Although only few postmenopausal women exhibit biochemical signs of hypovitaminosis D, vitamin D insufficiency has been shown to have adverse effects on bone metabolism and could be an important risk factor for osteoporosis and fracture. We determined serum levels of 25-hydroxyvitamin D [25(OH)D], intact parathyroid hormone (iPTH), bone turnover markers, dietary calcium intake, and bone mineral density (BMD; measured by dual X-ray absorptiometry) in 161 consecutive ambulatory women, healthy except for osteoporosis, referred to a bone metabolic unit. The prevalence of vitamin D insufficiency [25(OH)D ≤ 15 ng/ml] was 39.1%. 25(OH)D was lower in the osteoporotic subjects (15.7 ± 5.3 ng/ml vs. 21.8 ± 9.7 ng/ml; p < 0.001). After controlling for all other variables, lumbar spine (LS) BMD was found to be significantly associated with 25(OH)D, body mass index (BMI), and years after menopause (YSM) (R2 = 0.253; p < 0.001). For femoral neck (FN), significant independent predictors of BMD were YSM, BMI, iPTH, and 25(OH)D (R2 = 0.368; p < 0.001). The probability of meeting osteoporosis densitometric criteria was higher in the vitamin D insufficiency group (odds ratio [OR], 4.17, 1.83-9.48) after adjusting by YSM, BMI, iPTH, and dietary calcium intake. Our study shows that vitamin D insufficiency in an otherwise healthy postmenopausal population is a common risk factor for osteoporosis associated with increased bone remodeling and low bone mass.
OSTEOPOROSIS IS the most common human metabolic bone disorder. It has been defined as “a disease characterized by low bone mass and microarchitectural deterioration of bone tissue, leading to enhanced bone fragility and a consequent increase in fracture risk.”(1, 2) It is an important public health issue in postmenopausal women and if untreated, more than half of white women will experience fractures during their lifetime.
Normal bone metabolism depends on the presence of appropriate repletion of vitamin D. The importance of vitamin D insufficiency is related primarily to bone integrity. Although only few patients with osteoporosis exhibit obvious biochemical signs of hypovitaminosis D, vitamin D insufficiency has been shown to have adverse effects on calcium metabolism, osteoblastic activity, matrix ossification, bone mineral density (BMD), and bone remodeling.(3–6) Determination of serum 25-hydroxyvitamin D [25(OH)D] is the most clinically reliable indicator of vitamin D status. Thus, low serum 25(OH)D concentration is associated with secondary hyperparathyroidism, increased bone turnover, reduced BMD, and increased risk of osteoporotic fractures.(4–6) In addition, administration of vitamin D to the elderly slows bone turnover, increases BMD, and can reduce the rates of fragility fractures.(7) However, although previous studies have established the importance of vitamin D insufficiency in selected groups at risk(8–10) such as elderly patients who are housebound, living in a nursing home, or general medical inpatients, the role of this disorder on bone metabolism and postmenopausal bone loss is still controversial.(11–14) Therefore, we assessed the relationships between vitamin D insufficiency, parathyroid hormone (PTH) concentrations, BMD, and biochemical markers of bone turnover in a healthy (except for osteoporosis), community-dwelling, ambulatory postmenopausal women.
MATERIALS AND METHODS
From November 1998 to April 1999, 363 Spanish women were referred by general practitioners to our unit for osteoporosis screening in Granada, Spain (latitude 37°10′ N and longitude 3°35′ W). The study was approved by the Investigation Committee of the University Hospital “San Cecilio.” At the clinic, detailed medical histories and biochemical studies were performed to identify any causes for low BMD. After giving informed consent, 161 postmenopausal women were recruited. All were white; ambulatory; in good health except for osteoporosis; had normal values for serum calcium and phosphorus; and did not present renal, hepatic, gastrointestinal, or thyroid diseases or any other secondary causes for low BMD or serum vitamin D levels. None of them had been treated with calcium supplements, vitamin D preparations, hormone therapy, antiresorptive therapy, thiazides, steroids, or other medications that might affect BMD or vitamin D metabolism.
