7
992 Journals of Gerontology: MEDICAL SCIENCES Cite journal as: J Gerontol A Biol Sci Med Sci. 2013 August;68(8):992–998 doi:10.1093/gerona/glt008 Advance Access publication March 4, 2013 © The Author 2013. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: [email protected]. Association Between Serum 25-Hydroxyvitamin D Levels, Bone Geometry, and Bone Mineral Density in Healthy Older Adults Marco Mosele, 1 Alessandra Coin, 1 Enzo Manzato, 1 Silvia Sarti, 1 Linda Berton, 1 Francesco Bolzetta, 1 Alessandra Imoscopi, 1 Giulia Rinaldi, 1 Egle Perissinotto, 2 and Giuseppe Sergi 1 1 Department of Medicine, DIMED, Geriatrics Section, University of Padova, Italy. 2 Department of Cardiac, Thoracic and Vascular Sciences, University of Padova, Italy. Address correspondence to Giuseppe Sergi, Clinica Geriatrica - Ospedale Giustinianeo (2° piano), via Giustiniani 2, 35128, Padova, Italy. Email: [email protected] Purpose. The association between serum 25-hydroxyvitamin D values and cortical/trabecular bone parameters in older adults has been incompletely explored. This study was designed to investigate the relationship between serum 25-hydroxyvitamin D levels and bone parameters for the tibia and radius using peripheral quantitative computed tomog- raphy in free-living healthy older adults. Methods. The study involved 134 older adults attending a twice-weekly low-intensity fitness program. In addition to clinical history and serum parameters, we assessed fat-free mass using dual-energy X-ray absorptiometry, total bone and cortical bone cross-sectional areas, and trabecular and cortical bone mineral density for the tibia and radius by peripheral quantitative computed tomography. Results. After applying multivariate linear regression models, adjusting for sex, age, body mass index, fat mass and fat-free mass, and creatinine, the association between 25-hydroxyvitamin D and bone parameters was significant for total bone and cortical bone cross-sectional areas in the radius (partial R 2  = 0.05 and 0.09, respectively) and for trabecular bone mineral density and cortical bone cross-sectional area in the tibia (partial R 2  = 0.11 and 0.02, respectively). Conclusion. These findings support the idea that serum 25-hydroxyvitamin D levels and bone parameters are linked in older adults. Longitudinal studies are needed to establish whether vitamin D levels over time are associated with changes in these parameters. Key Words: 25-OHD—Cortical bone area—Bone density—Older adults—pQCT. Received March 16, 2012; Accepted January 8, 2013 Decision Editor: Stephen Kritchevsky, PhD W ITH aging, frailty fractures have become increasingly common worldwide. They now account for a large proportion of disability, death, and medical costs (1). Age- related changes in bone quality and geometry play an impor- tant part in causing frailty fractures. The bone loss related to aging is accompanied by a loss of trabeculae and fewer trabecular connections (2), more endocortical resorption and less periostial apposition (3,4). This is a key process in the onset of bone fragility in older age, especially in women (5). Dual-energy X-ray absorptiometry (DXA) is now used routinely to assess bone mineral density (BMD) at different bone sites and to predict the risk of fracture. DXA is insen- sitive to changes in the quantitative and geometric distribu- tion of trabecular and cortical bone tissues, however (6–8). Peripheral quantitative computerized tomography (pQCT) is a more reliable method for measuring volumetric BMD (mass/volume) because it distinguishes between cortical and trabecular bone components (5), which are known to respond differently to disease, loading and aging. pQCT also has the advantage of detecting bone geometry in distal segments of the limbs (9,10). Among the factors influencing frailty fractures, serum 25-hydroxyvitamin D (25-OHD) level has received more attention recently (11,12). One of the most important functions of serum 25-OHD is to maintain skeletal calcium balance by promoting calcium absorption in the intestines; it promotes bone resorption by increasing the number of osteoclasts, it maintains the calcium and phosphate levels needed for bone formation, and it enables the parathyroid hormone to function properly in maintaining serum calcium levels. Serum 25-OHD is important to bone and global health and in older adults because low serum levels of 25-OHD are common (13), predict reduced BMD (14) and a higher frailty at Florida Atlantic University on November 15, 2014 http://biomedgerontology.oxfordjournals.org/ Downloaded from

Association Between Serum 25-Hydroxyvitamin D Levels, Bone Geometry, and Bone Mineral Density in Healthy Older Adults

  • Upload
    g

  • View
    213

  • Download
    1

Embed Size (px)

Citation preview

Page 1: Association Between Serum 25-Hydroxyvitamin D Levels, Bone Geometry, and Bone Mineral Density in Healthy Older Adults

992

Journals of Gerontology: MEDICAL SCIENCESCite journal as: J Gerontol A Biol Sci Med Sci. 2013 August;68(8):992–998doi:10.1093/gerona/glt008 Advance Access publication March 4, 2013

© The Author 2013. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: [email protected].

Association Between Serum 25-Hydroxyvitamin D Levels, Bone Geometry, and Bone Mineral Density

in Healthy Older Adults

Marco Mosele,1 Alessandra Coin,1 Enzo Manzato,1 Silvia Sarti,1 Linda Berton,1 Francesco Bolzetta,1 Alessandra Imoscopi,1 Giulia Rinaldi,1 Egle Perissinotto,2 and Giuseppe Sergi1

1Department of Medicine, DIMED, Geriatrics Section, University of Padova, Italy.2Department of Cardiac, Thoracic and Vascular Sciences, University of Padova, Italy.

