A number of good-quality systematic reviews (SRs) have found significant effects of bisphosphonates (alendronate, risedronate and zoledronic acid) in reducing fracture risk. Few studies have directly compared different agents or classes of agents used to treat osteoporosis and hence the data are insufficient to determine the relative efficacy or safety of these agents.
Primary prevention of osteoporotic fractures with bisphosphonates
A pivotal good quality SR2 in 2002 included two randomised controlled trials (RCTs) (McClung et al and Hoskings et al, n = 1946)13,14 that reported the effect of alendronate 10–40 mg per day on fracture risk in postmenopausal women without osteoporosis. Alendronate was not associated with a reduction in the risk of vertebral fracture (relative risk [RR]: 0.45, 95% confidence interval [CI]: 0.06–3.15) or non-vertebral fractures (RR: 0.79, 95% CI: 0.28–2.24) compared to placebo. In the McClung trial,13 mean age was 51.8 years and T-score –1.8 and no patients had prevalent vertebral fractures, while in the Hosking trial,14 participant mean age was 53 years, mean T-score –1.8 and <10% had prevalent vertebral fractures. A Cochrane systematic review and MA in 200815 reported a reduction in the risk of vertebral fractures (RR: 0.55, 95% CI: 0.38–0.80) but no reduction in non-vertebral fracture risk (RR: 0.89, 95% CI: 0.76–1.04) with alendronate therapy in one study (Cummings et al, n = 4432).16 The mean T-score was –1.9, mean age 67.6 years and no patients had prevalent vertebral fractures. The fact that patients in the latter study were older by 15 years is likely to have contributed to the positive findings.
An MA of RCTs of risedronate17 in postmenopausal women (n = 111 in one trial by Mortensen et al)19 conducted in 2002 did not demonstrate reductions in vertebral (RR: 2.44, 95% CI: 0.12–49.45) or non-vertebral (RR: 0.49, 95% CI: 0.12–2.03) fractures. A 2008 Cochrane review18 on the effectiveness of risedronate at doses of 2.5 mg per day and 5.0 mg per day for a duration of two years for primary prevention of osteoporosis fractures included two RCTs (Mortensen et al and Hooper et al)19,20 with 327 early postmenopausal women (mean age 52.6). Results were not significant compared to placebo for either vertebral (RR: 0.97, 95% CI: 0.42–2.25) or non-vertebral fracture risk (RR: 0.81, 95% CI: 0.25–2.58). In the Mortensen and Hooper trials, the mean ages were 51.2 years and 52.6 years, the mean T-scores were –1.0 and –0.4, and 0% and 18% of the subjects had prevalent fractures, respectively.19,20
Treatment of postmenopausal women at high risk of osteoporotic fracture
A good-quality MA including six treatment trials showed a reduction in the risk of vertebral fracture for alendronate compared to placebo (RR: 0.53, 95% CI: 0.43–0.65), with no heterogeneity observed between trials.2 This translated to a number needed to treat (NNT) of 72 (95% CI: 61–99) to prevent one vertebral fracture over two years of treatment in women considered to be at high risk of vertebral fracture. The patients included were at high fracture risk, as indicated by a weighted mean age of 67.8 years (range: 59.5 to 71 years), weighted mean femoral neck (FN) T-score of –2.6 (range: –3.3 to –2.3) and prevalent vertebral fractures weighted mean 29% (range: 0% to 100%). Among five treatment trials included in the MA of non-vertebral fracture risk, the weighted mean age was 63.0 years (range: 59.5 to 64 years), weighted mean FN T-score of –2.7 (range: –2.8 to –2.3) and prevalent vertebral fractures weight mean 18% (range: 0% to 21%). The pooled RR for non-vertebral fracture was 0.49 (95% CI: 0.36–0.67), with no heterogeneity between trials. NNT to prevent one non-vertebral fracture over two years of treatment in women at high risk was 24 (95% CI: 19–37).2
A Cochrane review of RCTs in postmenopausal women compared risedronate 2.5 mg or 5.0 mg daily to placebo over 2–3 years.18 These trials were categorised as osteoporosis treatment (or secondary prevention) trials, based upon inclusion criteria of T-score ≤-2.0 or the presence of a prevalent vertebral fracture. Pooled data from three RCTs showed a 39% reduction in vertebral fractures (RR: 0.61, 95% CI: 0.50–0.76) for risedronate 5.0 mg per day with an estimated NNT of 48. The weighted mean age was 69.1 years (range: 64.7–71 years), weighted mean FN T-score of –2.7 (range: –2.9 to –2.4) and prevalent vertebral fractures weighted mean 79% (range: 30–100%). Pooled data from four RCTs showed a 20% reduction in non-vertebral fractures (RR: 0.80, 95% CI: 0.72–0.90), with an estimated NNT of 30. The weighted mean age was 76.9 years (range: 64.7–78 years), weighted mean FN T-score of –3.6 (range: –3.7 to –2.4) and prevalent vertebral fractures weight mean 47% (range: 42% to 100%). For hip fractures, there was a 26% reduction in risk (three RCTs, RR: 0.74, 95% CI: 0.59–0.94), with an estimated NNT of 202. The weighted mean age was 77.3 years (range: 69–78 years), weighted mean FN T-score of –3.6 (range: –3.7 to –2.4) and prevalent vertebral fractures weight mean 47% (range: 30–100%). The effect observed for 2.5 mg risedronate was not as large.18
