Cognitive effects, pharmacokinetics, and safety of zuranolone administered alone or with alprazolam or ethanol in healthy adults in a phase 1 trial

Abstract

Background:

Zuranolone is an oral, once-daily, 14-day treatment course approved for adults with postpartum depression in the United States.

Aims:

To assess cognitive effects, pharmacokinetics, and safety of zuranolone, alone or with alprazolam/ethanol.

Methods:

This was a phase 1, two-part, two-period, randomized, double-blind, placebo-controlled crossover trial. Participants received zuranolone 50 mg or placebo once daily for 9 days, and additionally received alprazolam (1 mg, Part A), ethanol (males: 0.7 g/kg; females: 0.6 g/kg, Part B), or corresponding placebo on days 1, 5, and 9. Within each part, participants received all treatment combinations. Cognition was assessed using a computerized test battery; pharmacokinetics and safety were also evaluated.

Results:

All participants (Part A, N = 24; Part B, N = 25) received ⩾1 dose of zuranolone/placebo. Compared to placebo, zuranolone produced small-to-moderate cognitive decline (Cohen’s |d| = 0.126–0.76); effects were larger with alprazolam (Cohen’s |d| = 0.523–0.93) and ethanol (Cohen’s |d| = 0.345–0.88). Zuranolone coadministration with alprazolam (Cohen’s |d| = 0.6–1.227) or ethanol (Cohen’s |d| = 0.054–0.5) generally worsened cognitive decline when compared with zuranolone alone. Maximal pharmacodynamic effects occurred at approximately 5 h and were resolved by 12 h postbaseline. No pharmacokinetic interactions were observed. Incidence of adverse events was similar between groups; most events were mild or moderate in severity.

Conclusion:

A general small-to-moderate magnitude decline in cognition occurred with zuranolone alone. Coadministration with alprazolam/ethanol increased the magnitude, but not the duration, of effects compared with single-agent administration. Zuranolone prescribers and patients should be aware of the potential for increased central nervous system-depressant effects if coadministered with GABAergic active compounds such as alprazolam and ethanol.

Introduction

Postpartum depression (PPD), one of the most common complications associated with pregnancy and childbirth (Taple et al., 2022), is defined as a major depressive episode with the onset of symptoms occurring during pregnancy or the postpartum period, which ranges from 4 weeks to 12 months depending on the guideline referenced (ACOG Committee on Clinical Practice Guidelines, 2023; American Psychiatric Association, 2013; El-Den et al., 2022). Individuals with PPD experience changes in maternal behavior, which can be associated with altered family functioning, mother–infant bonding, and emotional and cognitive development of the infant, and may have long-term effects on infant health and well-being (Frieder et al., 2019; Letourneau et al., 2012; Saharoy et al., 2023). PPD is often treated with selective serotonin reuptake inhibitors (Frieder et al., 2019), which are not approved specifically for PPD and typically have delayed onset of action, with time to response varying from 2 to 14 weeks, and only a small proportion of patients achieve response (29%–52%) or remission (28%–46%) (Artigas et al., 2018; Kaufman et al., 2022; Machado-Vieira et al., 2010; Molyneaux et al., 2014; Thase et al., 2001; Trivedi et al., 2006).

Zuranolone is a positive allosteric modulator of both synaptic and extrasynaptic gamma-aminobutyric acid type A (GABA_A) receptors and a neuroactive steroid. Zuranolone is approved as an oral, once-daily, 14-day treatment course for adults with PPD in the United States (ZURZUVAE™ [Prescribing Information], 2024) and under investigation for adults with major depressive disorder. The approval for PPD was based on two phase 3 clinical trials: ROBIN and SKYLARK (Deligiannidis et al., 2021, 2023). As a GABA_A receptor modulator, zuranolone has central nervous system (CNS)-depressant effects (Althaus et al., 2020; Cutler et al., 2023). Consistent with these effects, adverse events of somnolence (15%–27%) and sedation (5%–11%) were observed in phase 3 clinical trials with zuranolone 30 and 50 mg in patients with PPD (Deligiannidis et al., 2021, 2023). Therefore, further understanding of the CNS-depressant effects of zuranolone on other indices of brain function, such as cognition, is warranted, as is the extent to which other GABA receptor-targeting agents may interact with the CNS effects of zuranolone. Two such agents—alprazolam (a GABA_A receptor positive allosteric modulator) and ethanol—influence the CNS through modulation of GABA neurotransmission (Cutler et al., 2023; Davies, 2003). Both agents have been shown to cause impairment in cognition within 1–2 h of dosing that resolves over the ensuing several hours, making them suitable compounds for use in a clinical research setting (Ito et al., 2022; Landry et al., 2022; Pietrzak et al., 2012). Furthermore, the common societal use of both alprazolam and ethanol raises the possibility that they may be used by adults who receive zuranolone. As such, coadministration of zuranolone with these CNS-active medications could lead to increased CNS-depressant effects, including effects on cognition.

Here we report the results from a phase 1 placebo-controlled study that assessed cognitive effects of zuranolone, administered alone or in combination with alprazolam or ethanol. As zuranolone is intended for daily administration for 14 days and reaches a steady state by 5 days (ZURZUVAE™ [Prescribing Information], 2024), the primary aim of the study was to investigate the effect on cognition of repeated daily doses of zuranolone, alone or in combination with a single dose of alprazolam or ethanol administered after 5 and 9 days of daily zuranolone dosing. Because GABA_A receptor modulators can give rise to cognitive impairment following single-dose administration (Ito et al., 2022; Landry et al., 2022; Pietrzak et al., 2012), the effect of single-dose administration of zuranolone on cognition was also included as an exploratory endpoint. Finally, the pharmacokinetics and safety of zuranolone, alone and in combination with alprazolam or ethanol, were also assessed in this study.

