Glucocorticoids, prescribed for their antiinflammatory and immunosuppressive features, are the most common cause of secondary osteoporosis. Used to treat a variety of diseases affecting almost every bodily system, glucocorticoids can result in bone loss irrespective of the disease being treated or the age or sex of the person receiving the drug. A detailed discussion of the pathogenesis of bone loss in glucocorticoid-induced osteoporosis (GIOP) is beyond the scope of this editorial, but recent reviews are available on this topic (1, 2).

Community survey data estimate the frequency of oral use of glucocorticoids as 0.5% of the general population and 1.75% of women older than age 55 (3). The estimated incidence of GIOP is 50% in patients treated for >6 months (4, 5), and fractures occur in approximately one-third of those patients undergoing long-term oral glucocorticoid therapy (>1 year) (6, 7). Bone loss and fractures ensue most commonly in postmenopausal women (8), but men and menstruating women are not spared these complications (2, 9).

There is rapid onset of trabecular bone loss (10–20%) as early as 3 months after initiation of therapy, followed by a slower rate of 2% per year thereafter. In contrast, bone loss in the femoral neck proceeds more slowly (2–3% in the first year), but continues over time (10). In the year prior to initiation of glucocorticoid treatment, the incidence of nonvertebral fractures was 1.6 per 100 person-years, and it increased to 2.0 per 100 person-years within 3 months of starting therapy and then remained constant thereafter (8). In this same study, similar patterns of increasing fracture rates were noted for hips, vertebrae, and forearms. Parallel trends in loss of bone mineral density (BMD) equivalent to 1 SD or doubling of the fracture risk in postmenopausal women (11) and an increased risk of hip (22%) and vertebral (51%) fractures were documented in individuals receiving inhaled glucocorticoids compared with nonusers (12).

Low BMD and subsequent fracture are related in part to dosage of glucocorticoid (8). The relative rate of vertebral fractures increased from 1.55 to 5.18 in patients receiving low doses (<2.5 mg daily) compared with patients in the high-dose (>7.5 mg daily) group. Even the relative rates of nonvertebral fractures increased from 1.17 to 1.36 from the lowest to the highest dose groups. These data suggest that even low doses of glucocorticoids are associated with an increased risk of fracture.

However, other factors may be important in determining fracture risk in patients exposed to glucocorticoids. Bone loss may also be accelerated when the underlying disease process compromises bone architecture, either from the inflammation process itself or from immobility due to the disease, such as in rheumatoid arthritis (RA), systemic lupus erythematosus, inflammatory bowel disease, or asthma (7, 13–15). Advancing age is also a risk factor for both men and women taking oral glucocorticoids (16), as is the risk of falling (8).

The possibility of decreased fracture risk and recovery back to baseline levels within 2 years after glucocorticoids were discontinued has been noted in several studies; however, lifetime risk of fracture remains increased even after glucocorticoid treatment is withdrawn (8, 17, 18). Even though the chronic nature of the diseases requiring treatment often precludes discontinuing medications, physicians consistently strive to accomplish the objective of tapering glucocorticoids to the lowest possible dose. Patients sustaining osteoporotic fractures have a significant decrement in their quality of life, and the economic cost of these events is estimated at US $50–100 million per annum per million population. Extrapolating these observations to prevent fracture in GIOP suggests that prophylactic therapy is also likely to be cost effective in those patients requiring long-term glucocorticoid treatment (19, 20).

In this issue of Arthritis & Rheumatism, Solomon et al present their retrospective medical record review of rheumatologists' osteoporosis management in patients with RA who were taking a mean daily dosage of 8.8 mg of prednisone over 15 months. Only 23% of patients had BMD testing, 42% were taking prescription medications for osteoporosis, and no more than 25% had documentation of taking calcium and/or vitamin D acknowledged in their electronic medical record. According to regression models, men and menstruating women were less likely to have BMD testing or to receive treatment for osteoporosis, the authors' benchmarks for assessing adequate osteoporosis management. Physician factors and glucocorticoid dosage did not explain the lack of testing or treating those RA patients receiving glucocorticoids for long periods in this study (21).

