No definitive explanation was ever provided, though the editors and authors said that the action was prompted by last minute questions from reviewers at the NIH, which helped fund the study. (The episode was covered in detail on CardioBrief (here and here) and on RetractionWatch.)
Now a new version of the paper has been published in Circulation: Cardiovascular Quality & Outcomes. The first author is Robert Schneider, from the Institute for Natural Medicine and Prevention at Maharishi University of Management in Maharishi Vedic City, Iowa. The co-authors are from the same institution and from the department of medicine at the Medical College of Wisconsin in Wisconin. It is clearly the same study of 201 African American patients randomized to TM or health education (HE) and followed for 5.4 years, though some of the numbers have changed in important ways between the earlier and later publications. One change involves the primary endpoint: the new paper in Circ:CVQ&O reports 52 primary endpoint events (the composite of death, MI, or stroke). Of these, 20 events occurred in the TM group and 32 in the HE group. By contrast, in the previous Archives version there were 51 primary endpoint events: 20 in the TM group and 31 in the control group.
In both papers the difference in the primary endpoint did not achieve statistical significance until after the investigators adjusted for baseline differences. The unadjusted hazard ratio (HR) for the new study was 0.64, (confidence interval 0.3701.12, p = 0.12); after adjusting for age, gender, and lipid lowering medications the HR drops, presto-chango, to 0.52 (0.29-0.92, p=0.025) and achieves statistical significance..
Here are some other key questions that I have asked AHA, Circulation editors, and other experts to address:
— Was the AHA or the Circulation staff aware at any time that an earlier version of this paper had been scheduled for publication in Archives of Internal Medicine and withdrawn only 12 minutes before the scheduled publication time? Should this information have been disclosed by the study authors when submitting the paper?
— As mentioned above, the original publication in Archives appeared to have been cancelled because of questions raised by NIH reviewers. Have the Circulation: Cardiovascular Quality & Outcomes editors received any assurance that these questions have been addressed in the new paper?
— This trial was started all the way back in 1998 but was not registered on ClinicalTrials.Gov until February 2011 (shortly before the intended Archives publication). Why was it not registered earlier than 2011?
— June 2007 is listed in ClinicalTrials.Gov as the final data collection date for primary outcome measure. Why are the primary endpoint numbers different between the Archives and new version of this paper?
Update, November 13-- Harlan Krumholz, the editor of Circulation: Cardiovascular Quality & Outcomes, sent the following statement:
“We had no prior knowledge of what transpired with the Archives of Internal Medicine. The Schneider paper went through rigorous peer review, statistical review and editorial discussions and the authors of the article were responsive to the review process. As a result, the paper was accepted for publication and we are going ahead as planned. If you have any further questions, we suggest you contact the researchers directly.”
Update, November 18– An earlier version of this article contained several mistakes. First, the difference in the number of primary endpoint events between the two papers is much smaller than I had stated. The original paper reported 51 primary endpoint events. The new paper reports 52 endpoint events. I am still unclear about why there should be any difference between the two, since, as mentioned above, June 2007 was the final data collection date for the primary endpoint. I also incorrectly claimed that some data reported in the original Archives paper were not included in the new paper. In fact, as a reader pointed out in the comments section below, these data were reported in Figure 1. I apologize for this mistake and regret the error. –LH
I asked Sanjay Kaul for his perspective on the paper. Here are his points:
1. All cause mortality constituted nearly 80% (41 of 52) of the primary composite endpoint. This is highly unusual as most trials yield a higher proportion of nonfatal events compared with fatal events. This leads me to question the fidelity of the adjudication process to pick up nonfatal events.
2. It is difficult to attribute a nearly 50% reduction in hard CV outcomes to a <5 mm Hg difference in systolic BP. Total anger was significantly reduced, but attributable risk of anger or emotional distress for MI or death is very small.
3. It is interesting to note that the study was originally powered for a 36% reduction in a broader endpoint, but a larger 50% reduction in the revised narrower endpoint. This is difficult to justify.
4. It is not clear if the adjusted analyses were adjusted for multiple comparisons. Given the wide CI and less than robust P value, it is likely the results would be no longer significant.
5. Even if one accepts the adjusted p value, the strength of evidence is not robust.
Disclosure: I work closely with the editor of Circulation: Cardiovascular Quality & Outcomes, Harlan Krumholz, on the CardioExchange website. Harlan has had no editorial control or input on this story, though I have contacted him, and other AHA officials and additional experts, for comment and perspective about this story.