Sleep and Physiological Dysregulation: A Closer Look at Sleep Intraindividual Variability

Abstract

Study objectives: Variable daily sleep (ie, higher intraindividual variability; IIV) is associated with negative health consequences, but potential physiological mechanisms are poorly understood. This study examined how the IIV of sleep timing, duration, and quality is associated with physiological dysregulation, with diurnal cortisol trajectories as a proximal outcome and allostatic load (AL) as a multisystem distal outcome.

Methods: Participants are 436 adults (Mage ± standard deviation = 54.1 ± 11.7, 60.3% women) from the Midlife in the United States study. Sleep was objectively assessed using 7-day actigraphy. Diurnal cortisol was measured via saliva samples (four/day for 4 consecutive days). AL was measured using 23 biomarkers from seven systems (inflammatory, hypothalamic-pituitary-adrenal axis, metabolic glucose and lipid, cardiovascular, parasympathetic, sympathetic) using a validated bifactor model. Linear and quadratic effects of sleep IIV were estimated using a validated Bayesian model.

Results: Controlling for covariates, more variable sleep timing (p = .04 for risetime, p = .097 for bedtime) and total sleep time (TST; p = .02), but not mean sleep variables, were associated with flatter cortisol diurnal slope. More variable sleep onset latency and wake after sleep onset, later average bedtime, and shorter TST were associated with higher AL adjusting for age and sex (p-values < .05); after controlling for all covariates, however, only later mean bedtime remained significantly associated with higher AL (p = .04).

Conclusions: In a community sample of adults, more variable sleep patterns were associated with blunted diurnal cortisol trajectories but not with higher multisystem physiological dysregulation. The associations between sleep IIV and overall health are likely complex, including multiple biopsychosocial determinants and require further investigation.

Keywords: allostatic load; cortisol; health; intraindividual variability; physiological dysregulation; sleep.

Figure 1

Sample distributions of intraindividual variability (IIV) for bedtime (BT), risetime (RT), time-in-bed (TIB), sleep onset latency (SOL), wake after sleep onset (WASO), and sleep efficiency (SE). The sum of areas under the curve is 1 (100%); x-axis shows IIV (ie, model estimated individual standard deviation). Shaded regions indicate 95% credible intervals around the estimate.

Figure 2

Cortisol trajectories for individuals with high (+1 SD) and low (−1 SD) total sleep time (TST) intraindividual variability (IIV) based on the adjusted model. Covariates adjusted for are: sex, age, race, education, presence of bed partner, smoking history, perceived stress, and the number of chronic major medical conditions. CAR = cortisol awakening response. Please note that all sleep IIV variables were treated as a continuous (rather than categorical) variable in analyses; this figure displays +/− 1 SD in TST IIV to illustrate the difference in trajectories between high/low IIV.