SEARCH

SEARCH BY CITATION

Keywords:

  • vitamin D;
  • PTH;
  • bone turnover;
  • BMD;
  • latitude

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Poor vitamin D status is common in the elderly and is associated with bone loss and fractures. The aim was to assess worldwide vitamin D status in postmenopausal women with osteoporosis according to latitude and economic status, in relation to parathyroid function, bone turnover markers, and BMD. The study was performed in 7441 postmenopausal women from 29 countries participating in a clinical trial on bazedoxifene (selective estrogen receptor modulator), with BMD T-score at the femoral neck or lumbar spine ≤ −2.5 or one to five mild or moderate vertebral fractures. Serum 25(OH)D, PTH, alkaline phosphatase (ALP), bone turnover markers osteocalcin (OC) and C-terminal cross-linked telopeptides of type I collagen (CTX), and BMD of the lumbar spine, total hip, femoral neck, and trochanter were measured. The mean serum 25(OH)D level was 61.2 ± 22.4 nM. The prevalence of 25(OH)D <25, 25–50, 50–75, and >75 nM was 5.9%, 29.4%, 43.5%, and 21.2%, respectively, in winter and 3.0%, 22.2%, 47.2%, and 27.5% in summer. Worldwide, a negative correlation between 25(OH)D and latitude was observed. With increasing 25(OH)D categories of <25, 25–50, 50–75, and >75 nM, mean PTH, OC, and CTX were decreasing (p < 0.001), whereas BMD of all sites was increasing (p < 0.001). A threshold in the positive relationship between 25(OH)D and different BMD parameters was visible at a 25(OH)D level of 50 nM. Our study showed a high prevalence of low 25(OH)D in postmenopausal women with osteoporosis worldwide. Along with latitude, affluence seems to be an important factor for serum 25(OH)D level, especially in Europe, where it is strongly correlated with latitude.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Vitamin d deficiency is common in elderly people and in patients with osteoporosis.(1) It causes secondary hyperparathyroidism, high bone turnover, bone loss, mineralization defects, and fractures. Other consequences include myopathy and falls.(2,3) Low vitamin D status may play role in diabetes mellitus,(4) cancers, multiple sclerosis, and other autoimmune diseases,(5) and was associated with poorer physical performance,(6) falls and fractures,(3) and a greater risk of nursing home admission(7) in older men and women.

Vitamin D3 is synthesized in human skin after the photoisomerization of 7-dehydrocholesterol (7DHC) to previtamin D3, under the influence of UV B (UVB) radiation (wavelength, 280–315 nm). The major factors influencing this process are either environmental (latitude, season, time of day, ozone and clouds, reflectivity of the surface) or personal (skin type, age, clothing, use of sunscreen).(8) Oily fish also contains vitamin D3, and margarine, milk, some breads, and yogurts, at least in the United States, are fortified either with vitamin D3 or with vitamin D2.(9) With higher latitudes, there is an increase in the length of the “vitamin D winter,” when no previtamin D3 is produced in the skin: for example, it lasts from November through February at latitude 42° (Boston, MA, USA) and from October through March at 52° (Edmonton, Canada).(10) Vitamin D is metabolized in the liver to 25-hydroxyvitamin D [25(OH)D], and the measurement of circulating level of 25(OH)D is used to determine vitamin D status. Although the vitamin D-replete and –deficient states have been defined,(2,11) there is still no consensus on a cut-off value for the definition of low 25(OH)D status or a definition for successful repletion of vitamin D. Approximately 80 nM of 25(OH)D has been recently suggested to be sufficient.(12) There is also growing evidence from the international literature about high prevalence of unrecognized vitamin D deficiency worldwide in different age groups.(13–22) Surprisingly, the levels of 25(OH)D are often higher in the United States, Canada, and Scandinavia(23–25) than in the countries located at lower latitudes. These international differences are partially explained by different sunshine exposure, skin pigmentation, air pollution, skin covering, and vitamin D intake with diet,(26) as well as supplement use and fortification policies.(27) The use of different assays for the measurement of 25(OH)D also impairs the comparison between countries.(28)

