Estimating Cardiac 5-HT2B Safety Margins for Repeated Low-Dose Psilocybin Using an Exposure-Response Model

Psilocybin development has mostly focused on single, supervised high-dose administrations, which may be effective but are hard to scale because they require specialist settings and extensive monitoring. The authors note that repeated low, sub-perceptual dosing on an outpatient basis could be more practical, but this shifts the exposure pattern from isolated peaks to frequent receptor engagement over months or years. That change raises an important safety question: chronic serotonergic stimulation can cause drug-induced valvular heart disease through agonism at the cardiac 5-HT2B receptor. The authors argue that earlier concern about psilocybin-related valve risk has relied too heavily on receptor binding affinity and has not adequately accounted for functional efficacy or exposure duration. The paper’s aim is to estimate the cardiac 5-HT2B safety margin of a repeated low-dose psilocybin regimen using a transparent exposure-response model. Tyler and colleagues set out to compare a candidate 3 mg daily psilocybin schedule with drugs and endogenous serotonergic states already linked, or not linked, to valvular disease, and to support the model with pre-existing rat histopathology data. The authors present this as a framework for judging whether repeated low-dose psilocybin could be developed more safely than earlier serotonergic agonists intended for chronic use.

Tyler and colleagues used a pharmacology-based modelling approach rather than a new clinical trial. They built a one-compartment pharmacokinetic model for oral psilocin with first-order absorption and elimination, parameterised using data from their own Phase I single-ascending-dose study in healthy adults (n = 56; 0.5–4.0 mg psilocybin). The model estimated a dose-proportional peak psilocin concentration of about 3.7 ng/mL at 3 mg, with a terminal half-life of about 2.5 h. Matching assay data from a human 5-HT2B G-protein dissociation assay were used to assign potency and efficacy values for psilocin, serotonin, LSD and norfenfluramine; pergolide and cabergoline were included using representative parameters from the literature and were described as lower-confidence inputs. To represent cardiac valve signalling more realistically, the authors modelled psilocin as competing with endogenous serotonin at the receptor rather than acting on an empty receptor. They set a nominal free valvular serotonin concentration of 1 nM and used this in a competitive two-ligand Hill framework. The primary output was ΔTIA, the time-integrated increment in 5-HT2B Gq activation above baseline over a 24 h steady-state dosing interval, expressed in percent-hours per day. They also reported a duty cycle, meaning the number of hours per day with an increment of at least 10 percentage points. A conservative single-ligand sensitivity version that ignored endogenous competition was also calculated. The computations were performed in Python with numerical integration, and sensitivity analyses were used around uncertain quantities such as free serotonin and comparator concentrations. The paper also incorporated pre-existing cardiac histopathology from a sponsored rat study that was not designed for this paper. In that study, male Long-Evans rats received continuous subcutaneous psilocin from osmotic minipumps for 12 days, and the authors obtained hearts from 6 vehicle and 6 psilocybin-pump animals for blinded examination by an independent certified histopathologist. Longitudinal haematoxylin-and-eosin sections of all four valves and of the myocardium, endocardium and epicardium were assessed for degeneration, thickening and other abnormalities.

