ATLANTA— A continuing source of controversy in neuropsychiatry is whether the selection of antidepressant medications should be based on how they affect different symptoms or on other factors such as their side-effect profile. Finding the weight of evidence supporting the view that efficacy is generally comparable between classes of antidepressants and among agents in the same class, the current American Psychiatric Association (APA) Practice Guideline for the Treatment of Patients with Major Depressive Disorder states that the initial selection of an antidepressant should be based largely on “the anticipated side effects, the safety or tolerability of those side effects for the individual patient, patient preference, quality of clinical trial data, and its cost.”

Yet, despite the official sanction of this “one-effect-fits-all” approach, recent survey data of antidepressant drug selection by psychiatrists suggest that clinicians continue to choose an agent based on such factors as the patient’s symptoms or the drug’s “profile,” including its purported effectiveness.

What are the reasons for this stark disconnect between evidence-based protocols and clinical practice? Two presentations made at the 158th Annual Meeting of the APA help provide insight into this question. Although they were held two days apart and were not programmed by the meeting’s organizers to represent the pro and con of this issue, the two talks cross-referenced many of the same studies and together constitute the essence of this important clinical debate.

TARGETING IN

Do different antidepressant medications work better in different kinds of depression?

Yes, said Richard J. Metzner, MD—when antidepressants with different mechanisms of action are targeted at different depressive subtypes. Drawing on research studies dating from the 1960s and a tripartite model of depression developed in the early 1980s, Dr. Metzner and colleagues at the University of California, Los Angeles (UCLA), have developed a new approach to antidepressant selection, known as the targeted treatment of depression, including a self-administered, computer-scored, 17-item implementation of this new model called the Targeted Treatment of Depression Inventory (TTDI).

Unlike other depression tests that use scores of zero or higher, the TTDI employs two bimodal scales that can detect both positive and negative values, allowing for the identification of drug-induced blunted and manic states. The inventory’s “M” scale quantifies the potential need for modulation of increased negative affects (eg, anxiety or panic) with serotonergic antidepressants, while its “A” scale measures the potential need for activation of decreased positive affects (eg, apathy or reduced libido) with catecholaminergic agents.

THE TTDI IN ACTION

Preliminary studies involving several hundred patients indicated that the TTDI rapidly provides a quantified measure of demodulation and deactivation that can help guide antidepressant selection, according to Dr. Metzner, who is Clinical Professor of Psychiatry and Biobehavorial Sciences at UCLA. In a study involving more than 1,000 patients with depression seen between 1985 and 2000 as outpatients in a naturalistic private practice setting, Dr. Metzner and colleagues found that 96% of 100 patients receiving targeted treatment with one or more selective antidepressants showed improvement, compared with 65% of 55 patients who received nontargeted treatment with a tricyclic antidepressant or a monoamine oxidase inhibitor.

Two-thirds improvement is typical of nontargeted treatment approaches, which tend to lump all depressed patients together regardless of symptoms, Dr. Metzner maintained. He called for better studies and mood-measuring tools to help resolve the debate over the two opposing approaches, and he invited participation from his audience in a TTDI development group devoted to continuing validation studies on the inventory to derive normative, comparative, and predictive data.

“What we’re really looking to do is to get the wisdom of people who are experts at mood measuring to participate in further development of this idea,” Dr. Metzner said. “Then we can finally have a way to do prescriptive studies—studies that actually make an effort to say, ‘This subtype of depression, this set of core features, is being tested with a drug that may be more appropriate and a drug that may be less appropriate, in a double-blind fashion, and let’s see whether or not these people do, in fact, respond more to one or the other [drug].’

“Until we have studies that don’t simply lump depressed people together, where we don’t know which subtype they represent, we’re not going to know the answer to that question.”

