National spending for all antidepressants increased 600% during the 1990s.^[1] In 2000, annual expenditures for SSRIs alone exceeded $7 billion.^[2] According to statistics from the National Ambulatory Medical Care Survey comparing data from 1999 with similar data from 1985, antidepressants accounted for 13.5% of the entire increase in pharmaceutical prescribing.^[3] Yet, a recent study conducted by researchers in northwest Italy showed nearly 50% of antidepressant users filled a prescription for such an agent (or any other monthly antidepressant prescription) only once or twice.^[4] Data from the United States are not much different; almost 50% of patients stop taking their medication as early as 3 months after beginning antidepressant therapy.^[5]

This short-term use of antidepressants contrasts with almost all clinical indications for this class of medication. Whether one is prescribing for mental health disorders - such as major depression, dysthymia, panic disorder, obsessive-compulsive disorder, generalized anxiety disorder, or social phobia - or other conditions - such as migraine headache, chronic pain, fibromyalgia, or overactive bladder - the story is the same. Antidepressants are slow to work and most useful in treating chronic conditions. Aside from possible short-term (a few months) use to aid smoking cessation or to manage certain forms of sleep disorder, antidepressants need to be taken for at least 6 months (per the guidelines for the minimal treatment of major depression^[6]).

So what is going on? A recent article by Fava^[7] reminds us that Carroll^[8] warned about the inappropriate use of antidepressant drugs 2 decades ago: "[We] strongly suspect that many patients who are simply unhappy or dysphoric receive these drugs, with predictable consequences in terms of morbidity from side effects, mortality from overdose, economic waste, and irrational, unproductive clinical management."

Fava continues by warning that to the same extent that tardive dyskinesia has limited inappropriate use of antipsychotics and that antibiotics should not be routinely prescribed for minor viral ailments, inappropriate use of antidepressant drugs may lead to a deterioration of clinical course with no clear benefit. There is even a theoretical possibility that antidepressants may make patients more likely to relapse. Whether or not that is true, studies clearly show that antidepressants are no more effective than placebo in treating mild depression. In the National Institute of Mental Health Treatment of Depression Collaborative Research Program, participants with mild depression who received imipramine plus clinical management fared no better than those who received clinical management alone.^[9] Similar results have been obtained in primary care patients treated with paroxetine.^[10]

Not only is there inappropriate use of antidepressants for subsyndromal conditions, but widespread off-label use is also worrisome. Streator and Moss^[11] performed a retrospective claims analysis of 1080 patients in a network model HMO and found that 56% of medical claims for members taking an SSRI were linked to non-FDA-approved indications. An acknowledged weakness of this study was the inability to determine whether claim-form diagnoses accurately reflected the real reason for the antidepressant prescription.

Within the Practice Partner Research Network, 6.3% of a total of 149,327 active patients of 389 participating primary care physicians (PCPs) received a prescription for an antidepressant in 1996, according to data from an electronic patient record system.^[12] Of these patients, 40% did not receive a diagnosis of depression. None of this off-label use is surprising if we remember that antidepressants are no longer limited to use in specialty mental health care. Most antidepressants are prescribed by PCPs rather than psychiatrists: a national MCO recently reported that 77% of all antidepressant prescriptions were written by PCPs participating in the MCO's 31 health plans.^[13]

If what we are seeing is a pattern of widespread antidepressant prescribing for a multitude of subsyndromal, amorphous, patient complaints, it suggests that antidepressants have become the modern-day sugar pill, or placebo. It is quite likely that antidepressants have largely replaced benzodiazepines in this regard. Since the 1980s, benzodiazepines - previously the drugs of choice for anxiety, sleep disorders, and nonspecific emotional upset - have been vilified as addictive and harmful. It seems that antidepressants, SSRIs in particular, have replaced benzodiazepines as the drugs of choice when the physician is at a loss for what to do to get a patient out of the office.

Barbui and colleagues^[14] wonder if the definition of depression used by clinicians to justify an antidepressant prescription may have changed in order to accommodate the large population of primary care patients who present with relatively mild symptoms of sleep disturbance and anxiety. Whether such use of antidepressants represents good practice I leave for others to determine, especially since SSRIs are associated with numerous side effects. I will discuss the topic of SSRI-related side effects in next month's column.

Although my feelings about the possible overuse of antidepressants are clear in this article, I cannot end without also warning against underuse of antidepressants. It may be that while some practitioners are overtreating with antidepressants, many symptomatic persons never go to a clinician's office, or when they do show up are not recognized as having depression or some other antidepressant-responsive disorder.

I agree with Croghan's cautionary comments in the March/April 2001 issue of Health Affairs: "Spending for antidepressants has increased by about 600%, or more than $6 billion during the 1990s, driven in large part by increased use. Although restrictions on this use might seem good fiscal management, we must challenge ourselves to remember the equally important problem that depressive disorders are undertreated. We do not yet know the degree to which unnecessary care contributes to rising expenditures. Until we understand the magnitude of the problem and how to encourage proper matching of treatment and patient, overzealous cost-cutting directed at reducing utilization could result in reducing medication treatment for those truly in need."^[1]