Could we develop a fast-acting, serotonin-based antidepressant if we wanted to?
Frustratingly for clinicians and patients, traditional antidepressants that act via serotonin and other monoamine pathways, such as SSRIs, do not work terribly quickly. Even individuals on these drugs who see a large reduction in symptoms after just 3 weeks of treatment still seem to need at least a full week before they experience a meaningful reduction in symptoms. A substantial proportion will not see anything more than token improvements until 6 weeks or later. The delayed onset led to the search for and the development of antidepressant medications, such as ketamine and brexanolone, that begin to show meaningful effects within hours and days, and that act via mechanisms that go beyond the monoamine system.
An intriguing outcome of this development has been that as we have gotten better at creating fast-acting antidepressants, we have also gotten quite good at creating fast-acting placebos. The placebo groups in controlled trials of fast-acting antidepressants (typically delivered intravenously or intranasally) show faster improvement than people usually do on standard oral antidepressants. It’s as if by collectively imagining that fast-acting antidepressants exist, we were able to mimic that effect through the magic of expectancy.
In phase 3 studies of intranasal Spravato (esketamine) for depression with suicidality, patients with placebo—who were also receiving the standard of care, i.e., a newly initiated or optimized oral antidepressant—experienced as much improvement by day 4 as patients experience by week 6 in phase 3 trials of oral antidepressants!
We see a similar story with IV brexanolone for postpartum depression:
This is a massive placebo response! Again, you get more improvement within 72 hours with IV placebo than you get with a standard oral antidepressant in 6 weeks.
This leads me, and many others, to wonder: is there a way of leveraging this fast-acting placebo response to make standard antidepressants such as SSRIs produce faster results? We should at least be able to get as much response as a fast-acting placebo under the right conditions, right? Furthermore, perhaps we can even design the trial and select primary outcomes in a manner that could be used to show statistically significant, even if marginal, separation from placebo? Can we squint at the data in a way that makes it look fast-acting?
For example, we already know that citalopram and paroxetine can show statistically significant separation from placebo after 1 week of treatment on the depressed mood item (although not on the total score of depression rating scales). And we also know that antidepressants show a greater effect size on core symptoms of depression compared to the total score on a depression rating scale such as HDRS-17.
So if we wanted to explore this further, how would we go about it?
We need rapid entry into the brain and rapid achievement of steady-state level, so intravenous (IV) formulation would make sense. The use of IV would also enhance expectancy of effect.
Use the inpatient treatment setting to enroll patients with high symptom severity and to provide a therapeutic environment round the clock with daily assessments. This would enhance expectancy, non-specific therapeutic support, as well as regression to the mean.
The goal would be to show improvement within 3-7 days, with substantial improvement possibly even at 24 hours
Outcomes would emphasize HDRS or MADRS subscores that focus on core depression and anxiety symptoms.
We could also consider 5-HT1A receptor agonism, which can rapidly desensitize autoreceptors in the raphe nucleus and, in theory, counteract the initial suppression of 5-HT neurotransmission following inhibition of the SERT thought to underlie delayed antidepressant response.
It should not surprise us that smart people figured this out a long time ago and that this strategy has already been studied rigorously with IV vortioxetine (Trintellix).