Address correspondence to Giuseppe Sergi, Clinica Geriatrica - Ospedale Giustinianeo (2° piano), via Giustiniani 2, 35128, Padova, Italy. Email: [email protected]

Purpose. The association between serum 25-hydroxyvitamin D values and cortical/trabecular bone parameters in older adults has been incompletely explored. This study was designed to investigate the relationship between serum 25-hydroxyvitamin D levels and bone parameters for the tibia and radius using peripheral quantitative computed tomog-raphy in free-living healthy older adults.

Methods. The study involved 134 older adults attending a twice-weekly low-intensity fitness program. In addition to clinical history and serum parameters, we assessed fat-free mass using dual-energy X-ray absorptiometry, total bone and cortical bone cross-sectional areas, and trabecular and cortical bone mineral density for the tibia and radius by peripheral quantitative computed tomography.

Results. After applying multivariate linear regression models, adjusting for sex, age, body mass index, fat mass and fat-free mass, and creatinine, the association between 25-hydroxyvitamin D and bone parameters was significant for total bone and cortical bone cross-sectional areas in the radius (partial R2 = 0.05 and 0.09, respectively) and for trabecular bone mineral density and cortical bone cross-sectional area in the tibia (partial R2 = 0.11 and 0.02, respectively).

Conclusion. These findings support the idea that serum 25-hydroxyvitamin D levels and bone parameters are linked in older adults. Longitudinal studies are needed to establish whether vitamin D levels over time are associated with changes in these parameters.

Key Words: 25-OHD—Cortical bone area—Bone density—Older adults—pQCT.

Received March 16, 2012; Accepted January 8, 2013

Decision Editor: Stephen Kritchevsky, PhD 

WITH aging, frailty fractures have become increasingly common worldwide. They now account for a large

proportion of disability, death, and medical costs (1). Age-related changes in bone quality and geometry play an impor-tant part in causing frailty fractures. The bone loss related to aging is accompanied by a loss of trabeculae and fewer trabecular connections (2), more endocortical resorption and less periostial apposition (3,4). This is a key process in the onset of bone fragility in older age, especially in women (5).

Dual-energy X-ray absorptiometry (DXA) is now used routinely to assess bone mineral density (BMD) at different bone sites and to predict the risk of fracture. DXA is insen-sitive to changes in the quantitative and geometric distribu-tion of trabecular and cortical bone tissues, however (6–8). Peripheral quantitative computerized tomography (pQCT) is a more reliable method for measuring volumetric BMD (mass/volume) because it distinguishes between cortical

and trabecular bone components (5), which are known to respond differently to disease, loading and aging. pQCT also has the advantage of detecting bone geometry in distal segments of the limbs (9,10).

Among the factors influencing frailty fractures, serum 25-hydroxyvitamin D (25-OHD) level has received more attention recently (11,12). One of the most important functions of serum 25-OHD is to maintain skeletal calcium balance by promoting calcium absorption in the intestines; it promotes bone resorption by increasing the number of osteoclasts, it maintains the calcium and phosphate levels needed for bone formation, and it enables the parathyroid hormone to function properly in maintaining serum calcium levels.

Serum 25-OHD is important to bone and global health and in older adults because low serum levels of 25-OHD are common (13), predict reduced BMD (14) and a higher frailty

at Florida Atlantic U

niversity on Novem

ber 15, 2014http://biom

edgerontology.oxfordjournals.org/D

ownloaded from

Page 2: Association Between Serum 25-Hydroxyvitamin D Levels, Bone Geometry, and Bone Mineral Density in Healthy Older Adults

25-HyDRoxyvItAMIN D AND BoNE PARAMEtERS 993

fracture risk (15), and are associated with an increased risk of disability (16) and cognitive decline (17,18).

Long-standing underdiagnosed serum 25-OHD deficien-cies may already exist in earlier adulthood, exacerbating the negative effects of aging on bone.

Although there is a shortage of studies concerning the correlation between pQCT parameters and serum levels of 25-OHD, this association is well established from DXA studies. In a large cohort of more than 13,000 subjects, Heike and colleagues (19) found a significant positive asso-ciation between 25-OHD and bone mineral density, with no gender-related differences. Focusing on older adults in particular, Von Muhlen and colleagues conducted a multiple linear regression analysis and found a positive association between serum 25-OHD levels and BMD in a large group of old women (20).

Hence, our interest in identifying any relationship between serum 25-OHD levels and bone parameters (meas-ured using a reference method such as pQCT) in older adults. So far, only two studies have used pQCT to show an association between serum levels of vitamin D and bone parameters in older adults (2,21), but the information they provide on this relationship is incomplete. In fact, the InCHIANTI study found a relationship between 25-OHD and the cortical BMD of the tibia, but only in older women (2). Barbour and colleagues (21) found a positive associa-tion between 25-OHD levels and bone parameters concern-ing the radius and tibia in older Caucasian men, but they provided no data on older women.

Based on the hypothesis that higher serum 25-OHD lev-els would help to preserve bone structure in older adults, we used pQCT to investigate the relationship between serum 25-OHD levels and bone parameters for the tibia and radius in free-living, healthy older men and women.

Materials and Methods

SubjectsThis cross-sectional cohort study was conducted at the

Geriatrics Department of Padova University. Individuals older than 65 years who attended a twice-weekly low-inten-sity fitness program of aerobic and/or resistance exercises at public gyms in Padova were recruited on a voluntary basis. Trained medical personnel established that participants were healthy according to their clinical history, clinical examina-tion, and biochemical test results.

