Black women have a lower incidence of vertebral and hip fractures than white women, possibly due to differences in skeletal and mineral metabolism. One suggested mechanism is that blacks have decreased skeletal sensitivity to parathyroid hormone (PTH). To test this hypothesis, we infused h(1–34)PTH in healthy premenopausal black (n = 15) and white (n = 18) women over 24 h and measured serum and urine indices of bone turnover and calcium metabolism throughout the infusion. At baseline, the mean 25-hydroxyvitamin D (25(OH)D) concentration was significantly lower in black women (46%). There were also nearly significant trends toward higher PTH and lower urinary calcium and pyridinoline levels in black women. During infusion, there were no racial differences in the mean (1–34)PTH levels achieved or in resultant elevations of serum calcium or 1,25-dihydroxyvitamin D (1,25(OH)2D) levels. Endogenous parathyroid suppression (measured by (1–84)PTH levels) was also similar between blacks and whites. There was an initial decline in urinary calcium/creatinine in both groups with a greater reduction in black women early in the infusion period (p < 0.05 at 8 h). Furthermore, blacks had lower levels of urinary calcium/creatinine throughout the infusion (p < 0.05 group difference). Bone formation markers (carboxy-terminal propeptide of type I procollagen and osteocalcin) decreased within 8 h and continued to decline throughout the infusion with no distinguishable racial differences (p < 0.05 time trend for both). The most dramatic difference between black and white women in response to PTH infusion was represented by the bone resorption markers. Three separate metabolites of bone resorption (cross-linked N-telopeptide of type I collagen, cross-linked C-telopeptide of type I collagen, and free pyridinoline) all showed substantially greater elevations in white (mean peak increments 399, 725, and 43%) compared with black women (mean peak increments 317, 369, and 17%) during the infusion (p < 0.05 group differences for all three variables). These data strongly suggest that blacks have decreased skeletal sensitivity to the acute resorptive effects of increased PTH. This finding indicates that calcium homeostasis may be accomplished in blacks (during times of relative calcium deficiency) by greater conservation of calcium from nonskeletal sources (most likely renal) with relative preservation of skeletal tissue. These differences in calcium economy could account, at least in part, for the increased bone mass and lower incidence of osteoporotic fractures in black women.
Racial differences in the incidence of osteoporotic fractures are well established, with blacks having a lower incidence of vertebral compression fractures(1) and hip fractures(2–7) than whites. This difference has been attributed at least in part to a higher bone mineral density in blacks.(8–15) Investigations into the etiology of the racial disparity in bone mass have revealed differences in calcium metabolism and bone turnover, with the most consistent finding being lower urinary calcium excretion in blacks compared with whites.(13–19) In addition, numerous studies have documented lower 25-hydroxyvitamin D (25(OH)D) levels in black women(15,17–23) with higher parathyroid hormone (PTH) levels also reported in several studies.(15,17,20) Lower levels of osteocalcin,(14,15,17) bone-specific alkaline phosphatase,(15) and urinary hydroxyproline(14,15) suggest that black women might have lower bone turnover despite higher PTH levels. Consequently, it was theorized, initially by Bell et al.,(17) that blacks have decreased skeletal sensitivity to PTH. In the present investigation, this hypothesis was tested directly by infusing h(1–34)PTH and evaluating the skeletal response, using biochemical remodeling indices, in comparable groups of healthy black and white premenopausal women.
MATERIALS AND METHODS
Healthy premenopausal black (n = 15) and white (n = 18) women between the ages of 25 and 40 years were recruited from New York City, and Rockland and Orange Counties (NY, U.S.A.), through poster, newspaper, radio, and television advertisements. Participants were required to have a normal menstrual history and be in good health as assessed by medical history, physical examination, electrocardiogram, blood count, and serum chemistries, including liver and thyroid function tests. Women were required to be within 80–120% of ideal body weight for height and frame size as determined by the 1983 Metropolitan Life Insurance tables.(24) Potential subjects with medical conditions or drug exposure known to affect bone metabolism were excluded. Specifically, women with a recent history (within 6 months prior to participation) or current use of oral contraceptive or corticosteroid medications were excluded. No subjects were on thyroid hormone or calcium supplements. All patients gave written informed consent, and the study was approved by the Institutional Review Board of Helen Hayes Hospital.
