Abstract
Objective. To determine whether children or adolescents who are exposed to inhaled corticosteroids (ie, beclomethasone, budesonide, fluticasone) are at a higher risk of having bone fractures compared with nonexposed individuals.
Methods. We performed a population-based nested case-control analysis using data from the United Kingdom–based General Practice Research Database. Within a base population of 273 456 individuals aged 5 to 79 years, we identified by International Classification of Diseases codes children or adolescents who were aged 5 to 17 years with a fracture diagnosis and up to 6 control subjects per case matched to cases on age, gender, general practice attended, calendar time, and years of history in the GPRD. We compared use of inhaled steroids before the index date between fracture cases and control patients.
Results. We identified 3744 cases and 21 757 matched control subjects aged 5 to 17 years. Current exposure to inhaled steroids did not reveal a substantially altered fracture risk compared with nonusers, even in individuals with current longer term exposure (ie, ≥20 prescriptions; adjusted odds ratio 1.15; 95% confidence interval: 0.89–1.48). In individuals with current or previous exposure to oral steroids, the adjusted odds ratio for current long-term inhaled steroid use compared with nonusers was 1.21 (95% confidence interval: 0.99–1.49).
Conclusions. Exposure to inhaled steroids does not materially increase the fracture risk in children or adolescents compared with nonexposed individuals.
Current guidelines recommend the use of inhaled corticosteroids for children and adolescents with mild persistent or more severe forms of asthma.1,2 Inhalation targets corticosteroids directly to the site of airway inflammation with fewer adverse effects compared with oral administration.3 However, a considerable portion of an inhaled steroid dose is deposited in the oropharynx and subsequently swallowed and absorbed from the gastrointestinal tract. Some fraction will be bioavailable to the systemic circulation,3 thereby potentially increasing the risk of adverse systemic corticosteroid effects.
Although inhaled steroids have been established as the preventive treatment of choice, few studies have been conducted to assess the risks of inhaled steroid therapy. The major safety concern of long-term inhaled steroid therapy—beside potential effects on adrenal function and growth4—relates to the possible adverse effects on bone metabolism and/or bone mineral density. In adults, a causal association between oral steroids and osteoporosis and a subsequently increased fracture risk has been shown,5–7 but data on inhaled steroid therapy are limited. Although several studies in children or adolescents showed that inhaled steroid therapy does not have a significant effect on bone turnover or bone mineral density,8–10 some studies suggest a suppression of bone formation or negative effects on bone mineral density associated with long-term or moderate- to high-dose therapy.11–13 A recent cross-sectional study suggested that use of inhaled steroids in the previous year might be a risk factor for fractures in prepubertal children.14 Otherwise, information on fracture risk in children and adolescents who have asthma and are exposed to inhaled steroid therapy during bone growth is scarce. To assess the risk of fractures associated with inhaled steroid therapy in children or adolescents and to explore possible duration effects, we conducted a nested case-control analysis using data from the General Practice Research Database (GPRD).
METHODS
Study Population and Data Source
Data were derived from the United Kingdom–based GPRD, which has been described in detail elsewhere.15,16 More than 3 million people in the United Kingdom are enrolled with selected general practitioners who use office computers and have agreed to provide data for research purposes. General practitioners have been trained to record medical information, including demographic data, medical diagnoses, details of hospital stays, and deaths in a standard, anonymous form. The physicians generate prescriptions directly with the computer, and this information is automatically transcribed into the computer record. A modification of the Oxford Medical Information System classification (similar to the International Classification of Diseases, Eighth Revision [ICD-8]) is used to enter medical diagnoses, and a coded drug dictionary based on the United Kingdom's Prescription Pricing Authority dictionary is used for recording prescriptions. The recorded information on drug exposure and diagnoses has been validated and proved to be of high quality.17,18
The GPRD has been the source for numerous epidemiologic studies in recent years, including studies on fractures in relation with drug therapy,19 adverse effects associated with oral or topical corticosteroid use including fractures,7,20,21 and studies in respiratory epidemiology.22 The study was approved by the Scientific and Ethical Advisory Group.
