Too Few Diabetes Trials are Seeking Answers to Issues Important to Patients, Including Hard CV End Points

Shelley Wood

June 04, 2008

June 4, 2008 — Too many randomized controlled trials of diabetes drugs are designed with surrogate end points and not with the kinds of end points that actually matter to patients. That's the conclusion of a new analysis published in the June 4, 2008 issue of the Journal of the American Medical Association [ color="blue">1].

"There's an ethical obligation to justify the research enterprise by saying that at least a fraction of this--and a meaningful fraction--will provide information that will fundamentally change the lives of the people who are suffering from this condition," senior author on the study, Dr Victor M Montori (Mayo Clinic, Rochester, MN), told heartwire . "If a large majority of your research is either for knowledge's sake or for marketing purposes, that corrupts the evidence base and makes you wonder whether you really are doing the right thing."

If a large majority of your research is either for knowledge's sake, or for marketing purposes, that corrupts the evidence base and makes you wonder whether you really are doing the right thing.

The new study comes hot on the heels of regulatory discussions and congressional investigations into blockbuster drugs like ezetimibe and rosiglitazone and the trials conducted to assess them. Both drugs sparked fresh debates over the role of surrogate end points in major clinical trials. In the case of rosiglitazone, the drug appears to lower hemoglobin A1c but may increase the risk of both MI and death, while the entire glitazone class appears to increase the risk of heart failure.

"We really know very little about the effects of these new diabetes drugs and even about some of the older drugs on patient outcomes," Montori explained. "We know what they do on glucose control and from that we can infer all sorts of things, but those inferences could be right on or they could be fantasy. It's very difficult for patients and physicians to make decisions about the benefits of these agents without knowing how they will make patients feel, whether they will extend their lives, or whether they will prevent complications effectively."

Trials in the Pipeline Fall Short

For their study, Montori, with first author Dr Gunjan Y Gandhi and colleagues, reviewed relevant databases to identify all phase 2 to 4 randomized controlled trials of diabetes interventions, a search that yielded 436 trials. They found that "patient-important" outcomes--including cardiovascular events, death, pain, function, and quality of life--were chosen as primary outcomes in just 18% of trials and as primary or secondary outcomes in less than half of the studies.

While large trials and trials with longer follow-up were more likely to include outcomes important to patients, the bulk of trials used either surrogate outcomes, such as glycated hemoglobin or cholesterol levels, or physiological/laboratory outcomes (eg, insulin levels, C-peptide levels, etc). Of note, trials of patients with type 2 diabetes were significantly less likely than trials of type 1 diabetes or type 1 and 2 combined to report outcomes important to patients.

Authors of the current analysis had conducted an earlier review of published diabetes studies showing that patient-relevant outcomes were reported in just 21% of published trials [ color="blue">2]; this new research extends their findings to trials still in progress. "We wanted to see what's coming down the pipeline, and we found that 20% of the studies will be large enough and will measure a patient-important outcome," Montori told heartwire . "That's very good, but it's incredibly insufficient given the lack of evidence overall for diabetes. And while 45% of the studies will measure these outcomes,if the outcomes turn out not to be significant, they are unlikely to get published. So even if we go back and try to do meta-analyses, we're going to find that a lot of these studies have not reported the outcomes that matter."

According to Montori, there are two likely reasons why trials are continually designed with surrogate end points. The first is to find out how drugs work. "Those are necessary studies, but it's problematic if they become the majority of the work that you do, because then patients participate in this research and you get a lot of knowledge, but it's not translatable into practice; people don't benefit." The second is that trials are designed primarily to gain market approval or new market indications. "These tend to be brief, and they tend to focus on surrogate markers that then are sold as being important," he said.

Montori acknowledges that surrogate markers play an important role in expediting the drug-approval process, but he thinks the "stringent" rules for validating a surrogate marker are frequently bypassed. "It is unfortunate, because everyone wants to have newer, more potent, and safer agents, and it would slow the whole process down if we were to demand that for every one of these agents we want patient-important outcomes to be measured. Yet when we don't do it, we find ourselves crashing into a wall, realizing that we may have caused harm to people by again trying to rush to have these innovative agents on the market."

Reaching Consensus on Meaningful End Points

He and his colleagues hope that their study will spur researchers, clinical trialists, and diabetes health-advocacy groups to come together and reach some consensus about the types of outcomes needed in clinical trials of diabetes interventions. They point out that rheumatology researchers and patient groups launched just such an initiative, dubbed Outcome Measures in Rheumatology (OMERACT); as a result, researchers can compare outcomes across trials and appreciate the impact of different interventions on patients.

"By showing that only a fraction of published studies matter and that the ones coming down the pipeline are not going to be any better, we're hoping to raise the issue such that we can move people toward a consensus," says Montori. "Right now, with the potpourri of outcomes that we have, many of which are not validated and many of which have no relevant information in clinical decision making, that objective cannot be achieved."


The paper lists "none reported" for author financial disclosures.

  1. Gandhi GY, Murad MH, Fujiyoshi A, et al. Patient-important outcomes in registered diabetes trials. JAMA 2008; 299:2543-2549.

  2. Montori VM, Wang YG, Alonso-Coello P, Bhagra S. Systematic evaluation of the quality of randomized controlled trials in diabetes. Diabetes Care 2006; 29:1833-1838. Abstract

face="Verdana" size="1">The complete contents of Heartwire , a professional news service of WebMD, can be found at, a Web site for cardiovascular healthcare professionals.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: