In a speech last month, Dr. Margaret Hamburg, the new Food and Drug Administration commissioner, promised the agency would begin to take a tougher stance on public health risks. "The FDA must be vigilant, the FDA must be strategic, the FDA must be quick, and the FDA must be visible," she said.
This approach is certainly laudable when it comes to getting rid of bogus or obviously dangerous products. But our new, progressive public health administrators should be mindful that their interventions can have unintended consequences. This seems to be the case with the FDA's warnings about suicide risk associated with using antidepressants.
The FDA's action against antidepressants began in 2003, following a British government investigation into suicides associated with the use of paroxetine (Paxil). An FDA panel that examined data from dozens of clinical trials of different selective serotonin reuptake inhibitors—like Paxil, Prozac, Zoloft, and Celexa—didn't find a single suicide in the nearly 1,500 children and young adults enrolled. But it did find that those who got the drug exhibited higher rates of suicidal thoughts and behaviors than those who took the placebo. So in October 2004, the FDA ordered that a "black box warning" about pediatric use of the drugs be put at the top of the prescribing insert.
The warning was subtle and well-reasoned, as was an expanded 2007 message that applied to all patients under 25. The warnings didn't tell doctors not to prescribe the drugs, and they noted that untreated depression itself is associated with an increased risk of suicide. They urged those prescribing the drugs to "monitor appropriately and observe closely for clinical worsening, suicidality or unusual changes in behavior."
These were sage words. However, they were directed to an idealized health care system in which patients regularly see psychiatrists who closely monitor their mental health. In the real world of U.S. health care, visits to psychiatrists are highly rationed, so primary care physicians do much of the prescribing of psychiatric meds. These doctors get about 15 minutes to spend with their patients, and a lot of them apparently decided it was too risky to prescribe medications bearing black box warnings, with all their connotations of liability and disaster. So they stopped prescribing them, especially to new patients. A recent study in the Archives of General Psychiatry found that new diagnoses of depression fell 44 percent for children, 37 percent for young adults, and 29 percent for adults in the eight years before June 2007. Other studies show more modest, though still substantial, declines.
Shrinks assess these declines more or less according to philosophical positions about antidepressant use. Some therapists believe prescribing had gotten out of hand and that many children receiving medications were not clinically depressed. It's certainly possible that part of the drop is attributable to better diagnosing of children, though it's less easy to believe with older patients. A depressed 22-year-old is hard to miss.
Concern about the impact of the warnings led another FDA panel that was writing up a suicide warning for anti-epileptic drugs last year to decline the "black box" approach. Instead, they chose to incorporate the warning in the section of the insert on side effects. The panel determined that a patient who stopped using anti-epileptic drugs would be at much greater risk than one who used them cautiously. "The black box is a blunt instrument," says Robert Valuck, a psychopharmacology researcher in Denver. "It's the last step before withdrawing a drug."
Did the warning have any impact on suicide rates in teens? That's the bottom line. Here, unfortunately, the evidence is especially murky. In 2004, the rate of suicide in young people under 20 increased 18 percent over 2003—the first increase in 12 years. The rate decreased somewhat from 2004 levels in 2005 and 2006 but remained above the 2003 level.
When FDA panels met in 2004 to discuss the black box warnings, they heard from relatives of young patients who'd taken antidepressants and then committed suicide—truly disheartening anecdotes. Yet the few studies that have examined the blood of young suicides rarely find even traces of antidepressant use. A study of 38 youth suicides in New York City from 1999-2002 found metabolic traces of antidepressants in the blood of only one—a homeless 16-year-old who died of an intentional drug overdose. Many suicidal teenagers probably never see a therapist. Others do but don't take the medications they've been prescribed.
About 10 percent of Americans (2.5 percent of those under 17) took antidepressants in 2005, compared with 6 percent (1.4 percent of youth) in 1996, according to a recent survey. Suicide rates in the United States have declined marginally since 1990, and more steeply in most of Western Europe, but none of the suicide studies I've seen is detailed enough to determine cause and effect. Over the same period, teen drug use has declined, and more teen counseling and depression awareness may have played a role.
An FDA spokeswoman said the agency appreciates studies of the impact of its warnings, adding that "our drug safety discussions, for any type of product including SSRIs, are always open and ongoing."
There's one obvious conclusion to be drawn here: The FDA warnings, which came about because of worry that antidepressants might cause young people to commit suicide, have not caused suicide rates to drop and may have even caused doctors to withhold treatment for those who need it. Perhaps we can't fault the FDA for this. Anxiety about the use of SSRIs has grown with their use because of our profound philosophical concerns about the prescribing of mind-altering medications, especially in children. Perhaps the black box warnings were as good a method as any to test a hypothesis on public health. Once the impact of the warning becomes clearer, it may also help us clarify our thoughts about the role of pills in shaping our thoughts and actions.