Efficacy and Safety of Daily Risedronate in the Treatment of Corticosteroid-Induced Osteoporosis in Men and Women: A Randomized Trial


  • Supported by Procter & Gamble Pharmaceuticals and Hoechst Marion Roussel.


Long-term use of high-dose corticosteroids often results in bone loss, which may lead to osteoporosis-related fractures. This was a multicenter, double-blind study in which 290 ambulatory men and women receiving high-dose oral corticosteroid therapy (prednisone ≥ 7.5 mg/day or equivalent) for 6 or more months were randomized to receive placebo, risedronate 2.5 mg/day, or risedronate 5 mg/day for 12 months. All patients received calcium 1 g and vitamin D 400 IU daily. The primary endpoint was lumbar spine bone mineral density (BMD) at month 12. Additional measurements included BMD at the femoral neck and trochanter and the incidence of vertebral fractures. Overall, there were statistically significant treatment effects on BMD at 12 months at the lumbar spine (p < 0.001), femoral neck (p = 0.004), and trochanter (p = 0.010). Risedronate 5 mg increased BMD at 12 months by a mean (SEM) of 2.9% (0.49%) at the lumbar spine, 1.8% (0.46%) at the femoral neck, and 2.4% (0.54%) at the trochanter, whereas BMD was maintained only in the control group. Although not powered to show fracture efficacy, we observed a reduction in the incidence of vertebral fractures of 70% in the combined risedronate treatment groups, relative to placebo (p = 0.042). Risedronate was well tolerated, had a good safety profile, and was not associated with gastrointestinal adverse events. We conclude that risedronate increases BMD and potentially reduces the incidence of vertebral fractures in patients with corticosteroid-induced osteoporosis.


CORTICOSTEROIDS ARE USED extensively for their immunosuppressive and anti-inflammatory properties. Long-term therapy with high-dose oral corticosteroids (prednisone ≥ 7.5 mg/day or equivalent) often results in bone loss and corticosteroid-induced osteoporosis (CIO).(1, 2)

Bone loss is greatest in the first 12 months of corticosteroid therapy, continues at a lower rate thereafter, and predominantly affects trabecular bone.(1, 2) Bone mineral density (BMD) at the lumbar spine has been shown to decrease by 8% after 20 weeks of treatment with prednisone at an average daily dose of 7.5 mg;(3) reductions in vertebral trabecular BMD approaching 40% have also been reported.(4)

CIO is associated with an increased risk of vertebral and hip fractures: approximately 30% of adults receiving corticosteroids for 5 years or longer experience CIO-related fractures.(1) Vertebral fractures may occur at higher BMD thresholds in patients with CIO than in women with postmenopausal bone loss.(5)

Risedronate, a potent pyridinyl bisphosphonate,(6) induces biochemical remission in Paget's disease of bone,(7–9) and maintains or increases BMD in postmenopausal osteoporosis,(10) and chemotherapy-induced bone loss in postmenopausal women.(11)

The aim of this double-blind, placebo-controlled, parallel-group, multicenter study was to evaluate the efficacy, safety, and tolerability of risedronate in patients who had been receiving high-dose oral corticosteroid therapy for at least 6 months.


Role of the funding source

Employees of Procter and Gamble Pharmaceuticals participated in the study as coinvestigators. They took part in the design of the study with the other investigators and worked with the external research organization to implement the protocol and coordinate data collection and statistical analysis. They also contributed to the writing of this study. Data interpretation and decisions about the content of the report and submission for publication resided with the entire group of investigators.


Ambulatory men and women aged 18–85 years were enrolled from 23 study centers in Europe. All patients had been receiving oral corticosteroids (mean daily dose of prednisone ≥ 7.5 mg, or equivalent) for at least 6 months. Patients were stratified as follows: (1) men; (2) premenopausal women, using an acceptable form of birth control or who were surgically sterile; and (3) women who had been postmenopausal for at least 1 year.

Exclusion criteria included: conditions that might interfere with the evaluation of spinal osteoporosis; history of hyperparathyroidism, hyperthyroidism, or osteomalacia within a year before the study; or history of sarcoidosis or cancer. Patients were also excluded if they had taken (within 6–12 months, depending on the medication) or were still taking medication known to affect bone metabolism, including hormone replacement therapy. Low-dose vaginal estrogen (17-β-estradiol ≤ 0.2 mg/day; estropitate ≤ 1.5 mg/day) was permitted. Patients were not excluded on the basis of current or previous gastrointestinal illness or use of concomitant medications associated with gastrointestinal symptoms.

