OUP user menu

Use of bisphosphonates and dual-energy X-ray absorptiometry scans in the prevention and treatment of glucocorticoid-induced osteoporosis in rheumatology

E. Wall, K. Walker-Bone
DOI: http://dx.doi.org/10.1093/qjmed/hcm126 317-323 First published online: 12 February 2008


Background: Patients treated with steroids are at risk of glucocorticoid-induced osteoporosis. Appropriate investigations and therapeutic agents can decrease rate of bone loss and fracture.

Aim: To review adherence to current UK guidelines for the prevention of glucocorticoid-induced osteoporosis in rheumatology outpatient clinics.

Design: Retrospective case note review.

Methods: The management of patients taking glucocorticoids who attended outpatient rheumatology clinics at a Teaching Hospital NHS Trust over a 4-week period was reviewed against current UK recommendations for prevention and treatment of osteoporosis (Bone and Tooth Society, National Osteoporosis Society, Royal College of Physicians. Glucocorticoid-induced osteoporosis: guidelines for prevention and treatment. London: Royal College Physicians, 2002).

Results: Over the study period, 519 patients attended rheumatology outpatient clinics, amongst whom 104 were current glucocorticoid users. Most patients had been taking oral steroids for over 12 months (n = 79, 76%). The majority had also received steroids by at least one other route (n = 67, 64.4%). According to the guidelines, 51 patients, at relatively low risk of osteoporosis (<65 years, no previous fragility fracture) should have been referred for bone density assessment; of these, 27 (53%) had received a DEXA scan. In total, 58 subjects fulfilled criteria for bisphosphonates (>65 years, fragility fracture, T-score <−1.5) and, of these, 51 (87.9%) were appropriately treated. In 21 cases, a DEXA scan had been performed when guidelines recommended that treatment could commence without further assessment.

Discussion: The findings indicate a high level of awareness of glucocorticoid-induced osteoporosis amongst UK rheumatologists. Most patients identified to be at high risk of bone loss were offered treatment. Although encouraging, current practice could potentially be improved, particularly through more targeted use of DEXA scanning.


Therapy with oral glucocorticoids is associated with bone loss and an increased risk of fracture.1–3 Greatest bone loss is found at higher doses of steroids and with longer duration of treatment;2–4 however even low-dose glucocorticoid therapy has been associated with bone loss and fragility fracture.3 During the past decade, increasing evidence has accrued that therapeutic agents can prevent glucocorticoid-induced bone loss and reduce the associated risk of fracture and National and International guidelines in the prevention of osteoporotic fracture among patients taking glucocorticoids have been published.5–8

Previous studies among inpatients and outpatients across different specialties have shown that the majority of glucocorticoid-treated patients have not been evaluated for risk of osteoporosis or commenced on treatment to prevent accelerated bone loss and future fracture.9–16 In rheumatology, glucocorticoids are widely used in the management of many conditions including connective tissue diseases, rheumatoid arthritis and vasculitis. Rheumatologists are among those with the greatest experience of glucocorticoid therapy and have been shown to be more likely to report that they would prescribe preventive therapy17 and more likely than other specialist doctors and general practitioners to measure bone density and offer preventive measures and treatment of glucocorticoid-induced osteoporosis.18–21

Despite guidelines, several studies in the USA and UK have suggested that significant numbers of patients receiving glucocorticoids do not have bone density assessment or receive prophylactic treatment when appropriate.19–22 In the UK, the Royal College of Physicians have recommended that the management of patients taking glucocorticoids be reviewed against guidelines to prevent osteoporosis in primary and secondary care.1 This study explores the level of adherence to the current British guidelines for the prevention of glucocorticoid-induced osteoporosis among a random sample of rheumatology outpatients taking low to moderate doses of glucocorticoids.


All adult rheumatology outpatients attending a clinic at one of two hospitals in a British Teaching Hospital NHS Trust during the 4-week study period (9 October to 4 November 2006) were eligible for inclusion in this study. All subjects who were current users of oral glucocorticoids were included. For each patient, information was obtained from the case notes and recorded onto a proforma which was designed to extract information about disease, demography, oral glucocorticoid exposure (minimum, maximum, cumulative and current doses), other glucocorticoid exposure (inhaled, intramuscular, intra-articular, intravenous), risk factors for osteoporosis [smoking, alcohol intake, menopause <45 years, previous fragility fracture, dose steroids >15 mg/day, family history of osteoporosis, body mass index (BMI) <19 kg/m2], investigations including bone densitometry and treatments prescribed.

