Bone Loss in Patients with Untreated Chronic ... - Wiley Online Library

JOURNAL OF BONE AND MINERAL RESEARCH Volume 16, Number 11, 2001 © 2001 American Society for Bone and Mineral Research

Bone Loss in Patients with Untreated Chronic Obstructive Pulmonary Disease Is Mediated by an Increase in Bone Resorption Associated with Hypercapnia HANS P. DIMAI,1 WOLFGANG DOMEJ,1 GEORG LEB,1 and K.-H. WILLIAM LAU2

ABSTRACT This study sought to determine whether the bone loss in untreated chronic obstructive pulmonary disease (COPD) is associated with hypercapnia and/or respiratory acidosis. Bone mineral density (BMD) measured at the distal forearm of the nondominant arm (with peripheral quantitative computed tomography [pQCT]) and serum markers of bone turnover were determined in 71 male patients with untreated COPD and 40 healthy male subjects who matched the patients in age, weight, and body mass index (BMI). The COPD patients, compared with controls, had reduced pulmonary functions, lower arterial pH, and elevated arterial partial pressure of CO2 (PCO2). The BMD (in T score) was significantly lower in COPD patients than that in control subjects (ⴚ1.628 ⴞ 0.168 vs. ⴚ0.058 ⴞ 0.157; p < 0.001). The BMD of COPD patients correlated positively with arterial pH (r ⴝ 0.582; p < 0.001), negatively with PCO2 (r ⴝ ⴚ0.442; p < 0.001), and negatively with serum cross-linked telopeptide of type I collagen (ICTP), a bone resorption marker (r ⴝ ⴚ0.444; p < 0.001) but not with serum osteocalcin, a bone formation marker. Serum ICTP, but not osteocalcin, correlated with PCO2 (r ⴝ 0.593; p < 0.001) and arterial pH (r ⴝ ⴚ0.415; p < 0.001). To assess the role of hypercapnia, COPD patients were divided into the hypercapnic (PCO2 > 45 mm Hg; n ⴝ 35) and eucapnic (PCO2 ⴝ 35– 45 mm Hg) group (n ⴝ 36). Patients with hypercapnia had lower BMD, lower arterial pH, and higher serum ICTP than did patients with eucapnia. Arterial pH and serum ICTP of eucapnic patients were not different from those of controls. To evaluate the role of uncompensated respiratory acidosis, COPD patients with hypercapnia were subdivided into those with compensatory respiratory acidosis (pH > 7.35; n ⴝ 20) and those with uncompensated respiratory acidosis (pH < 7.35; n ⴝ 15). The BMD and serum ICTP were not different among the two subgroups. In conclusion, this study presents the first associative evidence that the bone loss in COPD is at least in part attributed to an increased bone resorption that is associated primarily with hypercapnia rather than uncompensated respiratory acidosis. (J Bone Miner Res 2001;16:2132–2141) Key words:

chronic obstructive pulmonary disease, respiratory acidosis, hypercapnia, bone resorption, bone loss (humans)

INTRODUCTION bone and mineral metabolism.(1,2) There are two major types of acidosis: (1) metabolic acidosis—reduction in pH of body fluids caused by reduction of bicarbonate concentration ([HCO⫺ 3 ]); and (2) respi-

A

CIDOSIS AFFECTS

ratory acidosis—reduction in pH of body fluid caused by an increase in partial pressure of carbon dioxide (PCO2), that is, hypercapnia. Metabolic acidosis increases osteoclastic bone resorption,(3,4) accelerates physicochemical dissolution of bone minerals,(5,6) is associated with increases in bone resorption

1

Department of Endocrinology and Pneumology, University of Graz Medical School, Graz, Austria. Departments of Medicine and Biochemistry, Loma Linda University, and Musculoskeletal Disease Center, J.L. Pettis Memorial VA Medical Center, Loma Linda, California, USA. 2

2132

HYPERCAPNIA AND BONE LOSS

and urinary calcium (Ca) excretion,(7,8) and decreases in bone formation(9) and bone mineral content.(10,11) Metabolic acidosis may play a key role in the bone loss associated with a number of metabolic bone diseases such as renal osteodystrophy(12–14) and osteoporosis.(1,15) Information regarding the effects of respiratory acidosis on bone and mineral metabolism has been limited and conflicting. Past in vivo studies with prolonged exposure to elevated atmospheric CO2 content as a model for respiratory acidosis showed both stimulatory and inhibitory effects on urinary Ca excretion,(16,17) serum Ca levels,(18) and bone Ca content.(18,19) Although respiratory acidosis induced bone Ca efflux, its effect on the release of Ca from bones in animals was much less than that of metabolic acidosis.(20) These in vivo observations are in conflict with the in vitro findings that the bone resorptive activity of isolated rat osteoclasts was more sensitive to the stimulation induced by the CO2 acidosis (a model of respiratory acidosis) than that induced (21) by the [HCO⫺ 3 ] acidosis (a metabolic acidosis model). Chronic obstructive pulmonary disease (COPD) describes chronic bronchitis or a combination of predominant chronic bronchitis and secondary emphysema. Clinical hallmarks include chronic cough, wheezing, expectoration, and dyspnea during physical activities. The airflow in COPD patients is progressively reduced with time, leading to an increasing imbalance between alveolar ventilation and blood flow and, thereby, an inefficient exchange of O2 and CO2. Patients with untreated COPD often develop chronic hypoxemia and hypercapnia(22) and are prone to respiratory acidosis.(23) Thus, untreated COPD is a good model for investigations of chronic respiratory acidosis on bone metabolism. COPD is a risk factor for osteoporosis because the patient with the disease often loses significant amounts of cortical and trabecular bone.(24) Patients with COPD are sometimes treated with large doses of corticosteroids (to reverse bronchoconstriction) that could cause bone loss.(25) However, it is unclear whether the disease itself or related complications such as respiratory acidosis and/or hypercapnia play a role in the bone loss.(26) This study sought to determine whether uncompensated respiratory acidosis and/or chronic hypercapnia would adversely affect bone mineral density (BMD) and bone turnover in 71 male patients with untreated COPD who had never received corticosteroid therapy. A group of 40 male control subjects matched in age, weight, and body mass index (BMI) were recruited for comparison. We investigated if there is an association between BMD and hypercapnia and/or uncompensated respiratory acidosis and if the effect on BMD is associated with increased bone resorption or reduced bone formation. To evaluate the effect of hypercapnia on bone metabolism, the patients with COPD were divided into the hypercapnic (i.e., PCO2 ⬎ 45 mm Hg) and eucapnic (i.e., PCO2 ⱕ 45 mm Hg) subgroups. To assess if uncompensated respiratory acidosis had an effect on bone metabolism, the patients with hypercapnia were divided further into the subgroup of patients with uncompensated respiratory acidosis (arterial pH ⱕ 7.35) and the subgroup with compensated acidosis. BMD and serum bone turnover parameters among the subgroups of patients with COPD were determined and compared.