The level of physical activity was explored using the NASA/Johnson Space Center Physical Activity Rating (PA-R) questionnaire, which contains seven categories of questions on a person’s regular exercising habits, scored according to the time spent on the physical activity and the type of activity (22). According to this scale, all participants were in class B, meaning they engaged regularly in a mod-erate physical activity.

Individuals with severe cardiovascular or pulmonary dis-eases, renal impairment (serum creatinine > 130 µmol/L), uncontrolled metabolic disease (diabetes, anemia, or thy-roid disease), electrolyte abnormalities, cancer, or inflam-matory conditions were excluded. Any use of drugs (corticosteroids, hormones, etc.) that might interfere with bone and body composition was also considered a reason for exclusion. A  hundred and thirty-four older adults (86 women and 48 men) met the inclusion/exclusion criteria and were enrolled in this study.

Using the variance estimates for each bone characteristic obtained by pQCT in earlier studies conducted at our clinic, we performed a power analysis to assess the adequacy of the sample size in our statistical analysis. The sample size of 134 enabled us to estimate the bone characteristics to be within 4%, based on two-sided 95% confidence inter-vals. It also had a power of 82% in detecting an incremen-tal R-squared value of 0.05 attributed to one independent variable in a multiple regression model, after controlling for six independent variables with a total R-squared of at least 0.15.

The study was designed in accordance with the Helsinki Declaration. All participants were fully informed about the study’s design and purpose, procedures, and risks, and they gave their informed consent.

Study DesignAll patients underwent the following tests:Blood assays: Fasting venous blood samples were col-

lected for all patients to measure 25-OHD, parathyroid hormone (PTH), interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-α), and creatinine. 25-OHD was tested by RIA (Incestar Corp, Stillwater, MN), PTH by IRMA (Allegro TM Intact, Nicholas Institute, S. Juan Capistrano, CA), and IL-6 and TNF-α by IMMUNOLITE one (Medical System S.p.A., Genova, Italy), a solid-phase, enzyme-labeled, chemiluminescent sequential immunometric assay.

Anthropometric measurements: Body weight was meas-ured to the nearest 0.1 kg and height to the nearest 0.1 cm using a standard balance and stadiometer (Seca, Germany), with participants wearing light clothing and no shoes. Body mass index (BMI) was calculated as their weight in kilo-grams divided by their height in square meters.

Bone measurements: pQCT was performed on the dominant forearm and right tibia using the Norland/Stratec XCT-3000 scanner (Stratec Medizintechnik GmbH, Pforzheim, Germany), following a standardized patient positioning and scanning protocol. Forearm length was measured from the olecranon to the ulna styloid process, and tibia length from the medial malleolus to the medial condyle. A pQCT scout view was obtained to establish an anatomical reference line bisecting the medial edge of the end of the distal radius or tibia. Starting from this reference line, scans were obtained at sites 4% and 66% along the

at Florida Atlantic U

niversity on Novem

ber 15, 2014http://biom

edgerontology.oxfordjournals.org/D

ownloaded from

Page 3: Association Between Serum 25-Hydroxyvitamin D Levels, Bone Geometry, and Bone Mineral Density in Healthy Older Adults

994 MoSELE Et AL.

length of the shaft for the radius and at sites 4%, 14%, 38%, and 66% along the length of the tibia. The following bone parameters were derived from the pQCT images:

Trabecular volumetric BMD (BMDt), that is, the mean den-sity of the trabecular bone area detected at the 4% site on the radius and tibia.

Cortical volumetric BMD (BMDc), that is, the mean den-sity of the cortical bone measured at the 66% and 38% sites on the radius and tibia, respectively.

Total bone cross-sectional area (CSA), that is, the area within the circumference delimiting all cortical bone tis-sues with a density higher than 180 mg/cm3, measured at the 4% site on the radius and tibia (this is a measure of bone size).

Cortical bone cross-sectional area (CSAc), assessed as the cross-sectional area of the voxels with a density higher than 710 mg/cm3, measured at the 38% site on the tibia and at the 66% point on the radius.

Body composition: We measured fat-free mass (FFM) and fat mass (FM) with Dual Energy X-ray Absorptiometry (DXA) using fan-beam technology (Hologic QDR 4500 W, Inc.). The DXA method has a good reproducibility in deter-mining soft tissue composition, achieving consistent results in older adults too (23).

Illness: The Cumulative Illness Rating scale (CIRS) assesses clinical status, comorbidities, and illness sever-ity levels, classifying comorbidities among 13 organ sys-tems and grading each condition from 1 (no problem) to 5 (severely incapacitating or life-threatening conditions). The comorbidity index (CI) is given by the number of condi-tions graded for severity as ≥3 (24). The Illness Severity Index (SI) is given by the average of all CIRS items.

Statistical AnalysisStatistical analyses were performed using the SPSS for

Windows, rel. 17.0 (SPSS Inc. Chicago). For the data analy-sis, serum 25-OHD was treated as both a quantitative and a categorical variable. Four levels of serum 25-OHD values were classified as follows: ≤25; 25–50; 50–75; >75 nmol/L. These cut-offs have been amply used in the literature and they roughly corresponded to quartiles in our sample popu-lation (Q1 = 28 nmol/L, Q2 = 52 nmol/L, Q3 = 75 nmol/L). To choose the best data analysis approach, analysis of variance was used to test the interaction between sex and 25-OHD levels. Because the interaction term was never sta-tistically significant, all statistical methods were applied to the whole sample and by 25-OHD level, whereas sex was included as a covariate in the multivariate analyses. Only the results regarding the bone characteristics measured by pQCT at the different sites were performed on the whole sample and by gender.

