|
||||||
|
|
|
|
|||||||||||
|
|||||||||||||
| ||||||||||||||||||||||||||||||||||||||
The results of the PROactive study were presented September 12 at the 41stAnnual Meeting of the European Association for the Study of Diabetes (EASD) inAthens, Greece, and were published in The Lancet on October81 ("DiabetesDrug Reportedly Cuts Death, Heart Attack, Stroke Risk," seepage 1). The study designand baseline data had been published previously in DiabetesCare.2 Thus,the diabetes community was aware of the potential importance of PROactive.
The study was designed to determine the effects on macrovascular disease(atherosclerosis) in type 2 diabetes of the addition to standard therapy ofthe peroxisone proliferator–activated receptor (PPAR)-gamma agonistpioglitazone (Actos,Takeda/Lilly).
At the outset, in the study design paper, the investigators stated:
"A composite cardiovascular disease end point is used because the aimof the study is to evaluate the overall effects on macrovascular disease. Theprimary end point variable is the time from randomization to the firstoccurrence of any of the events in the following composite: all-causemortality; nonfatal MI [myocardial infarction]; acute coronary syndrome;cardiac intervention, including coronary artery bypass graft, or percutaneouscoronary intervention; stroke; major leg amputation (above the ankle); bypasssurgery; or revascularization in the leg. An independent panel adjudicates theend points. Secondary end points include the individual components of theprimary end point and cardiovascularmortality."2
That is what one expected to hear about, but that was not to be.
The primary end point—based on the efficacy evaluation stated in thestudy design paper—failed to reach statistical significance. Thus,PROactive was a negative study. But one surely cannot glean that from theinvestigators' presentation. The investigators—and many in thepress—have discarded the primary end point because it did not show theeffect that was sought.
REVISIONIST END POINTS
So, how did this negative study become positive? The authors have crafted anew concept, that of "principal secondary end point," but there isno mention of this in the study design paper. This newly appearing end pointincluded only all-cause mortality, nonfatal MI, and stroke. At the EASDpresentation in Athens—but not in the Lancet paper—theinvestigators declared that they inserted this measure into the statisticalplan prior to breaking the code and analyzing the data.
Yet PROactive appropriately had a data safety and monitoring board (DSMB),which usually would have access to the emerging data to make an assessment ofrisk versus benefit. Thus, the statisticians at the coordinating center, theDSMB, and perhaps the study chairman, had earlier access to end-point data.Under these circumstances, one has to wonder who might have learned of thenegative primary end point and whether such knowledge could have led to thelast-minute creation of a "principal secondary end point" thatconveniently could be said to be positive. Even when secondary end points arestated a priori, they should only be considered meaningful when the primaryend point is positive. Otherwise, in the face of a negative primary end point,secondary end points should be considered exploratory or hypothesisgenerating.
The authors now argue that a composite end point encompassing onlyall-cause mortality, nonfatal MI, and stroke avoids those components of theprimary end point that may be "in part determined by a decision tointervene based on local surgical or medical practice." That raises thequestion of why it wasn't selected as the primary end point in the firstplace. I suspect the other components were included because, in thisevent-driven study design, investigators wanted to be sure there were enoughevents in the study's relatively short time frame. Yet including these extracomponents in diluted the potential impact of the study, resulting in anegative overall study. All of the investigators' claims about the importanceof their secondary end point cannot negate that. The correct statisticalinterpretation of the study is that it is negative. Period.
Adding insult to injury, the investigators—and the press releaseabout the study—assert that the beneficial effects were due topioglitazone per se. Hold on a minute. The glycated hemoglobin (A1C) level atthe end of the study was 7% in the pioglitazone group versus 7.6% in theplacebo group. Any result seen could merely be a consequence of improvedglycemic control rather than a unique effect of pioglitazone. In order to testwhether there is a unique effect of pioglitazone not attributable to glycemiccontrol, this effect must be evident when there is equivalent glycemic controlin the comparison group, that control being achieved by other glucose-loweringagents.
And the investigators knew that. Indeed, the study design paper states"the investigators are encouraged to maintain glycemia within the limitsoutlined in the International Diabetes Federation (IDF) Europe Guidelines(<6.5%), which was highlighted and circulated to allinvestigators."2The results paper again states, "we drew particular attention to theneed to reach an HbA1c concentration below the recommended target(<6.5%)."1Thus, there is no unique effect of pioglitazone, even if one were to acceptthe flawed premise that the secondary end point is positive.
In addition, in a subgroup analysis—which also probably is notwarranted in a negative study but was included in the EASDpresentation3 yetnot in the Lancetpaper1—thealleged beneficial effects for both the primary and principal secondary endpoints were seen in patients not using statins and were not seen in those whoused statins. Since statin use in patients with type 2 diabetes is highlydesirable, their use may obviate the need for adding pioglitazone, even if oneaccepts the flawed notion that a benefit of pioglitazone has beendemonstrated.
Finally, as noted by Hannele Yki-Järvinen of the University ofHelsinki in her commentary that accompanied The Lancet's publicationof PROactive, the subjects treated with pioglitazone had 115 more episodes ofheart failure and 221 more episodes of edema than the placebo subjects. Inaddition, weight gain was 4 kg (8.8 lb) greater in the pioglitazone group thanin the placebogroup.4 The questionYki-Järvinen raises is about the overall impact of pioglitazone onhealth. There were 58 fewer primary end points (57 fewer principal secondaryend points) with pioglitazone, but "from the patient's perspective, isit better to have healthy arteries in the heart than a failing heart?"she writes, pointing out that the prognosis of heart failure is particularlypoor in patients with type 2 diabetes.
The PROactive investigators conclude, "in patients with type 2diabetes who are at high cardiovascular risk, pioglitazone improvescardiovascular outcome, and reduces the need to add insulin toglucose-lowering regimens compared with placebo," and "we believeour results are generalizable to all patients with type 2diabetes."1For the reasons reviewed above, as a disciplined clinical trialist, I wouldargue that such conclusions statistically and clinically are not justified andmake a mockery of our system of evidence-based medicine. To me, this is a sadtale. 
Footnotes
|
References
2. Charbonnel B, Dormandy J, Erdmann E, et al., for the PROactiveStudy Group: The Prospective Pioglitazone Clinical Trial in MacrovascularEvents (PROactive): Can pioglitazone reduce cardiovascular events in diabetes?Study design and baseline characteristics of 5,238 patients.Diabetes Care 27:1647–1653, 2004.
3. European Association for the Study of Diabetes. Results ofPROactive, Athens, Greece, 12-Sept-05. Available atwww.easd-lectures.org/athens/index.php?menu=lectures&id=27.Accessed October 17, 2005.
4. Yki-Järvinen H: The PROactive study: Some answers, manyquestions. Lancet 366:1241–1242, 2005.[Medline]
CiteULike 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. S. Skyler PROactive: A Sad Tale of Inappropriate Analysis and Unjustified Interpretation Clin. Diabetes, April 1, 2006; 24(2): 63 - 65. [Full Text] [PDF]
|
|
Read all eLetters
| ||||||||||||||||||||||||||||||||||||||
|
||||||
|
| DOC News | Diabetes | Diabetes Care | Clinical Diabetes | Diabetes Spectrum |
