Denosumab for osteoporosis, October 2020

Page last updated: 16 March 2021

Drug utilisation sub-committee (DUSC)

October 2020



Naik-Panvelkar et al. (2020) used MedicineInsight data to examine osteoporosis management in general practice, reporting that 80% of patients who ceased denosumab treatment did not appear to have any record of subsequent bisphosphonate prescription.[1] This is a concern because there is evidence of lower bone mineral density (BMD) and increased risk of multiple vertebral fractures shortly after discontinuation of denosumab. For this reason, it is suggested that denosumab not be discontinued without considering a substitute treatment.[2]

At its June 2020 meeting, DUSC noted that the findings of Naik-Panvelkar et al. suggest that Australian patients who cease denosumab are at risk of fractures. DUSC requested that the utilisation of denosumab, including use of bisphosphonates or other osteoporosis medicines after denosumab discontinuation, be reviewed in the context of the total osteoporosis market using both PBS dispensing data and MedicineInsight data.

The osteoporosis medicines included in this analysis are:

  • denosumab;
  • bisphosphonates alone and in combination with colecalciferol and/or calcium carbonate (alendronate, risedronate and zoledronic acid[3]);
  • raloxifene and teriparatide;
  • strontium ranelate (delisted in 2016); and
  • calcitriol.

Date of first listing on the Pharmaceutical Benefits Scheme (PBS) for the treatment of osteoporosis

  • Calcitriol: 1 December 1991
  • Disodium etidronate and calcium carbonate: 1 August 1996 (delisted September 2012)
  • Alendronate: 1 November 1996
  • Raloxifene: 1 November 1999
  • Risedronate: 1 February 2001
  • Risedronate and calcium carbonate: 1 April 2006
  • Alendronate with colecalciferol: 1 August 2006
  • Strontium: 1 April 2007 (delisted 1 August 2016)
  • Risedronate and calcium carbonate with colecalciferol (1 May 2008)
  • Zoledronic acid: 1 December 2008
  • Teriparatide: 1 May 2009
  • Alendronate with colecalciferol and calcium carbonate: 1 June 2010
  • Denosumab: 1 December 2010

Data Source / methodology

PBS Data

Patient counts and patient level analysis data were extracted from the Services Australia PBS 10% sample database for prescriptions supplied from January 2012 to June 2019. Prescription analyses were based on prescriptions supplied from January 2014 to June 2019.

MedicineInsight Data

This paper reports on a retrospective cohort study using MedicineInsight data to explore the use of osteoporosis medicines in general practice, with a particular interest in the use of denosumab. It uses de-identified patient data from the clinical information system (CIS) of 506 participating general practices and provides information on the following patient cohorts:

  • 896,548 regularly attending patients aged 50+ years who visited their general practice at least 3 times between January 2018 and December 2019 (regular patient prevalence study population)
  • 120,388 patients aged 50+ years who visited their general practice at least 3 times between January 2014 and December 2017 and had a recorded diagnosis of osteoporosis prior to 1 January 2018 (historical osteoporosis study population);
  • 22,256 patients who were started on denosumab between January 2014 and December 2017 regardless of whether they had been prescribed another osteoporosis medicine first (denosumab initiator population); and
  • 11,122 patients who were started directly on denosumab between January 2014 and December 2017 without having been prescribed another osteoporosis medicine first (initiated directly on denosumab population).

