Effects of three years of low-dose thiazides on mineral metabolism in healthy elderly persons

Transcription

1 DOI /s ORIGINAL ARTICLE Effects of three years of low-dose thiazides on mineral metabolism in healthy elderly persons S. M. Ott & A. Z. LaCroix & D. Scholes & L. E. Ichikawa & K. Wu Received: 31 May 2007 / Accepted: 12 December 2007 # International Osteoporosis Foundation and National Osteoporosis Foundation 2008 Abstract Summary In this clinical trial of 12.5 or 25 mg/day of hydrochlorothiazide, the urine calcium showed significant decreases from placebo in men at one year, but the effects had waned by 3 years. Serum bicarbonate was consistently greater in the thiazide than in the placebo groups throughout the three years. These effects could be beneficial to the skeleton. Introduction Previous studies have shown increased bone density and reduced risk of fracture in patients taking thiazide diuretics. The long-term effects of low-dose thiazides on mineral metabolism have not been reported in normal subjects. Methods We conducted a randomized, double-blinded trial in normals aged years, using hydrochlorothiazide 12.5 or 25 mg/d or placebo for three years. Subjects were encouraged to maintain calcium intake of 1,000 to 1,500 mg/day. Measurements of serum and urine calcium metabolism were done at baseline, six months, and yearly. Data were analyzed in 88 men and 177 women who had taken study medication. Adjusted change in the measurements from baseline to one and three years were compared among groups. Results The calcium intake increased in all groups. Urine calcium per day was significantly lower in thiazide than S. M. Ott (*) University of Washington, 1959 NE Pacific Street, P. O. Box , Seattle, WA , USA smott@u.washington.edu A. Z. LaCroix : D. Scholes : L. E. Ichikawa Group Health Cooperative of Puget Sound, 1730 Minor Ave, 16th Floor, Seattle, WA 98101, USA K. Wu University of Washington, Box , Seattle, WA , USA placebo groups in men at one year but not at three years; in women the changes were not significantly different. Serum bicarbonate was higher in thiazide compared to placebo groups at one and three years. No changes were seen in serum calcium, phosphate, parathyroid hormone, sodium or magnesium. Conclusions The results suggest that both increased calcium availability from a hypocalciuric effect and reduction in acidinduced bone buffering could be mechanisms for the beneficial skeletal effects. Keywords Bicarbonate. Clinical trial. Hypercalciuria. Thiazides Introduction Thiazide diuretics reduce blood pressure and prevent cardiovascular complications of hypertension. They are also used to treat nephrolithiasis, because they reduce urine calcium excretion. Observational studies have shown fewer hip fractures in elderly persons who are taking thiazides. In a prospective cohort of 9,518 elderly men and women from multiple communities the relative risk for hip fracture was 0.6 in those using thiazides [1]. A meta-analysis in 1995 found an overall odds ratio of 0.82 ( ) for hip fracture in current thiazide users [2]. Since then, significant reductions in fractures in long-term thiazide users have been reported in a ten-year study of 83,728 woman [3], and in case-control studies of 30,601 fracture cases in the United Kingdom [4] and 64,699 from Denmark [5]. A recent meta-analysis of 25 studies found the reduction of risk for any fracture was 0.86 (95% CI ) [6]. Bone density is higher in thiazide users [7, 8] and randomized clinical trials have reported modest increases in the bone density over 1 to 3 years [9 12].

