Safety Problems Challenge Clinical Trial Design

FDA eyes new approaches to testing drugs for children and adults, and clarifies electronic data policies for clinical trials.

Questions about who knew what when about adverse events in several widely prescribed drugs has again focused the spotlight on clinical trial data and design issues. Members of Congress have grilled Food and Drug Administration (FDA) officials about why they approved antidepressants for children when most studies showed signs that these drugs weren't efficacious and might induce adolescents to contemplate suicide. Also, the research community is considering the need for more clinical data prior to approving new chronic-use drugs following Merck's surprise decision to pull Vioxx off the market due to long-term safety issues. These events have raised concerns about the conduct of clinical research and FDA oversight of the process.

Rethinking antidepressants
Charges that FDA officials hid serious effects data on antidepressants and kids prompted Congressional hearings and FDA deliberations this fall. The issue first grabbed public attention in June when New York Attorney General Eliot Spitzer challenged GlaxoSmithKline for allegedly concealing data on the safety and efficacy in children of Paxil (paroxetine) [see "View from Washington," October 2004]. The Senate Finance Committee began investigating whether FDA improperly withheld data on suicide-related adverse events in children taking selective serotonin re-uptake inhibitors. The House Energy and Commerce Committee held two hearings on the issue in September. At the same time, a joint meeting of FDA's Psychopharmacologic Drugs and Pediatric Advisory Committees weighed evidence on the safety and efficacy of antidepressants, and recommended that these products carry a black box warning on the possibility of suicidal behavior in young patients.

Beyond all the accusations about hiding data and FDA cover-ups, the debate has prompted policy changes designed to encourage more informative clinical trials on antidepressants. At the second House hearing, Robert Temple, Director of FDA's Office of Medical Policy, explained why it is difficult for many clinical trials of antidepressants to show efficacy. He noted that of seven approved products studied in pediatric patients with major depressive disorder, the data supported efficacy only for Eli Lilly's Prozac (fluoxetine).

Consequently, in future pediatric studies for antidepressants, Temple says that FDA plans to request three-arm studies that compare the test drug to both placebo and Prozac. If the test drug beats placebo, but not Prozac, then there's probably something wrong with the study, Temple explains. If the test drug fails to beat both arms, then the sponsor will know that the drug is ineffective-and not that the study was poorly designed.

Less pressure
Temple also questioned whether FDA's pediatric exclusivity program may encourage sponsors to conduct inadequate trials in order to meet patent deadlines. He observed that sponsors of pediatric studies are not under as much pressure to obtain positive results as when they are conducting clinical trials to bring a new product to market.

Under the Best Pharmaceuticals for Children Act of 2002, which amended provisions in the FDA Modernization Act of 1997, sponsors can obtain six-month patent extensions by testing an approved drug on children as outlined in an FDA request. The studies do not have to show efficacy or result in new labeling.

While Temple stated in his testimony that sponsors generally design these studies "with good intent and according to high standards," he noted that the failure of a supplemental pediatric study to show efficacy is not nearly as devastating as failure in a pre-approval efficacy trial. Sponsors of pediatric studies thus may support too few trials to obtain clear data, or may recruit too few patients or be careless in recruiting patients with the proper diagnosis. At the FDA advisory committee meeting, experts suggested that sponsors may rush to complete high-quality pediatric studies before key patents expire.

While these comments are directed at pediatric study design, FDA may extend its recommendations for more complex and comprehensive research protocols to adult studies. FDA is reviewing a database of 234 randomized clinical trials for 20 drugs to uncover any information on actual suicides in all the studies. Up until now, researchers in the United States and the United Kingdom have not found data indicating increased risk of suicide in adults, but have no clear explanation for the differing responses in adults and children. All parties are anxious to see if new trial designs can clarify these issues.

Chronic use concerns
Similarly, the withdrawal of Merck's COX-2 inhibitor Vioxx is raising broader questions about testing drugs likely to be used by patients for years. Merck made its startling decision based on data indicating that patients may have increased risk for heart attack and stroke if they use the drug for more than a few months, results that are generating massive lawsuits and sending shock waves throughout the pharmaceutical industry.

