When patients do not respond to treatment with one antidepressant, a common clinical strategy is to switch antidepressants. Although a change to an antidepressant of the same class may be efficacious, especially if the second agent offers a milder side-effect burden, most clinicians believe a switch from one antidepressant class to another would yield a greater chance of benefit (Fredman et al. 2000; Nierenberg and Amsterdam 1990). Results from the clinical studies of switching strategies for patients with treatment-resistant unipolar depression are compiled in
How should antidepressants be switched? When patients do not respond to treatment with an antidepressant alone, switching from a TCA to an SSRI or vice versa does not typically require washout but rather a taper of the first antidepressant and initiation of the second one. When time is an issue, clinicians can start the new drug while tapering the first, although the period of overlap increases the possibility of drug interaction side effects. Before switching to an MAOI, clinicians should wait at least 1 week after discontinuing either TCAs (with the exception of protriptyline, which requires 3 weeks of washout) or atypical antidepressants and at least 2 weeks after discontinuing SSRIs (with the exception of fluoxetine, which requires at least 5 weeks of washout). Clinicians switching from an MAOI to an SSRI, a TCA, or an atypical antidepressant should wait at least 2 weeks before starting the second antidepressant. When clinicians switch to an antidepressant of the same class, no washout is necessary, except for the change from phenelzine to tranylcypromine.
How long should the new agent be continued after a response? A reasonable approach is to maintain the patient on the antidepressant for at least 9 months following response and then gradually discontinue the drug. However, if the patient has a history of highly recurrent or chronic depression, consideration should be given to an indefinite duration of maintenance treatment.
What reports suggest efficacy of switching to specific agents? Switching from SSRIs to TCAs Even though the switch to TCAs has also been shown to be effective (response rate = 47%) among SSRI nonresponders in a large randomized and controlled crossover study (Kocsis et al. 1995), the popularity of this strategy has declined because of the improved safety profile of the newer agents.
Switching from TCAs to SSRIs Although systematically ascertained controlled data are limited, the switch to an SSRI from a TCA may prove effective. In an open-label study of 132 patients who had not responded to TCAs or had not tolerated side effects, investigators observed a rate of response to fluoxetine between 51.4% and 62.1%, depending on the definition of response (Beasley et al. 1990). A large crossover study has also shown that 63% of imipramine nonresponders responded to sertraline (Kocsis et al. 1995). Thus, patients who have not responded to TCAs may show a favorable response to an SSRI.
Another SSRI whose efficacy in treatment-resistant depression has been investigated is fluvoxamine. In an open study, 8 (29%) of 28 depressed patients who did not respond to standard TCAs responded to treatment with fluvoxamine alone (Delgado et al. 1988). In a double-blind partial crossover study of 71 patients who had not responded to earlier treatment with TCAs, 10% of patients treated with the SSRI fluvoxamine responded, in contrast to 39% of patients treated with the selective norepinephrine reuptake inhibitor oxaprotiline (Nolen et al. 1988a). In an open study, 9 (82%) of 11 patients who had not responded to desipramine responded to treatment with fluvoxamine (White et al. 1990).
Switching from SSRIs to SSRIs Switching patients from one SSRI to another does not appear to be the first choice for most psychiatrists, according to our recent survey (Fredman et al. 2000). However, this strategy is probably more common among primary care physicians than among psychiatrists and is supported by uncontrolled studies that have shown response rates ranging from 40% to 75% (W. A. Brown and Harrison 1995; Joffe et al. 1996; Thase et al. 1997, 1999; Zarate et al. 1996). However, these studies have several methodological limitations, including the retrospective (and not prospective) determination of nonresponse and the overlap between intolerance and nonresponse.
Switching to bupropion Even though switching to bupropion appears to be a popular strategy among psychiatrists (Fredman et al. 2000), there is very little literature on it. In two small uncontrolled studies, Goodnick et al. (1992) and Walker et al. (1993) found significant improvement on switching SSRI-treated patients to bupropion. The main advantage of this strategy is probably the reduced risk of weight gain and of sexual dysfunction; in fact, the study by Walker et al. (1993) showed improvement in sexual functioning (and depression) on switching to bupropion among 31 patients who had discontinued fluoxetine because of sexual side effects. Finally, an older study by Stern et al. (1983) had shown improvement among patients treated with bupropion after nonresponse to TCAs. In that study, 33 outpatients with a history of nonresponse or nonresponse plus intolerance to TCAs markedly improved after open treatment with the atypical antidepressant bupropion (Stern et al. 1983).
Switching to venlafaxine Nierenberg and colleagues (1994) found a 30%-33% response among 84 consecutive patients with treatment-resistant depression who had undergone at least three failed trials. At lower doses, venlafaxine is more of a selective serotonin reuptake inhibitor than a selective norepinephrine reuptake inhibitor and may work better in TCA nonresponders and MAOI nonresponders than among SSRI nonresponders (de Montigny et al. 1999), but at higher dosages (above 150 mg/day) it appears to be a true dual-uptake inhibitor. A Canadian multicenter study by de Montigny and colleagues (1999) observed a 58% response to venlafaxine among 152 treatment-resistant depressed patients, in further support of the findings by Nierenberg et al. (1994). Further controlled trials are needed.
