Effects of solriamfetol on on‐the‐road driving performance in participants with excessive daytime sleepiness associated with obstructive sleep apnoea

Abstract Objective To evaluate the impact of solriamfetol, a dopamine and norepinephrine reuptake inhibitor, on on‐the‐road driving in participants with excessive daytime sleepiness (EDS) associated with obstructive sleep apnoea (OSA). Methods Eligible participants were aged 21–75 years with OSA and EDS (Maintenance of Wakefulness Test mean sleep latency <30 minutes and Epworth Sleepiness Scale score ≥10). Participants were randomised 1:1 to solriamfetol (150 mg/day [3 days], then 300 mg/day [4 days]) or placebo for 7 days, before crossover to the other treatment paradigm. On Day 7 of each period, standardised on‐road driving tests occurred (2 and 6 hours postdose). Standard deviation of lateral position (SDLP) was the primary endpoint. Results Solriamfetol significantly reduced SDLP at 2 (n = 34; least squares mean difference, –1.1 cm; 95% CI, –1.85, –0.32; p = 0.006) and 6 hours postdose (n = 32; least squares mean difference, –0.8 cm; 95% CI, –1.58, –0.03; p = 0.043). Two hours postdose, 4 placebo‐treated and 1 solriamfetol‐treated participants had incomplete driving tests; 6 hours postdose, 7 and 3 participants, respectively, had incomplete tests. Common treatment‐emergent adverse events included headache, nausea, and insomnia. Conclusions Solriamfetol 300 mg/day significantly improved on‐the‐road driving performance in participants with EDS associated with OSA.

Conclusions: Solriamfetol 300 mg/day significantly improved on-the-road driving performance in participants with EDS associated with OSA.

K E Y W O R D S
excessive daytime sleepiness, obstructive sleep apnoea, on-the-road driving, solriamfetol, Sunosi

| INTRODUCTION
Obstructive sleep apnoea (OSA) is a sleep-related breathing disorder that is estimated to affect nearly 1 billion adults worldwide (Benjafield et al., 2019). Excessive daytime sleepiness (EDS) is a common symptom of OSA (Dongol & Williams, 2016;Pagel, 2009) and can severely impact patients' lives, causing impairments in mood, quality of life (QoL), cognitive function, work productivity, and safety (Garbarino et al., 2016;Gasa et al., 2013;Mulgrew et al., 2007;Pepin et al., 2009;Stepnowsky et al., 2019;Zhou et al., 2016). In addition, patients with OSA and EDS havee2.5 times the risk of motor vehicle accidents compared with healthy controls (Tregear et al., 2009), and shorter sleep latency as measured with the Maintenance of Wakefulness Test (MWT) is significantly correlated with sleepiness-related motor vehicle accidents and near misses in patients with sleep disorders (Philip et al., 2021).
Primary OSA therapy, such as continuous positive airway pressure (CPAP), can reduce symptoms of EDS; nevertheless, persistence of EDS has been reported in 9% to 22% of patients, despite their use of CPAP (Gasa et al., 2013;Pepin et al., 2009). Pharmacologic treatment can complement primary airway therapy for the alleviation of residual EDS associated with OSA (Marra et al., 2019). In laboratory studies, the wake-promoting agents (WPAs) modafinil and armodafinil (approved in the United States, but not the European  (Chapman et al., 2014;Kay & Feldman, 2013;Williams et al., 2010). In a retrospective cohort study, use of methylphenidate or modafinil was associated with a 20% reduction in the risk of hospitalisation attributable to a motor vehicle accident in patients with OSA (Lin et al., 2020). However, studies demonstrating that a pharmacologic treatment can result in specific improvements in on-the-road driving performance in sleepy patients with OSA are lacking. Solriamfetol (SUNOSI™) is a dopamine and norepinephrine reuptake inhibitor approved in the United States and European Union to improve wakefulness in adult patients with EDS associated with OSA (approved dose range, 37.5-150 mg/day) (Sunosi™ (solriamfetol) tablets Prescribing Information, 2021; Sunosi™ (solriamfetol) tablets Summary of Product Characteristics, 2020).
Solriamfetol was investigated in participants with EDS associated with OSA in short (12 weeks) and longer-term (up to 52 weeks) clinical trials, where treatment with solriamfetol at doses ranging from 37.5 to 300 mg/day was associated with reduced EDS and improvements on measures of daily functioning, work productivity, and QoL (Malhotra et al., 2020;Schweitzer et al., 2019;Weaver et al., 2020;Weaver et al., 2019).
In parallel with the 12-week phase 3 trial, the current study was conducted to evaluate the effects of solriamfetol on on-the-road driving performance in participants with EDS associated with OSA. basis, and ability to operate a vehicle with a manual transmission. As part of the inclusion criteria, participants were also required to take a practice driving test at screening, and to complete it without any safety concerns.