All patients gave written informed consent to participate in the study, which was conducted in accordance with Good Clinical Practices and the Helsinki Declaration. The protocol was approved by the appropriate ethics committees.

Dosage and administration

Patients were randomized to receive placebo or risedronate (2.5 or 5 mg) film-coated tablets, once daily for 12 months. Patients were instructed to take the study drug with at least 240 ml of water, on an empty stomach, 30–60 minutes before breakfast and remain upright for at least 1 h. All patients were required to take vitamin D 400 IU daily with breakfast and elemental calcium 1 g (as calcium carbonate) with lunch or dinner.

Bone mineral density assessment

BMD at the lumbar spine, femoral neck, femoral trochanter, and forearm was measured by dual-energy X-ray absorptiometry (DXA) with Lunar (Madison, WI, U.S.A.) or Hologic (Waltham, Massachusetts, U.S.A.) densitometers at 0, 6, and 12 months. All scans were analyzed at the Department of Radiology, University of California, San Francisco (UCSF), U.S.A. Standardized lumbar spine BMD was calculated at baseline to adjust for instrument differences.(12, 13) To ensure consistent instrument performance, DXA phantom data were analyzed by UCSF throughout the study.(14)

Fracture assessments

Vertebral fractures were assessed from thoracic and lumbar (T4-L4) lateral and anterior-posterior spinal radiographs taken at baseline, and lateral spinal radiographs taken at month 12. Prevalent vertebral fractures were diagnosed as described by Melton et al.(15) In accordance with the recommendations of the National Osteoporosis Foundation Working Group on Vertebral Fractures,(16) potential incident fractures were identified using quantitative morphometry, defined as a reduction of ≥ 15% in vertebral height in a previously intact vertebra or a reduction of ≥4 mm in a previously fractured vertebra. Fractures identified in this manner were subsequently visually verified by a radiologist.

Adverse events

Patient-reported adverse events were recorded. A physical examination was performed at 6 and 12 months. Endoscopy was performed at the investigator's discretion in patients who reported moderate or severe upper gastrointestinal adverse events. Clinical laboratory tests were performed at 0, 1, 3, 6, and 12 months.

Statistical analysis

The planned enrollment was 273 patients, with an expected withdrawal rate of 35% over 12 months. The study had at least 90% power to detect a between-treatment group difference of ≥3% in mean percentage change from baseline in lumbar spine BMD at month 12, with a two-sided α = 0.05 significance level and a standard deviation of 5%.

Statistical analysis of the results was performed on the intent-to-treat population, which included all patients who were randomized and received at least one dose of study medication. At baseline, variables were compared by analysis of variance (ANOVA) with treatment, investigator, and stratum as factors, or by Cochran-Mantel-Haenszel test, stratified by investigator.

At 6 and 12 months, ANOVA was used to investigate the overall treatment effect on the mean percentage change from baseline in BMD with treatment, investigator, and stratum as factors. ANOVA was also used to explore effects within strata and for subgroups of underlying disease. Supplementary analysis of covariance (ANCOVA) was performed. The model included treatment group, investigator, stratum, mean dose, duration of prestudy corticosteroid therapy, and mean concomitant corticosteroid dose. Fisher's exact test was used for analysis of vertebral fracture.


Patient population

A total of 290 patients were enrolled (mean age 59 years; 109 men, 181 women). Fourteen percent of the women were premenopausal. Baseline characteristics were similar in the three treatment groups (Tables 1 and 11 ). One third of patients had at least one vertebral fracture at baseline (Table 2) and the mean lumbar spine T-score was −1.6. The duration and mean daily doses of prednisone (or equivalent) before and during the study were similar across treatment groups.

Table Table 1.. Baseline Characteristics and Corticosteroid Therapy
original image
Table Table 2.. Baseline Bone Mineral Density and Vertebral Fracture Prevalence
original image

Of the 285 patients who received study drug, 78% completed 12 months of treatment. Twelve percent of patients discontinued because of adverse events and 7% withdrew voluntarily. Another 3% did not complete the study because of protocol violation or were lost to follow-up. Reasons for discontinuation were similar across the treatment groups.

