Denosumab used for the treatment of osteoporosis is generally well tolerated. The subcutaneous mode of administration avoids the gastrointestinal side effects associated with oral bisphosphonates and ensures excellent bioavailability. RCT data indicate no significant increase in adverse events with long-term denosumab treatment, including infection, malignancy, pancreatitis, cardiovascular disease, peripheral vascular disease, MRONJ and AFFs.2,3 Cellulitis has been more frequently reported with denosumab compared with placebo, although the incidence remains low (<0.2 events per 100 subject-years for long-term denosumab).1 Hypocalcaemia following denosumab administration is a significant risk in patients with severe renal impairment (chronic kidney disease Stage 4 or 5) or in patients receiving dialysis.
The development of multiple vertebral fractures following discontinuation of denosumab therapy due to rebound bone resorption is now well recognised,9 especially in those with previous vertebral fractures.10 In the 1001 participants who discontinued denosumab during the FREEDOM study or its extension, the vertebral fracture rate increased from 1.2 per 100 participant-years during the on-treatment period to 7.1 per 100 participant-years following denosumab cessation.10 The odds of developing multiple vertebral fractures after stopping denosumab were 3.9 (95% CI: 2.1–7.2)-fold higher in those with than without prior vertebral fractures.10 The mechanism for this remains uncertain, but may be due to a pool of osteoclasts that are activated following loss of the inhibitory effect of denosumab.11
A delay in denosumab administration of more than six months was associated with increased risk of rebound fracture.12 Although definitive measures to minimise the risk of rebound vertebral fractures remain unclear at the time of writing, denosumab should either be continued long-term or cessation should be followed by another antiresorptive medication (i.e., 12 months of oral bisphosphonate or one or more infusions of zoledronate).13,14
Evidence Statement
The first randomised placebo-controlled trial of denosumab with fracture as a primary outcome was the FREEDOM trial, published in 2009.1 In that trial, 7668 women aged 60–90 years with a DXA T-score at the hip or spine of –2.5 to –4.0 were randomised to either 60 mg denosumab or placebo subcutaneously every six months for 36 months. Relative to placebo, denosumab reduced the risk of new radiographic vertebral fractures by 68% (cumulative incidence in treatment and placebo groups 2.3% and 7.2%, respectively; RR 0.32; 95% CI 0.26–0.41; P<0.001), hip fractures by 40% (cumulative incidence in treatment and placebo groups 0.7% and 1.2%, respectively; HR 0.60; 95% CI: 0.37–0.97; P=0.04) and non-vertebral, non-hip fractures by 20% (cumulative incidence in treatment and placebo groups 6.5% and 8.0%, respectively; HR 0.80; 95% CI: 0.67–0.95; P=0.01).1 The FREEDOM trial was extended for a further seven years (total trial length 10 years), and outcomes of the first two,2 three,3 four, and five4 years of the extension have been reported. The FREEDOM extension used a crossover design. Women who completed three years of denosumab treatment in the original trial were eligible to continue denosumab treatment, whereas those in the placebo group ‘crossed over’ to receive denosumab for the duration of the extension. After five years of the extension (1542 long-term subjects completing eight years of denosumab treatment and 1462 subjects crossing over to receive five years of denosumab treatment), the respective annual incidence of new vertebral fractures in long-term subjects was 1.5%, 1.3% and 1.3% during extension years 4–5, 6 and 7–8 and the respective annual incidence in crossover subjects was 0.9%, 1.6% and 1.8%.4 The annual incidence of non-vertebral fractures also remained low in both the long-term and crossover groups during the extension years, varying between 0.7% and 1.8%, and 1.2% and 2.6%, respectively. The cumulative incidence of hip fractures over the five-year extension was 0.7% in the long-term group and 1.1% in the crossover group (mean age 79 years at year 8 of extension).4