Methods

Trial design and treatments

This was a two-part, two-period, randomized, placebo-controlled, crossover trial in healthy adults. In Period 1 of Part A, which followed a screening period of ⩽28 days, participants were randomized to receive zuranolone 50 mg or matching placebo capsules once daily for 9 days. Participants were additionally randomized 1:1 to receive alprazolam 1 mg or matching placebo on days 1, 5, and 9. In Period 2, participants were crossed over to zuranolone or placebo treatment once daily for 9 days and received concomitant alprazolam or placebo on days 1, 5, and 9. Part B followed the same treatment schedule, with ethanol 0.6 g/kg (females) or 0.7 g/kg (males) administered in place of alprazolam. On day 9 of both Periods 1 and 2, participants received the alternate study drug administered on days 1 and 5 (alprazolam or placebo (Part A); ethanol or placebo (Part B)). This design allowed each participant to receive each of the following four treatment regimens: daily doses of zuranolone alone, daily doses of zuranolone with alprazolam or ethanol, alprazolam or ethanol alone, and placebo. Eligible participants completed two 11-day inpatient treatment periods separated by a washout period of at least 7 days (Figure 1).

Figure 1.

Study design.

Note: Participants enrolled in Part A received alprazolam and those in Part B received ethanol. Alprazolam 1 mg, ethanol (males: 0.7 g/kg; females: 0.6 g/kg), or matching placebo was administered as a single dose on days 1, 5, and 9 in each period.

The dose levels of alprazolam and ethanol administered in this study were consistent with those shown to impair performance across a range of cognitive tests in healthy adults (Lowe et al., 2014; Pross et al., 2015; Snyder et al., 2005). Alprazolam and ethanol (or matching placebos) were each administered so that the approximate time of maximum systemic concentrations would coincide with the time of zuranolone maximum systemic concentrations. For both parts, zuranolone or matching placebo was administered at baseline (0 h). In Part A, alprazolam (or matching placebo) was administered 4 h after zuranolone (or matching placebo), and in Part B, ethanol (or matching placebo) was administered 4.5 h following zuranolone (or matching placebo) administration.

Parts A and B were conducted at separate study sites. Participants, site personnel (except pharmacist or qualified designee), and the sponsor were blinded to the treatment sequence. Randomization schedules for each part were generated by an independent statistician and provided to the site pharmacist or designee to facilitate the preparation and dispensation of the study drug. The randomization schedules were kept confidential and accessible only to authorized personnel until the time of unblinding. The study was unblinded after the database had been locked.

Participants

Participants could be enrolled in either Part A or Part B but not both. Following the screening and completion of the informed consent form, patients fulfilling prespecified eligibility criteria were randomized to the treatment sequence. Eligible participants were adults aged ⩾18 years for Part A (⩾21 years for Part B) and considered by the investigator to be in good general health. Key exclusion criteria included a history of suicidal behavior within 24 months or, in the opinion of the investigator, risk of suicide based on their answers to questions on the Columbia-Suicide Severity Rating Scale (C-SSRS) at screening. Additional inclusion and exclusion criteria are available in the Supplemental Appendix.

Assessments

Cognition was assessed using a set of neuropsychological tests from the Cogstate Early Phase Battery that measured five domains of cognition: (1) Detection (DET) test to measure psychomotor function; (2) Identification (IDN) test to measure attention; (3) One Card Learning (OCL) test to measure visual learning and memory; (4) International Digit Symbol Substitution Test (IDSST) to measure processing speed; and (5) Groton Maze Learning (GML) test to measure executive function (Davis et al., 2021; Ito et al., 2022; Williamson et al., 2022). Participants completed the cognitive test battery on days 1, 5, and 9 prior to receiving zuranolone (i.e., baseline), and at 5, 6, 8, and 12 h following zuranolone administration (corresponding to 1, 2, 4, and 8 h following alprazolam dosing and 0.5, 1.5, 3.5, and 7.5 h following ethanol dosing). At each scheduled assessment, participants completed all cognitive tests using a laptop computer. A practice assessment of the entire battery occurred on day −1 of each period, allowing participants to familiarize themselves with the tests and minimize potential practice effects at the subsequent assessment time points. Test data from practice assessments were discarded.

Pharmacokinetics

Blood samples for determining plasma zuranolone concentrations were collected on days 5 and 9 of both study parts prior to dosing and every hour for the first 6 h and then at 8, 12, 16, and 24 h postdose. In Part A, blood samples were collected on days 5 and 9 for plasma alprazolam concentrations at the following time points relative to alprazolam dosing: predose and 0.5 h, every hour for the first 6 h, and then 8, 12, and 20 h postdose. In Part B, blood samples were collected on days 5 and 9 for ethanol concentrations prior to ethanol dosing and at 0.5, 1, 1.5, 2, 3.5, 5, 7.5, and 11.5 h after ethanol dosing. Noncompartmental pharmacokinetic parameters were determined for zuranolone, alprazolam, and ethanol, including maximum observed concentration (C_max), time to reach C_max after dosing (t_max), and area under the concentration–time curve from time 0 to 24 h (AUC_0–24; zuranolone) or from time 0 to the time of the last quantifiable concentration (AUC_0–last; alprazolam and ethanol).