The aim of this study was to describe 25(OH)D status according to season in postmenopausal women with osteoporosis, in different countries with different economic status all over the world using a central laboratory facility, and to investigate the relationship between 25(OH)D status and parathyroid function, bone turnover markers, and BMD.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Study population

For this study, baseline data were used from the Fracture Incidence Reduction and Safety of Bazedoxifene Acetate Compared with Placebo and Raloxifene in Osteoporotic Postmenopausal Women Clinical Trial. Bazedoxifene is one of the selective estrogen receptor modulators (SERMs).

The total study population of this multicenter trial consisted of 7491 postmenopausal women. For this study, baseline data were available from 7455 women, 50–85 yr of age (mean, 66.4 ± 6.7 yr), from 29 countries on six continents (North America, South America, Europe, Asia, Africa, Australia).

All women who participated in this trial were at least 2 yr postmenopausal and had either osteoporosis according to WHO criteria (BMD T-score at the lumbar spine or femoral neck was < –2.5) or one to five mild or moderate asymptomatic vertebral fractures with a T-score > −3.5. Women with a history of diseases that may affect bone metabolism other than osteoporosis, severe prevalent vertebral fractures, postmenopausal symptoms requiring treatment, known history or suspected cancer of the breast, malignancy within the last 10 yr, history of venous thromboembolic events, active renal lithiasis, endocrine disorders requiring treatment (except well-controlled diabetes mellitus type 2 or hypothyroidism), and untreated malabsorption disorders were excluded. Patients with the use of the following drugs within 6 mo before screening were also excluded: systemic corticosteroids, systemic estriol > 2 mg/d, topical estrogen more often than three times a week, progestogens, androgens, calcitonin, bisphosphonates, PTH, SERMs, cholecalciferol (>50,000 IU/wk), and antiseizure drugs. In addition, the following subjects were excluded from this study based on laboratory measurements: high serum 25(OH)D (n = 2), high serum C-terminal cross-linked telopeptides of type I collagen (CTX; n = 1), low serum calcium (n = 2), high serum calcium (n = 8), and low serum phosphorus (n = 1), leaving 7441 women with known levels of serum 25(OH)D. From these women, 87.3% (n = 6495) were white, 6.5% (n = 482) were black, 4.6% (n = 346) were Hispanic, 1.3% (n = 100) were Asian, and 0.3% (n = 18) had other ethnicities.

The protocol was approved by the ethical review board at each center, and written informed consent was obtained from all participants in accordance with the Declaration of Helsinki.

The women were enrolled between December 2001 and September 2003 at 206 centers in 29 countries (23 in the northern and 6 in the southern hemisphere; a list of centers is available on request). Fasting blood samples were obtained at baseline, and after centrifugation, the serum samples were kept frozen until determination. Serum 25(OH)D was measured at the Covance Central Laboratory by the DiaSorin 25(OH)D assay with an interassay CV between 8.2% and 11.0%. Serum PTH concentrations were measured using the DiaSorin N-tact PTH SP immunoradiometric assay (IRMA), with an interassay CV of 3.4–4.9%. The alkaline phosphatase (ALP) and serum calcium assays were performed on the Roche Hitachi analyzers, with an interassay CV of 2.5–5.2% for ALP and 1.4–1.5% for calcium. The phosphorus, creatinine, and albumin assays were performed on the Roche Modular analyzer, with an interassay CV of 1.6–1.8%, 1.7–2.3%, and 2.3–2.6% for phosphorus, creatinine, and albumin, respectively. All assays on bone markers were performed in a specialized centralized laboratory (Synarc, Lyon, France) under the direction of Dr Patrick Garnero. Serum total osteocalcin (OC) and CTX were measured by automated analyser (Elecsys; Roche Diagnostics) with an interassay CV <7.2% for OC and <5.7% for CTX.

