Dispelling confusion about de‐prescribing bisphosphonates

Stopping bisphosphonate therapy in people at risk of bone fractures must be carefully considered

The major rationale for starting anti‐resorptive drug treatment in people with osteoporosis is that the risk of further fractures is 2–4 times and that of mortality 2–3 times as great for someone with a fragility fracture as for other people of the same age. This applies to virtually all fractures, and particularly during the 5 years after a fragility fracture.1,2,3

In people who have had an osteoporotic fracture, anti‐resorptive agents reduce the risk of subsequent fractures by about 50%, and in some studies they also improve survival.4,5,6,7 In most people at risk of fragility fractures, bone mineral density is not in the osteoporotic range (defined as a T‐score of –2.5 or less; that is, at least 2.5 standard deviations [SDs] below the mean density for a 30‐year‐old woman). Although the risk of fracture increases with declining bone density, most fractures occur in individuals with bone mineral density in the osteopenic range (T‐score between –2.5 and –1.0) because of the larger number of people in this range. Some individuals at risk of fractures but with bone mineral density values above the osteoporotic range may benefit from treatment before they have had a low trauma fracture. A number of risk factors can indicate increased risk of fracture and deserve investigation when osteoporosis is suspected, including measurement of bone mineral density and evaluation of biochemical parameters (Box).

Any fracture in a young healthy person that occurs under circumstances in which a fracture would not be expected should be carefully investigated as a potential low trauma fracture, including assessment of bone density.

Why stop bisphosphonate therapy?

In cases of chronic disease, specific treatments are withdrawn if they are ineffective, there is a better alternative, the condition is cured or in remission, the risk exceeds the benefit, or, after consulting the patient, a palliative approach is deemed more appropriate. These considerations rarely apply when treating people with osteoporosis.

The idea of stopping bisphosphonate therapy was motivated by the results of an extension study of a randomised controlled trial which suggested that the benefits of 5 years’ continuous treatment persisted for 5 years after ceasing therapy.7 Even for the potent bisphosphonates, this interpretation is not accurate. There is some persistence of benefit, but the effect rapidly declines and the risks of different types of fracture again increase. Increased risk has been reported after stopping treatment with alendronate and zoledronic acid,8,9,10 and the increase is possibly more rapid after withdrawing risedronate, which, after 7 years’ treatment, leads to bone turnover returning to pre‐treatment levels within 6 months.11 Vertebral fractures are the most frequent type after ceasing therapy; they are potentially disabling, but have responded better to therapy in all published randomised controlled trials.

The desire to stop treatment is driven by anxiety about potential side effects of anti‐resorptive therapy, including osteonecrosis of the jaws and atypical femur fractures. The frequency of these events is very low, probably fewer than one event per 10 000 person‐years, and concerns about these adverse events have been exaggerated. Although osteoporotic fractures are associated with increased mortality risk, this was not found to apply to atypical femur fractures in a large population‐based study.12 However, one should also note that the risk of fracture after stopping treatment is very low, at least during the first 2 years, when bone density is above the osteoporosis threshold T‐score of –2.5. As the risk of atypical femur fractures decreases by about 70% for each year after a patient ceases anti‐resorptive therapy,13 one could accept a 2‐year break from therapy for an individual whose fracture risk is low; that is, for someone whose T‐score is better than –1.5, who have not had any recent fractures, and who have no important risk factors, such as ongoing glucocorticoid or sex hormone ablative therapies.

Not all anti‐resorptive treatments, however, share the same mechanism of action or pharmacokinetic characteristics. Denosumab, for example, is not a bisphosphonate; it is a monoclonal antibody that blocks the activity of RANKL (the receptor activator of nuclear factor‐κB ligand), preventing the activation of osteoclasts. If a denosumab dose is missed, there is a rebound return to rapid rates of bone loss.14,15,16 Multiple vertebral fractures after discontinuing denosumab have been reported,17 although the absolute risk has not yet been established. Accordingly, denosumab should probably not be started if long term adherence is potentially a problem.

Monitoring patients after ceasing bisphosphonate therapy

There are no guidelines for monitoring patients after ending a period of bisphosphonate therapy. Length of time since discontinuing treatment is probably more appropriate than measuring bone mineral density for estimating risk. In one follow‐up study, bone density at the completion of alendronate therapy and time since its cessation were the best predictors of fracture risk.8 Markers of bone turnover, although used by some osteoporosis specialists, have not been found to be reliable predictors of risk.8

Some doctors advocate drug holidays from anti‐resorptive therapy, but the term itself is misleading, as in most cases the duration of any break in therapy is not defined in advance. Should a patient sustain significant early bone loss sufficient to return their T‐score to the osteoporotic range, the break in anti‐resorptive therapy should certainly be ended and treatment reinstated.

Summary

Explaining long term therapy to people with osteoporosis is straightforward. Firstly, the risk–benefit balance of anti‐osteoporotic therapy with an effective anti‐resorptive agent is clear for a patient who has had a low trauma fracture and whose bone density is in the osteopenic or osteoporotic ranges. Intravenous zoledronic acid at 18‐month intervals reduces fracture risk in women with osteopenia, even if they have not experienced fragility fractures.18 While justifiable as part of healthy lifestyle recommendations, adequate calcium, vitamin D, and protein intake, physical activity, and reducing the risk of falls have not been found to reduce the risk of fracture.19

The number needed to treat depends largely on the individual's underlying risk of fracture. The greater the individual's underlying risk of fracture, the greater the potential benefit of treatment. The risk of bone fracture can be estimated, for example, with the calculator provided by the Garvan Institute (www.garvan.org.au/bone-fracture-risk). The risk of osteonecrosis of the jaw or atypical femur fracture is estimated to be 1–10 cases per 10 000 person‐years of treatment.13 For individuals at moderate to high risk of fragility fractures, the benefit‐to‐risk ratio (estimated to be 50–100:1) is clearly in favour of treatment to avert fractures.

Secondly, temporarily ceasing bisphosphonate therapy is sometimes advocated, but is justifiable only for patients judged to be at low risk of fracture on the basis of clinical factors, the absence of recent fractures, and bone density above the osteoporosis threshold. It is justifiable even for these patients only if they have been treated with bisphosphonates, as the effect of these agents on bone persists to some degree for 1–2 years. Royal Australian College of General Practitioners osteoporosis guidelines advise reconsidering the need for bisphosphonate therapy every 5–10 years.20 In our opinion, this advice is valid only for patients at low risk. A break in therapy is contraindicated for people receiving denosumab; the accumulated bone density benefit of even years of treatment is lost within 6–12 months during the rapid decline after missing a dose. Even potent anti‐resorptive bisphosphonates can ameliorate but not prevent this loss.21

Thirdly, re‐instituting therapy (with anti‐resorptive or possibly bone‐forming agents) after a hiatus of 1–2 years is appropriate if clinical characteristics and bone density measurements indicate that a patient is at significant risk of fracture.