2133

MATERIALS AND METHODS Study subjects Seventy-one male patients with COPD were recruited from a single pneumology unit at the University Hospital of the Graz Medical School during December 1994 to January 1996. COPD was diagnosed according to the criteria defined by the World Health Organization (WHO; i.e., decreased pulmonary functions and exhibition of clinical symptoms, e.g., a history of chronic coughing, wheezing, expectoration, and dyspnea during mild physical activities). No subject had received any forms of therapy for at least 3 months before enrollment. Specifically, no one had received any drugs that could affect bone metabolism (e.g., calcitonins, bisphosphonates, fluorides, androgens, Ca, thiazides, or thyroid hormones) for more than 3 months within the past 3 years. Moreover, no subject had received corticosteroids (oral, inhalative, or injection) previously. Patients with primary or tertiary hyperparathyroidism were excluded, but patients with secondary hyperparathyroidism (i.e., elevated serum intact parathyroid hormone [iPTH], normal serum total and ionized Ca, and decreased serum phosphorus [PO⫺2 4 ]) were not excluded unless the cause was associated with other clinical conditions than COPD (e.g., malabsorption). Patients with elevated serum or urine creatinine or with diabetes mellitus were excluded also, but patients with ␣1-antitrypsin deficiency were not. For comparison, a control group of 40 apparently healthy male subjects who matched the patients with COPD in age, weight, and BMI were recruited from the same clinic. The control subjects did not have any apparent health problems that would affect bone and mineral metabolism. The study protocol (which followed the 1989 revision of the Declaration of Helsinki) was approved by the Institutional Review Board of the University of Graz Medical School, and signed informed consent was obtained from each subject before participation.

Assessment of pulmonary function of patients Spirometric parameters of the patients were obtained with a body plethysmograph (MGC 1070; Medical Graphics Corp., St. Paul, MN, USA). All values were averages of three independent measurements performed during the initial admission of the patient to the pneumology unit.

Physical activity index Each subject was asked the average number of hours per day spent in five levels of physical activity over the past 15 years: (1) basal activity (e.g., sleeping or lying down), (2) sedentary activity (e.g., sitting or standing, reading, listening to music, or watching television), (3) slight activity (e.g., walking on level ground or window shopping), (4) moderate activity (e.g., gardening, carpentry, dancing, or playing golf), and (5) heavy activity (e.g., shoveling or digging, tennis, swimming laps, or lawn mowing). The physical activity index (PAI) was calculated according to Young et al.(27)

2134

All subjects in this study also were asked for previous experience in immobilization and/or hospitalization. No subjects had suffered from any ailments previously that required long-term (i.e., ⬎1 month) hospitalization and/or immobilization.

BMD analysis BMD was measured at the distal forearm of the nondominant arm, using peripheral quantitative computed tomography (pQCT; XCT 900; Stratec Medizintechnik GmbH, Pforzheim, Germany). The densitometer was calibrated daily with a reference phantom. Repeated measurements showed a CV of ⬍1%. The total and trabecular BMDs also were reported in T scores (i.e., comparison with the peak BMD of normal subjects from central Europe). The STRATEC software (Stratec Medizintechnik GmbH), which calculated the T scores, did not provide T scores for cortical BMD.

DIMAI ET AL.

assayed with a two-site immunoradiometric assay (Nichols Institute, San Juan Capistrano, CA, USA). 25(OH)D3 was assayed with a commercial RIA kit (Immunodiagnostic Systems–IDS, Boldon, UK). Calcitonin was measured with a commercial RIA kit (Diagnostic System Laboratories, Webster, TX, USA). IL-1␤, IL-2, IL-6, and TNF-␣ were determined with each respective immunoradiometric assay (IRMA; Medgenix Diagnostics SA, Fleurus, Belgium). IL-8 was measured with an ELISA (Bender MedSystems, Vienna, Austria). Serum ␣1-antitrypsin was measured by laser nephelometry (Behringwerke AG, Marburg, Germany) based on a reference preparation for proteins in human serum. Histamine was analyzed with RIA (Pharmacia, Uppsala, Sweden). Neopterin was determined with the IMMUtest Neopterin RIA (Henning Berlin GmbH, Berlin, Germany). Myeloperoxidase (MPO) was assayed with RIA (Pharmacia). Intercellular adhesion molecule 1 (ICAM-1) was measured with ELISA (Bender MedSystems). Angiotensin converting enzyme was measured with a radioenzymatic assay (Buhlmann Laboratories AG, Basel, Switzerland).

Blood sample collection All blood samples were drawn between 7:00 and 9:00 a.m. with an overnight fast. Arterial blood samples were obtained from the radial artery. Freshly drawn arterial blood samples were used immediately for determination of pH, blood gases, and Ca2⫹ and magnesium (Mg2⫹). Venous blood samples were collected immediately after collection of arterial blood samples to prepare serum samples. Total Ca and Mg concentrations were measured in the serum on the same day of collection. Remaining serum samples were kept frozen at ⫺20°C until assay. Serum osteocalcin, iPTH, 25-hydroxyvitamin D3 [25(OH)D3], carboxy-terminal cross-linked telopeptide of type I collagen (ICTP), proinflammatory cytokines (i.e., interleukin-1 [IL-1], IL-2, IL-6, IL-8, and tumor necrosis factor ␣ [TNF-␣]), and other serum markers related to COPD were measured within 2 months. Aliquots of venous blood were covered with tin foil to avoid light exposure for the neopterin assay.

