Risk factors of osteoporosis in healthy elderly with unrecognized obstructive sleep apnea: role of physical activity

Sleep Medicine 22 (2016) 25–32

Contents lists available at ScienceDirect

Sleep Medicine j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / s l e e p

Original Article

Risk factors of osteoporosis in healthy elderly with unrecognized obstructive sleep apnea: role of physical activity Emilia Sforza a,*, Magali Saint Martin a, Thierry Thomas b, Philippe Collet b, Martin Garet a, Jean Claude Barthélémy a, Frédéric Roche a a

Service de Physiologie Clinique et de l’Exercice, Faculté de Médecine Jacques Lisfranc, COMUE de Lyon, Université Jean Monnet, Saint-Etienne, France Service de Rhumatologie, CHU Saint-Etienne, Faculté de Médecine Jacques Lisfranc, Inserm U1059 LBTO, COMUE de Lyon, Université Jean Monnet, Saint-Etienne, France b

A R T I C L E

I N F O

Article history: Received 12 January 2016 Received in revised form 13 April 2016 Accepted 19 April 2016 Available online 7 June 2016 Keywords: Osteoporosis Elderly OSA Obesity Metabolic factors Physical activity

A B S T R A C T

Objective: Several studies suggest a relationship between bone mineral density (BMD) anthropometric and metabolic variables, and obstructive sleep apnea (OSA); all of these factors have an effect on osteoporosis (OS) risk. This cross-sectional study explores these associations in a large sample of older subjects with and without OSA. Methods: Volunteers were recruited from the PROgnostic indicator OF cardiovascular and cerebrovascular events survey. A total of 461 subjects, aged 68.7 ± 0.8 years, were examined, blood samples were taken, and they were subjected to home polygraphy, assessment of daily energy expenditure (DEE), and dual-energy X-ray absorptiometry. Results: Osteopenia (OP) was detected in 44% of subjects at the femoral and 39% at the vertebral level, while the prevalence of OS was lower at the femoral (4%) and vertebral (12%) levels. As expected, women had a higher prevalence of OP and OS. Subjects with OP and OS had a tendency to have lower DEE and values of obesity, apnea–hypopnea index (AHI), and indices of hypoxemia (ODI). At the correlation analyses, anthropometric factors and DEE were significantly related to BMD with a slight effect of indices of OSA severity. After adjustment for confounding variables, univariate and multivariate regression analyses showed a strong significant association between femoral and lumbar BMD and T-score and DEE without contribution of metabolic data and with a slight negative effect of respiratory factors. Conclusions: In this sample of the elderly, physical activity was the best predictor of OS with a slight effect of body mass index. The indices of OSA confirm their protective effect on bone mineral density. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Bone remodeling is a lifelong process of bone tissue resorption by osteoclasts and new bone formation by osteoblasts. To remain functional, bone apposition must balance the resorption of bone tissue, a process critical in the repair of bone injury. Osteoporosis (OS) is a disorder characterized by bone mass loss and microarchitectural deterioration of bone tissue, resulting in skeletal fragility and fracture susceptibility, both listed as significant risk factors for fractures, morbidity, and mortality in the elderly [1]. There is ample evidence showing that anthropometric factors, hypertension (HT) [2], diabetes [3], obesity [4], low physical activity [5], and obstructive sleep apnea (OSA) [6] are implicated in the decreased bone mineral density (BMD) in the elderly. The association between

Clinical trial registrations: NCT 00759304 and NCT 00766584. * Corresponding author. Service de Physiologie Clinique et de l’Exercice, CHU Nord, Niveau 6, F-42055 Saint Etienne, France. Tel.: +33 4 77 82 83 00; fax: +33 4 77 82 84 47. E-mail address: [email protected] (E. Sforza). http://dx.doi.org/10.1016/j.sleep.2016.04.010 1389-9457/© 2016 Elsevier B.V. All rights reserved.

each of these factors and OS prevalence has been extensively studied, however, with conflicting results in particular for OSA. Considering diabetes mellitus (DM), it is generally recognized that while DM1 is associated with reduced BMD, DM2 is linked to normal or increased BMD [7,8]. Fat mass and obesity [9] may promote bone formation via higher mechanical load [10,11], hence playing a protective role in OS occurrence. When we consider the effects of OSA, clinical [12] and experimental studies [13] suggest that intermittent hypoxia acts as a stimulator of formation and activation of monocytes and macrophages, which are closely related to osteoclast formation, and thus bone resorption [6,14]. Moreover, stimulation of chemo- and baroreceptors during OSA-related intermittent hypoxia induces a rise in sympathetic activity that directly or via the leptin pathway play a role in bone remodeling [15,16]. Finally, in patients with chronic obstructive pulmonary disease associated with or without OSA [17–19], loss of energy expenditure leads to bone loss and, in turn, increased fracture risk. Thus a variety of factors may act on bone metabolism, which possibly explains the controversial and heterogeneous results in the literature and precludes a clear assessment of whether physical,

26

E. Sforza et al. / Sleep Medicine 22 (2016) 25–32

metabolic, vascular, and OSA factors are really associated with BMD dysfunction. Since, in the elderly, the aging processes increase the risk of both OS and OSA, we decided to perform a more detailed cross-sectional study in a healthy older population to: (1) confirm previous clinical and experimental studies describing a protective role of OSA on bone metabolisms; and (2) evaluate if other factors, ie, metabolic factors, obesity, and physical activity, might play a greater role than OSA in the OS risk in healthy elderly. 2. Subjects and methods 2.1. Subjects The study sample was recruited from the PROOF (PROgnostic indicator OF cardiovascular and cerebrovascular events) survey [20], a longitudinal study on a population-based cohort of 1011 volunteers aged ≥65 years in the city of Saint-Etienne (France). The first step of the study was conducted among the inhabitants of the city of SaintEtienne, France, from 2001 to 2003. Participants were eligible if aged 65 years at the date of inclusion. Among the 3983 eligible participants, 11% declined participation and 67% did not reply. The final study population included 1011 participants for whom clinical interview and neurological and cardiological examinations were available. Three years later, an ancillary study addressing the association between OSA assessed by at-home polygraphic study, and cardiovascular, cerebrovascular, metabolic, and BMD morbidity during a seven-year followup (Synapse study) was proposed to the overall group but accepted by only 854 participants, 58.5% women and 41.5% men, aged 68.7 ± 0.8 years at the study entry. Exclusion criteria were as follows: previous myocardial infarction, arrhythmia, cardiac pacemaker, stroke, neurological or psychiatric disease, DM1, chronic obstructive pulmonary disease, and cerebral magnetic resonance imaging results suggesting neurological disease or initial dementia. In order to exclude a possible bias on the longitudinal assessment, participants with previous or actual OS diagnosis and treatment were excluded by the analysis. After exclusion of subjects under anti-OS medication or selfreported OS, and those refusing polygraphy and/or analysis of energy expenditure and dual-energy X-ray absorptiometry (DXA), 461 subjects aged 68.7 ± 0.8 years were available for this study. The PROOF and the Synapse studies were approved by the University Hospital and the IRB-IEC (CCPRB Rhône-Alpes Loire). The National Committee for Information and Liberty (CNIL) approved the data collection. All subjects gave written consent prior to participation in the study. 2.2. Methods 2.2.1. Clinical assessment Detailed clinical assessment was focused on cardiac and cerebrovascular diseases, HT, DM2, as well as respiratory, neurological, and psychiatric disorders. Current medication use was analyzed with regard to antihypertensive, antidiabetic, hypolipidemic, as well as hypnotic, anxiolytic, or antidepressant therapy, and assessed by selfreport and/or medical prescription. None of the subjects were taking medication known to influence bone mineralization such as corticosteroids, heparin, anticonvulsants, biphosphonates, and vitamin D. Subjects were defined as normotensive if they did not report a history of HT or antihypertensive treatment or, at ambulatory blood pressure monitoring, did not show a mean systolic blood pressure >135 mmHg and a mean diastolic pressure >85 mmHg. For ethical reasons, subjects did not discontinue hypolipidemic, diabetic, or hypertensive medication for the study. 2.2.2. Anthropometrical measurements During clinical evaluation, height (without shoes) and weight (light clothing) were measured; body mass index (BMI) was cal-

culated as weight/height squared (kg/m2). In accordance with the National Institutes of Health clinical guidelines, an overweight adult was defined as BMI > 25 kg/m2 and obesity as BMI > 30 kg/m2. Neck circumference (NC) was measured at the midway point of the neck between mid-cervical spine and mid-anterior neck to 0.5 cm below the laryngeal prominence. Waist circumference (WC) was measured midway between the lower rib margin and the iliac crest. Hip circumference (HC) was measured over minimal clothing at the level of greatest protrusion of the upper hip bone. The ratio between hip and waist circumference was also calculated (HC/WC). 2.2.3. Measurement of metabolic risk factors Blood was collected for high-sensitivity assessment of biological factors the morning after polygraphy. Serum levels of glycemia, lipids, including total cholesterol, high-density lipoprotein cholesterol (HDL), calculated low-density lipoprotein cholesterol (LDL), and triglycerides (TRG) were assessed using enzymatic kits (Roche Diagnostics). Normal ranges for biochemical parameters were set as follows: serum TRG (<1.5 g/L), serum cholesterol (<2 g/L), serum LDL (<1.60 g/ L), serum HDL (<0.40 g/L), and serum glycemia (0.70–1.05 g/L). 2.2.4. Measurement of body composition by DXA Subjects were referred to the St Etienne Rheumatologic Research Group of the University Hospital for bone metabolism evaluation, measured by whole body DXA scanning (Hologic QDR2000, software version V5.67A, Hologic Inc., Bedford, MA, USA). The standard procedures described in the literature for DXA measurement were applied [21]. BMD was measured in all subjects at the proximal femur and lumbar spine (L1–L4) with a coefficient of variation of 0.8% and <1.2%, respectively. BMD was expressed in g/cm2 and as peak bone mass percentage in normal subjects (T-score). BMD results at the femoral neck and lumbar spine were classified into three groups according to WHO existing criteria: normal (T-score > −1.0 SD), OP (Tscore −1.0 to −2.5 SD), and OS (T-score < −2.5 SD) [22]. 2.2.5. Daily energy expenditure Daily energy expenditure (DEE) was assessed by a selfadministered physical activity questionnaire concerning seven main dimensions of everyday life with specific emphasis on autonomy and perceived exertion [23]. This questionnaire provided a quantitative picture of an individual’s mean habitual activities, based on calculation of DEE (kJ/24 h) = sum of intensity activity (J/min per kg) and duration of specific activity (min/day) according to age, weight, severity of the condition considered, and autonomy. 2.2.6. Ambulatory home polygraphy All subjects underwent a full-night ambulatory recording using a polygraphic system (HypnoPTT, Tyco Healthcare, Puritan Bennett, USA). The following parameters were included: sound measurement, electrocardiography, pulse transit time, nasal pressure, respiratory effort, and body position. Oxygen saturation (SaO2) was measured by pulse oximetry with a sampling rate of 250 Hz. To minimize potential overestimation of sleep duration, subjects completed the St Mary’s Hospital questionnaire, while wakefulness before lightsoff was excluded by the analysis. All examinations were visually validated and manually scored for respiratory events and nocturnal SaO2 according to the Chicago criteria [24] by a single scorer (F.R.), with intrascorer reliability for each evaluation being of 87%. Hypopnea was defined as a 50% or greater reduction in airflow from the baseline value lasting ≥10 s and associated with at least 3% oxygen desaturation. Because of the differences in the literature for oxygen desaturation degree, we compared the indices of oxygen desaturation putting the threshold, respectively, at 3 and 4% without, however, significant changes in the results. Apnea was defined as

E. Sforza et al. / Sleep Medicine 22 (2016) 25–32

the absence of airflow on the nasal cannula lasting for ≥10 s. The absence of rib cage movements associated with apnea defined the event as central, while a progressive increase in pulse transit time and respiratory efforts allowed definition of the episode as obstructive. Apnea–hypopnea index (AHI) is defined as the ratio of the number of obstructive apneas and hypopneas per hour of reported sleeping time. The indices of nocturnal hypoxemia were as follows: mean SaO2; percentage of recording time with a SaO2 ≤90%; minimum SaO2 value recorded during sleep (minimum SaO2); and oxygen desaturation index (ODI), ie, the number of episodes of oxyhemoglobin desaturation per hour of reported sleep time during which blood oxygen level fell by 3% or more. Pulse transit time was continuously monitored, and autonomic respiratory-related and total autonomic arousal indices were calculated after visual correction. According to recent data in elderly subjects [25] an AHI >15 with at least 50% of events scored as obstructive was considered diagnostic of OSA. Cases were subsequently stratified as mild-tomoderate (15 ≥ AHI < 30) and severe SDB cases (AHI ≥ 30). 2.3. Statistical analysis The results are presented as means ± standard deviation (SD) for continuous variables and percentages for categorical variables. Statistical significance was assessed by chi-square test for categorical variables and Student’s t-test for continuous variables. ANOVA was used to assess differences in subjects, according to severity of the AHI and BMD changes. Potential risk factors for changes in BMD, anthropometric measurements, AHI, indices of hypoxemia, DEE, and metabolic values were estimated using Pearson’s correlation coefficient. Univariate and multivariate regression analyses were then performed to analyze the relationship between BMD and T-scores at femoral and lumbar values (dependent variables) and gender, DEE, and anthropometric, metabolic, and polygraphic data. Data were analyzed using the Statistical Package for the Social Sciences version 17 for Windows (SPSS Inc., Chicago, IL, USA) and statistical significance was set at p < 0.05. 3. Results Clinical, anthropometric, and metabolic data for the total group and for men and women are summarized in Table 1. The total group