Safety and tolerability

Safety and tolerability were assessed throughout the study with respect to the frequency, nature, and severity of treatment-emergent adverse events (TEAEs). Sedation was evaluated on days 1, 5, and 9 at the same time points as the cognitive battery and at 24 h following zuranolone administration (days 6 and 10; corresponding to 20 h following alprazolam dosing and 19.5 h following ethanol dosing on days 5 and 9 (days 6 and 10, respectively)) using the Modified Observer’s Assessment of Alertness/Sedation (MOAA/S) scale, a validated 6-point scale ranging from 0 (“no response after painful trapezius squeeze”) to 5 (“responds readily to name spoken in normal tone”) (Chernik et al., 1990). Additional safety evaluations included the C-SSRS, clinical laboratory measurements, vital signs, end-tidal carbon dioxide (EtCO₂) levels, and 12-lead electrocardiograms (ECGs). At 7 ± 2 days after discharge following Period 2, participants received a phone call from the site to evaluate the presence of any TEAEs and collect information on any new concomitant medication or procedure.

Statistical analysis

Approximately 24 participants were to be randomized to each part (48 total participants) to have at least 20 participants complete both periods within each part. A priori power computations estimated that 20 participants would be sufficient to detect effect sizes of 0.67, assuming 80% power, two-tailed hypotheses, and the level of statistical significance set at 0.05. Data from Parts A and B were analyzed separately. Data within each part were pooled for treatment summaries and statistical comparisons, as needed.

The primary aim of the study was to determine the effect of multiple days of dosing of zuranolone on cognition. Data from days 5 and 9 were used for the analysis, which occurred in three stages for each study part. First, the mean change from baseline performance score for each cognitive test over the 12-h testing interval was plotted and inspected to determine the time point at which the maximum effect occurred. The maximal effects occurred at the 5- or 6-h assessment; therefore, in the second step, the test scores at the 5- and 6-h time points were analyzed independently for each test in each study part using a mixed-effect model for repeated measures. The cognitive test score was the dependent variable, with period, treatment, day, time point, treatment-by-period interaction, treatment-by-day interaction, and treatment-by-time point interaction as fixed effects and participant as a random effect. The predose test score was included as a covariate. The difference in the least squares means, 95% confidence interval (CI), and associated p value were derived for each comparison of interest. In Part A, the treatment comparisons of interest were zuranolone compared with placebo, alprazolam compared with placebo, and zuranolone plus alprazolam compared with each drug administered alone. Similar treatment comparisons were conducted in Part B with ethanol in place of alprazolam. Because this is a new drug and all information is important, no adjustment was made for multiple comparisons. Third, for each comparison of interest, the standardized mean difference was computed to provide an estimate of effect size (Cohen’s d) to facilitate interpretation of the effect and to reduce the risk of Type I error. Effect sizes for the DET, IDN, and GML tests were multiplied by (−1) for consistency with other measures such that a value <0 represents impairment for each cognitive test. The magnitude of effect sizes was classified as follows: |d| < 0.2 trivial, 0.2–0.5 small, >0.5–0.8 moderate, and >0.8 large (Cohen, 1988). No outcome data were excluded from the analysis and reporting. Missing data were not imputed.

The effects of a single dose of zuranolone, alone and in combination with alprazolam or ethanol, on day 1 were evaluated in an exploratory manner using similar methods as described above.

Pharmacokinetic interactions were evaluated using separate mixed-effects models on log-transformed C_max and AUC_0–24 (zuranolone) or AUC_0–last (alprazolam and ethanol), with treatment, period, and treatment-by-period as fixed effects and participant as a random effect. For each analysis, combination administration was considered the test treatment, and single-agent administration was considered the reference treatment. The differences between test and reference treatments and 90% CIs were back-transformed to obtain the ratios and 90% CI for each comparison in Part A (i.e., zuranolone ± alprazolam and alprazolam ± zuranolone) and Part B (i.e., zuranolone ± ethanol and ethanol ± zuranolone). All safety observations were summarized descriptively by treatment. Additional statistical methods are available in the Supplemental Appendix.

Trial oversight

The trial was conducted according to Good Clinical Practice guidelines developed by the International Council for Harmonisation and in compliance with the trial protocol. The trial protocol was approved by the central institutional review board. All participants provided written informed consent per Declaration of Helsinki principles.

Results

Participants

The trial was conducted from July to November 2021 with a database lock in December 2021. Of the 84 participants who consented to the study (Part A, N = 39; Part B, N = 45), 24 in Part A and 25 in Part B who met eligibility criteria were randomized and received at least one dose of the study drug. Most participants completed the study (Part A: 19/24 (79.2%); Part B: 23/25 (92.0%)). In Part A, five participants (20.8%) withdrew from the study prematurely due to an adverse event and discontinued treatment. In Part B, two participants (8.0%) withdrew prematurely (“withdrawal by the participant”) and discontinued treatment (Supplemental Figure S1).

Baseline participant characteristics were generally comparable in both parts of the study, with a younger median age (49.0 (range 18–79) years) in Part A compared with Part B (53.0 (range 26–68) years), as expected, owing to the differences in eligibility ages between parts (Table 1).

Table 1.

Baseline participant demographics.