2 Studies of thiazides in normal individuals have generally involved short-term exposure, except one study of postmenopausal women that lasted two years [11]. The mineral effects of thiazides have been studied in patients with hypertension or hypercalciuria, but these patients may have abnormalities in mineral metabolism, so the results may not apply to normal people. None of the long-term studies have examined mineral effects of low doses (less than 25 mg/day). We report here the measurements of serum and urine minerals in healthy men and women who participated in a three-year randomized, double-blinded clinical trial of lowdose thiazide. Subjects took placebo, hydrochlorothiazide (HCTZ) 25 mg/day or HCTZ 12.5 mg/day. Methods Subjects Details of the trial, study design, and data collection have been published previously [10]. The study was conducted at a large non-profit health maintenance organization. The target population for the trial was healthy men and women aged years at the time of entering the study. Blood pressure was lower than 160 mmhg systolic and 90 mmhg diastolic. Subjects were excluded if they had a history of congestive heart failure, low density lipoprotein cholesterol level >190 mg/dl, diabetes, or moderate ST-T wave changes on an electrocardiogram. Baseline hip bone density was within two standard deviations for the age, using the NHANES database [13]. We excluded potential subjects for contraindications to thiazide therapy, conditions and diseases known to influence bone loss or make completion of the trial unlikely. Age-eligible subjects were recruited through direct mailings. We enrolled 205 women and 115 men including 304 whites (95%), nine Asian-Americans (2.8%), three Hispanic-Americans, two African-Americans and two with race/ethnicity unclassified. The Human Subjects Review Committees at Group Health Cooperative and the University of Washington approved the study and each participant provided written informed consent. Study design Subjects were randomized, stratified by gender, into three groups: placebo, HCTZ 12.5 mg/day and HCTZ 25 mg/day. Study medication was taken for 3 years. Subjects were seen in the research clinic every six months. Fasting blood was drawn at yearly intervals in all subjects. Food frequency questionnaires were administered at baseline and after 3 years. At the beginning of the study we gave all subjects a brochure of foods with high potassium content, and advised them to include these foods in their diet. They were told the results of the calcium intake from their questionnaire, and advised to maintain a total calcium intake of 1,000 to 1,500 mg/day, using foods containing calcium or calcium carbonate supplements of their choice. The study was double-blinded. The patients were not told the results of their follow-up bone density readings until the study was completed. They did receive copies of their lab tests after each visit. If serum potassium was less than 3.4 meq/dl they were given potassium chloride supplements. Measurements The dietary intakes of calcium were assessed using the Fred Hutchinson Cancer Research Center Dietary Intake Questionnaire [14]. The serum and urine laboratory tests were done in the clinical laboratory at the University of Washington Medical Center using standard methods. Urine was collected for 24 hours. The assay for PTH was an immunochemiluminescent assay of intact PTH and osteocalcin an immunoradiometric assay (both from Nichols Institute, San Juan Capistrano, CA) and the Osteomark urine ELISA for N-telopeptide (Ostex International, Seattle, WA). Statistical methods The purpose of this analysis was to examine the potential mechanisms through which thiazide acts on the skeleton. Because calcium metabolism biomarkers were not primary outcomes of the trial an intent-to-treat analysis was not done. Instead, the analysis was restricted to the 265 subjects (177 women and 88 men) who took the study medication throughout the three-year study period, which allowed investigation of short and long-term effects of thiazide treatment on these measurements during active treatment. Baseline characteristics of subjects who did and did not discontinue the study medication were compared using t-tests for continuous variables and chi-squared tests for categorical variables. Among the subjects included in the analysis, we tested for a difference in baseline characteristics by treatment group. The mean change from baseline in serum and urine minerals and total calcium intake at the annual visits was calculated for men and women separately. A linear regression model was used to test for differences in the 12-month and 36-month change by treatment. Previous studies have suggested that the changes in urine calcium seen initially with therapy could return towards baseline with longer use [15], so we measured both the short-term (one year) and long-term (three years) change. For all serum and urine minerals except urine calcium, urine phosphate, and PTH, the model adjusted for continuous measures of baseline age, weight, height, total calcium

