Thiazides and seasonal bone change in healthy postmenopausal women

Bone and Mineral, 21 (1993) 41-51 © 1993 Elsevier Scientific Publishers Ireland Ltd. All rights reserved. 0169-6009/93/$06.00

41

BAM 00544

Thiazides and seasonal bone change in healthy postmenopausal women Bess Dawson-Hughes and Susan Harris USDA Human Nutrition Research Center on Aging at Tufts University, 711 Washington St, Boston, MA 02111, USA (Received 20 July 1992) (Accepted 2 October 1992)

Summary Thiazide diuretic use has been associated with reduced rates of bone loss and bone density has been found to fluctuate with time of year. Here we examine associations between thiazide use, calcium regulating hormones, and the pattern of bone change during the 6-month intervals of summer/fall when bone density increases and winter/spring when bone density declines. Of the 246 postmenopausal women who completed a l-year calcium and vitamin D supplement trial, 25 used a thiazide diuretic. In the winter/spring, bone loss was reduced in the thiazide users (for L2-4:0.46 ± 0.59 (SE)% for thiazide users vs. -1.02.4- 0.17% for non-users, P = 0.017; for whole-body: -0.13 4- 0.25% vs. -0.67 4- 0.08, P = 0.043). The benefit at the whole body was dose related. Increases in bone density were similar in the two groups in the summer/fall. Associations between thiazide use and net spinal change and between thiazide dose and net whole-body bone change were positive. Thiazide users had lower serum levels of PTH (26.2 4- 7.6 (SD) ng/l vs. 31.0 4- 11.7 ng/l, P = 0.009) and 1,25-dihydroxyvitamin D (60.8 4- 16.2 pmol/l vs. 77.0 4- 18.7 pmol/l, P < 0.001) in the winter/spring and, in the subset measured, reduced levels of osteocalcin year round (summer/fall: 2.89 4- 0.82 ttg/l, n = 10, vs. 3.65 4- 1.02, n = 82, P = 0.019; winter/spring: 2.54 4- 0.80 ttg/l vs. 3.47 4- 1.07/~g/l, P = 0.005). Sodium excretion in the two groups did not differ in the winter/spring. In conclusion, beneficial bone effects of thiazide diuretics occur in the winter/spring and they may result from a decrease in PTH-stimulated bone resorption and an associated reduction in the bone turnover rate. Key words: Thiazide diuretics; Osteocalcin; Parathyroid hormone; Sodium; Season; Bone density

Introduction Thiazide use has been associated with higher bone mineral density (BMD) [1-6] and in some [7-10] but not other [11,12] studies, with reduced hip fracture rates. In the one placebo-controlled trial reported I13], early postmenopausal women treated for 3 years with a low dose of thiazide had a reduced rate of bone loss over the first 6 months but not thereafter. Studies relating higher doses of thiazide to rate of bone loss are not available. Correspondence to: Bess Dawson-Hughes, USDA Human Nutrition Research Center on Aging at Tufts University, 711 Washington St, Boston, MA 02111, USA.

42

A favorable effect of thiazide diuretics on BMD implies a positive net effect of thiazides on calcium balance. The mechanism(s) by which thiazides favor preservation of BMD are uncertain but thiazide use results in decreased calcium excretion [14-16], increased serum calcium [16-18], reduced parathyroid activity [19-20] and reduced serum 1,25-dihydroxyvitamin D (1,25(OH)2D) [20,21]. Calcium absorption has been found to be decreased [16,20,22-23] or unchanged [24] by thiazides. Reduction of the bone turnover rate is a potential means by which thiazides mitigate bone loss. Lower serum levels of osteocalcin, a marker of the bone turnover rate, have been associated with reduced rates of bone loss in early postmenopausal women who were in the phase of rapid bone loss [25-27] and in women with osteoporosis [28,29]. PTH stimulates bone resorption and can increase the bone remodeling rate in healthy and osteoporotic women [30]. We recently completed a 1-year vitamin D trial in healthy postmenopausal women during which seasonal variation in rates of bone change and in several hormone levels was observed [31] over 6-month intervals of summer/fall and winter/spring. In these women, we now examine associations between thiazide diuretic use and both rates of bone change and serum concentrations of osteocalcin and calcium regulating hormones in order to increase insight into how thiazides may affect bone. Specific objectives are to determine whether thiazide use is associated with bone change in this population and if so, whether any change observed is season-specific, dose related and/or associated with differences in osteocalcin or calcium-regulating hormone levels.

