METHODS
Study participants
A cross-sectional survey representative of the Korean general population (KNHANES IV to V) was carried out from 2007 to 2012 by the Division of Chronic Disease Surveillance, Korea Centers for Disease Control and Prevention. This study was based on data acquired from the second year (2008) of KNHANES IV to the first year (2010) of KNHANES V. The survey consisted of an evaluation of medical history, nutritional status, and health. After an initial interview at home, participants visited mobile centers, where they underwent an extensive physical examination, laboratory testing, BMD measurement, and spirometry.
A total of 29,965 adults over 20 years of age participated in the survey. The analysis included 4,501 subjects who were older than 50 years with complete data on both pulmonary function tests and BMD measurements. Regions were categorized as urban and rural. In the adult population, occupations were classified into two groups based on the different levels of sunlight exposure. Outdoor work included primary and secondary industries (agriculture, forestry and fishery, manual labor, engineering, assembling, and technical work), whereas indoor work consisted of tertiary industries, including managers, specialists, clerical workers, salespersons, service personnel, housewives, and students. Unemployed persons and others with unclassifiable work were grouped as indoor work. During the home interview, information was collected on a wide range of variables, including age, gender, smoking history, daily food intake, frequency of vitamin or mineral supplement intake, and frequency of undertaking a range of common physical activities. Dietary intakes of calcium (Ca) and phosphorous (P) were estimated using a 24-hour dietary recall method. High-risk drinking was defined as “yes,” when men consumed seven cups of alcohol or five cans of beer at one time, or when women consumed five cups of alcohol or three cans of beer at one time. Smoking status was categorized as “current smoker,” “former smoker,” and “never smoker.” Never smokers were defined as those who smoked fewer than five packs during their entire lives. Exercise was indicated as “yes,” when the subject exercised regularly at moderate or vigorous levels (> 30 minutes at a time and > 5 times per week, in the case of moderate physical activity, and > 20 minutes at a time and > 3 times per week, in the case of vigorous physical activity). Women who had not menstruated for 1 year were considered postmenopausal. Body weight and anthropometrics were measured with light clothes. BMI was calculated as weight divided by height squared (kg/m2).
Blood sampling
Blood samples were collected after an overnight fast and were immediately refrigerated and transported in cold storage to the Central Testing Institute (Neodin Medical Institute, Seoul, Korea). Samples were analyzed within 24 hours after transportation. The concentrations of plasma glucose, serum calcium, phosphorus, creatinine, alkaline phosphatase (ALP), total cholesterol, and parathyroid hormone (PTH) were measured using standard laboratory techniques. The serum levels of 25-hydroxyvitamin D (25[OH]D), which are considered the biomarker for vitamin D metabolic status, were measured using a radioimmunoassay (DISCOVERY-W, Hologic Inc., Bedford, MA, USA) equipped with APEX software (APEX Corporation Software, Pittsburg, PA, USA). PTH levels were determined by chemiluminescence (LIAISON, Diasorin Inc., Stillwater, MN, USA) using the N-tact PTH assay (Diasorin Inc.).
Body composition and BMD
Whole-body dual energy X-ray absorptiometry (DXA) was performed with a BDR Discovery fan beam densitometer (Hologic Inc.) located in the mobile examination centers, and analyses were performed using Hologic Discovery software version 13.1. BMD was expressed as an absolute value and as a T-score (standard deviations from a young, sex-specific reference mean BMD). Osteopenia is defined as a T-score between –1.0 and –2.5. Osteoporosis is defined as a T-score below –2.5. Percentage body fat (fat mass / total mass × 100) as well as BMD at the lumbar spine (L1–4) and proximal femur (total hip region) were measured using DXA. FFMI was calculated from DXA-measured fat-free mass in kilograms divided by height in meters squared (kg/m2).
Pulmonary function test
Spirometry was performed by specially trained technicians according to the 2005 American Thoracic Society and European Respiratory Society recommendations, using a dry rolling seal spirometer (SensorMedics Model 1022, Sensor Medics, Palm Springs, CA, USA). We analyzed only data from subjects with acceptable spirometry performance. Spirometric values were prebronchodilator measurements, and relative values were expressed as the percentage predicted from reference values. Values used in the study were forced vital capacity (FVC), forced expiratory volume in 1 second (FEV1), and FEV1 to FVC ratio (FEV1/FVC). We categorized the subjects by severity of airflow obstruction. Airway obstruction was deemed to be present when the FEV1/FVC was less than 0.7. When the FEV1/FVC was greater than 0.7, the subjects were categorized as “no obstruction.” Severity of airway obstruction was determined on the basis of FEV1 : mild obstruction, FEV1 ≥ 80% of predicted value; moderate obstruction, FEV1 < 80% and FEV1 ≥ 50% of predicted value; and severe obstruction FEV1 < 50% of predicted value.
