Abstract
The use of inhaled corticosteroids has been associated with a dose-related increased risk of fracture. This may be related to systemic absorption. However, several studies have found that patients with more severe reductions in pulmonary function had reduced bone mineral density, independent of inhaled corticosteroids. The objective of this study was to evaluate the relationship between disease severity and fracture risk.
A large case–control study (108,754 cases) was conducted using data from the UK General Practice Research Database. It was found that higher doses of inhaled corticosteroids were associated with greater risks of fracture. The crude odds ratio of fracture among patients exposed to >1,600 μg beclomethasone equivalents per day was 1.95 (95% confidence interval (CI) 1.68–2.27). When adjustments were made for disease severity and use of bronchodilators, the initial dose–response relationship between inhaled corticosteroids and fracture risk disappeared (adjusted odds ratio of 1.19 (95% CI 1.01–1.41)).
In conclusion, patients with severe obstructive airway disease are at risk of fracture. However, adequate adjustment for disease severity is essential when the association between the use of inhaled corticosteroids and risk of osteoporotic fracture is studied in observational research.
Inhaled corticosteroids are frequently prescribed to patients with obstructive airway disease (OAD) 1. Although administered locally, inhaled corticosteroids undergo systemic absorption and may induce a suppression of plasma cortisol, especially at higher doses 2. Patients using oral corticosteroids have an increased risk of osteoporosis and fractures 3. For users of inhaled corticosteroids, an association between daily dose and increased risk of fractures has been reported 4–7. However, adjustment for underlying disease severity was limited in these studies. In the largest case–control study, adjustment for underlying disease severity was restricted to prior use of oral corticosteroids 5. The importance of controlling for severity of underlying disease in this type of research has recently been emphasised 8, 9. Also, it has been reported that there was no increased risk of fracture in patients using inhaled corticosteroids after adjustment for the presence of OAD 10 or use of bronchodilators 11. Patients with OAD have been found to have a reduced bone mineral density (BMD), independent of the use of respiratory medication 12–14. Therefore, the objective of this study was to evaluate the relationship between disease severity of OAD and the risk of fracture.
METHODS
Study population
A case–control study was conducted using data obtained from the General Practice Research Database (GPRD) 15. This database comprises the entire computerised medical records of a sample of general practitioners in the UK. Several independent validation studies have confirmed a high level of completeness and validity of the GPRD with regard to the recording of fractures 16, 17. In this study, cases were patients aged ≥18 yrs with a first record during GPRD follow-up (January 1987 to July 1999) of an osteoporotic fracture (defined as a fracture of the radius/ulna, femur/hip, ribs, humerus, vertebrae or clavicle) 18. The date of the first fracture was defined as the index date. For each case, one control patient without a history of a fracture was matched to a case by age, sex, medical practice and calendar time. A history of OAD before the index date was examined in both cases and controls. The definitions for asthma and chronic obstructive pulmonary disease (COPD) were the same as those used in a previous GPRD validation study 19.
Definitions of variables
Use of medication for the treatment of OAD was examined, including bronchodilators (British National Formulary (BNF) chapter 3.1 or 3.3) and inhaled corticosteroids (BNF chapter 3.2) 20. Current users were patients who had received at least one prescription in the 6 months before the index date. Past users had received their last prescription >6 months prior to the index date. For each current user, the average daily dose of bronchodilators or inhaled corticosteroids was estimated by dividing the total amount of respiratory medication by the treatment time (the time between the first and last prescription of respiratory medication). Dose equivalences were expressed as beclomethasone diproprionate equivalents (inhaled corticosteroids) or salbutamol equivalents (bronchodilators). Equivalents of different compounds were calculated with defined daily doses 21.