Despite the inherent limitations of a retrospective chart review, most notably missing or incomplete data, the authors concluded that academic rheumatologists' management of GIOP was suboptimal (21). These observations are consistent with results of other recent studies in which physician-based (3, 22–24) and patient-based (25) surveys worldwide highlight deficiencies in obtaining BMD testing in high-risk patients and/or in prescribing medication for the prevention and treatment of GIOP.

Furthermore, of all doctors queried in another study, rheumatologists were among the best of the physicians surveyed in implementing diagnostic and therapeutic options for patients at risk for low BMD and subsequent fracture due to GIOP (22). Indeed, the rates of osteoporosis treatment excluding calcium and vitamin D in previous studies ranged from 6% to 29%, so that the 42% reported in the current article may actually reflect improvement. However, the authors fail to comment on what benchmark they recommend for achieving adequate testing and treatment of GIOP, or over what time frame these management strategies should be achieved (21). Suggested criteria would assist those planning future studies in assessing the status of management practices in GIOP.

Practice guidelines have been published for multiple clinical problems, but, as shown in the current (21) and previous studies (3, 23, 24), significant disparities remain between published recommendations and actual practice in the management of osteoporosis (26). The next steps are to evaluate attitudes and opinions toward practice guidelines and to assess the evidence base used to develop those recommendations as potential barriers to their implementation.

Studies of practicing physicians indicate that experienced clinicians may be implementing guidelines selectively. Many clinicians are concerned that guidelines are based on randomized trials and do not reflect the complexity of the real world, in which the context and framework of a decision are important (27). In a survey of osteoporosis identification and treatment among primary care physicians treating patients hospitalized for fractures (28), 39% reported that BMD measurement was unnecessary or that their patients were too frail. In the current article, Solomon et al also noted that having at least 1 comorbid condition was one of the factors associated with not receiving treatment (21). Additionally, potential adverse effects of medication and cost of drug were cited as the main factors limiting treatment (28). Even though the primary care physician response was low in this survey, deterrents to care focused on cost and potential adverse effects for patients at high risk for future fractures.

In another survey, two-thirds of those in general practice in the UK remained unconvinced of the efficacy of drug therapy for osteoporosis. Furthermore, these general practitioners expressed a need for more education on osteoporosis (29). Other investigators and the authors of the current article (21, 30) have also recommended education as part of the approach to heighten awareness of GIOP.

Examining the knowledge base from which the American College of Rheumatology (ACR) recommendations (31, 32) were developed may also assist our evaluation of barriers, even though the study by Solomon et al (21) was not specifically designed to test implementation of these recommendations. It is also important to note that the current study surveyed physician behaviors from 1998 to 2000, when data were continuously emerging from multiple clinical trials on GIOP prevention and treatment.

Advances in the measurement of BMD by dual x-ray absorptiometry (DXA) improved precision in the areas of interest (lumbar spine, femoral neck, and forearm), and recommendations regarding treatment options and approaches to several patient scenarios were reviewed in 1996 in the ACR's published guidelines for the prevention and treatment of GIOP (31). For example, these guidelines state that it is reasonable to measure BMD in all individuals who are likely to take glucocorticoids for at least 6 months, and to begin preventive therapies as soon as the clinical situation permits. First-line treatment recommendations included calcium and vitamin D, hormone replacement therapy (oral contraceptives for young women with oligomenorrhea or amenorrhea and testosterone for hypogonadal men), and lifestyle modifications (weight-bearing exercise, smoking cessation, and limitation of alcohol consumption). However, only 5 years after the initial ACR recommendations for managing GIOP were published, important legislation affecting reimbursement for BMD measurement and emerging treatment options necessitated a revision of these guidelines (32).

In 1997, the US Congress passed the Bone Mass Measurement Act, and this legislation included Medicare reimbursement for persons receiving glucocorticoid treatment, thus relieving financial barriers to BMD testing for these individuals (33). A review published in 1998 outlined clinical indications for BMD measurements, provided guidelines for the interpretation of results obtained with the DXA technique, and also compared DXA with other available radiographic methods (34). In 2001, the ACR committee reinforced its initial recommendations to obtain BMD measurements at baseline and as part of longitudinal assessments at 6–12-month intervals (32).