Me~o~

Subjects The 246 women in this study completed a 1-year calcium and vitamin D supplement trial [31]. All women received 377 mg per day of supplemental calcium, largely as calcium citrate malate; half were randomized to placebo and half to treatment with 400 IU of vitamin D daily. The women were in good health and had no medical conditions or medications (e.g., estrogen, glucocorticoids) known to alter bone metabolism. Thiazide diuretic use (preparation, duration, dosage) was assessed at enrollment and again during the last half of the study. Twenty-five women used a thiazide diuretic during the study and the duration of use ranged from 0.5 to 29 years (mean 11 ± 7 (SD) years in 24, unknown in 1 woman). Twenty-four women used a thiazide throughout the study. One started taking a thiazide sometime during the study and for calculation of intrastudy dosage, we assumed that she started at the midpoint. Three women changed dosage during the study and were assumed to have changed at the midpoint. Twenty-two women used hydrochlorthiazide, in average doses ranging from 12.5 to 50 mg daily, two used chlorthiazide, 500 mg daily and one used hydrodiazide. Clinical characteristics of the 25 thiazide users and 221 non-users are shown in Table 1. Whole-body and spinal BMD was measured first in June or July, then in December or January and finally in the following June or July, and semi-annual rates of change over the first 6-month period (summer/fall) and the second 6-month period (winter/spring) were calculated. During the intervals of August through

43

Table 1 Clinical profile at enrollment of thiazide users and non-users Characteristic

Non-users

Thiazide users

n Age (years) Years since menopause Weight (kg) Body mass index (kg/m2) % Taking 400 IU vitamin D supplement % Smokers % With calcified aorta Bone mineral density (g/cm2) Spinec Whole-bodyc

221 61 -4- 5a 13 ~- 7 67 + 12 26 4- 4 49

25 65 4- 5b 16 4- 5b 72 4- 12b 29 4- 5b 52

8 8 1.03 4- 0.01 (199)d 1.06 4- 0.01

8 20b 1.17 4- 0.03 (18)b 1.11 ± 0.02b

aValues are mean ± SD. bGroups differ at P _<0.05. cValues are mean 4- SEM. dNumber with complete data shown in parentheses.

November and February through May each woman returned for assessment of diet, level of physical activity, cigarette use and for blood and urine tests. Blood was drawn between 07:30 h and 09:00 h, after an overnight fast. Weight was measured on each visit. Lateral lumbar radiographs were taken at enrollment for new recruits and up to 3 years prior to enrollment for those women who had participated in an earlier calcium trial [32]. Measuremen ts

Intake of calcium and vitamin D was assessed by a food frequency questionnaire. Current physical activity was measured with a questionnaire adapted from Kriska et al. [33] and is reported as the sum of kilocalories expended daily on walking out of doors, stair climbing and on sports, housework and leisure activities. Current smoking was assessed by questionnaire. Plasma 25-hydroxyvitamin D (25(OH)D) was measured by the method of Preece et al. [34] with intra- and interassay coefficients of variation of 5.0% and 7.3%, respectively. Plasma 1,25(OH)2D was measured by the competitive protein-binding method of Rheinhardt [35] with intra- and interassay coefficients of variation of 4.9% and 7.7%, respectively. Serum P T H was measured with Allegro Intact P T H kits from Nichols Institute (San Juan Capistrano, CA) with intra- and interassay coefficients of variation of 5.6% and 6.6%, respectively. Serum ionized calcium was measured with a N o v a 7 analyzer (Nova Biomedical, Waltham, MA). Serum phosphorus and creatinine levels were measured by colorimetry with a Cobas Mira and urine creatinine levels were measured with a Cobas Fara Centrifugal analyzer (Roche Instruments, BelleviUe, N J). Urine calcium and sodium were measured by direct-current plasma emission spectroscopy with a Spectrascan 6 (Beckman Instruments, Palo Alto, CA). In a subset of 92 women assigned to the vitamin D treatment arm during the trial, osteocalcin was measured at the end of the study in serum