Seasons were defined as follows: summer, April-September; winter, October-March in the northern hemisphere and the reverse in the southern hemisphere. For each of the 206 centers, latitude was searched in a database of geographic coordinate information (http://www.tageo.com/index.htm and http://www.geonames.org). The data on vitamin D were presented per country according to geographical regions. For each country, gross domestic product per capita (GDP) was searched for 2003 in The World Economic Outlook (WEO) Database April 2003 (World Economic And Financial Surveys, International Monetary Fund, http://www.imf.org/external/pubs/ft/weo/2003/01/data/index.htm). GDP was expressed in current U.S. dollars (USD) per person. Data were derived by first converting GDP in local currencies to USD and dividing GDP by the total population in 2003. GDP was used as an indicator of affluence in different countries.

BMD of lumbar spine and hip (total hip, femoral neck, and trochanter; in g/cm2), were measured by DXA on Hologic, Lunar, or Norland densitometers and standardized.(29)

Statistical analysis

SPSS 12.0 was used to perform statistical analyses. ANOVA was used to assess the seasonal differences in serum 25(OH)D in different countries and geographical regions. ANOVA was also used to assess differences in serum PTH, bone turnover markers, and BMD according to different cut-points for serum 25(OH)D (i.e., <25, 25–50, 50–75, and >75 nM). Pearson's correlation coefficients were calculated between serum 25(OH)D, latitude, and GDP and BMD, according to season. Partial correlation was performed with the calculation of Pearson's correlation coefficient to study the relationship between vitamin D and latitude controlling for GDP. Locally weighted regression smoothing (LOESS) plots were performed in STATA to study the relationship between 25(OH)D, PTH, and BMD.

Potential confounders included age, BMI, serum creatinine, and season. First, unadjusted analyses were performed. Subsequently, potential confounders were added to the models.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

The mean serum 25(OH)D level in 7441 postmenopausal women was 61.2 ± 22.4 nM. There was a significant seasonal difference in mean serum 25(OH)D, with higher levels in the summer than in the winter for most countries and geographical regions. Table 1 presents these data together with the percentages of women in different serum 25(OH)D level groups (<25, 25–50, 50–75, and >75 nM) per season in different countries. The prevalence of serum 25(OH)D level <25 nM was higher in winter for all regions, with the highest prevalence in south and southeastern Europe, both for winter (up to 34.4% for Romania and 28.2% for Croatia) and for summer (10.9% for Croatia and 10.5% for Romania).

Table Table 1.. Seasonal Differences in Serum 25(OH)D and in Prevalence of Serum 25(OH)D <25, 25–50, 50–75, or >75 nM According to Country
Thumbnail image of

Fugure 1 shows the significant negative correlation between 25(OH)D and latitude worldwide. As expected, this correlation was stronger in winter (r = −0.36, p < 0.001 without controlling, and rc = –0.40, p < 0.001 with controlling for GDP per capita) than in summer (r = −0.09, p < 0.001 and rc = –0.13, p < 0.001, respectively). Conversely, in Europe, a significant positive correlation was observed between 25(OH)D and latitude (r = 0.21, p < 0.001). We examined this finding more closely, trying to explain it with available data. Although GDP per capita in 2003 was strongly correlated with latitude when Europe was taken as a whole (r = 0.51, p < 0.001), it differs substantially between European countries. Fugure 2 shows mean values of 25(OH)D, latitude, and GDP per country and differences between them. In the group of countries with GDP > 10.000 USD, the distribution of GDP was strongly correlated with latitude (r = 0.75, p < 0.001). The positive correlation between 25(OH)D and latitude in these countries (r = 0.13, p < 0.001) almost disappeared and became nonsignificant (rc = 0.01, p = 0.626) when controlling for GDP. In the countries with GDP < 10.000 USD (Eastern European economies), the correlation between GDP and latitude was less strong (r = 0.17, p < 0.001), and the positive correlation between 25(OH)D and latitude (r = 0.17, p < 0.001) remained significant when controlling for GDP (rc = 0.14, p < 0.001). The positive correlation between GDP and 25(OH)D, however, stayed significant for both groups of countries after controlling for latitude (rc = 0.10, p < 0.001 and rc = 0.18, p < 0.001 for the groups of countries with GDP <10.000 USD and >10.000 USD, respectively).

thumbnail image

Figure Figure 1. Relationship between latitude and serum 25(OH)D worldwide. (A) Winter. (B) Summer.