Characteristic	Part A (alprazolam), N = 24	Part B (ethanol), N = 25
Age (years), median (range)	49.0 (18–79)	53.0 (26–68)
Sex, n (%)
Male	13 (54.2)	15 (60.0)
Female	11 (45.8)	10 (40.0)
Ethnicity, n (%)
Not Hispanic or Latino	16 (66.7)	18 (72.0)
Hispanic or Latino	8 (33.3)	7 (28.0)
Race, n (%)
White	14 (58.3)	9 (36.0)
Black or African American	9 (37.5)	15 (60.0)
Asian	1 (4.2)	0
Multiple	0	1 (4.0)
Body mass index (kg/m2), median (range)	26.8 (22–33)	28.0 (21–33)

Effect of repeat-dose zuranolone on cognition

The mean change values from baseline scores for each cognitive test following repeated administration of zuranolone in Part A and Part B are shown in Figures 2 and 3, respectively. The effects of single agent or coadministration of study drugs on cognition were similar across Parts A and B; stable performance was observed with placebo administration across all postbaseline time points. The greatest decline in cognition under zuranolone, alprazolam, and/or ethanol treatment conditions occurred at the 5- or 6-h time point (Figures 2 and 3), which coincided with the time of maximal plasma concentrations for each study drug. Changes from baseline in Cogstate battery scores were generally greater (i.e., worsening cognition) following coadministration of zuranolone and alprazolam compared with each study drug alone. Coadministration with ethanol resulted in slightly greater changes in Cogstate battery scores from baseline compared with ethanol alone; changes were greater compared to zuranolone alone. Across all cognitive tests and treatments, performance eventually returned to baseline, with all observed impairments resolving by 12 h postbaseline.

Figure 2.

Time course of the mean (SD) change from baseline for the Cogstate battery—Part A. (a) Detection test (psychomotor function), (b) identification test (simple attention), (c) One Card Learning test (visual episodic memory), (d) International Digit Symbol Substitution Test (processing speed), and (e) Groton Maze Learning test (executive function).

SD: standard deviation; t_max: time to reach maximum observed concentration after dosing.

^aTime of zuranolone or placebo administration (baseline (0 h)).

^bTime of alprazolam administration (4 h).

Figure 3.

Time course of the mean (SD) change from baseline for the Cogstate battery—Part B. (a) Detection test (psychomotor function), (b) identification test (simple attention), (c) One Card Learning test (visual episodic memory), (d) International Digit Symbol Substitution Test (processing speed), and (e) Groton Maze Learning test (executive function).

SD: standard deviation; t_max: time to reach maximum observed concentration after dosing.

^aTime of zuranolone or placebo administration (baseline (0 h)).

^bTime of ethanol administration (4.5 h).

The comparisons of interest for each cognitive test at the 5-h time point are summarized in Table 2, and the associated effect sizes are summarized in Figure 4. Compared with placebo, repeat dosing with zuranolone resulted in a decline in performance on each cognitive test, with the exception of the DET test in Part A. However, except for the IDSST, the zuranolone-related decline was less than that observed following treatment with alprazolam or ethanol and was not large enough in magnitude to reach statistical significance. Performance on each of the cognitive tests declined significantly in Part A following alprazolam administration compared with placebo, with differences yielding effect sizes that ranged in magnitude from 0.5 (IDN) to 0.9 (IDSST; Table 2 and Figure 4). Similarly, ethanol administration was associated with a significant decline in performance on each cognitive test in Part B, with corresponding effect sizes ranging from 0.3 (DET) to 0.9 (IDSST).

Table 2.

Performance on cognitive tests at the 5-h time point on days 5 and 9.

Part A
Test	Zuranolone vs placebo		Alprazolam vs placebo		Zuranolone and alprazolam vs zuranolone		Zuranolone and alprazolam vs alprazolam
	LSMD (95% CI)	ES a	LSMD (95% CI)	ES a	LSMD (95% CI)	ES a	LSMD (95% CI)	ES a
Detection (psychomotor function)	−0.027 (−0.078, 0.024)	0.126	0.54 b (0.001, 0.106)	−0.536	0.139 b (0.073, 0.205)	−1.227	0.059 (−0.008, 0.126)	−0.503
Identification (simple attention)	0.034 (−0.006, 0.074)	−0.377	0.050 b (0.011, 0.090)	−0.523	0.102 b (0.048, 0.156)	−0.727	0.086 b (0.038, 0.133)	−0.662
One Card Learning (visual episodic memory)	−0.040 (−0.114, 0.033)	−0.374	−0.089 b (−0.157, −0.020)	−0.589	−0.134 b (−0.207, −0.061)	−0.874	−0.086 b (−0.170, −0.002)	–0.648
IDSST (processing speed)	−9.9 b (−13.2, −6.5)	−0.76	−12.9 b (−16.4, −9.4)	−0.93	−12.5 b (−18.6, −6.3)	−0.80	−9.4 b (−15.6, −3.2)	−0.65
Groton Maze Learning (executive function)	31.2 (−5.8, 68.1)	−0.35	59.3 b (39.6, 81.8)	−0.66	49.8 b (6.5, 93.1)	−0.60	21.6 (−11.7, 54.9)	−0.22
Part B
	Zuranolone vs placebo		Ethanol vs placebo		Zuranolone and ethanol vs zuranolone		Zuranolone and ethanol vs ethanol
	LSMD (95% CI)	ES a	LSMD (95% CI)	ES a	LSMD (95% CI)	ES a	LSMD (95% CI)	ES a
Detection (psychomotor function)	0.008 (−0.046, 0.063)	−0.134	0.060 b (0.011, 0.110)	−0.366	0.075 b (0.029, 0.122)	−0.496	0.023 (−0.027, 0.073)	−0.279
Identification (simple attention)	0.028 (−0.012, 0.068)	−0.395	0.054 b (0.15, 0.094)	−0.345	0.056 b (0.014, 0.098)	−0.408	0.029 (−0.014, 0.072)	−0.400
One Card Learning (visual episodic memory)	−0.068 (−0.141, 0.004)	−0.425	−0.086 b (−0.153, −0.018)	−0.491	−0.026 (−0.087, 0.035)	−0.054	−0.008 (−0.072, 0.055)	−0.000
IDSST (processing speed)	−5.5 b (−10.4, −0.7)	−0.57	−10.2 b (−15.6, −4.9)	−0.88	−5.5 b (−10.3, −0.7)	−0.42	−0.8 (−5.4, 3.9)	−0.32
Groton Maze Learning (executive function)	−0.9 (−23.6, 21.9)	−0.17	18.9 b (2.9, 35.0)	−0.39	34.9 b (10.6, 59.2)	−0.50	15.1 (−1.2, 31.4)	−0.32

