Vitamin D and calcium are both required for maintenance of musculoskeletal health, among others.1 Furthermore, the adequate nutritional intake of both calcium and vitamin D is also the first step in the management of patients with osteoporosis,2, 3 the most prevalent disease in the western hemisphere that renders bone brittle and prone to fracture.4, 5 The mechanisms by which calcium and vitamin D exert their positive effects on bone were initially centered primarily on their role in offsetting calcium losses from the body by increasing the intestinal absorption and its efficiency, respectively, thereby preventing bone from being broken down to scavenge its calcium for protection of the body's serum calcium level. Additional mechanisms in the bone protection repertoire of this couple that are of similar importance include reduction of excessive bone remodeling that increases bone fragility,6–8 reduction of fracture rates when calcium and vitamin D act together,9 and improvement in neuromuscular function by vitamin D that leads to a reduction in falls.10–13
Given the obvious paramount importance of calcium and vitamin D, it is surprising that although circulating 25-hydroxyvitamin D [25(OH)D] is regarded as the best predictor of the vitamin D status, there is no general consensus yet as to what 25(OH)D levels should be achieved.1 The estimated required circulating 25(OH)D to maintain skeletal health ranges between 30 and 100 nmol/L (12 and 40 ng/mL).14 This wide range is based on many studies that have used primarily the relationship between low 25(OH)D and increased secretion of parathyroid hormone (PTH) to determine a plausible threshold.15–17 PTH, however, is only one measurable index of skeletal health, and we reasoned that histomorphometric analysis of iliac crest bone biopsies with a strong focus on the presence or absence of mineralization defects (i.e., increased osteoid indices and osteomalacia) would be another and even more direct approach to assess bone health and address the required minimum 25(OH)D level.
We reasoned that a cohort from the German population with its known high prevalence of vitamin D deficiency18, 19 would be ideal to address the latter question using osteoid indices as a primary readout. Furthermore, a study design with inclusion of patients throughout the entire year also would allow us to take expected seasonal changes of 25(OH)D serum levels through different cutaneous synthesis after solar exposure into account. Thus, by analyzing 25(OH)D serum levels and transiliac crest bone biopsies from 675 patients, including individuals from the third to the tenth decades of life, in all of whom diseases causing secondary bone disorders were excluded, we confirm the high prevalence of vitamin D deficiency in northern Germany. Moreover, here we show that manifest mineralization defects of bone are present in some 20% of the population analyzed, and most important, the latter mineralization defects were absent in all individuals who presented with 25(OH)D serum levels above 75 nmol/L. Thus these data strongly argue in favor of a minimum 25(OH)D serum level of at least 75 nmol/L (30 ng/mL) to guarantee bone health.
Materials and Methods
Transiliac bone specimens were obtained from 675 individuals (401 males, mean age 58.72 ± 16.99 years, and 274 females, mean age: 68.26 ± 17.27 years) (see Fig. 2A). Age distribution is given in Fig. 1A. All specimens were taken during outright autopsies at the Department of Legal Medicine, University Medical Center Hamburg-Eppendorf, Germany. In fact, full autopsy allowed excepting all individuals suffering from cancer, renal diseases, primary hyperparathyroidism, and Paget's disease or showing any other circumstances, such as immobilization or hospitalization, potentially leading to secondary bone diseases from the study. The different circumstances leading to death were motor vehicle or train accidents, assaults, suicides, and other unnatural or unexpected causes. In all cases, bone specimens were obtained within 48 hours of death. An adequate representation of various possible influences to health, such as age and socioeconomic background, was guaranteed because of the high number of individuals.
This study was approved by the Medical Association Ethics Committee of the State of Hamburg (OB-024/05).
The transiliac crest biopsies were taken at the standard site for bone biopsies, 2 cm below the iliac crest and 2 cm behind the anterosuperior iliac spine, as initially described by Bordier.20 The samples consisted of outer and inner cortices and the intervening cancellous bone. Specimens were fixed in 3.7% formaldehyde for 3 days, dehydrated and embedded undecalcified in methyl methacrylate,21 and cut on a Microtec rotation microtome (Techno-Med GmbH, Munich, Germany). Sections of 4 µm were subsequently stained by toluidine blue, trichrome Goldner, and a combination of the von Kossa and van Gieson method, thereby staining mineralized bone matrix in black and nonmineralized osteoid in red.