Infusions were performed during the follicular phase of the menstrual cycle (within 2 weeks of the last menstrual period). Between 3:00 and 4:00 p.m., two baseline blood samples and one urine sample were obtained. Beginning at 4:00 p.m., synthetic h(1–34)PTH (Rhone-Poulenc Rorer, Horsham, PA, U.S.A.) dissolved in normal saline was infused continuously by indwelling intravenous catheter at a rate of 0.55U/kg/h for a period of 24 h as previously described.(25,26) The women maintained light activity and ingested no more than 600 mg of calcium and 1000 mg of phosphorus throughout the investigation, as ordered by a nutritionist. Blood and urine samples were collected at 4-h intervals for a total of 24 h, after which the intravenous catheters were removed and the patients discharged home.
Circadian rhythm assessment protocol
To assure that changes in bone turnover markers were due to PTH infusion rather than the underlying diurnal rhythm of bone remodeling, a subgroup of black (n = 5) and white (n = 4) women who underwent the PTH infusion were also admitted for serial blood and urine sampling (4 hourly) over a 24-h period using an otherwise identical protocol but without infusion. These investigations were also performed during the follicular phase of the menstrual cycle and at least 8 weeks after the initial PTH infusion procedure.
Serum (1–34)PTH was analyzed using chicken antibody CK67 in a radioimmunoassay(27) and (1–84)PTH by immunoradiometric assay (Allegro Intact PTH, Nichols Institute, San Juan Capistrano, CA, U.S.A.). Serum 25(OH)D and 1,25-dihydroxyvitamin D (1,25(OH)2D) were analyzed by competitive protein binding and radioreceptor assays.(28,29) Serum ionized calcium was analyzed by standard techniques (NOVA 8 ionized calcium analyzer, Newton, MA, U.S.A.). Quality control information for these assays have been published previously.(25,26)
Serum samples were also obtained for analysis of bone formation markers, including osteocalcin (BGP) by immunoradiometric assay (human osteocalcin, Nichols Institute, CA) and carboxyterminal propeptide of type I procollagen (PICP) by radioimmunoassay (INCSTAR Corp., Stillwater, MN, U.S.A.). The minimum detectable concentrations were 0.01 nmol/l for BGP and 1.2 ng/ml for PICP. Interassay and intra-assay coefficients of variation were 10.6 and 6.1% for BGP, and 8.6 and 4.9% for PICP.
Urine calcium and creatinine were analyzed by standard techniques. Calcium and creatinine clearance were calculated using standard formulas.
Three urine bone resorption markers were all measured by ELISA assays, with results corrected for urine creatinine: free pyridinoline (PYD/Cr, Pyrilinks, Metra Biosystems, Mountain View, CA, U.S.A.)(30); cross-linked N-telopeptide of type 1 collagen (NTX/Cr, Osteomark, Ostex Int., Seattle, WA)(31); and cross-linked C-telopeptide of Type I collagen (CTX/Cr, Crosslaps, Osteometer, Glerupvej, Denmark).(32) The interassay and intraassay coefficients of variation were 13.4 and 2.6%; 9.2 and 6.1%; 9.2 and 6.2%, respectively.
All statistical analyses were performed using the SAS statistical program (SAS Institute, Cary, NC, U.S.A.). The mean and standard error are reported for each baseline variable. Unpaired Student's t-tests were used to compare baseline values between black and white women for normally distributed variables and nonparametric tests used for variables that were not normally distributed. Repeated measures analysis of variance was used to determine significance of time trends, overall group differences, differences in incremental or decremental changes between groups at specific time points, and group differences in the temporal course of responses (time/group interactions). To determine whether baseline biochemical levels could account in part for skeletal responsivity, Pearson correlation coefficients were determined between baseline variables and peak increments in skeletal resorption indices.
Table Table 1. Characteristics of Study Populations Mean ± SEM
Descriptive characteristics of the subjects are shown in Table 1. There were no significant differences between the two groups in age, height, educational level, or parity. Mean body weight and body mass index of the black women were slightly higher than those of the white women (p < 0.05 for the latter variable). Routine chemistry, thyroid, gonadal, and hematology profiles were normal in all women and similar between the groups (data not shown). Characteristics of the subgroups of black and white women admitted for assessment of diurnal rhythmic variation in bone turnover markers did not differ significantly from their respective total groups or from each other (data not shown).
For each patient, two basal values were averaged (Table 1) to minimize fluctuation due to pulsatile hormone secretion. The major difference between the groups was a significantly lower mean basal 25(OH)D level in black women (46% difference, p = 0.006). In general, mean basal (1–84)PTH was lower in whites, and mean urinary calcium/creatinine lower in blacks, but these differences did not quite meet statistical significance (p < 0.07). In general, mean indices of bone formation and resorption were similar between the two infusion groups and between the two control subgroups (data not shown), although PYD was slightly lower in black women (p < 0.07).