CI: confidence interval; ES: effect size; IDSST: International Digit Symbol Substitution Test; LSMD: least squares means difference.

^aBased on Cohen’s d; d < 0 signals impairment in all tests. |d| < 0.2 considered trivial, 0.2–0.5 small, >0.5–0.8 moderate, and >0.8 large.

^bp < 0.05 (not adjusted for multiplicity).

Figure 4.

Effect size for the treatment difference at the 5-h time point. (a) Part A (zuranolone with and without alprazolam) and (b) Part B (zuranolone with and without ethanol).

Note: Effect size, based on Cohen’s d; d < 0 signals impairment for all cognitive tests. |d| < 0.2 considered trivial, 0.2–0.5 small, >0.5–0.8 moderate, and >0.8 large.

Compared with zuranolone alone, at 5 h postbaseline, coadministration of zuranolone and alprazolam resulted in a decline in performance on each cognitive test that was statistically significant and moderate to large in magnitude (Table 2 and Figure 4). Similarly, compared with zuranolone alone, administration of zuranolone with ethanol resulted in a statistically significant decline that was small to moderate in magnitude on all tests except the OCL. Finally, zuranolone plus alprazolam was also associated with a statistically significant decline in cognition that was moderate in magnitude for the IDN and OCL tests as well as IDSST when compared to alprazolam alone. However, when compared with ethanol, treatment with zuranolone plus ethanol resulted in no statistically significant increase in cognitive decline for any test at 5 h postbaseline, with the magnitudes of differences between treatment conditions for each test small to trivial.

Differences observed for the comparisons of interest for the 6-h time point are provided in Supplemental Tables S1 (Part A) and S2 (Part B). The differences observed at the 6-h assessment were generally similar in magnitude or smaller than those at 5 h. There were two instances in which differences at the 6-h assessment were greater than those observed for the same comparison at the 5-h assessment. First, for the GML, compared with alprazolam alone, treatment with zuranolone plus alprazolam resulted in a statistically significant and large-magnitude difference (Cohen’s |d| = 0.85). Second, for the OCL, compared with ethanol alone, treatment with zuranolone plus ethanol resulted in a statistically significant moderate difference (Cohen’s |d| = 0.43).

Exploration of single-dose zuranolone on cognition

Given the study design, the sample size for each treatment on day 1 was 12, with the number of participants receiving both treatments for each comparison of interest no greater than 6. Because of the smaller sample size, and because the study was designed to detect treatment effects of repeat zuranolone dosing, inferences about treatment effects on day 1 were based on the effect sizes for the comparisons of interest, rather than on the statistical significance. Compared with placebo, a single dose of zuranolone was associated with a small to moderate decline on each cognitive test, with the largest effect size observed for the IDSST (Cohen’s |d| = 0.82 at the 5-h time point in Part A; Supplemental Tables S3 and S4). A single dose of zuranolone was also associated with cognitive decline that was greatest at the 5-h time point and generally equivalent in magnitude to that observed after repeated administration. The cognitive decline seen with zuranolone plus alprazolam or zuranolone plus ethanol tended to be greater than that with zuranolone alone or with alprazolam or ethanol alone. Again, all drug-related cognitive decline had resolved by the 12-h assessment.

Pharmacokinetics

Zuranolone plasma concentrations over time were comparable in the zuranolone and zuranolone plus alprazolam groups (Figure 5(a)) and in the zuranolone and zuranolone plus ethanol groups (Figure 5(b)). Similarly, concentrations of alprazolam over time were comparable in the alprazolam and zuranolone plus alprazolam groups (Figure 5(c)), and concentrations of ethanol over time were comparable in the ethanol and zuranolone plus ethanol groups (Figure 5(d)). Zuranolone and alprazolam concentrations were measurable in all participants throughout the sampling period, with the exception of one participant in Part B, in whom the zuranolone concentration was below 1.00 ng/mL (lower limit of detection of the assay) at 24 h after that participant received zuranolone. Ethanol concentrations declined rapidly after reaching the peak 1 h postdose, and mean concentrations were below the assay detection limit (10 mg/dL) 3.5 h postdose. The pharmacokinetic parameters for zuranolone, alprazolam, and ethanol are summarized in Table 3. The majority of geometric least squares mean ratios were between 0.9 and 1.00, with 90% CIs within the 0.8–1.25 range. Treatment with zuranolone resulted in a slightly lower C_max (14%) and AUC_last (19%) for ethanol, which was deemed not clinically important. Overall, no relevant changes in pharmacokinetic parameters of zuranolone, alprazolam, or ethanol were observed.

Figure 5.

Mean (SD) zuranolone, alprazolam, and ethanol concentrations versus time. Error bars indicate SD. (a) Zuranolone (Part A), (b) zuranolone (Part B), (c) alprazolam (Part A), and (d) ethanol (Part B).

