Experts in neurology and psychiatry agreed with the Food and Drug Administration's (FDA) assertion that antiepileptic drugs, as a class, increase the risk of suicidal thoughts and behaviors, but they were concerned about the negative consequences of imposing a black-box warning on the labeling information.FIG1

At a joint meeting on July 10, members of the peripheral and central nervous system and psychopharmacology advisory committees voted in favor of adding warnings and precautions about the suicidality risk to the package inserts of all antiepileptic drugs—but not in the form of a black-box warning.

In 2005, the agency began asking the makers of 11 drugs approved for long-term treatment of epilepsy to submit data on the drugs and suicide-related events reported in randomized, parallel-arm, placebo-controlled trials the companies conducted. Earlier this year, the FDA announced the results of its preliminary analysis and warned about a suspected increase in suicidality risk in patients taking antiepileptics (Psychiatric News, March 7).

The 11 drugs included in the analysis were carbamazepine, divalproex sodium, felbamate, gapapentin, lamotrigine, levetiracetam, oxcarbazepine, pregabalin, tiagabine, topiramate, and zonisamide. Many other drugs that carry the indication for treating epilepsy were not included because they are sold in generic formulations by multiple manufacturers, which have little incentive to conduct clinical trials. Also, the original clinical trial data for these drugs may be too old, nonrandomized, without placebo control, or difficult to obtain.

Suicidality was defined as completed suicide, attempted suicide, preparatory acts toward imminent suicidal behavior, or suicidal ideation.

At this meeting, FDA staff presented updated analyses based on pooled data from 199 clinical trials involving 27,863 patients on active drugs and 16,029 on placebo. The prevalence of any suicidality event was 0.22 percent in patients on placebo and 0.37 percent in those on active drugs. The odds ratio between them was 1.80. The difference between placebo and antiepileptics was statistically significant. The pooled trial data encompassed various indications such as epilepsy, psychiatric disorders, and pain.

Based on the statistical analyses, this risk difference is equivalent to an increase of 1.9 patients with suicidality in every 1,000 patients taking antiepileptic drugs compared with those on placebo.

The FDA had enough data for analysis up to 24 weeks, and the risk of suicidality appears to persist throughout this period. The risk difference between placebo and active drugs is unknown between patients with epilepsy and bipolar disorder who may take the drugs for years.

At the advisory committees' meeting, Pfizer representatives vehemently argued that its two drugs, pregabalin and gabapentin, were different from the other antiepileptics in the analyses and should be exempt from carrying the suicidality warning. These two drugs, however, were not the least risky of the 11 drugs analyzed. Carbamazepine and divalproex had odds ratios below 1 in the analyses, meaning that the risk of suicidality associated with these drugs was lower than for placebo in clinical trials of these two drugs. In addition, the individual risk with felbamate could not be calculated because no suicidal behavior or ideation was reported in its trials.

Nevertheless, the FDA position is that this "signal" of increased suicidality is a class effect, even though these 11 antiepileptic drugs have different direct biological effects on the nervous system. Mark Levenson, Ph.D., the FDA reviewer who conducted the analyses, pointed out that felbamate and carbamazepine had the two smallest datasets among the 11 drugs (both under 1,000 patients), and the low risk observed may be a result of the small sample size.

While 20 of the 21 members of the advisory committees concurred with the FDA analyses and the conclusion about an elevated suicidality risk, 18 members voted "yes" to the question of whether the increased risk should be applied to all 11 antiepileptic drugs; three members voted" no."

The agency then asked the advisory committees whether they believed this suicidality risk should be extrapolated to all drugs approved for treating epilepsy, even though there are no randomized, placebo-controlled clinical data for direct analyses except for these 11 drugs.

With a 15-5 vote and one abstention, the committees agreed with the agency's recommendation, showing a growing unease with expanding the conclusion to other drugs without direct evidence to support it. This vote was, in part, driven by the worry that some clinicians may be inclined to prescribe older antiepileptics if a suicidality warning was imposed on only newer drugs, even though there is no evidence that the older drugs are safer.

To further complicate the debate, many of the drugs in this class are prescribed for a variety of indications. Carbamazepine and lamotrigene, for example, are major treatments for bipolar disorder. Pregabalin is approved for neuropathic pain, and topiramate for migraine. Off-label use is common among these drugs. The risk-benefit ratio can differ dramatically depending on the illness for which the drugs are used.

"We are concerned about the clinical impact of reduced use of mood stabilizers," said Darrel Regier, M.D., M.P.H., director of the American Psychiatric Institute for Research and Education, in his testimony to the advisory committees and the FDA. He urged the agency to refrain from requiring the black-box warning and pointed to the substantial risk of suicide if bipolar patients go without adequate treatment.

"The data in the FDA analyses do not seem to suggest that the rate of suicidal thoughts and behaviors [linked to antiepileptics] outweighs the potential harm from possible medication discontinuation as a result of a black-box warning," he said.

Laurence Greenhill, M.D., president of the American Academy of Child and Adolescent Psychiatry, echoed Regier's concerns in his testimony. "As we have learned from previous experience, the FDA's decision to use a black-box warning significantly influenced prescribing in psychiatry and impacted patient care." Patients with serious neurological and psychiatric disorders could face life-threatening risks if they choose to stop taking these necessary treatments out of fear linked to the black-box warning, he noted.

"There is a need for prospective, systematic analysis of suicidality as opposed to relying on retrospective, spontaneous reports of suicidal events in clinical trials," regier said.

Many members on the advisory committees expressed concerns about the unintended consequences of black-box warnings on patient care. The documented reduction in antidepressant prescribing in pediatric and adult patients with depression after the black-box warning about suicidality was issued in 2004, and the warning's impact on public health, continues to worry some clinicians.

In the end, the advisory committees voted 14-4, with three abstentions, to recommend against imposing a black-box warning for all antiepileptic drugs.

The FDA faces a difficult decision because it has limited means to communicate nuanced risk information to the public and clinicians. Russell Katz, M.D., director of the agency's Division of Neurology Products, told reporters after the meeting that the final decision on the warning for antiepileptic drugs will be announced soon. In the past, the FDA has frequently, but not always, followed the recommendations of its advisory committees. ▪