Based on the 2002 guidelines, all patients with exposure or commitment to exposure to glucocorticoids were included. Those aged 65 years and over, or with prior fragility fracture, should have received prophylactic treatment for glucocorticoid-induced osteoporosis. Patients aged <65 years without prior fragility fracture should be assessed for their risk of fragility fracture by dual energy X-ray absorptiometry (DEXA) scanning and then treated if appropriate.


In total, 519 rheumatology outpatients were seen during the 4-week study period amongst whom 104 were identified as current users of oral glucocorticoids (prevalence of glucocorticoid use among rheumatology outpatients 20.0%) (Table 1). In total, 77 women and 27 men (mean age 61.8 years) were currently taking glucocorticoids. Where stated, the majority of subjects were Caucasian (n = 41, 39.4%) but ethnicity was generally poorly documented (n = 60, 57.7%). Rheumatoid arthritis (n = 61) was the most common single diagnosis requiring glucocorticoids but a significant minority of patients had more than one diagnosis requiring glucocorticoids (n = 22, 21.2%). Table 2 summarizes the dose, duration and route of steroid treatment. Most subjects (n = 61, 58.7%) were taking <7.5 mg of daily prednisolone at the time of the study, but the mean daily dose amongst all subjects was 8.56 mg. The majority of patients (n = 67, 64.4%) had also been exposed to steroids by at least one alternative route: 47 (45.2%) had received intramuscular glucocorticoids; 43 (41.3%) had been treated with intra-articular glucocorticoids; 12 (11.5%) intravenous glucocorticoids; and 15 (14.4%) inhaled glucocorticoids. The majority of study subjects had been taking glucocorticoids in excess of 12 months (n = 79, 76%) and only six participants had commenced treatment within the last 3 months.

View this table:
Table 1

Characteristics of participants

Mean age (years)61.8
    Asian Indian21.9
    Asian Oriental11
    Not specified6057.7
    Rheumatoid arthritis6158.7
    Polymyalgia rheumatica/temporal arteritis2422.4
    Connective tissue disease1312.5
    Other rheumatological condition1918.3
    Other non-rheumatological condition76.7
View this table:
Table 2

Steroid exposure

Steroid exposureNumber(%)
Current dose <7.5 mg/day6158.7
Current dose 7.5–15 mg/day2927.9
Current dose >15 mg/day1413.5
Mean current daily dose glucocorticoids (mg)8.56
Mean total cumulative dose of glucocorticoids (mg)11488.6
Highest cumulative dose of glucocorticoids(mg)84217.9
Duration of oral glucocorticoids <1 month65.8
Duration of oral glucocorticoids 1–6 months54.8
Duration of oral glucocorticoids 6–12 months1413
Duration of oral glucocorticoids >12 months7976
Exposure or commitment to oral glucocorticoids >3 months10096.2
Ever had steroids by non-oral route6764.4
Intramuscular steroids4745.2
Intra-articular steroids4341.3
Intravenous steroids1211.5
Inhaled steroids1514.4

Risk factors and preventive factors were obtained from the case notes where available (Table 3). Cigarette smoking and alcohol intake were recorded in most cases; 18 patients were documented to be current cigarette smokers (17.3%); only one subject was considered to have an excessive alcohol intake (1.0%). Evidence of a fall in the last 6 months was found among five patients (4.8%). Exercise habits were infrequently documented, with only three subjects recorded as taking exercise on a regular basis (three or more times/week) (2.9%). Among the female patients, 13 (16.9%) were either pre-menopausal or were taking HRT. According to the guidelines, patients were deemed to be at ‘high risk’ of osteoporosis if: aged >65 years (n = 45, 43.3%); taking >15 mg/day steroids (n = 14, 13.5%); have a positive family history of osteoporosis (defined as the presence of any osteoporotic fracture affecting either parent; n = 1, 1.0%) have low BMI (<19 kg/m2) (n = 6, 5.7%); or, if female, had experienced the menopause at age <45 years (n = 6, 7.8% female subjects). A previous low-impact fracture was documented in the notes of 12 subjects (five vertebral, two hip, two radius, two rib, one humerus). In many cases, risk factors and preventive factors were poorly documented, most notably exercise, menopausal age, family history and previous falls. A past history of fragility fracture was also very difficult to elicit reliably from the notes.