Analytical methods Arterial blood was handled under anaerobic conditions using a micropuncture set (AVL Microsampler, AVL Austria, Graz, Austria) designed for blood pH/gas, oximetry, and electrolyte measurements. pH, PO2, and PCO2 were measured within 30 s of blood collection using an AVL 995-HB automated blood gas analyzer (AVL Austria), and [HCO⫺ 3] was calculated. Blood Ca2⫹ and Mg2⫹ concentrations were measured immediately with ion-selective, flow-through, liquid-membrane electrodes on an automated electrolyte analyzer (AVL 988 – 4 Electrolyte Analyzer; AVL Austria).(28) Total serum Ca and Mg were assayed by the cresolphthalein complexone method(29) and the xylidyl blue method,(30) respectively, with a Hitachi 717 spectrophotometer (Boehringer Mannheim, Mannheim, Germany). ICTP was measured with the Telopeptide Radioimmunoassay (RIA; Orion Diagnostica, Espoo, Finland). Osteocalcin was determined by the OSTKR-PR RIA (CIS BioInternational, ORIS Group, Cedex, France). iPTH was

Statistical analyses Results are shown as mean ⫾ SEM. Statistical significance of the difference between groups was determined with analysis of variance (ANOVA) followed by Tukey post hoc analysis. Multiple correlation analysis was performed with the Pearson correlation matrix test. Multiple linear regression analyses were performed to determine the partial correlation coefficient of multiple factors to assess relative contribution of each factor on a given parameter. Statistical analyses were performed with the SYSTAT statistical program (SYSTAT, Inc., Evanston, IL, USA). The difference was statistically significant when p ⬍ 0.05. Power analysis, performed with the PASS 6.0 program (NCSS, Kaysville, UT, USA), indicated that the statistical power to detect a 20% difference in BMD between 71 patients with COPD and 40 controls was 0.995 with an ␣ of 0.01 and a ␤ of 0.005 and the power between the 35 COPD patients with hypercapnia and 36 eucapnia was 0.921 with an ␣ of 0.05 and a ␤ of 0.079. The power to detect a 20% difference in BMD between the subgroup of COPD patients with hypercapnia with compensated respiratory acidosis (n ⫽ 20) and the subgroup with uncompensated respiratory acidosis (n ⫽ 15) was reduced to 0.649 with an ␣ of 0.05 and a ␤ of 0.351.

RESULTS Physical characteristics, blood chemistries, and BMD of the study population Table 1 summarizes the physical characteristics of the study population, which consisted of 71 male patients with COPD who had received no therapy for the disease for at least 3 months before the study and 40 healthy male control subjects. The control subjects matched the patients with COPD in age, weight, and BMI. As expected, the ventilatory functions (i.e., percent forced expiratory volume in 1 s per vital capacity [FEV1%/VC]) of the patients with COPD were significantly (p ⬍ 0.001) reduced compared with that


2135

TABLE 1. CHARACTERISTICS

OF THE

STUDY POPULATIONS

Parameter (mean ⫾ SEM)

Age- and weight-matched controls (n ⫽ 40)

Patients with untreated COPD (n ⫽ 71)

Age BMI Height (cm) Weight (kg) PAI FEV1%/VC Pack-years Total BMD (g/cm3) T score Trabecular BMD (g/cm3) T score Cortical BMD (g/cm3) Blood pH Blood PO2 (mm Hg) Blood PCO2 (mm Hg) Blood HCO⫺ 3 (mmol/liter) Ca Total (mmol/liter) Ionized (mmol/liter) Mg Total (mmol/liter) Ionized (mmol/liter) PO⫺2 (mmol/liter) 4 iPTH (ng/liter) 25(OH)D3 (␮g/liter) Calcitonin (ng/liter) Osteocalcin (␮g/liter) ICTP (␮g/liter)

64.9 ⫾ 1.5 25.2 ⫾ 0.4 175.4 ⫾ 1.3 77.5 ⫾ 1.5 28.6 ⫾ 0.7 79.5 ⫾ 0.7 9.8 ⫾ 2.1 327.0 ⫾ 7.5 ⫺0.058 ⫾ 0.157 210.4 ⫾ 7.5 ⫺0.225 ⫾ 0.238 418.7 ⫾ 11.9 7.398 ⫾ 0.003 90.6 ⫾ 0.7 39.6 ⫾ 0.3 23.8 ⫾ 0.2

64.5 ⫾ 0.9 25.8 ⫾ 0.4 171.9 ⫾ 0.8* 76.4 ⫾ 1.4 28.5 ⫾ 0.3 55.6 ⫾ 1.9‡ 32.6 ⫾ 3.7‡ 249.4 ⫾ 7.8‡ ⫺1.628 ⫾ 0.168‡ 141.4 ⫾ 5.8‡ ⫺2.297 ⫾ 0.182‡ 335.5 ⫾ 10.0‡ 7.377 ⫾ 0.004‡ 65.3 ⫾ 1.4‡ 47.0 ⫾ 1.1‡ 24.7 ⫾ 0.2†

* p ⬍ 0.05;

†

p ⬍ 0.01;

‡

2.362 ⫾ 0.026 1.205 ⫾ 0.011

2.333 ⫾ 0.014 1.195 ⫾ 0.007

0.938 ⫾ 0.025 0.620 ⫾ 0.015 0.7 ⫾ 0.1 40.7 ⫾ 2.0 24.1 ⫾ 1.7 18.9 ⫾ 1.3 9.5 ⫾ 0.5 3.96 ⫾ 0.31

0.804 ⫾ 0.013‡ 0.557 ⫾ 0.011‡ 0.9 ⫾ 0.0‡ 45.7 ⫾ 2.6 21.2 ⫾ 1.0 34.1 ⫾ 2.1‡ 7.4 ⫾ 0.5† 5.19 ⫾ 0.24†

p ⬍ 0.001 (compared with controls).

of control subjects. The primary etiology of the disease in these patients was cigarette smoking, because patients with COPD consumed significantly more cigarettes than controls. Although ␣1-antitrypsin deficiency is a potential cause of COPD, only 4 of the patients were ␣1-antitrypsin deficient. No subject suffered from severe malnutrition. Over the past 15 years, the PAI of the patients with COPD was not different from that of the age- and weight-matched controls. We did not analyze the recent physical activities (e.g., within the past 6 months) of this study population. However, it is highly likely that the recent PAI of these patients with COPD was lower than that of the healthy control subjects, because a large number of the COPD patients had become seriously ill only within the past few months before admittance. On the other hand, we reasoned that long-term physical activity probably would have a more important impact on BMD than recent short-term changes in physical activity. Thus, we only analyzed PAI over the past 15 years rather than a more recent short-term period to emphasize that the physical activities of the patients with COPD before the onset of symptoms of the disease were not different from those of the control subjects. The COPD patients had significantly lower total trabecular and cortical BMD compared with control subjects (Table 1). According to the WHO definition, 20 of the 71

patients with COPD (28%) were considered osteoporotic (T scores ⬍ ⫺2.5) and 30 (42%) were osteopenic (T scores between ⫺1.0 and ⫺2.5). In contrast, only 6 of the 40 control subjects (15%) had BMD values that were considered mildly osteopenic (T scores slightly ⬍ ⫺1.0) and none was osteoporotic. These patients with COPD showed evidence for hypoxemia and hypercapnia because they, as a group, had significantly lower arterial PO2 (by 28%) and higher arterial PCO2 (by 19%) compared with control subjects (Table 1). Although the patients also had a slightly elevated [HCO⫺ 3 ] (by 4%), the arterial pH correlated significantly with PCO2 but not with [HCO⫺ 3 ], indicating that they suffered from respiratory acidosis rather than metabolic acidosis. There was a highly significant correlation between total BMD (T scores) and arterial pH (Fig. 1A) and between BMD and PCO2 (Fig. 1B) in these COPD patients, suggesting that the uncompensated respiratory acidosis and hypercapnia are two important contributing factors to the bone loss in these patients with COPD. A very similar correlation between trabecular BMD (T scores) and arterial pH or PCO2 was observed (data not shown). Neither the total BMD (Fig. 1C) nor the trabecular BMD correlated with [HCO⫺ 3 ] in these patients with COPD. Serum total Ca and blood Ca2⫹ in the patients with COPD (as a group) were not different from those of the

2136

DIMAI ET AL.