Effects of PTH infusion
During the infusion, serum concentrations of PTH(1–34) rose promptly (Fig. 1), with peak concentrations of 16.62 ± 2.6 pmol/l in white women and 18.67 ± 2.1 pmol/l in black women, with no significant racial differences. Levels remained elevated throughout the infusion. Serum ionized calcium rose in both groups within 4 h and continued to rise throughout the investigation with no differences between black and white women. In both groups, endogenous PTH(1–84) declined significantly below baseline within 4 h, with continued suppression throughout the study. Serum 1,25(OH)2D levels increased significantly in both groups within 8 h and continued to rise through 20 and 24 h in blacks and whites, respectively. There were no racial differences. Serum 25(OH)D levels remained unchanged throughout the infusion in both racial groups (data not shown).
Significant time trends were seen in both groups for urinary calcium/creatinine (UCa/Cr, p < 0.01). Levels decreased below baseline in black and white women for at least 16 h, but within 20 h UCa/Cr levels were elevated at which time mean serum calcium levels exceeded 1.39 mmol/l in both groups. Throughout the investigation, UCa/Cr levels were lower in black than in white women with an overall group difference (p < 0.05). Decrements from baseline differed at the 8 h time with black women showing greater renal calcium conservation than white women (p < 0.05). Analysis of calcium clearance corrected for creatinine clearance revealed the same results as UCa/Cr (data not shown).
Bone formation markers:
In control patients undergoing the serum sampling protocol for assessment of diurnal rhythms in biomarkers (with no infusion), there were slight fluctuations (maximum of approximately 10%) in both serum BGP and PICP levels (Fig. 2), over the 24 h period. There were no significant differences between groups for either of these variables under control conditions.
With PTH infusion, both serum BGP, and even more markedly, serum PICP, decreased in both groups within 8 h and declined to nadir levels at 20 h (time trends p < 0.05). Group responses were very similar.
Bone resorption markers:
All resorption markers showed evidence of a natural diurnal rhythm, with peak levels at 8 a.m. and nadir levels at 4 p.m. to midnight (Fig. 3). Black and white women did not differ significantly in this circadian pattern or the degree of variation in resorption marker levels over 24 h.
In response to PTH infusion, NTX/Cr and CTX/Cr increased in both white and black women (time trend p < 0.005; Fig. 3). Responses of these peptide-bound pyridinoline markers were substantially higher, however, in white compared with black women (p < 0.003 group difference for both NTX/Cr and CTX/Cr). Mean peak NTX/Cr in white women was 112 ± 16 versus 74 ± 8 nMBCE/mMCr in black women, corresponding to 399 and 317% increments over baseline, respectively. Substantial differences in mean peak levels of CTX/Cr were also seen (in white, 448 ± 24 vs. 303 ± 12 μg/l/mMCr in black, or 725 and 369% increments over baseline, respectively). Incremental changes at multiple specific times were also higher in white than black women, and an overall racial difference in the temporal course of response (time/group interaction) was seen for CTX/Cr (p < 0.02).
Free PYD/Cr increased above diurnal rhythm variation only in white women with a peak increment of 43% over baseline. In black women, PYD/Cr did increase with time (mean peak 17%) but this increase did not differ from diurnal variation. Both an overall group difference (p < 0.003) and a racial difference in the temporal course of response (time/group interaction, p < 0.07) were seen for this variable.
The relationship between free PYD/Cr and peptide-bound pyridinolines (NTX/Cr and CTX/Cr) differed depending on the total level of bone resorption. Figure 4 illustrates that, in both races, the ratio of free PYD/Cr to total measured pyridinolines in urine (free + peptide bound) decreases substantially as total pyridinolines increase. There were no racial differences in the relationship between free and peptide-bound pyridinolines.
Interactions between basal vitamin D status and bone resorption response:
Linear correlations were determined between baseline 25(OH)D and peak resorption marker increments as well as baseline PTH and peak resorption increments. In black women, there were no significant or even close to significant relationships. In white women, there were weak positive correlations between 25(OH)D and resorption marker increments and weak negative correlations between PTH and resorption marker increments. None of these was statistically significant. In black and white women combined, correlation coefficients between 25(OH)D and resorption marker increments varied from 0.30 for PYD/Cr to 0.38 for NTX/Cr (p < 0.03 for NTX/Cr, p < 0.09 for PYD/Cr and CTX/Cr). Relationships between basal PTH levels and peak resorption marker increments were somewhat weaker. Only the correlation with the peak CTX/Cr increment and basal PTH was close to significant (r = −0.33, p = 0.08 vs. r = −0.13 for PYD/Cr and −0.14 for NTX/Cr).