Histologic and histomorphometric analysis
Histologic and histomorphometric analysis was performed using a Zeiss microscope (Carl Zeiss Vision GmbH, Germany) and the semiautomatic image-analyzing computer programs OsteoMeasure (Osteometrics, Inc., Atlanta, Georgia) and Bioquant (Bioquant Image Analysis, Inc., Nashville, TN) according to ASBMR standards.22 Bone volume per trabecular volume (BV/TV [%]), trabecular number (Tb.N. [/mm]), trabecular thickness (Tb.Th. [µm]), and trabecular separation (Tb.Sp. [µm]) and osteoid volume per bone volume (OV/BV [%]) and osteoid surface per bone surface (OS/BS [%]) were measured.
Beside bone biopsies, blood samples also were taken at autopsy. The blood was centrifuged the same day, and the serum gained was stored at −80°C for further analysis. Serum concentrations of 25(OH)D were measured for all patients at the Department of Clinical Chemistry, University Medical Center Hamburg-Eppendorf, Germany, using a radioimmunoassay (DiaSorin, Stillwater, MN) with an interassay coefficient of variation of between 8.2% and 11%. The sensitivity of this method is 1.5 µg/L. Unlike PTH or calcium, for instance, 25(OH)D was found to be stable in various experiments for at least 10 days postmortem.
Statistical analysis was performed using the unpaired Student's t test because a Gaussian distribution was granted. Levels of significance were defined as p ≤ .05 → *, p ≤ .005 → **, and p ≤ .0005 → ***.
BV/TV decreased significantly with increasing age beginning from the third decade of life. Interestingly, there were three significant steps of decline every two decades starting at the fourth (Fig. 2B). The reduction of BV/TV was caused mainly by a decrease of Tb.N. rather than a decrease of Tb.Th. (see Fig. 2B). Tb.Th., on the other hand, was found to be fairly stable throughout the whole lifetime. As expected and in reverse to the correlation of Tb.N. and age, Tb.Sp. increased significantly with increasing decades of life (see Fig. 2B).
Beside mineralized bone matrix, we also examined the nonmineralized bone and found an unexpected high number of biopsies revealing mineralization defects, that is, OV/BV of greater than 2%. Of the cases examined, 25.63% showed high OV/BV of up to 17.44%, with a mean twice as high as a conservative lower level of pathologic osteoid accumulation (Fig. 3A–C). An OV/BV of greater than 5% was found in 4.89% of the cases (33 of 675), and in 1.04% of the cases (7 of 675), the osteoid volume was even greater than 10%. Lowering the threshold of pathologic lack of mineralization to 1.2%, as described by Delling in 1975,23 the fraction of osteomalacia increases up to 43.41%. Irrespective of whether one considers an osteoid volume of greater than 1.2%, of greater than 2.0%, or else of greater than 5% as the borderline toward manifest osteomalacia, mineralization defects were found independent of bone mass throughout all ages and both sexes, respectively. Furthermore, buried osteoid as a potential morphologic indicator of previous periods of low 25(OH)D serum levels was found in 2% of cases with an OV/BV of 2% or less and in 36% of the cases with an OV/BV of greater than 2% (see Fig. 3B, E). Moreover, 36.15% of the cases presented with a high OS/BS of greater than 20% and 23.29% of the cases with an OS/BS of greater than 25%, respectively. Regardless of the thresholds of 20% or 25%, again, no elevated osteoid surface was found when 25(OH)D exceeded 75 nmol/L (Fig. 4E). Finally, examining the increase in osteoid thickness (O.Th.), a cornerstone in the diagnosis of mineralization defects,24 we found a similar effect. When setting the upper limit of normal O.Th. to 12 µm, no pathologic cases were found with 25(OH)D levels above 75 nmol/L (see Fig. 4F).
As expected, the population of northern Germany presents with a high prevalence of vitamin D deficiency because unlike other countries of similar latitude, fortification of food with vitamin D is not allowed in Germany. Low levels of 25(OH)D were found independent of sex (see Fig. 4A) or age (see Fig. 4C) except for subtle seasonal changes (see Fig. 4B). Blood samples taken during the summer showed significantly higher levels of 25(OH)D than those taken in winter or spring, although the mean values for every season were far below 75 nmol/L (see Fig. 4B).
While we could not establish a minimum 25(OH)D level that inevitably was associated with manifest mineralization defectsm we did not find pathologic accumulation of osteoid in any patient with circulating 25(OH)D levels above 75nmol/L (30 ng/mL). Therefore, our combined analysis of vitamin D levels and histomorphometric data provides direct evidence that the bone tissue level of circulating 25(OH)D should reach a minimum threshold of 75 nmol/L (30 ng/mL) to ensure bone health.
