Sleep disorders and osteoporosis






Introduction


Sleep is critical for life. Sleep and its relationship with our internal circadian system is increasingly recognized as an important factor in optimal human health—the 2017 Nobel prize was awarded to three scientists for their discoveries of the molecular mechanisms controlling the circadian clock because of its impact on human health and disease . Sleep and circadian disturbances (e.g., insufficient sleep duration, shift work, and jet lag) are associated with increased risk of insulin resistance, obesity, cardiovascular disease, and impaired cognition . Normal bone physiology , clock gene deletion studies in animals , and sleep restriction studies in animals and humans all suggest that sleep and circadian disturbance may also be novel risk factors for impaired bone health .



Normal sleep and circadian rhythms


Normal sleep is regulated by two independent, but related processes, sleep (“S”) and circadian (“C”) . Process “S” involves the build-up of sleep pressure with wakefulness. The drive to sleep is influenced by duration of prior wakefulness (sleep deprivation leads to greater homeostatic drive for sleep) and physical and mental activity during wake (brain areas that are more active require more recovery during sleep). Sleep pressure (as measured by adenosine accumulation in the brain) decreases nonlinearly during sleep, dissipating more quickly at the beginning . Process C serves to counteract the accumulating sleep drive from process S across the waking day.


The circadian system is an endogenously generated rhythmic cycle of ~24 hours. In humans the central circadian pacemaker is located in the hypothalamic suprachiasmatic nucleus (SCN) with peripheral clocks located in every cell of the body . The SCN uses environmental signals, principally light exposure, to synchronize its internal period (e.g., 24.1 hours for a “night owl” or 23.8 hours for an “early bird”) to the environmental 24-hour day. When circadian phase shifts occur (e.g., shift work and jet lag) the SCN uses these same signals to reentrain with its environment. The SCN coordinates with peripheral clocks, which oscillate independently and are particularly sensitive to fasting/feeding cycles , via direct neural and hormonal signals (melatonin, core body temperature, cortisol) including the sympathetic nervous system . The circadian wake drive increases across the waking day and the circadian sleep drive increases across the sleeping night. Therefore process “C” helps individuals (1) stay awake later in the day when process “S” has increased sleep drive and (2) stay asleep in the second half of the night when sleep pressure has dissipated. The relationship between processes “C” and “S” is not additive —when homeostatic sleep drive is high (e.g., sleep deprivation), the circadian process has little influence but when sleep drive is low process C dominates. As a result of this complex interaction, there is an optimal window for sleep when it is initiated in the biological evening. Sleep disruption is more common when work, social, or environmental influences (e.g., shift work, jet lag, caffeine, and medications) require that sleep occur at other times.


Circadian rhythms, including the rhythmic expression of proteins and their receptors, are critical for aligning internal biological processes (e.g., digestion) with the organism’s external environment (e.g., food availability) . A core set of clock genes, brain and muscle Arnt-like protein ( Bmal1 ), circadian locomotor output cycles kaput ( Clock ), Period ( Per1, Per2 ) and Cryptochrome ( Cry1, Cry2 ), drive circadian rhythms via a transcriptional-translational feedback loop that cycles every ~24 hours . BMAL and CLOCK dimerize and translocate to the nucleus where they promote transcription of Per and Cry . PER and CRY proteins then dimerize in the cytoplasm and return to the nucleus to inhibit the BMAL1/CLOCK complex . Degradation of PER and CRY alleviate this inhibition and the cycle starts again . Therefore BMAL1/CLOCK are positive transcriptional factors, whereas PER/CRY are repressors . These cycling clock genes influence the transcription and translation of other genes/proteins in various organs and tissues to enable optimal timing of cellular processes .


As with bone metabolism, sleep changes across the lifespan. The SCN deteriorates with age and rhythms become blunted . Sleep complaints increase with age, which is likely multifactorial, including the accumulation of comorbidities (and associated medications) that affect sleep and an increasing incidence of certain sleep disorders (e.g., restless leg syndrome, periodic limb movements) . Although total sleep time may not change with aging itself, changes in sleep architecture (less slow wave sleep) and sleep timing (earlier sleep onset and earlier wake) often result in less restorative sleep. Menopause, a time of increased bone turnover and accelerated bone loss, also significantly impacts sleep independent of aging with increases in awakenings and sleep latency (i.e., the time it takes to fall asleep) . It is possible that menopausal sleep disruptions augment the associated decline in bone mineral density (BMD) seen across the menopausal transition.