The modelling results suggested that functional efficacy and exposure duration, rather than binding affinity alone, are the main determinants of 5-HT2B-related valvular risk. Psilocin was characterised as a low-efficacy partial agonist at human 5-HT2B, with an efficacy ceiling of 51.8% of the serotonin maximum in the matched assay, compared with 96.2% for norfenfluramine and 62.1% for LSD. The authors emphasise that a partial agonist cannot exceed its ceiling, so psilocin should be intrinsically limited in the amount of 5-HT2B signalling it can produce. Across the calibrated exposure-response model, every exposure known to cause human valvular heart disease scored ΔTIA at or above +172 %•h/day. Examples included norfenfluramine (+1042), pergolide (+465), daily high-dose cabergoline (+172), and carcinoid heart disease (+1108). Exposures not known to cause valvular heart disease scored at or below +28, including normal serotonin tone (+0), weekly cabergoline (+21) and daily sub-hallucinogenic 10 µg LSD (+28). A candidate 3 mg daily psilocybin regimen scored ΔTIA +3, which the authors describe as roughly two orders of magnitude below the weakest valvulopathic exposure and below even the negative controls in their calibrated scale. Under a conservative model variant that ignored endogenous competition, the margin between 3 mg daily psilocybin and fenfluramine was still about 5.5-fold. The exposure-duration analysis indicated that low-dose psilocin would produce brief daily receptor engagement rather than sustained activation. At the projected 3 mg daily dose, psilocin was modelled as reaching near-zero levels for about twenty hours per day, with no accumulation at steady state. The authors report zero hours per day of sustained elevated 5-HT2B drive for the candidate regimen. By contrast, valvulopathic drugs were characterised by sustained or repeated activation. The paper also noted that daily 10 µg LSD and weekly cabergoline fell within the upper end of the model’s non-valvulopathic range, while 20 µg daily LSD would score higher (+44) but was still said to sit well below the valvulopathic threshold supported by external safety data. The preclinical histopathology was concordant with the model. Rats receiving continuous psilocin for 12 days had stable plasma levels around 8.7–10.0 ng/mL, which the authors state was about 2.4-fold above the projected human peak for 3 mg daily psilocybin. In the blinded pathology review, no lesions were observed in the myocardium, endocardium, epicardium, aortic valve or mitral valve, and the tissues were described as indistinguishable from vehicle controls. The authors report that all cardiac tissues appeared normal and that there was no histological evidence of valvulopathy. They also state that this continuous exposure corresponded to ΔTIA +11 within the model’s safe band.

The authors interpret the combined evidence as indicating that repeated low-dose psilocybin has a favourable cardiac 5-HT2B safety margin. They argue that the main determinants of valvular fibrosis risk are functional receptor efficacy and exposure duration, not receptor binding affinity alone. In their view, psilocin’s low-efficacy partial agonism at 5-HT2B and its short half-life together mean that a 3 mg daily regimen should be far below exposures associated with known human valvular heart disease. They present the rat histopathology, which showed no valve abnormalities after 12 days of continuous psilocin exposure, as supportive but not definitive corroboration. They place these findings in the context of earlier research, noting that a recent human LSD microdosing trial with echocardiographic monitoring found no evidence of valvulopathy after eight weeks of twice-weekly dosing, and that a large observational cohort reported only a small and adjustment-dependent association between lifetime hallucinogen use and valvular heart disease. The authors caution that such epidemiological data cannot establish a causal effect for controlled low-dose psilocin, but they see the overall pattern as reassuring. They also stress that psilocin is less efficacious at 5-HT2B than LSD in their matched assay data, making the null LSD trial a relatively stringent human precedent. Several limitations are acknowledged. First, the assay data used here show psilocin as a low-efficacy partial agonist on the Gq pathway, but at 5-HT2B it appears to recruit β-arrestin2 with higher efficacy than it activates Gq/calcium signalling, so the fibrotic relevance of this arrestin-preferring profile remains uncertain. Second, the rat study covered only 12 days, whereas the intended clinical use is for months to years, so it cannot exclude a slowly emerging fibrotic effect. Third, the model depends on uncertain parameters, including the free psilocin fraction, the free serotonin concentration at the valve, and comparator efficacies from the literature. The authors therefore say the qualitative conclusion seems stable, but the precise safety margin is not fixed. On the basis of these uncertainties, the authors recommend continued clinical investigation with quantified vigilance rather than dismissal of the concern. They state that repeated low-dose psilocybin trials should include prospective echocardiography from baseline onwards to track valve structure, valve function and pulmonary-artery pressures, and that participants with existing valvulopathy or pulmonary hypertension should be excluded. They also suggest that intermittent dosing schedules may reduce cumulative exposure. Overall, they conclude that repeated low-dose psilocybin merits continued, carefully monitored investigation for neuropsychiatric disorders.