Cohort Definition
Within the GPRD, we initially identified a population consisting of 3 separate groups: group 1 included all individuals who were aged 5 to 79 years and had a diagnosis of asthma (ICD-8 code 493.x) or chronic obstructive pulmonary disease (COPD; OXMIS code 5199Gx) with at least 1 prescription for an inhaled corticosteroid (ie, beclomethasone, budesonide, fluticasone) and/or an oral corticosteroid. Group 2 encompassed all individuals who were aged 5 to 79 years and had a diagnosis of asthma or COPD and had no exposure to inhaled or oral corticosteroids in their record. Group 3 encompassed a random sample of 50 000 individuals who were aged 5 to 79 years and had neither a diagnosis of asthma or COPD nor a prescription recorded for inhaled and/or oral corticosteroids. We excluded all patients with parenteral corticosteroid therapy and/or use of nasal or rectal corticosteroids (eg, enema, suppositories). Furthermore, we excluded all patients with a computer-recorded diagnosis of osteoporosis, osteomalacia, cancer (excluding nonmelanoma skin cancer), or alcoholism. In addition, we excluded patients who were prescribed bisphosphonates (as indicator for osteoporosis or bone metastases) before the index date. From the initial base population (age range: 5–79 years) consisting of 273 456 individuals, we identified 65 779 individuals aged 5 to 17 years to form the base population for this study.
Case Definition and Nested Case-Control Analysis
Within the base population, we identified all patients who developed a first-time diagnosis of a bone fracture (by ICD-8 codes 800.x-829.x). We identified 3744 patients who were aged 5 to 17 years and selected up to 6 control subjects per case matched on age (±3 years), gender, general practice attended, calendar time (by using the same index date as for cases), and years of history in the GPRD (matching on the number of years in the GPRD ± 1 year). The same exclusion criteria were applied to control as to case patients.
In a previous study,19 we reviewed 200 randomly selected case records by hand to verify the computer-recorded diagnosis and quantify the proportion of individuals with fractures as a result of severe trauma (eg, motor-vehicle trauma). Within this sample, there was only 1 case with evidence of fracture as a result of severe trauma, a proportion that we considered negligible; hence, we decided to include all cases identified by computer. Moreover, several previous studies used GPRD data to assess the risk of fractures associated with corticosteroids (or other drugs) in which the fracture diagnosis was validated by sending a questionnaire to the general practitioners, by looking at hospital discharge summaries or radiographic reports.7 Because the results of these validations were good, we did not perform any additional validation procedures for the present study.
Statistical Analysis
We conducted analyses using conditional logistic regression to explore the association between fracture risk and type of exposure (inhaled steroids, oral steroids, or no corticosteroid therapy), exposure timing (current use, defined as having had the last prescription ≤180 days before the index date; and past use >180 days before the index date), and exposure to individual inhaled steroids (beclomethasone, budesonide, and fluticasone). We assessed the number of prescriptions for inhaled steroids and/or oral steroids as a proxy to investigate the exposure duration.
Age, gender, general practice, calendar time, and years of recorded history in the GPRD before the index date were controlled by matching. We further controlled for specific comorbidities that are potentially associated with fractures, including chronic renal failure, hyperthyroidism, hyperparathyroidism, inflammatory bowel disease, malnutrition, and malabsorption. We also controlled for other drugs that are commonly used in patients with asthma or COPD (eg, β-agonists, anticholinergic drugs, xanthines, leukotriene antagonists, mast cell stabilizers) and drugs that are potentially associated with the risk of fractures (benzodiazepines, neuroleptics, antidepressants, antihypertensives, calcium, fluoride, and vitamin D preparations).
The statistical analysis was conducted using the software program SAS, Version 8.0 (SAS Institute, Inc, Cary, NC). Relative risk estimates are presented as odds ratios (ORs) with 95% confidence intervals (CIs); P values are 2-sided and considered statistically significant at <.05.
RESULTS
Our data set included 3744 fracture cases and 21 757 matched control subjects aged 5 to 17 years. Of the 3744 cases patients, 29 (0.8%) had multiple fractures recorded (28 patients with 2 fractures, 1 patient with 3 fractures). The characteristics of case and control patients with regard to age and gender distribution and exposure to other asthma drugs are displayed in Table 1. More than half of the cases were between 10 to 14 years of age, and case patients were predominantly (65.6%) male. The exposure prevalence of drugs potentially associated with an altered fracture risk (eg, benzodiazepines, neuroleptics, antidepressants) was negligibly small in these young case and control patients; exposure to any of these drugs was not associated with a material change of the fracture risk, and we did not find any relevant alteration of the fracture risk in association with potential comorbidities (data not shown).
Characteristics of 3744 Cases and 21 757 Matched Control Subjects and Their Independent Effects on the Fracture Risk in Univariate Analyses
The location of fractures in case patients was as follows: 2021 (54.0%) had a fracture of the hand/lower arm, 357 (9.5%) had a fracture of the humerus, 313 (8.4%) had a foot fracture, 237 (6.3%) had a fracture of the clavicle, 130 (3.5%) had a fracture of the malleolus, 25 (0.7%) had a fracture of the femur, 8 (0.2%) had a vertebral fracture, and 653 (17.4%) had a fracture at another location.