Bone mass

At 12 months there were significant differences in mean percentage change from baseline among the groups at the lumbar spine (p < 0.001), femoral neck (p = 0.004), and trochanter (p = 0.010) (Table 3). BMD was maintained at each skeletal site in the placebo group. Risedronate 5 mg increased BMD (mean ± SEM) from baseline at these sites by 2.9 ± 0.49%, 1.8 ± 0.46%, and 2.4 ± 0.54%, respectively (Fig. 1). The improvements in BMD, compared with placebo, in the risedronate 5 mg group were 2.7 ± 0.67% (p < 0.05), 1.9 ± 0.67% (p < 0.01), and 1.6 ± 0.81% (p < 0.05), respectively, at these sites. The corresponding differences for the risedronate 2.5 mg group were 1.2 ± 0.68%, −0.1 ± 0.68%, and −0.9 ± 0.82%, respectively, none of which was significant. Although risedronate 2.5 mg did increase spinal BMD relative to baseline at the lumbar spine (1.9 ± 0.5%, p < 0.01), there was no significant improvement compared with placebo at this site and no significant change compared with placebo or baseline at the hip sites.

Table Table 3.. Percentage Change from Baseline in Bone Mineral Density at Month 12
original image
Figure FIG. 1..

Mean (SEM) change from baseline in bone mineral density of the lumbar spine (A), the femoral neck (B), and the femoral trochanter (C) by study visit in patients receiving placebo (○), risedronate 2.5 mg (▴), or risedronate 5 mg (▪).#Significantly different from placebo (p < 0.05) based on a three-way ANOVA model; *significantly different from baseline (p < 0.05) based on a one-sample t-test.

Subgroup analysis by stratum at 12 months showed that risedronate 5 mg increased BMD, relative to placebo, at all sites in men and postmenopausal women. Because only eight premenopausal women were included in the risedronate 5 mg group, the results were not considered meaningful. Risedronate 5 mg increased BMD from baseline by 2.1–4.8% across the skeletal sites in men, and by 1.7–2.4% in postmenopausal women (Table 3). The change from baseline in BMD was numerically greater in men than in postmenopausal women at all three skeletal sites, with the greatest difference observed at the lumbar spine.

Risedronate 5 mg increased BMD at 12 months in patients with rheumatoid arthritis, lung diseases, and all other diseases combined (Fig. 2). Despite the small numbers reported in some of the disease subgroups, these results do show a general trend to a positive treatment effect across a wide range of conditions requiring corticosteroids, although the changes were not statistically significant in the lung disease group.

Figure FIG. 2..

Mean percentage change (SEM) from baseline in lumbar spine bone mineral density at month 12 in patients receiving placebo (white bar) or risendronate 2.5 (dotted bar) or 5 mg (striped bar), categorized according to underlying disease.Asthma, chronic obstructive pulmonary disease, and chronic interstitial lung disease;polymyalgia rheumatica, systemic lupus erythematosus, temporal arteritis, vasculitis, polymyositis, pemphigoid, pemphigus, Behçet's disease, dermatomyositis, and eczema; *significantly different from baseline (p < 0.05; one-sample t-test).

Risedronate 5 mg maintained BMD of the distal radius at 12 months (−0.6%), whereas there was a loss in the placebo group. There were no significant differences between risedronate and placebo at the midshaft or distal radius (Table 3). Directionally, there were nonsignificant decreases, with the exception of risedronate 5 mg at the midshaft radius (−0.5%, p < 0.05) and placebo at the distal radius (−2.0%, p < 0.05).


Nine out of 60 (15%) patients in the placebo group and three out of the 60 (5%) patients in each treatment group experienced incident vertebral fractures by month 12. The estimated reductions in the incidence of vertebral fracture, relative to placebo, were 70% (p = 0.125) in both the risedronate 2.5 mg group and the risedronate 5 mg group, with significance being achieved when the active treatment groups were combined (p = 0.042).

Adverse events

Overall, the proportions of patients reporting adverse events, drug-related adverse events, serious adverse events, and upper gastrointestinal adverse events were similar in each of the three treatment groups (Table 4). Thirty-three patients withdrew from the study because of adverse events (11 from each treatment group).

Table Table 4.. Summary of Overall and Upper Gastrointestinal Adverse Events
original image

Back pain (23% vs. 10%) and arthralgia (24% vs. 16%) were reported more frequently in the risedronate 5 mg group than in the placebo group. In general, these events were not considered to be related to the study drug and were mild in severity. No differences were observed in the incidence of moderate or severe back pain in the three groups. Furthermore, no patients withdrew from the study because of back pain or arthralgia.

The overall incidence of upper gastrointestinal adverse events was similar in the risedronate 5 mg and placebo groups, with fewer in the risedronate 2.5 mg group. Most events were of mild severity (Table 4). Eleven patients had endoscopies (four placebo, four risedronate 2.5 mg, and three risedronate 5 mg). The endoscopy results showed that in the placebo group, one patient had a hiatal hernia and ulceration and erosion of the stomach, one had lower esophageal inflammation that was normal at a later endoscopy, one had a hiatal hernia, and one had no abnormalities. In the risedronate 2.5 mg group, the endoscopies revealed one case each of: hiatal hernia; inflammation of the midesophagus and stomach antrum; duodenal erosions; and esophageal inflammation with a hiatal hernia. In the risedronate 5 mg group, the endoscopies revealed stomach erosions and a hiatal hernia in the first patient, duodenal ulceration in the second patient, and no abnormalities in the third patient.

