Abstract Introduction Sleep health, a multidimensional construct, including quality, duration, efficiency, timing, and regularity, may influence glycemic regulation during pregnancy. African American pregnant women (AAPW) experience disproportionate sleep disturbances and metabolic risk, yet these relationships remain under-explored. This study investigates between-person and within-person associations between sleep health and glycemic outcomes in non-diabetic AAPW. Methods Data from 55 non-diabetic AAPW enrolled in the BETTER lifestyle counseling study (NCT05234125) were analyzed. Sleep health was assessed using the Pittsburgh Sleep Quality Index (PSQI) for sleep quality, and 7-day Fitbit-derived measures for sleep duration, efficiency, timing (sleep midpoint), and regularity (SD of sleep midpoint). Glycemic outcomes (mean glucose, coefficient of variation CV, and mean amplitude of glycemic excursions MAGE) were measured through 7-day continuous glucose monitoring. Between-person associations were examined using multivariable regression, while within-person night-to-next-day associations were tested using mixed-effects models. All models were adjusted for maternal age, gestational age, BMI, daily steps, and education. Results Poorer sleep quality was significantly associated with greater glycemic variability. Specifically, higher PSQI scores predicted higher CV (β= 0.44, p=0.005), and MAGE (β=1.10, p=0.003). For within-person associations (n=51), higher sleep efficiency predicted higher next-day mean glucose (β=0.22, p=0.029). No other sleep health dimensions were significantly associated with glycemic outcomes in either between-person or within-person models. Conclusion Poor sleep quality may contribute to greater glycemic variability during pregnancy, while daily sleep-glucose associations appear complex. Further research with larger, more variable samples should explore mechanisms and consider sleep health as a modifiable target for improving maternal cardiometabolic outcomes. Support (if any) This study is funded by the NIH/NIMHD R01MD015724
Irsheed et al. (Fri,) studied this question.