SD: standard deviation.

Table 3.

Pharmacokinetic parameters for zuranolone, alprazolam, and ethanol.

Zuranolone parameter	Zuranolone		Zuranolone + alprazolam		GMR a	90% CI
	N	LS means	N	LS means
Effect of alprazolam on zuranolone
Cmax (ng/mL)	21	106.40	21	98.23	0.923	(0.776, 1.098)
AUC0–24 (ng∙h/mL)	20	1407	21	1330	0.946	(0.827, 1.081)
Zuranolone parameter	Zuranolone		Zuranolone + ethanol		GMR	90% CI
	N	LS means	N	LS means
Effect of ethanol on zuranolone
Cmax (ng/mL)	24	93.50	24	95.77	1.024	(0.910, 1.152)
AUC0–24 (ng∙h/mL)	23	1326	24	1332	1.004	(0.914, 1.103)
Alprazolam parameter	Alprazolam		Alprazolam + zuranolone		GMR	90% CI
	N	LS means	N	LS means
Effect of zuranolone on alprazolam
Cmax (ng/mL)	21	13.50	21	12.90	0.956	(0.864, 1.057)
AUClast (ng∙h/mL)	21	159.1	21	157.6	0.991	(0.952, 1.031)
Ethanol parameter	Ethanol		Ethanol + zuranolone		GMR	90% CI
	N	LS means	N	LS means
Effect of zuranolone on ethanol
Cmax (mg/dL)	24	71.20	24	61.50	0.865	(0.789, 0.947)
AUClast (mg∙h/dL)	24	170.9	24	138.7	0.812	(0.713, 0.923)

AUC_0–24: area under the curve from time 0 to 24 h; AUC_last: area under the curve from time 0 to the last measurable concentration; CI: confidence interval; C_max: maximum concentration; GMR: geometric mean ratio; LS: least squares.

^aCombination/single agent.

Safety

In Part A, 20 of 24 participants (83.3%) had at least one TEAE, with the majority (18/20 (90%)) reporting mild or moderate events. The administration of alprazolam with zuranolone did not meaningfully increase the incidence of TEAEs compared with zuranolone alone (52.2% vs 50.0%; Table 4). The incidence and severity of TEAEs were greater with zuranolone plus alprazolam than with alprazolam alone; two participants who received zuranolone plus alprazolam had severe TEAEs (anxiety (considered zuranolone related), n = 1; substance-induced psychotic disorder (not zuranolone related), n = 1). The TEAE of a substance-induced psychotic disorder (verbatim term: exacerbation of methamphetamine psychosis) occurred 9 days after administration of zuranolone plus alprazolam and was considered serious. Upon clinical evaluation, the participant had an ongoing history of methamphetamine-induced psychotic disorder that was not disclosed at screening. Five participants had at least one TEAE that resulted in treatment discontinuation and withdrawal from the study. Four were in the zuranolone plus alprazolam group, with the following TEAEs reported: anxiety (n = 1); tremor, dysarthria, and vertigo (n = 1); disorientation (n = 1); and substance-induced psychotic disorder (n = 1). One participant who received zuranolone alone reported dizziness and somnolence.

Table 4.

Summary of treatment-emergent adverse events, experienced by two or more participants in any treatment group.

n (%)	Part A				Part B
	Zuranolone alone (n = 22)	Alprazolam alone (n = 21)	Zuranolone and alprazolam (n = 23)	Placebo (n = 21)	Zuranolone alone (n = 24)	Ethanol alone (n = 24)	Zuranolone and ethanol (n = 24)	Placebo (n = 24)
Any TEAE	11 (50.0)	4 (19.0)	12 (52.2)	1 (4.8)	5 (20.8)	2 (8.3)	5 (20.8)	1 (4.2)
TEAE by highest severity
Mild	5 (22.7)	3 (14.3)	8 (34.8)	1 (4.8)	2 (8.3)	2 (8.3)	1 (4.2)	0
Moderate	6 (27.3)	1 (4.8)	2 (8.7)	0	3 (12.5)	0	4 (16.7)	1 (4.2)
Severe	0	0	2 (8.7)	0	0	0	0	0
Nervous system disorders
Somnolence	9 (40.9)	2 (9.5)	9 (39.1)	1 (4.8)	0	0	0	0
Dizziness	3 (13.6)	0	0	0	4 (16.7)	0	3 (12.5)	0
Headache	1 (4.5)	2 (9.5)	0	0	0	1 (4.2)	0	0
Tremor	1 (4.5)	0	2 (8.7)	0	0	0	0	0
Lethargy	0	0	1 (4.3)	0	2 (8.3)	0	2 (8.3)	0
Gastrointestinal disorders
Diarrhea	0	0	2 (8.7)	0	0	0	0	0
Nausea	0	0	0	0	0	1 (4.2)	3 (12.5)	0

TEAE: treatment-emergent adverse event.

In Part B, 10 of 25 (40%) participants had at least one TEAE, all of which were mild or moderate in severity, with none leading to discontinuation. Zuranolone plus ethanol compared with zuranolone alone did not meaningfully increase the incidence (20.8% in both groups) or severity (mild, 4.2% vs 8.3%; moderate, 16.7% vs 12.5%, respectively) of TEAEs. The incidence and severity of TEAEs were greater with zuranolone plus ethanol than with ethanol alone (incidence: 20.8% vs 8.3%; severity: mild, 4.2% vs 8.3%; moderate, 16.7% vs 0%, respectively).