View this table:
Table 3

Risk factors for osteoporosis

Yes (%)No (%)Not specified (%)
Risk factors
    Fall in the last 6 months5 (4.8)099 (95.2)
    Smoker18 (17.3)80 (76.9)6 (5.8)
    Alcohol intake above recommended limits1 (1.0)94 (90.4)9 (8.7)
Preventive factors
    Male/premenopausal female40 (38.5)61 (58.7)3 (2.9)
    regular exercise3 (2.9)23 (22.1)79 (76.0)
High risk
    Dose steroids >15 mg/day14 (13.5)90 (86.5)0
    Age >65 years45 (43.3)59 (56.7)0
    Low Body Mass Index6 (5.7)96 (9.2)2 (1.9)
    Early menopause (female only)6 (7.8)26 (33.8)45 (58.4)
    Family History of osteoporosis1 (1.0)0103 (99.0)
    Previous Low impact fracture12 (11.5)1 (1.0)91 (87.5)
Low impact fracture (n = 12)n (%)
    Vertebral5 (4.7)
    Hip2 (1.9)
    Radius2 (1.9)
    Rib2 (1.9)
    Humerus1 (1.0)

When reviewing data according to the 2002 guidelines, four subjects were excluded from further analysis of investigation and treatment of glucocorticoid-induced osteoporosis, since it was unclear whether there was commitment to continue steroids for 3 months. Of the 100 patients who had commitment or exposure to oral steroids for >3 months, 48 had been referred for a bone density scan (Table 4). According to the guidelines, all patients aged <65 years without evidence of a fragility fracture should have been referred for DEXA. In total, 51 subjects in this study fulfilled this criterion, amongst whom, 27 (53%) had been referred for at least one DEXA, 17 of these within 12 months of commencement of the glucocorticoids (Figure 1). DEXA scan had been performed in an additional 21 subjects (21%) in whom the guidelines recommend preventive treatment without measurement of bone mineral density.

Figure 1.

Summary of adherence to guidelines.

View this table:
Table 4

Investigations and treatment of patients with >3 months commitment to steroids

Investigations and treatmentNumber(%)
Ever had DEXA
    <65 years31/5655.3
    >65 years17/4438.6
DEXA indicated and performed27/5152.9
DEXA not indicated and performed21/4942.9
General measures
    Vitamin D54/10054
    Stopped treatment due to side effects2/1002
    Recommended to GP but not commenced1/1001
Patients taking bisphosphonate
    Also taking calcium and vitamin D41/7554.7
    Taking calcium or vitamin D2/752.7
Treatment declined22
No treatment documented2020
Number documented as on treatment where guidelines indicate need51/5887.9
Number who declined treatment when guidelines indicate need2/583.4
Number who were not documented to be offered/taking treatment5/588.6
Number on treatment where not indicated11/1861.1
Number on treatment where need uncertain16/2466.7

Overall, 75 (75%) patients were currently documented as taking bisphosphonates (Table 4), the most popular being once weekly alendronate (n = 48, 48%), followed by weekly risedronate (n = 24, 24%) and cyclical etidronate (n = 3, 3%). Comparison with the guidelines suggested that 58 patients fulfilled criteria for bisphosphonates (aged >65 years, past history of fragility fracture, T-score <−1.5), amongst whom 51 (87.9%) were treated appropriately; two ‘at risk’ patients (3.4%) were recommended but had declined treatment and in five cases (8.6%), there was no evidence that treatment had been recommended. No patients were identified in whom the use of bisphosphonates was contraindicated. In total, 18 patients in this sample did not need bisphosphonates according to the guidelines, (age <65 years, without previous fragility fracture, T-score >−1.5) amongst whom 11 patients (61.1%) were receiving bisphosphonates. In an additional 16 cases, the patient was receiving bisphosphonates but whether or not this was over-treatment could not be fully assessed due to the absence of a DEXA scan.

In total, 54 (54%) patients were receiving calcium and vitamin D supplementation, recommended as general measures for the prevention and treatment of glucocorticoid osteoporosis. This included 41 of the 75 patients taking a bisphosphonate (54.7%), one patient was documented to be taking just calcium and another to be receiving vitamin D on its own. There was no record of adjunctive supplementation in 44 patients in this sample, 32 of whom were documented to be taking a bisphosphanate.


