Associations between fears related to safety during sleep and self-reported sleep in men and women living in a low-socioeconomic status setting.

South Africans living in low socioeconomic areas have self-reported unusually long sleep durations (approximately 9-10 h). One hypothesis is that these long durations may be a compensatory response to poor sleep quality as a result of stressful environments. This study aimed to investigate whether fear of not being safe during sleep is associated with markers of sleep quality or duration in men and women. South Africans (n = 411, 25-50 y, 57% women) of African-origin living in an urban township, characterised by high crime and poverty rates, participated in this study. Participants are part of a larger longitudinal cohort study: Modelling the Epidemiologic Transition Study (METS)-Microbiome. Customised questions were used to assess the presence or absence of fears related to feeling safe during sleep, and the Epworth Sleepiness Scale, Pittsburgh Sleep Quality Index (PSQI) and Insomnia Severity Index were used to assess daytime sleepiness, sleep quality and insomnia symptom severity respectively. Adjusted logistic regression models indicated that participants who reported fears related to safety during sleep were more likely to report poor sleep quality (PSQI > 5) compared to participants not reporting such fears and that this relationship was stronger among men than women. This is one of the first studies outside American or European populations to suggest that poor quality sleep is associated with fear of personal safety in low-SES South African adults.

Measurement of body composition in postpartum South African women living with and without HIV infection.

Background: While several methodologies are available to measure adiposity, few have been validated in sub-Saharan African (SSA) and none in postpartum African women living with HIV (WLHIV). We compared bioelectrical impendence analysis (BIA) and air displacement plethysmography (ADP) against dual x-ray absorptiometry (DXA) in South African women and examined differences by HIV and body mass index (BMI) status. Methods: Lin's concordance correlation coefficient (CCC) test was used to examine fat mass (FM), fat free mass (FFM), and total body fat percent (%BF) difference between BIA vs. DXA, and ADP vs. DXA in women living with HIV (n = 57) and without HIV (n = 25). The Bland Altman test was used to assess mean differences and the direction of bias. Results: The median age was 31 years (IQR, 26-35) and months postpartum were 11 (IQR, 7-16), 44% of the women had obesity. Lin's CCC for BIA and ADP vs. DXA were both 0.80 for %BF and 0.97 for FM, and 0.86 and 0.80 for FFM, respectively. Mean differences (DXA-BIA and ADP estimates) were 0.22 ± 4.54% (p = 0.54) and 3.35 ± 3.27% (p < 0.01) for %BF, -0.82 ± 3.56 kg (p = 0.06) and 1.43 ± 2.68 kg (p = 0.01) for FM, -1.38 ± 3.61 kg (p = 0.01) and - 3.34 ± 2.37 kg (p < 0.01) for FFM, respectively. BIA overestimated %BF in WLHIV and underestimated it in women with obesity. Conclusion: Body composition measurements using BIA and ADP correlated well with DXA, thereby providing alternative, safe tools for measuring postpartum FM and FFM in SSA women, including WLHIV.