In the descriptive analyses, the prevalence of specific conditions was expressed as a percentage, and differences

among groups were assessed using the χ2 test. Normally distributed variables were reported as means ± standard deviations. The median and the first and third quartiles were calculated for ordinal variables (scores). Differences among mean values were assessed using Student’s unpaired t test, whereas differences among median values were ana-lyzed using the unpaired Mann–Whitney test. The means of the quantitative variables for the different 25-OHD lev-els were compared by analysis of variance for unbalanced groups. The same procedure was used to obtain sex- and age-adjusted p values.

Pearson’s product-moment correlation coefficient (r) was applied to measure simple linear associations between pQCT parameters and serum 25-OHD values, age, BMI, FFM, FM, PTH, creatinine, albumin, IL6 and TNFα, comor-bidity, severity of diseases, and physical activity score.

Multiple linear regression models were used with a step-wise forward procedure to assess the independent asso-ciation among the bone parameters at each site and the 25-OHD values, including the factors found significantly associated with the previous simple linear analysis as covar-iates (ie, age, BMI, FM, FFM, serum creatinine, and sex). Partial and total R2 coefficients were estimated, and R2 was used as a fit criterion. A variance inflation factor was com-puted for each independent variable in the model to test for multicollinearity: Values more than 2 were used to indicate a multicollinearity problem in the model.

In all analyses, a p value less than .05 was considered statistically significant.

ResultsTable 1 shows the descriptive statistics for the anthropo-

metric measurements, body composition, and serum param-eters, for the whole sample and stratified by serum 25-OHD levels.

Participants averaged approximately 73  years of age, with a BMI around 26 and a mean albumin level of 44 g/L. Hypovitaminosis D (<75 nmol/L) was identified in about 75% of the sample.

Basically, there were no significant differences among the groups divided by 25-OHD level in terms of their gen-eral and clinical characteristics, except for age (p = .03) and height (sex- and age-adjusted p = .004).

The findings regarding comorbidities (CI) and severity of diseases (SI), as expressed by the CIRS, consistently confirmed that our study population was in substantially good health (the first and the third quartiles ranged from 0 to 2 for CI and from 1.10 to 1.35 for SI), with no significant differences among the 25-OHD groups.

The NASA scale also confirmed the sample’s homoge-neity in terms of the participants’ physical activity scores, which ranged between 3 and 4.

Figure  1 shows the mean bone features identified by pQCT by sex. As expected, statistically significant

at Florida Atlantic U

niversity on Novem

ber 15, 2014http://biom

edgerontology.oxfordjournals.org/D

ownloaded from

Page 4: Association Between Serum 25-Hydroxyvitamin D Levels, Bone Geometry, and Bone Mineral Density in Healthy Older Adults

25-HyDRoxyvItAMIN D AND BoNE PARAMEtERS 995

differences emerged among the genders for all bone char-acteristics, with significantly higher mean values in men (p = .002 and .02 for BMDc at the radius and tibia, respec-tively; p < .0001 for the other comparisons at both sites).

Table  2 shows the pQCT results for the bone charac-teristics in our whole sample and by 25-OHD level. After adjusting for sex and age, there were statistically significant

differences for CSA (p = .0004) and CSAc (p < .0001) at the radius, and BMDt (p = .04), CSA (p = .02) and CSAc (p = .01) at the tibia.

Table  3 shows the simple correlations between the bone characteristics and a number of potential covariates. 25-OHD levels were linearly associated with all bone characteristics at the radius and tibia. Other covariates

1072

152

360

62

1097

195

487

93

1037

182

1102

229

1079

211

1402

331

0

200

400

600

800

1000

1200

1400

1600

BMDcmg/cm3

BMDtmg/cm3

CSAmm2

CSAcmm2

BMDcmg/cm3

BMDtmg/cm3

CSAmm2

CSAcmm2

BMDcmg/cm3

BMDtmg/cm3

CSAmm2

CSAcmm2

BMDcmg/cm3

BMDtmg/cm3

CSAmm2

CSAcmm2

WOMEN MEN WOMEN MEN

Radius Tibia

Figure 1. Mean values of bone characteristics by site and gender (vertical lines indicate the upper limit of the 95% confidence interval).

Table 1. Mean Values of Anthropometry, Body Composition, Biochemical Parameters, Comorbidity Indices (CIRS), and Physical Activity Level (NASA) in the Sample as a Whole and by 25-OHD Levels

VariablesTotal Sample,

N = 134

25(OH)D Levels (nmol/L)

