Underrepresented Populations in Clinical Trials

Shelley Wood; Sidney C Smith Jr, MD; Roxana Mehran, MD


December 05, 2013

This feature requires the newest version of Flash. You can download it here.

Ms Shelley Wood: Hi. I am Shelley Wood, the managing editor for heartwire , and my guests here today are Dr Sidney Smith and Dr Roxana Mehran, and our topic is underrepresented populations in clinical trials. My question is, how do we make sure that the results that we are seeing at these major meetings, the results I am reporting on for theheart.org, are representative of the patients that you are seeing every day in your practices? I wanted to start with research that you yourself have done, Dr Mehran, dealing with women in drug-eluting stent trials. If you want to just take a few minutes to tell us what you found in this recent research.

Dr Roxana Mehran: First, I want to thank you, Shelley, for paying attention to this study because it is a study of underrepresented population and it is usually underreported. The study is out in the Lancet as of September 2,[1] and what we did is a collaborative effort across 26 randomized trials that enrolled patients with drug-eluting stents and bare-metal stents from first-generation to next-generation stents, so a whole evolution of interventional cardiology over the past decade, and we found that only about 26% of the population were women. We used the raw data, patient-based data, to actually look at the safety and efficacy of drug-eluting stents, bare-metals, and next-generation drug-eluting stents in the female patient population without having to compare them with the male counterparts.

Ms Wood: And happily, the results you found were that indeed these were just as safe and effective in women.

Dr Mehran: And in fact, the next-generation drug-eluting stents, as they have been shown in the general population to be safe and effective and safer than the first generation is exactly replicated in the female patient population, so it was fantastic and a great way to share that data with everyone.

Ms Wood: This is an important piece of information, but in a way it is something we have assumed at least for what now—eight years—these have been implanted in women for years, so it is a piece of information we are getting quite a bit after the fact. Of course, women aren't the only group underrepresented in trials. Dr Smith, I want to address ethnic groups, there are different populations, and different parts of the world who may be treating patients on the basis of trials conducted half a world away. How do we make sure that the information that we are getting from the biggest, most popular studies are relevant to everyone they need to be relevant to?

Sidney Smith, MD: Thank you, Shelley. I think the other title for this might be how do I know if this trial is right for me and my patients, and so the answer is look at the inclusion criteria very carefully: is there representation of women and men; what are the ages represented; what about ethnicity; and what about other variables, which might be confounding—diabetes, smoking, hypertension, lipids, and the extent to which all of these things are treated? If you are lucky, the average trial will have 30% women, 70% men, roughly in that area. Why? Well, it turns out that most of the companies that sponsor trials don't want to have a lot of older patients because of comorbidities, so frequently they will exclude patients over 75, and we know that women become at risk 10 years older than men, so if you take a group of people in their 60s that have coronary disease you are going to find more men than women. It has nothing to do with the fact that I don't like women or women are less likely to be involved in trials or what have you. As we get into international work, a major concern of mine in regard to the work I am doing now in China has been the lack of Asian, both East Asian and South Asian, patients in these trials. Right now we are working on dual antiplatelet studies in Asia, we have three agents on the market, there's actually more, but three that have gone through major trials—clopidogrel, ticagrelor, and prasugrel—but the representation of Asian patients, who are more likely to bleed, is very low for the ticagrelor and prasugrel trials. So what do you do if you are in China and none of your people are represented? You know they have more bleeds. How do you interpret the data?

Ms Wood: Right. How do you?

Dr Smith: Well, I think you need to insist on getting some trials, and even with things like platelet reactivity there is the Asian paradox in that there is a shift in the curve. Asians are much more likely to have platelet resistance on clopidogrel, but they also have a lower level of thrombogenicity. They are less likely to clot, and so it turns out that Asians treated in this particular case with clopidogrel do quite well regardless of platelet-reactivity levels that might lead them to bleed.

Ms Wood: Can I just get you to rewind? You said you need to insist on trials in these populations, but I am not sure how do you go to a drug company and say, great, your drug is approved. It is approved internationally. Can you please go back and do the trial now in this population? How often does that happen?