Indicators of more severe OAD included records of OAD exacerbations and oxygen use in the 12 months before the index date, respiratory problems (bacterial respiratory tract infections, coughing, presence of sputum, haemoptysis, dyspnoea, tachypnoea, shortness of breath, acute bronchitis or wheezing), use of oral corticosteroids in the 6 months before the index date and a body mass index (BMI) <20. Next, the current authors controlled for various general risk factors of fracture, including history of diabetes mellitus, rheumatoid arthritis, hyperthyroidism, congestive heart failure, seizures, anaemia, dementia, depression, psychotic disorder and cerebrovascular accident. Moreover, prescriptions in the 6 months before the index date for anticonvulsants, nonsteroidal anti-inflammatory drugs, methotrexate, hormone-replacement therapy, thiazide diuretics, anxiolytics/hypnotics, antipsychotics, antidepressants and anti-Parkinson's drugs were also considered as potential confounding variables. BMI and smoking status (current or ex-smoker, nonsmoker or unknown) were included if entered in the database.
Statistical analyses
Odds ratios (ORs) of fracture were estimated using conditional logistic regression. Final regression models were determined by backward elimination using a significance level of 0.05. The ORs of these models were compared with the ORs of the models including all variables to identify confounding by an eliminated variable, with inclusion of this variable in cases of confounding. Current users of inhaled corticosteroids or bronchodilators with a history of OAD before the index date were compared with patients who had never been exposed to inhaled corticosteroids or bronchodilators and who did not have a diagnosis of OAD.
RESULTS
The study population included 108,754 adult patients who sustained an osteoporotic fracture (44,201 radius/ulna, 22,250 femur/hip, 16,189 rib, 14,646 humerus, 8,712 vertebral and 3,908 clavicle fractures). Baseline characteristics are listed in table 1⇓. The median time of enrolment before the index date was 2.2 yrs. In the overall study population, higher BMI (>25) was associated with a reduced risk of a hip fracture (crude OR 0.61; 95% confidence interval (CI) 0.56–0.66), compared with patients with a BMI ranging 20–25. For patients with low BMI (<20), the reverse was observed (crude OR 1.78; 95% CI 1.58–2.00). Oral corticosteroid use was also associated with an increased risk of hip fracture (crude OR 1.75; 95% CI 1.58–1.95). Of the cases with current oral corticosteroid use, 54% received five or more oral corticosteroid prescriptions in the previous year (for controls, this was 43%). The percentage of patients with continuous oral corticosteroid use for >6 months was 44% among the cases and 36% among the controls (continuous use was defined as receiving a repeat oral corticosteroid prescription within 3 months of the previous prescription).
The risk of osteoporotic fracture was increased in patients with asthma (crude OR 1.28; 95% CI 1.23–1.32), in patients with COPD (crude OR 1.61; 95% CI 1.52–1.71), and in patients who had both diagnoses recorded (crude OR 1.72; 95% CI 1.60–1.85). Patients with an indicator of more severe OAD generally had increased risks of fracture (fig. 1⇓) compared with patients without the indicator who did not have a history of OAD. For example, the risk of osteoporotic fracture among patients with OAD exacerbations was 2.02 (95% CI 1.82–2.23) and 2.54 (95% CI 1.90–3.40) among patients with a BMI <20. Patients with OAD who also smoked had an increased risk of osteoporotic fracture (crude OR 1.57; 95% CI 1.49–1.67). Those who smoked <20 cigarettes·day−1 had a crude OR of 1.57 (95% CI 1.41–1.76) compared with an OR of 1.79 (95% CI 1.24–2.57) in patients with OAD who smoked >20 cigarettes·day−1. Among patients with a history of OAD, the risk of osteoporotic fracture was increased in both current (crude OR 1.53; 95% CI 1.46–1.59) and past users (crude OR 1.36; 95% CI 1.27–1.47) of inhaled corticosteroids, compared with those who had never used inhaled corticosteroids and were not diagnosed with OAD. Among the current users, the risk of fracture increased with daily dose (table 2⇓). Patients exposed to bronchodilators also had an increased risk of fracture. When the current authors adjusted for general risk factors, bronchodilator use and indicators of disease severity, current use of inhaled corticosteroids was no longer associated with an increased risk of osteoporotic fracture (adjusted OR 1.04; 95% CI 0.97–1.11). Furthermore, the dose response with inhaled corticosteroids observed in the crude analysis was no longer statistically significant. Current users of beclomethasone, budesonide or fluticasone had a comparable risk of fracture.