The multi-centre international Health Outcomes and Reduced Incidence with Zoledronic Acid Once Yearly (HORIZON)–Pivotal Fracture Trial (PFT) has followed the safety and efficacy of zoledronic acid in a cohort of postmenopausal women with osteoporosis for nine years. In the initial trial, 7765 patients (mean age 73) were randomised to receive either placebo or a single infusion (5 g) of zoledronic acid at trial baseline, 12 months and 24 months. At 36 months from baseline, zoledronic acid treatment reduced the risk of morphometric vertebral fracture by 70%, compared to placebo (3.3% versus 10.9%, RR: 0.30, 95% CI: 0.24–0.38), and reduced the risk of hip fracture by 41% (1.4% versus 2.5%, hazard ratio [HR]: 0.59, 95% CI: 0.42–0.83).21 A post hoc analysis of the data from patients who received only one dose of zoledronic acid at baseline revealed a similar reduction in vertebral fracture risk (68%) at 18 months follow up, compared with placebo.22
Treatment following hip fracture
An annual infusion of zoledronic acid within three months after a hip fracture was associated with a reduction in the rate of new clinical fractures and improved survival in women and men with an average age of 74.4 years, followed for a median of 1.9 years.23 In this randomised, double-blind, placebo-controlled trial, the rates of any new clinical fracture were reduced by 35% (P = 0.001) from 13.9% in the placebo group to 8.6% in the zoledronic acid group. The respective rates of a new clinical vertebral fracture were 1.7% and 3.8% (P = 0.02), and the respective rates of new non-vertebral fractures were 7.6% and 10.7% (P = 0.03). There was also a reduction of 28% in deaths from any cause in the zoledronic acid group (P = 0.01). No adverse effects on the healing of fractures were noted. The rates of renal and cardiovascular adverse events, including atrial fibrillation and stroke, were similar in the two groups.23 On the basis of this trial, the US Endocrine Society Practice guideline for osteoporosis in men suggests treatment with IV zoledronic acid in men with a recent hip fracture.24
Duration of therapy
The Fracture Intervention Trial Long-term Extension (FLEX) trial demonstrated a reduction in clinical (not morphometric) vertebral fractures among those who continued alendronate for 10 years, compared to those who discontinued after five years.25 A post hoc analysis26 revealed that among postmenopausal women without a vertebral fracture at FLEX baseline, an FN T-score of –2.5 or less at FLEX baseline were associated with non-vertebral fracture risk reduction (RR: 0.50, 95% CI: 0.26–0.96). A further post hoc analysis of the FLEX trial27 showed that among women who discontinued alendronate after five years, the predictors of fracture were age (HR: 1.54 [95% CI: 1.26–1.85] per five-year increase) and FN T-score (lowest tertile of baseline FN DXA versus other two tertiles relative HR: 2.17, 95% CI: 1.38–3.41). Change in BMD after one year was not a predictor of further fracture.27
In an extension of the HORIZON–PFT study, 1233 women who had received three annual doses of zoledronic acid in the original trial were randomised to receive either zoledronic acid for another three years under the same annual, three-dose regimen, or placebo.28 At 36 months follow-up, the incidence of new vertebral fractures was lower in women who received six years of zoledronic acid, compared with those who had received the drug for only three years (14 versus 30, odds ratio [OR]: 0.51, P = 0.035), but there was no change in the fracture rate in the placebo group.28 A further three-year extension of the trial did not show a significant difference in fracture rates between women taking zoledronic acid for a full nine years, compared with those who had taken the drug for six years followed by three years placebo.29 These results indicate that maximum benefit of zoledronic acid may be achieved in some patients after six years of therapy (for reduction of vertebral fracture risk), and that for most patients, benefits are maintained for a further three years once therapy is stopped.
Treatment of osteoporosis in men
One RCT30 found a significant reduction (P = 0.02) in the risk of vertebral fractures in older men with osteoporosis (n = 241) compared to placebo for alendronate 10 mg per day for two years. The effect on non-vertebral fractures was not significant. An RCT to assess the effectiveness of risedronate 5 mg daily versus vitamin D/calcium in men with osteoporosis (n = 316) with a baseline mean lumbar spine T-score of –3.3 and a prevalent vertebral fracture rate of 50%,31 found a significant 60% RR reduction (P = 0.028) in new morphometric vertebral fractures and statistically significant increases in lumbar spine and hip BMD at one year of follow-up. A placebo-controlled RCT of risedronate involving 284 men over two years32 with a baseline mean lumbar spine T-score of –3.2 and 25% of subjects with prevalent vertebral fractures demonstrated improved lumbar spine and hip BMD with risedronate. However, there was no significant effect on vertebral or non-vertebral fractures, although the study was underpowered to detect differences in fracture rates.32
Other data on the efficacy of zoledronic acid in reducing fracture risk in men are rare. In a multi-centre doubleblinded trial of 1199 men with osteoporosis 50 to 85 years of age, randomised to receive either placebo or 5 mg zoledronic acid at baseline and at 12 months, zoledronic acid reduced the rate of morphometric vertebral fracture by 67% at 24 months follow-up (RR: 0.33, 95% CI: 0.16–0.70, P = 0.002).33 The rate of non-vertebral fractures was also lower in the treatment group, but did not reach significance.33