Relationships of Serum 25-Hydroxyvitamin D to Bone Mineral Density and Serum Parathyroid Hormone and Markers of Bone Turnover in Older Persons
The Journal of Clinical Endocrinology & Metabolism April 2009 Vol. 94, No. 4 1244-1250
Natalia O. Kuchuk, Saskia M. F. Pluijm, Natasja M. van Schoor, Caspar W. N. Looman, Johannes H. Smit and Paul Lips
Department of Endocrinology (N.O.K., P.L.), VU University Medical Center, 1007 MB Amsterdam, The Netherlands; Department of Public Health (S.M.F.P., C.W.N.L.), Erasmus MC, 3015 CE Rotterdam, The Netherlands; EMGO Institute (N.M.v.S., P.L.), VU University Medical Center, 1007 MB Amsterdam, The Netherlands; and Department of Sociology and Social Gerontology (J.H.S.), VU University, 1007 MB Amsterdam, The Netherlands
"In conclusion, low serum 25(OH)D concentrations are very common in the elderly. Bone health and physical performance in older persons are likely to improve when serum 25(OH)D is raised over at least 50-60 nmol/liter. The implication for our older population is that at least 64% should receive vitamin D supplements because they had a serum 25(OH)D level lower than 60 nmol/liter."
Context: Serum 25-hydroxyvitamin D [25(OH)D] may influence serum PTH and other parameters of bone health up to a threshold concentration, which may be between 25 and 80 nmol/liter.
Objective: The aim of the study was to assess the threshold serum 25(OH)D with regard to PTH, bone turnover markers, and bone mineral density (BMD)..
Design and Setting: This was part of the Longitudinal Aging Study Amsterdam, an ongoing cohort study.
Participants: A total of 1319 subjects (643 men and 676 women) between the ages of 65 and 88 yr participated in the study.
Main Outcome Measures: Serum 25(OH)D, PTH, osteocalcin, urinary deoxypyridinoline/creatinine, quantitative ultrasound of the heel, BMD of lumbar spine and hip, total body bone mineral content, and physical performance. The relationship between the variables was explored by analysis of covariance and the locally weighted regression (LOESS) plots.
Results: Serum 25(OH)D was below 25 nmol/liter in 11.5%, below 50 nmol/liter in 48.4%, below 75 nmol/liter in 82.4%, and above 75 nmol/liter in 17.6% of the respondents. Mean serum PTH decreased gradually from 5.1 pmol/liter when serum 25(OH)D was below 25 nmol/liter to 3.1 pmol/liter when serum 25(OH)D was above 75 nmol/liter (P < 0.001) without reaching a plateau. All BMD values were higher in the higher serum 25(OH)D groups, although only significantly for total hip (P = 0.01), trochanter (P = 0.001), and total body bone mineral content (P = 0.005). A threshold of about 40 nmol/liter existed for osteocalcin and deoxypyridinoline/creatinine, 50 nmol/liter for BMD, and 60 nmol/liter for physical performance.
Conclusions: Low serum 25(OH)D concentrations are common in the elderly. Bone health and physical performance in older persons are likely to improve when serum 25(OH)D is raised above 50-60 nmol/liter.
Vitamin D deficiency is common in the elderly. Long-lasting severe vitamin D deficiency may cause mineralization defects resulting in osteomalacia. In an earlier stage, vitamin D deficiency causes secondary hyperparathyroidism, resulting in high bone turnover and bone loss (1). Both osteomalacia and secondary hyperparathyroidism may cause fractures, especially hip fracture. Vitamin D and calcium supplementation can decrease the incidence of hip fracture and other nonvertebral fractures in nursing home residents, as was shown in two trials in Lyon, France (2, 3). The results of vitamin D supplementation in community-dwelling elderly have been controversial with positive and negative clinical trials (4, 5, 6). The negative outcome of some trials has been attributed to noncompliance and to the achievement of suboptimal levels of serum 25-hydroxyvitamin D [25(OH)D] (7). The required serum 25(OH)D for adequate bone health has been debated. The level of serum 25(OH)D leading to suppression of PTH was estimated according to some vitamin D supplementation studies as 30 or 50 nmol/liter (1, 8), but other studies showed that complete PTH suppression was obtained at higher serum 25(OH)D, about 75-80 nmol/liter (9, 10). The population-based National Health and Nutrition Examination Survey III study showed that BMD of the hip increased with a higher serum 25(OH)D up to about 80 nmol/liter (11). A roundtable conference stated that the optimal serum 25(OH)D level was somewhere between 50 and 80 nmol/liter (12). In a previous study from the Longitudinal Aging Study Amsterdam (LASA), a positive relationship was found between serum 25(OH)D and physical performance, significant up to serum 25(OH)D of 50 nmol/liter (13).. The aim of our study was to estimate the threshold serum 25(OH)D with regard to serum PTH, bone turnover markers, BMD, bone ultrasound parameters, and physical performance in a population-based LASA.