The relative risk estimates (OR) for all fractures combined in users of inhaled steroids who were never prescribed oral steroids stratified by timing of exposure (current vs past use) is displayed in Table 2. Of 1010 total cases exclusively exposed to inhaled steroids, 544 (53.9%) were current and 466 (46.1%) were past users.
Exposure to Inhaled Corticosteroids Only, in Cases and Control Patients, Stratified by Exposure Timing and Association With Fracture Risk
The overall adjusted OR for current inhaled steroid use was 1.01 (95% CI: 0.90–1.13). Additional analyses stratified by number of prescriptions did not provide evidence of a substantially increased fracture risk, even in long-term users (ie, ≥20 prescriptions; adjusted OR: 1.15; 95% CI: 0.89–1.48, compared with the reference group of nonusers of steroids). The median number of prescriptions in current long-term inhaled steroid users was 26 (interquartile range: 22–33). Additional stratification by gender and age (age groups 5–9 years, 10–14 years, and 15–17 years, respectively) to assess potential effect modification did not materially change the risk estimates. We also performed an additional analysis in which we included only cases and control subjects who were age matched within 1 year; the results were closely similar. The risk estimates for various fracture sites (lower vs upper extremities) were closely similar to the unstratified risk estimates of all fractures combined (data not shown). Current inhaled steroid users were stratified further by exposure to individual inhaled steroids. Among individuals who were currently exposed to inhaled steroids only, 76.2% were exposed to beclomethasone, 21.7% to budesonide, and 2.1% to fluticasone. The risk for fracture did not differ materially among the various inhaled steroids. Among patients who were currently exposed to oral steroids in the absence of inhaled steroids (27 cases and 181 control subjects), the risk estimates tended to increase with increasing duration of use (OR: 6.07; 95% CI: 0.38–97.44, in individuals with ≥20 prescriptions compared with nonusers of steroids).
Because a considerable proportion of individuals who were exposed to inhaled steroids were also prescribed oral steroids at some point in time (mixed group; n = 516), we conducted an additional analysis in which we assessed the effect of inhaled steroids on the fracture risk, adjusted for the use of oral steroids. The risk estimate for inhaled steroid users with ≥20 prescriptions adjusted for use of other antiasthmatic drugs and oral steroid use was 1.21 (95% CI: 0.99–1.49) compared with individuals without inhaled steroid exposure. Among those with long-term exposure to oral steroids (≥20 prescriptions), the risk estimate adjusted for use of other antiasthmatic drugs and inhaled steroid use compared with individuals without oral steroid exposure was 3.65 (95% CI: 0.61–22.04).
DISCUSSION
Some observational and randomized trials in adults suggested that inhaled steroids may decrease bone mineral density or increase biochemical markers of bone resorption.23–25 So far, there is little information to demonstrate whether the effect leads to an increased fracture risk. Observational studies assessing the risk of fractures in association with inhaled steroid use in adults showed conflicting results. Although 1 retrospective cohort study did not find an altered risk of hip fractures among elderly women who were exposed to inhaled steroid compared with nonexposed,26 2 other observational studies found evidence of a slightly increased fracture risk associated with inhaled steroid exposure.21,27 Data in children or adolescents exploring the association between inhaled steroid use and the fracture risk are limited. With the exception of a recent cross-sectional study in children,14 we are unaware of any other study assessing the risk of bone fractures associated with inhaled steroid use in children or adolescents.
The present study provides evidence that exposure to inhaled steroids without concomitant or previous oral steroid exposure is not associated with an increased fracture risk in children or adolescents, not even after long-term use. However, because of small sample size in the individual strata and, as a result of that, wide CIs, we cannot exclude with certainty that especially long-term exposure with inhaled corticosteroid might be associated with a slightly increased fracture risk.
Assuming that in general 1 prescription for an inhaled steroid corresponds to 200 puffs, the median number of 26 prescriptions found in long-term users in our study would correspond to >7 years of exposure assuming 2 puffs a day or almost 2 years of exposure assuming 8 puffs a day if taken on a regular basis and not only intermittently.
