To evaluate the risk of fracture in patients receiving intermittent therapy with high-dose oral glucocorticoids (GCs).
To evaluate the risk of fracture in patients receiving intermittent therapy with high-dose oral glucocorticoids (GCs).
The study group comprised 191,752 patients from the UK General Practice Database who were 40 years of age and older and received therapy with GCs. The followup time period was divided into the categories of “current” and “no exposure.” The daily dose and cumulative dose for each time period were determined. Relative risks were estimated using Cox proportional hazards models, adjusted for age, sex, body mass index, smoking, disease history, and drug history. Fractures of the radius/ulna, humerus, rib, femur/hip, pelvis, or vertebrae were included in the evaluation.
Patients who intermittently received high-dose GCs (daily dose ≥15 mg) and had no or little previous exposure to GCs (cumulative exposure ≤1 gm) had a small increased risk of osteoporotic (but not hip/femur) fracture; this risk increased substantially with increasing cumulative exposure. Among patients who received a daily dose ≥30 mg and whose cumulative exposure was >5 gm, the relative risk (RR) of osteoporotic fracture was 3.63 (95% confidence interval [95% CI] 2.54–5.20), the RR of fracture of the hip/femur was 3.13 (95% CI 1.49–6.59), and the RR of vertebral fracture was 14.42 (95% CI 8.29–25.08).
Intermittent use of high-dose oral GCs (daily dose ≥15 mg and cumulative exposure ≤1 gm) may result in a small increased risk of osteoporotic fracture. Conversely, patients who receive several courses of high-dose GCs (daily dose ≥15 mg and cumulative exposure >1 gm) have a substantially increased risk of fracture.
Oral glucocorticoids (GCs) are frequently prescribed to patients with inflammatory disorders, such as obstructive airway disease, rheumatoid arthritis (RA), and inflammatory bowel disease (IBD). Treatment with oral GCs has been associated with increases in the risk of fractures, particularly fractures of the hip and vertebrae (1–3). This effect is dose dependent and occurs rapidly after the start of treatment (2–4). However, previous studies have not distinguished between the effect on fractures of intermittent therapy and long-term therapy with oral GCs. Several studies have evaluated the association between intermittent use of high-dose GCs and bone mineral density (BMD). A study in patients with Crohn's disease (5) demonstrated a rapid loss of BMD following short-term treatment with oral GCs, but this effect was not observed in 3 other studies (6–8). Those studies included only a small number of patients and differed in terms of the extent of GC exposure. Importantly, fracture is the clinically more important end point, and it has been reported that the risk of fracture in patients receiving oral GCs may be partially independent of BMD (1). The objective of the current study was to evaluate the risk of fracture in patients receiving intermittent treatment with high-dose GCs.
We conducted a retrospective cohort study among patients from the UK General Practice Research Database (GPRD). The GPRD contains computerized medical records, prescription data, specialist referrals, and hospital discharge summaries of patients enrolled in 683 general practices in England and Wales (9). A cohort of patients ages 40 years and older, all of whom received oral GCs between January 1987 and December 1997, was created. This cohort was previously analyzed (10), but that analysis focused on the effects of continuous use of oral GCs on the risk of fractures, with the daily dose of oral GCs averaged over the total treatment period.
Because patients may change their daily dose of oral GCs over time, and because high doses of oral GCs are used for only short periods of time, the current analysis evaluated the association between the daily dose of oral GCs and fracture risk, using time-dependent analyses and correlating the daily dose for each GC treatment episode to the risk of fracture. In addition, the current study compared patients currently receiving GCs with patients who had received GCs in the past, while our previous study compared patients receiving oral GCs with controls. This approach was used in order to minimize confounding by indication.
In this study, the daily dose of GCs was estimated separately for each GC prescription. The total period of followup for each patient was taken as the time period from the first GC prescription until the end of data collection for the GPRD. This total period of time was then divided into periods of current exposure and past exposure, with patients moving between current and past exposure. Each period of current exposure started with a GC prescription and ended 3 months after the expected duration of GC therapy. The expected duration of GC therapy was based on how the drug was supplied and the prescribed daily dose (as determined by instructions from the patient's general practitioner). In cases in which patients received a repeat prescription within a current exposure period, this period was then extended using the expected duration of the repeat prescription. In the event of overlap between 2 prescriptions (i.e., a repeat prescription given within the duration of use for a previous prescription), the “overlap” days were added to the duration of the repeat prescription. When data on the prescribed quantity or daily dose were missing, the median expected duration of treatment (based on data for patients of similar age and sex) was used.