Dr Smith: I think what happens is they see some market issues. In China, for instance, clopidogrel has been used safely in the COMMIT trial,[2] which is close to 46 000 patients, and we have two other agents that appear to be more effective in terms of outcomes—ticagrelor, prasugrel. If those companies want to have any market position in the largest country in the world where there is an agent that is being used safely and there is a concern about bleeding [with these agents], you need to do the studies. You absolutely do. The other thing that concerns me is the sound bite "what is the message here," and you really have to be careful. At this ESC meeting, we had two trials that looked at the [dipeptidyl peptidase-4] DPP4 inhibitors in terms of treating diabetes, came up with conclusions that they both do a good job of lowering glucose, they don't have pancreatic side effects and maybe a congestive heart failure issue, lower microalbumin.[3,4] They look like they are pretty good, and so the message goes out these are safe for patients with diabetes, but wait a minute, and it doesn't improve cardiovascular outcomes.

Ms Wood: Right.

Dr Smith: How long were the trials continued? Two years. Can you really learn anything about what is going to happen in cardiovascular outcomes in two years? Who were the people in the trials? Well, gee, all of them had high-risk coronary disease or recent coronary symptoms. Does this apply to a young patient with diabetes who has no disease?

Ms Wood: Now we are talking about trials and how do we include a representative population of just Americans or just that type of thing. There is layer after layer to this subject.

Dr Smith: And one other thing. If you want to show something works, you want to intervene on a high-risk group, so even our hypertension trials are going to take patients that have starting blood pressures at 160 or 170 mm Hg. The big argument now is what you should do at 150 mm Hg, go down lower than 140 mm Hg, but very few trials will start treating patients at this level because it is a large number of patients, and the drug companies are not going to invest in that.

Ms Wood: So if investigators aren't going to get it completely right for the patients that are easy to enroll, how are we sure that trials are going to get it right with the patients that are difficult to enroll, either different ethnic groups, different country populations, even women? Roxanna, do you want to take that one?

Dr Mehran: Well, I think it is also important to note that it is very hard to have a trial that is completely representative of every single subgroup and every single population, and I think that is the reason why some of these meta-analytical works can help us. Of course, it is not as good as a randomized study, and I agree with Dr Smith that if you want the drug to be used in the largest country of the world, you really do have to see the safety and efficacy in that population, but I think this is much more complex than that, Shelley. That that is why registries and working with registries that are representative of the entire healthcare system can help. We saw an example at this meeting, where you can do randomization blocks and really understand comparative-effectiveness research across the board in those large populations in healthcare registries. That could actually be an answer.

Dr Smith: Very strong point, Dr Mehran. If you look at what the Swedish trial did with the registry, I think it is going to be the way forward with electronic medical records.[5]

Ms Wood: This was in the TASTE trial on thrombus aspiration. We are not necessarily going to convince the people who fund the trials to make sure everyone is in there, but you both feel that we could perhaps after the fact or perhaps with these prospective registries make sure we can pull that information out a little bit later.

Dr Mehran: The trials should at least try to bring in a certain number and have some prespecified analyses for the subgroups that are important—patients with diabetes, patients with chronic kidney disease—they are totally ignored, the racial differences as well as the female patient population. That is first and foremost. Then, I think it is really important that we continue to include these patients in the trial, but what we don't want to do is to lag behind innovation in getting a great drug to the market because we don't have enough women or some other subgroup.

We need to make sure, and the FDA guidance document now for devices and for drugs is asking for exactly these kinds of very specific analyses for health disparities, if you will, for those patients who are underrepresented, and I think women are there, patients with chronic kidney disease, etc. Then these next methodologies, like patient-based pooled analyses, are another way to look at them.

Dr Smith: And I would add the elderly to this underrepresented group or excluded group for understandable reasons. And what we end up doing is we extrapolate something that happened and did a good job in a 63-year-old to an 83-year-old and they have comorbidities. They are on multiple other meds. It doesn't follow that what works in a 60-year-old is necessarily going to be right for an 80-year-old. So I think we have got to really get into that issue in terms of management.

Dr Mehran: I couldn't agree more with that. It is an aging world.

Ms Wood: I don't think we have any easy answers to the question I have posed here, and we have just scratched the surface, but I suppose for the practicing cardiologist, keep in mind that your patient may not have been in that trial and think it through and perhaps wait for more data as the years roll on. Who knows? Here is hoping. Thanks so much for speaking with me today.


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.