Switching to nefazodone Thase et al. (1998) recently presented results of a multicenter study in which patients with poor response to SSRIs improved on switching to nefazodone. The main disadvantage of the nefazodone switch is that this is a drug that is at times underdosed in clinical practice due to its need for dose titration. On the other hand, treatment with nefazodone is associated with fewer sexual side effects than is treatment with the SSRIs (Feiger et al. 1996).
Switching to mirtazapine Catterson and Preskorn (1996) found that 59% of 49 amitriptyline nonresponders exhibited good response on switching to mirtazapine in a crossover phase. A multicenter study (M. Fava et al. 1999) has recently shown a 47% response rate to mirtazapine switch (15-45 mg/day) among 103 patients who had not tolerated nor responded to SSRI treatment. The efficacy of mirtazapine was comparable among SSRI nonresponders (n = 76) and SSRI-intolerant patients (n = 18). Sedation and appetite increase/weight gain were the most common side effects, but an interesting advantage of the switch to mirtazapine was that by switching abruptly from the short-acting SSRI paroxetine to mirtazapine there were fewer discontinuation-emergent symptoms than occurred when a washout period was employed. Similarly, the abrupt switch from SSRIs to mirtazapine was equally efficacious as the switch after a brief washout. In addition, there was a significant improvement in sexual functioning in a substantial proportion of patients with SSRI-induced sexual dysfunction (M. Fava et al. 1999).
Switching to reboxetine A multicenter study suggested efficacy for the switch to reboxetine among patients who have not responded to fluoxetine (M. Fava et al. 2000c). A possible advantage of this switch is a potentially distinctive effect on social functioning of reboxetine (Massana 1998).
Switching to MAOIs Switching to MAOIs is a very effective strategy for refractory depression that was popular in the 1970s and 1980s, but nowadays it is typically considered at the end of an algorithm, primarily because of the dietary restrictions and the risk of (spontaneous and not) hypertensive crises. Several studies have examined the efficacy of MAOIs in the treatment of patients who had not responded to TCAs. In an open-label study, 23 (55%) of 42 depressed patients who had not responded to sustained, adequate treatment with imipramine and IPT responded to treatment with either tranylcypromine or phenelzine (Thase et al. 1992). Response rates were greater in patients with anergic/atypical depression (67%) than in those with typical depression (31%), which is consistent with the reported greater efficacy of MAOIs compared with TCAs in atypical depression (Liebowitz et al. 1988). Similarly, in a crossover study of patients with atypical depression, 4 (29%) of 14 nonresponders to phenelzine responded to imipramine, but 17 (65%) of 26 nonresponders to imipramine responded to phenelzine (McGrath et al. 1987).
Of 47 patients with major depression who had not responded to treatment with at least two different antidepressants, 1 (5%) of 22 patients treated with either l-5-hydroxytryptophan or nomifensine responded, whereas 13 (52%) of 25 patients treated with tranylcypromine responded (Nolen et al. 1988b). In addition, 8 (67%) of 12 nonresponders to l-5-hydroxytryptophan and 5 (62%) of 8 nonresponders to nomifensine eventually responded to tranylcypromine in the crossover phase of the study. Thus, the switch from a TCA to an MAOI can be a very efficacious alternative to augmentation for patients with treatment-resistant depression.
Switching to naturally occurring compounds: l-5-hydroxytryptophan; S-adenosyl-l-methionine The efficacy of l-5-hydroxytryptophan, a serotonin precursor, in the treatment of nonresponders was initially suggested in an open trial (Takahashi et al. 1975). However, another report indicated that none (0%) of 12 patients with major depression who had not responded to treatment with at least two antidepressants responded to treatment with l-5-hydroxytryptophan (Nolen et al. 1988b). Finally, S-adenosyl-l-methionine (SAMe), a naturally occurring methyl donor observed to increase serotonergic turnover, was found in an open study to induce remission in 2 (22%) of 9 patients with treatment-resistant depression (Rosenbaum et al. 1990). SAMe has recently become available in the United States as a neutraceutical agent and is frequently used in patients with depression that is resistant to standard antidepressants. Although l-5-hydroxytryptophan and SAMe may be helpful in patients with refractory depression, the efficacy of these agents has not yet been adequately investigated.
Switching to anticonvulsants Of 9 patients with treatment-resistant depression, 4 (44%) showed moderate or marked response after treatment with the anticonvulsant carbamazepine alone (Cullen et al. 1991). In another open study, 11 (92%) of 12 patients with chronic depression resistant to TCAs and MAOIs had a definite improvement after treatment with carbamazepine alone at dosages of 400-600 mg/day (Prasad 1985). In a double-blind, placebo-controlled study, 5 (45%) of 11 unipolar depressed patients responded to carbamazepine after having received other antidepressant treatments that were not effective (Post et al. 1986). There are no controlled, published data on the efficacy of valproic acid or other anticonvulsants in the treatment of patients with refractory depression.
Switching to steroid-suppressing agents Six (75%) of eight patients with treatment-resistant depression who completed 2 months of treatment with one or more steroid-suppressing agents (aminoglutethimide, ketoconazole, and/or metyrapone) showed full response (Murphy et al. 1991). The efficacy of this type of treatment may derive from normalization of hypercortisolism in depression, but the data are very preliminary.
Revision date: July 3, 2011
Last revised: by Janet A. Staessen, MD, PhD