Reader Comments

Post a new comment on this article

Trial reports vs Publications: Support from the Epoetin story

Posted by DCoyne on 13 Apr 2012 at 21:48 GMT

Drs. Doshi, Jefferson, and De Mar make an extremely important case for the review of complete clinical trial reports by Federal Agencies (e.g. the CDC) and their advisory committees (e.g. ACIP). I would add that these reviews are also essential by any Medical organization writing clinical practice guidelines. Such guidelines have a huge impact on the use of medications and treatments, can be used by the Federal Government to set clinical performance measures which further increase medication use, and intentionally or not, are extremely lucrative to the Pharmaceutical companies that fund such guideline organizations.
As an example of what I believe is both the selective publication by Pharma to obscure risks and overstate benefits, and Pharma’s manipulation of Guidelines by this selective publication, I point readers to my article on the use of Epoetin in US dialysis patients. The open access article can be read at
Briefly, the article compares the FDA-filed clinical trial report I obtained through a Freedom of Information Act (FOIA) request to the New England Journal of Medicine publication on the only outcomes study performed by Amgen using epoetin in dialysis patients. The so called Normal Hematocrit Trial randomized 1265 patients to Hemoglobin of 9-11g/dl or 13-15 g/dL. The trial was stopped by the Data Safety monitoring board in 1996 after the high target arm demonstrated a reported trend toward increased risk of death or nonfatal MI, according to the 1998 New England Journal of Medicine publication. While the 1998 NEJM results said the primary endpoint death + MI only trended up (Hazard Ratio 1.3, 95% Confidence interval 0.9 to 1.8), this was actually an adjusted confidence interval. The clinical trial report filed with the FDA in late 1996 showed the risk of death + MI (final HR 1.28, 95%CI 1.06-1.56) and all-cause death (final HR 1.27 95%CI 1.04 – 1.54) were significantly higher. It was not until June 2011 that these unadjusted confidence intervals appeared in the package insert for all Epo products. I received the trial report 2 weeks later in early July 2011. I had filed my FOIA request in January 2008.
The 1998 NEJM also replaced predefined endpoint analyses showing higher Epoetin targeting higher hemoglobin increased thrombotic events and accelerated the time to first hospitalization with post hoc analyses showing no effect on these endpoints. Most importantly, rather than report that targeting higher hemoglobin had absolutely no effect on improving Health-related Quality of Life (HRQoL) scores, they 1998 NEJM article reported only the association of higher hemoglobin with higher scores.
Amgen funded KDOQI guidelines beginning in the mid-1990’s. The first guideline released by this Amgen-funded group was late 1997. Serial guidelines written by KDOQI (in 1997, 2001, 2006, and 2007), and now KDIGO (draft guidelines in 2011) accepted Amgen’s claims about the Normal Hematocrit trial based on the 1998 NEJM publication, and the corresponding author, an Amgen employee. The claim higher hemoglobin improved HRQoL served as the cornerstone for KDOQI to say in 2006 all kidney patients with anemia needed to hemoglobin >11 g/dL.
If guidelines are going to be valid, all such guideline organizations must review the entire clinical trial report, not the selective publications of the pharmaceutical industry. Additionally, physicians should have access to the clinical trial reports to keep the guideline groups honest.

No competing interests declared.