Ironically, these new anti-inflammatory medicines made huge market gains in the last five years because they appeared less likely to cause the ulcers and gastrointestinal bleeding found in older painkillers such as aspirin and ibuprofen. But FDA approved COX-2 inhibitors based on clinical trials running less than a year, some only six to nine months. The serious side effects, including elevated blood pressure, cardiovascular events and liver toxicity, appeared months later.

To address rising safety concerns, Merck assessed these side effects in a three-year post-approval study to test Vioxx for efficacy in preventing the recurrence of colorectal polyps. The company had to halt the 2600-patient study after 18 months when preliminary results revealed that 3.5% of patients taking Vioxx had thrombotic events, compared to 1.9% in the placebo arm.

FDA officials are examining additional data from postmarketing studies involving other anti-inflammatory drugs to see if similar safety problems extend beyond Vioxx. The findings may influence how much more long-term safety data FDA is likely to require for new drugs in this class coming down the pipeline. Merck has underway an extensive safety trial for Arcoxia (eterocoxib), a follow-on compound to Vioxx, but it may require even more testing to ever come to market. Novartis is running extensive studies with thousands of patients in efforts to gain approval of its new COX-2 therapy, Prexige (humiracoxib).

Proposals to greatly increase the length of studies and number of patients tested always create a dilemma for FDA. Overly extensive data demands can block a new drug from market if test costs escalate too much, but new drug approvals based on limited data can subject the agency to attack if safety problems later emerge.

Clarity on Part 11
Longer and larger trials mean more complex data filed with FDA, making it even more important for sponsors, investigators, and reviewers to be able to send and receive study information electronically. FDA aims to encourage the use of computerized systems in clinical research through new guidance on how clinical trial data is affected by recent revisions in its Part 11 rules governing electronic documents and signatures.

The guidance gives recommendations to contract research organizations, data management centers, and sponsors about using computerized systems to create, modify, maintain, archive, retrieve, or transmit clinical data intended for submission to FDA. The overriding aim is to ensure that data provided as the basis for FDA decisions on the safety and effectiveness of new drugs and medical products are of high quality and integrity.

When finalized, the document will replace an April 1999 guidance and will align FDA's thinking on clinical trial operations with its revised Part 11 policy and with international harmonization efforts. In August 2003, FDA issued guidance clarifying the agency's plan to narrowly interpret the scope of Part 11 and to exercise enforcement discretion regarding requirements for validation, audit trails, record retention, and record copying. While reiterating that FDA does not intend to enforce compliance with part 11 requirements while that policy is undergoing revision, the guidance offers recommendations for establishing SOPs for computerized systems, for ensuring proper data entry into these systems, for maintaining audit trails and other security measures, and for retrieving data and retaining records.

The clinical data guidance was one of several documents issued by FDA in September to conclude the first phase of its initiative to modernize current good manufacturing practices (GMPs). The aim is to encourage manufacturers to adopt new production technologies and quality-based systems, while also promoting risk-based management approaches for industry and FDA itself.

A final report on the two-year initiative summarizes changes in how FDA conducts plant inspections, reviews applications, oversees postapproval changes, and deals with electronic submissions. The agency also established a new FDA Council on Pharmaceutical Quality to provide a central forum to address ongoing issues related to inspections and application review, as well as further policy development. In the next phase of the initiative, FDA officials plan to consider the need for a "sliding scale" of GMPs for clinical supplies as one way to ensure the quality of test therapies used in clinical studies.

Sidebar: No Consulting with NIH
In the face of continuing criticism over lucrative liaisons between drug companies and National Institutes of Health (NIH) scientists, NIH director Elias Zerhouni imposed a temporary one-year ban on paid outside consulting activities involving agency staffers and industry. NIH says it will use the hiatus to establish a more effective oversight system for outside activities by federal researchers. NIH employees still may advise pharma companies and research organizations, but just can't take any money for doing so. Some researchers fear that the ban will prompt talented scientists to leave the government, but others find the temporary consulting ban acceptable. It remains to be seen how long the ban lasts and if it ends up being only temporary.-J.W.