Protecting participants in first-in-human trials
Review preclinical studies for treatments
The first-in-human clinical trials raise difficult ethical issues for researchers and review boards because of the uncertainty that accompanies them. Did the preclinical studies that preceded them provide enough information about effectiveness and risks and benefits to proceed with human volunteers?
Too often, positive expectations from those earlier tests do not end up translating well to results in human research, says Jonathan Kimmelman, PhD, an associate professor of biomedical ethics at McGill University in Montreal.
"I've studied areas like gene transfer, and there's an example of a field that's been characterized by this kind of boom-bust cycle where very promising preclinical trials come forward, they're rushed into human trials, and the agent turns out not to have the activity that everyone was expecting and hoping for from the preclinical studies," he says.
Kimmelman says review boards need to be able to look critically at proposed first-in-human trials, particularly in the area of balancing risks and benefits. "The vast majority of drugs that enter into human trials never survive to licensure," he says. "So there's a lot at stake in making good decisions at the point of initiating human studies. If you make a bad decision at that point, you end up exposing many patients to a drug that has an unfavorable risk-benefit balance, and you end up investing many resources into developing a drug that turns out not to be promising enough for licensure."
Review boards might mistakenly think that if the Food and Drug Administration (FDA) has approved the drug for human trials, it has adequately vetted its potential for efficacy, he says. "The FDA is mainly concerned about making sure that the investigators have a pretty good handle on the toxicity of the drug," Kimmelman says. "It's really up to review boards to be vetting first-in-human studies for clinical promise."
Preclinical studies key
The key to evaluating first-in-human trials is having reliable information about prior non-human studies, Kimmelman says.
In an article in the journal PLoS Medicine, Kimmelman and his colleague, Alex John London of Carnegie Mellon University, argue that predictions about the potential of an investigational drug are too often based on prior studies that might have problems themselves (such as lack of randomization in an animal study, for example) or that might not be comparable to the proposed human study under review.1
Sorting this issue out can be difficult for review boards, due to the complexity of early phase trials. "These are not garden-variety clinical trials," Kimmelman says. "They involve a very high level of expertise, not just in the clinical realm, but also in understanding the preclinical realm. One of the challenges of reviewing early phase studies is that it can be very difficult for review boards to have qualified personnel to review those studies."
Review boards should be ready to seek out the necessary expertise to deal with a particular study, he says. Also, they should look, wherever possible, to outside resources such as the National Institutes of Health's (NIH's) Recombinant DNA Advisory Committee, which reviews human gene transfer research. "In certain areas, such as areas of gene transfer, centralized oversight really provides an opportunity for experts to weigh in on the quality of preclinical studies," Kimmelman says.
Review boards also should hold investigators and sponsors accountable for ensuring that all relevant studies have been included.
"There is some pretty solid evidence that not all preclinical data get published," Kimmelman says. "And it's not entirely clear what proportion of preclinical data actually end up getting shared with review boards. If you're a review board member, you want to be sure that you're not just looking at the most positive outcomes from the preclinical studies; you want to make sure that you're looking at all the preclinical studies."
- Kimmelman J, London AJ. Predicting harms and benefits in translational trials: ethics, evidence and uncertainty. PLos Med 2011; 8:e1001010. Doi:10.1371/journal.pmed.1001010.
Issues concerning informed consent
Casting a bigger net for first-in-human trials
When there's a lack of strong data on a new experimental agent for first-in-human studies, researchers should cast the net wider by providing information to review boards about other drugs that work in the same way or that have other important similarities to the study drug, says Jonathan Kimmelman, PhD, associate professor of biomedical ethics at McGill University in Montreal.
Those other relevant studies shouldn't just influence the evaluation of a review board; they should be included in the informed consent in a trial going forward, Kimmelman says. Because first-in-human studies deal with such unknown quantities, informed consent must give participants any information possible that can aid in a decision.
"To simply say, 'We don't know how well this drug is going to perform,' you're telling a patient that there's somewhere between a 0 and 100% chance that the drug is going to have the clinical activity that's predicted in preclinical studies," Kimmelman says. "That basically has no informational content at all. What we're saying is if there is a record of similar interventions going forward, and not succeeding in clinical trials, patients should be told that."
Review boards should consider carefully the type of subject who is to be enrolled in a first-in-human trial, he says. One approach taken in many studies is to limit enrollment to patients for whom there has been no effective treatment to that point.
"If you enroll patients who have advanced disease, they have less to lose if there is no clinical activity or if there turn out to be risks that were unanticipated," Kimmelman says.
On the other hand, a person whose illness is well managed by existing treatments has greater potential for his or her quality of life to be harmed by participating in a first-in-human trial. "You want to have a good justification for enrolling patients who have other treatment options," Kimmelman says. "You want to be able to say that your new agent is competitive with those other treatment options, and in order to do that, you need to be able to make reliable predictions about the clinical activity of your drug. If you apply the procedures that we recommend and you conclude that effects that you saw in animals are unlikely to generalize to human beings, then I think you should be very cautious about enrolling patients who have other treatment options."