Statistical analyses
Statistical analyses were performed using SAS version 9.2 (SAS Institute Inc., Cary, NC, USA). To correct the effects of gender on BMD, study participants were divided into two groups by sex and then separately analyzed. BMI was further divided into four categories, which corresponded to the International Classification suggested by the World Health Organization: underweight (BMI < 18.5), normal (18.5 ≤ BMI < 25), overweight (25 ≤ BMI < 30), and obese (BMI ≥ 30). Serum 25(OH)D levels were also divided into quartiles: < 16.69, 16.79 to 21.44, 21.14 to 26.32, and ≥ 26.32 ng/mL cutoff values were used for men; < 13.75, 13.75 to 17.90, 17.90 to 23.42, and ≥ 23.42 ng/mL cutoff values were used for women. To compare the baseline characteristics between males and females, the Student t test or the chi-square test was used for comparisons. Bivariate correlation analysis was used to evaluate the association between possible predictive factors and BMD. Multivariate linear regression analysis was used to evaluate the relationships between BMD and selected covariates that are considered clinically important risk factors for osteopenia. Continuous variables were expressed as the mean ± SD. All tests were two-sided, and p < 0.05 was considered statistically significant.
RESULTS
The baseline demographics of the study populations are given in
Table 1. The subjects were comprised of 2,100 (46.7%) males and 2,401 (53.3%) females. The mean age of males was 61.9 years, and the mean age of females was 62.8 years. Most participants were in their 50s. The mean BMI was 24.0 kg/m
2 in males and 24.3 kg/m
2 in females. Most subjects were of normal weight in both genders, followed by being overweight. All female subjects were postmenopausal. In both genders, the proportion living in an urban area was about 68.0%. The proportion of those with high-risk alcohol intake was 22.0% in males and 3.2% in females. Of males, 17.7% were never smokers compared to 92.4 % of female. The proportion of the regular exercise “yes” group was 28.3% in males and 22.8% in females (
Table 1).
The mean 25(OH)D level was 22.0 ng/mL in males and 19.1 ng/mL in females. The mean FEV
1 (% of predicted value) was 88.6% in males and 94.5% in females. While the no airway obstruction group comprised 71.8% of subjects, the mild obstruction group comprised 13.6% and the moderate to severe obstruction group was 14.6% of males; most female participants belonged to the no obstruction group (91.3%), followed by the moderate to severe obstruction group (5.2%) and the mild obstruction group (3.5%) (
Table 2).
The subjects who lived in urban areas appeared to have lower BMDs than those dwelling in rural areas, but this trend was statistically significant only in females. The indoor work group was shown to have significantly higher BMDs of the lumbar spine than the outdoor work group in both genders, whereas there was no significant correlation between occupation and BMD of the femur. The regular exercise “yes” group appeared to have higher BMDs than the “no” exercise group in both males and females, except for BMDs of the L-spine (
Table 3).
The BMD level was lower with increased age, number of smoking pack-years, and levels of PTH and ALP, whereas it was significantly higher with increased BMI, total FFMI, body fat percentage, daily calcium intake, and daily phosphorus intake in both genders. The levels of 25(OH)D did not correlate with BMDs of the femur neck, total femur, or lumbar spine. The FEV
1/FVC ratio was positively correlated with BMDs of all bone regions in both males and females, except for the BMD of the L-spine in males with FEV
1/FVC. In males, FEV
1 (% of predicted value) was increased with increased BMD of total femur, whereas the FEV
1(% of predicted value) did not correlate with any region in females. The FVC (% of predicted value) was lower with increasing BMDs of the lumbar spine in males, whereas the FVC (% of predicted value) was higher with increasing BMDs of the femur neck and total femur in females. BMDs of other regions did not show a significant correlation with FVC (% of predicted value) (
Table 4).