| METHODS
Key study exclusion criteria included an unwillingness to try to use a primary OSA therapy, occupational nighttime shift work, usual bedtime after 1:00 A.M., a clinically relevant medical or psychiatric disorder (other than OSA) associated with EDS, a history or presence of an unstable medical or psychiatric condition, or pregnancy. Additional exclusion criteria were excessive caffeine use (>8 cups of coffee/day), smoking >10 cigarettes/day, use of medication that could affect sleep-wake functions within 7 days before screening, use of a monoamine oxidase inhibitor within 14 days or 5 half-lives before screening, use of an investigational drug within 30 days or 5 half-lives before baseline, anticipated use of any of these substances during the study, or previous use of solriamfetol.

| Design
A randomised, double-blind, placebo-controlled, 2-period crossover study design was used. Eligible participants were randomly assigned 1:1 to receive either solriamfetol (150 mg/day for 3 days, followed by 300 mg/day for 4 days) or placebo for 7 days (Period 1) and then cross over to the other treatment for 7 days (Period 2); there was no washout between periods. Solriamfetol 150-and 300-mg tablets and placebo tablets were supplied in identical opaque gelatin capsules to ensure adequate blinding. This study was initiated before regulatory approval or dosing recommendations were finalised. Therefore, the 300-mg/day dose used was based on prior phase 2 study data (Bogan et al., 2015;Ruoff et al., 2016), consistent with the maximum dose used in pivotal trials of solriamfetol for patients with OSA (Malhotra et al., 2020;Schweitzer et al., 2019).

| Procedures
The study included a screening/washout period of ≤28 days prior to the first dose of study treatment: Eligibility was assessed (including general safety assessments; in addition, a 40-minute MWT and ESS were assessed at visit 2), prohibited medications were washed out, and participants completed a practice driving test (at baseline/visit 3). On Day 7 and 14 (ie, Day 7 of each period), visits were conducted to evaluate driving performance. A safety follow-up visit was conducted approximately 1 week after completion of Period 2 ( Figure 1a).
Participants were instructed to take a single capsule once daily, within 1 hour of waking in the morning, on an empty stomach, and then to wait ≥30 minutes before having breakfast. On driving test days, the capsule for that day was administered at the driving test site in the presence of an investigator at 8:45 A.M. (2 hours before the start of the first drive); 30 minutes after administration, participants received a light breakfast. Throughout the study, caffeine users were instructed to not increase their use during the study, and nicotine users were instructed to maintain a consistent level of use. In addition, on driving test days, 1 cup of black coffee was permitted prior to arrival at the test site, with no additional consumption until after the second driving test; nicotine use was restricted to 1 cigarette in the morning ≥1 hour before the first MWT trial and 1 cigarette on waking on driving test days, with no other use until after the study procedures were completed on those days.
At the end of each treatment period, a standardised on-road driving test (Verster & Roth, 2011) was conducted at 2 hours and 6 hours after administration of study treatment (Figure 1b). For each test ( e1 hour in duration), participants drove a specially instrumented vehicle over a 100 km ( e62 miles) primary highway circuit; they were accompanied by a licensed driving instructor with access to dual controls (brakes, clutch, accelerator). Participants were instructed to maintain both a steady lateral position between the delineated boundaries of the slower (right) traffic lane and a constant speed of 95 km/h ( e59 mph). Participants were permitted to deviate from these instructions only to pass a slower vehicle, to respond to slower traffic ahead, or to exit and reenter the highway at the turnaround point (these events were later removed for the purposes of the analysis of driving parameters by 2 experienced raters). Vehicle speed and lateral distance to the left-lane line were continuously recorded, and the data stored on an onboard computer. The driving test could be stopped by the participant or by the accompanying driving instructor if either considered it unsafe to continue.

| Assessments and outcomes
The primary outcome assessment from the driving tests was standard deviation of lateral position (SDLP) in centimetres-a measure of "weaving" or road-tracking control (Ramaekers, 2017;Verster & Roth, 2011). Data were analysed for all driving tests (completed or incomplete) with data available; for incomplete driving tests, SDLP data from the part of the test that was completed were analysed.
Standard deviation of speed and number of lane drifts (defined as deviations >100 cm from the absolute lateral position within an 8-second window) were also determined from driving test data.
The Toronto Hospital Alertness Test (THAT) is a 10-item selfreport questionnaire that measures perceived alertness over the previous week; scores can range from 0 to 50, with higher scores indicating greater alertness (Shapiro et al., 2006). This assessment was administered at baseline and on driving test days, prior to administration of study treatment.
Safety assessments included a physical examination, electrocardiogram, clinical laboratory tests, and assessment of adverse events (AEs).
Participants using a primary OSA therapy (PAP or oral appliance) at screening recorded their primary OSA therapy usage and the estimated duration of use (more than half of the night, less than half of the night, or don't know) on a daily basis. VINCKENBOSCH ET AL.