NS) (Fig. 1). There also was no significant correlation between serum ICTP and osteocalcin (r ⫽ ⫺0.092; p ⫽ NS) in these patients with COPD. Multiple linear regression analysis shows that serum ICTP but not serum osteocalcin had a significant (p ⫽ 0.003) partial correlation coefficient with the BMD, suggesting that increased bone resorption probably is a more important contributor to the bone loss than suppression of bone formation in these patients with COPD.

Effects of COPD-associated hypercapnia on BMD and bone metabolism

FIG. 1. Scatterplot of total BMD (in T scores) versus (A) arterial pH, (B) arterial PCO2, (C) arterial [HCO⫺ 3 ], (D) the serum bone resorption markers serum ICTP, (E) and the serum bone formation marker serum osteocalcin.

controls. Serum total Mg and blood Mg2⫹ of patients with untreated COPD were lower (by 14% and 10%, respectively; p ⬍ 0.001 for each) than those of control subjects. Serum PO⫺2 4 and calcitonin of the patients with COPD were increased (by 33% and 80%, respectively; p ⬍ 0.001 for each), but neither serum iPTH nor serum 25(OH)D3 in patients with COPD was significantly different from that of the control group. Serum ICTP, a bone resorption marker, was elevated (by 31%; p ⬍ 0.01), whereas serum osteocalcin, a bone formation marker, was decreased (by 22%; p ⬍ 0.01) compared with control subjects (Table 1), suggesting that these patients with untreated COPD, as a group, could have an increase in bone resorption along with a decrease in bone formation. Markers of inflammation (e.g., neopterin and MPO) and serum proinflammatory cytokines (e.g., IL-1, IL-2, and TNF-␣) were increased in the patients with COPD (data not shown), supporting the premise that these patients had experienced respiratory inflammation.

Relationship between bone turnover and bone loss in patients with untreated COPD There was a strong, negative correlation between total BMD and serum ICTP (r ⫽ ⫺0.444; p ⬍ 0.001) but not between total BMD and serum osteocalcin (r ⫽ 0.111; p ⫽

To evaluate if hypercapnia would negatively affect BMD and bone metabolism, the patients with COPD were divided into two groups: patients with hypercapnia (i.e., arterial PCO2 ⬎ 45 mm Hg) and patients with eucapnia (i.e., arterial PCO2 was between 35 and 45 mm Hg). Of the 71 patients with untreated COPD, 35 displayed the symptom of hypercapnia. The two groups of patients with COPD did not differ in age, BMI, weight, and PAI (Table 2). Both groups of COPD patients exhibited similar significant reduction in reduced FEV1%/VC compared with the control group. The total BMD of the two groups of patients with COPD was significantly lower than that of controls. However, the reduction in total BMD was significantly (p ⬍ 0.001) more severe in patients with hypercapnia than in those patients without the hypercapnia (Fig. 2). A more severe (p ⬍ 0.001) reduction in trabecular BMD also was seen in patients with hypercapnia than that seen in patients with eucapnia (data not shown). Table 2 shows that patients with COPD with hypercapnia had an elevated arterial PCO2, a corresponding decrease in PO2, and a small increase in [HCO⫺ 3 ] compared with patients with eucapnia and also compared with control subjects. Because of the increased PCO2, patients with COPD with hypercapnia, as a group, exhibited significantly lower arterial pH than the group with eucapnia. In contrast, the arterial pH and PCO2 in patients with COPD with eucapnia were comparable with those of controls. None of the 36 patients with COPD with eucapnia had an arterial pH ⱕ 7.35. Seven patients with eucapnia and 5 patients with hypercapnia had developed secondary hyperparathyroidism. Although total Ca and Ca2⫹ concentrations of either COPD group were not different from those of the control group, the Ca2⫹ concentration in the group with hypercapnia was significantly (p ⬍ 0.05) lower than that in patients with normal PCO2. There were no significant differences in serum total Mg, Mg2⫹, serum PO⫺2 4 , iPTH, and 25(OH)D3 among the two groups of patients with COPD. Conversely, serum calcitonin of the group with hypercapnia was significantly (p ⬍ 0.05) lower than that of the group with eucapnia. Serum concentration of each proinflammatory cytokine was not significantly different between the two groups of patients with COPD (data not shown). Although serum ICTP was markedly elevated in the group with hypercapnia compared with the control group, it was only slightly but not significantly increased in the group with eucapnia. Serum osteocalcin of the group with hypercapnia was not signifi-


TABLE 2. COMPARISON

OF

2137

PHYSICAL CHARACTERISTICS OF PATIENTS WITH UNTREATED COPD WITH EUCAPNIA TO THOSE PATIENTS WITH HYPERCAPNIA

Parameter (mean ⫾ SEM)

Patients with untreated COPD with eucapniaa (n ⫽ 36)

Patients with untreated COPD with hypercapniab (n ⫽ 35)

Age BMI Height (cm) Weight (kg) PAI FEV1%/VC Pack-years Blood pH Blood PO2 (mm Hg) Blood PCO2 (mm Hg) Blood HCO⫺ 3 (mmol/liter) Ca Total (mmol/liter) Ionized (mmol/liter) Mg Total (mmol/liter) Ionized (mmol/liter) PO⫺2 (mmol/liter) 4 iPTH (ng/liter) 25(OH)D3 (␮g/liter) Calcitonin (ng/liter) Osteocalcin (␮g/liter) ICTP (␮g/liter)

63.9 ⫾ 1.3 26.5 ⫾ 0.6 170.4 ⫾ 0.9* 76.6 ⫾ 2.0 28.7 ⫾ 0.5 55.0 ⫾ 3.0‡ 37.9 ⫾ 6.6‡ 7.396 ⫾ 0.004 68.2 ⫾ 2.0‡ 40.3 ⫾ 0.4 23.9 ⫾ 0.2