There is no defined threshold yet as to the minimum 25(OH)D serum level that guarantees skeletal health.25 Current estimates are based largely on the observation of secondary hyperparathyroidism in vitamin D deficiency through the use of serum PTH levels as a surrogate parameter for the optimal 25(OH)D serum level. However, the various studies show different threshold 25(OH)D levels ranging from 30 to 100 nmol/L (12 to 40 ng/mL) as where secondary hyperparathyroidism starts. Another surrogate parameter for determination of the optimal 25(OH)D level that has been used is bone mineral density (BMD) because decreased BMD values are associated with low 25(OH)D levels. However, depending on the study and on the population included, the 25(OH)D threshold to low BMD varied from 30 nmol/L26, 27 to 75 nmol/L.28 There is, however, expert consensus that 25(OH)D should be above 50 nmol/L (20 ng/mL) in all individuals,29, 30 whereas some experts advocate a mean population value of 75 nmol/L (30 ng/mL) or greater.14 Interestingly, a large population-based regression analysis confirmed that higher 25(OH)D levels were associated with higher BMD values throughout the entire range from 22.5 to 92.6 nmol/L (9 to 37 ng/mL) in all subgroups.28 Therefore, we reasoned that the direct assessment of bone mineralization at the tissue level in correlation with 25(OH)D serum levels would be another and maybe even better surrogate parameter to assess vitamin D status and thus might help to harmonize the consensus on the 25(OH)D serum level required for skeletal health. Bone mineralization defects are best quantified by histomorphometry. However, this requires bone biopsies and undecalcified processing of the latter. Since performing an iliac crest biopsy is a rather invasive diagnostic technique, at least as compared with DXA osteodensitometry, a population-based study, such as, for example, Third National Health and Nutrition Examination Survey (NHANES III) with 13,432 individuals included, seemed impossible. Thus it was a prerequisite that such a study be carried out in a population with a known endemic vitamin D deficiency, as is Germany.18, 19 Furthermore, another obstacle for a biopsy study is that, in general, iliac crest bone biopsies are taken only from patients who present with suspicious processes indicative of bone diseases. However, to establish the effect of various 25(OH)D serum levels on bone mineralization, other confounding factors such as primary hyperparathyroidism,31 renal failure,32 renal phosphate leak,33 and Paget's disease34 that might cause hyperosteoidosis have to be minimized, and thus only patients without diseases affecting the skeleton should be included. Therefore, we identified forensic autopsy cases to be the best population for this study because this allowed us to include a large number of individuals of all ages and both sexes. Most important, autopsy cases allowed us to exclude any individual who presented with diseases known to affect the skeleton.
There also are, however, certain clear limitations of our study that need to be considered. First, tetracycline labeling, as one gold standard for quantification of bone formation, is missing. Second, serum analysis is limited to the assessment of 25(OH)D, whereas other laboratory data, such as serum calcium, phosphate, creatinine, PTH, and alkaline phosphatase levels, are missing owing to a lack of stability, and thus some circumstances that potentially interfere with bone mineralization, such as moderate renal dysfunction or mild primary hyperparathyroidism, might remain undetected at the tissue level. And third, one could argue that the cohort studied here might not represent a population sample because the mean age in this study was 58.72 ± 16.99 years for males and 68.26 ± 17.27 years for females, and thus the cohort seems to be rather young. However, since 25(OH)D levels generally are considered to be somewhat lower in the elderly, this should not be a major bias for interpretation of this study's results.
Compared with some previous reports on the 25(OH)D serum levels in Germany,35–37 the 675 individuals from Hamburg studied here seem to have, on average, very low vitamin D serum levels at first glance. One has to take into account, however, that the 25(OH)D levels in the 476 patients studied by Woitge and colleagues do not represent the German population because they were all recruited from the southwestern part of Germany, whereas Hamburg represents the northern part of Germany. Kuchuk and colleagues studied 7455 patients in a large international study, but the data for Germany are based on 66 patients only, and moreover, patients from northern Germany were specifically excluded from the study.
Finally, the data on 25(OH)D levels in Germany provided by Paul Lips and colleagues are based on only 31 osteoporotic females who were included into the Multiple Outcomes of Raloxifene Evaluation (MORE) trial in Germany. However, there are two large recent studies from the Robert Koch Institute, the German National Health Interview and Examination Survey (GNHIES, 4030 patients) and the German National Health Interview and Examination Survey for Children and Adolescents (KIGGS, 10,015 patients), that study the vitamin D status in a population-based sample. For both cohorts, Hintzpeter and colleagues18, 19 report an endemic vitamin D deficiency, and the, on average, slightly higher 25(OH)D levels as compared with our cohort are most likely explained by the fact that patients from all parts of Germany were included, whereas our patients were recruited from northern Germany only.