44 that had been stored at -70°C for between 1 and 2 years. The 92 women were preselected from the larger group of 124 women for thyroid studies and so aspirin users and those with abnormal TSH levels were excluded. Osteocalcin radioimmunoassay kits from INCSTAR (Stillwater, MN) were used; intra- and interassay coefficients of variation are reported to be 2.9% and 6.5%, respectively [36]. BMD was measured with a model DPX dual-energy x-ray absorptiometer from Lunar Radiation Corp. (Madison, WI) with a precision (assessed in 6 subjects measured 6 times each, with repositioning, on 2 occasions, 9 months apart) of 1.0% for the lumbar spine (L2-4) and 0.6% for the whole body [37]. The women had two scans at each site at each timepoint and were repositioned between the scans. The mean of the two measures was used in the analyses. As described previously [31], women who, on lateral spine radiographs, had calcification of the aorta, osteophytes or other calcifications in the spine scan field were excluded from analyses of the spine data. Phantom measurements were stable during the study [31].

Statistical methods Two-tailed t-tests of simple means and adjusted means from analysis of covariance were used to compare subject characteristics, biochemical values and rates of bone change in the thiazide users compared with non-users and in thiazide users on different hydrochlorthiazide doses. Paired t-tests were used to evaluate seasonal changes within thiazide use groups. Spearman rank order correlation coefficients were calculated to examine associations of dose-years and duration of thiazide use with initial BMD. P-values less than 0.05 were considered to indicate statistical significance.

Results

The thiazide users were older, heavier and had higher initial BMD at the spine and whole body (Table 1) than non-users. The differences in initial BMD were not altered by adjustment for differences in age and logarithm of body mass index. A disproportionate number of thiazide users had calcification of the aorta (Table 1). Among the 22 hydrochlorthiazide users, there was no association between current dose, duration, or current dose × duration and initial spinal or whole-body BMD, either before or after controlling for differences in size and age. Rates of change in BMD (adjusted for age, logarithm of BMI and initial BMD) were examined by season. In the winter/spring, thiazide users had no significant bone loss at the spine or whole body whereas non-users lost mineral from both sites (Table 2, Fig. 1). In the summer/fall, thiazide users and non-users had similar increases in spinal BMD whereas non-users alone had a significant increase in whole-body BMD. Overall there was a net increase in spinal BMD in both groups, however the gain in the thiazide users tended to be greater (1.63 q- 0.63 (SE)% vs. 0.39 4- 0.18%, P = 0.062). There was no net change in whole-body BMD in either group. During each season, the thiazide users and non-users had similar levels of physical activity, changes in weight and intakes of calcium and vitamin D (Table 3). Among the 25 thiazide users, we identified no significant influence of calcium or vitamin D intake on overall bone change at either site. Of the 22 hydrochlorthiazide users, 6 took 50 mg per day and 16 took 25 mg per day or less (mean 23 4- 5 rag, range 11-25 mg). Those on the higher dose had a

45

Table 2 Adjusted mean (+SE) percent change in bone mineral density a, by season and overall, in thiazide users and non-users Bone site and season

Non-users

Spine Summer/faU Winter/spring Overall Whole-body Summer/faU Winter/spring Overall