65.2 ⫾ 1.4 25.0 ⫾ 0.6 173.5 ⫾ 1.3 76.1 ⫾ 2.1 28.4 ⫾ 0.4 56.3 ⫾ 2.5‡ 27.2 ⫾ 3.2* 7.357 ⫾ 0.004‡,¶ 62.3 ⫾ 1.9‡,§ 54.0 ⫾ 1.3‡,¶ 25.6 ⫾ 0.3‡,¶

2.343 ⫾ 0.019 1.212 ⫾ 0.008

2.323 ⫾ 0.021 1.177 ⫾ 0.010§

0.805 ⫾ 0.018‡ 0.569 ⫾ 0.015* 1.0 ⫾ 0.0‡ 41.8 ⫾ 4.1 23.2 ⫾ 1.7 38.7 ⫾ 3.1‡ 7.8 ⫾ 0.7 4.26 ⫾ 0.25

0.804 ⫾ 0.020‡ 0.545 ⫾ 0.017† 0.9 ⫾ 0.1* 49.6 ⫾ 3.0 19.2 ⫾ 1.0 29.3 ⫾ 2.6†,§ 7.0 ⫾ 0.6† 6.10 ⫾ 0.34‡,¶

a

Clinically COPD patients with normal arterial PCO2 (between 35– 45 mm Hg). Clinically COPD patients with elevated arterial PCO2 ⬎ 45 mm Hg. * p ⬍ 0.05; † p ⬍ 0.01; ‡ p ⬍ 0.001 (compared with age- and weight-matched controls in Table 1); § p ⬍ 0.05; ¶ p ⬍ 0.001 (compared with patients with untreated COPD with eucapnia). b

with COPD with hypercapnia compared with patients without the hypercapnia (Fig. 2).

Association between bone resorption and hypercapnia in patients with COPD

FIG. 2. Comparison of total (left panel) and trabecular (right panel) BMD among the subgroup of patients with untreated COPD with eucapnia (middle bars), the subgroup of patients with untreated COPD with hypercapnia (right bars), and age- and weight-matched control subjects (left bars).

cantly different from that of the subgroup of patients with COPD without hypercapnia (Table 2). The apparently bigger effect of hypercapnia on bone turnover corresponded well with the lower BMD (i.e., greater bone loss) in patients

Multiple correlation analysis between serum ICTP or serum osteocalcin and blood gas parameters shows that serum ICTP correlated significantly with arterial pH, PCO2, or [HCO⫺ 3 ] but not with PO2. No correlation between serum osteocalcin and any test blood gas parameter was observed (Table 3). Multiple linear regression analyses of the relationship between serum ICTP and the blood parameters reveals that PCO2 but not the other blood gas parameters (pH, PO2, or [HCO⫺ 3 ]) had a significant partial correlation coefficient (p ⫽ 0.003). These findings suggest that among the test blood parameters, hypercapnia might be the most important contributing factor to the increased bone resorption in these patients with COPD. Multiple correlation analysis of the association between serum ICTP and circulating concentrations of several known effectors of bone resorption, such as Ca2⫹, Mg2⫹, iPTH, 25(OH)D3, PO⫺2 4 , calcitonin, or resorptive cytokines (IL-1, IL-2, IL-6, IL-8, and TNF-␣) in these patients with COPD shows that serum ICTP exhibited a significant negative correlation with blood Ca2⫹ (r ⫽ ⫺0.325; p ⫽ 0.006)

2138

DIMAI ET AL.

TABLE 3. MULTIPLE CORRELATION ANALYSESa BETWEEN SERUM BONE TURNOVER MARKERS AND BLOOD GAS PARAMETERS IN 71 PATIENTS WITH COPD Parameter Arterial pH Arterial PCO2 Arterial PO2 Arterial HCO⫺ 3

Serum ICTP

Serum osteocalcin

⫺0.415 (p ⬍ 0.001) 0.593 (p ⬍ 0.001) ⫺0.196 (p ⫽ NS) 0.405 (p ⫽ 0.001)

0.118 (p ⫽ NS) ⫺0.208 (p ⫽ NS) 0.109 (p ⫽ NS) ⫺0.241 (p ⫽ 0.042)

b

a Multiple correlation analyses were performed with the Pearson correlation matrix test. b Correlation coefficient (r values).

but not with other test parameters. However, because there was no significant difference in blood Ca2⫹ concentrations between the patients with COPD and the control subjects; we tentatively conclude that blood Ca2⫹ probably was not the major factor in the COPD-associated stimulation of bone resorption.

Effects of uncompensated respiratory acidosis on BMD and bone metabolism Multiple linear regression analysis reveals that the partial correlation coefficient of arterial pH (p ⬍ 0.001) but not that of PO2, PCO2, or [HCO⫺ 3 ] was significant in association with BMD (data not shown). Thus, we next subdivided the patients with COPD with hypercapnia into two subgroups: one with uncompensated respiratory acidosis (blood pH ⱕ 7.35 and blood PCO2 ⬎ 45 mm Hg) and one that apparently had compensated respiratory acidosis (blood pH ⬎ 7.35 and blood PCO2 ⬎ 45 mm Hg). Of the 35 patients with COPD with hypercapnia, 15 had uncompensated respiratory acidosis. The two subgroups of patients with COPD with hypercapnia did not differ in the age, BMI, weight, height, cigarette consumption, or PAI (data not shown). The FEV1%/VC also was not significantly different between the uncompensated acidosis subgroup (52.8 ⫾ 3.5%) and the compensated acidosis group (59.1 ⫾ 3.1%). The BMD in both subgroups was significantly lower than that of control subjects but it was not statistically significant between the two subgroups (Fig. 3). Both subgroups of patients had significantly lower arterial pH, lower PO2, elevated PCO2, increased [HCO⫺ 3 ], reduced serum total Mg and blood Mg2⫹, and elevated serum IL-2 when compared with the age- and weight-matched control subjects (Table 4). It is interesting to note that, other than the arterial pH (which was expected), none of the other test parameters was significantly different between the two subgroups. More importantly, serum ICTP in the subgroup with uncompensated respiratory acidosis was not significantly different from that in the subgroup with compensated respiratory acidosis. However, we should emphasize that the number of patients

FIG. 3. Comparison of total (left panel) and trabecular (right panel) BMD among the subgroup of patients with untreated COPD with hypercapnia with compensated respiratory acidosis (middle bars), the subgroup of patients with untreated COPD with hypercapnia with uncompensated respiratory acidosis (right bars), and age- and weightmatched control subjects (left bars).

in these two subgroups was rather small (n ⫽ 15 and 20, respectively). It is likely that there may not be sufficient statistical power to detect small differences in some of the test parameters.