Physiologic relationships between sleep/circadian systems and bone metabolism



Bone turnover marker rhythmicity


Serum bone turnover markers (BTMs) should be checked on fasted morning samples because of their diurnal variations and the influence of food intake . Bone resorption markers [C-telopeptide of type I collagen (CTX), N-telopeptide of type I collagen (NTx)] display a large amplitude sinusoidal circadian rhythm, peaking overnight in the early morning hours with a nadir in the late afternoon . This rhythm persists independent of the light/dark cycle, age, sex, menopausal status, posture/bedrest, or parathyroid hormone (PTH) . The presence of clock genes in bone cells and the persistent circadian rhythm in ex vivo murine bone culture suggests that this diurnal variation is due to an endogenous circadian rather than exogenous (behavioral and environmental) rhythm. The endogenous circadian rhythmicity of bone resorption markers was recently confirmed in humans where the 24-hour rhythm of urinary NTx persisted in a “constant routine” protocol where environmental and behavioral stimuli (food intake, posture, light exposure, etc.) are minimized or evenly distributed throughout the day to separate the influence of environmental from endogenous circadian rhythms . Food intake decreases levels of bone resorption markers , possibly mediated by release of glucagon-like-peptide-2 . Bone formation markers [osteocalcin, Procollagen Type 1 N-propeptide (P1NP)] display a much smaller amplitude diurnal variation with overnight peaks in the early morning , similar to bone resorption markers. The amplitude of osteocalcin’s diurnal variation is larger than P1NP , but often larger studies are necessary to demonstrate a statistically significant 24-hour variation in bone formation markers .



Clock gene knockout animal models


Clock genes are present in osteoclasts , osteoblasts , and osteocytes . Numerous animal gene deletion studies have been performed to elucidate the effect of clock genes on bone phenotype ( Table 60.1 ). The altered bone phenotypes observed in most of these studies reinforce that circadian rhythms are important for normal bone turnover and bone mass.



Table 60.1

Bone phenotypes observed with clock gene knockout animal models.






































































































































Clock gene Knockout location Animal/age/sex Bone phenotype Comments Reference
Bmal1 and clock gene knockout models
Bmal1 Intestine Male mice Low bone mass Impaired transcellular calcium absorption; elevated sympathetic tone; bone resorption activated, bone formation suppressed
Bmal1 Germline Mice, 8 weeks, male Low bone mass (seen in trabecular and cortical compartments by micro CT)
Bmal1 Osteoclast Mice, 8 and 12 weeks, male Normal
Bmal1 Mesenchymal cells Mice. 8 weeks, male and female Low bone mass (seen in trabecular and cortical compartments by micro CT) Observed high P1NP and CTX and increased number of osteoclasts
Bmal1 Germline Mice, various ages (mostly >5 months old), male Low bone mass affecting trabecular and cortical compartments; shorter long bone compared to wild type; observed inferior bending strength
Bmal1 Germline Mice, B weeks Low bone mass
Bmal1 Osteoblast Mice Low bone mass High bone turnover
Bmal1 Osteocyte Mice Normal
Bmal1 Osteoclast Mice, 12 weeks High bone mass
Bmal1 Germline Mice Normal
Bmal1 Germline Mice, 8 weeks High bone mass not observed Increased bone formation and resorption; Bmal1 KO mice were hypogonadal
Clock Germline Mice, 12 weeks, male and female Low bone density
Cry and Per gene knockout models
Cry1 and Cry2 Germline Mice High bone mass Skeletal effects mediated by leptin and sympathetic nervous system
Per1 and Per2 Germline Mice High bone mass
Per2 Germline Mice, various (3, 12, 24, 48 weeks), male and female (no sex differences so only female data presented) High bone mass Higher bone formation rate
Cry2 Germline High bone mass Decrease bone resorption
Per2 and Cry2 Germline Normal


As seen in Table 60.1 the clock gene knocked out, knockout location (i.e., germline, cell specific), age, and sex of the mice studied likely influence the observed bone phenotype. In general, germline deletion of the positive molecular clock regulators ( Bmal1, Clock ) results in a low bone mass phenotype and compromised bone strength . Cell specific (osteoblast and mesenchymal cell) Bmal1 deletion studies suggest this low bone mass phenotype is due to high bone turnover affecting both cortical and trabecular compartments . Mesenchymal Bmal1 deficient mice have an increased number of osteoclasts , possibly due to mesenchymal cell RANKL expression as Bmal1 has been reported to control RANKL expression in osteoblasts . In fact, the CTX rhythm amplitude is blunted in Bmal1 mesenchymal knockouts suggesting that mesenchymal cells (e.g., osteoblasts and osteocytes) regulate osteoclast rhythms . The effects of osteoclast specific Bmal1 deletion are mixed ( Table 60.1 ) . Interestingly, intestinal Bmal1 knockout male mice also have low bone mass, in part due to impaired intestinal absorption of calcium and elevated sympathetic tone . Unlike the germline or mesenchymal cell Bmal1 deletions, bone turnover in these Bmal1 intestinal knockouts is uncoupled with suppressed bone formation and increased bone resorption that is partially rescued by propranolol .