Unadjusted p*Sex–Age,

Adjusted p*

≤25, N = 27, Mean ± SD

25–50, N = 36, Mean ± SD

50–75, N = 38, Mean ± SD

>75, N = 33, Mean ± SD

25-OHD (nmol/L) 53.2 ± 28.1 19.6 ± 5.0 36.4 ± 7.5 63.5 ± 7.5 93.1 ± 14.3 (77–152) < .001 —Sex f (%) 64.2 70.4 66.7 52.6 69.7 n.sAge (years) 72.9 ± 5.2 74.5 ± 5.3 73.7 ± 6.1 72.8 ± 4.3 70.9 ± 4.6 .03Height (cm) 161.4 ± 9.0 156.8 ± 7.4 161.6 ± 9.0 164.6 ± 8.7 161.4 ± 9.2 .006 .004Weight (kg) 67.7 ± 11.1 64.6 ± 11.5 67.9 ± 12.7 69.2 ± 8.2 68.3 ± 11.9 n.s n.sBMI (kg/m2) 25.9 ± 3.2 26.2 ± 3.9 25.9 ± 3.2 25.6 ± 2.6 26.1 ± 3.4 n.s n.sAlbumin (g/L) 44.3 ± 2.3 44.2 ± 2.4 43.6 ± 2.9 44.7 ± 2.0 44.4 ± 1.9 n.s n.sFat mass (kg) 20.7 ± 6.1 21.1 ± 7.1 20.6 ± 6.1 20.5 ± 6.2 20.6 ± 5.4 n.s n.sFat-free mass (kg) 46.3 ± 9.1 42.6 ± 7.4 46.3 ± 9.3 48.6 ± 8.6 46.8 ± 10.1 .07 n.s

Creatinine (µmol/L) 78.5 ± 17.4 77.7.1 ± 20.1 81.8 ± 19.7 78.5 ± 12.9 75.3 ± 17.2 n.s n.s

IL6 (ng/L) 3.2 ± 2.2 3.9 ± 3.3 3.0 ± 1.6 3.4 ± 2.6 2.7 ± 1.0 n.s n.sTNF (ng/L) 10.8 ± 5.0 10.9 ± 7.2 10.8 ± 4.1 10.0 ± 2.9 11.6 ± 5.6 n.s n.sPTH (ng/L) 88.2 ± 31.3 92.1 ± 32.8 86.5 ± 33.4 90.1 ± 26.9 88.2 ± 31.3] n.s n.sComorbidity index,

median (Q1–Q3)1 (0–2) 1 (1,2) 1 (0–2) 1 (0–2) 1 (0–2) — —

Illness severity index median (Q1–Q3)

1.21 (1.11–1.35) 1.21 (1.14–1.42) 1.28 (1.14–1.36) 1.14 (1.04–1.32) 1.21 (1.07–1.29) —

NASA, median (Q1–Q3) 3(3) 3(3) 3(3) 3(3) 3(3) ns

Notes: BMI = body mass index; PTH = parathyroid hormone; IL-6 = interleukin-6; TNF-α = tumor necrosis factor-alpha; 25-OHD = 25-Hydroxyvitamin D; NASA = NASA/Johnson Space Center Physical Activity Rating (PA-R) questionnaire.

*p values obtained from a general linear model (F test) with 25-OHD levels included in the model as a categorical variable.

at Florida Atlantic U

niversity on Novem

ber 15, 2014http://biom

edgerontology.oxfordjournals.org/D

ownloaded from

Page 5: Association Between Serum 25-Hydroxyvitamin D Levels, Bone Geometry, and Bone Mineral Density in Healthy Older Adults

996 MoSELE Et AL.

significantly associated with at least one bone characteristic were BMI, FFM, FM, and creatinine. As expected, given the homogeneity of our sample population, the comorbid-ity, disease severity, and physical activity indexes did not correlate with bone characteristics.

Table 4 shows the results of the multivariate linear regres-sion models evaluating the independent association between the 25-OHD levels and the bone characteristics as depend-ent variables, adjusting for sex, age, BMI, FM and FFM, and creatinine. The association between 25-OHD levels and bone parameters was significant for CSA and CSAc at the radius (partial R2 = 0.05 and 0.09, respectively), and for BMDt and CSAc at the tibia (partial R2 = 0.11 and 0.02, respectively).

DiscussionThis study identified a strong correlation between serum

25-OHD levels and bone parameters assessed by pQCT, suggesting an important but still incompletely elucidated

role for vitamin D in contributing to bone density and bone area in older adults.

The high prevalence of older adults with low basal serum levels of 25-OHD in our study population has already been reported in many other publications (2,25) and is attribut-able to an inadequate intake, scarce exposure to sunlight, and a declining vitamin D synthesis in the skin (25).

Our sample’s mean BMI corresponded to the 50th percen-tile of a reference older adult population in the ILSA study (26). Our exclusion criteria meant that our older adults’ total cumulative illness rating scores were low, indicating minimal comorbidities, so our participants could be seen as a sample of healthy older adults with a good nutritional status.

The association between 25-OHD and bone density was only evident for the trabecular bone regions in the tibia. Previous findings in rats showed that vitamin D deficiency results in a significant decline in trabecular bone density, but the reasons why 25-OHD is differently associated with

Table 2. Bone Mean Values of the Total Sample and by 25-OHD Levels by Peripheral Quantitative Computerized Tomography

