[HSF] More on the Traysylol saga

Wed Oct 4 18:51:46 EDT 2006

This is the problem with sophisticated statistical techniques - 
Multivariate analysis, propensity,  scoring, "metanalysis"etc etc which 
99.9 % of readers just don't understand/verify.
Unfortunately these are used to surmise and answer problems which can 
never be more accurate than the "worst estimate" and yet are being 
propounded as the "ultimate" in statistics. The discussion paper in the 
Mangano paper was more of an "axe to grind" type of writing rather than 
a truly "temperate" scientific sermon. The worst thing is when lay press 
gets access and worse still when publishers and authors parade their 
results.
This is why we need to strictly curb medical advertising but 
unfortunately I see this on the rise.  You then start getting back from 
what you dole out and in this case unfortunate publicity , litigation 
and a whole lot of heart burn.
Prasanna
Ani Anyanwu wrote:
> Well this confirms what several of us have maintained for the last 6 months. I must say I was amazed at the response from the surgical committee to stop using a drug based on a paper they never read - I say never read because anyone who did read the paper would have immediately discounted its findings because of glaring study flaws. Indeed when I read the paper, I only read the methods section and because the methods were so flawed, I never wasted my time reading the results or discussion. Of course arguing this point was difficult because I would be reminded "it was in the NEJM...".
>
> The morale of the surgery is not to base our practice on what is in the mass media or on rumors but to study the data ourselves. Like we often quote the Loop 1986 NEJM paper as evidence for benefit of IMA grafting, but I bet most have never read the paper; anyone who has would agree that it is anything but evidence, also a very flawed study that would not be publishable today. Indeed one of my mentors maintains that there is no definitive evidence that IMA is superior to VG for grafts to the LAD - I must reluctantly agree he is telling the truth.
>
> Hopefully the only lawsuit that will come out of this will be against the multimillion dollar establishment called the NEJM for misleading the medical fraternity, causing harm to thousands of patients who were denied the drug and causing millions of dollars of loss to Bayer.
>
> For people unconvinced about the dubious business of medical publishing of which NEJM is at the forefront, see
>
>  Richard Smith  The highly profitable but unethical business of publishing medical research
> J. R. Soc. Med. 2006 99: 452-456. 
>
> Ani
>   ----- Original Message ----- 
>   From: Tdmartin2000 at aol.com<mailto:Tdmartin2000 at aol.com> 
>   To: OpenHeart-L at lists.hsforum.com<mailto:OpenHeart-L at lists.hsforum.com> 
>   Sent: Tuesday, October 03, 2006 11:00 PM
>   Subject: Re: [HSF] More on the Traysylol saga
>
>
>   FDA committee reassures on aprotinin safety 
>   from Heartwire - a professional news service of WebMD 
>
>
>   Sue Hughes 
>     
>   September 27, 2006Rockville, MD - At a meeting of the US FDA 
>   Cardiovascular and Renal Drugs Advisory Committee last week, the panel 
>   supported the overall safety and efficacy of the antifibrinolytic drug 
>   aprotinin (Trasylol, Bayer) if used according to its labeled 
>   indication--to reduce bleeding complications in high-risk CABG patients. 
>
>
>   The September 21, 2006 meeting was held to discuss safety concerns 
>   about the drug raised by two observational studies published earlier 
>   this year. 
>
>
>
>   The first study, by Mangano et al, published in the New England Journal 
>   of Medicine, showed that CABG patients who received aprotinin had higher 
>   rates of renal failure, MI, and stroke compared with those treated with 
>   other drugs to prevent bleeding or with no treatment [1]. The second 
>   study, by Karkouti et al, published in Transfusion, reported more cases 
>   of reduced kidney function in patients treated with aprotinin compared 
>   with another treatment to prevent bleeding, but no excess risk of 
>   cardiovascular events [2]. 
>
>
>
>   The cardiorenal committee reviewed both these studies at the meeting, 
>   along with the randomized data from clinical trials supplied by Bayer. 