In individuals who were exposed to inhaled steroids and were also prescribed oral steroids at some point in time, our results again did not suggest a materially increased risk. There was some suggestion of a slightly increased fracture risk in long-term inhaled steroid users with additional exposure to oral steroids even after adjusting the analysis for oral steroid use. In addition, the data suggested that individuals with longer term current exposure to oral steroids may have an increased fracture risk compared with nonexposed individuals. It is known from studies in adults that exposure to oral steroids may lead to osteoporosis and an increased bone fracture risk.5–7 Data are again limited with regard to the risk in children or adolescents, but a recent observational study showed an increased risk in children who were aged 4 to 17 years and had oral steroid exposure in association with different underlying diseases.28
Although we are unable to compare our results with regard to inhaled steroid exposure and fracture risk to other similar studies, the present study confirms results of previous studies showing that the fracture risk in children and adolescents is generally higher in boys than in girls and increases with age.14,29,30 Furthermore, the distribution of the fracture sites is compatible with other studies showing that fractures of the lower arm, carpal and metacarpal region, and the phalanges of the hand are most common.29,30 The study has a number of limitations. Although the smoking status of patients is generally recorded in the GPRD, this information was not available for the vast majority of children and adolescents included in this analysis; we also did not have any information on environmental tobacco smoke exposure. Although it can be speculated that active smoking is a negligible confounder in children under the age of 10, smoking might be much more prevalent in prepubertal children and adolescents. Therefore, we were unable to control for this potential confounder, which may be associated with asthma or adverse effects on lung growth and development31,32 and eventually also with exposure to corticosteroids. Moreover, at least in adults, smoking has been associated with the occurrence of certain fractures.33
Furthermore, the database does not provide information on physical activity. Physical activity might be a potential confounder because it may be related to asthma and exposure with inhaled steroids as well as the fracture risk. Previous findings showed that sport participation in prepubertal children is associated with an increased bone mineral density, particularly in boys, which might be expected to result in a decreased fracture risk in later life.34 However, increased physical activity and sports participation may increase the fracture risk in children.14,29,35 Moreover, it has to be kept in mind that the database includes information on drug prescriptions, which may differ to some degree from the actual drug use by the patient. In case of a hospitalization, the drugs given during the hospital stay are not, or only incompletely, recorded by the general practitioner. This may lead to some underestimation of the use of oral corticosteroids in patients who are hospitalized as a result of asthma exacerbation necessitating a course of oral corticosteroids while in hospital. In addition, noncompliance does occur but cannot be reliably quantified and therefore cannot be taken into account in the analysis.
Another limitation related to the structure of the database is the lack of information on socioeconomic status as well as on dietary factors (eg, regular intake of dairy products). In addition, although height and weight are routinely recorded in adults, these parameters are often missing in children; therefore, we could not take body mass index into account in the analysis.
In conclusion, our study shows that treatment with inhaled steroids does not seem to materially affect the risk of bone fractures in children or adolescents. However, there was some evidence that longer term exposure to inhaled steroids in individuals with concomitant or past oral steroid exposure might slightly increase the fracture risk. Additional longitudinal follow-up studies in children or adolescents exposed to inhaled steroids over several years may further address this issue and quantify the fracture risk. This would be especially interesting if studies are focused on inhaled corticosteroids other than beclomethasone, which was the drug with the highest exposure prevalence in the present study.
Acknowledgments
The Boston Collaborative Drug Surveillance Program is partly supported by grants from Hoffmann-La Roche. Dr Meier is the recipient of a grant from the Swiss National Science Foundation (grant 32–056 751). The present study was not funded.
Footnotes
- Received May 23, 2003.
- Accepted November 5, 2003.
- Reprint requests to (C.R.M.) Basel Pharmacoepidemiology Unit, Division of Clinical Pharmacology and Toxicology, University Hospital, Petersgraben 4, CH-4031 Basel, Switzerland. E-mail: meierch{at}uhbs.ch
REFERENCES
- Copyright © 2004 by the American Academy of Pediatrics
In this issue
Jump to section
Related Articles
Cited By...
- Bone mineral density and fracture risk with long-term use of inhaled corticosteroids in patients with asthma: systematic review and meta-analysis
- Ambulatory intravenous ceftriaxone in paediatric A&E: a useful alternative to hospital admission?
- The use of inhaled corticosteroids in the wheezy under 5-year-old child
- Effect of Long-term Corticosteroid Use on Bone Mineral Density in Children: A Prospective Longitudinal Assessment in the Childhood Asthma Management Program (CAMP) Study
- Global strategy for asthma management and prevention: GINA executive summary
- Childhood onset arthritis is associated with an increased risk of fracture: a population based study using the General Practice Research Database