Reduced cerebral blood flow and white matter hyperintensities predict poor sleep in heart failure

Heart failure (HF) affects greater than 5 million American adults and is associated with poor outcomes such as cognitive impairment, decreased quality of life, rehospitalizations, and increased morbidity and mortality risk [1‐3]. As many as 80% of HF patients also report experiencing sleep difficulties [4] that are multifaceted in nature, including difficulties falling asleep, insomnia, interrupted sleep at night (e.g., nocturia), restless sleep, and breathing difficulty during sleep [5‐10]. Poor sleep quality in HF, in-turn, is associated with decreased quality of life, depression, impaired self-care, and decreased participation in key treatment recommendations (e.g., physical activity) [11, 12].

Much attention has been paid to correlates and modifiers of poor sleep quality in HF. Sleep quality in HF is affected by an array of demographic, medical, and clinical factors. Demographic variables that negatively influence sleep in HF include older age and being female [5, 7, 13]. In terms of medical factors, taking beta-blockers and medical comorbidities such as respiratory disease, stroke, hypertension, and coronary artery disease have been linked with sleep disturbances [5]. Many of these comorbidities are theorized to negatively affect circadian rhythms subsequent to associated disabilities and limited daytime activity [14]. Known psychosocial modifiers of sleep in HF include depression, anxiety, and perceived health status [5, 13]. Despite these findings, the literature on factors that influence sleep quality in HF is not entirely consistent [5, 7, 13], suggesting poor sleep quality in HF may be more complicated than typically believed.

Although not yet examined, a likely mechanism of poor sleep in HF may involve structural and functional brain abnormalities. Adverse brain changes such as brain atrophy, white matter hyperintensities (WMH), and cerebral hypoperfusion are common in HF and believed to underlie cognitive impairment in this population [15, 16]. Interestingly, recent work also shows poor sleep is associated with reduced cognitive function in HF [11] and HF patients exhibit reduced structural (e.g., atrophy) and functional (e.g., impaired axonal projections) brain integrity among key brain regions that help regulate sleep (e.g., raphe magnus, hypothalamus, medial temporal lobe structures) [17‐20]. Cerebral hypoperfusion is commonly proposed to underlie adverse brain changes in HF, including white matter hypertintensities, and may also impact sleep quality in HF through its impact on the brain and sleep regulation [15, 21]. Indeed, altered cerebral hemodynamics, brain atrophy, and WMH have all been independently correlated with sleep difficulties and regulation (e.g., excessive daytime sleepiness, difficulties with breathing during sleep, arousals during sleep) in other populations [19, 22‐25].

Despite these findings, the mechanisms of sleep difficulties in HF patients remain poorly understood. The purpose of the current study was to examine the effects of cerebral blood flow, WMH, and total brain volume on sleep disturbances in older adults with HF. This study also examined the possible association between sleep quality and volumes of regional brain structures believed to be important for sleep (e.g., thalamus and brain stem) [19, 26]. Based on past work in other samples, we hypothesized that adverse structural brain changes (e.g., greater WMH, smaller total and regional brain volume) would be associated with greater sleep disturbances in HF due to their known role in sleep regulation. We also hypothesized that reduced cerebral blood flow would predict poorer sleep quality in HF in light of its known association with brain insult (e.g., WMH) in this population.

Consistent with the extant literature, the current study shows that reported sleep difficulties are prevalent in older adults with HF. Past work has demonstrated that a series of demographic (e.g., older age) and medical factors (e.g., HF severity) are correlated with poor sleep quality in HF [5], though findings are not entirely consistent across studies [7]. The current findings extend past work by showing that adverse structural and functional brain changes may be important contributors to sleep disturbances in patients with HF.

The current study shows that greater WMH volume is associated with poorer sleep quality in older adults with HF. The exact mechanisms underlying this association are unclear, though there are several possible explanations. Disrupted circadian rhythms have been shown to occur in vascular dementia patients, believed to be the result of deafferentation [41, 42], and WMH may lead to similar circadian rhythm disturbances in HF by interfering with cortical and subcortical neuronal connections to key brain regions that help regulate sleep (e.g., brain stem, suprachiasmatic nucleus) [43‐47]. Supporting this possibility is the current association between WMH and altered core body temperature (i.e., feeling too hot as assessed by item 5 g on the PSQI) and past work that demonstrates HF patients exhibit impaired integrity of axonal projections among brain structures important for sleep, including connections among the basal forebrain, hypothalamus, raphe magnus, and brain stem [20, 48]. WMH are also closely linked with amyloid beta deposition [49] and amyloid beta is a significant contributor to poor sleep quality in pre-clinical Alzheimer’s disease (AD) patients [50]. This is noteworthy, as HF patients are at risk for AD and declining sleep quality may be indicative of AD pathogenesis in this population [50, 51]. Lastly, the known influence of WMH on psychiatric symptoms (e.g., depression) in HF may also contribute to decreased sleep quality [13, 52].