Assessment of dietary calcium intake
Dietary calcium intake was assessed by a detailed weekly food-frequency interview performed by a single expert dietitian. Each subject was asked to indicate the type and quantity of all food and drink consumed over a 1-week period and the dietary calcium content was calculated using a local food composition table.(15) They were entered in a database and analyzed by a single trained operator using a computer program to obtain the average daily intake of calcium.
Bone density measurements and radiography
BMD was measured by dual-energy X-ray absorptiometry (Hologic QDR1000; Hologic, Inc., Waltham, MA, USA) at lumbar spine (LS) and femoral neck (FN). The in vivo precision (CV) was better than 2% at both sites of measurement. A total of 2552 healthy normal subjects (1331 females and 1221 males) served to establish the mean BMD in the healthy Spanish population and to calculate the T score (number of SDs of the patient's value from the mean of young control population) and Z score for each BMD measurement (number of SDs of the patient's value from the mean control population in a 5-year-age band). The characteristics of this reference population have been described previously.(16) According to the World Health Organization (WHO) criteria,(17) we considered osteoporosis when the LS and/or FN T score was equal or below −2.5 SD. Vertebral fractures, calcifications, or other significant degenerative changes were assessed by X-ray study and corrected according to the method of Orwoll.(18)
Anteroposterior and lateral X-rays of the thoracic and lumbar spine were taken for all women to assess the presence of fractures. The anterior, central, and posterior height of each of the 13 vertebral bodies from Th4 to L4 were measured. Vertebral fracture was defined as a reduction of as least 20% in the anterior, central, or posterior vertebral height according to the following criteria: (a) anterior wedge deformity—ratio of anterior to posterior height <80%; (b)concavity deformity—ratio of central to posterior height <80%; (c) crush deformity—ratio of posterior height to posterior height of the adjacent vertebra <80%.
Morning fasting samples of venous blood were taken. Serum 25(OH)D3 (25-hydroxyvitamin D125I radioimmunoassay [RIA]; reference value (RV), 27.8 ± 9.45 ng/ml; Incstar Corp., Stillwater, MN, USA) and intact PTH (iPTH; Immulite Intact PTH; RV, 10-72 pg/ml; Diagnostic Products Corp., Los Angeles, CA, USA) were assayed. Serum total alkaline phosphatase (tALP; Hitachi 704 autoanalyzer; RV, 100-280 IU/liter;Boehringer Mannheim, Mannheim, Germany), bone ALP (bALP; Tandem-T, Ostase ImmunoRadioMetric Assay; RV, 11.6 ± 4.11 μg/ml; Hybritech Europe, Liege, Belgium), osteocalcin (OC; Osteocalcin125I RIA; RV, 1.8-6.6 ng/ml; Incstar Corp.), tartrate-resistant acid phosphatase (TRAP; Hitachi 704 autoanalyzer; RV: ≤7 IU/liter; Boehringer Mannheim), fasting urine cross-linked carboxy-terminal telopeptide of type I collagen corrected by urinary creatinine (CTX; α-CrossLaps RIA; RV postmenopausal women, 429 ± 225 μg/mmol; Osteometer Biotech, Herlev, Denmark), and fasting urine calcium/creatinine ratio were determined as bone turnover markers. The samples were analyzed in multiple analysis; the interassay CVs were 8.6-12.5% for 25(OH)D, <7% for iPTH, 3.7-6.7% for bALP, 2.2% for tALP, <3.5% for OC, 13.7% for CTX, and 2.2% for TRAP.
Definition of vitamin D insufficiency
We selected 15 ng/ml (37 nM) as the lower 25(OH)D cut-off in our analysis, based on published data showing that serum iPTH is increased in patients who have 25(OH)D concentrations equal to or below 15 ng/ml.(8, 19, 20) Moreover, a specific cut point was calculated as the 25(OH)D concentration at which maximum signification is reached comparing bone turnover markers and iPTH (28 ng/ml).