3 intake, and protein. The aim of the trial was to see if thiazides could improve bone density in patients already taking the recommended amount of calcium, so for many of the subjects the calcium intake increased. Therefore, the model for urine calcium, urine phosphate, and PTH included change in total calcium intake instead of baseline calcium since these outcomes are directly related to the change in the calcium intake. A global test was used to determine if any treatment differences were significant. If the global test was significant, we looked at the difference in the 12.5 mg thiazide treatment compared to the placebo group and in the 25 mg thiazide treatment compared to the placebo group. The measurements that we included in this study have been shown by others to be associated with thiazides, so we did not make any adjustments for multiple comparisons. The analysis for this paper was generated using Base SAS and SAS/STAT software, Version 9.1 of the SAS System for Windows. (Copyright SAS Institute Inc. SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc., Cary, NC, USA). All tests were twosided and performed at the p=0.05 significance level. Results The baseline characteristics of the 265 subjects who took their medication during the study interval are shown in Table 1. The reasons for discontinuation have been published [10]. Compared to per-protocol subjects, discontinuers of the study medication had a lower mean serum sodium (142.6 mmol/l vs mmol/l) and a higher mean urine phosphate (28.6 mmol/d vs mmol/d) at baseline; otherwise there were no significant baseline differences among the subjects who did and did not continue their study medication. Subjects whose baseline urine calcium was greater than 7.5 mmol/day were removed from the trial and are not included in this report. The only baseline differences by treatment group were serum potassium and serum phosphate among males and total calcium intake among females (Table 1). The women taking placebo had higher intake of calcium at baseline, and their intake remained statistically significantly higher at each time point except the two-year visit. During the study, the calcium intake increased in all groups, shown in Fig. 1. This was most likely a direct result of advising subjects to maintain a daily calcium intake of 1,000 to 1,500 mg/day. The amount of increase in mean calcium intake between the baseline and one-year visit was not significantly different between treatment groups for either men or women. Figure 2 shows the unadjusted changes in urine calcium at time points throughout the study. In men, the urine calcium increased in the placebo group but decreased in both thiazide groups. After two years the urine calcium returned towards baseline. In women, the urine calcium did Table 1 Baseline characteristics±se Gender Male Female Treatment (n) Placebo (31) HCTZ 12.5 (34) HCTZ 25 (23) Placebo (57) HCTZ 12.5 (60) HCTZ 25 (60) Age, years 67.3± ± ± ± ± ±0.6 Weight, kg 83.2± ± ± ± ± ±1.5 BMI, kg/m2 26.7± ± ± ± ± ±0.5 Calcium intake, mg/day* 772±76 925±88 849± ± ± ±84 Protein intake, g/day 62.5± ± ± ± ± ±3.4 Smoking, current, % 9.7± ± ± ± ± ±2.3 Serum values: Calcium, mmol/l 2.36± ± ± ± ± ±0.01 Phosphate, mmol/l* 1.00± ± ± ± ± ±0.02 Magnesium mmol/l 0.58± ± ± ± ± ±0.01 Bicarbonate mmol/l 27.0± ± ± ± ± ±0.3 Potassium, mmol/l* 4.47± ± ± ± ± ±0.04 Sodium, mmol/l 142.8± ± ± ± ± ±0.2 PTH, ng/l 32.8± ± ± ± ± ±1.8 Osteocalcin, ng/ml 7.1± ± ± ± ± ±0.4 Urine values/ day: Calcium, mmol 3.97± ± ± ± ± ±0.22 Sodium, mmol 7.30± ± ± ± ± ±0.30 Phosphate, mmol 29.3± ± ± ± ± ±0.8 N-telopeptide, Beq/mg creat 34.6± ± ± ± ± ±2.1 *=significant differences among groups at baseline