The influence of baseline 25(OH)D was also investigated by analyzing the NTX/Cr and CTX/Cr responses to PTH infusion with 25(OH)D as an independent variable. The racial group difference persisted for CTX/Cr (p < 0.003) and just missed significance for NTX/Cr (p = 0.10). There was a significant time/race interaction term for both CTX/Cr (p = 0.007) and NTX/Cr (p = 0.043) with 25(OH)D considered as a covariate.
To elucidate further the possible influence of basal vitamin D status on resorption response to PTH infusion, subgroups of subjects were identified whose 25(OH)D levels could be matched with those from the other racial group. Due to the large racial difference in 25(OH)D, only five women within each race had comparable 25(OH)D levels, with resultant mean levels of 50.9 nmol/l in white and 50.3 nmol/l in black women. The response of NTX/Cr to PTH was elevated in the white subgroup compared with the black subgroup to the same extent as that seen for the whole groups. Mean peak NTX/Cr was 122 ± 24 nMBCE/mMCr in whites (407% elevation) and 72 ± 16 nMBCE/Cr in blacks (313% elevation; group difference, p < 0.058). Similarly, the CTX/Cr response was greater in these white (peak 429 ± 49 μg/l/mMCr) versus black (peak 321 ± 42) women (group difference, p = 0.032). As in the larger groups, a slightly greater PYD/Cr increment was also seen in the subgroup of white women compared with the black subgroup.
Based on previous findings of higher PTH with lower levels of bone turnover markers (hydroxyproline and osteocalcin) in black women, it was hypothesized that black women might be resistant to the skeletal effects of parathyroid hormone.(17) Similar racial differences observed in more recent investigations(14,15,20) supported this hypothesis. We now present dynamic data providing the first direct evidence that black women are less sensitive to the bone resorbing effects of acute increments in PTH, with substantially lower increments in biochemical markers of bone resorption in response to PTH infusion in black compared with white women. We also affirmed the finding of overall superior renal calcium conservation in black women(13–19) and demonstrated a somewhat greater renal calcium conservation response to PTH infusion, particularly in the early part of the protocol. Even small differences in renal calcium handling have the potential to result in significant effects on calcium homeostasis and help preserve the skeleton.
Our baseline data confirmed previous findings of significantly lower 25(OH)D levels in blacks compared with whites.(15,17–23) In our study, lack of significant differences in mean basal PTH, urinary calcium, and bone turnover variable levels between blacks and whites was probably due to our small sample size. This study was designed to investigate changes in response to dynamic perturbations and may not have been large enough to detect baseline differences.
Serum BGP and PICP, both markers of osteoblastic activity, have well documented circadian rhythms with a nocturnal rise and morning decline.(33–38) Control subjects of both races in this study exhibited little rhythmicity in either serum BGP or PICP. This might have been due to our small sample size and an overall smaller absolute change in levels expected with bone formation variables, in contrast to resorption variables. More frequent serum sampling might have made the diurnal variation in BGP and PICP more obvious. The circadian rhythm of the resorption markers,(39–41) with peak levels in the early morning and the nadir in late afternoon or early evening, was seen in our investigation and did not differ between blacks and whites.
