Effect of HIV Infection on Sleep and Chronotype in an Ageing Rural South African Cohort

Introduction: Sleep disturbances are a well-established consequence of HIV infection. The pathophysiology of these disturbances have yet to be experimentally defined, but theories suggest that HIV infection may both impact the homeostatic and circadian systems of sleep-wake regulation. Studies to date have primarily explored sleep quality using the Pittsburgh Sleep Quality Index (PSQI), with limited objective data available to probe specific sleep characteristics in people living with HIV (PLWH). Those that have employed objective measures have utilised PSG, which describes aspects of sleep architecture, but does not allow for longer term monitoring of sleep habits in the home environment unlike actigraphy. Crucially, much of the research on HIV and sleep has been conducted in industrialised societies with greater access to HIV education and healthcare. Therefore, the aims of this study are to utilise actigraphy to explore sleep parameters in PLWH, and assess whether HIV infection impacts chronotype in a rural dwelling South African population.

Materials and Methods: Participants (N = 688; aged 45-100y, mean 66.4 ± 12.07y; 426 women, 166 HIV+) from the Agincourt Socio-demographic and Health Surveillance System (Mpumalanga, South Africa) were selected randomly for inclusion in this study. Participants were required to complete the Munich Chronotype Questionnaire (MCTQ), and a subset of these participants (N = 172; aged 45-93y, mean 67.06 ± 11.6y; 99 women, 31 HIV+) wore an accelerometer for a minimum of 5 nights of actigraphy (ActTrust, Condor Instruments). MCTQ data were processed in Rstudio using the ‘mctq’ package. ANOVA and subsequent multiple linear regressions were performed in RStudio to determine the relationship between HIV status and both actigraphy and MCTQ parameters, controlled for age and sex.

Results: Actigraphy analyses showed no significant relationship between HIV status and measures of sleep efficiency. However, there was a significant relationship between HIV status and total sleep time, with HIV+ individuals sleeping significantly less (F(3,168) = 2.69; P=.0482). Analysis of the MCTQ showed that the effects of HIV infection were most prominent on working days, with HIV+ individuals going to bed earlier (F(3,599) = 15.17; P≤.001) and spending more time in bed F(3,599) = 18.79; P≤.001). This effect was most pronounced in HIV+ men, and was not observed on free days. Analyses also revealed that HIV status had an interesting interaction with age on MCTQ derived chronotype (P=.002). In HIV+ individuals, chronotype was significantly later before the age of 60, but shifted earlier with age, whereas the opposite relationship was observed in HIV- individuals (F(4,264) = 3.24; P=.012).

Conclusions: Together, these data suggest that PLWH are more fatigued by work than HIV- individuals, and their earlier bedtimes may reflect an effort to combat . However, the reduced actigraphically derived total sleep time suggests that sleep needs may not be met, resulting in a cycle of sleep restriction and fatigue. Moreover, HIV may impact phase of the internal biological clock producing a shift in chronotype. Analysis of circadian phase markers will complement these data.

Acknowledgements: Supported by Academy of Medical Sciences