The results of this study suggest a high level of awareness of the risks of glucocorticoid-induced osteoporosis among practising UK rheumatologists. A prevalence of glucocorticoid use of 20% among rheumatology outpatients is a figure similar to that found in other rheumatology surveys,22–24 but a relatively high proportion (41%) of the patients in this sample were currently receiving doses of >7.5 mg/day. According to the current UK guidelines, our sample contained 58 patients at ‘high risk’ of osteoporotic fracture for whom bisphosphonates should be prescribed. In our study, 53 (91%) had been appropriately recommended treatment, although two patients had declined and in only five ‘high risk’ cases (8.6%) was there no evidence that treatment had been suggested. Compared with the results of other studies among glucocorticoid patients in the general population,9,,25 ophthalmology clinics.14 medical wards12,,26 and medical outpatient clinics.15 the results of this study suggest that rheumatologists are very much more likely to consider the risk of glucocorticoid-induced bone loss and prescribe treatment.

Calcium and vitamin D were recorded as being taken by just over half of subjects (54%). This figure is similar to other studies conducted in rheumatology outpatient departments in the UK,22–23 but better than a study in primary care where only 34.1% patients were co-prescribed supplements despite recommendations.27 A total of 32 patients were receiving a bisphosphonate without documented supplementation with calcium and vitamin D, an unexpected finding that may reflect true clinical practice or may be contributed to by poor documentation in clinical notes.

Our results suggest that rheumatologists were not optimizing the use of DEXA scans in practice. In total, 21 subjects in this survey had been referred for a DEXA scan when the guidelines suggest that treatment should be initiated without measurement. Furthermore, according to the guidelines, a DEXA scan was indicated in the assessment of another 24 patients for whom one had not been performed. Bone loss is maximal early in the course of glucocorticoids and therefore bone densitometry should be measured early. Even taking account of delays between requesting the investigation and the scan being carried out and reported, it was not anticipated that this would be more than 6 months. The results of our study showed that in only a third of those for whom scans were indicated had the DEXA actually been carried out within 12 months of commencement of glucocorticoids; this was largely due to delayed referral for scanning. It appears therefore that DEXA scans are neither being targeted appropriately nor being requested early enough into the course of glucocorticoids in most patients. These findings were not expected and may reflect the changing nature of the GIOP prevention guidelines in the UK of which four sets were published (1995–2002).1,5,6,28 It may be that in practice, rheumatologists are very aware of risk and requirement for prevention and are tending to request a scan for most patients without having appreciated that the most high-risk patients should be treated early with bisphosphonates without waiting for a scan. This issue will be addressed by additional education in the department and will be reviewed in 12 months.

Our findings suggest that rheumatologists are tending to prescribe therapy for more patients than recommended by the guidelines (at least 11 patients in this sample were receiving bisphosphonates outside the guidelines). It is possible that this may represent over-treatment in some cases but it was not within the scope of this study to investigate this. Guidelines are however only for guidance, unable to take account of individual factors, such as patient anxiety and it is possible that 10% of patients in real life practice fall outside of common circumstances as encompassed by guidelines.

When interpreting the results of this study, account must be taken of several limitations. This was a retrospective case note review. Such a study design allowed an accurate assessment of real-life current practice, without the introduction of bias by influencing usual practice. However, eliciting information from case notes is notoriously problematic—written documentation may frequently under-represent oral communication between patient and physician. Risk factors for osteoporosis were often not documented; ethnicity, exercise, menopausal age, family history and previous falls and fragility fractures were often not clearly recorded. Similarly, investigations may have been requested or treatment commenced without this being documented in the notes. It is noteworthy that there was marked heterogeneity in the information in patients’ notes. In particular, patients assessed in nurse clinics, where a standardized proforma is used, tended to have much more complete information including things not recorded elsewhere, such as lifestyle, risk factors and medication.

A significant proportion (64.4%) of the study population had been exposed to glucocorticoids through intramuscular, intra-articular, intravenous and inhaled routes as well as orally. There is some evidence that long-term use of high-dose inhaled steroids impacts on bone density, although evidence of the effects of conventional doses are not established.29 Similarly, pulsed doses of intramuscular30 and intravenous31 steroid are associated with glucocorticoid-induced osteoporosis. Less evidence exists for an influence of intra-articular injection on bone density.32 Given the evidence that non-oral steroid therapy may affect bone density, calculated cumulative steroid doses may be inaccurate when based only on oral glucocorticoid intake.

The results of this study are promising and indicate a high level of awareness of glucocorticoid-induced osteoporosis amongst practising rheumatologists. Clinicians are however tending to treat some patients with bisphosphonates outside the guidelines, either failing to use DEXA scanning as a screening tool or, in other cases, undertaking unnecessary scans. With the numbers involved, our results suggest that it would be virtually cost-neutral to improve our practice and conform with the guidelines by targeting DEXA scans to those who do need them in order to assess risk.


View Abstract