TEAEs experienced by at least two participants in any treatment group are shown in Table 4. The three most common TEAEs were somnolence, dizziness, and headache in Part A and dizziness, nausea, and lethargy in Part B. Of note, in both Part A and Part B, gastrointestinal disorders were reported only with alprazolam alone or ethanol alone or when either drug was combined with zuranolone.

The 6-point MOAA/S scale was used to evaluate the sedating effects of the study treatments. In Part A, at days 5 and 9 across all treatments and time points including baseline, most participants had an MOAA/S score of 5 (“responds readily to name spoke in normal tone”; Figure 6(a)). The lowest score recorded was 2 (“responds only after mild prodding or shaking”), observed in a single participant 1 h after treatment with alprazolam. All except three participants (two with zuranolone and one with alprazolam) had a score of 5 by 12 h postbaseline. In Part B, across nearly all treatments and time points including baseline, the majority of participants had an MOAA/S score of 5 (Figure 6(b)). The lowest score recorded was 1 (“responds only to painful trapezius squeeze”), observed in a single participant 5 h after receiving a placebo. Upon repeat assessment 15 min later, the MOAA/S score in this participant was 3 (“responds only after the name is called loudly and/or repeatedly”). Overall, most participants in Part A and Part B showed little or no signs of reduced alertness on MOAA/S scores.

Figure 6.

MOAA/S. Day 5 and 9 data. (a) Part A and (b) Part B. Score responsiveness: 5 Responds readily to name spoken in normal tone; 4 Lethargic response to name spoken in normal tone; 3 Responds only after name is called loudly and/or repeatedly; 2 Responds only after mild prodding or shaking; 1 Responds only after painful trapezius squeeze; 0 No response after painful trapezius squeeze. Times are reported relative to baseline (i.e., zuranolone administration), which correspond to 1, 2, 4, 8, and 20 h following alprazolam dosing and 0.5, 1.5, 3.5, 7.5, and 19.5 h following ethanol dosing.

MOAA/S: Modified Observer’s Assessment of Alertness/Sedation.

No clinically meaningful changes in respiratory rate or oxygenation and no clinically meaningful differences in mean EtCO₂ values, as measured by capnography, were observed among the treatment groups. Clinical chemistry, hematology, urinalysis, coagulation, vital signs, and ECG parameters were all unchanged across treatment groups. No participants in Parts A or B reported suicidal ideation, suicidal behavior, or non-suicidal self-injurious behavior during their lifetime, in the past 24 months, or at any time during the study, as assessed by the C-SSRS (data not shown).

Discussion

This study evaluated the effects of single or repeat doses of zuranolone, with or without alprazolam or ethanol, on cognition in healthy adults. The study design allowed the effects of zuranolone to be evaluated in two independent cohorts (Parts A and B). The findings from these two cohorts were qualitatively and quantitatively similar, showing that repeated dosing of zuranolone was associated with a general decline in each aspect of cognition assessed, which reached a maximum at 5–6 h postdose, coinciding with the t_max, and had then resolved by 12 h after dosing. By convention, the effects of zuranolone on cognition were classified as being trivial to small in magnitude (i.e., Cohen’s |d| < 0.5), with the exception of a moderate decline observed for processing speed in each cohort (Cohen’s |d| = 0.76 (Part A) and 0.57 (Part B)). A single dose of zuranolone was also associated with cognitive decline that was greatest at the 5-h time point and equivalent in magnitude to that observed after repeated doses, again with the greatest decline observed for processing speed.

The study results also demonstrated that the coadministration of alprazolam or ethanol with zuranolone exacerbated the cognitive decline when compared to zuranolone alone. Coadministration of zuranolone with alprazolam gave rise to moderate to large effects (Cohen’s |d| = 0.6–1.227) compared to zuranolone, whereas the effects on cognition were less (Cohen’s |d| = 0.054–0.5) following coadministration of zuranolone with ethanol compared to zuranolone. Similarly, when compared to alprazolam alone, the coadministration of zuranolone and alprazolam was associated with a moderate and generally greater cognitive decline. However, when compared to ethanol alone, coadministration of zuranolone and ethanol resulted in only a small incremental decline in cognition. Thus, while the magnitude of decline varied across these comparisons, coadministration of two GABAergic active compounds (i.e., CNS-depressant drugs) generally increased cognitive decline versus single-drug administration.

Each study part included a treatment arm with either alprazolam or ethanol, administered at doses for which the pharmacokinetic and pharmacodynamic effects are well described (Ito et al., 2022; Landry et al., 2022; Pietrzak et al., 2012). As expected, a generalized decline in cognition was observed 30–60 min after administration of alprazolam and ethanol (corresponding to the 5-h postbaseline assessment). While this cognitive decline confirms the sensitivity of the experimental assay, it also provides a benchmark for understanding the importance of the cognitive decline observed following treatment with zuranolone alone. For example, when compared to placebo, the magnitudes of cognitive decline observed at commonly used doses of both alprazolam and ethanol were systematically greater than that associated with zuranolone, either given as a single dose or as repeated dosing. These data also suggest that the cognitive effects of single and repeat doses of zuranolone may not differ substantially and that while an acute zuranolone-related cognitive impairment does occur, this is less than that associated with other GABAergic active compounds used commonly in society.

In this study, coadministration of two CNS-depressant drugs generally increased the cognitive decline compared with single-drug administration. Neuroactive steroids interact with synaptic and extrasynaptic GABA_A receptors, whereas benzodiazepines bind primarily to synaptic receptors (Chuang and Reddy, 2018). Both zuranolone and benzodiazepine have been shown to interact with synaptic GABA_A receptors in vitro through different mechanisms (Althaus et al., 2020). Consequently, coadministration of zuranolone and alprazolam could produce physiological interactions with potential clinical manifestations. While the specific binding site on the GABA_A receptor for ethanol is less clear, both neuroactive steroids and ethanol have been shown to interact with multiple GABA_A receptor subunits in vitro (Althaus et al., 2020; Davies, 2003). Coadministration of zuranolone and ethanol could result in increased effects via simultaneous actions at multiple GABA_A receptors comprising different subunits, or they could bind to the same receptor and act through different mechanisms. The data reported herein are consistent with noncompetitive interaction at GABA_A receptors between zuranolone, a neuroactive steroid, and alprazolam or ethanol.

The estimated pharmacokinetic parameters after repeat dosing of zuranolone were consistent with previous reports (Srinivas et al., 2023) and were comparable across both study parts. Zuranolone systemic exposures (C_max and AUC_0–24) remained relatively unchanged upon coadministration with single-dose alprazolam or ethanol. Similarly, repeat dosing of zuranolone did not affect the exposure of alprazolam or ethanol in any meaningful way. The absence of a pharmacokinetic interaction is consistent with the reported metabolic characteristics for each compound. Zuranolone and alprazolam are both metabolized by CYP3A4; however, neither agent is considered an inhibitor or inducer of the isozyme (Dunbar et al., 2023; U.S. Food & Drug Administration, 2023; XANAX^® Alprazolam Tablets, USP [Prescribing Information], 2021). In addition, moderate acute ethanol intake has been reported to have no clinically relevant pharmacokinetic interactions with CYP3A substrates (Gazzaz et al., 2018). Ethanol metabolism occurs primarily via oxidative pathways, including aldehyde dehydrogenase and alcohol dehydrogenase (Zakhari, 2006). While the effect of zuranolone on these pathways has not been evaluated, the ethanol pharmacokinetic results in this study suggest the impact, if any, is likely minor.

The majority of TEAEs were mild to moderate in both parts of the study; severe TEAEs were reported only in Part A following coadministration of zuranolone and alprazolam (8.7%), and the incidence of TEAEs leading to discontinuation was 20.8%. Zuranolone, alprazolam, and ethanol are each associated with adverse events reflective of CNS depression (Althaus et al., 2020; Cutler et al., 2023; Davies, 2003); however, there was no evidence of respiratory depression or changes in respiratory ventilation with zuranolone administered alone or in combination with alprazolam or ethanol. In addition, most participants showed minimal or no signs of reduced alertness, as assessed by MOAA/S scores, following the administration of zuranolone, alone or in combination with alprazolam or ethanol. Clinically, the more pronounced cognitive decline observed following coadministration of zuranolone plus alprazolam or ethanol may suggest an increased risk of diminished decision-making ability in adults who take zuranolone plus either of these agents.

This study has some limitations. It was conducted in a relatively small population of healthy adults, which may limit its generalizability to adults with PPD. In addition, while the doses of alprazolam and ethanol used in the study were consistent with those that previously demonstrated negative cognitive effects (Ito et al., 2022; Snyder et al., 2005; Sun et al., 2015) and are considered safe to administer to healthy participants, doses of these agents may be larger in therapeutic or recreational settings. Therefore, the effects of coadministration with zuranolone may differ from those observed in this study. The absence of statistical significance for some of the main treatment comparisons may have arisen because these experimental effects were truly small, and therefore, the sample size included in the study did not provide sufficient statistical power to reject the relevant null hypotheses. However, because this study was the first to investigate the cognitive effects of a new drug therapy, conclusions about the effects were considered with respect to the effect sizes reflecting the comparisons of interest. Furthermore, any interpretations are tempered by reference to outcomes from the tests of statistical significance uncorrected for familywise error. These effect sizes can also be integrated directly into future meta-analyses of the effects of zuranolone on cognition. Finally, while efforts were made to ensure participants and clinical site staff were blinded to the treatment, physical effects of zuranolone, ethanol, or alprazolam may have compromised blinding, similar to studies with psychoactive substances in which participants were able to guess whether they received the active drug (Wilsey et al., 2016).

Of note, the design of the present study does not reflect how zuranolone is intended for clinical use, given that zuranolone was administered in the morning while the US prescribing information for zuranolone recommends it be administered in the evening (ZURZUVAE™ [Prescribing Information], 2023). The maximal effect coincided with the time of peak plasma concentrations (5–6 h postdose for zuranolone), which should therefore occur late at night or during sleep hours. By 8–12 h postdose, there was a reduction in the systemic concentrations and dissipation of cognitive decline. This suggests that the cognitive decline observed with zuranolone, alone or in combination with alprazolam or ethanol, is temporary and fully reversible. Regardless, prescribers and patients should be aware of the potential risks from the combined use of zuranolone with other CNS-depressant medications.

Conclusion

Healthy adults who received zuranolone 50 mg once daily for up to 9 days experienced small to moderate declines in cognition, which peaked at 5 h and resolved by 12 h postbaseline. Coadministration of zuranolone with alprazolam or ethanol increased the magnitude, but not the duration, of decline compared with the drugs administered alone. No clinically relevant pharmacokinetic interactions between zuranolone and alprazolam or ethanol were observed. All adverse events were consistent with the known pharmacological profiles of the agents administered, and most adverse events were mild or moderate in severity. Prescribers and patients should be aware of the potential for increased CNS-depressant effects from the combined use of zuranolone with other CNS-depressant agents, such as alprazolam or ethanol.