Group means were compared using the two-tailed nonpaired Student's t-test. Simple linear regression analysis was used to study the relationship among 25(OH)D, iPTH, and BMD measurements. Multiple linear regression was used to estimate adjusted β-coefficients for changes in BMD; also, partial correlations were completed to measure percentage variance explained by each covariate. Logistic regression was used to estimate crude and adjusted odds ratio (OR) for the dichotomous outcomes of osteoporosis. The Hosmer and Lemeshow goodness-of-fit test used in the models resulted in nonsignificant (p > 0.3). The years since menopause (YSM) variable was entered in the model in the following categories: 0-10 YSM, 10-20 YSM, 20-30 YSM, and >30 YSM, the first being used as the reference category. The iPTH levels variable was entered in the model in two categories using a cut-off at 72 pg/ml, the upper limit of the normal range measured in our laboratory. The reference category was the iPTH < 72 pg/ml group. The body mass index (BMI; kg/m2) variable was entered in the model in two categories, with (BMI ≥ 30 kg/m2) or without (BMI < 30 kg/m2) obesity,(21, 22) with the obese group as the reference category. The dietary calcium intake variable was entered in the model in three categories: below to 500 mg/day, between 500 and 1000 mg/day, and greater than 1000 mg/day, with the upper category as the reference. We tested for differences in the relationship of vitamin D levels to change in these measurements by including interaction terms for vitamin D status in the models. None of these interaction terms were significant. Analyses were performed with adjustment for YSM, BMI, iPTH, and dietary calcium intake. Statistical significance was defined as a value of p < 0.05. All analyses were performed in SPSS, version 9.0.1 (SPSS Ltd., Chicago, IL, USA) using the multiple regression program and logistic regression program when calculating 95% CIs and p values.
The characteristics of the patients are shown in Table 1. The mean (±SD) serum 25(OH)D concentration for all 161 patients was 18.8 ± 8.4 ng/ml. The prevalence of vitamin D insufficiency [circulating 25(OH)D equal or less than 15 ng/ml] was 39.1%. According to the WHO criteria,(18) 83 patients (51.6%) were considered nonosteoporotic (normal bone mass or osteopenic; T score > −2.5) and 78 patients (48.4%) were osteoporotic (T score ≤ −2.5). Forty-five women were diagnosed as having vertebral fracture, 19 (30.2%) in the vitamin D insufficiency group and 26 (26.6%) in the normal vitamin D group. Thirty-four subjects (43.6%) of the osteoporotic group had prevalent vertebral fractures. The dietary calcium intake was below the recommended dietary allowance (1500 mg/day for postmenopausal women)(23) in the entire sample.
Table Table 1.. Clinical Characteristics of 161 Postmenopausal Women
There were significant linear correlations between 25(OH)D, age, and YSM (age vs. 25(OH)D: r = −0.247, p = 0.002; YSM vs. 25(OH)D: r = −0.317, p < 0.001) and between iPTH, age, and YSM (age vs. iPTH: r = 0.213, p = 0.007; YSM vs. iPTH: r = 0.256, p = 0.001).
25(OH)D, iPTH, and BMD
The average serum 25(OH)D was lower in the osteoporotic group (Table 1). Serum iPTH levels were higher both in the osteoporotic and in the vitamin D insufficiency groups (Table 1) and even higher in subjects with vitamin D insufficiency independently of bone status (Fig. 1). More interestingly, the subjects with vitamin D insufficiency had a significantly lower BMD, with an average that fell 0.81 SD below the expected for their age at LS.
A strong negative linear correlation was found between iPTH and 25(OH)D (Fig. 2). There was a significant linear correlation between 25(OH)D and BMD at LS and FN (Fig. 3). Also, there was a significant linear correlation between iPTH and BMD at LS (iPTH vs. BMD [g/cm2]: r = −0.305, p < 0.001) and FN (iPTH vs. BMD [g/cm2]: r = −0.351, p < 0.001).
Bone turnover markers
The mean values of all biochemical markers of bone turnover exhibited no differences between 25(OH)D ≤ 15 ng/ml and 25(OH)D > 15 ng/ml groups or osteoporotic and nonosteoporotic groups except CTX, which was higher in the vitamin D insufficiency group (CTX: 433 ± 279 μg/mmol vs. 590 ± 429 μg/mmol; p = 0.012) and correlated with 25(OH)D (r = −0.240; p = 0.002). Also, there were internal significant correlations among bone turnover markers and between iPTH and bone turnover markers (Table 2). However, CTX, tALP, bALP, and iPTH were higher in patients with 25(OH)D < 28 ng/ml (CTX: 522 ± 352 μg/mmol vs. 320 ± 310 μg/mmol, p = 0.012; tALP: 195 ± 50 IU/liter vs. 170 ± 62 IU/liter, p = 0.045; bALP: 13.5 ± 6.4 μg/ml vs. 9.5 ± 4.3 μg/ml, p = 0.006; iPTH: 57.8 ± 17.9 pg/ml vs. 39.2 ± 16.1 pg/ml, p < 0.001).
Table Table 2.. Correlations Among iPTH and Biochemical Markers of Bone Turnover
Bivariate and multivariate analysis
After controlling for all other variables, LS BMD was found to be significantly associated with serum 25(OH)D levels, BMI, and YSM (Table 3). This final model explained 25.3% of the observed variance in lumbar spine BMD. For FN BMD, significant independent predictors of BMD were YSM, BMI, iPTH, and serum 25(OH)D levels. The final model explained 36.8% of the observed variance in FN BMD (Table 3). Overall, the serum 25(OH)D accounted for a 10.04% of the variance in BMD at LS after adjusting for BMI and YSM and a 2.52% at FN after adjusting for YSM, iPTH, and BMI.
Table Table 3.. Multiple Linear Regression Analysis of BMD
Taking into account the total effect of YSM, BMI, dietary calcium intake, iPTH levels, and 25(OH)D levels, the YSM variable was associated the most strongly with OP followed by 25(OH)D levels and BMI (Table 4). The effects of iPTH and dietary calcium intake were not statistically significant, but they should not be ignored because they increase the global significance of the model.
Table Table 4.. Bivariate and Multivariate Analysis
Our study shows that vitamin D insufficiency is a common risk factor for osteoporosis, associated with increased bone remodeling and postmenopausal osteoporosis, in healthy (except for osteoporosis), community-dwelling, ambulatory postmenopausal women.
In previous studies of elderly housebound people, medical inpatients, and nursing home residents, one-quarter to one-half had low vitamin D levels.(8–10, 24) However, in our study, with ambulatory subjects without secondary causes or medications that might affect bone density, the prevalence of vitamin D insufficiency was 39.1%. The high frequency of vitamin D insufficiency in postmenopausal women in this sunny region should be of concern.
The main factors contributing to vitamin D insufficiency are poor vitamin D intake, low sunlight exposure (staying indoors), and disorders or drugs known to affect vitamin D metabolism.(25) In our population, secondary causes affecting vitamin D metabolism were excluded. Serum 25(OH)D was measured from November to April, a winter period where there is normally no cyclical seasonal variation. In other countries in which dairy products are fortified with vitamin D, it is well known that calcium-deficient diets tend to be vitamin D-deficient too, because a single food, milk, is the principal dietary source of both of these nutrients.(26) In our country, this finding may be justified by the lack of a governmental mandate that food be supplemented with vitamin D. Hence, cutaneous synthesis would be the major source of the vitamin in ambulatory women of this age group. This could be an explanation for the low serum levels of 25(OH)D found in our sample.