DISCUSSION In this study, a majority of the 71 patients with untreated COPD showed characteristics of hypercapnia and respiratory acidosis and had lower BMD (in absolute density or T scores) compared with a group of 40 healthy subjects who matched the patients in age, weight, sex, and BMI. This confirms the previous finding that the patients with COPD lost a significant amount of bone.(25,26) Bone loss could result from an increase in bone resorption, a reduction of bone formation, or both. Three observations in this study support our tentative conclusion that increased bone resorption was a more important contributing factor to the bone loss than reduced bone formation in these patients with COPD. First, there was a strong, negative correlation between BMD and the resorption marker, serum ICTP. In contrast, the BMD in these patients did not correlate with the bone formation marker, serum osteocalcin. Second, multiple linear regression analysis indicates that serum ICTP but not serum osteocalcin showed a highly significant partial correlation coefficient with the BMD. Third and more importantly, patients with COPD with hypercapnia who lost much more BMD than patients with eucapnia had significantly elevated serum ICTP (bone resorption) when compared with patients with eucapnia. Serum osteocalcin (bone formation) in patients with COPD with hypercapnia was not different from that of patients with COPD who had eucapnia. However, we should emphasize that this group of


2139

TABLE 4. COMPARISON OF BLOOD PARAMETERS OF PATIENTS WITH UNTREATED COPD WITH HYPERCAPNIA WITH COMPENSATED ACIDOSIS WITH THOSE WITH UNCOMPENSATED ACIDOSIS Parameter Blood pH Blood PO2 (mm Hg) Blood PCO2 (mm Hg) Blood HCO⫺ 3 (mmol/liter) Ca Total (mmol/liter) Ionized (mmol/liter) Mg Total (mmol/liter) Ionized (mmol/liter) PO⫺2 (mmol/liter) 4 iPTH (ng/liter) 25(OH)D3 (␮g/liter) Calcitonin (ng/liter) Osteocalcin (␮g/liter) ICTP (␮g/liter)

Patients with COPD with hypercapnia with compensated acidosisa (n ⫽ 20)

Patients with COPD with hypercapnia with uncompensated acidosisb (n ⫽ 15)

7.373 ⫾ 0.003* 62.5 ⫾ 2.6* 52.9 ⫾ 1.7* 25.7 ⫾ 0.5*

7.335 ⫾ 0.003*,§ 62.2 ⫾ 3.0* 55.4 ⫾ 2.2* 25.4 ⫾ 0.4†

2.304 ⫾ 0.032 1.174 ⫾ 0.013

2.349 ⫾ 0.025 1.181 ⫾ 0.017

0.782 ⫾ 0.022* 0.521 ⫾ 0.020* 1.0 ⫾ 0.1† 55.0 ⫾ 4.1‡ 18.9 ⫾ 1.2 33.2 ⫾ 3.8† 6.7 ⫾ 0.7‡ 5.68 ⫾ 0.33†

0.834 ⫾ 0.037‡ 0.577 ⫾ 0.030 0.8 ⫾ 0.1 42.5 ⫾ 4.7 19.5 ⫾ 1.9 24.1 ⫾ 3.2 7.4 ⫾ 1.1 6.67 ⫾ 0.66*

Clinically COPD patients with elevated PCO2 ⬎ 45 mm Hg but normal blood pH (7.35–7.45). Clinically COPD patients with elevated PCO2 ⬎ 45 mm Hg and acidosis (blood pH ⱕ 7.35). * p ⬍ 0.001; † p ⬍ 0.01; ‡ p ⬍ 0.05 (compared with age- and weight-matched subjects in Table 1); § p ⬍ 0.001; (compared with patients with compensated respiratory acidosis). a

b

COPD patients was rather heterogeneous and the etiology of the bone loss in such a heterogeneous group of patients probably is multifactorial. In this regard, we cannot exclude the possibility that the bone loss in some patients also might be caused by reduced bone formation, because a number of patients with COPD in this study also had reduced serum osteocalcin, in addition to elevated serum ICTP (compared with control subjects). Nevertheless, because of the strong correlation between serum ICTP and the BMD and the lack of a correlation between serum osteocalcin and the BMD, we tentatively conclude that increased bone resorption probably is a more important contributor to the bone loss than reduced bone formation in these patients with COPD. The etiology of the apparent increase in bone resorption in patients with untreated COPD is unknown. Airway inflammatory symptoms of COPD caused increases in serum concentrations of proinflammatory cytokines.(31) Although many of these cytokines are potent stimulators of bone resorption,(32) we do not believe that these cytokines were likely to be major contributors to the observed increase in bone resorption and/or bone loss, because of the lack of correlation between any of these cytokines and serum ICTP or BMD in these patients with COPD. Low serum Ca2⫹ is a potent stimulator of bone resorption.(33) However, there was no significant correlation between serum Ca2⫹ and serum ICTP or between serum Ca2⫹ and the BMD in this group of patients with COPD. In addition, there was no significant decrease in blood Ca2⫹ in the patients with COPD compared with the age- and weightmatched controls. This lack of an effect of COPDassociated hypercapnia and/or respiratory acidosis on circulating Ca2⫹ is consistent with the previous findings that chronic respiratory acidosis and hypercapnia had no signif-

icant effect in blood Ca2⫹ in the rat.(34) Moreover, the action of hypocalcemia to stimulate bone resorption is mediated largely through an increase in PTH secretion.(35) Although a total of 12 out of 71 patients with COPD (17%) had developed secondary hyperparathyroidism, the serum iPTH in this patient population, as a group, was not significantly increased. This finding is consistent with the previous observation that chronic hypercapnia and respiratory acidosis did not affect serum iPTH and vitamin D metabolites in the rat.(34) Moreover, there was no significant correlation between serum iPTH and the BMD or serum iPTH and serum ICTP. Importantly, the serum iPTH in patients with COPD with hypercapnia who lost significantly more BMD and exhibited a more elevated serum ICTP (i.e., bone resorption) than patients with eucapnia was not significantly different from that in patients with COPD with eucapnia. Together, these findings argue strongly against a direct role for either hypocalcemia or secondary hyperparathyroidism in the increased bone resorption in this COPD patient population. In this study, a number of patients with COPD had an elevated serum concentration of calcitonin, which is a potent inhibitor of bone resorption.(36) Increased calcitonin production should cause an inhibition, rather than the observed stimulation, of bone resorption. Thus, the increased serum calcitonin is not likely to be responsible for the increase in bone resorption and/or the BMD loss in these patients with untreated COPD. On the other hand, increased bone resorption frequently led to an increased secretion of calcitonin. Accordingly, it is possible that the observed increase in serum calcitonin was the consequence, rather than the cause, of an increased bone resorption in the patients with COPD. Serum Mg2⫹ also has been suggested