Thiazide users

1.44 -4- 0.17 (199) b -1.02 4. 0.17 (199) b 0.39 4. 0.18(199) c

1.50 4. 0.60 (18) c 0.46 4. 0.59 (18) d 1.63 4. 0.63 (18) c'e

0.65 4. 0.08 (221) b -0.67 q- 0.08 (221) b -0.04 4. 0.09 (221)

0.34 4. 0.26 (25) -0.13 4- 0.25 (25) f 0.12 4. 0.27 (25)

aMeans adjusted for age, logarithm of body mass index and initial bone mineral density. Values in parentheses are sample size. bDiffers from zerO, P < 0.001. CDiffers from zero, P < 0.05. aGroups differ, P = 0.017. eGroup comparison, P = 0.062. fGroups differ, P = 0.043. Spine

*,:~ ~t

t 1 >,

•

0

cO ¢0

E3

$ 0 nn

Whole Body

._= (D

c~ t-

2

t-

L) o~

•

Non-Users

[]

Thiazide

1

-1

t

-2 L ~ Summer/ Winter/ Overall Fall Spring Fig. 1.Mean rates of change in bone mineral density (adjusted for age, logarithm of body mass index and initial BMD) by season and overall in thiazide users and non-users. Bars labeled with asterisks (*) differed from zero (P < 0.05); daggers (T) indicate a difference between treatment groups (P < 0.05) and the double dagger (*) a treatment group comparison for which P = 0.062. T bars show the standard error.

46

Table 3 Dietary and physical measures by season in thiazide users and non-users a Measure and season Calcium intake (rag/day) b Summer/fall Winter/spring Dietary vitamin D (IU/day) Summer/fall Winter/spring Physical activity (kcal/day) Summer/fall Winter/spring Change in weight (kg) Summer/fall Winter/spring

Non-users

Thiazide users

757 4. 215 764 4. 208

794 -4- 220 814 4- 211

94 4. 67 98 4. 70

108 ± 71 107 -4- 71

323 -4- 372 191 4- 207 c

344 4. 434 200 4. 262

0.59 4- 2.26 -0.09 4- 2.38 c

0.86 4. 1.83 -0.14 4. 2.36

aValues are mean 4. SD. blncludes 377 mg of supplemental calcium. eDifiers from summer/faU value, P < 0.005.

similar whole-body bone gain in the summer/fall (0.53 + 0.37 vs. 0.52 ± 0.37%), tended to lose less in the winter/spring (+0.50 4- 0.49 vs. -0.57 4- 0.32%, P = 0.098) and had a better net whole-body bone response (1.01 4- 0.35 vs. -0.14 4- 0.27%, P = 0.024). There were too few on the higher dose with spine data (n = 3) to analyze. Thiazide use was associated with several biochemical differences including reduced levels of PTH and 1,25(OH)2D in the winter/spring but not in summer/fall (Table 4). Adjustment for differences in age and logarithm of body mass index (BMI) did not appreciably alter mean values of 1,25(OH)vD and PTH in either group. Serum ionized calcium and phosphorus levels were similar in the two groups during both seasons (in winter/spring, Ca ++ 1.29 4- 0.04 mmol/1, n = 24 in thiazide users vs. 1.28 4- 0.04 mmol/l, n = 205 in non-users; phosphorus 1.23 4- 0.15 mmol/1, n = 25 vs. 1.21 4- 0.14 mmol/1, n = 219). Plasma 25(OH)D varied with season and was similar in the thiazide users and non-users (Table 4) and did not significantly influence the bone response to thiazide therapy. Urine content of calcium, sodium and creatinine was examined. In thiazide users, the urine calcium to creatinine ratio was 19% lower in the summer/fall (P = 0.017) and 12% lower in the winter/spring (not significant (NS)) than in non-users. The urine sodium to creatinine ratio was higher in thiazide users than in non-users in the summer/fall (P = 0.036), but similar in the two groups in the winter/spring (Table 4). Urine calcium and urine sodium were correlated in summer/fall (r = 0.29, P < 0.001, n = 246) and in winter/spring (r = 0.33, P < 0.001), however neither was correlated with bone change during either season. Sodium excretion did not vary with season in either group and creatinine clearance in the two groups was similar (Table 4). Serum osteocalcin levels in 92 women treated with vitamin D are shown in Table