Download figure to PowerPoint

thumbnail image

Figure Figure 2. Relationship between latitude, serum 25(OH)D, and gross domestic product per capita (GDP) in USD in Europe. (A) Latitude and serum 25(OH)D. (B) Latitude and GDP. (C) Serum 25(OH)D and GDP.

Download figure to PowerPoint

Table 2 shows the mean values of bone markers and BMD measurements in different 25(OH)D groups. In groups with increasing mean 25(OH)D of <25, 25–50, 50–75, and >75 nM, mean PTH was 4.5 ± 1.5, 4.1 ± 1.5, 3.7 ± 1.2, and 3.5 ± 1.2 pM, respectively (by ANCOVA, p < 0.001). Also, the other parameters of bone turnover, such as serum levels of OC (bone formation) and CTX (bone resorption), were significantly lower in the highest serum 25(OH)D group (except for ALP, a marker of bone formation, which did not significantly change). All BMD values were significantly higher in the highest serum 25(OH)D group.

Table Table 2.. Mean Values (±SD) of Serum PTH, Bone Turnover Markers, and BMD for Different Groups Arranged According to Serum 25(OH)D of <25, 25–50, 50–75, and >75 nM
Thumbnail image of

The relationship between 25(OH)D and some bone parameters is presented in Fig. 3. LOESS plots, corrected for age, BMI, serum creatinine, and season, show the mean values of serum PTH and BMD for each value of serum 25(OH)D. In the inverse relationship between serum PTH and 25(OH)D, there was a steep decrease of PTH up to 50 nM and a more slow decrease between 50 and 100 nM. This relationship between PTH and 25(OH)D showed no plateau with serum 25(OH)D up to 100 nM. Although only significant for BMD of hip trochanter, the LOESS plot of the relationship between 25(OH)D and BMD parameters seemed to show a threshold around the 25(OH)D value of 50 nM. A similar shape of LOESS plot was observed for BMD of femoral neck (data not shown).

thumbnail image

Figure Figure 3. Relationship of serum 25(OH)D with PTH and BMD. (A) With PTH, p < 0.001. (B) With BMD of hip trochanter* (p = 0.01; dotted line represents 95% CI). *Similar shape was observed for the BMD of the femoral neck (p = 0.19; data not shown).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

This study allows one to compare serum 25(OH)D and PTH levels in postmenopausal women with osteoporosis throughout the world. A central laboratory facility was used to perform the measurements in 7441 women who were enrolled in one double-blind, randomized, controlled clinical trial, thereby eliminating the variation in assay and methods between the laboratories in different countries.

We found considerable differences in vitamin D status in different countries depending on season and latitude (Table 1), with high prevalence of serum 25(OH)D < 25 and 25–50 nM in many countries. The results are consistent with the previous global study with similar design, in which the highest prevalence of vitamin D deficiency [25(OH)D < 25 nM] was observed in countries of central and southern Europe.(23) In our study, more countries of southeastern Europe were represented, all of them with high prevalence of low serum 25(OH)D, especially in winter.