>   Chair of the committee, Dr William Hiatt (University of Colorado Health 
>   Sciences Center, Denver), told heartwire that the main concern about 
>   increased cardiac events was seen only in the study by Mangano et al, 
>   but because of questions about some of the methodologies used in this 
>   study and the fact that the data from this study had not been 
>   independently reviewed by the FDA, the committee could not endorse its 
>   findings. 
>
>
>
>   The suggestion of a slight reduction in renal function, however, was 
>   seen in all three data sets and is probably real, Hiatt noted. "Both 
>   observational studies showed an increase in renal creatinine, and this 
>   was also seen in some of Bayer's data, so there is a signal for some 
>   reversible renal insufficiency across the data," he commented. 
>
>
>
>   At the end of the session, the committee voted (18 in favor, none 
>   against, and one abstention) that the totality of clinical data 
>   presented supports acceptable safety and efficacy for aprotinin among 
>   CABG patients. 
>
>
>
>   Mangano study flawed? 
>
>
>
>   Elaborating on the concerns voiced about the Mangano study, Hiatt 
>   explained that several committee members were of the opinion that 
>   Mangano et al did not conduct the propensity score properly and 
>   therefore the adjustment for confounding factors could have been flawed. 
>   Added to this, Mangano apparently refused to allow the FDA unrestricted 
>   access to his data, imposing restrictions on what the agency could and 
>   could not see. "Because there was no independent review of his data, the 
>   committee could not substantiate his claims," Hiatt told heartwire. 
>   "Mangano et al concluded in their paper that the use of aprotinin was 
>   not advised based on their data, but we could not condone this and some 
>   members of the panel said this was an irresponsible statement," he 
>   added. 
>
>
>
>   Indeed, in a somewhat unusual turn of events, one member of Mangano's 
>   original study group, Dr Linda Shore-Lesserson (Montefiore Medical 
>   Center, Bronx, New York), spoke at the advisory committee meeting and 
>   told how she had actually removed her name from the NEJM paper because 
>   she believed the conclusions had been overstated. 
>
>
>
>   Hiatt said he would like to see more data on aprotinin, but it was 
>   unlikely that any more placebo-controlled trials would be conducted. 
>   "More observational data are essential, but they must be analyzed 
>   properly," he stressed. He added that the panel has no problems with the 
>   Karkouti study, which, despite also being observational, was more 
>   credible than the Mangano study because transparency and independent 
>   review of the primary data by the FDA were allowed to occur. 
>   Other panel members agreed with this view. Dr Robert Harrington (Duke 
>   University, Durham, NC) commented to heartwire: "The Karkouti study was 
>   well done from a methodological view, and while it did raise questions 
>   about an association with renal toxicity, the authors were appropriately 
>   cautious about their conclusions, given that it was an observational 
>   study. But the Mangano study raised a lot more discussion about the 
>   statistical methods used, and the panel did not like the fact that 
>   Mangano had not allowed the FDA unrestricted access to his data." 
>
>
>
>   FDA offered "chaperoned" view of data 
>
>
>
>   An FDA spokesperson confirmed that the agency had been offered a 
>   "chaperoned" view of the Mangano data, which in its view was 
>   insufficient. Some members of the advisory committee were vocal in their 
>   criticism of Mangano for this, with one pointing out that "science does 
>   not get done by hiding your data." 
>
>
>
>   Harrington said the take-home message for him from the meeting was that 
>   aprotinin does appear to do what it is supposed to do--reduce bleeding 
>   in CABG patients, with the best effect in highest-risk patents. And 
>   although observational data are important, possible flaws in the Mangano 
>   study mean that its conclusions may not be reliable. But Harrington also 
>   pointed out that while concerns about cardiac events had not been seen 
>   in Bayer's randomized trials, these studies were quite dated, and new 
>   data from contemporary studies were needed. 
>
>
>
>   He also noted that the panel did question whether, if aprotinin is so 
>   good at reducing bleeding and transfusions after CABG, why a mortality 
>   reduction has not been seen. And it was suggested that there were not 
>   enough data at this point to be able to conclude anything about 
>   mortality. 