47

Table 4 Laboratory values (mean 4. SD) by season in 25 thiazide users and 219 non-users Measure and season Plasma 25(OH)D (nmol/1) Summer/fall Winter/spring Plasma 1,25(OH)2D (pmol/l) Summer/fall Winter/spring Serum parathyroid hormone (ng/l) Summer/fall Winter/spring 24-h urinary Ca:Cr ratio (mmol/mol) Summer/fall Winter/spring 24-h Urinary Na:Cr ratio (mol/mol) Summer/fall Winter/spring Creatinine clearance (ml/s) Summer/fall Winter/spring Serum osteocalcin (/zg/1)f Summer/fall Winter/spring

Non-users

Thiazide users

88.2 4- 25.1 a 76.0 4- 30.7

93.1 4- 8.1 a 76.4 4- 29.8

75.2 4- 19.9 77.0 4- 18.7

69.0 4- 19.9 60.8 4- 16.2 b

28.6 4- 10.0 a 31.0 4- 11.7

26.2 4- 710 26.2 4- 7.6 c

542 4- 240 (221) a'd 510 4- 216 (221)

441 4- 185 e 448 4- 190

13.0 4- 5.2 13.4 4- 5.0

15.9 ± 6.4 e 14.3 4- 5.9

1.34 4- 0.31 (218) 1.30 4- 0.25 (218)

1.33 4- 0.38 1.27 4- 0.29

3.65 4- 1.02 (82) a 3.47 4- 1.07 (82)

2.89 4- 0.82 (10) a,e 2.54 4- 0.80 (10) c

aSeasons differ at P < 0.02. bGroups differ at P < 0.001. CGroups differ at P ~ 0.005. dNumber with complete data shown in parentheses. eGroups differ at P < 0.05. fMeasured only in those taking the 400 IU vitamin D supplement.

4. After adjustment for age and logarithm of BMI, mean serum osteocalcin in thiazide users was 3.02 ± 1.65/~g/l (vs. 3.63 ± 1.63/zg/1 in non-users, P = 0.088) in the summer/fall and 2.75 4- 1.70 /~g/1 (vs 3.44 4- 1.63 /~g/1, P = 0.059) in the winter/spring. This subset of thiazide users had reductions in their 1,25(OH)2D and PTH levels in the winter/spring that were similar to those of the larger group (Table 4). Both groups displayed a modest but significant seasonal variation in serum osteocalcin (Table 4).

Discussion

In healthy late postmenopausal women, thiazide use was associated with a reduced rate of bone loss from the spine and whole body in the winter/spring. Thiazide users had the typical summer/fall gain in BMD at the spine whereas gain at the whole body was attenuated. Associations between thiazide use and seasonal bone changes were examined over the relatively short period of one year, however cumulative benefit from these diuretics is suggested by the higher initial BMD of thiazide users in this