2140

to have a regulatory role in bone and mineral metabolism because Mg deficiency impaired bone growth and mineralization.(37) Although the patients with untreated COPD had a significantly lower blood Mg2⫹ concentration compared with control subjects, the blood Mg2⫹ did not correlate with serum ICTP or the BMD in this COPD patient population. Moreover, we previously reported that decreased blood Mg2⫹ is associated with a reduction rather than an enhancement of bone resorption in healthy adult young men.(38) Therefore, we do not believe that the COPD-associated reduction in blood Mg2⫹ is a major cause of increased bone resorption in this COPD patient population. There is circumstantial evidence that the etiology of the COPD-associated increase in bone resorption and the consequent bone loss could be in part associated with the chronic hypercapnia. Accordingly, there was a highly significant inverse correlation between the BMD at the distal forearm and arterial PCO2 in these patients. The fact that the subgroup of patients with untreated COPD who exhibited hypercapnia had significantly lower BMD and higher serum ICTP than age- and weight-matched control subjects further supports an association between hypercapnia and the increased bone resorption (and bone loss) associated with untreated COPD. Consequently, these findings are entirely consistent with the previous speculation that the bone loss in patients with COPD could be associated with chronic hypercapnia and/or the associated respiratory acidosis.(39) We should point out that there also was a significant positive correlation between the BMD and arterial pH, raising the possibility that uncompensated respiratory acidosis also might be involved in the COPD-related bone loss. However, patients with COPD who had uncompensated respiratory acidosis also had elevated arterial PCO2 (i.e., hypercapnia). Thus, the increased bone resorption and bone loss in patients with COPD with uncompensated respiratory acidosis could be caused by chronic hypercapnia rather than uncompensated respiratory acidosis. Two important observations support the contention that the increase in bone resorption (and the bone loss) might be associated more closely with chronic hypercapnia than with uncompensated respiratory acidosis. First, patients with untreated COPD who experienced chronic hypercapnia exhibited a much more significant increase in serum ICTP and bone loss compared with patients who did not have the symptom. Second, when the patients with hypercapnia were subdivided into two groups based on whether they had compensated or uncompensated respiratory acidosis, no significant difference in serum ICTP or the BMD was noted between the two subgroups. It is possible that the lack of difference could be caused by the lack of sufficient statistical power owing to the relatively small number of patients in the two subgroups. Nevertheless, even if there were a difference, the difference was probably small. Consequently, these observations raise the possibility that the uncompensated respiratory acidosis might not be as important a contributing factor as chronic hypercapnia to the increased bone resorption and bone loss in patients with untreated COPD. We can only speculate on the mechanism whereby chronic hypercapnia stimulates osteoclastic resorption. It is possible that osteoclasts might have a “sensor” or a “recep-

DIMAI ET AL.

tor” for CO2 in manners analogous to nitric oxide.(40) Alternatively, it has been shown that carbonic anhydrase II is involved in stimulation of osteoclastic resorption.(41) It is conceivable that increased blood concentration of CO2 might activate the carbonic anhydrase II in osteoclasts, leading to subsequent activation of osteoclastic resorption. We must emphasize that currently there is no evidence for either mechanism. Additional work is required to decipher the mechanism by which CO2 stimulates bone resorption. Finally, we should be reminded that patients with untreated COPD even without hypercapnia also have lost significant amounts of BMD, particularly at trabecular bone sites (Fig. 2). Accordingly, inasmuch as hypercapnia may be an essential factor contributing to the bone loss, it is obvious that factors other than hypercapnia also could cause bone loss in patients with untreated COPD. The identity of these factors has not been determined. However, potential candidates include cigarette smoking, inflammatory cytokines, hypocalcemia, increased stress, and reduced physical activity level. In addition, sex steroids are known to affect bone metabolism. We did not assess the gonadal function (e.g., testosterone levels) of the patients in this study. Thus, one cannot rule out the possibility that reduced gonadal function caused by the disease also could contribute to the bone loss. Much further work is needed to identify these factors. In summary, this report provides strong evidence that patients with untreated COPD could lose significant amounts of bone minerals at both cortical and trabecular bone sites. We provide associative evidence that the bone loss in patients with untreated COPD is at least in a major part mediated by an increase in bone resorption that is associated with chronic hypercapnia rather than uncompensated respiratory acidosis. In conclusion, this study shows for the first time that chronic hypercapnia could be a significant risk factor for osteoporosis.

ACKNOWLEDGMENTS The authors thank Ms. Lisa Stach for her excellent technical assistance in performing pQCT measurements and acknowledge the support from the Departments of Endocrinology, Nuclear Medicine, and Pneumology of the University of Graz and from the Musculoskeletal Disease Center of the Jerry L. Pettis Memorial V.A. Medical Center. This work was supported in part by funding from the University of Graz Medical School (H.P.D.) and by a research grant from the National Institute of Dental and Craniofacial Research (DE13097; K.-H.W.L.).

REFERENCES 1. Barzel US 1972 The buffering function of bone. J Oral Med 27:40 – 42. 2. Lemann J Jr, Lennon EJ 1972 Role of diet, gastrointestinal tract and bone in acid-base homeostasis. Kidney Int 1:275– 279. 3. Goldhaber P, Rabadjia L 1987 H⫹ stimulation of cell-mediated bone resorption in tissue culture. Am J Physiol 253:E90 –E98. 4. Krieger NS, Sessler NE, Bushinsky DA 1992 Acidosis inhibits osteoblastic and stimulates osteoclastic activity in vitro. Am J Physiol 262:F442–F448.

