Sleep Outcomes and Multidimensional Assessment With Antiretroviral Therapy in People Living With HIV on Dolutegravir- vs Doravirine-Based Regimens

Background With the successful scale-up of antiretroviral therapy (ART), HIV infection has transitioned from a fatal disease to a chronic, manageable condition. This has led to substantial reductions in AIDS-related mortality globally and marked improvements in life expectancy among people living with HIV (PLWH), including in South Africa where ART coverage and viral suppression rates continue to rise. As a result, HIV care has increasingly shifted toward long-term health optimization, with growing attention to non-communicable diseases (NCDs) such as cardiovascular disease, obesity, and metabolic dysfunction.

Sleep disturbance is increasingly recognized as a potentially important but underappreciated contributor to morbidity in PLWH. Poor sleep is highly prevalent in this population and has been associated with adverse cardiometabolic, neurocognitive, and functional outcomes. Reported prevalence estimates of sleep disturbance in PLWH are as high as 50% or more, significantly exceeding that of the general population. However, most existing data rely on subjective questionnaires such as the Pittsburgh Sleep Quality Index (PSQI), Insomnia Severity Index (ISI), Epworth Sleepiness Scale (ESS), and screening tools for obstructive sleep apnoea (OSA). While these tools are useful for identifying symptoms, they are less sensitive to subtle or early physiological changes in sleep and may not adequately capture sleep as a multidimensional construct.

Contemporary sleep science conceptualises sleep health as a multidimensional construct encompassing satisfaction, alertness, timing, efficiency, duration, and regularity, rather than simply the absence of sleep disorders. Instruments such as RU-SATED have been developed to capture these domains; however, few studies have applied multidimensional sleep health frameworks in PLWH, and even fewer have evaluated how sleep health differs across antiretroviral regimens.

Objective sleep assessment using overnight polysomnography (PSG) provides a gold-standard method for characterising sleep architecture and physiology. PSG enables measurement of sleep stages (N1, N2, N3, and REM sleep), sleep continuity, respiratory events, oxygen desaturation, arousals, and other physiological parameters. Sleep architecture reflects the distribution of time spent across sleep stages, each of which plays distinct physiological roles. In healthy adults, sleep typically consists of approximately 50% N2, 20% N3 (slow-wave sleep), 25% REM sleep, and 5% N1 sleep, cycling throughout the night. N3 sleep is particularly important for physical restoration and metabolic regulation, while REM sleep is involved in emotional regulation and cognitive processing. Disruption of these stages, as well as sleep fragmentation and altered sleep continuity, has been associated with adverse health outcomes including cardiovascular disease, neurocognitive decline, and metabolic dysfunction.

Although PSG studies in PLWH are limited, existing evidence suggests consistent abnormalities in sleep architecture and continuity. Studies have reported reduced sleep efficiency, increased wake after sleep onset, altered distribution of sleep stages, reduced slow-wave sleep, and disrupted REM patterns in PLWH compared with HIV-negative controls. These abnormalities appear to persist even in virologically suppressed individuals on ART. Some evidence suggests that PLWH may exhibit increased vulnerability to sleep disruption in the presence of comorbid conditions such as obstructive sleep apnoea, with more pronounced effects on sleep architecture than in HIV-negative populations. Importantly, most PSG studies in PLWH have been conducted in predominantly male cohorts, limiting generalisability to women.

There is also significant interest in the potential role of antiretroviral agents in influencing sleep physiology. Efavirenz (EFV), a non-nucleoside reverse transcriptase inhibitor (NNRTI), is well known to cause neuropsychiatric side effects, including insomnia and alterations in REM and slow-wave sleep. In contrast, dolutegravir (DTG), an integrase strand transfer inhibitor (INSTI), is widely used due to its efficacy and tolerability but has been associated with neuropsychiatric symptoms and weight gain in some studies. While weight gain associated with DTG is recognised, the potential role of sleep disturbance as a mechanistic pathway contributing to metabolic outcomes has not been well explored. Given the bidirectional relationship between sleep and weight regulation, it is plausible that changes in sleep architecture may contribute to or exacerbate metabolic risk in individuals receiving DTG-based regimens.

Doravirine (DOR), a newer NNRTI, appears to have a more favourable neuropsychiatric and metabolic profile compared with older agents. Clinical trials have reported lower rates of sleep-related adverse events in participants receiving DOR compared with EFV, and improvements in neuropsychiatric symptoms following switches to DOR-based regimens. However, no studies to date have directly assessed objective sleep architecture using polysomnography in individuals receiving DOR or compared its effects to DTG.

A critical gap therefore exists in understanding whether specific ART regimens differentially influence sleep health and sleep architecture over time, and whether these effects are clinically meaningful. In particular, it remains unclear whether sleep disturbance in PLWH is primarily driven by HIV infection itself, by antiretroviral agents, or by an interaction between the two.

Study Rationale This study is designed to address key gaps in understanding the relationship between ART, sleep health, and metabolic outcomes in PLWH. Specifically, it aims to determine whether DTG-based and DOR-based regimens differentially affect subjective sleep health and objective sleep architecture over time.

Given the increasing global burden of obesity and cardiometabolic disease in PLWH, understanding potential mechanistic links between ART, sleep, and metabolic outcomes is essential. Emerging evidence suggests that sleep quality, rather than sleep duration alone, may be more strongly associated with body mass index and long-term metabolic health. Therefore, identifying whether ART regimens differentially affect sleep quality may have important implications for long-term treatment selection and risk reduction.

Study Design This is a prospective, randomized, comparative study evaluating sleep outcomes in adults living with HIV initiating or switching to either a dolutegravir-based ART regimen or a doravirine-based ART regimen.

Participants will be assigned to one of two treatment groups:

DTG-based regimen (TDF/3TC/DTG; standard of care) DOR-based regimen (TDF/3TC/DOR)

The study will evaluate both subjective and objective sleep outcomes over time.

Significance This study addresses a major gap in HIV research by integrating multidimensional subjective sleep assessment with gold-standard objective sleep physiology measurements in a randomized comparative ART context in South Africa. It will provide novel evidence on whether DTG- and DOR-based regimens differentially affect sleep health, and whether these effects are clinically meaningful and associated with metabolic risk.

Findings will inform the design of a future fully powered clinical trial, including refinement of endpoints and sample size calculations, and may contribute to more individualized ART selection strategies that consider long-term sleep and metabolic health in PLWH.

• Between-Arm Difference in RU-SATED Total Sleep Health Score at Week 12 [Time frame: Week 12]
• Change from Baseline to Week 12 in Polysomnography-Derived Sleep Stage Distribution (N1, N2, N3, REM) [Time frame: Baseline and Week 12]
• Change from Baseline to Week 12 in Polysomnography-Derived Sleep Latency [Time frame: Baseline and Week 12]
• Change from Baseline to Week 12 in Polysomnography-Derived REM Latency [Time frame: Baseline and Week 12]
• Change from Baseline to Week 12 in Polysomnography-Derived Total Sleep Time [Time frame: Baseline and Week 12]
• Change from Baseline to Week 12 in Polysomnography-Derived Sleep Efficiency [Time frame: Baseline and Week 12]
• Change from Baseline to Week 12 in Polysomnography-Derived Wake After Sleep Onset [Time frame: Baseline and Week 12]
• Change from Baseline to Week 12 in Polysomnography-Derived Apnea-Hypopnea Index [Time frame: Baseline and Week 12]
• Change from Baseline to Week 12 in Polysomnography-Derived Oxygen Desaturation Events [Time frame: Baseline and Week 12]
• Change from Baseline to Week 12 in PROMIS Sleep Disturbance Short Form 8a T-score [Time frame: Baseline and Week 12]
• Change from Baseline to Week 12 in PROMIS Sleep Related Impairment Short Form 8a T-score [Time frame: Baseline and Week 12]
• Participant Recruitment Rate [Time frame: From initiation of recruitment until enrollment of the final participant, assessed over approximately 12 months]