48 and other [1-5] studies and by the inverse association between hip fractures and duration of thiazide use reported by Ray et al. [81. Within the limits of the small sample size, a dose of 50 mg per day of hydrochlorthiazide appears to have a greater bone-sparing effect than 25 mg or less, at the whole body. This is consistent with the finding of Felson et al. [10] that pure thiazide diuretics are more effective in reducing hip fracture rates that mixed diuretics and their observation that the latter generally contain no more than 25 mg of hydrochlorthiazide. The presence of calcification of the aorta in a disproportionate number of thiazide users can probably be attributed to aortic injury from hypertension. The increase in serum calcium that results from chronic thiazide use [16], although not documented here, may also play a role in this process. Available hormone and biochemical data allow certain speculation about the mechanisms by which thiazides affect bone. Calcium and sodium excretion are linked [38] and the hypocalciuric effect of thiazides is related to enhanced calcium reabsorption in the distal renal tubule [39,40]. The thiazide users excreted less calcium than non-users, but only in the summer/fall. Since the skeletal benefit of thiazides occurred in the winter/spring, it seems unlikely that reduced calcium excretion alone accounts for the observed benefit of these diuretics. This winter/spring benefit cannot be attributed to treatment group differences in sodium intake in the winter/spring since sodium excretion and changes in weight in the two groups were similar during this season. Thiazide users had substantially lower PTH concentrations in the winter/spring and, in contrast to the non-users, their PTH levels did not increase between summer/fall and winter/spring. The increase in PTH in the non-users is expected to activate new remodeling sites by initiating bone resorption [30]. The lower mean osteocalcin level in the thiazide users in the winter/spring is consistent with this interpretation. A similarly reduced osteocalcin in the summer/spring suggests that thiazides reduce the remodeling rate year round. The reduction in the summer/fall may be a carry over effect related to the 3-6-month time course of the remodeling period [41] or it may involve other mechanisms. The higher BMD in thiazide users appears to result from a reduced remodeling rate. This reduced remodeling rate appears to result in higher BMD among thiazide users and may or may not increase fracture resistance. Reported effects of thiazide use on rates of hip fracture have been contradictory [7-12]. The lower winter/spring PTH level in the diuretic users is presumably mediated by a rise in serum calcium. This was not documented here but as noted earlier, has been observed by others [16-18]. The mean 1,25(OH)2D level was reduced in the thiazides users in the winter/spring as would be expected with their lower PTH levels. This is consistent with findings of others [20,21[, although these studies assessed 1,25(OH)zD levels without reference to season. Lower 1,25(OH)2D levels in the thiazide users may have caused some reduction in calcium absorption in the winter/spring [20], however, in these women with mean calcium intakes of about 800 mg per day, the reduction was not sufficient to cause the urine calcium to drop below the summer/fall level or to prevent the more positive calcium balance implied by their better bone performance in the winter/spring. All of the women in this study were taking calcium supplements, and we therefore can draw no conclusions about possible effects of thiazide use on rates of bone loss in women with low calcium intakes. Vitamin D has previously been shown to in-

49

fluence 25(OH)D and bone change in this cohort [31]. Although subset analyses were limited by the small number of thiazide users, vitamin D intake did not appear to influence the effect of thiazides on net bone change in this study. In conclusion, in healthy postmenopausal women, thiazide use is associated with higher initial BMD and with reduced rates of spinal and whole-body bone loss in the winter/spring. The effect of thiazides on net whole-body bone change appears to be dose related. When compared with non-users, thiazide users had lower levels of osteocalcin year round and lower levels of PTH and 1,25(OH)2D in the winter/spring. The skeletal benefit from thiazides may result from a decrease in PTHstimulated bone resorption and an associated decrease in the rate of bone remodeling.

Acknowledgements This study was supported by the USDA Human Nutrition Research Center on Aging at Tufts University (Contract No. 53-3K06-5-10). The contents of this publication do not necessarily reflect the views or policies of the US Department of Agriculture, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government.