4 lower than 3.4 meq/l. This occurred in one man in the placebo group, one man and one woman in the 12.5 mg group, and two men and 20 women in the 25 mg group. The adjusted change in urine calcium was significantly different by treatment group only among men at one year with both thiazide groups having a significant decrease compared to the placebo group. The change in urine sodium was significantly different by treatment group only among men at three years in the 25 mg thiazide group. There were no differences among the groups at any time point for serum calcium, phosphate, magnesium, sodium, creatinine or parathyroid hormone. Urine phosphate also was not changed, and the tubular reabsorption of phosphate (data not shown) was not different. The urine N-telopeptide, which is a marker of bone resorption, showed decreases at 6 months (data not shown), which were significant when data for men and women were combined. At one year, there was an overall significant treatment difference among men. However, this difference was between the two thiazide groups and not between Fig. 1 Change in calcium intake during study, mean ± SE not show significant differences between thiazide and placebo groups. The serum bicarbonate results are plotted in Fig. 3. The values in thiazide groups were generally higher than in placebo groups. The subjects returned one year after discontinuation of treatment, and the serum bicarbonate values in the former thiazide groups were not different than in the former placebo groups. Tables 2 and 3 show the adjusted changes in the laboratory measurements after one and three years of treatment in men and in women. Significant changes were seen in the serum bicarbonate, serum potassium, urine calcium and urine sodium. In men, the increase in serum bicarbonate was greater than placebo for both 12.5 mg and 25 mg doses at one and three years. In women, the increase in serum bicarbonate was greater than placebo at one year in the low dose group and at three years in the higher dose group. The serum potassium decreased at both doses and time points for women; in men the only significant change was in the 25 mg dose at one year. These values should be interpreted with caution because subjects were given potassium chloride supplements if their serum values were Fig. 2 Unadjusted changes in urine calcium in women and men during 3 years of thiazide or placebo use, mean ± SE

5 thiazide and placebo. By three years the differences in N-telopeptide were no longer significant. A decrease in osteocalcin, a marker of osteoblast activity, was significant in women at 1 year only in the high dose group. The hip bone density change from baseline to 36 months in this per-protocol analysis in men was 1.28% in the 25 mg group, 1.46% in the 12.5 mg group, and 0.61% in the placebo group. In women, these changes were 0.78%, -0.13% and -0.88%, respectively. We looked more closely at urine calcium and serum bicarbonate due to the observed differences by thiazide group. The change in serum bicarbonate or in urine calcium did not show a significant association with the change in hip BMD. The differences in bone density among the thiazide groups were similar to the above results when they were adjusted by the change in urine calcium or the change in serum bicarbonate; there was no attenuation of effect. We grouped the 3-year change in urine calcium and serum bicarbonate in tertiles and examined the association of the 3-year change in these lab values with the 3-year change in hip bone mineral density. In a model that combined men and women and included gender, age, weight, height, protein, dietary calcium, and treatment group as covariates, the 3-year change in hip bone mineral density was not associated with the 3-year change in urine calcium or serum bicarbonate. There was also no association when using 3-year change in N-telopeptide instead of hip BMD. Fig. 3 Unadjusted changes in serum bicarbonate in women and men during 3 years of thiazide or placebo use, and one year after discontinuation of study medication, mean ± SE Discussion In this 3-year study of low (25 mg/day) and very low (12.5 mg/day) doses of thiazide in normal elderly men and Table 2 Changes in laboratory measurements in women after one and three years of treatment Serum values Placebo (n=57) HCTZ 12.5 mg/day (n=60) HCTZ 25 mg/day (n=60) One year Three years One year Three years One year Three years Calcium, mmol/l -0.01± ± ± ± ± ±0.01 Phosphate,mmol/L -0.03± ± ± ± ± ±0.02 Magnesium mmol/l 0.08± ± ± ± ± ±0.01 Bicarbonate mmol/l -0.16± ± ± ±0.49* 1.19±0.35* 2.65±0.47 Potassium, mmol/l 0.04± ± ±0.05* -0.26±0.05* -0.47±0.05* -0.41±0.05* Sodium, mmol/l -0.73± ± ± ± ± ±0.38 PTH, ng/l a -4.70± ± ± ± ± ±2.03 Osteocalcin, ng/ml -1.5± ± ± ± ±0.3* -2.0±0.3 Urine values/ day Calcium, mmol a -0.17± ± ± ± ± ±0.23 Sodium, mg 0.04± ± ± ± ± ±0.36 Phosphate, mmol a -0.6± ± ± ± ± ±1.1 N-telopeptide, Beq/mg creat b -8.1± ± ± ± ± ±2.8 Adjusted for age, weight, height, baseline protein intake, baseline calcium intake a adjusted for change in calcium intake instead of baseline value b half sample at one year *=significantly different from placebo group, p<.05