Bone Characteristics

Total Sample, N = 134,

Mean ± SD≤25, N = 27, Mean ± SD

25–50, N = 36, Mean ± SD

50–75, N =38, Mean ± SD

>75, N = 33, Mean ± SD Unadjusted, p*

Sex–Age, Adjusted p*

Radius BMDc (mg/cm3) 1081.5 ± 47.7 1060.1 ± 48.2 1077.4 ± 45.3 1090.3 ± 47.9 1093.8 ± 44.8 .03 n.s BMDt (mg/cm3) 167.8 ± 54.6 145.9 ± 44.3 166.5 ± 66.3 178.0 ± 52.8 175.8 ± 46.8 .09 n.s CSA (mm2) 409.6 ± 86.8 350.0 ± 82.8 392.3.8 ± 81.7 435.6 ± 88.3 433.6 ± 76.0 .004 .0004 CSAc (mm2) 74.1 ± 22.2 50.2 ± 16.0 70.1 ± 17.6 80.1 ± 18.4 85.0 ± 22.0 <.0001 <.0001Tibia BMDc (mg/cm3) 1051.6 ± 99.3 1016.6 ± 49.6 1042.7 ± 51.3 1051.4 ± 53.5 1088.9 ± 173.4 .02 n.s BMDt (mg/cm3) 192.1 ± 41.0 172.9 ± 35.7 191.1 ± 43.6 192.9 ± 37.8 207.6 ± 41.6 .05 .04 CSA (mm2) 1219.6 ± 193.9 1144.1 ± 201.6 1160.6 ± 165.8 1264.0 ± 204.1 1267.3 ± 188.1 .06 .02 CSAc (mm2) 268.4 ± 59.2 229.2 ± 47.2 259.6 ± 55.8 284.8 ± 55.6 279.4 ± 63.5 .02 .01

Notes: BMDc = cortical bone mineral density; BMDt = trabecular bone mineral density; CSA = total cross-sectional area; CSAc = cortical cross-sectional area.*p values obtained by general linear models F test, with 25-OHD level included in the model as categorical variable.

Table 3. Simple Correlation Between Bone Parameters at Radius and Tibia Sites and Independent Variables

Radius Tibia

BMDc BMDt CSA CSAc BMDc BMDt CSA CSAc

Age −0.29** −0.18* 0.03 −0.18 −0.32*** −0.20* 0.11 −0.01BMI 0.26** −0.02 −0.03 0.05 0.28** 0.09 −0.02 0.15FFM 0.25* 0.46*** 0.69**** 0.73**** 0.48**** 0.36*** 0.75**** 0.87****FM −0.06 0.05 −0.30** −0.22* −0.13 0.13 −0.40*** −0.17PTH 0.13 0.10 0.02 0.10 0.07 0.01 0.15 −0.03Creatinine 0.07 0.20* 0.36*** 0.37*** 0.17 0.17 0.40*** 0.37***Albumin 0.09 0.03 0.10 0.18 0.01 −0.06 −0.02 0.01IL6 −0.10 0.11 0.05 0.01 −0.01 −0.07 0.11 −0.03

TNFα 0.02 0.14 0.10 0.06 −0.01 0.05 0.01 0.06

CI −0.09 −0.11 −0.01 −0.06 −0.06 −0.05 −0.18 −0.16SI −0.10 −0.10 0.09 0.06 −0.05 −0.08 −0.13 −0.13NASA 0.07 0.02 0.05 0.06 0.03 −0.04 −0.11 0.0125-OHD 0.20* 0.22* 0.31** 0.53*** 0.23* 0.33** 0.26* 0.30***

Notes: BMDt =  trabecular bone mineral density; BMDc = cortical bone mineral density; CSA =  total cross-sectional area; CSAc = cortical cross-sectional area; FFM = fat-free mass; FM = fat mass; BMI = body mass index; IL-6 = interleukin-6; TNF-α = tumor necrosis factor-alpha; 25-OHD = 25-Hydroxyvitamin D; CI = Comorbidity Index; SI = Severity index; NASA = NASA/Johnson Space Center Physical Activity Rating (PA-R) questionnaire.

*p < .05; **p < .01; ***p < .001; ****p < .0001.

at Florida Atlantic U

niversity on Novem

ber 15, 2014http://biom

edgerontology.oxfordjournals.org/D

ownloaded from

Page 6: Association Between Serum 25-Hydroxyvitamin D Levels, Bone Geometry, and Bone Mineral Density in Healthy Older Adults

25-HyDRoxyvItAMIN D AND BoNE PARAMEtERS 997

cortical and trabecular bone is still not clear. Trabecular bone may be more exposed to the endocrine effects of cir-culating hormones such as 25-OHD.

The association between 25-OHD and bone area was especially evident in the cortical bone area. This effect was stronger for the radial bone parameters, explaining about 9% of their cortical area variability (on multivariate linear regression modeling), whereas the contribution of 25-OHD to these parameters in the tibia explained only 2% of their variability. The metabolic effect of vitamin D on bone was more evident at the radial site. The gravitational effect of body mass is weaker on the radius than on the tibia, and the mechanical strength is less influential on bone parameters than the endocrine-metabolic component.

We surmise that other factors, such as exercise, muscle mass and a greater gravitational load, contribute to deter-mining bone parameters at the tibial site (especially in men) and that the effects of vitamin D are not actually site and gender related, but its action may be masked by other fac-tors at some sites (such as the tibia).

Cortical bone area is a good measure of total cortical bone mass and a valid marker of bone resistance to compression and tensile loads (27). It has also been demonstrated in rats that vitamin D supplementation has the effect of increasing a bone’s size rather than its density (28). This was confirmed by a pQCT analysis that demonstrated a close relationship between cortical CSA and bending fracture load to x and y axes (R2 ranged from 0.50 to 0.60 at both sites and in both genders; data not shown).

This unexpected association between vitamin D and several bone components measured using pQCT in older adults prompts several considerations. The lack of data in the literature on the relationship between 25-OHD and the bone parameters assessed by pQCT in older adults make our results difficult to compare. Chronic hypovitaminosis D

would most likely have a negative impact on bone, and find-ings based on a single measurement may not reflect an indi-vidual’s real long-term exposure. On the other hand, we had measured 25-OHD levels in a subgroup of 92 individuals 3 years before the present study, and there was no difference between the two values (concordance coefficient  =  0.85; unpublished data).