The two-year Denosumab Fracture Intervention Randomised Placebo Controlled Trial (DIRECT)15 measured fracture incidence with denosumab treatment versus placebo in Japanese men and women aged >50 years with 1–4 prevalent fractures and mean DXA T-scores of –2.8 at the lumbar spine and –2.0 at the hip. Over 24 months, the incidence of new or worsening vertebral fracture was 3.6% in the denosumab group, compared with 10.3% in the placebo group, a risk reduction of 65.7% (P=0.0001). Subgroup analysis of female subjects showed that the risk of new or worsening vertebral fracture at 24 months was reduced by 63.2% in the denosumab compared with placebo group (HR 0.37; 95% CI: 0.21–0.65; P=0.0004). The incidence of new vertebral fracture was reduced by 74% (P<0.0001). Subgroup analysis of male subjects showed a new or worsening vertebral fracture incidence at 24 months of 0% in denosumab-treated men, compared with 12.5% in men treated with placebo. However, this difference did not reach statistical significance (P=0.07) due to the small sample size (23 men in the denosumab arm and 24 in the placebo arm).15 A one-year crossover extension (n=775) of the DIRECT trial showed maintenance of low-fracture rates, with no difference in annualised fracture incidence between two and three years of treatment in the long-term group.16 As expected, the incidence of new and worsening vertebral fractures was reduced in the crossover group after commencement of denosumab treatment; the RaRs comparing years 2 and 1 and years 3 and 1 were 2.87 (P=0.003) and 0.23 (P=0.0003), respectively.16 These results suggested that the magnitude of effect on fracture risk reduction by denosumab depended on treatment duration. Overall, in men with low BMD treated with denosumab, increases in BMD were similar to those seen in postmenopausal women.5
Safety
The original three-year FREEDOM trial showed no significant increase in the incidence of cancer or infection compared with placebo.1 There was no increase in serious adverse events, including coronary heart disease and stroke, compared with placebo, but a significant increase in cellulitis requiring hospitalisation was reported (0.3% in the denosumab group compared with <0.1% in the placebo group; P=0.002). No cases of MRONJ or AFFs were reported. In the five-year extension study, adverse events for the duration of the FREEDOM extension, including cellulitis and other serious infection, were similar to those in the denosumab group in the original FREEDOM trial, with no increases over time.3 A total of two cases of AFF occurred in year 3 (in the crossover group) and year 7 (long-term group) of denosumab treatment, and a total of eight cases of MRONJ occurred in years 2 and 4 (in the crossover group) and years 6 and 7 (long-term group) of denosumab treatment. The cumulative incidence rates during the FREEDOM extension were 4.2 per 10,000 subject-years for MRONJ and 1.0 per 10,000 subject-years for AFF.4 Adverse event rates were similarly low in the two-year DIRECT trial and one-year DIRECT extension, with no significant difference between treatment and placebo groups.15,16 One case of MRONJ occurred during the extension in a crossover subject (one year of denosumab treatment).15,16 Although no head-to-head trials have been published, a systematic review of nine RCTs (n=4890) comparing the safety and efficacy of denosumab with bisphosphonate treatment for up to two years found no statistically significant difference between groups in terms of fracture risk or adverse events.17
Definitive measures to minimise the risk of rebound vertebral fractures remain unclear at the time of writing. To avoid this phenomenon, denosumab could be continued long term. However, if denosumab cessation is required, an observational study suggested that after 2–5 years of denosumab, 5 mg IV zoledronate administered six months following the last denosumab injection protected against the occurrence of multiple vertebral fractures.13 Zoledronate also appeared to retain 66% of the BMD gained with denosumab at the lumbar spine and 49% at the total hip.13 However, a recent randomised open-label study found that IV zoledronate did not fully prevent the increased bone turnover and bone loss observed following denosumab cessation, even when IV zoledronate was readministered following a rise in bone turnover markers or a fall in BMD.14 More evidence is required to guide therapy in this area.