The proportions of patients with osteopenia and osteoporosis gradually increased according to the severity of airway obstruction in both males and females. In severe obstruction, the proportion of osteoporosis was 21% of males and 38% of females. Indeed, all female participants in the severe airway obstruction group had osteopenia and osteoporosis. The proportion of patients with osteoporosis was higher in females than in males at all stages of airway obstruction (
Fig. 1).
The total BMDs at different regions of the femur and lumbar spine in the “obstruction” group were significantly lower than for the “no obstruction” group in both genders, except for the total BMD of the lumbar spine in males (
Fig. 2).
To adjust confounding factors, data were analyzed by multivariate linear regression analysis adjusted for age, BMI, total FFMI, body fat percentage, smoking amount, exercise, PTH, 25(OH)D, daily intake of calcium and phosphorus, and ALP. Results revealed that BMDs of the femur neck, total femur, and lumbar spine in both genders were reduced significantly as age increased. When daily calcium intake increased, BMDs of all regions increased significantly in males. As ALP increased, BMDs of all regions s increased in both males and females. Severity of airway obstruction did not show a significant correlation with BMDs of any region in males or females (
Table 5).
DISCUSSION
In this population based, cross-sectional study, BMD was not lower significantly with decreased pulmonary function in the general population after adjustment. Associations between pulmonary function and osteoporosis have been conflicting in different studies [
7-
9]. A meta-analysis reported that the prevalence of osteoporosis was higher in COPD patients than in healthy subjects [
7]. Some researchers have reported that BMDs of the total hip and femoral neck significantly decreased with impairments in lung function [
8]. As airway obstruction becomes more severe, the proportion of patients with osteoporosis increases [
9]. Indeed, one study proposed that the presence of COPD is one of the parameters for increasing risk of osteoporosis in males, but many studies investigated COPD patients who came to the hospital for treatment, so study populations have consisted of moderate to severe airway obstruction. Shepherd et al. [
10] used data from the National Health and Nutrition Examination Survey III of the U.S. population. They validated the Male Osteoporosis Risk Estimation Score (MORES) that included three variables—age, weight, and history of COPD all showed excellent predictive value for developing osteoporosis [
10]. Lower FEV
1 showed a correlation with the degree of osteopenia regardless of glucocorticoid use, although for comparable degrees of pulmonary impairment, the subjects treated with prednisone had more frequent osteopenia. Indeed, a cross-sectional study investigated healthy women, and the BMDs of the lumbar spine and proximal femur revealed a significant positive correlation with FEV
1 in postmenopausal women [
11]. Two general population-based studies suggested that FEV
1 becomes higher with increasing BMD in middle-aged to older people. Additionally, this association remained significant after adjusting for age, weight, height, respiratory disease, cardiovascular diseases, smoking status, and medical history of steroid use and hormone replacement therapy. However, two studies did not adjust for daily vitamin intake, and it is known that serum 25(OH)D, PTH, and ALP levels can all influence BMD. These particular studies also did not quantify smoking amount or use of corticosteroids, and these variables may have dose-related influences on BMD or pulmonary function. Physical activity, alcohol consumption, and nutritional status were also not explored [
8,
12]. Sin et al. [
13] also analyzed data from Caucasian participants in the Third National Health and Nutrition Examination Survey, conducted in the United States between 1988 and 1994. Airflow obstruction was associated with increased odds of osteoporosis compared with the no airflow obstruction group (odds ratio, 1.9; 95% confidence interval, 1.4 to 2.5). The risk of osteoporosis increased in a linear “severity-dependent” fashion [
13]. These authors used quantified smoking status (pack-years) and surveyed medication usage (bronchodilators, inhaled or intranasal corticosteroids, estrogens, diuretics, etc.). BMD levels increased gradually with age, showing a peak in the third decade of life, and gradually decreasing thereafter. Their study populations were older than 20 years, and menopause in women was not reflected.
Contrary to many other studies, some researchers have failed to find statistically significant difference between BMD values of COPD and non-COPD groups [
14,
15]. Twenty-eight male COPD patients with a moderate degree of airway obstruction and 20 male volunteer subjects with normal lung function participated in the study. There was no difference in baseline characteristics (age, BMI, dietary parameters) and the level of physical activity between the two groups, except for smoking history. There was no statistically significant difference in the prevalence of osteoporosis between COPD patients and the control healthy group [
15]. These discrepancies may be derived from differences in the characteristics of subjects among the studies, such as race, age, gender, menopausal status, history of corticosteroid exposure, and other anthropometric variables. These parameters can contribute to the pathophysiology of osteoporosis in COPD patients [
15].
Our data from a large population-based sample with or without airflow obstruction are a nationally representative sample, as conducted by KNHANES, compared with other studies in which the data were collected from a local hospital. Thus, a different proportion of patients with severe lung function impairment is one reason for the conflicting results. As we analyzed the age group older than 50 years, we eliminated the influences of premenopause and inconsistencies associated with aging on BMD [
9]. Our data contained more detailed information on the demographic and socioeconomic factors, such as complete smoking history and physical activities, as measured by weekly, quantitative energy consumption, so we could analyze the data more deeply and could adjust for various confounding factors. After adjusting for confounding factors, including age, BMI, total FFMI, body fat percentage, smoking amount, exercise, PTH, 25(OH)D, daily intake of calcium and phosphorus, and ALP, the association between airway obstruction and BMD disappeared. These results suggest that the high prevalence of osteoporosis in COPD patients is associated with other factors and not with an impairment of lung function. The strengths of this study, compared with most previous studies of lung function and BMD, include a larger number of participants and the rigorous centralized quality control of the BMD scans. These factors may account for the discrepancy among the studies.
We explored possible contributions of body composition, such as FFMI and body fat percentage. Osteoporosis can be influenced by low body weight, fat mass or fat-free mass, and FFMI, depending on age, sex, and menopausal status. Additionally, body composition changes can cause complications, including impaired respiratory muscle function and loss of skeletal muscle mass with reduced force generation. Many researchers have suggested that FFMI is associated with disease severity, exercise capacity, quality of life, and mortality in patients with COPD [
16-
18]. The linkage between loss of FFMI and BMD may be a causative mechanism that can explain osteoporosis in COPD patients. Previous analyses of data from KNHANES IV to V showed that both BMI and FFMI values decreased significantly in subjects with severe airway obstruction, whereas those subjects with mild to moderate airway obstruction had values equivalent to those of the control group.
Several limitations of this study deserve consideration. We used 25(OH)D levels for analysis, which were found to be associated with lung function from the Third National Health and Nutrition Examination Survey in the USA [
19]. A previous analysis of the KNHANES III data showed a relationship between serum concentrations of 25(OH)D and lung function. Contrary to previous studies, our analysis from the KNHANES III and IV data did not show such an association. A possible explanation is that these discordances may be derived from differences in age of subjects, season, region, and exposure to sunlight, which are important determinants of BMD. We did not investigate each subject’s amount of sunlight exposure. We assumed that those who work outdoors would have more sunlight exposure, but our data revealed conflicting results between occupation and BMD in each region. Thus, we could not estimate the amount of sunlight exposure. These differences may influence the relationship between BMD and lung function.
The measurements of BMD using DXA may be confounded by several factors [
20]. In particular, measurements of lumbar spine BMD using DXA becomes less useful with increasing age due to confounding effects, such as vertebral collapse, osteophytes, and aortic calcification, all leading to a “pseudo-normalizing” of the BMD. Different calculating patterns have been used for each DXA scan, and thus, these discrepancies should each be considered [
21].
The data did not contain information about current medications, such as steroid use, dosage, and duration. Although the effects of inhaled corticosteroids on osteoporosis and fractures remain controversial, systemic corticosteroids could affect BMD by decreasing new bone formation while increasing bone loss. Indeed, the rate of bone loss correlates loosely with the daily dose and duration of therapy [
22]. In addition, pharmacological treatments of osteoporosis such as bisphosphonates can influence BMD. Smith et al. [
23] suggested that alendronate improves significantly T- and Z-scores for lumbar spine BMD, but not for hip BMD, in patients with asthma or COPD. Further studies are needed to understand the effects on BMD of corticosteroids and medications for osteoporosis.
In conclusion, BMD appeared lower in subjects with airflow limitation on pulmonary function tests in the general Korean population aged 50 years and older. Osteopenia and osteoporosis are more common in adult with airflow limitation. However, the results of this study suggest that lower BMD levels are not associated with deficits in lung function after adjusting for confounding factors. To determine the association between impaired lung function and osteoporosis, more prospectively controlled studies are needed.