27

had a mean age of 68.7 ± 0.8 years and consisted of 57.5% females; 52% of the subjects showed normal weight and 8% obesity. DM2 was found in 5% of participants, dyslipidemia in 33%, and HT in 40%. The mean AHI reached 20.0 ± 14.2 and the mean ODI 8.5 ± 8.8. Comparison analysis between women and men revealed that men were more frequently overweight (50%) and obese (16%). As expected, gender differences were found for anthropometric, DEE, and vascular factors with men having greater NC, WC, DEE, and HT. When metabolic data were considered, men had higher values of glycemia and TRG, while women showed higher total HDL and LDL cholesterol values. Polygraphic data revealed that men had more severe nocturnal hypoxemia and greater AHI (p < 0.001). Table 2 reports the clinical, anthropometric, metabolic, and polygraphic data for subjects stratified according to AHI. Subjects without OSA (AHI < 15) represented 44% of the total population, 37% was labeled as mild-moderate (≥15 AHI <30), and 19% as severe cases (AHI ≥ 30). Comparison analysis between groups revealed a tendency to progressive and significant increase in BMI and the other indices of obesity paralleling the severity of OSA. As expected, there were significant differences for AHI and indices of nocturnal hypoxemia both in moderate (p = 0.001) and severe (p < 0.001) cases. No differences between groups were found for all metabolic variables. Table 3 illustrates the clinical, anthropometric, metabolic, and polygraphic data for subjects with normal BMD and for subjects with OS and OP at femoral and lumbar sites. At the femoral level, 44.2% had OP and 4.8% OS. There was a trend toward a progressive decline for DEE, BMI, WC, HC, and LDL-cholesterol associated with the BMD changes. Considering the indices of OSA severity, subjects with OP and OS were those having lower AHI (p = 0.03) and ODI (p < 0.006). No effect was found for the other metabolic parameters. At the lumbar site, 49% of the sample had normal BMD, 38.8% OP, and 12.1% OS. Again, subjects with OS had lower levels of DEE and indices of obesity, ie, BMI, WC, HC, and NC. Any of the metabolic variables differ between groups. Concerning OSA severity, again subjects with OS had a lower mean value of AHI as well as ODI. Fig. 1 illustrates the prevalence of normal BMD, OP, and OS at femoral and vertebral sites in the total group and according to gender. While at the femoral level, gender differences were found for subjects with normal BMD and OP, the latter greater in women, at the vertebral site the only difference present was for OS being slightly significant in women.

Table 1 Clinical, anthropometric, metabolic, and polygraphic data for the whole group and for women and men separately (mean ± SD).

Age (y) BMI (kg/cm2) NC (cm) WC (cm) HC (cm) HC/WC ratio GLY (g/L) TRG (g/L) Cholesterol (g/L) HDL (g/L) LDL (g/L) Diabetes (%) Dyslipidemia (%) HT (%) DEE (kJ·24 h−1) AHI (n/h) ODI (n/h) Time SaO2 < 90% (min) Minimal SaO2 (%)

Total (n = 461)

Women (n = 265)

Men (n = 196)

p

68.7 ± 0.8 25.1 ± 3.6 37.0 ± 3.9 85.5 ± 11.2 98.1 ± 8.2 0.87 ± 0.08 0.99 ± 0.22 1.06 ± 0.53 2.2 ± 0.33 0.6 ± 0.18 1.4 ± 0.31 4.6 33.3 40.2 10,342.8 ± 1,915.6 20.0 ± 14.2 8.5 ± 8.8 1.9 ± 6.9 89.9 ± 4.1

68.7 ± 0.8 24.8 ± 4.0 34.8 ± 2.5 81.6 ± 10.4 96.6 ± 9.0 0.84 ± 0.07 0.96 ± 0.17 0.98 ± 0.4 2.3 ± 0.32 0.67 ± 0.18 1.38 ± 0.31 3.8 37.7 38.1 9,705.2 ± 1,758.6 16.5 ± 12.1 6.4 ± 7.0 1.8 ± 8.1 90.5 ± 3.8

68.7 ± 0.7 25.6 ± 2.9 40.0 ± 2. 90.6 ± 10.2 98.1 ± 6.9 0.92 ± 0.07 1.02 ± 0.23 1.16 ± 0.65 2.1 ± 0.33 0.56 ± 0.16 1.34 ± 0.30 5.6 27.2 43.1 11,201.5 ± 1,781.8 24.815.4 11.4 ± 10.8 2.0 ± 4.7 89.1 ± 4.4

ns 0.006 <0.001 <0.001 ns <0.001 0.001 0.001 <0.001 <0.001 0.04 ns 0.01 0.01 <0.001 <0.001 <0.001 ns <0.001

Abbreviations: BMI: body mass index; NC: neck circumference; WC: waist circumference; HC: hip circumference; GLY: glycemia; TRG: triglycerides; HDL: high density lipoprotein cholesterol; LDL: low density lipoprotein cholesterol; HT: hypertension; DEE: daily energy expenditure; AHI: apnea––hypopnea index; ODI: oxygen desaturation index; SaO2: oxygen saturation. p: Student’s t-test or chi-square differences between women and men.

28

E. Sforza et al. / Sleep Medicine 22 (2016) 25–32

Table 2 Clinical, anthropometric, metabolic, and polygraphic data for the three groups of subjects stratified according to the apnea–hypopnea index (mean ± SD).