HYPERCAPNIA AND BONE LOSS 5. Bushinsky DA, Sessler NE 1992 Critical role of bicarbonate in calcium release from bone. Am J Physiol 263:F510 –F515. 6. Bushinsky DA, Wolbach W, Sessler NE, Mogilevsky R, LeviSetti R 1993 Physiochemical effects of acidosis on bone calcium flux and surface ion composition. J Bone Miner Res 8:93–102. 7. Bleich HL, Moore MJ, Lemann J Jr, Adams ND, Gray RW 1979 Urinary calcium excretion in human beings. N Engl J Med 301:535–541. 8. Sutton RAL, Wong NLM, Dirks JH 1979 Effects of metabolic acidosis and alkalosis on sodium and calcium transport in the dog kidney. Kidney Int 15:520 –533. 9. Kraut JA, Mishler DR, Singer FR, Goodman WG 1986 The effects of metabolic acidosis on bone formation and bone resorption in rat. Kidney Int 30:694 –700. 10. Barzel US 1969 The effect of excessive acid feeding on bone. Calcif Tissue Res 4:94 –100. 11. Lemann J Jr, Litzow JR, Lennon EJ 1966 The effects of chronic acid loads in normal man: Further evidence for the participation of bone mineral in the defense against chronic metabolic acidosis. J Clin Invest 45:1608 –1614. 12. Kraut JA 1995 The role of metabolic acidosis in the pathogenesis of renal osteodystrophy. Adv Ren Replace Ther 2:40 – 51. 13. Mora Palma FJ, Ellis HA, Cook DB, Dewar JH, Ward MK, Wilkinson R, Kerr DNS 1983 Osteomalacia in patients with chronic renal failure before dialysis or transplantation. QJM 52:332–348. 14. Cunningham J, Fraher LJ, Clemens TL, Revell PA, Papapoulos SE 1982 Chronic acidosis with metabolic bone disease. Effect of alkali on bone morphology and vitamin D metabolism. Am J Med 73:199 –204. 15. Barzel US 1976 Osteoporosis II: An overview. In: Barzel S (ed.) Osteoporosis II. Grune and Stratton, New York, NY, USA, pp. 1–12. 16. Canzanello VJ, Bodvarsson M, Kraut JA, Johns CA, Slatopolsky E, Madias NE 1990 Effect of chronic respiratory acidosis on urinary calcium excretion in the dog. Kidney Int 38:409 – 416. 17. Gray SB, Morris JEW, Brooks CJ 1973 Renal handling of calcium, magnesium, inorganic phosphate, and hydrogen ions during prolonged exposure to elevated carbon dioxide concentrations. Clin Sci Mol Med 45:751–764. 18. Schaefer KE 1982 Effects of increased ambient CO2 levels on human and animal health. Experientia 38:1163–1168. 19. Messier AA, Heyder E, Braithwaite WR, McCluggage C, Peck A, Schaefer KE 1979 Calcium, magnesium, and phosphorus metabolism and parathyroid-calcitonin function during prolonged exposure to elevated CO2 concentrations on submarines. Undersea Biomed Res 6(Suppl):S57–S70. 20. Bushinsky DA, Sessler NE, Krieger NS 1992 Greater unidirectional calcium efflux from bone during metabolic, compared with respiratory, acidosis. Am J Physiol 262:F425–F431. 21. Arnett TR, Boyde A, Jones SJ, Taylor ML 1994 Effects of medium acidification by alteration of carbon dioxide or bicarbonate concentrations on the resorptive activity of rat osteoclasts. J Bone Miner Res 9:375–379. 22. Erbland ML, Ebert RV, Snow SL 1990 Interaction of hypoxia and hypercapnia on respiratory drive in patients with COPD. Chest 97:1289 –1294. 23. Begin P, Grassino A 1991 Inspiratory muscle dysfunction and chronic hypocapnia in chronic obstructive pulmonary disease. Am Rev Resp Dis 143:905–912. 24. Aris RM, Neuringer IP, Weiner MA, Egan TM, Ontjes D 1996 Severe osteoporosis before and after lung transplantation. Chest 109:1176 –1183.

2141 25. Libanati CR, Baylink DJ 1992 Prevention and treatment of glucocorticoid-induced osteoporosis: A pathogenic perspective. Chest 102:1426 –1435. 26. Riancho JA, Gonzalez Macias J, Del Arco C, Amado JA, Freijanes J, Anton M 1987 Vertebral compression fractures and mineral metabolism in chronic obstructive lung disease. Thorax 42:962–966. 27. Young DR, Sharp DS, Petrovitch H, Curb JD 1995 Internal validity of the physical activity index over 26 years in middleaged and older men. J Am Geriatr Soc 43:999 –1006. 28. Hristova EN, Cecco S, Niemala JE, Rahak NN, Elin RJ 1994 AVL and NOVA analyzers compared for ionized calcium, ionized magnesium, pH, and sodium. Clin Chem 40:1093– 1099. 29. Zak B, Epstein E, Baginski ES 1975 Review of calcium methodologies. Ann Clin Lab Sci 5:195–215. 30. Chromy V, Svoboda V, Stepanova I 1973 Spectrophotometric determination of magnesium in biological fluids with xylidyl blue II. Biochem Med 7:208 –217. 31. Hasday JD, McCrea KA, Meltzer SS, Bleeker ER 1994 Dysregulation of airway cytokine expression in chronic obstructive pulmonary disease and asthma. Am J Resp Crit Care Med 150:S54 –S58. 32. Manolagas SC 1995 Role of cytokines in bone resorption. Bone 17:63S– 67S. 33. Thompson ER, Baylink DJ, Wergedal JE 1975 Increases in number and size of osteoclasts in response to calcium or phosphorus deficiency in the rat. Endocrinology 97:283–289. 34. Canzanello VJ, Kraut JA, Holick MF, Johns C, Liu CC, Madias NE 1995 Effect of chronic respiratory acidosis on calcium metabolism in the rat. J Lab Clin Med 126:81– 87. 35. Stauffer M, Baylink DJ, Wergedal JE, Rich C 1973 Decreased bone formation, mineralization and enhanced resorption in calcium deficient rats. Am J Physiol 225:269 –276. 36. Friedman J, Raisz LG 1965 Thyrocalcitonin, inhibitor of bone resorption in tissue culture. Science 150:1465–1467. 37. Sojka JE, Weaver CM 1995 Magnesium supplementation and osteoporosis. Nutr Rev 53:71– 80. 38. Dimai HP, Porta S, Wirnsberger G, Lindschinger M, Pamperl I, Dobnig H, Wilders-Truschnig M, Lau K-HW 1998 Daily oral magnesium supplementation suppresses bone turnover in young adult males. J Clin Endocrinol Metab 83:2742–2748. 39. Frassetto LA, Morris RC Jr, Sebastian A 1996 Effect of age on blood acid-base composition in adult humans: Role of agerelated renal functional decline. Am J Physiol 271:F1114 – F1122. 40. Mittal CK 1995 Oxygen-radical/nitric oxide mediate calciumdependent hormone action on cyclic GMP system: A novel concept in signal transduction mechanisms. Mol Cell Biochem 149 –150:257–262. 41. Laitala T, Vaananen HK 1994 Inhibition of bone resorption in vitro by antisense RNA and DNA molecules targeted against carbonic anhydrase II or two subunits of vacuolar H⫹-ATPase. J Clin Invest 93:2311–2318.

Address reprint requests to: K.-H. William Lau, Ph.D. Musculoskeletal Disease Center (151) Jerry L. Pettis Memorial V.A. Medical Center 11201 Benton Street Loma Linda, CA 92357, USA Received in original form October 16, 2000; in revised form March 7, 2001; accepted May 9, 2001.