In both blacks and whites, PTH infusion resulted in a small decline in BGP levels and a more marked decrement in serum PICP. Differences in responsivity of PICP versus BGP might be due to a variety of factors including expression at different points during osteoblast development, differential sensitivity of gene expression of the two molecules in response to PTH stimulation, or differential stability in the circulation and/or clearance rates of the two compounds. Decreases in bone formation variables have been seen with previous PTH challenge studies in humans,(42,43) as well as in the rat(44) and in vitro studies.(45) This acute suppression of bone formation contrasts with the effects of chronic endogenous elevations of PTH such as in primary hyperparathyroidism,(46) and with results of exogenous PTH administration by subcutaneous injection, where both bone formation and resorption variables are increased.(47)
Major racial differences in response to PTH infusion were seen with bone resorption indices where significantly higher elevations of all three variables (NTX/Cr, CTX/Cr, PYD/Cr) occurred in white compared with black women, despite the same increment in PTH levels. In black women, the lack of response of free PYD/Cr to infusion differed from the responses of the peptide-bound pyridinoline resorption markers, where increments above diurnal rhythm peaks were seen. Similarly, increments seen in white women were of much lower magnitude with PYD/Cr than with either NTX/Cr or CTX/Cr. Although all of these markers have validity as measures of bone resorption based on either histomorphometric or kinetic studies,(33,46) clearly they do not reflect exactly the same process, or are metabolized differently. Perhaps during acute substantial increments in bone resorption, less pyridinoline is excreted in its free form due to less complete collagen digestion at the osteoclast level or less time for peripheral metabolism. Proportionately more, at that time, would be liberated in the peptide-bound form. Consistent with this hypothesis is the finding that the fraction of free pyridinoline to total pyridinoline becomes smaller as total pyridinoline, and thus total resorption, becomes greater. This could explain why the free pyridinoline response in white women (43% peak elevation above baseline) was less robust than the peak N-telopeptide or C-telopeptide responses (399–725% for white women). If this is true, the significantly lower free pyridinoline response expected (in contrast to peptide-bound pyridinoline) in black women might have been difficult to see with this assay.
A traditional interpretation of the phenomenon of skeletal resistance to PTH in black women is that it developed as an adaptive response to help protect the skeleton from relative vitamin D deficiency. Indeed, in our larger study of normal premenopausal black and white women, 72% of blacks but only 11.5% of whites had 25(OH)D levels ≤37.5 nmol/l (unpublished data), consistent with data from NHANES, a large population-based study.(23) Our investigation suggests that these differences in basal vitamin D between black and white women contributed only weakly, if at all, to the difference in resorption response to PTH. We considered the possibility that, in black women, increased ambient levels of PTH in basal conditions (induced by low vitamin D) might result in down-regulation of PTH receptors. This down-regulation, in turn, could decrease skeletal responsivity to exogenously administered PTH. The lack of relationships between basal 25(OH)D and PTH with resorption marker increments in black women did not support this thesis. There was a suggestion that basal vitamin D status in white women might be associated with resorption response to exogenous PTH, suggesting perhaps a more important physiological role for 25(OH)D in white women. Even when the data from the black and white women were combined, the weak relationships detected between basal 25(OH)D and skeletal responsivity would explain only 10–15% of the variance in skeletal responsivity at best. This combined analysis, however, cannot eliminate the effect of race itself and does not therefore directly address the issue of the 25(OH)D level as a predictor of skeletal response to PTH. Furthermore, in the full racial groups, when 25(OH)D levels were considered as independent variables in the analysis of variance of the resorption marker response to PTH infusion, the racial differences persisted. Similarly, when we analyzed small subgroups of black and white women, who were matched by basal 25(OH)D levels, greater increments in resorption markers in response to PTH infusion were seen in white compared with black women, with differences of very similar magnitude to those seen in the full groups. This again indicates that racial differences in basal vitamin D make at most a minor contribution to skeletal resistance to PTH in black women.
Moreover, the concept of black skeletal physiology as an adaptation from white physiology must be questioned, given the superior skeletal mass in blacks compared with whites. Adaptations in medical physiology do not usually overcompensate for the original condition that caused them to occur. If PTH resistance developed only to preserve the skeleton, it is unlikely that the black skeleton could have become stronger than the white skeleton which did not require any adaptive changes in hormonal sensitivity. Therefore, it is likely that skeletal resistance to PTH is at least partially independent of the relative vitamin D deficiency in black women.
Racial differences in skeletal metabolism might be better viewed in terms of white women's physiology adapting from the black physiology. If white women were not as good at conserving calcium from extraskeletal sources (despite their higher mean vitamin D levels), they might have adapted by increasing skeletal sensitivity to PTH to increase the calcium supply from the skeleton in order to maintain calcium homeostasis. Preservation of skeletal mass might have been sacrificed in the process.
In conclusion, if calcium homeostasis can be maintained effectively in blacks through superior renal calcium conservation and perhaps increased gastrointestinal absorption (which has been seen in only one(22) of several studies,(21,48,49) with less calcium liberation from skeletal tissue, higher bone mass could be maintained. These racial differences in skeletal and mineral metabolism might therefore explain, at least in part, the lower incidence of osteoporosis in black women.
This work was supported in part by National Institutes of Health grants AR41386 and DK46381 and AR39191.