At the time that each course of GCs was prescribed, the prior cumulative GC exposure within the current exposure period was calculated. We assumed that a period of intermittent use was characterized by high daily doses and low cumulative exposure. Each patient was followed up until he or she sustained a clinical osteoporotic fracture (i.e., a fracture of the radius/ulna, femur/hip, vertebra, humerus, or pelvis) following the first GC prescription, or until the patient's transfer out of the general practice, death, or the end of the study (whichever came first).
The following risk factors were considered as potential confounders: prior prescriptions for anticonvulsants, antiarrhythmics, hypnotics/anxiolytics, antidepressants, antiparkinson drugs, or nonsteroidal antiinflammatory drugs, and the occurrence of falls 6 months prior to the GC prescription. In addition, a history of cerebrovascular disease, dementia, fractures, or anemia prior to the GC prescription was evaluated as a potential confounder. Smoking status and body mass index were also determined.
Cox proportional hazards models were used to estimate the relative risk (RR) of fracture. The past exposure periods were considered the reference category, in order to minimize confounding by indication. Spline regression analyses were used to summarize the association between time between repeat GC prescriptions and the risk of fractures. This method has been advocated as an alternative to categorical analysis (11). All analyses were performed using SAS version 9.1.3 software (SAS Institute, Cary, NC).
Baseline characteristics of the study subjects are shown in Table 1. The cohort included 191,752 patients who received GCs, of whom 111,349 (58.1%) were women. During and after therapy, patients who received oral GCs sustained a total of 7,412 osteoporotic fractures, of which 2,144 were hip/femur fractures and 1,269 were vertebral fractures. A total of 114,125 patients had received at least 1 prescription for a daily dose of ≥15 mg. The mean age of patients who received high-dose GCs was 64 years, and the most frequent indication for GC treatment was obstructive airway disease.
|Characteristic||At least 1 prescription for <15 mg GCs (n = 131,259)||At least 1 prescription for ≥15 mg GCs (n = 114,125)|
|Female sex||77,027 (58.7)||65,245 (57.2)|
|Age at baseline, years|
|40–59||42,032 (32.0)||40,116 (35.2)|
|60–79||65,309 (49.8)||56,473 (49.5)|
|≥80||23,918 (18.2)||17,536 (15.4)|
|Diseases requiring GCs|
|Obstructive airway disease||64,312 (49.0)||61,663 (54.0)|
|Polymyalgia rheumatica||11,822 (9.0)||7,332 (6.4)|
|Inflammatory bowel disease||7,006 (5.3)||4,902 (4.3)|
|Rheumatoid arthritis||6,859 (5.2)||2,525 (2.2)|
|Treatment 6 months before first GC prescription|
|Bronchodilators||54,701 (41.7)||50,291 (44.0)|
|NSAIDs||25,177 (19.2)||19,731 (17.3)|
|Hypnotics/anxiolytics||22,624 (17.2)||18,495 (16.2)|
|Rectal GCs||2,390 (1.8)||1,776 (1.6)|
As shown in Table 2, patients receiving GCs at a dosage of <7.5 mg/day had a 60% increased risk of osteoporotic fracture. With a daily dose of 7.5–15 mg, this risk was increased by 115%. Further increases in the daily dose were not associated overall with larger increases in risk (the risk of osteoporotic fractures was increased by 62% overall in patients receiving a daily dose ≥15 mg). In patients who received high-dose GCs (≥15 mg/day), a strong association with cumulative GC exposure was observed. Patients who intermittently received high-dose GCs but who previously had no or low cumulative exposure to GCs (daily dose ≥15 mg and cumulative exposure ≤1 gm) had a small increased risk of osteoporotic and vertebral fracture (but not hip/femur fracture), while the risk increased substantially in patients with high cumulative exposure (Figure 1). Among patients who were receiving a daily dose ≥30 mg and whose cumulative exposure was >5 gm, the RR was 3.63 (95% confidence interval [95% CI] 2.54–5.20) for osteoporotic fracture, 3.13 (95% CI 1.49–6.59) for hip/femur fracture, and 14.42 (95% CI 8.29–25.08) for vertebral fractures.