Reduced cerebral blood flow also emerged as a significant predictor of sleep quality in the current sample of HF patients, and may interact with increased WMH to exacerbate sleep disturbances. Cerebral hypoperfusion and subsequent ischemia is commonly proposed to underlie WMH in HF [15, 21] and the current findings showed an inverse association between cerebral perfusion and WMH. Thus, it is likely that chronic disruptions in cerebral hemodynamics affects sleep in HF through its association with WMH, though the cross-sectional design of the current study precludes empirical test of such mediation. Our findings are in the opposite direction of past work examining cerebral blood flow and sleep quality in AD patients [24]. Similarly, sleep spindles help to prevent arousal during sleep, and in contrast to the suggested direction of our findings, past work shows that cerebral blood flow and spindle activity are negatively correlated [25, 53‐56]. The exact reason for the directionality of our findings is not clear, but it is plausible that decreased cerebral blood flow in HF may reflect greater HF severity such as increased number of medical comorbidities and cardiac dysfunction, which are all significant correlates of poor sleep quality in this population [5]. Interestingly, past work shows that cerebral blood flow can be improved through physical activity interventions [57]. This is noteworthy, as improved systemic perfusion in HF has been linked with better neurocognitive outcomes [58]. Because physical activity is a key treatment recommendation in the management of HF, future studies should examine whether greater physical activity results in better sleep quality among HF patients due to improved cerebral circulation.

The current study found no association between total and regional brain volume and sleep quality. These findings further suggest interference with cortico-cortical and cortico-subcortical connections as a likely underpinning for poor sleep quality in HF given the observed impact of WMH on reduced sleep quality in the current study. Although brain atrophy has been linked with reduced sleep quality in other medical populations [19], it is likely that atrophy in the current sample did not reach threshold to produce clinical manifestations. For instance, brain volume loss has been shown among patients with greater HF severity (i.e., LVEF <30) [20] while LVEF in this sample fell within the average range (i.e., mean LVEF = 40). Longitudinal studies that examine the effects of brain atrophy on sleep in HF are needed, particularly as it involves WMH.

The current findings are limited in several ways. The cross-sectional design does not permit causal inferences and prospective studies are needed to confirm our findings. For instance, it is possible that insufficient sleep in HF may lead to reduced cerebral perfusion [23], though this is unlikely in this population in light of the negative effects of cerebral hypoperfusion on the brain [15]. In addition, self-report of sleep quality is limited by biases [59] and future work should use objective assessments of sleep quality (e.g., polysomnography). Similarly, future studies that employ imaging techniques examining axonal tractography would also help elucidate the underlying mechanisms between WMH and sleep quality through direct examination of neuronal pathways that may be occluded by the presence of WMH. We also found an inverse association between physical fitness and sleep quality. Past work shows that exercise interventions in HF lead to improved sleep quality [60] and future work should examine whether such findings are a result of the positive benefits of exercise on the brain. Lastly, although it is notable that the current study controlled for the effects of diagnosed sleep apnea, such history was obtained through self-report and medical record review and information regarding duration, treatment, and severity of sleep apnea is unknown. Consequently, it is possible that undiagnosed sleep apnea or sleep disordered breathing was present in many of the HF participants [10], which may have introduced possible confounds. Future work that employs more objective assessments of sleep quality is needed to clarify and confirm the current findings. Likewise, case controlled studies with larger samples that are more diverse in disease severity and utilize direct measurements of comorbid conditions are much needed to better understand the independent effects of brain abnormalities on sleep in HF.

In brief summary, the current study shows that WMH and decreased cerebral blood flow are associated with poor sleep quality in older adults with HF. Prospective studies are needed to confirm these findings, clarify mechanisms by which WMH and reduced cerebral blood flow disrupt sleep, and examine whether sleep difficulties are modifiable through interventions that target improved cerebral blood flow (e.g., cardiac rehabilitation).