It has been well established that vitamin D deficiency causes secondary hyperparathyroidism and bone loss. The results of this LASA study confirm those of other studies, i.e. low serum 25(OH)D levels are associated with an increase of serum PTH, increased bone turnover, lower BMD, and lower physical performance. However, the cut-points or thresholds where serum PTH and bone resorption started to increase and BMD started to decrease varied highly in different studies. In our study, serum PTH did not reach a plateau with increasing serum 25(OH)D. The optimal serum 25(OH)D should be higher than 100 nmol/liter according to serum PTH. However, a steep decrease of serum PTH was observed until serum 25(OH)D was above 50 nmol/liter, and a more gradual decrease was seen between 50 and 100 nmol/liter. To prevent increased bone turnover, i.e. increased urine DPD/Cr excretion as a marker of bone resorption and increased serum OC as a marker of bone formation, serum 25(OH)D should be higher than 40 nmol/liter.. Furthermore, BMD and total body BMC increased up to serum 25(OH)D levels of 50-60 nmol/liter, whereas physical performance increased up to 25(OH)D levels of 60 nmol/liter.
These results have several implications. Most important, higher bone turnover and lower BMD at a serum 25(OH)D below 50 nmol/liter suggest deleterious effects of low serum 25(OH)D concentrations. The different thresholds for the different bone health parameters may represent the different nature of response of each parameter to serum 25(OH)D.
The fact that in our study all BMD parameters and total body BMC increased up to the serum 25(OH)D level of at least 50 nmol/liter indicates that the optimal serum 25(OH)D should be at least 50 nmol/liter. This finding is in line with a previous LASA study on vitamin D status and physical performance, in which we showed that serum 25(OH)D concentrations below 50 nmol/liter were associated with poorer physical performance and a greater decline in physical performance in older men and women (13).
Recent discussions have been focused on which circulating level of 25(OH)D is appropriate. This question is important because it has implications for the prevention of osteoporosis and fractures and, on the other side, for public health strategies including food fortification with vitamin D and the use of supplements. The required serum 25(OH)D has usually been established by assessing the threshold serum 25(OH)D below which serum PTH starts to rise. It has been suggested that the simultaneous measurement of serum PTH may aid in interpreting the circulating level of serum 25(OH)D because of an inverse relationship between serum 25(OH)D and PTH.. However, the increases of serum PTH associated with vitamin D deficiency are usually within the normal reference range; serum PTH has a short half-life and depends on calcium intake, so different data sets could lead to different conclusions (1). This threshold might vary between 30 nmol/liter in Amsterdam vitamin D study (19), 37.5 nmol/liter in a study in American hospital inpatients (20), 50 nmol/liter in an American study on vitamin D supplementation (21), 75 nmol/liter in the French SUVIMAX study (9), or 80 nmol/liter in an American rural female population (22). In a study that did provocative testing, which is often used in endocrinology to determine ultimately the homeostatic level for a particular hormone, 50 nmol/liter was considered the minimum 25(OH)D value for vitamin D sufficiency in the age group of 49-83 yr (21).
In a study conducted in southern California among ambulatory community-dwelling women (74.6 ± 10.0 yr), only 2% of these older women had 25(OH)D below 50 nmol/liter, whereas in 18% of them, the serum PTH levels were indicative of hyperparathyroidism (>5.86 pmol/liter or 65 ng/liter), suggesting that the criterion for vitamin D insufficiency (<50 nmol/liter) might be too low (23). In that study, there was little seasonal variation for PTH, probably reflecting the overall adequate values of 25(OH)D with a mean of 102.0 (±35.0) nmol/liter, and all PTH values were below 5.9 pmol/liter (65 pg/ml) above a threshold serum 25(OH)D level of 120 nmol/liter. An English study found that at a serum 25(OH)D level of 80 nmol/liter, 1.5% had an elevated serum PTH; at 50 nmol/liter, 8% had raised serum PTH; and at 30 nmol/liter, 13% had a raised serum PTH (24). In that study, no patient had a serum PTH above the normal range (4.32 pmol/liter or 48 pg/ml) at a serum 25(OH)D level of 100 nmol/liter. Interestingly, in an Italian study in 104 centenarians, where circulating serum 25(OH)D was undetectable in 99 subjects and detectable in five subjects only with a mean of 16.0 nmol/liter, serum PTH was not exceeding the upper normal limit of 5.9 pmol/liter or 65 pg/ml in 27% (25). However, this could be due to a state of functional hypoparathyroidism and could possibly be related to magnesium depletion (26), because PTH release is impaired during magnesium deficiency (27).