Histologic and structural histomorphometric analysis revealed that there is a gradual decline in bone volume within the iliac crest during aging. Structural analysis demonstrates that this age-dependent bone loss is mainly due to ae reduction in structural elements, as represented by the significant decrease in trabecular number. In contrast, trabecular thickness remains relatively constant during aging, with a mean trabecular thickness of approximately 120 µm throughout life. This is in line with previous histomorphometric studies of human bone samples performed at other skeletal sites, including the spine.38 A clear strength of this study is that the sampling site within the iliac crest was highly standardized, thus minimizing a bias through the known structural heterogeneity within the iliac crest itself39 and limiting the structural differences detected to inter-individual differences. Even more striking than the results of structural element analysis is the quantification of osteoid indices. An osteoid volume above 2% (up to 17.4%) was detected in 25.63% of the 675 biopsies, and in line with the latter finding, surface osteoidosis with an OS/BS of 20% and above was present in 36.15% of the biopsies. This is indeed remarkable because an osteoid volume of above 1.2% has been defined previously as the pathologic threshold to osteomalacia.23
Since the latter mineralization defects were found independent of bone mass throughout all ages and affected both sexes, it was of interest to correlate the osteoid indices with the vitamin D status of the patients. Therefore, 25(OH)D was measured in all patients because it is regarded as the best predictor of the vitamin D status.1 Our results on vitamin D show low levels of 25(OH)D independent of age and sex, thereby confirming that vitamin D deficiency is a major public health issue in Germany.18 Indeed, the GNHIES, which included 7124 men and women aged 18 to 79 years, a population sample reflecting the noninstitutionalized adult population in Germany, has documented the extent of vitamin D deficiency in Germany, and this was further confirmed by the KiGGS survey of children and adolescence with a prevalence of 25(OH)D serum levels below 75 nmol/L in 90% of the 3- to 17-year-old population.19 There are essentially at least two reasons for these low 25(OH)D serum levels. The first is that although there are seasonal variations, with the highest 25(OH)D levels in the summer, overall solar exposure is not sufficient to synthesize 25(OH)D from 7-dehydrocholesterol in Germany. This is understandable if one remembers that Webb and colleagues have shown an inadequate sun irradiation to generate vitamin D in Boston (latitude 42°N),40 which is about the same latitude as Rome in Europe (41°N), whereas, in fact, Hamburg (latitude 53°N) is as far north as is Edmonton, with a greatly diminished ultraviolet spectrum. Thus, even in summer, there is insufficient solar exposure in Germany to synthesis enough 25(OH)D in the skin. The other reason is that in Germany, unlike the situation in the United States, Great Britain, and the Scandinavian countries, it is largely prohibited to fortify food with vitamin D.
What is a major health problem in Germany is an advantage for the question posed in our study because the aim was to detect vitamin D deficiency–associated mineralization defects. The mineralization defects that were detected ranged from surface osteoidosis, to volume osteoidosis, to combined surface and volume osteoidosis, to buried osteoid, indicating variations in calcium metabolism and mineralization status. For the sake of clarity, we chose to use volume osteoidosis with an OV/BV of above 2% as the most conservative threshold to a pathologic mineralization status, that is, osteomalacia. Importantly, we did not find pathologic accumulation of osteoid in any patient with circulating 25(OH)D above 75 nmol/L (30 ng/mL). Interestingly, the same threshold of 75 nmol/L (30 ng/mL) has been put forward by results of some studies using PTH as a surrogate parameter for skeletal health. In contrast, we could not establish a minimum 25(OH)D level that inevitably was associated with mineralization defects, an observation that is completely in line with the finding from cross-sectional studies that have shown secondary hyperparathyroidism in many, but not all, vitamin D–deficient individuals using PTH as a surrogate parameter to define variable 25(OH)D level thresholds.25 As a limitation of our study design, we were not able to comeasure PTH, which, unlike 25(OH)D, is unstable.
Taken together, our histomorphometric bone analysis supports experts who argue in favor of 25(OH)D levels greater than 75 nmol/L (30 ng/mL),1, 14 and the data presented here strongly suggest that indeed a minimum 25(OH)D serum level of 75 nmol/L (30 ng/mL) in conjunction with a sufficient calcium intake should be ensured to guarantee skeletal health at the tissue level. Especially osteoporotic patients with insufficient calcium intake and/or vitamin D insufficiency require adequate calcium and vitamin D supplementation in combination with their anabolic or antiresorptive prescription medicine.
The authors state that they have no conflicts of interest.