Choosing a First-Line Antidepressant: Title and subTitle BreakEqual on Average Does Not Mean Equal for Everyone

Patients, physicians, health insurers, and pharmaceutical manufacturers all have a considerable interest in the initial selection of an antidepressant drug. Patients and physicians hope to minimize the trial and error needed to find the treatment with the greatest benefit and the fewest risks or adverse effects. Health insurers hope to satisfy patients and insurance purchasers while minimizing drug acquisition costs. Manufacturers hope to demonstrate a unique advantage for a specific drug, either for all patients or specific subgroups of patients. Such an advantage would reduce the need to compete based on lowest price.

The study by Kroenke et al¹ in this issue of THE JOURNAL compared the effectiveness of the 3 most commonly prescribed selective serotonin reuptake inhibitor (SSRI) antidepressants (fluoxetine, paroxetine, and sertraline) among primary care patients initiating antidepressant treatment. In designing this randomized trial, the investigators attempted to replicate the conditions of everyday practice. Participating physicians represented a wide range of community practices. Patients were included based on the treating physicians' decision to prescribe an antidepressant rather than a specific diagnosis or score on a research assessment. Those with comorbid medical illness, anxiety disorder, or alcohol use disorder were not excluded. Following initial randomization treatment assignment, treating physicians were free to adjust, discontinue, or change antidepressant medication as clinically indicated.

Several findings deserve emphasis. First, more than two thirds of the 573 patients beginning treatment recovered during 9 months of follow-up. Second, approximately 20% of patients switched medication one or more times. Third, initial medication assignment had no significant effect on probability of continuing treatment, clinical outcomes, or functional outcomes. Fourth, the relevant patient characteristics of age and anxiety level did not predict better outcomes with any of the 3 drugs.

The primary findings of the study by Kroenke et al are consistent with a large body of previous research. The efficacy and effectiveness of SSRI antidepressants are well established.^{2 - 3} Switching between antidepressant medications is quite common.⁴ Numerous randomized trials have found no significant difference in clinical effectiveness between individual SSRI antidepressants or between SSRI drugs and other classes of antidepressants for outpatient treatment of depression.^{2 - 3} Furthermore, individual patient characteristics do not reliably predict better or worse response to a particular drug or drug class. For example, higher levels of anxiety or insomnia do not predict better response to drugs often considered more sedating.^{5 - 9} One exception to this rule is the superiority of monoamine oxidase inhibitor antidepressants for patients with atypical symptom patterns (oversleeping and overeating).¹⁰ However, given the risks and inconvenience associated with monoamine oxidase inhibitor drugs, this finding has limited relevance for current practice.

Two other issues are important for the choice among antidepressants. First, the fact that SSRI drugs are equally effective on average does not mean that they are equally effective for individual patients. Among patients who do not respond to one SSRI antidepressant, half or more will experience significant benefit from another drug in the same class.^{11 - 12} Similarly, intolerable adverse effects from one SSRI drug do not necessarily predict intolerable adverse effects from a similar medication.^{11 ,13} Second, SSRI antidepressants may differ in potential for drug interactions; fluoxetine and paroxetine are more likely to inhibit action of the cytochrome P450 2D6 enzyme.¹⁴ This potential should be considered when prescribing SSRI drugs to patients also using drugs, such as β-blockers, antipsychotic agents, and some antiarrhythmics, that are metabolized by the P450 2D6 enzyme (approximately 15% of patients treated with antidepressants¹⁵ ). However, adverse events monitoring has not shown differences in rates of clinically significant interactions.¹⁶

During the next decade, selection and prescribing of antidepressant drugs will probably depend more and more on knowledge of drug-gene interactions. Both the dosing of specific antidepressant drugs and the potential for drug interaction may be predicted based on genetic variability in the cytochrome P450 enzyme system.¹⁷ Recent research has also identified significant genetic variability in the serotonin transporter protein, the presumed site of action for SSRI antidepressants. This serotonin receptor polymorphism may explain some of the variability in response to SSRI antidepressants and might eventually be used to guide medication selection.^{18 - 20}

In the meantime, however, randomized trials offer clinicians little specific guidance regarding initial choice of antidepressant medication. For patients with no history of exposure to SSRI antidepressants, expectations regarding probability of clinical response, speed of response, and probability of adverse events do not differ among available SSRI drugs. Individual patient characteristics do not clearly argue for or against prescribing any particular drug. First-line treatment choices should be based on anticipated average drug effects. Selective serotonin reuptake inhibitor antidepressants are, on average, equally effective and equally well tolerated.

Research on antidepressant treatment offers specific guidance regarding care after the initial prescription. Consistent with previous research, less than half of patients in the primary care trial by Kroenke et al experienced a good outcome with the original medication. The remainder discontinued or changed medication because of intolerable adverse effects or lack of significant benefit. With appropriate dose adjustments or medication changes, the majority of these initial treatment failures can be converted to successes. Identifying potential treatment failures and making appropriate adjustments require regular follow-up contact. Although research offers little guidance regarding specific second-line medication options, randomized trials consistently demonstrate the benefits of regular follow-up care, systematic application of evidence-based guidelines, and appropriate use of specialty consultation.^{21 - 23}

The new findings by Kroenke et al,¹ considered along with previous research, also have clear implications for drug coverage or formulary decisions. Given the available evidence, it is difficult to criticize formulary policies recommending or requiring one SSRI drug for first-line treatment. When selecting drugs for patients with no previous exposure to SSRI antidepressants, clinicians cannot reliably predict that one drug will have greater effectiveness, fewer adverse effects, or fewer risks. However, practitioners can reliably predict differences in prescription costs to patients or third parties. All other considerations being equal, an initial choice based on prescription costs is prudent, ethical, and clinically reasonable.

Conversely, it is difficult to defend formulary policies that preclude or restrict use of any SSRI antidepressant for second- or third-line treatment. For a substantial minority of patients, initial treatment with any SSRI drug will prove unsuccessful. Half or more of these patients could be successfully treated with an alternative medication. Given the personal and societal burden of persistent depression, prohibiting the use of such alternatives seems neither ethical nor prudent.