Key Findings

PBS Data

  • In 2014, 32,277 patients in the 10% PBS sample were dispensed an osteoporosis medicine at least once. In 2018, this number had risen to 49,451 patients – a 50% increase.
  • Patterns of dispensing between 2014 and 2018 changed with denosumab accounting for 76.1% of the patients dispensed any osteoporosis medicines in 2018 compared with 36.5% of patients in 2014. Over the same period, the bisphosphonates decreased from being dispensed to 46.5% of all patients dispensed an osteoporosis medicine in 2014 to 18.6% in 2014.
  • Initiations to denosumab have driven an increase in the overall number of patients initiating osteoporosis therapy from 7,202 patients initiating in 2014 to 9,514 patients in 2018, representing a 32% increase (PBS 10% sample).
  • There has been a 2.8-fold increase in the rate of prevalent prescribing of denosumab. In 2014, the prevalent rate of prescribing was 4.5 per 1,000 persons. In 2018, the prevalent rate was 12.5 per 1,000 persons.
  • Females aged 70+ years accounted for almost two-thirds of all patients dispensed denosumab in the 10% PBS sample between January 2014 and June 2019.
  • However, between 2014 and 2018 the number of male prevalent denosumab patients increased 4.6-fold compared with 3.0-fold for women and the number of male denosumab initiators increased by 60% compared with 40% for women.
  • The number of all patients newly started on denosumab in the 10% PBS sample each year has stabilised in recent years. However, the number of patients who start on denosumab despite not having used any osteoporosis medicine in the prior two years is still increasing. This suggests that denosumab is being used more and more frequently as a first-line therapy for osteoporosis.
  • The proportion of patients who were directly initiated on denosumab (without previous osteoporosis therapy) increased from 55% in 2014 to 76% in 2018. However, in the MedicineInsight study, among all patients who had been prescribed denosumab at least once between 2014 and 2017, 50% were directly initiated on denosumab. The lower proportion of patients directly initiated on denosumab in the MedicineInsight study compared with the PBS 10% study could indicate that specialists are more likely to start patients directly on denosumab than GPs. The PBS 10% study may have misclassified (and overestimated) patients as new to denosumab therapy if they had previously used other osteoporosis drugs prior to 2012 which could partly explain the magnitude of the difference between the data sources.
  • The median average duration on any initiating osteoporosis medicine was 2.1 years, including breaks or 1.8 years excluding treatment breaks.
  • The longest duration of therapy was seen for patients who initiate denosumab with a median average treatment duration (including breaks) of 2.5 years. In the subset of patients with at least 3.5 years of follow-up available (sensitivity analysis), the median average treatment duration, including breaks, was 3.9 years. The MedicineInsight data reported a 3.2-year median treatment duration (including breaks).
  • Patients who initiated zoledronic acid had a median average duration of therapy of 1.0 years including or excluding treatment breaks. Alendronate and risedronate had similar treatment durations, at 1.1 years including breaks and 0.7 years excluding breaks. Of the 11,219 patients who experienced a treatment break from the very first osteoporosis medicine they were dispensed, patients on zoledronic acid and denosumab had the longest median time to first treatment break, at approximately 1 year, reflecting one supply of zoledronic acid and two supplies of denosumab.
  • Of the 11,533 of the patients who were started on denosumab (regardless of whether they had been dispensed a different osteoporosis medicine beforehand), 35.2% had a treatment break. During this break only 2.5% were covered by another prescription for osteoporosis therapy.
  • Just over a third (33.8%) of patients directly initiated on denosumab (without being dispensed a different osteoporosis medicine beforehand) had ceased therapy by the end of the study period and in the MedicineInsight study almost a quarter of patients (24.6%) had ceased denosumab treatment by study end.
  • Of the PBS patients who ceased denosumab only 5.0% had a subsequent record of osteoporosis treatment and of the MedicineInsight patients who ceased denosumab only 13.8% had a subsequent record of osteoporosis treatment. Patients who were directly initiated on bisphosphonates were more likely to cease therapy than patients initiated on denosumab. They were also more likely to start treatment with an alternative osteoporosis medicine (probably denosumab) after ceasing their original medicine.

MedicineInsight Data

  • The prevalence of osteoporosis among regularly attending MedicineInsight patients, (aged 50+ years, for all analyses), was 13.6% (95% confidence interval [CI] 13.0 to 14.1). As expected, the patient prevalence of osteoporosis was higher in women and rose with increasing age in both sexes.
  • 60.5% (95% CI 59.5 to 61.4) of all regularly attending patients with a record of osteoporosis had been prescribed an osteoporosis medicine at least once.
  • Just over half (52.7%; 95% CI 51.2 to 54.1) of regularly attending men with recorded osteoporosis had been prescribed an osteoporosis medicine at least once, compared with 62.5% (95% CI 61.5 to 63.5) of regularly attending women, suggesting that men may still be less likely to be treated for osteoporosis than women even when they have a recorded diagnosis.
  • The proportion of regularly attending patients who had no record of being prescribed an osteoporosis medicine was higher in this study (39.5%; 95% CI 38.6% to 40.5%) than that reported in the Naik-Panvelkar paper (23.5%). This is most likely due to differences in cohort selection, as the cohort in the Naik-Panvelkar paper was restricted to patients seen at the general practice at least once every year over an 8-year period.
  • Among the 22,256 patients in the historical osteoporosis cohort who had been prescribed denosumab at least once, 50.0% were started on denosumab without any record of having been previously prescribed any other osteoporosis medicine.
  • The median duration of treatment of denosumab treatment (including treatment holidays) was 1,154 days or 3.2 years (interquartile range: 887 days [1st quartile] to 1,679 days [3rd quartile]).
  • 66.0% (95% CI 64.8 to 67.1) of the patients started on denosumab between January 2014 and December 2017 were still on denosumab treatment on 31 December 2019. Almost a quarter of patients (24.6%; 95% CI 23.6 to 25.6) had ceased denosumab treatment by this date.
  • 86.2% of the historical osteoporosis cohort had no record of osteoporosis therapy after denosumab cessation. This is higher than that reported by Naik-Panvelkar et al. (2020) but may be explained by the differences in the study cohorts.


[1] Naik-Panvelkar P, Norman, S, Elgebaly, Z, et al. Osteoporosis management in Australian general practice: an analysis of current osteoporosis treatment patterns and gaps in practice. BMC Fam Pract 2020; 21(1): 32.

[2] AMH. Australian Medicines Handbook. Adelaide: Australian Medicines Handbook Pty Ltd; 2020.

[3] A fourth bisphosphonate, etidronate, has not been included as it was removed from the PBS in 2012 (prior to the study period) and is no longer available in Australia.

Full Report