This proof-of-concept study demonstrated that a combination of dronabinol and acetazolamide taken nightly for 1 week reduces the severity of OSA without significant safety concerns. In this small sample, the improvement in OSA severity was not accompanied by changes in sleep architecture or OSA symptoms, mood, or quality of life, and the 35%–40% reduction in AHI was similar across low, medium, and high doses.

There is evidence that both dronabinol and acetazolamide, when taken as monotherapies, reduce the severity of OSA. In a randomized controlled trial, dronabinol reduced the AHI by approximately 7% and 16% at 2.5 and 10 mg doses, respectively, in comparison to a 36% increase in AHI in the placebo arm [15]. Acetazolamide 250 mg monotherapy has been demonstrated to reduce the AHI by 26% [16], which is less than the 39% reduction observed in our study when taken in combination with 5 mg dronabinol. No studies have investigated the effect of acetazolamide doses less than 250 mg on OSA severity. However, given the recognized dose response with acetazolamide monotherapy up to 500 mg [5], where there is an associated 38% reduction in AHI, it is hypothesized that 125 mg acetazolamide in the low dose in our study would not fully account for the 40% reduction in AHI. Therefore, it is likely that combination therapy has an additive or synergistic benefit that could be elucidated in further trials investigating both therapies alone and in combination. Future trials could also examine the combination of sulthiame, another carbonic anhydrase inhibitor, and dronabinol. Four weeks of 200 and 400 mg sulthiame have recently been shown to reduce the AHI by 33% and 40%, respectively [17], again supporting the notion that the response to the low dose of IHL-42X is unlikely to be due exclusively to one of the constituents.

The 40% reduction in OSA severity at the low dose in our study is one of the largest effect sizes reported for pharmacological treatment of OSA [4]. It is particularly encouraging given the low rate of adverse events reported at this dose (n = 4 vs. n = 5 when taking placebo) and their mild nature. However, it is intriguing that we did not observe additional clinical benefits at the medium and higher doses. Understanding the contribution of the pathophysiological mechanisms of OSA at each dose and with the monotherapies may explain the absence of a dose response. Notably, when taken with atomoxetine, increasing dronabinol from 5 to 10 mg also provided no additional benefit, although the combination reduced OSA severity whereas atomoxetine alone did not [18]. We also identified a reduction in ODI3% with the moderate dose and reductions in the total and respiratory arousal indices at all doses of IHL-42X. The reduction in arousal indices may reflect an increase in the arousal threshold, although we might expect to see a concomitant increase in event duration if this were the case. Alternatively, the reduction in arousal indices may be secondary to a reduction in the number of respiratory events. Arousal threshold does not appear to be modified by acetazolamide or sulthiame [19, 20], although the effect of dronabinol remains to be investigated. Although it is outside the scope of this preliminary trial, understanding the effect of IHL-42X on the pathophysiological mechanisms (i.e. endotypes) of OSA will be valuable for identifying those who are most likely to respond to it.

Despite improvement in OSA severity with IHL-42X treatment, we did not identify any improvement in symptoms, mood, or quality of life. This is perhaps not unexpected given the relatively asymptomatic cohort which we studied and the short duration of the treatment period. The mean baseline level of sleepiness was normal; therefore, due to a floor effect, we would not expect to observe a large improvement in symptoms. Furthermore, a single week of treatment may be inadequate for achieving symptom resolution or reduction. Perhaps also relevant is that although the 40% reduction in OSA severity across the whole night is encouraging, OSA was not abolished. It could be argued that CPAP, which has an efficacy close to 100% (for duration of use), would induce a larger reduction in OSA symptoms. However, it is becoming increasingly evident that despite the high efficacy of CPAP (for the duration of use), its effectiveness is limited by poor tolerance to continuous use, with the average nightly CPAP wear time being 4.6 hours [21].

The combination of dronabinol and acetazolamide was well tolerated with no serious adverse events and no participant withdrawals due to intolerance. Of the 35 adverse events reported across the placebo and three treatment arms, 74% were while taking the medium and high doses. Occurring in four of the nine participants, paraesthesia of the fingers or hands was the most frequently reported side effect of IHL-42X. This is remarkably consistent with a meta-analysis which determined that 2.3 participants needed to take acetazolamide for one person to experience paraesthesia [22]. However, it is somewhat lower than the incidence of paraesthesia associated with sulthiame treatment, which was 67% in those taking 200 mg and 79% in those taking 400 mg [17]. Gastrointestinal side effects have also been commonly reported in previous trials of acetazolamide and dronabinol when taken as monotherapies. However, when taking dronabinol in clinical trials, only dry mouth, dizziness, and headache have been reported to occur at a greater frequency than when taking placebo [9]. Importantly, in our study, side effects resolved within a day or soon after ceasing treatment.

The higher plasma THC levels in the evening, approximately 2 hours post-first dosing, than in the morning, approximately 10 hours post-final dosing, aligns with previous reports for a single dose of a product containing 10 mg of THC [23]. However, the average morning THC level (0.45 ± 0.81 ng/mL) across all doses in the current study reflects plasma levels after taking IHL-42X for 1 week, where accumulation, if it were to occur, would be expected. Plasma THC levels have a very weak relationship with functional impairment but are important in some jurisdictions for driving and workplace screening, where 1 ng/mL is considered one of the most conservative thresholds [24]. Although this suggests the likelihood of minimal daytime impairment and risk of screening positive for THC in the morning after taking IHL-42X the prior evening, it is notable that one participant, when taking the medium dose in the current study, had a morning THC level exceeding 1 ng/mL, suggesting individual variability that warrants investigation in future studies.

Sleep architecture was not significantly altered when taking IHL-42X compared to placebo. Increases in total sleep time, sleep efficiency, and proportion of NREM 3 sleep have been reported with acetazolamide use, although whether this is secondary to improvements in OSA is unknown [22]. Of the two previous studies that have investigated the effect of dronabinol on OSA [6, 15], the only change in sleep architecture identified was a small reduction in the amount of REM sleep (and inverse increase in NREM sleep) in one study for participants taking 2.5 mg dronabinol relative to placebo [15]. Despite the absence of significant improvements in sleep architecture when taking IHL-42X compared to placebo in our study, it is important to note that IHL-42X was not detrimental for sleep and all sleep architecture parameters changed in a favorable direction, which might be detectable with a larger sample size.

We used a randomized crossover study design for this proof-of-concept trial and the encouraging results suggest that IHL-42X may be an effective alternative therapy for some people. However, we report data from nine participants who used the therapy for 1 week; OSA is typically a life-long disorder, so larger and longer duration trials will be crucial for establishing IHL-42X as an efficacious and safe treatment option. Larger trials with multinight measurements will also serve to minimize concern that night-to-night variability [25] influenced outcomes; it is notable that one participant demonstrated significant unexplained variability between baseline and placebo nights, although it is encouraging that six of the nine participants had a reduction in AHI of at least 50% with one or more of the doses of IHL-42X relative to placebo. It is also worth noting that the participants in our trial were otherwise healthy and taking few other medications. Dronabinol and acetazolamide are contraindicated in some populations and have known interactions with other medications, some of which are associated with significant negative health consequences. Both medications, when taken as monotherapies, are also known to have distinct side effects which may have been perceived by participants and may have contributed to unblinding. However, although blinding will be important for minimizing bias on subjective measures of symptoms and function, the need for maintaining blinding when establishing OSA efficacy based on polysomnographic parameters is debatable [26]. Finally, it is possible that the improvements in OSA may be due to acetazolamide or dronabinol alone. Further investigation of the combined and individual effects of acetazolamide and dronabinol will elucidate their individual, combined, or synergistic effects on OSA and the OSA endotypes, which may allow for targeted therapy.

In conclusion, many people do not tolerate or fail to respond to the first-line therapies for OSA and many patients would like a pharmacological solution. This study provides proof-of-concept data, from a small sample, that a combination of acetazolamide and dronabinol, taken over a week, reduced the AHI and is well tolerated in participants with OSA. The reduction in AHI was accompanied by improvements in arousal indices and oxygen desaturation but was independent of changes in sleep architecture and subjective symptom and quality of life measures. These results are encouraging, although further studies with larger sample sizes and longer treatment durations are required to confirm long-term efficacy, safety, and tolerability and to determine mechanisms of action and rationale for the absence of a dose response.