Actigraphy in the Assessment of Sleep Patterns in Adult HIV-AIDS Patients in the Yaoundé Central Hospital: A Pilot Study

Nfor Leonard Njamnshi (
Internal Medicine, University of Yaounde 1
June, 2013
- 3 Months Online Training on Public Health: Quantitative Methods in Clinical & Public Health Research by Harvard School of Public Health. University of Harvard, USA.

- Training on the Integrated Management of Childhood Illnesses organized by the Ministry of Public Health (PCIME)

- Technology in Service of the Environment, TUNZA International Youth Conference Organized by UNEP and Bayer AG

Computer skills:
- Statistical analysis (e.g. EXCEL, SPSS, Epi-Info, Stata)
- Document and magazine design (e.g. Microsoft Word, Microsoft Publisher)
- Poster and slide preparation (Microsoft PowerPoint)


HIV-AIDS remains a serious public health challenge in sub-Saharan Africa(SSA) where almost 68% of the 33.4 million adults worldwide infected with HIV live. As new therapies extend life expectancy of HIV patients, its long term adverse impact on quality of life are common and need to be addressed. There has been increasing evidence of sleep problems in people living with HIV-AIDS (PLWHA). Little is known of the extent and pattern of these problems in SSA particularly using actigraphy a tool validated for sleep studies which is cheaper, less cumbersome, and more accessible for resource limited settings than polysomnography, the gold standard.
In this preliminary study we therefore had as objective to characterize sleep patterns and problems experienced by adults with HIV compared to age and sex matched HIV-negative controls using actigraphy. Specifically we sought first to determine the feasibility of conducting sleep pattern studies using actigraphy secondly to determine the prevalence and nature of actigraphy-defined sleep disorders in the sample and thirdly compare the sleep pattern in the both groups to determine any circadian rhythm changes.

To attain this objective, we carried out a cross sectional case control study of PLWHA and age and sex matched controls living within Yaoundé. Study participants were clerked and examined. Excluded from the study were individuals who did not give their informed consent. Sleep and activity were estimated with a noninvasive battery-operated wrist actigraph that detects movement and acceleration. The actigraph was worn continuously on the non-dominant wrist for 3 consecutive days during which they kept a sleep diary alongside. The Sadeh algorithm was used to calculate total sleep time(TST), number of awakenings, sleep ratio and sleep onset latency(SOL) using an automatic sleep scoring program Action4® and Matlab®. Cosinor rhythmometry was used to get parameters of Mesor, Amplitude, Acrophase time, Interdaily stability and Intradaily variability. The mean values for the complete days were entered into Microsoft Excel 2010 and analysed using SPSS version 20.0. Cut off of statistical significance was set at 0.05.

Demographic, clinical and actigraphy data for 16 PLWHA and 16 age and sex matched controls was successfully collected. The mean age in case group was 33.81 years (±7.33) and 31.75 years (±9.04) for controls (p=0.484). The mean total sleep time was 7.23±2.85hours for the HIV group and 9.7±2 hours for controls. HIV patients took more time to fall asleep (49.57±25.95mins) than controls (27.83±16.3mins) with p=0.003. Similarly patients in whom their regiment included Efavirenz took more time to fall asleep than their counterparts (p=0.0353). Cases had more numerous night awakenings (20.06 ± 11.072 vs 12.13±6.407, p=0.019) thus a more fragmented night sleep. The two main types of sleep disorders in the sample were insomnia and sleep fragmentation. With regards to Insomnia 62.5% of cases had sleep-onset insomnia and 68.8% had reduced sleep quantity against 25.0% and 12.5% in controls respectively (p=0.033,OR=5 and p=0.001 OR=4.8 respectively). Sleep fragmentation was present in 43.75% of the cases and 6.3% of controls (p= 0.014, odds ratio of 11.7). Cases had a less structured day-to-day rhythm revealed by a higher intradaily variability(0.235±0.1 vs 0.188 ±0.037, p=0.032) and a lower interdaily stability (0.962 ± 0.014 vs 0.971±0.006, p=0.031). Apart from the Acrophase time which was 1hour 10 mins in advance for cases (14:26 vs 13:14, p=0.013) other cosinor parameters were similar for both groups. Generally patients had a phase shift in their circadian rhythm which was in advance of about 1hour before normal controls and there was no particular associations with HIV disease progression.
Our results show a significant difference in sleep and circadian rhythm parameters in HIV-positive individuals compared to HIV-negative controls. Insomnia (more time to fall asleep and reduced sleep quantity) and sleep fragmentation were more common in cases. This thus adds to the existing evidence of sleep disorders in HIV-AIDS adults in Sub Saharan Africa.
We therefore recommend that further studies be done to investigate the degree and scope of the circadian rhythm disruption as well as examine its impact on quality of life and survival in correlation with biomarkers of disease. We also recommend that Health personnel should consider screening and providing appropriate management for sleep disorders in PLWHA.