Psychotic depression. Conrad Swartz shows us how to recognize this underdiagnosed illness, and why antipsychotics and antidepressants may not be the optimal treatment.
Published On: 06/20/2022
Duration: 18 minutes, 27 seconds
Referenced Article: “Psychosis During Depression,” The Carlat Child Psychiatry Report, June 2022
Chris Aiken, MD, Kellie Newsome, PMHNP, and Conrad Swartz, MD, PhD, have disclosed no relevant financial or other interests in any commercial companies pertaining to this educational activity.
Conrad Swartz wrote the book on psychotic depression. Literally. And today he shows us why it’s missed and how to recognize it.
Welcome to the Carlat Psychiatry Podcast, keeping psychiatry honest since 2003. I’m Chris Aiken, the editor in chief of the Carlat Report. And I’m Kellie Newsome, a psychiatric NP and a dedicated reader of every issue.
KELLIE NEWSOME: In 2000, a group of NIMH funded researchers began enrolling patients for what would become the largest prospective clinical trial of major depressive disorder ever conducted. They called it STAR-D, and part of its claim to fame was that the study enrolled patients with only minimal exclusion criteria, patients that would better reflect the ones we see in real life. Among the 4,041 subjects were all kinds of patients we usually don’t see in the aseptic world of clinical trials: Patients with chronic depression, suicidality, medical problems, and even substance use disorders, as long as they didn’t require detoxification.
But some were excluded from STAR-D – those with eating disorders or OCD, for example, and patients with psychotic or bipolar depression. But the STAR-D researchers are a lot like you and me – making tough judgment calls in the face of ambiguous information. Bipolar disorder and psychosis can’t be ruled out definitively – these are diagnoses that often reveal themselves gradually over careful, long-term follow-up. Which raises the question, “What if a few patients with bipolar or psychotic depression snuck into the study?” In 2011 Roy Perlis and colleagues looked into that question, and here’s what they found.
CHRIS AIKEN: Before enrolling in the trial, patients were given 6 screening questions for psychosis, and another 6 for mania. For psychosis, these were questions about being spied on, plotted against, having special powers, being controlled, or events that others said did not occur, as well as hearing voices or seeing things. For mania, they were questions about elevated mood, extreme self-confidence, decreased need for sleep, talkativeness, new projects, and impulsive activities.
These screening tests were followed up with a diagnostic interview, which was ultimately used as the deciding factor in whether or not to enroll the patient. But if you do this in practice – and we recommend you do – you’ll find a lot of cases where patients answer yes on paper but then say no when the same items are explored in a face-to-face interview.
So Dr. Perlis looked back at those who had answered YES for some of the manic or psychotic symptoms but were ultimately enrolled in STAR*D after being cleared with a diagnosis of non-psychotic major depression. And what he found surprised us – it was the soft-psychotic signs, not the soft-bipolar signs – that predicted whether or not a patient would respond to the antidepressant-driven therapies in the STAR*D trial. We would have guessed the opposite – that bipolarity contributed to non-response – but Perlis found no evidence of that, and subsequence STAR*D analysis came up with the same answer – classic markers of bipolarity like family history of bipolar disorder, subsyndromal manic symptoms, and even genetic markers for bipolarity did not predict response.
That’s not to say that bipolar features don’t matter. Other large studies have found that these soft bipolar signs predict a poor response to antidepressants. Few things are black-and-white in psychiatry, and different studies arrive at different conclusions for the same reason that different patients with the same diagnosis can have opposite responses to medications. What this STAR*D analysis did teach us is that we need to pay more attention to psychotic depression.
The common-sense idea about psychotic depression is that these patients are suffering from two syndromes – psychosis, in the form of delusions or hallucinations, and depression – and they need two types of treatments to get well: An antidepressant and an antipsychotics. But this month we spoke with Dr. Conrad Swartz. Dr. Swartz has spent his life studying psychotic depression, and he had a different take.