Gender (% women) BMI (kg/cm2) NC (cm) WC (cm) HC (cm) HC/WC ratio GLY (g/L) TRG (g/L) Cholesterol (g/L) HDL (g/L) LDL (g/L) Diabetes (%) Dyslipidemia (%) HT (%) DEE (kJ·24 h−1) AHI (n/h) ODI (n/h) Time SaO2 < 90% (min) Minimal SaO2 (%)

AHI < 15 (n = 205)

≥15 AHI <30 (n = 169)

AHI ≥ 30 (n = 87)

p

70% 24.5 ± 3.4 36.0 ± 3.2 82.8 ± 10.6 0.96.8 ± 7.7 0.86 ± 0.08 0.98 ± 0.2 1.06 ± 0.6 2.2 ± 0.33 0.7 ± 0.2 1.4 ± 0.3 4.9 38.4 35.6 109,956.4 ± 1,774.4 8.2 ± 4.0 3.1 ± 2.7 1.0 ± 5.7 91.6 ± 2.8

51% 25.4 ± 3.9 37.2 ± 3.8 86.4 ± 11.9 98.0 ± 8.8 0.84 ± 0.07 0.98 ± 0.18 1.03 ± 0.5 2.2 ± 0.34 0.62 ± 0.16 1.38 ± 0.31 4.7 29.6 40.2 10,391.0 ± 2,032.8 23.1 ± 6.1 9.5 ± 6.2 2.3 ± 8.5 88.8 ± 4.5

39% 26.3 ± 3.1 39.1 ± 3.5 89.1 ± 9.5 99.9 ± 7.5 0.90.0 ± 0.8 1.02 ± 0.21 1.10 ± 0.5 2.2 ± 0.31 0.57 ± 0.14 1.02 ± 0.21 3.5 27.9 51.2 11,159.9 ± 1,748.5 42.0 ± 11.1 19.2 ± 11.7 2.9 ± 5.7 88.0 ± 4.4

<0.001 <0.001 <0.001 <0.001 0.003 <0.001 ns ns ns <0.001 ns ns 0.01 0.05 <0.001 <0.001 <0.001 0.03 <0.001

Abbreviations: BMI: body mass index; NC: neck circumference; WC: waist circumference; HC: hip circumference; GLY: glycemia; TRG: triglycerides; HDL: high density lipoprotein cholesterol; LDL: low density lipoprotein cholesterol; DEE: energy expenditure; HT: hypertension; AHI: apnea–hypopnea index; ODI: oxygen desaturation index; SaO2: oxygen saturation. p: ANOVA (differences between groups).

Correlation analysis (Table 4) revealed that DEE was strongly associated with BMD and T-score at both femoral and lumbar levels. Moreover, BMD and T-score correlated positively with all anthropometric variables, particularly in HC and WC. Considering metabolic measurements, glycemia and cholesterol levels were negatively and significantly correlated with BMD and T-score at both bone levels, and greater at the femoral site. For polygraphic data, a slight effect

was found for ODI, in particular at the femoral level, with a low contribution of AHI. Univariate and multivariate regression analyses using a stepwise method were performed to examine the independent associations between the BMD and T-score at the femoral as well as lumbar sites and DEE, gender, AHI or ODI, glycemia, cholesterol, and anthropometric data. At the femoral level (Table 5), BMD was

Table 3 Anthropometric, biological, and polygraphic data of subjects with normal bone mineral density and subjects with osteopenia or osteoporosis at the femoral and vertebral levels (mean ± SD). Femur

Normal BMD (n: 235)

Osteopenia (n: 204)

Osteoporosis (n: 22)

p

DEE (kJ·24 h−1) BMI (kg/cm2) WC (cm) HC (cm) HC/WC ratio (cm) NC (cm) Glycemia (g/L) Triglycerides (g/L) Cholesterol (g/L) HDL (g/L) LDH (g/L) AHI (n/h) ODI (n/h) SaO2 minimal (%)

10,920.68 ± 1,834.31 26.1 ± 3.5 88.5 ± 10.9 100.0 ± 8.3 0.89 ± 0.08 37.9 ± 3.5 1.00 ± 0.2 1.07 ± 5.2 2.2 ± 0.3 0.61 ± 0.2 1.33 ± 0.3 21.1 ± 15.3 9.5 ± 9.6 89.2 ± 4.4

9,856.3 ± 1,822.3 24.3 ± 3.4 82.7 ± 10.5 95.9 ± 7.5 0.86 ± 0.08 36.1 ± 3.8 0.97 ± 0.2 1.06 ± 0.5 2.3 ± 0.3 0.64 ± 0.2 1.41 ± 0.3 19.5 ± 13.2 7.41 ± 7.9 90.5 ± 3.6

8,657.85 ± 1,442.5 22.3 ± 3.4 79.4 ± 12.1 92.3 ± 6.2 0.88 ± 0.09 35.9 ± 2.9 0.93 ± 0.1 0.94 ± 0.5 2.2 ± 0.3 0.64 ± 0.2 1.36 ± 0.3 14.3 ± 10.1 4.41 ± 5.3 90.7 ± 4.6

<0.001 <0.001 <0.001 <0.001 ns 0.01 0.05 ns ns ns 0.006 0.03 0.006 ns

Vertebral spine

Normal BMD (n: 226)

Osteopenia (n: 179)

Osteoporosis (n: 56)

p

10,747.26 ± 1,930.61 25.8 ± 3.5 87.0 ± 10.3 98.9 ± 8.3 0.88 ± 0.08 37.5 ± 3.6 1.00 ± 0.2 1.1 ± 0.6 2.2 ± 0.3 0.61 ± 0.2 1.36 ± 0.3 20.7 ± 17.9 8.9 ± 8.8 89.5 ± 4.3

10,205.55 ± 1,764.3 24.8 ± 3.4 85.4 ± 10.9 97.8 ± 7.7 0.88 ± 0.08 36.7 ± 3.7 0.98 ± 0.2 1.0 ± 0.5 2.2 ± 0.3 0.64 ± 0.2 1.36 ± 0.3 20.4 ± 14.2 8.7 ± 9.1 90.1 ± 4.0

9,209.07 ± 1,820.9 23.6 ± 3.8 79.9 ± 11.6 93.6 ± 7.6 0.85 ± 0.08 35.6 ± 3.3 0.94 ± 0.1 0.99 ± 0.5 2.3 ± 0.4 0.66 ± 0.2 1.42 ± 0.4 16.4 ± 10.6 6.5 ± 7.5 90.6 ± 3.9

<0.001 <0.001 <0.001 0.001 <0.001 0.0001 0.05 ns ns ns ns 0.04 0.05 ns

−1)

DEE (kJ·24 h BMI (kg/cm2) WC (cm) HC (cm) HC/WC ratio (cm) NC (cm) Glycemia (g/L) Triglycerides (g/L) Cholesterol (g/L) HDL (g/L) LDL (g/L) AHI (n/H) ODI (n/h) SaO2 minimal (%)

Abbreviations: DEE: energy expenditure; BMI: body mass index; WC: waist circumference; HC: hip circumference; NC: neck circumference; HDL: high density lipoprotein cholesterol; LDL: low density lipoprotein cholesterol; AHI: apnea–hypopnea index; ODI: oxygen desaturation index; SaO2: oxygen saturation. p: ANOVA.

