Association between androgen deprivation therapy and fracture risk:  A population-based cohort study in men with non-metastatic prostate cancer.

Reviewer: Charles Wood, MD
Abramson Cancer Center of the University of Pennsylvania
Last Modified: June 5, 2004

Presenter: M.R. Smith
Presenter's Affiliation: Massachusetts General Hospital
Type of Session: Scientific


  • Several small retrospective studies have reported that GnRH agonists increase fracture risk in prostate patients.  However, the studies have been limited by small sample size, short follow-up, and a potential recall bias.
  • This large-scale population-based cohort study was undertaken to more accurately characterize the link between GnRH agonist use and clinical fractures.

Materials and Methods


  • Utilizing a 5% national random sample of claims submitted by Medicare beneficiaries between 1991 and 2001, a study group of 3887 men with nonmetastatic prostate cancer who initiated use of GnRH agonist therapy between 1992 and 1994 was identified.  A comparison group of 7774 men with non-metastatic prostate cancer who did not receive GnRH therapy during the study period was identified and matched in a 1:2 ratio based on age, race, and Charlson co-morbidity index.
  • Neither group had undergone prior orchiectomy or had evidence of bone metastases.  The study group had no history of prior GnRH agonist use, and the comparison group did not have GnRH agonist therapy during the study period. 
  • Clinical fractures sustained between 1994 and 2001 were identified.


  • There was a significantly higher yearly incidence per patient of overall fractures in the GnRH agonist group (8.29% vs. 6.64%, RR=1.25).
  • There was a significantly higher yearly incidence per patient of hip or femur fractures in the GnRH agonist group (1.53% vs. 1.05%, RR=1.46).
  • There was a significantly higher yearly incidence per patient of vertebral fractures in the GnRH agonist group (1.19% vs. 0.73%, RR=1.63).
  • Longer duration of GnRH agonist use (>3 years) was associated with a significantly higher risk of fracture versus shorter duration (<1 year).
  • Overall, men in the study group were more likely to develop fractures versus men in the comparison group (RR=1.4).


Author's Conclusions

  • Non-metastatic prostate patients undergoing GnRH agonist therapy are more likely to sustain a clinical fracture, and longer duration of use denotes increased risk of fracture.
  • The magnitude in the increased risk is similar to other high risk medications (eg. corticosteroids).
  • GnRH agonist therapy should be considered a high risk medication.

Clinical/Scientific Implications
        This study highlights the significant side effect profile associated with GnRH agonist use in non-metastatic prostate cancer.  The methods used in patient selection and follow-up, however, were not able to identify use of bisphosponate therapy between the two groups, and intermittent versus continuous use of GnRH agonists in the study group was likewise unknown.  Nonetheless, physicians should consider the side effect profile of GnRH agonists and potentially take measures to prevent the inherent skeletal events via such options as bisphosphonate therapy and intermittent dosing of GnRH agonists. 

Oncolink's ASCO Coverage made possible by an unrestricted Educational Grant from Bristol-Myers Squibb Oncology.