Complexities of treatment-resistant depression: cautionary notes and promising avenues.

Depressive episodes can be of mild intensity and transient, but – especially in tertiary care settings – they are often chronic and/or relapsing. As clinicians we often see people towards the latter end of this spectrum, including “treatment-resistant depression” (TRD), and spend much of our efforts in treating them. McIntyre et al1 competently and comprehensively review the TRD definition, prevalence and management, and portray our ways forward. Here I present a few further perspectives on this topic. First, patients with TRD can still experience spontaneous full remission. In randomized controlled trials examining various pharmacological switching or augmentation strategies, one in four patients remitted even when they were allocated to the control conditions, i.e. continued the same antidepressants on which they had been judged refractory2. There is a paucity of systematic long-term prospective studies on the prognosis of TRD. One small study found that even among patients who were treatment-refractory and were depressed chronically over two years (average: 8.4 years), 8% (95% CI: 3-20) achieved complete remission in the next two years3. When nothing seems to help, hope can still be there. How to get out of TRD is naturally of utmost concern once patients are in there, and McIntyre et al provide a cutting-edge summary of various strategies available today and possibly in the future. As clinicians, however, our first concern should be about how not to let patients get there. In general, pharmacotherapies and psychotherapies are equally efficacious4, with only some small differences among the various antidepressants and no demonstrable differences among the various psychotherapies, as acute phase treatments. However, a recent systematic review and network meta-analysis5 found an important difference between psychotherapies and pharmacotherapies when we aim not only to make the patients well, but also to keep them well. Starting treatment of a new depressive episode with psychotherapies increased the proportion of patients with a sustained response (i.e., responding to the acute phase treatment and maintaining that response) by more than 10 percentage points over starting the treatment with antidepressants and keeping them on these medications after response. Scaling up psychotherapies may be one indirect yet important way to decrease the suffering due to TRD. It may also be important to point out that too aggressive pharmacotherapies may lead to what McIntyre et al call “pseudo-resistance”. Selective serotonin reuptake inhibitors (SSRIs) achieve the optimum balance between efficacy and side effects towards the lower end of their approved dose ranges6 and, despite the commonly accepted clinical wisdom, titrating up the dosage flexibly in view of the patient's response does not increase the response rate7. Dose increase after initial non-response does not help, but only increases dropouts due to side effects. One small caveat is needed concerning discussion of the incidence of TRD. An often cited source for this estimate is the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, arguing that failure of Level 1 and 2 treatments in that study corresponds to the US Food and Drug Administration (FDA)/European Medicines Agency (EMA) criteria for TRD, i.e. inadequate response to two or more antidepressants. The estimate of the TRD incidence is then said to be 55%. However, we should consider that 17% of the patients starting Level 1 treatment in that study had already received some antidepressant medication for the index episode before enrolment, and that the average length of the index episode was already over two years at baseline. The true estimate of non-response to antidepressants in hitherto untreated episodes of major depression could then be lower. Another large trial of antidepressant therapies (the Strategic Use of New generation antidepressants for Depression, SUN☺D) entered only untreated episodes of depression and estimated, after imputation for missingness, the cumulative remission rate to be 37% (95% CI: 35-39) by 9 weeks and 52% (95% CI: 50-54) by 25 weeks8. McIntyre et al highlight the lack of consensus on the definition of TRD. We agree that this makes the identification of risk factors and effective therapies more difficult. However, we also wonder whether such variability in the definition of TRD might be commensurate with the heterogeneity of depression itself. In other words, treatment resistance comes in different colors and in different shades. If the diagnosis of TRD is aimed to indicate the next treatment choices, could different definitions actually suggest different treatments? For example, could TRD type 1 be best treated with added antipsychotics; type 2 with added glutamatergic agents; type 3 by neurostimulation therapies, including electroconvulsive therapy or transcranial magnetic stimulation; and type 4 by combined psychotherapies? When TRD develops into persistent depressive disorder, exploratory analyses suggest that patients’ characteristics – including baseline severity of depression and anxiety, prior treatment and adverse childhood experiences – may moderate the relative efficacy of psychotherapies, pharmacotherapies or their combination9. We are yet far from personalized, differential therapeutics for TRD, but with systematic and consorted efforts we can perhaps reach this level of knowledge within the next decade or two.