|GC exposure||Osteoporotic fracture||Hip/femur fracture||Vertebral fracture|
|Adjusted for age and sex only||Adjusted for age, sex, and other factors†||Adjusted for age and sex only||Adjusted for age, sex, and other factors†||Adjusted for age and sex only||Adjusted for age, sex, and other factors†|
|Current use||1.68 (1.60–1.76)||1.60 (1.52–1.68)||1.60 (1.46–1.76)||1.49 (1.36–1.63)||3.51 (3.08–3.99)||3.21 (2.81–3.65)|
|<7.5 mg/day||1.60 (1.50–1.71)||1.53 (1.43–1.63)||1.51 (1.34–1.69)||1.40 (1.24–1.57)||3.16 (2.70–3.69)||2.89 (2.47–3.39)|
|7.5–15 mg/day||2.15 (1.97–2.34)||1.99 (1.82–2.16)||2.44 (2.11–2.82)||2.17 (1.87–2.51)||4.59 (3.79–5.57)||4.00 (3.29–4.86)|
|≥15 mg/day||1.62 (1.30–1.77)||1.56 (1.44–1.69)||1.50 (1.28–1.68)||1.43 (1.22–1.68)||3.63 (3.01–4.34)||3.35 (2.77–4.04)|
New users of high-dose oral GCs did not have a significantly increased risk of osteoporotic fracture (RR 0.97, 95% CI 0.81–1.16). This risk was still not significantly increased among patients who received 1–4 prescriptions for oral GCs (RR 1.21, 95% CI 0.92–1.59) before the start of a new treatment episode, whereas patients who received more than 4 previous GC prescriptions had a significantly increased risk of osteoporotic fracture (RR 1.51, 95% CI 1.19–1.92).
An analysis of the time since discontinuation of GC therapy showed that the RR of fracture was not increased in patients who had stopped receiving high-dose GCs more than 12 months previously. Moreover, we observed a rapid decrease in the fracture risk beginning 3 months after discontinuation of treatment (Figure 2). Among patients who received low-dose GCs (<15 mg/day), the risk of fracture was no longer increased in those who had received their most recent prescription for oral GCs at least 9 months previously (RR 1.23, 95% CI 0.90–1.70). Figure 3 shows the risk of osteoporotic fracture among users of high-dose GCs who had high cumulative exposure or low cumulative exposure. The risk of fracture returned to baseline levels after 6 months in patients who were exposed to ≤1 gm of prednisone equivalent. In contrast, patients who received high-dose GCs and were exposed to >1 gm of prednisone equivalent were not at increased risk of fracture if they had received their last prescription at least 15 months before the index date.
Table 3 shows the RR of fracture, stratified according to the indication for GC therapy. No substantial differences were observed between the indications for GCs in terms of the risk of fracture. Patients who received a daily dose ≥15 mg and whose cumulative exposure was >5 gm had a 3-fold increased risk of osteoporotic fractures, regardless of the underlying disease.
|Fracture type/disease at baseline||Past exposure||Current exposure, daily dose <15 mg||Current exposure, daily dose ≥15 mg|
|≤1 gm||1–5 gm||>5 gm|
|Obstructive airway disease||1.00||1.69 (1.58–1.82)||1.32 (1.10–1.58)||1.85 (1.45–2.37)||3.00 (2.30–3.91)|
|Arthropathy||1.49 (1.36–1.63)||1.21 (0.90–1.63)||1.92 (1.46–2.53)||2.69 (2.01–3.59)|
|Inflammatory bowel disease||1.56 (1.27–1.91)||1.34 (0.75–2.40)||1.57 (0.86–2.85)||3.20 (1.93–5.30)|
|Obstructive airway disease||1.00||1.37 (1.18–1.58)||0.90 (0.58–1.39)||1.37 (0.78–2.38)||1.94 (1.03–3.65)|
|Arthropathy||1.41 (1.20–1.66)||0.79 (0.40–1.55)||1.64 (0.96–2.78)||2.44 (1.42–4.19)|
|Inflammatory bowel disease||0.95 (0.66–1.38)||0.99 (0.35–2.84)||0.35 (0.04–2.62)||1.77 (0.63–4.91)|
|Obstructive airway disease||1.00||3.67 (3.08–4.36)||1.71 (1.08–2.73)||5.65 (3.74–8.53)||10.61 (6.98–16.12)|
|Arthropathy||3.24 (3.08–4.36)||1.38 (0.59–3.22)||8.12 (5.19–12.74)||10.13 (6.06–17.08)|
|Inflammatory bowel disease||2.89 (1.74–4.80)||1.45 (0.33–6.33)||6.06 (2.51–14.55)||11.58 (5.08–26.41)|
In this study, we observed that intermittent use of high doses of oral GCs was associated with only a small increase in the risk of osteoporotic fracture. Nevertheless, continued use of high-dose oral GCs (cumulative exposure >1 gm of prednisone equivalent) was associated with substantially higher risks of fracture. We did not observe major increases in the risk of fracture when short-term GC treatment was given at least 1 year after previous treatment. There were also no substantial differences in the relative increases in fracture risk between patients with obstructive airway disease, those with arthropathies, and patients with IBD.