In our study, the pathogenic role of vitamin D insufficiency in the decreased bone mass is suggested by the significant correlation between 25(OH)D and BMD in LS and FN. Moreover, the average serum 25(OH)D was lower in the osteoporotic group and women with vitamin D insufficiency had a significantly lower BMD. Secondary hyperparathyroidism is a well-known consequence of vitamin D insufficiency.(3, 24, 27–29) Our data show a strong negative correlation between iPTH and 25(OH)D and a significant correlation among iPTH and bone turnover markers. This biochemical picture suggests that increased concentrations of PTH might play an important role in the development of increased bone remodeling and bone loss in postmenopausal women with vitamin D insufficiency. Villarreal et al.(5) could not find in their population a correlation between 25(OH)D and iPTH at physiological concentrations. In contrast, in our study and others reported previously(8, 30) 25(OH)D is involved directly in the feedback control of PTH, even at physiological concentrations. We found that aging may lead to an increase in PTH levels and a decrease in vitamin D concentrations in postmenopausal women, as previous reports show.(31) Despite this fact, these findings are not likely to account entirely for the low bone mass found in the majority of our vitamin D insufficiency subjects. We observed that 25(OH)D and iPTH still predict independently BMD at LS and FN, respectively, after adjusting for YSM, dietary calcium intake, and BMI.
Vitamin D status appeared to be less related to proximal femur than to LS BMD. This may be caused by the fact that other mechanical factors (i.e., physical activity and weight) could influence proximal femur BMD to a higher extent. Moreover, the sex hormone deficiency makes LS bone more susceptible for vitamin D insufficiency. As it has been suggested previously, estrogens deficiency potentiates the effect of PTH excess because of vitamin D insufficiency.(32) Furthermore, hyperparathyroidism predisposes to cortical rather than cancellous bone loss, which would be more obvious at FN compared with LS and also may explain that iPTH was a significant independent predictor of BMD at FN, influencing to a lower extent the vertebral bone mass. This dual effect of PTH on bone has been described classically in patients with primary hyperparathyroidism.(33, 34)
There were no differences in biochemical markers of bone turnover between 25(OH)D ≤ 15 ng/ml and 25(OH)D > 15 ng/ml groups or osteoporotic and nonosteoporotic groups, except CTX. However, CTX, tALP, bALP, and iPTH levels were higher in patients with 25(OH)D < 28 ng/ml and there was a significant positive correlation between iPTH and bone turnover markers. This result supports previous data suggesting that the 25(OH)D threshold level needed to prevent secondary hyperparathyroidism and bone loss should be higher.(26, 27, 35–38)
Serum 25(OH)D is the most clinically available measurement of vitamin D status reflecting lifestyle and dietary habits.(25) Assuming a cut-off level of 15 ng/ml in our sample, the probability of meeting osteoporosis densitometric criteria was higher in the vitamin D insufficiency group, after adjusting for YSM, BMI, serum iPTH levels and dietary calcium intake. These results could suggest another effect of vitamin D on bone metabolism independently of PTH excess as described in animal models and in vitro studies.(39, 40)
Limitations of our study include those associated with cross-sectional studies. All the subjects studied attended our unit for screening of osteoporosis; so, our sample is not representative of the general population. Otherwise, we probably have underestimated the prevalence of hypovitaminosis D by not identifying subjects with low serum 1,25(OH)D concentrations despite adequate serum 25(OH)D concentrations.
Our most interesting finding shows that vitamin D contributes for 10% of the BMD variance at LS and 2.5% of BMD at FN and that the probability of meeting osteoporosis criteria is significantly higher in the vitamin D insufficiency group. Vitamin D contribution to the BMD is not negligible and represents an attractive candidate agent for its use in preventive interventions like systematic supplementation of dairy products. In conclusion, we found that vitamin D insufficiency in an otherwise healthy postmenopausal population is a common risk factor for osteoporosis associated with increased bone remodeling and low bone mass. Because replacement therapy is inexpensive, simple, and safe, a widespread screening for vitamin D insufficiency and the normalization of vitamin D status should alter our routine medical practice and have important public health implications even in southern latitude countries.
We express our gratitude to Andrew F. Stewart for his exciting critical review of this article. This work was supported by a grant from Servicio Andaluz deSalud (Exp 97/44-304) and the Hospital Clínico Foundation.