Worldwide, serum 25(OH)D was negatively correlated with latitude, a finding that we expected. For Europe, however, this correlation was paradoxically positive, as was also seen in other studies.(23–25) Thus, although the synthesis of vitamin D in the skin is known to be the most important source of vitamin D, there are more factors than latitude and sun exposure that determine the serum 25(OH)D level. The use of multivitamins, which is known to be associated with higher income, is common among the elderly in the United States and in Europe and was found to be 50–60%.(30,31) Dietary intake of vitamin D-rich food (oily fish like salmon, mackerel, herring, and sardines, which are considered to be the best sources) is also important to consider. The traditional use of cod liver oil or other supplements was up to 60% in a study from Iceland.(32) In a Danish study on the use of dietary supplements in 3707 women and 942 men, it was found that, in the age group of 60–65 yr, 78% of women and 60% of men were using some kind of supplement, whereas 18% of women and 14% of men also used fish oils.(33) In a Norwegian study on 37,226 women 41–55 yr of age, cod liver oil supplements were taken by 44.7%.(34)

In our study, GDP as a marker of affluence is positively correlated with 25(OH)D in Europe, also after correction for latitude. Therefore, in a group of European countries with GDP > 10.000 USD, in which GDP is strongly correlated with latitude (r = 0.75; p < 0.001), we expected to find a negative correlation between latitude and 25(OH)D after controlling for GDP. However, the positive correlation between latitude and 25(OH)D in these countries disappeared and became nonsignificant. We do not have data on income of participating women, neither on their diet, use of multivitamins, or holidays in sunny countries, which could all affect their serum 25(OH)D level. A low prevalence of low serum 25(OH)D in northern America and northern Europe was found before,(24,25) but as far as we know, the investigators did not try to explain this by differences in affluence between the countries. The indicators of poor income (lowest levels of income, food stamp use, food insufficiency) are known to be associated with lower dietary intakes among homebound older adults,(35) and high prevalence of diets with the lowest quartile in at least two of four musculoskeletal nutrients (vitamin D, calcium, magnesium, and phosphorus).(36) Therefore, it would be interesting to assess the relationship between income level per person, the diet, fortification policy, the use of supplements, and holidays in sunny destinations, and to study the share of each factor that has influence on the relation between latitude and serum 25(OH)D.

In our study, in the inverse relationship between serum PTH and 25(OH)D, no plateau was observed at serum 25(OH)D up to 100 nM. Above that level, there seems to be a plateau; but because 95% of our subjects have serum 25(OH)D <100 nM and the CIs >100 nM are wide, no conclusions can be drawn for serum 25(OH)D levels >100 nM. The results of a study on 25(OH)D, PTH, and calcium intake suggested that vitamin D sufficiency can maintain low serum PTH values even when the calcium intake level is <800 mg/d, whereas a high calcium intake (>1200 mg/d) is not sufficient to maintain low serum PTH, as long as vitamin D status is insufficient.(32) Interestingly, that study indicated that the variation of the relationship between 25(OH)D and PTH might be dependent on calcium intake, as well as different serum 25(OH)D levels needed for the inflection point of serum PTH. In fact, some studies find an inflection point in the relationship between serum 25(OH)D and serum PTH, whereas other do not. These different outcomes could be caused by different statistical techniques. For example, if a regression model is used, it suggests a negative relationship without a plateau. When locally weighted regression smoothing is used to show the shape of the relationship, some studies find a plateau at higher serum 25(OH)D levels. In our study, a plateau seems visible after serum 25(OH)D of 100–120 nM, but the CI above this level becomes wide, because there are not many subjects with such high serum 25(OH)D levels.

In contrast with another study,(34) PTH did not show any significant relationship with creatinine in our study, after controlling for weight, height, or both (data not shown), in the total group as well as in a group of respondents with 25(OH)D < 50 nM.

A recent review of the literature that reported a threshold for the relation between serum 25(OH)D and PTH found that most estimates were clustered between 40 and 50 nM or 70 and 80 nM. The same review found that, in the studies with a mean 25(OH)D of >50 nM, calcium intake did not affect PTH, but in studies with a mean 25(OH)D of <50 nM, dietary calcium was inversely related to PTH.(37) In an American study on vitamin D and calcium supplementation, which evaluated the effect of increasing 25(OH)D levels >25 nM with vitamin D therapy on blood concentrations of serum PTH, PTH levels did not substantially decline in subjects who had a starting blood level of 25(OH)D of at least 50 nM,(38) which is consistent with our observation.