An expected outcome of our study was the finding that FFM and FM correlated with some of the bone parameters in our participants. Using the pQCT method enabled us to investigate the influence of body composition on cortical and trabecular bones, better clarifying the different effects of FFM and FM on the bone parameters considered. In our sample, FFM correlated particularly with cortical bone, emphasizing the effect of mechanical stress on cortical bone, on which the muscles exert their action through their insertion in the bone.

The effect of FM on bone seems to derive from a dif-ferent mechanism (29–34). Adipose tissue is a possible source of estrogen, which may help to prevent bone min-eral loss. Estrogen appears to be a major regulator of bone metabolism in men and in women (35), through its effect on RANKL and on inflammatory cytokines such as TNF-α and IL-1β, and possibly on the Wnt inhibitor, sclerostin (35).

This study has several limitations. First, our participants were volunteers, not a random sample of older adults, so our findings may not be generally applicable to more frail, nonambulatory older adults in nursing homes. In addi-tion, we assessed the relationship between vitamin D and bone strength based on bone density and geometry, not the real fracture risk. The study’s cross-sectional design also prevents us from attempting any causal inferences on the effect of 25-OHD levels on the way in which bone param-eters change during an individual’s life span. On the other hand, our study has the advantage of being the first (to the best of our knowledge) to provide a complete picture of the correlation between vitamin D levels and bone parameters regarding the tibia and radius of healthy older adults.

In conclusion, vitamin D seems to have a marked influ-ence on bone parameters, and 25-OHD levels could be used as a simple marker of bone structure and strength. Longitudinal studies on 25-OHD levels might help to pre-dict changes in bone parameters and would make a valuable contribution to the literature, shedding light on the effect of chronic exposure to low or high vitamin D levels and extending this investigation to more specific outcomes, such as fracture risk.

Acknowledgment

Our sincerest gratitude goes to the gym coordinator Flavio Martinello for his help in organizing recruitment.

Authors’ roles: study design: GS and MM; conduction of the study: MM, AC, GS, LB; data collection: SS, AI, MM, FB, and GR; data analysis: SS, MM, EP; data interpretation: GS, MM, AC; drafting of the manuscript: MM GS; revision of manuscript content: GS; approval of the final version: MM GS EP EM; GS, AC, and EM take responsibility for the integrity of the data analysis.

Table 4. Results* From the Multiple Linear Models Evaluating the Association Between 25(OH)D Values and Each Bone

Characteristic at Radius and Tibia Sites

Dependent Variables Total R2 β SE

p for β Estimate Partial R2

Radius, BMDc 0.14 0.20 0.15 .19 †

BMDt 0.31 0.28 0.16 .07 †

CSA 0.64 0.74 0.25 .005 0.05 CSAc 0.79 0.25 0.04 <.0001 0.09Tibia, BMDc 0.06 0.86 0.46 .07 †

BMDt 0.30 0.47 0.12 .0001 0.11 CSA 0.71 0.68 0.48 .16 †

CSAc 0.82 0.25 0.12 .03 0.02

Notes: BMDc  =  cortical bone mineral density; BMDt  =  trabecular bone mineral density; CSA  =  total cross-sectional area; CSAc  =  cortical cross-sectional area. Values in bold indicate significant p values.

*Adjusted estimates for 25-OHD were obtained by multivariate linear regression models built with stepwise procedure, including sex, age, BMI, fat mass, fat free mass, and serum creatinine as covariates.

†25-OHD removed from the model.

at Florida Atlantic U

niversity on Novem

ber 15, 2014http://biom

edgerontology.oxfordjournals.org/D

ownloaded from

Page 7: Association Between Serum 25-Hydroxyvitamin D Levels, Bone Geometry, and Bone Mineral Density in Healthy Older Adults

998 MoSELE Et AL.

References 1. Marlsand D, Colvin PL, Mears SC, Kates SL. How to optimize patients

for geriatric fracture surgery. osteop Int. 2010;21:S535–S546. 2. Lauretani F, Bandinelli S, Russo CR, et al. Correlates of bone quality

in older persons. Bone. 2006;39:915–921. 3. Bouxsein ML, Karasik D. Bone geometry and skeletal fragility. Curr

osteoporos Rep. 2006;4:49–56. 4. Szulc P. Bone density, geometry, and fracture in elderly men. Curr

osteoporos Rep. 2006;4:57–63. 5. Lauretani F, Bandinelli S, Griswold ME, et al. Longitudinal changes

in BMD and bone geometry in a population-based study. J Bone Miner Res. 2008;23:400–408.

6. Guglielmi G, Schneider P, Lang TF, Giannatempo GM, Cammisa M, Genant HK. Quantitative computed tomography at the axial and peripheral skeleton. Eur Radiol. 1997;7:32–42.

7. Augat P, Gordon CL, Lang TF, Iida H, Genant HK. Accuracy of cor-tical and trabecular bone measurements with peripheral quantitative computed tomography (pQCT). Phys Med Biol. 1998;43:2873–2883.

8. Russo CR, Ricca M, Ferrucci L. True osteoporosis and frailty-related osteopenia: two different clinical entities. J Am Geriatr Soc. 2000;48:1738–1739.

9. Ferretti JL, Cointry GR, Capozza RF, et al. Effects of bisphosphonates on the mechanical efficiency of normal and osteopenic bones. Medicina (B Aires). 1997;57:83–92.