E. Sforza et al. / Sleep Medicine 22 (2016) 25–32

29

Fig. 1. Schematic representation of prevalence of subjects with normal bone mineral density, osteopenia, and osteoporosis in the total sample and according to gender. The prevalence of osteoporosis was low with a tendency to greater prevalence in women at the vertebral level. While osteopenia did not differ between sexes at the vertebral spine, women had greater osteopenia at the femoral level.

significantly related to DEE, BMI, indices of obesity (p < 0.0001), and cholesterol value (p < 0.002), as well as AHI (p = 0.0003) and ODI (p > 0.00001). At the multivariate analysis, the DEE was the best predictor of BMD (p < 0.00001) with a small contribution of AHI (p = 0.33). At the vertebral site (Table 6), univariate regression analysis showed the significant role of DEE and anthropometric values (p < 0.001) with an effect of glycemia (p = 0.0004), HDL cholesterol (p < 0.0001), AHI (p < 0.0001), and minimal AaO2 (p = 0.0004). Multivariate regression analysis for BMD at the vertebral site showed that DEE (p < 0.00001), HDL cholesterol (p = 0.02), and AHI (p = 0.04) were the only variables affecting lumbar BMD. Stepwise regression analysis between T-score at lumbar and femoral levels (Table 7) and physical activity, metabolic and nocturnal respiratory data, confirmed the key role of DEE with a small contribution of BMI. Interestingly, neither metabolic nor AHI and ODI contribute to the occurrence of T-score changes. 4. Discussion In this sample of healthy elderly individuals/participants over 65 years of age, we examined which factor mostly contributes to

OS risk, taking into account the presence of OSA and physical, metabolic, and anthropometric measurements, overall known to affect BMD. The first and major finding of our study was that in this sample of older subjects, physical activity as assessed by the DEE questionnaire was the most significant factor affecting BMD and T-score at femoral and lumbar levels with a little added contribution of BMI at the femoral level. Moreover, neither diabetes, dyslipidemia, and indices of obesity nor OSA parameters significantly contribute to the OS occurrence. Therefore, we can conclude that the increased risk of OS and perhaps, also, most likely fracture is strongly related to physical activity and lifestyle, with a small influence of BMI, and without any significant contribution of metabolic and sleeprelated breathing disorders. It is well known that BMD reaches a peak at approximately 40 years of age and subsequently gradually decreases [26] with a sharp decline after menopause in women, and at an age exceeding 60 years. Thus, in the aging population, OS is an important worldwide health problem with increasing fracture susceptibility and decreasing patient activity and quality of life [27]. The assessment of the factors contributing to OS/OP in the elderly represents an interesting background allowing prevention and therapeutic strategies.

30

E. Sforza et al. / Sleep Medicine 22 (2016) 25–32

Table 4 Pearson’s correlation coefficients of bone mineral density and T-score at two sites and clinical, metabolic, and polygraphic data.

Femoral BMD Femoral T-score Lumbar BMD Lumbar T-score HC (cm) NC (cm) WC (cm) HC/WC ratio BMI (kg/cm2) Glycemia (g/L) Triglycerides (g/L) Cholesterol (g/L) HDL (g/L) LDL (g/L) AHI (n/h) ODI (n/h) Minimal SaO2 (%) DEE (kJ·24 h−1)

Femoral BMD

Femoral T-score

Lumbar BMD

Lumbar T-score

1 0.864** 0.612** 0.551** 0.282** 0.439* 0.376** 0.266** 0.316** 0.190* 0.115* −0.149** −0.092* −0.144** 0.168** 0.268** −0.191** 0.453**

0.864** 1 0.600** 0.626** 0.345** 0.254** 0.376** 0.266** 0.359** 0.190** 0.115* −0.149** −0.090 −0.086 0.111* 0.196** −0.175** 0.386**

0.612** 0.600** 1 0.951** 0.242** 0.389** 0.320** 0.211** 0.281** 0.135** 0.120** −0.102* −0.221** −0.022 0.142** 0.149** −0.130** 0.396**

0.551** 0.626** 0.951** 1 0.204** 0.231** 0.217** 0.137** 0.248** 0.105* 0.060 −0.085 −0.113* −0.48 0.142** 0.149 −0.130** 0.294**

Abbreviations: BMD: bone mineral density; HC: hip circumference; NC: neck circumference; WC: waist circumference ratio; BMI: body mass index; HDL: high density lipoprotein cholesterol; LDL: low density lipoprotein cholesterol; AHI: apnea– hypopnea index, ODI: oxygen desaturation index; SaO2: oxygen saturation; DEE: energy expenditure. p = Pearson’s correlation coefficients: * = 0.05; ** = 0.01.

When mechanisms implicated in OS prevalence are considered, several studies suggest an association between OS/OP and metabolic disorders such as obesity [4], metabolic syndrome [7,8], diabetes [4,8], and lipids [4], with an effect either on bone formaTable 5 Univariate and multivariate regression analyses showing the association between bone mineral density at the femoral level and explicative polygraphic, anthropometric, and biological data. Univariate regression analysis of femoral BMD (continuous dependent variable) Independents

R

R2

F-value

p-value

DEE (kJ·24 h−1) BMI (kg/cm2) HC/WC ratio (cm) Glycemia (g/L) Triglycerides (g/L) Cholesterol (g/L) HDL (g/L) LDL (g/L) AHI (n/h) ODI (n/h) Time SaO2 < 90% (%) Mean SaO2(%) SaO2 minimal (%)

0.453 0.316 0.266 0.190 0.115 0.149 0.092 0.144 0.168 0.268 0.078 0.196 0.191

0.205 0.100 0.071 0.036 0.013 0.022 0.009 0.021 0.028 0.072 0.006 0.038 0.037

11.7 50.5 34.6 17.1 6.2 10.3 3.92 9.66 13.3 35.8 2.57 18.2 17.4

<0.0001 <0.0001 <0.0001 <0.0001 0.02 0.002 0.05 0.002 0.0003 <0.0001 0.11 <0.0001 <0.0001

