depression: drugs vs CBT

The Agency for Healthcare Research and Quality (AHRQ) just released their clinical review comparing nonpharmacological vs pharmacological treatments for patients with major depressive disorders (see https://www.effectivehealthcare.ahrq.gov/ehc/products/568/2155/major-depressive-disorder-report-151202.pdf​ for the full report, or https://www.effectivehealthcare.ahrq.gov/ehc/products/568/2303/major-depressive-disorder-clinician-160915.pdf for the summary). details:

--they reviewed 44 trials from 1990-2015 of patients with mild to severe major depressive disorder (MDD)
--cognitive behavioral therapy (CBT) is as effective as second-generation antidepressants (SGAs, which for this review includes SSRIs, SNRIs, bupropion , mirtazapine, nefazodone, and trazodone) for mild to severe MDD, with moderate level of evidence. The evidence does not support SGAs plus CBT as being better than SGAs alone, though this was based on low quality of evidence

--therapy (SGA vs CBT) was discontinued at equal rates, for either SGAs or CBT, for adverse effects or patient dissatisfaction:

    --common adverse effects: 

        --for SGAs: nausea/vomiting, diarrhea, sleepiness/fatigue, headache, insomnia, weight gain

        --for psychotherapies: dissatisfaction with treatment

        --Black box warnings by FDA for SGAs: increased risk of suicidal thinking and behavior for kids and up to age 24, during initial treatment

        ​--"patients  of all ages should be monitored appropriately and observed closely for clinical worsening, suicidal thinking and behavior, or unusual changes in behavior"

--no statistically significant benefit of augmenting an SGA with another SGA or non-SGA med (low quality of evidence) or with CBT (moderate quality of evidence)

--limited evidence of benefits and adverse effects for the following, either alone or in combo with SGAs (only a few studies done). However, there is low-level evidence that these interventions may be similar to SGAs (except for the omega-3 fatty acids, which seem to be worse)

    --interpersonal psychotherapy

    --psychodynamic therapy

    --acupuncture

    --omega-3 fatty acids

    --s-adenosyl-L-methionine

    --St John's wort

    --Exercise therapy

--no eligible studies compared SGAs to other therapeutic modalities, such as yoga, humanistic therapy, or mindfulness interventions

--also very limited evidence regarding treatment efficacy in patient subgroups (eg by sex, age, race/ethnicity, medical comorbidities, coexisting psych conditions) and very few studies have looked at long term benefits or adverse effects treatments

commentary:

--MDD affects 16% of US adults over their lifetime, and 7% (17.5 million in 2014) have an episode of MDD each year

--approximately 1/3 of patients with MDD are severely depressed, as measured by their symptom severity, functional impairment and level of distress

--but, about 40% of the patients do not respond to first-step treatment, and 70% do not achieve remission (reinforcing the need for a different treatment strategy, as mentioned in the document above)

--old numbers I have seen suggest that about 80% of depressed patients have their depression treated by primary care clinicians, with SGAs typically being the first option

--there were a couple of studies finding superiority of third-wave CBT, which seems to involve acceptance and commitment treatment, behavioral activation, cognitive behavioral analysis, mindfulness based CBT,... These studies did find them better than SGAs, but were felt not to have sufficient strength of evidence. I should note that this document did not define third-generation CBT, though it seems that the studies on depression focused mostly on "behavioral activation", which I gather is, in the great BF Skinner “operant conditioning” approach, mostly figuring out ways to increase environmental reinforcement and reduce environmental punishment (ie, asking the patient what activities are hardest for them to do, then positively reinforcing patients as they try to perform these activities; another approach is to try to identify "depression loops", when a coping strategy leads to increased depression, such as alcohol/drugs/escape/rumination, then try to identify and implement positive/reinforcing coping strategies)

--my general approach for those patients who are severely depressed is to try an SSRI first (and I tend to use fluoxetine in those with lots of psychomotor retardation, sertraline otherwise, though this is NOT evidence-based and small studies I have seen suggest that this really is not a useful approach…..) And, given the possibility of a significant adverse event (eg suicide attempt, or perhaps worsening in those with an unmasked bipolar disorder), I do see patients back within 1-2 weeks to see how they are doing, realizing that it takes 3-4 weeks to really begin to see an anti-depressant effect. And I do promote counseling/CBT either as an alternative to SGAs or in combination (also, not supported by the literature, as above)

--one very useful aspect of this report is that for us in primary care, we see only some of the studies on depression, since most don’t make it out of the psych literature (and probably a slanted cut of the studies at that, mostly the positive ones). My recollection over the years has been that there was some value to SSRI augmentation (eg, adding another SGA such as bupropion, or adding psychotherapy or other non-med treatment, or adding buspirone, tricyclics or other non-SGA “augmenters”). But on review of this AHRQ compilation, it becomes clear that although there are a few smaller studies which show benefit to SSRI augmentation, this is not a uniform conclusion, and some of the big studies (eg STAR*D) did not have the correct groupings to allow for clear guidance (eg, adding bupropion or buspirone seemed to help, but there was no way to know for sure if the small improvements were from adding the single new agent or from the combination/augmentation strategy). And given the large number of the recommendations in the “low strength-of-evidence” in the document, in some ways the real issue is that we just don’t have good enough studies to be able to promote a clear approach to patients with one of the most common and disabling conditions we treat. And, it would be great to have much more granular data: are there certain groups of patients who are more likely to respond to one drug over another, or some form of counseling vs drugs, or the combo of both???

--not surprisingly, this summary is not so different from the recently-published clinical practice guidelines published by the American College of Physicians (see http://gmodestmedblogs.blogspot.com/2016/02/adult-depression-guidelines.html), which includes some more in-depth commentary