The findings from this study extend previous results for the same study population (2, 3). In earlier studies with our study population, daily doses of oral GCs were averaged over the total treatment period. The current study focused on the time of use of high-dose oral GCs. Consistent with our earlier investigations that focused on the effects of continuous treatment with oral GCs (3), in the current study we observed a strong association between daily dose of oral GCs and risk of fracture in patients who continuously received GCs. However, we also observed that this strong dose response did not apply to patients who received high doses of oral GCs for short periods of time or intermittently. We observed a small increased risk of osteoporotic or vertebral fracture among patients who received high-dose oral GCs and whose cumulative intake was ≤1 gm of prednisone equivalent (intermittent users). The risk of hip/femur fracture was not increased.
To our knowledge, this is the first study to demonstrate that intermittent treatment with high-dose GCs may result in only a small increased risk of fractures. The current findings extend results from studies that evaluated the association between short-term, intermittent use of high-dose oral GCs and BMD (5, 7, 8, 12). Reduced BMD is an independent risk factor for fracture (13). Intermittent use of high-dose GCs (daily dose ≥10 mg and cumulative exposure ≤1 gm) did not change BMD in men with chronic obstructive pulmonary disease. In addition, significantly decreased femoral neck and lumbar spine BMD was observed after cumulative exposure >1 gm of prednisone equivalent (6). Conversely, longer duration of GC treatment (1,260 mg, with tapering over 2 months) resulted in decreased femoral neck BMD in patients with Crohn's disease (5).
The GC regimen corresponded with a daily dose of 22.5 mg, and these results are consistent with our observation of an increased risk of hip fracture with cumulative exposure >1 gm among patients who received high-dose GCs. Although we already observed an increased risk of vertebral fracture among patients with obstructive airway disease who had received a cumulative dose of <1 gm, our cumulative dose-response trend was similar to the cumulative exposure threshold of 1 gm of prednisone equivalent with regard to the risk of hip fracture.
Several mechanisms have been suggested for the accelerated bone loss and decreased bone formation rates that lead to an increased risk of fracture in patients receiving oral GCs. Bone resorption may be facilitated by decreased absorption of calcium and phosphate in the gut and renal tubules, increased parathyroid hormone secretion, and increased osteoclastic activity. Decreased bone formation can be the result of decreased osteoblast and preosteoblast proliferation, sex steroid deficiency, and increased apoptosis of osteoblasts and osteocytes (14, 15). However, the risk of fracture in patients receiving oral GCs is partially independent of decreased bone mass (1), and the increased fracture risk associated with GCs may occur 3 months after treatment is initiated (2). Increased apoptosis of osteoblasts and osteocytes and decreased osteocyte viability may lead to bone microdamage and, ultimately, bone fragility (16–19). Damage to the microarchitecture of trabecular bone has been shown to be a risk factor for vertebral deformities (20). In addition, it has been suggested that the degree of disruption of the trabecular network of bone correlates with cumulative exposure to oral GCs (21). GG-induced disruption is probably reversible, which is consistent with our observation of a rapid decrease in fracture risk 3–12 months after cessation of GCs.
Our study has several limitations. First, data regarding treatment with GCs during hospitalization and treatment prior to the start of data collection for the GPRD were not recorded, which may have resulted in an underestimate of the effects of oral GCs. Second, information on smoking and body mass index was available for only part of the study population. Information on potential confounders such as diet and exercise was not available. Last, the comparison group in this study consisted of past users of GCs. This approach was used to reduce confounding by indication. Previously, it was demonstrated that fracture risk decreased rapidly toward baseline levels after discontinuation of GC therapy (3). A sensitivity analysis with stratification for past use according to “recent past” and “distant past” showed no substantial effect on the results. Thus, it is unlikely that any residual effects of GCs would have resulted in a substantive underestimate of the effects of oral GCs.
In conclusion, intermittent use of high-dose oral GCs (daily dose ≥15 mg and cumulative exposure ≤1 gm) may result in a small increased risk of osteoporotic fracture. Conversely, patients who receive several courses of high-dose GCs (daily dose ≥15 mg and cumulative exposure >1 gm) are at substantially increased risk of fracture. Further investigations and preventative treatment might be conveniently targeted to these patients.
Dr. van Staa had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Mr. de Vries and Drs. Leufkens, Cooper, and van Staa.
Acquisition of data. Dr. van Staa.
Analysis and interpretation of data. Mr. de Vries and Drs. Bracke, Leufkens, Lammers, Cooper, and van Staa.
Manuscript preparation. Mr. de Vries and Drs. Bracke, Leufkens, Lammers, Cooper, and van Staa.
Statistical analysis. Mr. de Vries and Dr. van Staa.
Procter & Gamble Pharmaceuticals had no role in the study design or in the collection, analysis, or interpretation of the data.