Conversely, germline deletion of the negative regulators of the molecular clock machinery ( Cry1 , Cry2 , Per1 , Per2) results in a high bone mass phenotype . Individual deletion of Per2 or Cry2 results in high bone mass via two difference mechanisms . Per2 knockout mice display a higher bone formation rate, whereas Cry2 knockout mice have lower bone resorption . Bone mass is normal when both Per2 and Cry2 are deleted . The high bone mass phenotype observed with germline deletion of Cry1 and Cry2 or Per1 and Per2 is mediated by leptin and the sympathetic nervous system .


Thus clock genes appear to regulate bone turnover in animals using feedback loops to balance the ultimate skeletal effect. Bmal1/Clock (which decrease bone turnover to maintain/support bone mass) stimulate Per/Cry (which counter the effects of Bmal1/Clock by impairing bone mass via increased bone resorption or decreased bone formation) which, in turn, inhibit Bmal1/Clock . These functions may have pathophysiologic implications in the development of osteoporosis with aging and menopause. Murine models of estrogen deficiency demonstrate attenuation of circadian clock gene oscillations in bones and peripheral clocks in aged mice have impaired responses to exercise and stress .



Skeletal response to shift work and sleep duration changes



Animal studies


Bone phenotype is altered in interventional animal studies where sleep/wake or light/dark cycles are manipulated, further supporting the link between bone and the sleep/circadian systems.


Everson et al. exposed 18 male rats to 10 days of sleep restriction followed by 2 days of ad libitum sleep repeatedly over 72 days (~10%–15% of the expected rat lifespan) . Sleep restricted rats had significantly lower indices of bone formation compared to ambulation controls, including decreased osteoid thickness and decreased osteoblast numbers and activity . Despite lower bone formation, sleep restricted rats had higher levels of a bone resorption marker (TRACP 5b), resulting in lower BMD compared to controls .


Xu et al. similarly found significant decreases in BMD and several parameters of bone microarchitecture in 12, 5-month-old female rats restricted to 6 hours of sleep per day for 3 months compared to 12 female controls . Decreases in serum P1NP in sleep restricted compared to control rats were evident by 1 month . NTx levels were similar in sleep restricted compared to control rats until month 3, at which time they declined .


Mice ( N =134) exposed to continuous light exposure for 24 weeks to attenuate central clock rhythmicity had decreased trabecular bone volume compared to 119 age-matched controls exposed to normal 12 hours/12 hours light/dark cycles . Further reinforcing that normal circadian rhythmicity is important for optimal bone health, trabecular changes were reversed over the subsequent 24 weeks when exposure to normal light/dark cycles was restored .



Epidemiological and observational studies


Postmenopausal women in the Nurses’ Health Study who reported 20+ years of night shift work had a higher risk of hip and wrist fractures compared to those women who reported never working the night shift (RR 1.37; 95% CI 1.04, 1.80) . This relationship was particularly strong in those women with a BMI<24 and those who reported never using hormone replacement therapy . The 2009 Korean National Health and Nutrition Examination Survey (KNHANES) study identified more fractures in “other-than-daytime” workers compared to daytime workers but this was not statistically significant (2.1% vs 1.2%, P =.06) .


Three studies performed in Chile , Korea , and the United States have looked at the association between shift work and BMD. The smallest study identified lower L-spine and femoral neck BMD and an increased incidence of osteopenia and osteoporosis in 39 postmenopausal nurses who had rotating shifts compared to 31 postmenopausal nurses who worked daytime shifts . The women included in the study had maintained their respective work schedules for over 10 years; however, rotating workers had an increased prevalence of smoking and coffee use, which was not adjusted for in the analyses . The 2009 KNHANES found that those men and women who self-reported working “other-than-daytime” shifts had lower L-spine and total hip BMD than those working daytime shifts . The average age of participants was significantly younger than other studies, 36.4 years old and this risk was highest among night shift workers, even after adjustment for multiple confounders . Conversely, Santhanam et al. identified higher femoral neck BMD in male, but not female, shift workers using NHANES data, but this association was lost after adjustment for age, gender, and moderate physical activity, and there was no difference in fractures in men or women over 50 years old . Similar to other epidemiological shift work studies, the duration of shift work, age at which shift work is performed, and the presence of current shift work exposure are likely important when establishing associations with long-term outcomes such as BMD and fracture. Intervention studies performed in animals and humans suggest that sleep and circadian disruptions similar to shift work negatively alter bone metabolism and result in lower BMD (see 60.4.3 ).