Most current users of inhaled corticosteroids were also concomitantly exposed to bronchodilators (76% for cases and 72% for controls). No statistically significant differences could be found between the use of both drugs at the same time and single use. In the high-dose group (>1,600 μg beclomethasone equivalents per day), approximately half had not been exposed to oral corticosteroids in the previous 6 months. These patients did not have an increased risk of osteoporotic fracture (adjusted OR 1.19; 95% CI 0.96–1.47). The risk of fracture was increased in patients who were exposed to both oral corticosteroids and high-dose inhaled corticosteroids. For osteoporotic fracture, the adjusted OR was 1.79 (95% CI 1.40–2.29), for hip fracture 1.77 (95% CI 0.91–3.46), and 3.78 (95% CI 1.79–7.97) for vertebral fracture. The median number of prior oral corticosteroid prescriptions was 14 in these patients.
As shown in table 3⇓, patients with more severe OAD had higher risks of fracture, and these risks were comparable between users and nonusers of inhaled corticosteroids. The adjusted OR for osteoporotic fracture was 1.47 (95% CI 1.25–1.74) in nonusers and 1.48 (95% CI 1.29–1.71) in users of inhaled corticosteroids with more severe OAD.
DISCUSSION
In the current study, the association between exposure to high doses of inhaled corticosteroids and the risk of osteoporotic fracture was confirmed. Nevertheless, it was also found that the risk of fracture was related to indicators of severe OAD, and that patients using higher doses of bronchodilators also had increased risks of fracture. After adjustment for disease severity and use of bronchodilators, the initial association between use of inhaled corticosteroids and risk of osteoporotic fracture almost disappeared.
Two other studies, one using a case–control design and the other a cohort design, evaluated the risk of fracture of patients using inhaled corticosteroids in GPRD 4, 5. Both these studies found a dose-dependent increase in fracture risk, but they differed in their interpretation on the aetiology of this increased risk. The current authors were able to confirm the results of the case–control study 5 when the same definitions were applied (i.e. timing of prior use of inhaled corticosteroids was ignored by combining current and past users, and no adjustment was made for general risk factors, prior use of bronchodilators or disease-severity indicators, with the exception of oral corticosteroid use). However, when adjusting for disease-severity indicators, the hip-fracture risk in patients using inhaled corticosteroids was statistically comparable with nonusers, even at the higher doses. The present findings suggest an important role of the underlying respiratory disease in the aetiology of increased fracture risk.
There have been several studies that have evaluated the risk of fracture in users of inhaled corticosteroids. A case–control study conducted in Denmark found a nonsignificant trend between use of high-dose inhaled and intranasal corticosteroids and risk of hip fracture, but the authors did not adjust for comedication and severity indicators 7. In a nested case–control study among elderly American veterans with COPD, a dose–response relationship between short-term use of inhaled corticosteroids and nonvertebral fracture risk was found. Adjustment for disease severity was limited to use of oral corticosteroids, the number of hospitalisations and the number of outpatient visits 6. The importance of adjustment for the underlying disease severity was confirmed in a nested case–control study conducted in a Canadian population. Hip-fracture risk was not associated with exposure to high-dose inhaled corticosteroids over a 4-yr period. The analysis was adjusted for the number of dispensings of bronchodilators 11. In a cohort study consisting of Canadian females, exposure to inhaled corticosteroids was not associated with an increased risk of hip fracture. Results were adjusted for a wide range of underlying diseases, including a history of COPD 10.
The relationship between the use of inhaled corticosteroids and BMD was analysed in two meta-analyses, with inconsistent results. A Cochrane review of randomised clinical trials found no evidence for an effect of inhaled-corticosteroid exposure on BMD; however, the patient population included in the randomised trials had relatively mild respiratory disease and was young 22. A meta-analysis, which also included observational studies, found that inhaled-corticosteroid users had, on average, a lower BMD than expected for their age and sex. To address the effect of confounding by underlying disease severity, Richy et al. 23 classified studies on the basis of the type of controls to the inhaled-corticosteroid users: healthy population controls or controls with lung disease. It was found that inhaled-corticosteroid users had lower BMD in both sets of study populations and that these results were statistically similar. Nevertheless, the number of patients in the studies with healthy controls was quite small. Moreover, the largest differences between inhaled-corticosteroid users and controls were seen in the studies with healthy controls, although this was not statistically significant. In conclusion, this trend is consistent with the current hypothesis that underlying disease severity is important in the aetiology of reduced BMD in patients using inhaled corticosteroids 23.