However, a closer review of the evidence reveals important caveats when using BMD for diagnosis and for monitoring of therapeutic efficacy. The incidence of fracture is higher in patients taking glucocorticoids than in those with postmenopausal or involutional osteoporosis, even though BMD is relatively higher in the former group (35). As noted by Manolagas (36), bone mass alone does not explain the rapid onset (within 3 months) and offset (within 1 year) of fractures in patients taking glucocorticoids noted by Van Staa and colleagues (8). Reduction in the risk of vertebral fracture is seen soon after treatment with antiresorptive agents is initiated to prevent GIOP (37), and the rate becomes similar to that in postmenopausal women not receiving glucocorticoids (38, 39). Riggs and Melton hypothesize that the common efficacy thread among the spectrum of antiresorptive agents is that they normalize bone turnover (40). High bone turnover disrupts the microarchitecture of bone, and one could speculate that a similar mechanism operates in GIOP.

Microarchitecture is undoubtedly important for bone strength (41), but clinical methods for assessing bone architecture do not yet exist, and BMD does not measure fragility of bone. While BMD measurements are included in most guidelines, the limitations noted above may contribute to a reduction in the use of this test by practitioners.

When considering use of treatment, the years from 1996 to 2000 reflect a time of intense activity. The results from large randomized controlled trials and a meta-analysis of these studies contributed evidence that antiresorptive therapies were effective in GIOP (40). In these studies, change in BMD, not fracture, was used as the primary outcome measure. At the end of this period, development of bone-building drugs emerged.

In the trials of bisphosphonates, the percent difference in lumbar spine BMD between treatment and placebo groups ranged from 2.5% to 3.7% over 48–52 weeks across all studies (37, 42, 43). In the meta-analysis, the weighted mean difference of BMD between the treatment and placebo groups was 4.3% (95% confidence interval [95% CI] 2.7, 5.9) at the lumbar spine and 2.1% (95% CI 0.01, 3.8) at the femoral neck (44). No individual study has definitively addressed fracture reduction. However, when vertebral fracture data were pooled from all study participants for the meta-analysis, a 24% reduction in the odds of spinal fractures (odds ratio 0.76, 95% CI 0.37, 1.53) was noted (44). Although fracture data were limited, the reduction in vertebral fracture risk occurred primarily in postmenopausal women, the group at highest risk for this outcome (43), and a similar trend was also noted in men (37).

To put the fracture data in clinical perspective, Sambrook estimated that the number of postmenopausal women needing treatment with a bisphosphonate for 1 year to prevent 1 vertebral fracture ranged from 5 to 26, depending on the initial BMD measurement, previous fracture occurrence, and methodology to demonstrate a radiographic vertebral fracture (5).

In another meta-analysis, a modest beneficial effect of vitamin D with calcium versus no therapy or calcium alone was shown (effect size 0.60, 95% CI 0.34, 0.85) (45). Furthermore, this study also provided guidance on the relative effectiveness of bisphosphonates and calcitonin compared with vitamin D for GIOP for the outcome measure BMD. However, due to limited or no studies on estrogen or selective estrogen receptor modulator (SERM) efficacy on BMD improvement and none on fracture prevention in GIOP, comparison of estrogen or SERM use was not included in the comparative analysis of vitamin D versus other therapies. Similarly, there were limited data on BMD outcomes and no fracture data on calcitonin or hormone strategies for men exposed to exogenous steroids (5) reviewed in this comparative analysis. Specifically, bisphosphonates were more effective than vitamin D (effect size 0.57, 95% CI 0.09, 1.05), and calcitonin was similar in efficacy to vitamin D (effect size 0.03, 95% CI −0.39, 0.45). Although vitamin D was noted to be less effective than some GIOP therapies in improving BMD, these authors suggested that the minimum medication prescription for patients at risk for GIOP should be vitamin D and calcium, especially if low doses of steroids are used.

Although not included in these meta-analyses of antiresorptive medications, the best strategy for preventing or treating GIOP should be an anabolic agent, which would increase bone mass and bone strength (1). In recent clinical trials, daily subcutaneous injections of parathyroid hormone (PTH) were safe and effective in postmenopausal women (46) and in those receiving hormone replacement therapy and long-term glucocorticoid therapy (47). New information on the pathophysiologic effects of PTH suggests that an increased ratio of bone effectors, osteoprotegerin to receptor activator of nuclear factor κB ligand, permits bone formation and diminishes bone resorption (48).

The conclusions from the ACR task force and the 2 meta-analyses regarding treatment options parallel those by Sambrook, especially for postmenopausal women, and probably apply to older men receiving glucocorticoids too (5, 32). Manolagas derived similar conclusions from his report on GIOP, in which he shows that there is a mechanistic explanation for bisphosphonates preventing fractures and that the clinical data support the hypothesis that they are beneficial in fracture reduction in GIOP (1). However, the real benefit in treating GIOP is likely to be realized from anabolic medicines, such as PTH, which increase BMD by 9–13% and reduce the risk of new vertebral fractures by 65–69% when compared with placebo. These bone health benefits exceed benchmarks reported from the use of antiresorptive agents in similar patients. The strategies for prevention of GIOP include minimizing glucocorticoid doses, modifying lifestyle factors, and using a bisphosphonate followed by vitamin D with calcium, since calcium alone is not sufficient (5, 32). The evidence for calcitonin is limited, but it might be an alternative if other therapies are poorly tolerated (45). Estrogen or testosterone could be considered if women or men, respectively, are hypogonadal (49, 50). Since combination therapy with hormones and bisphosphonate enhanced BMD (51) in postmenopausal women not taking glucocorticoids, this approach could be considered in those postmenopausal women exposed to glucocorticoids. Recommendations regarding the use of anabolic agents are likely to be included in future guidelines.

In contrast, for premenopausal women and younger men receiving glucocorticoids, the data for decisionmaking regarding osteoporosis management are less robust. The assessment of risks for bone loss and consequent fracture in younger individuals exposed to glucocorticoids depends on baseline BMD, anticipated dosage and duration of steroid use, and other risk factors including future pregnancies. Sambrook even advocates a role for careful observation with serial BMD measurements in these younger individuals, assuming the absence of other factors such as prevalent vertebral fractures or the presence of a markedly low BMD (5). A recent study confirms that age is a risk factor for fracture independent of BMD (16) and supports observation as a possible strategy in younger individuals taking glucocorticoids.

Therefore, based on the above observations and the results of the study by Solomon et al (21), are we receiving conflicting signals concerning the utility of BMD measurements, especially in younger individuals? Do we also need more studies to assess risks and treatment options in younger patients? Even though antiresorptive agents reduce fractures, are we limited by toxicity and cost, as suggested by primary care physician surveys? Should we aim for even better efficacy and consider new treatments that form bone, such as anabolic agents, or combinations of currently available treatment options? Is the deficiency in the management of GIOP noted in numerous past studies and the current article just a reflection of inadequate education? Areas for future investigation include not only clarification of the items noted above, but also assessment of bone microarchitecture and turnover in these patients to refine risk reduction strategies, and the impact of management practices on the critical outcome variable, fracture.

The current article highlights the importance of continued vigilance in assessing our management practices (21), especially in the setting of rapidly changing science. Developing recommendations is only part of the task. Implementing and assessing the achievement of targets such as fracture reduction are equally essential goals. Rheumatologists do not want to miss an opportunity to lead the effort in minimizing the consequences of GIOP.

Addendum. Recent results from the Women's Health Initiative (WHI) document a modest hip fracture reduction, but this benefit was outweighed by the increased risk of breast cancer, stroke, coronary heart disease, and thromboembolic events, in postmenopausal women receiving estrogen and progestin. Continued use of hormone replacement therapy over time in the WHI trial was also associated with increasing risks of the adverse events noted above (52). Although the women in the WHI were not taking glucocorticoids, rheumatologists should now incorporate these new data when recommending hormone replacement therapy for long-term use in women at risk for GIOP.


  1. Top of page