6 Table 3 Changes in laboratory measurements in men after one and three years of treatment Serum values Placebo (n=31) HCTZ 12.5 mg/day (n=34) HCTZ 25 mg/day (n=23) One year Three years One year Three years One year Three years Calcium, mg/dl -0.05± ± ± ± ± ±0.02 Phosphate, mmol/l -0.05± ± ± ± ± ±0.03 Magnesium, mmol/l 0.06± ± ± ± ± ±0.01 Bicarbonate, mmol/l -1.42± ± ±0.42 * 1.86±0.53 * 0.32±0.51 * 1.78±0.63 * Potassium, mmol/l -0.16± ± ± ± ±0.08 * -0.25±0.08 Sodium, mmol/l -0.41± ± ± ± ± ±0.60 PTH, ng/l a 0.21± ± ± ± ± ±2.21 Osteocalcin, ng/ml -1.2± ± ± ± ± ±0.3 Urine values/ day Calcium, mmol a 0.42± ± ±0.26 * -0.72± ±0.31 * 0.17±0.43 Sodium, mmol -0.11± ± ± ± ± ±0.67 * Phosphate, mmol a 3.0± ± ± ± ± ±2.3 N-telopeptide, Beq/mg creat b -9.4± ± ± ± ± ±3.2 Adjusted for age, weight, height, baseline protein intake, baseline calcium intake a adjusted for change in calcium intake instead of baseline value b half sample at one year *=significantly different from placebo group, p<.05 women, the effects on urine calcium were inconsistent. All of the subjects were encouraged to maintain an adequate dietary calcium intake because we wanted to see the effects of thiazide in persons who were not calcium deficient; this resulted in an increase in dietary calcium that was greater in men than women but within each gender was similar in thiazide and placebo groups. In men, there were significant decreases in urine calcium (compared to the placebo group) at one year, but the effect was largely gone by three years. In women, there were no significant differences in urine calcium between thiazide and placebo groups. In both men and women, the most consistent finding was an increase in serum bicarbonate in the thiazide groups that persisted for three years, as long as the subjects were taking thiazide. After the drug was discontinued, the bicarbonate values were no longer different from those in subjects who had taken placebo. In patients with hypertension or hypercalciuria thiazides consistently decrease urine calcium, but in some cases the urine calcium returns towards baseline [15]. Decreased serum potassium and increased bicarbonate are welldescribed effects that are related to dose [16, 17]. Some biochemical changes that are reported in hypertensive patients taking thiazides were not apparent in our study. These include hyponatremia, hypomagnesemia, and mild hypercalcemia. We think this is probably because we used low doses. The fact that these were healthy people also could have played a role, since hypertensive patients may have different metabolism of sodium and potassium. Reid and colleagues [11] have conducted a randomized clinical trial of 50 mg/day thiazide for 2 years in postmenopausal women. They reported decreased urine calcium, unchanged PTH, lower potassium and decreased N-telopeptide which was maximal at 2 months. These were similar to our results. They found decreased serum sodium and increased serum calcium, which we did not see; these changes may reflect the larger dose used in their study. Serum bicarbonate was not reported. The mechanisms for the skeletal effects of thiazides are not completely understood. Thiazides directly inhibit the thiazide-sensitive sodium-calcium cotransporter (NCC) which is located in the distal tubules of the nephron and also in bone cells [18, 19]. Patients who have a defect in the NCC gene (Gitelman's disease) manifest symptoms that are similar to some of those caused by thiazides, including hypokalemia, hypomagnesemia, alkalosis, hypocalciuria, and high bone density [20]. The opposite effects are seen in patients with pseudohypoaldosteronism II, who have low bone density associated with WNK-4 mutations which fail to inhibit NCC [21]. Thiazides have a hypocalciuric effect [22]. It is frequently postulated that the resulting increase in calcium availability with thiazide use is one mechanism for beneficial effects of thiazides. Reid et al. [11] found that the increase in bone density in their trial was inversely related to the 2-year change in 24-hour urine calcium excretion. In a 2-year extension of that trial, the urine calcium results did not persist, similar to findings in our study [23]. The bone density, nevertheless, remained significantly higher than the placebo group at the total body, legs, and forearms, but not the spine. These study results suggest that decreased urine calcium is not the sole mechanism for skeletal benefit. In our women the reduction in urine calcium was not statistically significant, and in men the urine calcium values