Multivariate regression analysis (model: R = 0.490, R2 = 0.0241, F-value = 15.71, p < 0.0001) for femoral BMD (continuous dependent variable) Independents

Coefficient

SE/SD

t-value

p value

DEE (kJ·24 h−1) BMI (kg/cm2) HC/WC ratio (cm) Glycemia (g/L) Triglycerides (g/L) LDL (g/L) AHI (n/H) ODI (n/h) Mean SaO2(%) SaO2 minimal (%)

0.00003 −001 0.130 0.057 0.004 0.052 0.0005 −0.002 −0.06 0.002

0.0.000005/0.398 0.002/−0.034 0.083/0.075 0.033/0.079 0.013/0.016 0.037/0.062 0.0004/0.045 0.001/0.138 0.004/−0.080 0.002/0.066

6.92 −0.57 1.571 1.74 0.34 1.326 0.956 2.69 −1.582 1.16

<0.0001 0.57 0.12 0.08 0.73 0.12 0.33 0.008 0.12 0.25

Abbreviations: DEE: energy expenditure; BMI: body mass index; WC/HC ratio: waist circumference/hip circumference ratio; HDL: high density lipoprotein cholesterol; LDL: low density lipoprotein cholesterol; AHI: apnea–hypopnea index; ODI: oxygen desaturation index; SaO2: oxygen saturation.

Table 6 Univariate and multivariate regression analyses showing the association between bone mineral density at the lumbar level and explicative polygraphic, anthropometric, and biological data. Univariate regression analysis of lumbar BMD (continuous dependent variable) Independents

R

R2

F-value

p-value

DEE (kJ·24 h−1) BMI (kg/cm2) HC/WC ratio (cm) Glycemia (g/L) Triglycerides (g/L) Cholesterol (g/L) HDL (g/L) LDL (g/L) AHI (n/h) ODI (n/h) Time SaO2 < 90% (%) Mean SaO2 (%) SaO2 minimal (%)

0.396 0.281 0.211 0.135 0.120 0.102 0.221 0.022 0.208 0.193 0.08 0.148 0.164

0.157 0.080 0.044 0.0018 0.014 0.010 0.049 0.00005 0.043 0.037 0.006 0.002 0.027

84.8 39.2 21.2 8.4 6.7 4.8 23.3 0.225 20.7 17.6 2.9 10.2 12.6

<0.0001 <0.0001 <0.0001 0.0004 0.01 0.06 <0.0001 0.63 <0.0001 0.08 0.08 0.002 0.0004

Multivariate regression analysis (model: R = 0.430, R2 = 0.0185, F-value = 10.1, p < 0.0001 dependent variable) Independents

Coefficient

SE/SD

t-value

p value

DEE (kJ·24 h−1) BMI (kg/cm2) HC/WC ratio (cm) Glycemia (g/L) Triglycerides (g/L) HDL (g/L) AHI (n/H) ODI (n/h) Mean SaO2(%) SaO2 minimal (%)

0.00003 −0.000013 0.015 0.028 −0.002 −0.105 0.001 0.001 −0.04 0.001

0.000005/0.332 0.03/−0.003 0.098 0.038/0.034 0.015/−0.007 0.045/0.115 0.001/0.100 0.001/0.057 0.005/−0.042 0.002/0.027

5.56 −0.047 0.008 0.728 −0.151 −2.231 2.090 1.069 −0.801 0.482

<0.0001 0.96 0.87 0.46 0.88 0.02 0.04 0.28 0.42 0.63

Abbreviations: DEE: energy expenditure; BMI: body mass index; WC/HC ratio: waist circumference/hip circumference ratio; HDL: high density lipoprotein cholesterol; LDL: low density lipoprotein cholesterol; AHI: apnea–hypopnea index; ODI: oxygen desaturation index; SaO2: oxygen saturation.

tion or bone deterioration, both implicated in the age-related bone remodeling. In our sample, neither diabetes nor dyslipidemia affects BMD, the only anthropometric factor being BMI having a slight impact on T-score at the femoral level. This finding may be explained by the positive and protective effects of obesity on BMD [28–30] due to the gravitational loading and mechanical stimulation of bone tissue by increased weight. This effect is, however, modulated by the presence of a weight threshold [31,32], a BMI < 30 having a neutral or protective effect, and a BMI > 30 lowering BMD

Table 7 Stepwise regression analysis between T-scores at lumbar and femoral levels and explicative polygraphic, anthropometric, and biological data. T-score at the lumbar level: Independent variable in the model (R = 0.303; R2 = 0.092; p < 0.0001) Coefficient

SE

Standard coefficient

DEE 0.00022 0.00003 0.303 Variables not in the model: gender, age, BMI, WC/HC ratio, ODI, AHI, mean SaO2, min SaO2, time SaO2 < 90%, glycemia, HDL, and LDL. T-score at the femoral level: Independent variables in the model ((R = 0.424; R2 = 0.180; p < 0.0001) Coefficient

SE

Standard coefficient

DEE 0.00016 0.00003 0.299 BMI 0.047 0.017 0.162 Variables not in the model: gender, age, WC/HC ratio, ODI, AHI, mean SaO2, min SaO2, time SaO2 < 90%, glycemia, HDL, and LDL. Abbreviations: DEE: energy expenditure; BMI: body mass index; WC/HC ratio: waist circumference/hip circumference ratio; ODI: oxygen desaturation index; AHI: apnea– hypopnea index; SaO2: oxygen saturation; HDL: high density lipoprotein cholesterol; LDL: low density lipoprotein cholesterol.

E. Sforza et al. / Sleep Medicine 22 (2016) 25–32

status. In our subjects, despite a positive association between anthropometric measurements of obesity and BMD, only BMI contributes to OS/OP occurrence with, however, a slight effect likely related to the lack of obese subjects in our sample. Controversial data are present in literature concerning the rise of OS in patients with OSA. Given the high prevalence of OSA in the older population, it could be suggested [33,34] that apnea recurrence and falls in oxygen saturation can lead to OP/OS by several mechanisms, such as chronic increased sympathetic tonus, systemic inflammatory processes [35], oxidative stress [36,37], vitamin D deficiency, and glycemic and lipidic dysfunction [15]. Moreover, intermittent hypoxia as occurring in OSA stimulates bone resorption by increasing osteoclast formation and activity [14,38]. In contrast to this hypothesis, we found a slight association between BMD and indices of OSA, with AHI and ODI slightly contributing to the OS occurrence. In addition, subjects with OS/OP were those having low AHI and ODI values. These findings are in line with previous data published by our group [12] and a recent meta-analysis [39] showing the lack of effect of AHI and ODI on BMD. To explain the protective role of OSA on bone metabolism we have to remember that the recurrent hypoxia related to apneas is likely to affect bone cellular function by increasing biochemical markers, ie, mesenchymal cells that enhance physiological regeneration of tissue via differentiation of mesenchymal cells into osteoblasts and reduction of bone resorption [13]. Although speculative, we can suggest that in the elderly the lack of association between OSA and BMD alteration is related to the occurrence of protective mechanisms against intermittent hypoxia in elderly [40]. The most interesting finding of the current study is the strong role of physical activity on OS-risk subjects with OP and/or OS being those with lower DEE. These results are in line with previous studies on large cohorts [41,42] showing the key role of physical activity on bone density in young and older subjects as well as in women and men, again with a small contribution of BMI. We know that the rate of bone loss varies across the aging period via multiple complex mechanisms including genetic factors [43], nutrition [44], and physical activity habits [45]. Physical activity helps maintain mobility, physical functioning, BMD, and muscle strength balance, and therefore may help prevent falls and fractures among elderly. The positive effect of physical activity has been studied in a cohort of 3262 men [46] in which the risk of fractures was followed for 21 years and showed a significant reduction of fracture risk compared to subjects not participating in the project of intense physical activity. In a meta-analysis of 13 prospective cohort studies [47], the authors found a significant reduction of fracture risk of 45% in men and 38% in women, stressing the positive effects of moderate-to-vigorous physical activity on bone stability. The mechanisms by which physical activity contributes to bone remodeling and OS prevention are not completely understood, but a possible mechanism includes the stimulation of bone narrow mesenchymal stem cells and osteoblasts and reduction of osteoclast function. 4.1. Strengths and limitations Methodological strengths of this study were the large number of participants with an actual gender balance, the inclusion of blood samples, and use of DXA methodology in the objective evaluation of the presence of OS/OP, in addition to the assessment of sleep respiratory disorders and severity. However, some limitations of the study need to be mentioned. Firstly, the cross-sectional nature of the study is unable to assess a causative relationship between the examined factors and OS risk. Secondly, we examined a homogenous group of elderly subjects with similar age, ie, 68 years, which can be defined as “young” elderly, precluding the application of these results in more senior subjects who structurally have a higher OS prevalence. Thirdly, we examined a reduced number of subjects com-

31

pared to the original population. This was related to the fact that only a smaller sample accepts to fill-in the physical activity. However, the similarity of results in the original sample [12] and in this smaller group allows exclusion of a selection bias. Finally, subjects were examined by polygraphy and not polysomnography, the latter allowing analysis of sleep fragmentation. However, we introduce in our analysis the respiratory autonomic activations that are a surrogate marker of sleep disruption and we did not find any association between BMD and autonomic sleep fragmentation. 5. Conclusion While it is well established that aging is a major risk factor in the progression of various metabolic disorders, OS, and obstructive sleep apnea, little is known about the links between these factors and aging itself. In the present study, we confirm the lack of impact of metabolic disorders, obesity, and OSA on BMD, with the level of daily physical activity playing a key role on the prevention of OS. The strong contribution of physical activity in OS risk might support recent therapeutic models of association of nutrition, lifestyle, and physical activity strategies to counteract both the agerelated risk of OS and the severity of OSA. Large-scale longitudinal epidemiological studies with specific BMD assessment, physical activity, and quality of life as endpoints are needed to confirm the key role of physical activity on OS risk in OSA populations. Conflict of interest All authors declare no financial or other potential conflicts of interest. The ICMJE Uniform Disclosure Form for Potential Conflicts of Interest associated with this article can be viewed by clicking on the following link: http://dx.doi.org/10.1016/j.sleep.2016.04.010. Acknowledgments The authors would like to thank all the subjects included in the present study, and Mr. Olivier Grataloup and Dr. Stéphane Chomienne (CHU Saint-Etienne, France) for their expert help in data acquisition. References [1] Consensus development conference: prophylaxis and treatment of osteoporosis. Am J Med 1991;90:107–10. [2] Varenna M, Manara M, Galli L, et al. The association between osteoporosis and hypertension: the role of low dairy intake. Calcif Tissue Int 2013;92:217–27. [3] Dominiquez LJ, Muratore M, Quarta E, et al. Osteoporosis and diabetes. Reumatismo 2004;56:235–41. [4] Migliaccio S, Greco EA, Fornari R, et al. Is obesity in women protective against osteoporosis? Diabetes Metab Syndr Obes 2011;4:273–82. [5] Johansson J, Nordström A, Nordström P. Objectively measured physical activity is associated with parameters of bone in 70-year-old men and women. Bone 2015;81:72–9. [6] Chen YL, Weng SF, Shen YC, et al. Obstructive sleep apnea and risk of osteoporosis: a population-based cohort study in Taiwan. J Clin Endocrinol Metab 2014;99:2441–7. [7] Montagnani A, Gonnelli S, Alessandri M, et al. Osteoporosis and risk of fractures in patients with diabetes: an update. Aging Clin Exp Res 2011;23:84–90. [8] Kanazawa I, Tugimoto T. Bone disease caused by impaired glucose and lipid metabolism. Clin Calcium 2013;23:1605–11. [9] El Maghraoui A, Rezqi A, El Mrahu S, et al. Osteoporosis, vertebral fracture and metabolic syndrome. BMC Endocr Disord 2014;14:93. [10] Kim BJ, Ahs SH, Bae S, et al. Association between metabolic syndrome and bone loss at various skeletal sites in postmenopausal women: a 3-year retrospective longitudinal study. Osteoporos Int 2013;24:2243–52. [11] Xue P, Gao P, Ly Y. The association between metabolic syndrome and bone mineral density: a meta-analysis. Endocrine 2012;42:546–54. [12] Sforza E, Thomas T, Barthélémy JC, et al. Obstructive sleep apnea is associated with preserved bone mineral density in healthy elderly subjects. Sleep 2013;36:1509–15.

32

E. Sforza et al. / Sleep Medicine 22 (2016) 25–32

[13] Carreras A, Rojas M, Tsapikouni T, et al. Obstructive apneas induce early activation of mesenchymal stem cells and enhancement of endothelial wound healing. Respir Res 2010;11:91. [14] Arnett TR, Gibbons DC, Utting JC, et al. Hypoxia is a major stimulator of osteoclast formation and bone resorption. J Cell Physiol 2003;196:2–8. [15] Chakhtoura M, Nasrallah M, Chami H. Bone loss in obesity and obstructive sleep apnea: a review of literature. J Clin Sleep Med 2015;11:575–80. [16] Swanson CM, Shea SA, Stone KL, et al. Obstructive sleep apnea and metabolic bone disease: insights into the relationship between bone and sleep. J Bone Miner Res 2015;30:199–211. [17] Sergi G, Coin A, Marin S, et al. Body composition and resting energy expenditure in elderly male patients with chronic obstructive pulmonary disease. Respir Med 2006;100:1918–24. [18] Tsai YW, Yun LL, Pai CC, et al. Association of bone mineral density and obstructive sleep apnea in chronic obstructive pulmonary disease. Int J Chron Obstruct Pulmon Dis 2015;10:231–7. [19] McNicholas WT. Chronic obstructive pulmonary disease and obstructive sleep apnea: overlaps in pathophysiology, systemic inflammation and cardiovascular diseases. Am J Respir Crit Care Med 2009;180:692–700. [20] Barthélémy JC, Pichot V, Dauphinot V, et al. Autonomic nervous system activity and decline as prognostic indicators of cardiovascular and cerebrovascular events. The PROOF Study. Neuroepidemiology 2007;29:18–28. [21] Salamone LM, Fuerst T, Visser M, et al. Measurement of fat mass using DEXA: a validation study in elderly adults. J Appl Physiol 2000;89:345–52. [22] Orimo H, Hayashi Y, Fukunaga M, et al. Diagnostic criteria for primary osteoporosis: year 2000 revision. J Bone Miner Metab 2001;19:331–7. [23] Garet M, Barthélémy JC, Degache F, et al. A questionnaire-based assessment of daily activity in heart failure. Eur J Heart Fail 2004;6:577–84. [24] American Academy of Sleep Medicine Task Force. Sleep-related breathing disorders in adults: recommendations for syndrome definition and measurement techniques in clinical research. Sleep 1999;22:667–89. [25] Pavlova MK, Duffy JF, Shea SA. Polysomnographic respiratory abnormalities in asymptomatic individuals. Sleep 2008;31:241–8. [26] Mora S, Gilsanz V. Establishment of peak bone mass. Endocrinol Metab Clin North Am 2003;32:39–63. [27] Kanis JA, McCloskey EV, Johansson H, et al. European Guidance for the diagnosis and management of osteoporosis in postmenopausal women. Osteoporos Int 2013;24:23–57. [28] Albala C, Yanez M, Devoto E, et al. Obesity as a protective factor for postmenopausal osteoporosis. Int J Obes Relat Metab Disord 1996;20:1027– 32. [29] Felson DT, Zhang Y, Hannan MT, et al. Effect of weight and body mass on bone mineral density in men and women. J Bone Miner Res 1993;8:567–73. [30] Richards JB, Valdes AM, Burlink K, et al. Serum adiponectin and bone mineral density in men and women. J Clin Endocrinol Metab 2007;92:1517–23.

[31] Liu PY, Illich JZ, Brummel-Smith K, et al. New insights into fat, muscle and bone relationships in women: determining the threshold at which body fat assumes a negative relationship with bone mineral density. Int J Prev Med 2014;5:1452– 63. [32] Greco EA, Fornari R, Rossi F, et al. Is obesity protective for osteoporosis? Evaluation of bone mineral density in individual with high body mass index. Int J Clin Pract 2010;64:817–20. [33] Uzkeser H, Yildirim K, Aktan B, et al. Bone mineral density in patients with obstructive sleep apnea. Sleep Breath 2013;17:339–42. [34] Tomiyama H, Okzaki R, Inoue D, et al. Link between obstructive sleep apnea and increased bone resorption in men. Osteoporos Int 2008;19:1185–92. [35] Hatipoglu H, Rubistein I. Inflammation and obstructive sleep apnea syndrome pathogenesis: a working hypothesis. Respiration 2003;70:665–71. [36] Basu S, Michaelsson K, Olofssons H, et al. Association between oxidative stress and bone mineral density. Biochem Biophys Res Commun 2001;288:275–9. [37] Ostman B, Michaëlsson K, Helmersson J, et al. Oxidative stress and bone mineral density in elderly men: antioxidant activity of alpha-tocopherol. Free Radic Biol Med 2009;47:668–73. [38] Utting J, Robins S, Brandao-Burch A, et al. Hypoxia inhibits the grown, differentiation and bone-forming capacity of rat osteoblast. Exp Cell Res 2006;312:1693–702. [39] Upala S, Sanguankeo A, Congrete S. Obstructive sleep apnea is not associated with an increased risk of osteoporosis: a systematic review and meta-analysis. J Clin Sleep Med 2015;11:1069–70. [40] Lavie P, Lavie L. Unexpected survival advantages in elderly people with moderate sleep apnoea. J Sleep Res 2009;18:397–403. [41] Langsetmo L, Hitchcock CL, Kingwell EJ, et al. Physical activity, body mass index and bone mineral density-associations in a prospective population-based cohort of women and men: the Canadian Multicentre Osteoporosis Study (CaMos). Bone 2012;50:401–8. [42] Chastin SF, Mandrichenko O, Helbostadt JL, et al. Associations between objectively-measured sedentary behaviour and physical activity with bone mineral density in adults and older adults, the NHANES study. Bone 2014;64:254–62. [43] Moayyeri A, Hammond CJ, Hart DJ, et al. Effects of age on genetic influences on bone loss over 17 years in women: the healthy ageing twin study (HATS). J Bone Miner Res 2012;27:2170–8. [44] Ishimi Y. Osteoporosis and lifestyle. J Nutr Sci Vitaminol 2015;61:S139–41. [45] Moayyeri A. The association between physical activity and osteoporotic fracture: a review of evidence and implication for future research. Ann Epidemiol 2008;18:827–35. [46] Kujala UM, Kaprio J, Kannus P, et al. Physical activity and osteoporotic hip fracture risk in men. Arch Intern Med 2000;160:705–8. [47] Yuan Y, Chen X, Zhang L, et al. The roles of exercise in bone remodelling and in prevention and treatment of osteoporosis. Prog Biophys Mol Biol 2015;9.