We found a continuous decline in serum PTH with increasing serum 25(OH)D and no plateau, consistent with other studies (28, 29, 30). In our study, a small increase of serum PTH when serum 25(OH)D fell from 100 to 50 nmol/liter was interpreted as a physiological compensatory mechanism to increase renal 1- hydroxylation and serum 25(OH)D to maintain a stable serum calcium level. The tendency of establishing the required serum 25(OH)D by assessing the threshold below which serum PTH starts to rise could lead to the use of almost the upper limit of serum 25(OH)D level in different data sets as a threshold, therefore establishing the required serum 25(OH)D level at 100-120 nmol/liter.
We believe that in establishing the required serum 25(OH)D level, different aspects of bone health must be considered. Previous LASA studies on vitamin D status and clinical outcomes found that a serum 25(OH)D level lower than 25 nmol/liter was associated with an increased risk of falling (31), and serum 25(OH)D levels below or equal to 30 nmol/liter were associated with an increased fracture risk in the age group of 65-75 yr (32). However, the multifactorial etiology of falls and fractures makes it difficult to use these parameters for determination of the required serum 25(OH)D.
A significant positive association between 25(OH)D levels and BMD was observed in the NHANES III study, where a threshold of about 80 nmol/liter was found, representing serum 25(OH)D level above which BMD increased more slowly (11). A recent review suggests that a desirable serum 25(OH)D concentration for optimal health begins at 75 nmol/liter, with the best concentration being 90-100 nmol/liter (33). Our study yielded a lower threshold (50-60 nmol/liter) which may be partly due to a difference between assays for serum 25(OH)D (34). Different thresholds of serum 25(OH)D for serum PTH, bone turnover, BMD, and neuromuscular function may also be explained by the extrarenal hydroxylation of 25(OH)D to the active metabolite 1,25-dihydroxyvitamin D in different organs because it is known that 1,25-dihydroxyvitamin D is capable of acting via autocrine and paracrine mechanisms (35).
The strengths of our study are in its prospective cohort design including men and women, being a representative group of the Dutch population, and the combination of serum PTH, OC, urinary DPD/Cr, different bone density measurements, as well as physical performance as parameters of bone health. The limitations are in the relatively smaller subgroup of persons in whom BMD measurements were performed. Furthermore, we adjusted for the season of blood collection, but because this study is cross-sectional, we could not adjust for the cyclic variations in each parameter that were observed in other studies (36, 37).
The question remains which dose is recommended. A recent review on this matter suggests that the dose of vitamin D in the management of osteoporosis should be no less than 700-800 IU per day (38). By supplementation of 400 IU per day (5, 39), a mean level of only 60 nmol/liter was achieved. In a study on supplementation of vitamin D in institutionalized elderly with a very good compliance, 90% of participants achieved a serum 25(OH)D above 50 nmol/liter after 4 months of a daily supplementation with 600 IU of vitamin D (40). As was shown in a meta-analysis on vitamin D and fracture prevention, a dose of 700-800 IU per day appears to reduce the risk of hip and any nonvertebral fractures in older persons (7). Indeed, with a daily supplementation of 800 IU, the mean levels of 25 (OH)D became 75 nmol/liter, being above 60 nmol/liter in 80% of participants with good compliance after 6 months (41). In a global study, vitamin D supplementation had more effect on serum 25(OH)D and serum PTH when baseline serum 25(OH)D was lower (42). Therefore, the supplementation dose should be 600-800 IU per day when sunshine exposure is insufficient, and a decision about it should be made on an individual basis with respect to baseline serum 25(OH)D level, lifestyle, and diet.