Several studies have investigated the association between sleep duration and BMD in adults. These studies, including two meta-analyses, have reported that low BMD is associated with both long and short sleep durations (as reviewed in ). Two additional studies found no association between sleep duration and BMD . The most recent meta-analyses identified an increased risk of osteoporosis with short ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-1-Frame class=MathJax style="POSITION: relative" data-mathml='≤’>

7 hours/night) and long ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-2-Frame class=MathJax style="POSITION: relative" data-mathml='≥’>

9 hours/night) sleep duration with the lowest risk observed in middle-aged and elderly adults sleeping 8 hours/night . All of the prior studies used self-reported sleep duration (with or without naps) and used variable cutoffs to define short, normal, and long sleep. Furthermore, the population studied and the anatomical site and method used to determine BMD varied considerably [Dual-energy X-ray Absorptiometry (DXA), quantitative ultrasound (QUS), peripheral quantitative computed tomography (pQCT)], making it difficult to ascertain how BMD at sites that predict fracture relates to sleep duration. In the Study of Osteoporotic Fractures (SOF), longer 24-hour sleep duration (nocturnal sleep+naps) assessed objectively by wrist actigraphy was associated with lower total hip BMD . Only one study has evaluated the association between sleep duration and bone mass accrual in children and found that <8 hours of sleep per night may impair bone mass accrual , potentially impacting lifelong skeletal health.



Human intervention studies


There have been a limited number of sleep/circadian intervention studies in humans. Although the reported outcomes in these studies are limited to BTMs, they are in line with the above animal studies and indicate that sleep and circadian disruption negatively alter BTM balance.


US Army Rangers exposed to sleep deprivation (4 h/night) and energy shortfall (2500–4500 kcal/day expenditure with 1000–4000 kcal/day intake) during their 8-week training had significantly lower levels of bone formation markers [Bone specific alkaline phosphatase (BSAP), osteoclacin] and an increase in one bone resorption marker (TRAP5b) . The changes in osteocalcin and TRAP5b but not BSAP returned to baseline 2–6 weeks posttraining.


A secondary analysis examined the effect of a protocol that induced cumulative sleep restriction combined with circadian disruption (SRCD) on BTMs . Men ( N =10) and women ( N =9) in both younger and older age groups were exposed to recurring 28-hour days (instead of the standard 24-hour day) with 6.5 hours sleep opportunity (~5.6 hours sleep opportunity per 24 hours) over ~3 weeks. Compared to baseline, men had a significant 18%–28% decline in P1NP despite no change in CTX after ~3 weeks of SRCD exposure. The decrease in P1NP was greater in the younger men (20–27 years old) who had higher levels of BTMs at baseline compared to older men (55–65 years old). A similar pattern of BTM change was observed in the women in that study where P1NP and osteocalcin were lower after SRCD and younger women also had an increase in CTX . The magnitude of P1NP decline was again greater in younger women (18–24 years old) despite no significant difference in baseline BTM levels. These data suggest that younger age at the time of sleep/circadian intervention, not baseline BTM level, influences the magnitude of SRCD’s impact on bone. Although participants were not acutely circadian misaligned between their baseline and post-SRCD BTM measurements, it is unclear if these changes were due to independent or combined effects of the sleep restriction or circadian disruption or any residual desynchrony between central and peripheral clocks.



Increased risk of falls and fracture


Sleep and circadian disturbances indirectly affect bone health through an increased risk of falls and fractures. Medications commonly used for insomnia or other sleep/circadian disorders are associated with an increased risk of falls and fractures . However, the underlying sleep/circadian disturbances are associated with an increased risk of falls independent of associated medication use . Nocturnal sleep disruptions are associated with numerous risk factors for fracture, including increased risk of falls, decreased motor function, and impaired cognition . Numerous epidemiological studies have demonstrated the risk between sleep disruptions (including altered sleep duration) and falls/fractures. An analysis of 157,000 women in the Women’s Health Initiative (WHI) identified an increased risk of recurrent falls in individuals with self-reported sleep duration of ≤5 hours (OR 1.27, 95% CI 1.22–1.33) and ≥10 hours (OR 1.24, 95% CI 1.08–1.42) compared to 7 h/night . In addition, short sleep duration (5 hours) was associated with an increased fracture risk out to 5 years of follow-up (HR 1.12, 95% CI 1.05–1.20) . Women with poor sleep quality, as assessed by the WHI Insomnia Rating Scale, also had an increased risk of recurrent falls (OR 1.18, 95% CI 1.15–1.21) . In the SOF, <SPAN role=presentation tabIndex=0 id=MathJax-Element-3-Frame class=MathJax style="POSITION: relative" data-mathml='≤’>