In our study, a significant positive correlation was found between 25(OH)D and BMD of the femoral trochanter. Furthermore, mean values of BMD at all measured sites (lumbar spine, total hip, femoral neck, and femoral trochanter) were higher in groups with higher serum 25(OH)D level (Table 2). These absolute differences in mean BMD between the different vitamin D groups might be small, but at the population level, these differences could mean a substantial reduction in fracture risk. Hip fractures were found to be strongly related to reduced BMD in all regions of the proximal femur, with a risk ratio of 2.5–2.7 for hip fracture (95% CI, 1.9–3.6) with each SD decrease of BMD at any site of the proximal femur, after adjustment for age.(39) Therefore, a decrease in BMD of the femoral neck from 0.729 to 0.700 g/cm2 (0.25 of SD) in our data may increase the RR for hip fracture by >50%. Indeed, the protective effect of vitamin D on fractures was found in several studies.(14,40) In addition, a positive association between 25(OH)D and BMD was established in The National Health and Nutrition Examination Survey III (NHANES III) in 13,432 subjects including whites, Hispanics, and blacks.(41)

There are different views on the optimal level of serum 25(OH)D. For the prevention of rickets, a serum 25(OH)D level >25 nM seems to be sufficient. For prevention of bone loss and other outcomes, the optimal level of serum 25(OH)D probably is >50 nM.(42,43) However, in a recent review on the estimation of optimal serum concentrations of 25(OH)D for multiple health outcomes, the most advantageous serum concentration of 25(OH)D was found to be >75 nM for outcomes including BMD, lower extremity function, risk of falls, fractures, and colorectal cancer.(44) The percentage of postmenopausal women with osteoporosis with serum 25(OH)D <75 nM in the winter approaches 90–100% in Europe and 80% in Canada and the United States. Even in Brazil, the country with the best 25(OH)D status in our study, the percentage of women with 25(OH)D levels of >75 nM is only 34.3% in winter and 43% in summer. Accepting the cut-off value of 75 nM would implicate that almost 80% of postmenopausal women with osteoporosis worldwide should be treated for hypovitaminosis D in winter and up to 75% in summer because of levels <75 nM. To achieve these levels, a high supplementation dose might be needed. If the required level of serum 25(OH)D is 50 nM,(2,11) 35% of postmenopausal women with osteoporosis worldwide should be treated for hypovitaminosis D in winter and up to 25% in summer, and this would be easier to achieve.

The strength of our study is in the central laboratory facility for all measurements from 206 centers in 29 countries, eliminating the variation between laboratories. Another strong point is the significant relationship between serum 25(OH)D and BMD at all measured sites (Table 2). A limitation is that the women participating in a clinical trial usually differ from the general population. Therefore, the results can not be generalized to all postmenopausal women with osteoporosis. Because it is well known that subjects who are participating in clinical trials usually are more conscious about their health, the estimation of a high prevalence of vitamin D deficiency in the general population is conservative. Another limitation is a small sample size for some countries.

Our study showed a high prevalence of low serum 25(OH)D in women with postmenopausal osteoporosis all over the world. These results confirm that hypovitaminosis D is a worldwide problem that needs to be addressed.(18,45–47) Vitamin D deficiency has such important health implications that measurement of serum 25(OH)D was proposed to be part of a routine physical examination for children and adults of all ages.(47) The U.S. economic burden caused by vitamin D insufficiency from inadequate exposure to solar UVB irradiance, diet, and supplements was estimated at 40–56 billion USD in 2004, whereas the economic burden for excess UV irradiance was estimated at 6–7 billion USD.(48) Besides sunshine exposure, vitamin D status can be improved by increasing dietary intake (e.g., by oily fish and cod liver oil, food fortification, and vitamin D supplements).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

We are grateful to Caspar Looman for statistical support.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES