Patients with chronic obstructive pulmonary disease (COPD) are at increased risk of osteoporosis. However, the prevalence, correlates and effectiveness of treatment of osteoporosis in COPD patients remain unclear.
We performed a systematic review of the literature to answer three questions. 1) What is the prevalence of osteoporosis in COPD? 2) What are identified correlates of osteoporosis in COPD? 3) What are the effects of treatment of osteoporosis in COPD? A computerised literature search in MEDLINE/PubMed and the Cochrane database was carried out. In addition, reference lists were searched by hand and authors were contacted if necessary.
The prevalence of osteoporosis and osteopenia varied 9–69% and 27–67%, respectively. Prevalence of osteoporosis was generally higher than in healthy subjects and some other chronic lung diseases. Correlates of osteoporosis in COPD are mainly measures of body composition, disease severity and the use of corticosteroids, although causality has not been proven. Effects of treatment of osteoporosis have not been investigated in samples consisting of COPD patients only.
Longitudinal follow-up to assess determinants of osteoporosis in COPD and randomised placebo-controlled trials on the effects of treatment of osteoporosis in patients with COPD only are warranted.
Chronic obstructive pulmonary disease (COPD) is characterised by a progressive airflow limitation that is not fully reversible and is associated with an abnormal inflammatory response of the lung to noxious particles and gases 1. The degree of airflow limitation can be assessed by spirometry and stratified in accordance with the Global Initiative for Chronic Obstructive Pulmonary Disease (GOLD) 1. Although primarily a pulmonary disease, there are significant extrapulmonary effects in COPD 2–5. Indeed, the GOLD guidelines incorporated these extrapulmonary effects in their definition of COPD 1. Examples of extrapulmonary effects are increased arterial stiffness 3, skeletal muscle atrophy 4, systemic hypertension 6 and osteoporosis 7.
Osteoporosis is a systemic skeletal disease characterised by a low bone mass and/or microarchitectural deterioration of bone tissue leading to increased bone fragility and increased fracture risk 8. Known risk factors for osteoporosis in the general population are, among others, female sex, advanced age, low body weight, chronic glucocorticoid therapy and endocrinological disorders such as hyperthyroidism and primary hyperparathyroidism 8–11. In COPD, the prevalence of osteoporosis is assumed to be two- to five-fold higher than in age-matched subjects without airflow obstruction 3, 12. Indeed, in a recently developed screening tool for males at risk for osteoporosis, the presence of COPD is one of the parameters increasing this risk almost four times 13.
The burden of osteoporosis varies with the incidence of fracture risk 8. Fractures of the hip, vertebrae and forearm are the most common fractures in patients with osteoporosis, although fractures of other body parts can also be the result of osteoporosis 8. The treatment of osteoporosis aims at fracture prevention and, according to the World Health Organization (WHO), should consist of lifestyle modification (such as smoking cessation, weight-bearing physical exercise and adequate calcium intake) and drug treatment 8. The latter should consist of bisphosphonates, calcium supplementation (in the case of low dietary intake) and vitamin D supplementation (in the case of vitamin D deficiency). Especially in COPD patients it is important to prevent vertebral fractures since they might result in a decreased forced vital capacity 14. In addition, (osteoporotic) hip fractures in COPD patients pose a greater problem than hip fractures in otherwise healthy subjects because of the increased operative risk in COPD patients 15–17.
The aim of this article was to perform a systematic review of the literature in order to answer the following questions. 1) What is the prevalence of osteoporosis in COPD? 2) What are identified correlates of osteoporosis in COPD? 3) What are the effects of treatment of osteoporosis in COPD?
A computerised literature search of the MEDLINE/PubMed and the Cochrane databases was performed. The time span was January 1988 to April 2008. For each question two groups of keywords were used to search for relevant articles.
To determine the prevalence of osteoporosis in COPD the keywords were: 1) COPD, chronic obstructive pulmonary disease, emphysema, chronic bronchitis; 2) osteoporosis, osteopenia, dual-energy X-ray absorptiometry (DXA) scan, bone mineral density, prevalence of osteoporosis, prevalence of osteopenia.
To identify correlates of osteoporosis in COPD the keywords were: 1) COPD, chronic obstructive pulmonary disease, emphysema, chronic bronchitis; 2) osteoporosis, osteopenia, risk factors for osteoporosis, determining factors for osteoporosis, screening for osteoporosis.
Finally, to determine the effects of treatment of osteoporosis in COPD the keywords were: 1) COPD, chronic obstructive pulmonary disease, emphysema, chronic bronchitis; 2) treatment of osteoporosis, lifestyle interventions in osteoporosis, bisphosphonates, calcium supplementation, vitamin D.
For each of the three searches, keywords from group 1 were combined with the keywords from group 2 by “AND”. Within each group, the keywords were combined using “OR”. In addition, reference lists of the identified articles were searched by hand in order to find relevant articles that were missed in the initial search strategy.
Non-English articles were excluded. In addition, we excluded review articles, although we did search their reference lists by hand for relevant articles. Finally, editorials, congress abstracts and case reports were not included.
Prevalence of osteoporosis in COPD
Only studies that enrolled COPD patients or where COPD patients could be isolated from the studied sample (in the case of inclusion of patients with other diseases besides COPD) were included. One of the outcomes had to be the prevalence of osteoporosis based on bone mineral density (BMD) measurements, thereby excluding studies with a prevalence of osteoporosis based on (osteoporotic) fractures. Moreover, a definition of osteoporosis had to be given by the authors. Cross-sectional, longitudinal and intervention trials were included. When the prevalence of osteoporosis was given per site only (e.g. hip or lumbar spine), we contacted the authors by e-mail to ask for the prevalence of the combined sites (i.e. if one of the sites was diagnosed as osteoporosis then the patient was defined as having osteoporosis). We determined the overall mean prevalence of osteoporosis in 13 identified studies (see online supplementary material).
Correlates of osteoporosis in COPD
Only studies investigating correlates of osteoporosis in COPD patients, or where COPD patients could be isolated from studied samples (in the case of inclusion of patients with other diseases besides COPD) were included. Correlation and/or regression analysis had to be carried out by searching for clinical outcomes associated with osteoporosis or BMD (thereby excluding studies with fractures as primary end-points). Cross-sectional, longitudinal and intervention trials were included.
Treatment of osteoporosis in COPD
Again, only studies where the results for COPD patients could be isolated from other groups of patients were included. In addition, only randomised placebo-controlled trials studying the effects of lifestyle interventions (e.g. more weight-bearing exercise, smoking cessation) and/or “osteoporosis medication” (e.g. bisphosphonates, calcium supplementation, vitamin D) were included.
To assess the methodological quality of identified trials, the Delphi list was used 18. The Delphi list is a comprehensive criteria list for quality assessment of randomised controlled trials (RCTs) for conducting systematic reviews. It consists of nine items all having a “yes”, “no” or “don't know” answer. If bias was unlikely, the criterion was rated positive (“yes”). In cases where information was lacking or insufficient and/or if bias was likely, the criterion was rated negative (“no” or “don't know”, respectively). All “yes” scores (1 point per “yes”) were summed to produce an overall quality score.
Prevalence of osteoporosis in COPD
The search resulted in 240 articles. Of these 42 articles were not in the English language, 48 were review articles and six were editorials or letters. Another 94 articles were excluded because they were about other topics (clear from the abstracts only). After reading the remaining articles, another 36 were excluded: 12 because COPD patients could not be isolated from other groups under investigation and 24 because they were about other topics. One article was excluded because a subgroup of patients used in a previous paper was investigated (see online supplement). Finally, 13 studies with a total of 775 COPD patients were included. In total, there were more male patients (67% versus 33%, n = 759). Moreover, patients had a mean±sd age (if reported) of 63.4±5.2 yrs (n = 775), a forced expiratory volume in 1 s (FEV1) of 46.7±13.5% predicted (n = 514), a body mass index (BMI) of 24.9±2.3 kg·m−2 (n = 721) and a fat-free mass index (FFMI) of 16.7±0.9 kg·m−2 (n = 311) (table 1⇓).
The prevalence of osteoporosis varies from 9% to 69% depending on the patients under study, the method used to assess BMD and the definition used to define osteoporosis (table 1⇑). The overall mean prevalence of osteoporosis for the 13 identified studies was 35.1% (272 of 775 patients). Patients with osteoporosis consisted of a higher proportion of women. In addition, FEV1 (% pred), BMI and FFMI were lower in the osteoporotic COPD patients (table 2⇓).
Univariate binary logistic regression analysis showed that females had an odds ratio (OR) of 1.968 for osteoporosis (p<0.001). A lower FEV1, BMI and FFMI increased the OR of osteoporosis, while age had no significant influence on osteoporosis in COPD (see online supplementary material, table E1).
Eight studies determined the prevalence of osteopenia, which varies 27–67%, resulting in an overall mean prevalence of osteopenia of 38.4% in COPD (table 1⇑).
Four studies included an age-matched control group of healthy subjects 3, 12, 23, 27. The prevalence of osteoporosis in COPD was increased compared with the healthy subjects in three trials (overall mean prevalence of osteoporosis of 31.7% in COPD versus 5.8% in healthy subjects, p<0.001) 3, 12, 27. However, Karadag et al. 23 did not find a significant difference in prevalence of osteoporosis between COPD patients and healthy subjects (fig. 1⇓). Nevertheless, the significant difference in the prevalence of osteoporosis between COPD patients and healthy subjects still remained after analyses of the four trials together: 32.5 versus 11.4%, respectively, p<0.001.
Four studies included a control group of patients with other chronic lung diseases 19, 25, 26, 29. The prevalence of osteoporosis was higher in patients with COPD than in patients with: asthma (50% versus 21%, respectively) 25; idiopathic pulmonary fibrosis (69% versus 43%, respectively) 26; pulmonary hypertension (69% versus 55%, respectively) 26; a mixed group of idiopathic pulmonary fibrosis, pulmonary hypertension, sarcoidosis, scleroderma or Kartagener's syndrome (45% versus 15%, respectively) 29; and a mixed group of α-1 antitrypsin deficiency, sarcoidosis, lymphangioleiomyomatosis, fibrosing alveolitis and bronchiectasis (59% versus 32%, respectively) (fig. 2⇓) 19. In addition, the prevalence of osteoporosis was lower in patients with COPD than in patients with cystic fibrosis in two studies: 69% versus 76%, respectively 26; and 45% versus 75%, respectively (fig. 2⇓) 29.
Correlates of osteoporosis in COPD
In total, 207 articles were found. Of these, 43 were not in the English language, 45 were review articles and three were editorials, comments or letters. In addition, eight articles were excluded because COPD could not be isolated from other patient groups, 92 because they were not about the topic and another four based on their statistical methods (no correlation and/or regression analysis performed). Twelve studies were included for this review.
Correlates of osteoporosis and/or (a low) BMD in COPD identified in mainly cross-sectional studies were body composition measures 12, 20, 22, 24, 25, 28, 30, 31, measures of disease severity 20, 27, 31, 34 and corticosteroids 24, 31 (table 3⇓). In addition, two longitudinal studies investigating change in BMD were identified (table 3⇓). Scanlon et al. 33 found inhaled corticosteroids to be a risk factor for decreasing BMD at both femoral neck and lumbar spine. In addition, they found an age of >56 yrs to be a predictor for decreasing BMD at the femoral neck, while >65 yrs of age and female sex to be predictors for decreasing BMD at the lumbar spine (table 3⇓). Mineo et al. 22 investigated COPD patients before and after lung volume reduction surgery and found significant correlations between the increase in lumbar BMD and the changes in the following parameters: residual volume (RV), diffusing capacity of the lung for carbon monoxide (DL,CO), BMI, fat-free mass (FFM), bone alkaline phosphatase (bone-AF), β-crosslaps, methylprednisolone (table 3⇓).
Treatment of osteoporosis in COPD
The search resulted in 136 articles, 27 were excluded because they were not in the English language, 34 because they were reviews, three were editorials, letters or comments and 69 because they were about another topic. No studies were identified investigating the treatment of osteoporosis in COPD patients alone.
The prevalence of osteoporosis in COPD patients varies 9–69%. In addition, the prevalence of osteopenia varies 27–67%. The prevalence of osteoporosis was higher in COPD patients than in healthy subjects. Identified correlates of osteoporosis or a low BMD in COPD patients are body composition measures, disease severity and treatment with corticosteroids. Treatment of osteoporosis has not been investigated in random samples consisting of only patients with COPD.
Prevalence of osteoporosis in COPD
Vrieze et al. 20 found a low prevalence of osteoporosis compared with the other studies (table 1⇑). This may, at least in part, be because none of the GOLD stage II patients had osteoporosis. In contrast, Bolton et al. 12 found a prevalence of osteoporosis in 20% of GOLD stage II patients. A possible explanation for these conflicting results could be the use of quantitative ultrasound by Vrieze et al. 20 instead of DXA scan, which is the gold standard to assess osteoporosis 8. The highest prevalence (69%) of osteoporosis in COPD was reported by Tschopp et al. 26. In two other studies conducted in COPD patients considered for lung transplantation, the prevalence of osteoporosis was lower (48% and 59%) 19, 29. This difference in prevalence may be due to differences in patient characteristics. However, in the absence of clinical characteristics for COPD patients only, this was hard to check 26. Three other studies 22, 25, 28 found a relatively high prevalence of osteoporosis in COPD (50%, 49% and 60%, respectively). However, the mean age of these patients was 72, 70 and 71 yrs, respectively, whereas in other studies the mean age was ≤67 yrs (table 1⇑). In addition, Katsura and Kida 25 defined osteoporosis according to the Japanese guidelines whereas most other studies used the WHO criteria to define osteoporosis (table 1⇑).
We found differences in sex distribution, FEV1, BMI and FFMI after stratification for presence/absence of osteoporosis. In addition, in a logistic regression analysis we identified male sex, decreasing FEV1, BMI and FFMI as possible determinants of osteoporosis in COPD. Nevertheless, heterogeneity of the study designs makes it somewhat difficult to draw strong conclusions from these analyses.
On average, the prevalence of osteoporosis and/or a low BMD was significantly higher in COPD patients than in healthy subjects (fig. 1⇑). Only one study did not find a significant difference in prevalence of osteoporosis in Turkish COPD outpatients (40% for lumbar spine and hip combined, 35% and 10% for lumbar spine and hip respectively) compared with their healthy peers (40% at lumbar spine and 15% at the hip) 23. Nevertheless, the prevalence of osteoporosis in the healthy group appears to be higher than the prevalence of osteoporosis in healthy subjects in other studies (0–25%) 3, 12, 27, 35. Indeed, the Turkish population has BMD values ∼1 sd lower than the Swedish population 36. The decreased BMD in the Turkish population may partly explain the lack of difference in prevalence of osteoporosis between COPD patients and healthy subjects 23.
Correlates of osteoporosis in COPD
To determine risk factors for osteoporosis, longitudinal (intervention) studies are preferred. To date, only two longitudinal studies that met the inclusion criteria have been identified 22, 33. The other studies are cross-sectional (table 2⇑). For this reason, interpretation of the results should be with caution as causality of the correlates (table 2⇑) needs to be confirmed. Several studies found body composition measures (low BMI, low FFMI and % of ideal body weight) to have a significant correlation with osteoporosis and/or BMD 12, 20, 25, 28, 30. In the general population, low body weight and/or low BMI have also been identified as risk factors for osteoporosis and incorporated in guidelines 8, 13. The link between low body composition and osteoporosis or low BMD in COPD could be increased inflammation, decreased physical activity and/or other mechanisms leading to proteolysis 37–42. Another explanation for more osteoporosis in patients with lower body composition measurements could be that bone formation is decreased because there is relatively low mechanical loading on these bones. Indeed, astronauts lose as much bone mass in a 1-month spaceflight as postmenopausal females in 1 yr 43. In addition, COPD patients have been shown to be physically inactive compared with age-matched healthy subjects 44.
Mineo et al. 22 have found a significant correlation between an increase in BMD and an increase in BMI and FFM after lung volume reduction surgery. More longitudinal studies are needed to investigate the influence of change in BMI and/or FFMI on change in BMD in COPD.
A higher GOLD stage and/or a lower FEV1 have been shown to be correlated with osteoporosis and/or a low BMD 20, 31, 34. Also, in subjects without COPD, significant correlations between FEV1 and BMD have been found 45–47. These relationships between lung function parameters and BMD are complex and not yet clear. Again, in COPD patients, systemic inflammation can be a key factor, as reduced lung function has been found to be associated with increased inflammatory markers, which is a risk factor for osteoporosis 48. A strong relationship between serum 25-hydroxyvitamin D and pulmonary function parameters was found in patients from the third National Health and Nutrition Examination Survey 49. This could be another link between lung function and BMD. It is also possible that there is no causal relationship between lung function and BMD. Perhaps reduced physical activity because of impaired lung function is the reason for reduced BMD 50. More longitudinal studies are needed to investigate the possible causality between lung function and BMD.
Only Scanlon et al. 33 have found age >56 yrs and female sex to be independent correlates of osteoporosis in COPD. Indeed, in the general population higher, age and female sex are two of the most important risk factors for osteoporosis 8. In COPD patients these risk factors may disappear after correction for other risk factors that are more important in COPD and less important in the general population. Other explanations could be limited sample size, cross-sectional design of most of the trials, inclusion of only males or only females and/or a selected age category. More large longitudinal trials are needed to investigate the influence of a higher age and/or female sex on BMD in COPD patients.
Treatment of osteoporosis in COPD
The pharmacological treatment of osteoporosis should consist of bisphosphonates in combination with calcium supplementation (in case of low dietary calcium intake) and with vitamin D supplements (in case of vitamin D deficiency) 8. The protective effect of bisphosphonates has been found in multiple studies 51–54. However, no studies were found investigating the drug treatment to prevent fractures in patients with osteoporotic COPD only. Two studies 55, 56 investigated alendronate in patients with asthma or COPD. Both studies treated the patients for 1 yr. Smith et al. 56 found a significant improvement in T- and Z-scores for lumbar spine BMD, but not for BMD at the hip in the alendronate group. Lau et al. 55 found an increase in BMD in the alendronate group and a decrease in the placebo group, and these changes in BMD were significantly different between the two groups. In these studies, different eligibility criteria were used (males and females versus females only, patients with high fracture risk versus patients on inhaled corticosteroids). In addition, patient characteristics were different (67 versus 49 yrs of age, percentage asthma and COPD unknown versus majority asthma patients). The results of the two studies could not be pooled. These two studies may indicate that alendronate improves BMD in COPD patients.
Mirzaei et al. 57 investigated the effect of rocaltrol in patients with a T-score <-1 and compared this with control patients. However, patients and treating physicians were not blinded for treatment, because the patients were assigned to the control group if they had contra-indications for rocaltrol or if they were unwilling to use rocaltrol. Again, nothing can be concluded about the effect of rocaltrol in COPD patients only. Considering the methodological quality of the three described trials, only Smith et al. 56 scored high on the Delphi scoring list (8 points) whereas the other two studies 55, 57 scored very low (4 and 2 points, respectively;see online supplementary material, table E2). RCTs in COPD patients only and with a long-term follow-up are needed.
No studies were found that investigated the effect of lifestyle changes on BMD in COPD patients. However, in an RCT, lung transplantation patients who performed 6 months of exercise on a lumbar extensor machine significantly gained lumbar BMD, in contrast to the control patients (also after lung transplantation without exercise) who lost lumbar BMD 58. Another RCT compared the effect of alendronate plus mechanical loading to alendronate alone and to control patients (without alendronate and without mechanical loading) in lung transplant recipients 59. Again, in control patients a significant decrease in BMD compared with baseline BMD was found after 8 months (−14.1%). In patients treated with alendronate the BMD increased (1.4%) and in the alendronate plus mechanical loading group the BMD increased even more (10.8%). In contrast, in a 4-yr RCT in middle-aged males regular aerobic exercise training had no significant effect on femoral BMD 60. Intervention studies in COPD patients are needed to investigate the possible short- and long-term effects of progressive resistance training on BMD 61.
Prevalence of osteoporosis and osteopenia seems to be high in COPD patients. Correlates of osteoporosis in COPD are body composition measurements, measures of disease severity and corticosteroids. Although causality has not been proven, based on the current results it seems reasonable to advise (chest) physicians to screen for osteoporosis in COPD patients with a low BMI (<21 kg·m−2) and/or a low FFMI (<16 kg·m−2 in males and <15 kg·m−2 in females). The effects of treatment of osteoporosis have not been investigated in patients with COPD only.
Further areas where research is needed
To determine risk factors of osteoporosis in COPD, more prospective studies are needed. Moreover, randomised (placebo-) controlled trials on the effects of pulmonary rehabilitation (including progressive resistance training and nutritional modulation 62) and/or relevant drug treatment on BMD and fracture risk reduction are warranted in osteoporotic patients with COPD.
Statement of interest
Statements of interest for E.F.M. Wouters and F.W.J.M. Smeenk can be found at www.erj.ersjournals.com/misc/statements.dtl
We thank E.J. Martens (PhD, statistical advisor for the Dept of Education and Research, Catharina Hospital Eindhoven, Eindhoven, the Netherlands) for her statistical advice. We thank the following authors for providing additional information: L. Førli (Dept of Medicine, Rikshospitalet, University of Oslo, Oslo, Norway), C.E. Bolton (Dept of Respiratory Medicine, School of Medicine, Cardiff University, Cardiff, UK), F. Karadag (Dept of Chest Diseases, Adnan Menderes University School of Medicine, Aydin, Turkey), E.F. Dubois (Dept of Pulmonary Diseases, Reinier de Graaf Groep Delft and Voorburg, the Netherlands), R.M. Aris (Dept of Medicine, Surgery and Transplant, The University of North Carolina, at Chapel Hill, NC, USA), N.R. Jørgensen (Dept of Clinical Biochemistry, Copenhagen University Hospital Hvidovre, Hvidovre, Denmark) and D.D. Bikle (Endocrine Unit, VA Medical Center, San Francisco, CA, USA).
This article has supplementary material available from www.erj.ersjournals.com
- Received August 24, 2008.
- Accepted December 22, 2008.
- © ERS Journals Ltd