References 1 Wasnich RD, Benfante R J, Yano K, Heilbrun L, Vogel JM. Thiazide effect on the mineral content of bone. N Engl J Med 1983;309:344-347. 2 Sowers MR, Wallace RB, Lemke JH. Correlates of mid-radius bone density among postmenopausal women: a community study. Am J Clin Nutr 1985;41:1045-1053. 3 Wasnich RD, Ross PD, Heilbrun LK, Vogel JM, Yano K, Benfante RJ. Differential effects of thiazide and estrogen upon bone mineral content and fracture prevalence. Obstet Gynecol 1986;67:457-462. 4 Lindsay R, Marshall B, Haboubi A, Herrington BS, Tohme J. Increased axial bone mass in women with hypertension: role of thiazide therapy. J Bone Miner Res 1987;2:Suppl 1:$29 (abstract). 5 Hunt I, Murphy N, Henderson C, Clark V, Coulson A. Bone mineral content in elderly women. Am J Epidemiol 1988;128:903 (abstract). 6 Wasnich RD, Davis J, Ross PD, Vogel JM. Effect of thiazide on rates of bone mineral loss: a longitudinal study. Br Med J 1990;301:1303-1305. 7 Hale WE, Stewart RB, Marks RG. Central nervous system symptoms of elderly subjects using antihypertensive drugs. J Am Geriatr Soc 1984;32:5-10. 8 Ray WA, Downey W, Griffin MR, Melton III, LJ. Long-term use of thiazide diuretics and risk of hip fracture. Lancet 1989;i:687-690. 9 LaCroix AZ, Wienpahl J, White LR, Wallace RB, Scherr PA, George LK, Cornoni-Huntley J, Ostfeld AM. Thiazide diuretic agents and the incidence of hip fracture. N Engl J Med 1990;322:286-290. 10 Felson DT, Sloutskis D, Anderson J J, Anthony JM, Kiel DP. Thiazide diuretics and the risk of hip fracture. JAMA 1991;265:370-373. 11 Adland-Davenport P, McKenzie MW, Notelovitz M, McKenzie LC, Prendergast JF. Thiazide diuretics and bone mineral content in postmenopausal women. Am J Obstet Gynecol 1985; 152:630-634. 12 Heidrich FE, Stergachis A, Gross KM. Diuretic drug use and the risk of hip fracture. Ann Intern Med 1991;115:1-6. 13 Transbol I, Christensen MS, Jensen GF, Christiansen C, McNair P. Thiazide for the postponement of postmenopausal bone loss. Metabolism 1982;31:383-386.