DR. SWARTZ: It’s pretty frustrating to treat psychotic depression with medications. For males I like to use bupropion. I have not had good success with tricyclics for males, but I have with bupropion (Wellbutrin). If these medications don’t work then I’ll add lithium. After that I don’t have much hope for medications. Of course, if I’m in a situation and I have sometimes practiced in a situation where ECT is not available and I have to use medications then I turn to using antipsychotics sometimes with or without benzos.
CHRIS AIKEN: In addition to the antidepressants?
DR. SWARTZ: Yes, in addition to the antidepressants. So my patients when the ECT is not available my patients with psychotic depression will have quite a list of medications.
TCPR: Okay, so it sounds like antipsychotics are second or third line for you?
DR. SWARTZ: Yes.
CHRIS AIKEN: Dr. Swartz was not surprised when he saw the STAR*D results, has he doesn’t see very good results with medications of any sort in psychotic depression. Instead, he prefers ECT, which has remission rates approaching 100% in psychotic depression.
Wait. Did we just say remission rates approach 100%? OK, that was to get your attention: The actual remission rate of psychotic depression in ECT is 95%. That figure comes from the large, NIMH sponsored “CORE” study of 253 patients with psychotic and non-psychotic depression who underwent ECT – remission rates – not response but remission – were 83% for non-psychotic and 95% for psychotic depression.
But it wasn’t just lack of efficacy that steers him away from antipsychotics – he also suspects they may interfere with functional recovery.
DR. SWARTZ: Yes, well the goal of diagnosis and treatment for psychotic depression is recovery to full premorbid functioning. This is not something that you expect for schizophrenia, for example. This goal is implied with treating any mood disorder, and that’s a big difference from treating schizophrenia, delusional disorder, or even schizoaffective disorder. So achieving this goal as I see it requires minimizing antipsychotic drugs at least the duration of the antipsychotic drugs, avoiding them long term.
there are these impairments caused by antipsychotic drugs. The frontal lobe syndromes: the apathy, the dysexecutive syndrome and rarely the orbitofrontal syndrome. But the dysexecutive syndrome is really pretty common. Patients have difficulty solving problems, dealing with the complexity of interpersonal relationships, taking initiative, multitasking. They become relatively dependent and passive and quiet. And then they don’t see this. They are satisfied taking the antipsychotics. Well they will discontinue them – that’s pretty common, but the ones who don’t discontinue accept taking it and their personalities are changed into being passive.
CHRIS AIKEN: Yeah, that makes sense. Not something that’s readily picked up on in empirical research though.
DR. SWARTZ: That’s right, and even the evaluations of the effects of antipsychotic drugs in treating mania and depression don’t look for these frontal lobe syndromes: the dysexecutive, the orbitofrontal, and the apathy. They talk about somnolence. The olanzapine talks about personality changes, but I think the personality changes they are referring to are actually dysexecutive syndrome and orbitofrontal syndrome that they are describing with euphemisms.
CHRIS AIKEN: Antipsychotics do have some benefits in psychotic depression, but nowhere near the effects of ECT. And if you use them, Dr. Swartz offered up this pearl: You usually need to raise the antipsychotic to the higher range – the kind of dose you’d use for mania or schizophrenia – to see results in psychotic depression. The doses used for antidepressant augmentation are much too small for this illness, so we’re talking olanzapine 10-20 mg, not 5mg; quetiapine 400-800 mg/day, not 150-300mg, and aripiprazole 15-30 mg/day, not 5mg.
I was surprised to hear an expert on psychotic depression question the use of antipsychotics, after all the APA guidelines recommend an antidepressant-antipsychotic combo as first-line therapy in their 2010 guidelines. But Dr. Swartz forced me to take a closer look at those guidelines, and I did not like what I saw.