10. Ferretti JL, Capozza RF, Zanchetta JR. Mechanical validation of a tomographic (pQCT) index for noninvasive estimation of rat femur bending strength. Bone. 1996;18:97–102.

11. Dawson-Hughes B, Heaney RP, Holick MF, Lips P, Meunier PJ, Vieth R. Estimates of optimal vitamin D status. osteoporos Int. 2005;16: 713–716.

12. Steingrimsdottir L, Gunnarsson O, Indridason OS, Franzson L, Sigurdsson G. Relationship between serum parathyroid hormone levels, vitamin D sufficiency, and calcium intake. JAMA. 2005;294:2336–2341.

13. Gennari C. Calcium and vitamin D nutrition and bone disease of the elderly. Public Health Nutr. 2001;4(2B):547–559.

14. Malavolta N, Pratelli L, Frigato M, Mulè R, Mascia ML, Gnudi S. The relationship of vitamin D status to bone mineral density in an Italian population of postmenopausal women. osteoporos Int. 2005;16: 1691–1697.

15. Sakuma M, Endo N, Oinuma T. Serum 25-OHD insufficiency as a risk factor for hip fracture. J Bone Miner Metab. 2007;25:147–150.

16. Houston DK, Neiberg RH, Tooze JA, et al. Low 25-Hydroxyvitamin D predicts the onset of mobility limitation and disability in community-dwelling older adults: the Health ABC Study. J Gerontol A Biol Sci Med Sci. 2013;68:181–187.

17. Annweiler C, Rolland Y, Schott AM, et al. Higher vitamin D dietary intake is associated with lower risk of alzheimer’s disease: a 7-year follow-up. J Gerontol A Biol Sci Med Sci. 2012;67:1205–1211.

18. Slinin Y, Paudel M, Taylor BC, et al. Study of Osteoporotic Fractures Research Group. Association between serum 25(OH) vitamin D and

the risk of cognitive decline in older women. J Gerontol A Biol Sci Med Sci. 2012;67:1092–1098.

19. Bischoff-Ferrari HA, Dietrich T, Orav EJ, Dawson-Hughes B. Positive association between 25-hydroxy vitamin D levels and bone mineral density: a population-based study of younger and older adults. Am J Med. 2004;116:634–639.

20. von Mühlen DG, Greendale GA, Garland CF, Wan L, Barrett-Connor E. Vitamin D, parathyroid hormone levels and bone mineral density in community-dwelling older women: the Rancho Bernardo Study. osteoporos Int. 2005;16:1721–1726.

21. Barbour KE, Zmuda JM, Horwitz MJ, et al. Osteoporotic Fractures in Men (MrOS) Research Group. The association of serum 25-hydroxy-vitamin D with indicators of bone quality in men of Caucasian and African ancestry. osteoporos Int. 2011;22:2475–2485.

22. Jackson AS, Blair SN, Mahar MT, Wier LT, Ross RM, Stuteville JE. Prediction of functional aerobic capacity without exercise testing. Med Sci Sports Exerc. 1990;22:863–870.

23. Baumgartner RN, Stauber PM, McHugh D, Koehler KM, Garry PJ. Cross-sectional age differences in body composition in persons 60+ years of age. J Gerontol A Biol Sci Med Sci. 1995;50:M307–M316.

24. Linn BS, Linn MW, Gurel L. Cumulative illness rating scale. J Am Geriatr Soc. 1968;16:622–626.

25. Harvey N, Dennison E, Cooper C. Osteoporosis: impact on health and economics. Nat Rev Rheumatol. 2010;6:99–105.

26. Sergi G, Perissinotto E, Pisent C, et al. ILSA Working Group. An ade-quate threshold for body mass index to detect underweight condition in elderly persons: the Italian Longitudinal Study on Aging (ILSA). J Gerontol A Biol Sci Med Sci. 2005;60:866–871.

27. Rittweger J, Beller G, Ehrig J, Jung C, Koch U, Ramolla J. Bone–mus-cle strength indices for the human lower leg. Bone. 2000;27:319–326.

28. Iwamoto J, Yeh JK, Takeda T, Sato Y. Effects of vitamin D supplemen-tation on calcium balance and bone growth in young rats fed normal or low calcium diet. Horm Res. 2004;61:293–299.

29. Reid IR. Relationships among body mass, its components, and bone. Bone. 2002;31:547–555.

30. Reid IR. Fat and bone. Arch Biochem Biophys. 2010;503:20–27. 31. Coin A, Sergi G, Benincà P, et  al. Bone mineral density and body

composition in underweight and normal elderly subjects. osteoporos Int. 2000;11:1043–1050.

32. Visser M, Kiel DP, Langlois J, et  al. Muscle mass and fat mass in relation to bone mineral density in very old men and women: the Framingham Heart Study. Appl Radiat Isot. 1998;49:745–747.

33. Reid IR, Plank LD, Evans MC. Fat mass is an important determinant of whole body bone density in premenopausal women but not in men. J Clin Endocrinol Metab. 1992;75:779–782.

34. Horstman AM, Dillon EL, Urban RJ, Sheffield-Moore M. The role of androgens and estrogens on healthy aging and longevity. J Gerontol A Biol Sci Med Sci. 2012;67:1140–1152.

35. Khosla S. Pathogenesis of age-related bone loss in humans. J Gerontol A Biol Sci Med Sci. 2012. doi:10.1093/gerona/gls163.

at Florida Atlantic U

niversity on Novem

ber 15, 2014http://biom

edgerontology.oxfordjournals.org/D

ownloaded from