The severity of pulmonary disease inversely correlated with BMD in three studies. An analysis of the Third National Health and Nutrition Examination Survey (NHANES) revealed that the risk of osteoporosis among males and females was inversely correlated with the degree of their airway obstruction. Adjustment for age, smoking, BMI, physical activity and different types of medication (among others, inhaled or oral corticosteroids, bronchodilators and oestrogens) did not change these results 13. In a group of elderly Japanese females with COPD who were not exposed to oral corticosteroids, the prevalence of osteoporosis was 50%, twice as high as a comparison group consisting of females of the same age with asthma 24. A cross-sectional study in a general-practice setting among British females aged 45–76 yrs also showed an inverse relationship between forced expiratory volume in one second and BMD at the hip, which remained after adjustment for body weight and length. Although use of inhaled corticosteroids was not measured, the results remained similar after exclusion of patients who had a history of respiratory disease 12.
Several mechanisms have been suggested for this possible relationship between OAD and increased risk of fracture or osteoporosis. They include lack of physical activity 25, 26, low BMI among patients with COPD 27, smoking 28, a decreased exposure to sunlight 13, decreased testosterone levels 29, hypercapnia 30, and chronic inflammation. Cytokines that are expressed in inflammatory diseases, such as asthma and/or COPD, include tumour necrosis factor (TNF)-α, transforming growth factor-β, interleukin (IL)-1β, IL-4 and IL-8. These cytokines have been shown to affect bone remodelling in vivo and in vitro 31–37. However, it is uncertain whether these cytokine levels are also increased in the osteoblasts and osteoclasts at the basic multicellular unit 38 in humans with respiratory disease. In a study among 68 post-menopausal females with osteoporosis, a correlation between levels of plasma IL-8 or TNF-α and BMD was not found 39. Many patients with OAD use β2-agonists. The adrenergic pathway may play a role in the regulation of bone formation in ovariectomised mice 40. The current authors stratified the bronchodilator group to β2-agonists, antimuscarinics, xanthines, cromones and adrenoceptor agonists (e.g. ephedrine), and were unable to find a clear relationship between any class of drugs and risk of fractures. An interesting finding in the present study was an increased risk of fracture among past users of inhaled corticosteroids. This might be related to a history of concomitant oral-steroid use or a history of OAD.
This study has several limitations. Data on physical activity and smoking were limited. It is not obligatory for participating practices to record smoking status. The current data on smoking were incomplete, and the present results on the relationship between smoking and risk of fracture were lower than those recently reported in a meta-analysis 41. There were no data on the loss of fat-free mass. Loss of fat-free mass has been associated with reduced BMD and more severe COPD, and might be a more concise marker of severity of OAD compared with BMI 42, 43. Although the current authors did not have data on BMI for all patients, the present finding of an inverse relationship between BMI and fracture risk was consistent with those previously reported 44, 45. Controlling for severity of the underlying disease could also be improved with spirometry data, which were not available in this study. Lastly, among the patients who used >1,600 μg beclomethasone equivalents per day, a large number (46%) were using oral corticosteroids (median number of oral-corticosteroid prescriptions was 14). Therefore, it may be difficult in this patient group to separate the effects of oral and inhaled corticosteroids.
In conclusion, patients using higher doses of inhaled corticosteroids have an increased risk of osteoporotic fracture. However, patients using bronchodilators and those with more severe obstructive airway disease also have an increased risk of fracture, and the dose–response relationship between inhaled corticosteroids and fracture risk almost disappears after adjustment for disease severity. Consequently, adequate adjustment for disease severity is essential when the association between use of inhaled corticosteroids and risk of fracture is studied in observational research.
- Received May 14, 2004.
- Accepted January 10, 2005.
- © ERS Journals Ltd