| Statistical analyses
The primary efficacy endpoint was SDLP at 2 hours postdose, and the secondary efficacy endpoints included SDLP at 6 hours postdose, percentage of participants with improved or impaired driving on solriamfetol compared with placebo, standard deviation of speed, lane drifts, and THAT score.
For the primary endpoint, the null hypothesis was that at 2 hours postdose the mean SDLP values for solriamfetol and placebo were equal; the alternative hypothesis was that they were not equal. The treatment difference in mean SDLP between solriamfetol and placebo at 2 hours postdose was tested; a 5% type I error rate (p < 0.05) was considered statistically significant. A sample size of 36 participants would provide 90% power to detect a mean difference of 2.0 cm on the primary outcome measure, SDLP (Ramaekers et al., 2006;Verster et al., 2008), assuming a standard deviation (of SDLP) of 3.25 cm and a 2-sided 0.05 significance level using paired t test. A study enrolment of 40 participants was planned in order to allow for  Figure 1b shows the instrumented vehicle during the driving test in actual traffic on a primary highway. The lateral position of the car relative to the white middle line is continuously measured during a 1-h drive by means of a camera that is mounted on the roof of the car (right upper panel). The mean SDLP over the entire ride is calculated offline after completion of the test using signal editing software. SDLP is a measure of weaving and indicates road-tracking control of the driver. Drugs that induce sleepiness and sedation cause significant increments in weaving motion and thus loss of vehicular control. 2.5, and 3.5 cm were tested (Ramaekers et al., 2006;Verster et al., 2008). In comparisons of solriamfetol and placebo, improvement was defined as a decrease in SDLP in participants treated with solriamfetol compared with placebo at the threshold, and impairment was defined as an increase in SDLP at the threshold, or failure to complete the driving test while on solriamfetol because of sleepiness or safety concerns regardless of their performance while on placebo (participants who failed to complete the driving test while on placebo but completed the test while on solriamfetol were not counted as impaired or improved).
The number of participants who failed to complete the driving test was summarised descriptively, as was the duration of the drive before stopping. Additional secondary efficacy measures (standard deviation of speed, number of lane drifts) were analysed with an ANOVA method similar to that used for SDLP. THAT scores were analysed using a mixed effect analysis of covariance (ANCOVA) model. No multiplicity adjustments were made in the efficacy analyses for multiple endpoints, and all p values are therefore nominal.
Demographic, OSA history, and safety data were summarised descriptively for the safety population, which included all participants who received ≥1 dose of study drug. No formal statistical testing was performed.

| RESULTS
Of 59 participants who were screened, 34 met the study inclusion criteria and were enrolled ( Figure 2). All participants received ≥1 dose of study treatment and comprised the safety population; 1 participant was withdrawn after study Period 1 and did not receive the study treatment (placebo) for Period 2. All enrolled participants were white, of non-Hispanic/Latino ethnicity, and located in the Netherlands (Table 1). Participants had a mean (standard deviation [SD]) ESS score of 14.4 (3.5) and a mean (SD) MWT sleep latency of 14.3 (7.3) minutes. Actigraphy and sleep diary data showed no differences in total sleep time between placebo and solriamfetol treatment (data not shown). Twenty-nine participants were using primary OSA therapy; the remaining 5 participants had attempted CPAP use but ultimately discontinued.
Use of primary OSA therapy was stable throughout the study.
The mean percentage of nights that participants used primary OSA therapy for more than half the night was 95.7% at baseline, 94.6% at the end of the placebo treatment period, and 92.7% at the end of the solriamfetol treatment period (n = 28 at each time point).
On the primary outcome measure, SDLP at 2 hours postdose, there was a statistically significant reduction with solriamfetol compared with placebo (least squares [LS] mean difference, -1.1 cm; p = 0.006; Table 2). The full set of ANOVA results is presented in Supplementary Table S1. An improvement with solriamfetol versus placebo was also observed at 6 hours postdose (LS mean difference, -0.8 cm; p = 0.043).
Individual driving performance with solriamfetol versus placebo is shown in Figure 3. Spaghetti plots of individual participant data for solriamfetol and placebo are shown in Supplementary Figure S1 for each time point.
Eight participants stopped ≥1 driving test prematurely. More participants had incomplete tests when receiving placebo compared with solriamfetol at 2 hours postdose and 6 hours postdose (Table 3) Table S2).
Treatment-emergent adverse events (TEAEs) were reported in approximately two-thirds of participants overall. The majority of TEAEs were mild to moderate in severity, and none led to study drug interruption or withdrawal. There were no serious TEAEs or deaths.