50 14 Lamberg B-A, Kuhlback B. Effect of chlorothiazide and hydrochlorothiazide on the excretion of calcium in urine. Scand J Clin Lab Invest 1959;11:351-357. 15 Middler S, Pak CY, Murad F, Bartter FC. Thiazide diuretics and calcium metabolism. Metabolism 1973;22:139-146. 16 LjunghaU S, Backman U, Danielson BG, Fellstrom B, Johannsson G, Wikstrom B. Calcium and magnesium metabolism during long-term treatment with thiazides. Scand J Urol Nephrol 1981;15:257-262. 17 Seitz H, Jaworski ZF. Effect of hydrochlorothiazide on serum and urinary calcium and urinary citrate. Can Med Assoc J 1964;90:414-420. 18 Stote RM, Smith LH, Wilson DM, Dube WJ, Goldsmith RS, Arnaud CD. Hydrochlorothiazide effects on serum calcium and immunoreactive parathyroid hormone concentrations. Ann Intern Med 1972;77:587-591. 19 Coe FL, Canterbury JM, Reiss E. Hyperparathyroidism in idiopathic hypercalciuria: primary or secondary? Trans Assoc Am Physicians 1971;84:152-161. 20 Sakhaee K, Nicar M J, Glass K, Zerwekh JE, Pak CYC. Reduction in intestinal calcium absorption by hydrochlorothiazide in postmenopausal osteoporosis. J Clin Endocrinol Metab 1984;59: 1037-1043. 21 Sowers MR, Wallace RB, Hollis BW. The relationship of 1,25-dihydroxyvitamin D and radial bone mass. Bone Miner 1990;10:139-148. 22 Nassim JR, Higgins BA. Control of idiopathic hypercalciuria. Br Med J 1965;1:675-681. 23 Donath A, Nordio S, Macagno F, Gatti R. The effect of hydrochlorothiazide on calcium and strontium transport in intestine and kidney. Helv Paediat Acta 1970;25:293-300. 24 Harrison AR, Rose GA. The effect of bendrofluazide on urinary and fecal calcium and phosphorus. Clin Sci 1968;34:343-350. 25 Slemenda C, Hui SL, Longcope C, Johnston CC. Sex steroids and bone mass. J Clin Invest 1987;80:1261 - 1269. 26 Johansen JS, Riis BJ, Delmas PD, Christiansen C. Plasma BGP: an indicator of spontaneous bone loss and of the effect of oestrogen treatment in postmenopausal women. Eur J Clin Invest 1988;18:191-195. 27 Sowers MR, Clark MK, Hollis B, Wallace RB, Jannausch M. Radial bone mineral density in preand perimenopausal women: a prospective study of rates and risk factors for loss. J Bone Min Res 1992;7:647-657. 28 Delmas PD, Wahner HW, Mann KG, Riggs BL. Assessment of bone turnover in postmenopausal osteoporosis by measurement of serum bone gla-protein. J Lab Clin Med 1983;102:470-476. 29 Gundberg CM, Lian JB, Gallop PM, Steinberg JJ. Urinary ~,-carboxyglutamic acid and serum osteocalcin as bone markers: studies in osteoporosis and Paget's disease. J Clin Endocrinol Metab 1983;57:1221-1225. 30 Kotowicz MA, Klee GG, Kao PC, O'Fallon WM, Hodgson SF, Cedel SL, Eriksen EF, Gonchoroff DG, Judd HL, Riggs BL. Relationship between serum intact parathyroid hormone concentrations and bone remodeling in type I osteoporosis: evidence that skeletal sensitivity is increased. Osteoporosis Int. 1990;1:14-22. 31 Dawson-Hughes B, Dallal GE, Krall EA, Harris S, Sokoll LJ, Falconer G. Effect of vitamin D supplementation on wintertime and overall bone loss in healthy postmenopausal women. Ann Intern Med 1991;115:505-512. 32 Dawson-Hughes B, Dallal GE, Krall EA, Sadowski L, Sahyoun N, Tannenbaum S. A controlled trial of the effect of calcium supplementation on bone density in postmenopausal women. N Engl J Med 1990;323:878-883. 33 Kriska AM, Sander RB, Cauley JA, LaPorte RE, Hom DL, Pambianco G. The assessment of historical physical activity and its relation to adult bone parameters. Am J Epidemiol 1988;127:1053-1063. 34 Preece MA, O'Riordan JL, Lawson DE, Kodicek E. A competitive protein-binding assay for 25hydroxycholecalciferol and 25-hydroxyergocalciferol in serum. Clin Chem Acta 1974;54:235-242. 35 Rheinhardt TA, Horst RL, Orf JW, Hollis BW. A microassay for 1,25-dihydroxyvitamin D not requiring high performance liquid chromatography: application to clinical studies. J Clin Endocrinol Metab 1984;58:91-98. 36 Caniggia A, Nuti R, Galli M, Lore F, Turchetti V, Righi GA. Effect of a long-term treatment with 1,25-dihydroxyvitamin D 3 on osteocalcin in postmenopausal osteoporosis. Calcif Tissue Int 1986;38:328-332.

51 Johnson J, Dawson-Hughes B. Precision and stability of dual-energy X-ray absorptiometry measurements. Calcif Tissue Int 1991;49:174-178. 38 Kleeman CR, Bohannan J, Bernstein D, Ling S, Maxwell M. Effect of variations in sodium intake on calcium excretion in normal humans. Proc Soc Exp Biol Med 1964;115:29-32. 39 Brickman AS, Massry SG, Coburn JW. Changes in serum and urinary calcium during treatment with hydrochlorothiazide: studies on mechanisms. J Clin Invest 1972;51:945-954. 40 Sutton RAL. Diuretics and calcium metabolism. Am J Kidney Dis 1985;5:4-9. 41 Parfitt AM. The physiologic and clinical significance of bone histomorphometric data. In: Recker RR, ed. Bone Histomorphometry: techniques and interpretation. Boca Raton, Florida: CRC Press, 1983;143-223. 37