>
>
>
>   Why did the NEJM publish Mangano's study? 
>
>
>
>   There was also much discussion as to why the New England Journal of 
>   Medicine published Mangano's study, given that several statisticians 
>   present at the meeting had questioned the statistical methodology used, 
>   and why the journal had allowed Mangano to use such emotional language 
>   in his discussion section. The consumer representative at the meeting 
>   made the comment: "Where were all the adults at the NEJM when this paper 
>   was going through peer review?" 
>
>
>
>   Harrington said that if Mangano had written the study up as an 
>   observational study with all the caveats that that entailed, it would 
>   have been better accepted. "But instead, he claimed that the rigor of 
>   his data collection--because he adjusted for so many 
>   variables--essentially mimicked randomization. But the panel did not 
>   agree with this. There is a reason why we do randomized trials," 
>   Harrington added. 
>
>
>
>   Another panel member, Dr Michael Lincoff (Cleveland Clinic, OH), 
>   defended the journal. "It's always easy to pick up flaws in a paper in 
>   retrospect. The peer-review process may not have had as much statistical 
>   expertise as the FDA panels and invited experts. The good thing about 
>   science is that it is self correcting--the truth comes out in the end." 
>   But Lincoff agreed that the discussion section of the Mangano paper 
>   could have been toned down somewhat. "The NEJM did allow a very dogmatic 
>   discussion section, with some very strong comments drawn from some 
>   really quite weak data," he commented to heartwire. 
>
>
>   One of the statisticians on the panel, Dr David Demets (University of 
>   Wisconsin, Madison, WI), declined to comment on the meeting other than 
>   to say that the discussion was "fascinating and has implications for 
>   future discussions about observational data relative to safety issues." 
>
>
>
>   Requests by heartwire for comment from Mangano received no response. 
>
>
>
>   Bayer notes that it is in discussion with the FDA regarding possible 
>   updates to the US label for aprotinin to include additional detail 
>   specifying that aprotinin should be used in patients at increased risk 
>   for blood loss and blood transfusion and revised guidance with regard to 
>   elevations in serum creatinine levels and hypersensitivity reactions. 
>
>
>
>   Mangano DT, Tudor IC, Dietzel C. The risk associated with aprotinin in 
>   cardiac surgery. N Engl J Med 2006; 354:353-65. 
>   Karkouti K, Beattie W, Dattilo K, et al. A propensity score 
>   case-control comparison of aprotinin and tranexamic acid in 
>   high-transfusion-risk cardiac surgery. Transfusion 2006; 46:327-338. 
>
>
>   The complete contents of Heartwire, a professional news service of 
>   WebMD, can be found at www.theheart.org<http://www.theheart.org/>, a Web site for cardiovascular 
>   healthcare professionals. 
>   _______________________________________________
>   OpenHeart-L mailing list
>
>   Send postings to:
>    OpenHeart-L at lists.hsforum.com<mailto:OpenHeart-L at lists.hsforum.com>
>
>   To UNSUBSCRIBE, to CHANGE email address, or to view archives:
>   http://mmp.cjp.com/mailman/listinfo/openheart-l<http://mmp.cjp.com/mailman/listinfo/openheart-l>
>
>   All messages transmitted by the OpenHeart-L are subject to the policies and 
>   disclaimers posted at:
>   http://www.hsforum.com/listdisclaim<http://www.hsforum.com/listdisclaim>
>   -----------------------------------------
> _______________________________________________
> OpenHeart-L mailing list
>
> Send postings to:
>  OpenHeart-L at lists.hsforum.com
>
> To UNSUBSCRIBE, to CHANGE email address, or to view archives:
> http://mmp.cjp.com/mailman/listinfo/openheart-l
>
> All messages transmitted by the OpenHeart-L are subject to the policies and 
> disclaimers posted at:
> http://www.hsforum.com/listdisclaim
> -----------------------------------------
>
>
>