In conclusion, low serum 25(OH)D concentrations are very common in the elderly. Bone health and physical performance in older persons are likely to improve when serum 25(OH)D is raised over at least 50-60 nmol/liter. The implication for our older population is that at least 64% should receive vitamin D supplements because they had a serum 25(OH)D level lower than 60 nmol/liter.
Serum 25(OH)D was significantly higher in men than in women, whereas mean PTH levels did not differ (Table 1). Bone turnover markers, serum OC, and urinary DPD/Cr, were higher in women than in men. As expected, bone ultrasound parameters and BMD measurements were significantly higher in men than in women.
Table 2 shows the mean values of bone markers, BMD measurements, and physical performance for different 25(OH)D level groups. Serum 25(OH)D was below 25 nmol/liter in 11.5%, below 50 nmol/liter in 48.4%, below 75 nmol/liter in 82.4%, and above 75 nmol/liter in 17.6% of the participants. In groups with increasing mean 25(OH)D of less than 25, 25-50, 50-75, and more than 75 nmol/liter, mean serum PTH, serum OC, and urine DPD/Cr were decreasing.. All bone density values were higher in the higher serum 25(OH)D groups than in the lowest serum 25(OH)D group, although this was only significant for the total hip, femoral trochanter, and total body BMC after all adjustments. In univariate analysis, differences in serum OC, BUA, SOS, and femoral neck BMD between lowest and highest serum 25(OH)D groups were also significant (data not shown), but differences disappeared after adjustment for potential confounders. Physical performance was low in the lower 25(OH)D level groups.
The relationship between 25(OH)D and some outcome parameters is presented in Fig. 1. LOESS plots with 95% confidence intervals, adjusted for age, sex, season of vitamin D determination, level of education, level of urbanization, BMI, number of chronic diseases, level of Cr, and smoking show the values of serum PTH, bone turnover markers, BMD of the total hip and femoral trochanter, and physical performance for each value of serum 25(OH)D. For higher levels of 25(OH)D, lower levels of serum PTH were observed up to values of 100 nmol/liter and higher, without reaching a plateau (Fig. 1A). Although only highly significant for the relationship between serum 25(OH)D and urine DPD/Cr, LOESS plots of serum 25(OH)D and both bone markers showed a steep decrease up to the serum 25(OH)D level of about 40 nmol/liter, followed by a plateau (Fig. 1, B and C).
For the relationship between serum 25(OH)D and BMD of total hip and femoral trochanter, a threshold appeared to exist around the serum 25(OH)D level of 50 nmol/liter (Fig. 1, D and E). This applied also to the total body BMC (LOESS plot not shown, P = 0.09). The LOESS plot of the relationship between 25(OH)D and physical performance showed a similar threshold around the serum 25(OH)D value of 60 nmol/liter (Fig. 1F).
Subjects and Methods
The study was conducted within the LASA. LASA is an ongoing cohort study on predictors and consequences of changes in autonomy and well-being in the aging population in The Netherlands (14). The sampling and data collection procedures have been described in detail elsewhere (15). Briefly, a random sample of older men and women (aged 55-85 yr), stratified by age, sex, and urbanization, was drawn from the population registers of 11 municipalities in the areas in the west (Amsterdam and its vicinity), northeast (Zwolle and vicinity), and south (Oss and vicinity) of The Netherlands. In total, 3107 subjects participated in this baseline examination (1992-1993).
The present study included participants who were born in or before 1930 (aged 65 yr and older as of January 1, 1996) and participated in a follow-up examination that took place in 1995-1996. After a main interview and a medical interview at home (n = 1509), participants were invited to the VU University Medical Center (respondents living in Amsterdam) or a health care center (respondents living in Zwolle or Oss), where blood and urine samples were obtained in 1328 persons. Serum levels of 25(OH)D, PTH, and bone markers were determined in 1319 persons (final study sample). BMD measurements of the lumbar spine and hip were obtained in a subsample including 517 participants in Amsterdam. All interviews were conducted by intensively trained and supervised interviewers, and were tape-recorded to monitor the quality of the data. Informed consent was obtained from all respondents. The study was approved by the Medical Ethics Committee of the VU University Medical Center.