RESEARCH PROTOCOLS AND CLINICAL FINDINGS

Depression is the world’s most disabling medical disorder, according to a 1996 report by the World Health Organization; yet a 2003 study by Kessler et al published in JAMA found that only one in five depressed patients in the United States receives proper treatment. Potential targets for treating depression can be categorized as psychotherapeutic, medical, and psychopharmacologic. Although selecting antidepressant medications based on a patient’s clinical subtype or symptoms may appear to be a reasonable approach, it is not currently endorsed by many leading psychiatric researchers nor is it a component of standard protocols such as the APA treatment guidelines.

“We have a big controversy in psychiatry when it comes to the psychopharmacologic side,” Dr. Metzner said. “The idea that one can select antidepressant treatment on the basis of neurophysiological mechanisms that may more effectively address the patient’s specific clinical pattern or subtype at this moment in time is not an accepted idea in academic psychiatry or official psychiatry.”

In Dr. Metzner’s view, the dividing line is clear between those who support the officially sanctioned, nontargeted approach to the treatment of depression and those who don’t. On one side, there are mostly psychiatric researchers who have done controlled studies using instruments such as the Hamilton Depression Rating Scale (HAM-D) and newer measures that have been developed to replace such established scales. On the other side are mainly clinical psychiatrists and other clinicians, most of whom will acknowledge—privately, if not for the record—that they do distinguish among antidepressants and find them to have, at times, profoundly different effects.

In fact, it is the experience of many primary care physicians that serves as one of the main reasons why Dr. Metzner finds it would be useful to have a test such as the TTDI to guide antidepressant prescription. “Most of the people prescribing antidepressants are not sitting in this room,” he pointed out to his audience. “Primary care physicians prescribe most of them, and if you’ve talked to them—as I have all over the country, I’ve spent time talking with thousands of them—they really don’t have a very good idea, most of them, about how to do this.”

Although Dr. Metzner is an academic researcher as well as a private practitioner, he clearly finds himself among the clinicians in this debate. “I attended medical school where William Osler had once been, and I was raised with the notion that for a physician to ignore what a patient says or what [the physician] sees is not good medicine,” he said. “But that’s in essence what we are currently being taught to do and what other medical specialists are being taught to do: Don’t pay attention to what you see as much as, well, just to accept the idea that they’re all the same except for side effects.”

BUILDING THE CASE FOR TARGETED TREATMENT

The case for nontargeted treatment is based on the view that the effectiveness of antidepressant medications is comparable among different classes of drugs. Dr. Metzner cited studies comparing serotonergic and noradrinergic antidepressants that revealed no differences in efficacy. In his view, such findings fly in the face of evidence from naturalistic, neurobiologic, and psychological studies supporting the idea that serotonin, norepinephrine, and dopamine do have differential effects—not to mention surveys of experienced practitioners showing that clinicians select antidepressants on the basis of clinically observed differences in patients’ symptoms.

Dr. Metzner asserted that reliance on undifferentiated depressed populations weakens the case for nontargeted treatment. Attention to subtypes is the key to the distinction between the two opposing approaches. “Using instruments like the HAM-D or the Montgomery, you really don’t know who the one in three patients is who didn’t get better with each drug. We just don’t have those data. In fact, when you talk to companies that do these studies, they say that the FDA mandates that they not look at subtypes.”

He cited a report by European investigators who noted that the Montgomery Asberg Depression Rating Scale measured improvement using more established antidepressant agents but did not differentiate the effects of the agents, despite their different mechanisms. Dr. Metzner then pointed to work by Gibbons et al suggesting that insensitive tools hamper the search for a preferential response from patients to different antidepressants.

“The tautology goes something like this: Standardized depression tools don’t discriminate among depressive subtypes,” he explained. “Clinical trials of antidepressants using those tools can’t detect differential effects on those subtypes. Lack of data for targeted treatment is presented as a problem with targeted treatment rather than with outdated tools. And the lack of good tools limits clinically relevant research.”

Dr. Metzner proceeded to cite several studies supporting the targeted approach, including work by Delgado et al showing the specific depressant effects of serotonin versus catecholamine depletion in patients who had improved when treated with antidepressants. In these studies, patients became depressed again when the neurotransmitter their medication had targeted was depleted but not when the nontargeted neurotransmitter was depleted. More recently, separate studies involving MRI and/or positron-emission tomography conducted in France and at Columbia University in New York correlated brain-scan data with different depressive subtypes, noting, for example, significantly reduced dopamine function in psychomotor-retarded patients.

“This alone would make one question the advisability of an algorithm that says, ‘Start everybody on an SSRI [selective serotonin-reuptake inhibitor],’” Dr. Metzner said. “Serotonergic agents tend to inhibit dopamine after a period of time. And if people start out with low dopamine—you can pick it up clinically, you don’t need to do a quantified EEG, you can see it with your eyes and ears—it might behoove us to be thinking about those symptoms as a reason to start people on a catecholaminergic agent.”

THE OTHER SIDE OF THE FENCE

Are therapeutic strategies such as the targeted treatment of depression supported by the clinical trial evidence? No, said Craig Nelson, MD—the great majority of randomized controlled trials comparing selective serotonin and norepinephrine agents in patients with major depression found that the drugs had nearly identical efficacy, with little evidence suggesting a differential effect on different symptoms or patient subtypes.

To begin with, the number of pertinent studies, in Dr. Nelson’s estimation, is small. Of the 15 double-blind studies that he reviewed comparing serotonin and norepinephrine drugs, 10 studies (involving about 1,300 patients) examined differences in symptom response. “Few differences were found, and those that were described were not consistent or replicated,” he said. “The reason I’m emphasizing this [regarding the 15 studies] is that that’s the database.”

What Dr. Nelson ultimately found is that agents selected for serotonin or norepinephrine reuptake inhibition have similar effects on core symptoms of depression, and similar residual symptoms remain after treatment with these selective agents. “In fact, as some, if not many, individuals have previously suggested, in depressed patients maybe these drugs really do act through a final common pathway,” Dr. Nelson said. “And it’s not the direct effect of the drug [on patients] so much as the effect of the drug on depression that results in similar symptoms.”

He pointed to a slide showing a Venn diagram of a more traditional view of symptom response as it relates to drug type, with relatively little overlap among serotonin, norepinephrine, and dopamine and their effects on such factors as anxiety, impulse, drive, and vigilance. He then contrasted the diagram with his own view of drug type and symptom response, in which the three classes of antidepressant agents appear to have similar effects on such core symptoms as mood, anxiety, and energy while demonstrating different effects on secondary factors.

“We know our antidepressants differ in terms of how sedating they are,” Dr. Nelson said. “They differ in their effects on sleep—though sleep is also a core symptom, so you’re getting some of both—they differ in some of their effects on weight and in their effects on sexual functioning. We generally view all of those effects as secondary effects or side effects, depending, for example, on whether you want the person to gain weight or not.”

WEIGHT OF EVIDENCE FAVORS NONTARGETED APPROACH

According to Dr. Nelson, Professor of Psychiatry at the University of California, San Francisco, clinicians frequently base antidepressant drug selection on their own assumptions about the drug’s profile or its effect on specific symptoms. “Most of the hypotheses about specific effects are based on animal studies or symptom correlates of biological variables,” he observed.

Yet, none of these findings was borne out in his review of the 15 existing comparison studies of selective serotonin and norepinephrine agents, in which the two classes of drugs had the same overall response rates.

“But the question is, Are they treating the same symptoms or different symptoms?” Dr. Nelson asked. “The same question can be asked a different way: Are they treating the same patients or different patients?”

Two other studies do provide evidence supporting symptom differences among antidepressants: a 1995 retrospective study by Filteau and colleagues, who examined symptom predictors in patients who had participated in prior drug studies at their site, finding that anxiety is best treated with a serotonergic agent, and a 2004 study in poststroke depression by Rampello and colleagues suggesting that psychomotor retardation responds better to a norepinephrine agent. “That’s about it, however, in terms of selective agents supporting this hypothesis,” Dr. Nelson emphasized.

Other studies have reported no differences or contradictory findings. Studies include:

• A 1995 trial by Burns and colleagues in which anxiety predicted a good response to lofepramine, a norepinephrine agent, but a poor response to fluoxetine, while psychomotor retardation and lack of energy predicted a poor response to lofepramine but did not predict a response to fluoxetine.

• A 2001 meta-analysis by Rush and colleagues that found that anxiety did not predict response to either bupropion or sertraline.

• Another half-dozen studies in which few symptom differences were found and, of those that were found, none was consistent or replicated.

Finally, Dr. Nelson presented the results of two secondary analyses he and his colleagues conducted of two randomized, double-blind studies comparing fluoxetine, a selective serotonin agent, and reboxetine, a selective norepinephrine reuptake inhibitor, in patients with major depression. Of the many comparisons that were performed, only five differences were statistically significant; given the number of tests performed, chance findings would be expected. “In no case did the symptom that differed in one study differ in the other study,” he said.

Depressed mood, loss of interest, and psychic anxiety showed the greatest change and similar magnitude of change with both drugs. In the second analysis, residual symptoms were examined in patients who responded to treatment. Again, similar residual symptoms were found in patients treated with fluoxetine and reboxetine.

Dr. Nelson discussed possible limitations of these studies, particularly involving the depression scales used. He also held out the possibility of identifying other dimensions or factors that might help to predict the effects of antidepressant drugs.

“When I started, I should confess, I was hoping to find differences and expecting to find them,” Dr. Nelson admitted. “I was hoping to support the validity of that approach, because I think we’re all interested in how we can decide that patient X ought to be treated with this drug rather than another drug. But the data just don’t support that approach at the present time.”

—Fred Balzac

Suggested Reading
American Psychiatric Association. Practice guideline for the treatment of patients with major depressive disorder (revision). Am J Psychiatry. 2000;157(4 suppl):1-45.
Andreoli V, Caillard V, Deo RS, et al. Reboxetine, a new noradrenaline selective antidepressant, is at least as effective as fluoxetine in the treatment of depression. J Clin Psychopharmacol. 2002; 22:393-399.
Burns RA, Lock T, Edwards DR, et al. Predictors of response to amine-specific antidepressants. J Affect Disord. 1995;35:97-106.
Delgado PL. How antidepressants help depression: mechanisms of action and clinical response. J Clin Psychiatry. 2004;65 Suppl 4:25-30.
Filteau MJ, Baruch P, Lapierre YD, et al. SSRIs in anxious-agitated depression: a post-hoc analysis of 279 patients. Int Clin Psychopharmacol. 1995;10:51-54.
Gibbons RD, Clark DC, Kupfer DJ. Exactly what does the Hamilton Depression Rating Scale measure? J Psychiatr Res. 1993;27:259-273.
Kessler RC, Berglund P, Demler O, et al. The epidemiology of major depressive disorder: results from the National Comorbidity Survey Replication (NCS-R). JAMA. 2003;289:3095-3105.
Massana J, Moller HJ, Burrows GD, Montenegro RM. Reboxetine: a double-blind comparison with fluoxetine in major depressive disorder. Int Clin Psychopharmacol. 1999;14:73-80.
Nelson JC, Portera L, Leon AC. Are there differences in the symptoms that respond to a selective serotonin or norepinephrine reuptake inhibitor? Biol Psychiatry. 2005;57:1535-1542.
Rampello L, Alvano A, Chiechio S, et al. An evaluation of efficacy and safety of reboxetine in elderly patients affected by “retarded” poststroke depression: a random, placebo-controlled study. Arch Gerontol Geriatr. 2005;40:275-285.
Rush AJ, Batey SR, Donahue RM, et al. Does pretreatment anxiety predict response to either bupropion SR or sertraline? J Affect Disord. 2001;64:81-87.

Return to table of contents