5 h/night of sleep (as measured by wrist actigraphy) was associated with an increased risk of falls, independent of other risk factors including benzodiazepine use (OR 1.52, 95% CI 1.03–2.24) . In a subsequent analysis, women with self-reported sleep duration of <SPAN role=presentation tabIndex=0 id=MathJax-Element-4-Frame class=MathJax style="POSITION: relative" data-mathml='≥’>

10 hours/24 hours period had an increased risk of nonvertebral fractures (HR 1.26, 95% CI 1.00–1.58) compared to women sleeping 8–9 hours . In addition, women who reported daily napping had an increased risk of ≥2 falls (OR 1.32, 95% CI 1.03–1.69) and were more likely to suffer a hip fracture (HR 1.33, 95% CI 0.99–1.78) compared to nonnappers . In Puerto Rican adults with insomnia an increased fall risk was observed in those over 60 years of age and in women . Similar relationships are observed in older men . The Osteoporotic Fractures in Men Study (MrOS) assessed both subjective (self-reported) and objective (wrist actigraphy) measures of sleep duration and quality. Actigraphic, but not self-reported short (≤5 hours and >5–7 h/night) sleep duration was associated with an increased risk of falls compared to those men sleeping >7–8 h/night (OR 1.79, 95% CI 1.22–2.60 and OR 1.42, 95% CI 1.08–1.89, respectively) . An increased risk of falls was also observed in men with excessive self-reported daytime sleepiness (OR 1.52, 95% CI 1.14–2.03), lower sleep efficiency (OR 1.32, 95% CI 1.01–1.72), and nocturnal hypoxemia (≥10% of sleep time spent with SaO 2 <90%; OR 1.62, 95% CI 1.17–2.24) . A recent meta-analyses of seven observational studies (including SOF) identified an increased risk of falls with both long and short sleep durations with the lowest risk of falls occurring with 7–8 hours of sleep per night (OR 1.35, 95% CI 1.17–1.56; OR 1.32, 95% CI 1.21–1.46, respectively) .


Shift work is also associated with risk factors for falls and an increased risk of fractures. The body’s cardiovascular response to postural stress is regulated by the circadian system, potentially increasing the risk for syncope and falls in night shift workers . In fact, postmenopausal women in the Nurses’ Health Study who reported working 20+ years of rotating night shift work had a significant (37%) increased risk of hip and wrist fractures compared to those who reported none .


It is also important to remember that the sleep–bone relationship can be bidirectional where osteoporosis and subsequent fractures affect sleep. It may not be surprising that prolonged pain after fracture affects sleep . However, individuals with osteoporosis were significantly more likely to report shortened sleep time (<6 h/night) in the 2003 National Sleep Foundation Survey .



Obstructive sleep apnea and bone


Obstructive sleep apnea (OSA), characterized by episodes of decreased (hypopnea) and absent (apnea) airflow during sleep, is a complex phenotype typically characterized by multiple comorbidities (obesity, insulin resistance, hypertension, hypogonadism, vitamin D deficiency, etc.), increased inflammation, nocturnal hypoxia, and increased sympathetic tone. Data suggest that OSA may affect bone metabolism . Tomiyama et al. were the first to report increased levels of urinary CTX in men with severe OSA compared to those with mild OSA and hypertensive age-matched controls . After 3 months of CPAP treatment, CTX levels declined . Similarly, obese men with severe OSA had higher levels of CTX and P1NP compared to obese controls . In that study, severity of OSA (as measured by the apnea hypopnea index, AHI) correlated with higher P1NP and PTH and lower vitamin D . These BTM data, exclusively in men, suggest that over time, OSA could be associated with bone loss and lower BMD; however, epidemiological studies have produced conflicting data regarding the relationship between OSA and BMD ( Table 60.2 ).


Oct 27, 2020 | Posted by in ENDOCRINOLOGY | Comments Off on Sleep disorders and osteoporosis

Full access? Get Clinical Tree

Get Clinical Tree app for offline access