Nonvertebral fractures occurred in six (6%) patients in the placebo group, eight (9%) in the risedronate 2.5 mg group, and eight (8%) in the risedronate 5 mg group, with similar distributions across skeletal sites.

Laboratory evaluations indicated no clinically significant changes in hematological, renal, and hepatic function. Mean values for serum calcium and phosphorus were similar in the treatment groups at baseline and remained similar throughout the study.


This study showed that 12 months of treatment with risedronate 5 mg increases BMD in patients receiving high doses of oral corticosteroids; the 2.5 mg dose was less effective. This was observed both in women and men with a variety of diseases requiring corticosteroid therapy. Furthermore, the incidence of vertebral fractures was lower in the active treatment groups compared with placebo. The combined data for the risedronate 2.5 and 5 mg groups revealed a statistically significant 70% reduction in fracture incidence (p = 0.042). A similar trend was observed in the risedronate 5 mg group alone, although this was not statistically significant (p = 0.125). These data suggest that 12 months of risedronate 5 mg will have a positive effect on incident fracture risk in patients with CIO.

There was a significant 3% increase in lumbar spine BMD from baseline after 12 months of treatment with risedronate 5 mg. Twelve months of treatment with risedronate 5 mg also produced significant increases from baseline in BMD at the femoral neck and trochanter. The estimated difference between the risedronate 5 mg and placebo groups was 1–2%, a difference that was statistically significant at the femoral neck. Increases from baseline in BMD at the femoral neck were observed in the men and postmenopausal women with risedronate 5 mg treatment; the increase in the small group of premenopausal women was not statistically significant.

In the placebo group, the absence of a significant reduction in BMD at any site or stratum over the 12-month period may have been a consequence of the administration of daily vitamin D and calcium supplements, because this may be beneficial in maintaining BMD but has not been shown to reduce fracture rates.(17) Also, BMD loss may have reached a plateau by the time therapy was initiated; bone loss is most rapid during the first year of treatment with corticosteroids, and, on average, patients had been on corticosteroid therapy for approximately 5 years. Despite the maintenance of BMD, the placebo-treated patients (receiving calcium and vitamin D) continued to experience a high incidence of fractures (15%) during the trial.

Risedronate was well tolerated in our study; the incidence and severity of adverse events were similar in the active treatment and placebo groups, although the incidence of mild back pain and arthralgia was greater in the risedronate 5 mg group.

Various therapies have been used in an attempt to reverse or prevent corticosteroid-induced bone loss. Some success in increasing BMD has been achieved with hormone replacement therapy in postmenopausal women and hypogonadal men,(18, 19) as well as with calcidiol,(20) calcitriol, and calcitonin.(21, 22) Other bisphosphonates, such as etidronate, pamidronate, and alendronate, have been found to prevent bone loss in patients with CIO.(1, 23) Studies with etidronate over 1–2 years have shown increases in BMD at the lumbar spine ranging from 0.6 to 7%.(24–26) Saag et al.(27) showed that 1 year of therapy with alendronate 10 mg resulted in increases in lumbar spine BMD of 3%. An increase in the incidence of gastrointestinal adverse effects (mostly abdominal pain) was noted in patients receiving alendronate, although there was no increase in the incidence of esophagitis. In our study, risedronate was not associated with an increase in the incidence of gastrointestinal adverse events, compared with the placebo group, despite the fact that patients with previous or current gastrointestinal diseases were not excluded from the study.

Recent recommendations stipulate that bisphosphonate therapy should be considered for prevention of bone loss in patients receiving oral corticosteroids, and for treatment of those patients who develop fractures while receiving corticosteroids.(28) Although it is not yet known whether the protective effects on bone density and the reduction in the incidence of vertebral fractures are sustained beyond 1 year, our study showed that 12 months of risedronate 5 mg effectively increased BMD.

In summary, 12 months of daily treatment with risedronate 5 mg was safe and well tolerated. Risedronate therapy increased BMD at the lumbar spine, femoral trochanter, and femoral neck in patients with CIO. Improvements in BMD at the lumbar spine after risedronate therapy were also associated with a reduction in incident vertebral fractures.


  1. 1

    For the European Corticosteroid-Induced Osteoporosis Treatment Study.