The APA cited 3 references to support their endorsement of the antipsychotic-antidepressant combination, but only one of those references supports the conclusion, and that one was a small study with so many design flaws – only 18 patients were enrolled for this 3 arm study that used no placebos - that the APA would normally not even mention it. Then the APA acknowledges that there is another study that refutes their conclusion – this one found nortriptyline worked just as well on its own as it did with the antipsychotic perphenazine – this study did use a placebo, and was twice as large as the one the APA sided with, but what about the other 2 studies they used as support?
One was a meta-analysis that found a slightly higher response rate with combination vs. antidepressant monotherapy, but the difference was not statistically significant, and I have never seen the APA draw recommendations from a negative p value. The third was a study comparing antipsychotic monotherapy with antidepressant-antipsychotic combination – concluding that yes antipsychotics don’t work well on their own in this mood disorder, but telling us nothing about antidepressant monotherapy.
The confusion in the APA’s recommendations reflect a more general confusion in the field. A 2012 report found that practice guidelines were all over the place in their recommendations on psychotic depression, as was an opinion poll of psychiatrists. And the confusion is understandable – we don’t have definitive studies in this common condition – which has the same prevalence as schizophrenia. So what have we learned recently?
After the 2010 APA guidelines, an updated meta-analysis came out in 2012 concluding that the antidepressant-antipsychotic combination has a small benefit over monotherapy – the effect size was 0.25 – which is a little larger than the effect of buspirone in generalized anxiety, and a little smaller than the effect of omega-3’s in depression. Arguably, you’d see a similar difference if you compared antipsychotic augmentation with antidepressant monotherapy in non-psychotic depression – so it’s not clear that the antipsychotic is uniquely addressing the illness.
But if you do start an antipsychotic, how long do you need to continue it? That was answered by a recent large trial, the STOP-PD study. They began by randomizing 259 patients with psychotic depression to either sertraline + olanzapine or olanzapine alone. At the end of the 12 week trial, remission rates were nearly double with the combination therapy, which tells us that antipsychotics don’t work well on their own – it says nothing about antidepressant monotherapy because they didn’t test that group.
A more interesting answer came in the second phase of STOP-PD, when they took the patients who responded to combination therapy and randomized them to either discontinue the antipsychotic and stay on sertraline, or stay on both medications after 2 months of recovery. Here are two take-aways from that study. Relapses were much higher 55% vs 20% in those who came off the antipsychotic, but nearly all of the relapses occurred soon after stopping the olanzapine. Why does that matter? Well we generally find that depressive relapses are much higher when antipsychotics are stopped after 2 months of recovery like they did in STOP-PD than if they are stopped after 6 months of recovery, and the fact that nearly all the relapses occurred during that vulnerable period of the first 6 months suggests they might have had better results if they had waited a little longer, and perhaps tapered a bit slower as withdrawal psychosis can occur. But why taper at all? In this study, physical outcomes were much better in those who came off the olanzapine – not just metabolic ones but the fall risk as well.
The bottom line: Antipsychotics and antidepressants can work in psychotic depression, but they are not as effective as we’d like. If you do use them, watch for tolerability problems on the antipsychotic, and if they start to stack up, consider a slow taper – like a 2-3 month taper – after 6 months of recovery.
Among the medication options, lithium and nortriptyline also have empiric support in psychotic depression, and Dr. Swartz has seen good results with both of these. But ECT is the gold-standard here, and these depressions are often severe enough to warrant an ECT referral. The problem, though, is that much like bipolar depression, psychotic depression is hard to diagnose; even in the rigorous setting of an academic medical center, one out of three cases of psychotic depression were missed. In our online interview, Dr. Swartz describes two types of psychotic depression – one where psychosis predominates, which can look like schizophrenia or delusional disorder, and one where depression dominates, which can look like uncomplicated major depression. He also pulls up revealing research about the overlap of trauma, psychotic depression, and PTSD.
KELLIE NEWSOME: Conrad Swartz is Professor Emeritus of Psychiatry at Southern Illinois University in Springfield, Illinois. He is the author, along with Edward Shorter, of the 2012 textbook Psychotic Depression from Cambridge University Press.