Blood samples were obtained in the morning and immediately centrifuged and frozen. Participants were allowed to have tea and toast, but no dairy products. The participants also brought an overnight urine. The serum and urine samples were stored at -20 C. Serum 25(OH)D was determined according to a competitive protein binding assay (Nichols Diagnostics, San Juan Capistrano, CA). The interassay coefficient of variation (CV) was 10%.. PTH was measured by means of immunoradiometric assay (Incstar Corp., Stillwater, MN), with an interassay CV of 12%. Serum osteocalcin (OC) was measured by an immunoradiometric assay (Biosource/Medgenix Diagnostics, Fleuris, Belgium) with an interassay CV of 8%. Urinary deoxypyridinoline (DPD) was measured by a competitive luminescence immunoassay (ACS180 System; Bayer Diagnostics, Mijdrecht, The Netherlands) with an interassay CV of 8%. Values were corrected for creatinine (Cr) concentration in the same urine sample. Serum Cr level was measured using the Jaffe alkaline picrate reaction with a Hitachi 747 analyser and was included as a marker for renal function. The analyses were carried out at the Endocrine Laboratory of the VU University Medical Center.
Assessment of ultrasound and BMD measurements
Quantitative ultrasound measurements were measured at the calcaneus with the CUBA clinical instrument (McCue Ultrasonics, Winchester, UK). The ultrasound system consisted of two transducers (emitting and receiving) faced with silicone rubber coupling pads. These were placed in direct contact on either side of the heel using a coupling gel. Broadband ultrasound attenuation (BUA; dB/MHz) and speed of sound (SOS; m/sec) were measured twice in both the right and left calcaneus. The feet were repositioned after the first measurement. Mean BUA and SOS were calculated from these four measurements. The CV, calculated in 20 healthy volunteers measured on five occasions consecutively within 1 h, was 3.4% for BUA and 1.3% for SOS (16).
The dual-energy x-ray absorptiometry scans were made at the Department of Nuclear Medicine, using a Hologic QDR 2000 scanner (Hologic Inc., Waltham, MA). The lumbar spine and right hip were scanned. Total body bone mineral content (BMC; g) and BMD (g/cm2) of the femoral neck, trochanter, and total hip were measured.
Physical performance was assessed by three tests including the walk test (time needed to walk 3 m, turn around, and walk back), five chair stands (time needed to stand up and sit down five times with arms folded), and the tandem stand (ability to stand with one foot behind the other in a straight line for at least 10 sec) (17). For the walk test and chair stands, 1 to 4 points were given, corresponding to the quartile of the distribution of time needed. The more time was needed, the lower the physical performance scores. Participants who did not complete the test were given a score of 0. For the tandem stand, zero points were given to those who could not perform the tandem stand, 2 points to those who stood less than 10 sec, and 4 points to those who stood at least 10 sec. The three items were summed up to a final score ranging from 0-12 points.
Potential confounders and effect modifiers
Age, sex, season of vitamin D determination, level of education, level of urbanization, body mass index (BMI), number of chronic diseases, level of Cr, and smoking were considered as potential confounders. Because vitamin D status is partly dependent on sunlight exposure, we adjusted for the season of data collection. Four periods during the year were distinguished: January to March, April to June, July to September, and October to December. Level of education was assessed by asking the respondent for the highest educational level completed, ranging from primary school to university. Body weight (kilograms) was measured in subjects wearing underwear only, using a calibrated balance beam scale. Height (meters) was measured using a stadiometer. BMI was calculated by weight divided by height squared (kilograms/meter2). The presence of chronic diseases was assessed with a detailed questionnaire that included chronic obstructive pulmonary disease, cardiovascular disease, stroke, peripheral arterial disease, diabetes mellitus, malignant neoplasms, and joint disorders including osteoarthritis and rheumatoid arthritis (18). Mild renal impairment could affect vitamin D/PTH homeostasis; therefore, we also adjusted for serum Cr levels. Smoking status was classified as smoker or nonsmoker (cigarettes, cigar, and pipe).
Baseline characteristics of the population were calculated for the total sample and for men and women separately. Differences between the groups were tested by Studentfs t test for normally distributed variables and by Mann-Whitneyfs test for variables with a skewed distribution. Differences in proportions were tested by the X2 test. We used analysis of covariance, adjusted for potential confounders, to examine differences in mean values of the outcome measures across four categories of serum 25(OH)D (<25, 25-50, 50-75, and >75). Subsequently, locally weighted regression smoothing (LOESS) plots with 95% confidence intervals were performed in S-Plus (MathSoft, Inc., Seattle, WA) to investigate the relationship between serum 25(OH)D and various outcome measures. First, unadjusted analyses were performed. Subsequently, potential confounders were added to the models.