| DISCUSSION
This double-blind crossover study evaluated the effect of solriamfetol treatment on driving performance in participants with EDS associated with OSA. Participants received 7 days of treatment and undertook an on-road driving performance test 2 and 6 hours after dosing. Solriamfetol (150 mg/day for 3 days followed by 300 mg/day for 4 days) significantly improved SDLP, an important measure of driving performance, at both time points compared with placebo.
Fewer participants completed the driving test on placebo than solriamfetol at both time points. Additionally, a numerically greater percentage of participants had improved SDLP than impaired SDLP with solriamfetol compared with placebo at 2 hours postdose.
Fifteen (11.5%) of 131 tests were stopped because the instructor or participant considered it unsafe to continue. This happened more frequently than in comparable studies assessing sedating drugs in healthy volunteers (3.1%) (Verster & Roth, 2012). Most incomplete tests in this study were stopped under placebo treatment (n = 11/15, 73.3%) and at the participant's request (n = 11/15, 73.3%; Table 3). In contrast, during the aforementioned studies, 3 to 4 times more tests were stopped by the instructor than the participant (Verster & Roth, 2012). This suggests that participants in our study were often aware of their potential impairment and careful to avoid further risks. The on-road driving test is the gold standard for assessing druginduced changes in driving (Jongen et al., 2017). However, studies with other WPAs in participants with OSA have examined only simulated driving (Chapman et al., 2014;Kay & Feldman, 2013;Williams et al., 2010). One such study in participants with OSA before CPAP initiation (Kay & Feldman, 2013)  Solriamfetol's safety profile in this study was consistent with the larger 12-week and 52-week studies (Malhotra et al., 2020;Schweitzer et al., 2019). Most TEAEs were mild or moderate in severity. Headache, nausea, insomnia, and dizziness occurred more frequently with solriamfetol than placebo. No TEAEs were serious or led to treatment/study discontinuation.
Limitations include the fact that the tested dose of solriamfetol (300 mg/day) exceeds the highest recommended dose (150 mg/day).
As previously noted, this study was conducted before regulatory approvals of solriamfetol in OSA, and dosing was based on previous and ongoing studies at the time this study was designed (Bogan et al., 2015;Ruoff et al., 2016;Schweitzer et al., 2019). Thus, it is unknown how the magnitude of functional improvements observed at 300 mg/day would translate to the highest approved dose (150 mg/day) in clinical practice. However, efficacy of the 150-mg and 300-mg doses in the overall OSA population in the phase 3 study was similar (Schweitzer et al., 2019). While SDLP is linked to accident risk (Ramaekers, 2017), how the functional improvements observed here might affect accident risk is unknown, as the study was not designed to directly assess this. Additionally, long-term effects on driving performance were not assessed. An open-label extension study indicated that solriamfetol's wake-promoting effects are maintained for up to 1 year (Malhotra et al., 2020); it is reasonable to expect improved driving performance would also be maintained. Finally, the study was homogeneous in terms of race and ethnicity, and the majority of participants were male, which may limit the generalisability of these findings to other groups.
Strengths of the study include the fact that participants' baseline characteristics reflected real-world OSA populations (eg, primarily male; mean age,e51 years; mean BMI,e29 kg/m 2 ) (Bailly et al., 2016;Tkacova et al., 2014). Additionally, the driving test was conducted in on-the-road traffic. The crossover design eliminated between-group differences in participant characteristics (eg, BMI, apnoeahypopnea index, hypoxemia) that predict motor vehicle accidents in drivers with OSA (Tregear et al., 2009).

| CONCLUSION
For participants with EDS associated with OSA, solriamfetol treatment was associated with significant improvement in on-the-road driving performance, as assessed by SDLP at 2 hours and 6 hours postdose; additional secondary outcome measures (THAT scores) also indicated greater alertness compared with placebo. These findings demonstrate that solriamfetol's wake-promoting efficacy observed across multiple clinical trials (Malhotra et al., 2020;Schweitzer et al., 2019) is also associated with improved real-world functional performance in this study.