7 (adjusted for calcium intake) were significantly lower in the thiazide groups than in the placebo group at one year, but the effect did not persist. Furthermore, there was no decrease in the PTH levels which would be expected if calcium-deficiency is corrected by increased calcium availability. Our subjects were not deficient in calcium intake, so results might be different in the setting of calcium insufficiency. Thiazides increased bicarbonate in our study, and this effect was more consistent, persistent and statistically significant than the changes in urine calcium. Mild metabolic alkalosis is beneficial to the bones. The bones buffer any excess protons to maintain the systemic hydrogen ion concentration [24]. Dietary intake of fixed acids, predominantly phosphates and sulfur from amino acids, causes a slight negative calcium balance and increase in bone resorption. Without adequate renal compensation, the bone loss due to acid could accumulate and contribute to the osteoporosis seen with aging [25]. Trials of potassium bicarbonate have shown dose-dependent improvement in calcium balance with mild alkalosis and decreased urine calcium [26]. A recent study of potassium citrate found increased bone density compared to subjects given potassium chloride [27]. In normal persons without hypercalciuria or hypertension, the mild alkalosis could be an important contributing factor to the increased bone density seen with thiazides. We did not see a significant relationship between individual change in serum bicarbonate and change in hip bone density. Neither was there a significant relationship between either baseline urine calcium or change in urine calcium and the change in hip bone density. The changes in bone density were modest and the measurement error for a single person is higher than the mean difference, which may explain the lack of significant findings or of attenuation of the effect when adjusted for change in bicarbonate or calcium. In our subjects, when hip bone density was measured twice on the same day, the absolute percent difference between measurements was 1.7%, which is approximately the same as the difference between the thiazide and placebo group in the women. Study limitations include a population of mostly Caucasian persons, and a higher discontinuation rate in the men on 25 mg/day. Also, the fact that calcium intake was not steady makes the results more difficult to interpret. We did not have direct measurements of gastrointestinal calcium absorption, which would have allowed more precise calculation of the calcium economy. Strengths include 3-year duration which is longer than most studies in normals, inclusion of both genders, and a randomized trial design with a low drop-out rate. This study differs from previous studies because we measured serum bicarbonate at every visit, we included men as well as women, and we examined the effects of low doses (25 and 12.5 mg/d). We think the bone and mineral effects of these low doses of thiazide are important because they have fewer side effects than higher doses. This study did not have enough power to determine if low doses of thiazides could reduce fracture incidence. The effect on preserving bone density was modest. While these effects are not of sufficient magnitude to replace use of treatments for established osteoporosis, these data support a biologically plausible link between low dose thiazide use and reduced risk of fracture. The majority of fractures in elderly people occur in those whose bone density is not low enough to be classified as osteoporosis. Observational studies suggest that thiazides could potentially be used to prevent fractures in these persons with mildly decreased bone density, and we think long-term clinical trials should be done in this population. In summary, we have shown that three years of low-dose thiazide decreases urine calcium and increases serum bicarbonate in normal elderly men and women, without changes in the serum calcium or PTH. The urine calcium effect did not persist and was no different from placebo by the end of three years, whereas the bicarbonate differences remained significant throughout the duration of thiazide exposure. Both of these effects could contribute to beneficial skeletal effects. Funding National Institutes of Health (AG09825) and University of Washington Clinical Nutrition Research Unit, National Institutes of Health (P30 DK35816). The study medications were provided by Ciba-Geigy. Conflicts of interest interest. References None of the authors have any conflicts of 1. LaCroix AZ, Wienpahl J, White LR, Wallace RB, Scherr PA, George LK, Cornoni-Huntley J, Ostfeld AM (1990) Thiazide diuretic agents and the incidence of hip fracture. N Engl J Med 322: Jones G, Nguyen T, Sambrook PN, Eisman JA (1995) Thiazide diuretics and fractures: can meta-analysis help? J Bone Miner Res 10: Feskanich D, Willett WC, Stampfer MJ, Colditz GA (1997) A prospective study of thiazide use and fractures in women. Osteoporos Int 7: Schlienger RG, Kraenzlin ME, Jick SS, Meier CR (2004) Use of beta-blockers and risk of fractures. Jama 292: Rejnmark L, Vestergaard P, Mosekilde L (2005) Reduced fracture risk in users of thiazide diuretics. Calcif Tissue Int 76: Wiens M, Etminan M, Gill SS, Takkouche B (2006) Effects of antihypertensive drug treatments on fracture outcomes: a metaanalysis of observational studies. J Intern Med 260:

8 7. Bauer DC, Browner WS, Cauley JA, Orwoll ES, Scott JC, Black DM, Tao JL, Cummings SR (1993) Factors associated with appendicular bone mass in older women. The Study of Osteoporotic Fractures Research Group. Ann Intern Med 118: Orwoll ES, Bevan L, Phipps KR (2000) Determinants of bone mineral density in older men. Osteoporos Int 11: Transbol I, Christensen MS, Jensen GF, Christiansen C, McNair P (1982) Thiazide for the postponement of postmenopausal bone loss. Metabolism 31: LaCroix AZ, Ott SM, Ichikawa L, Scholes D, Barlow WE (2000) Low-dose hydrochlorothiazide and preservation of bone mineral density in older adults. A randomized, double-blind, placebocontrolled trial. Ann Intern Med 133: Reid IR, Ames RW, Orr-Walker BJ, Clearwater JM, Horne AM, Evans MC, Murray MA, McNeil AR, Gamble GD (2000) Hydrochlorothiazide reduces loss of cortical bone in normal postmenopausal women: a randomized controlled trial. Am J Med 109: Wasnich RD, Davis JW, He YF, Petrovich H, Ross PD (1995) A randomized, double-masked, placebo-controlled trial of chlorthalidone and bone loss in elderly women. Osteoporos Int 5: Looker AC, Wahner HW, Dunn WL, Calvo MS, Harris TB, Heyse SP, Johnston CC Jr, Lindsay R (1998) Updated data on proximal femur bone mineral levels of US adults. Osteoporos Int 8: Kristal A, Feng Z, Coates R, Oberman A, George V (1997) Associations of race/ethnicity, education, and dietary intervention with the validity and reliability of a food frequency questionnaire: the Women's Health Trial Feasibility Study in Minority Populations. Am j Epidemiol 146: Preminger GM, Pak CY (1987) Eventual attenuation of hypocalciuric response to hydrochlorothiazide in absorptive hypercalciuria. J Urol 137: Odvina CV, Preminger GM, Lindberg JS, Moe OW, Pak CY (2003) Long-term combined treatment with thiazide and potassium citrate in nephrolithiasis does not lead to hypokalemia or hypochloremic metabolic alkalosis. Kidney Int 63: Greenberg A (2000) Diuretic complications. Am J Med Sci 319: Barry EL, Gesek FA, Kaplan MR, Hebert SC, Friedman PA (1997) Expression of the sodium-chloride cotransporter in osteoblast-like cells: effect of thiazide diuretics. Am J Physiol 272:C Lalande A, Roux C, Graulet AM, Schiavi P, De Vernejoul MC (1998) The diuretic indapamide increases bone mass and decreases bone resorption in spontaneously hypertensive rats supplemented with sodium. J Bone Miner Res 13: Nicolet-Barousse L, Blanchard A, Roux C, Pietri L, Bloch-Faure M, Kolta S, Chappard C, Geoffroy V, Morieux C, Jeunemaitre X, Shull GE, Meneton P, Paillard M, Houillier P, De Vernejoul MC (2005) Inactivation of the Na-Cl co-transporter (NCC) gene is associated with high BMD through both renal and bone mechanisms: analysis of patients with Gitelman syndrome and Ncc null mice. J Bone Miner Res 20: Mayan H, Vered I, Mouallem M, Tzadok-Witkon M, Pauzner R, Farfel Z (2002) Pseudohypoaldosteronism type II: marked sensitivity to thiazides, hypercalciuria, normomagnesemia, and low bone mineral density. J Clin Endocrinol Metab 87: Sutton RA (1985) Diuretics and calcium metabolism. Am J Kidney Dis 5: Bollard MJ, Ames RW, Horne AM, Orr-Walker BJ, Gamble GD, Reed IR (2007) The effect of treatment with a thiazide diuretic for 4 years on bone density in normal postmenopausal women. Osteoporos Int 18: Lemann J Jr, Bushinsky DA, Hamm LL (2003) Bone buffering of acid and base in humans. Am J Physiol Renal Physiol 285:F Bushinsky DA (2001) Acid-base imbalance and the skeleton. Eur J Nutr 40: Frassetto LA, Nash E, Morris RC, Sebastian A (2000) Comparative effects of potassium chloride and bicarbonate on thiazide-induced reduction in urinary calcium excretion. Kidney Int 58: Jehle S, Zanetti A, Muser J, Hulter HN, Krapf R (2006) Partial neutralization of the acidogenic Western diet with potassium citrate increases bone mass in postmenopausal women with osteopenia. J Am Soc Nephrol 17: