Gonadotropin-Releasing Hormone Agonists and Fracture Risk: A Claims-Based Cohort Study of Men With Nonmetastatic Prostate Cancer
http://www.100md.com
《临床肿瘤学》
the Massachusetts General Hospital, Boston, MA
Abt Associates Clinical Trials, Bethesda, MD
Novartis Pharmaceutical Corp, East Hanover, NJ
ABSTRACT
PURPOSE: Gonadotropin-releasing hormone (GnRH) agonists decrease bone mineral density, a surrogate for fracture risk, in men with prostate cancer. We conducted a claims-based cohort study to characterize the relationship between GnRH agonists and risk for clinical fractures in men with nonmetastatic prostate cancer.
PATIENTS AND METHODS: Using medical claims data from a 5% national random sample of Medicare beneficiaries, we identified a study group of men with nonmetastatic prostate cancer who initiated GnRH agonist treatment from 1992 to 1994 (n = 3,887). A comparison group of men with nonmetastatic prostate cancer who did not receive GnRH agonist treatment during the study period (n = 7,774) was matched for age, race, geographic location, and comorbidity. Clinical fractures were identified using inpatient, outpatient, and physician claims during 7 years of follow-up.
RESULTS: In men with nonmetastatic prostate cancer, GnRH agonists significantly increased fracture risk. The rate of any clinical fracture was 7.88 per 100 person-years at risk in men receiving a GnRH agonist compared with 6.51 per 100 person-years in matched controls (relative risk, 1.21; 95% CI, 1.14 to 1.29; P < .001). Rates of vertebral fractures (relative risk, 1.45; 95% CI, 1.19 to 1.75; P < .001) and hip/femur fractures (relative risk, 1.30; 95% CI, 1.10 to 1.53; P = .002) were also significantly higher in men who received a GnRH agonist. GnRH agonist treatment independently predicted fracture risk in multivariate analyses. Longer duration of treatment conferred greater fracture risk.
CONCLUSION: GnRH agonists significantly increase risk for any clinical fracture, hip fractures, and vertebral fractures in men with prostate cancer.
INTRODUCTION
Prostate cancer is the most common noncutaneous malignancy and the second leading cause of cancer death among men in the United States. Approximately 230,110 new cases of prostate cancer were diagnosed in the United States in 2004.1 Gonadotropin-releasing hormone (GnRH) agonists are the mainstay of treatment for metastatic prostate cancer and a routine part of management for many men with nonmetastatic prostate cancer. Early primary androgen-deprivation therapy improves survival for men with locally advanced nonmetastatic prostate cancer.2 Adjuvant androgen-deprivation therapy improves survival for men with locally advanced prostate cancer treated with radiation therapy3 and men with lymph node–positive prostate cancer treated with radical prostatectomy and pelvic lymphadenectomy.4 In addition, GnRH agonists are commonly administered in clinical settings where the effects of GnRH treatment on disease-specific and overall survival are unknown. These clinical settings include primary therapy for low or intermediate risk disease and salvage therapy for men with an increasing prostate-specific antigen level as the only indication of disease recurrence after surgery or radiation therapy for early-stage prostate cancer.
Osteoporosis has emerged as an important adverse effect of GnRH agonist treatment in men with prostate cancer. Bone mineral density, a surrogate for fracture risk, decreases during both initial GnRH agonist therapy5-7 and long-term GnRH agonist therapy.8 Several uncontrolled retrospective studies have reported that clinical fractures are common in GnRH agonist–treated men with prostate cancer.9-12 Potential limitations of these studies include small sample size, short follow-up period, and ascertainment bias. We conducted a large, claims-based cohort study to characterize more precisely the relationship between GnRH agonist therapy and risk for clinical fracture in men with prostate cancer.
PATIENTS AND METHODS
Data Source
The study compared fracture rates in men with nonmetastatic prostate cancer who initiated GnRH agonist treatment with rates in matched (for age, race, and comorbidity) patients with nonmetastatic prostate cancer who did not initiate GnRH agonist treatment during the follow-up period. Men with prostate cancer were identified from Medicare Public Use File databases that comprise the medical care claims of a 5% national random sample of Medicare beneficiaries. Men were included in the study cohort if they had at least two medical claims between 1992 and 1994 with a diagnosis code for prostate cancer (codes 185.0, 233.4, or 236.0 from International Classification of Disease, ninth revision [ICD9]). From this sample, codes J9217, J9218, J9202, and J1950 were used to identify men who initiated GnRH agonist treatment during the years 1992 to 1994 (Fig 1). Men with bone metastases at baseline, history of orchiectomy, or treatment with a GnRH agonist before the study period were excluded.
A total of 51,612 men 65 years of age and older living in the 50 U.S. states and the District of Columbia were identified with at least two medical claims with a diagnosis code for prostate cancer. From this group of men with prostate cancer, 5,298 men were identified as having received GnRH agonist treatment between 1992 and 1994. The study group (n = 3,887) was identified after excluding men with any claim for GnRH agonist treatment, bone metastases, or orchiectomy within 1 year before the index date. From the group of 51,612 men with prostate cancer, a cohort of 41,250 men was identified with no claims for GnRH agonist treatment, bone metastases, or orchiectomy within 1 year before the index date or during the 7-year study period. From this cohort, a control group of 7,774 men were identified by matching with the study group in a 2:1 ratio for age, race, and comorbidity as assessed by the Charlson comorbidity index.13 Patients in both groups were longitudinally tracked until death or for 7 years after the index date. A patient was considered deceased if no medical claims were recorded for a 1-year period. Inclusion and exclusion criteria were determined before any data analyses.
A compendium of all pertinent ICD9 codes and Current Procedural Terminology (fourth edition) codes was created to assess the outcomes of bone-related complications according to the protocol described by Yao et al.14 The outcome of interest was any fracture. Fractures were identified based on the following fracture codes: ICD9 733.1 through 733.19, 22325 through 22327, 805, 806, and 820 through 829). Each patient was assigned no more than one fracture, even if multiple events were observed during the follow-up period.
The primary analyses of fracture incidence included all men in the study and control groups. Additional analyses were performed to evaluate potential effect(s) of incident bone metastases on fracture risk. In these analyses, men who developed bone metastases (identified via diagnosis codes 170.2, 170.6, 198.3, and 198.5) before a fracture during follow-up were excluded from the analyses. After excluding men with incident bone metastases, the total sample size decreased from 11,661 (study group, n = 3,887; control group, n = 7,774) to 10,617 (study group, n = 3,346; control group, n = 7,271).
Statistical Analyses
Person-year fracture incidence rates were calculated to account for varying follow-up duration and to decrease the potential bias associated with differences in median follow-up between the control and study groups. The numerator for the incidence rate was the number of fracture events in the population. The denominator was the sum of each individual's time at risk. Relative risk was defined as the incidence of fracture in the study group divided by the corresponding incidence in the control group.
Cox proportional hazards analyses were also used to evaluate fracture risk. Age, race, geographic location, cardiovascular disease, diabetes mellitus, GnRH agonist treatment, and duration of GnRH agonist treatment were included as covariates in the model. The duration of GnRH agonist treatment was specified as less than 1 year versus 1 year to distinguish between short-term neoadjuvant therapy and longer-term treatment. The multivariate analyses included all men in the study group (n = 3,887) and control group (n = 7,774). Men with incident bone metastases were censored at time of diagnosis with bone metastases.
The Kaplan-Meier method was used to estimate time to first fracture. In these analyses, men who either died or had any claim for bone metastases before fracture were censored from the analyses.
RESULTS
A total of 11,661 eligible patients were included in the analyses (study group, n = 3,887; comparison group, n = 7,774). Demographics are shown in Table 1. Thirty-five percent of the patients in the study group received less than 1 year of GnRH agonist treatment. The median follow-up was 4.7 years for the study group and 5.4 years for the comparison group.
GnRH agonist treatment was associated with significantly increased fracture risk (Table 2). Rates of any fracture were 7.88 per 100 person-years at risk in the study group and 6.51 per 100 person-years at risk in the control group (relative risk [RR], 1.21; 95% CI, 1.14 to 1.29; P < .001). Rates of vertebral fractures were 0.98 per 100 person-years in the study group and 0.68 per 100 person-years in the control group (RR, 1.45; 95% CI, 1.19 to 1.75; P < .001). Rates of hip/femur fractures were 1.26 per 100 person-years in the study group and 0.97 per 100 person-years in the control group (RR, 1.30; 95% CI, 1.10 to 1.53; P = .002)
Additional analyses were performed to control for the potential effect of incident bone metastases on fracture risk. In these analyses, all patients who developed bone metastases during the observation period were excluded. After controlling for incident bone metastases, GnRH agonist treatment was associated with increased fracture risk (Table 3). Rates of any fracture were 8.29 per 100 person-years at risk in the study group and 6.64 per 100 person-years at risk in the control group (RR, 1.25; 95% CI, 1.09 to 1.45; P < .001). Rates of vertebral fractures were 1.02 per 100 person-years in the study group and 0.68 per 100 person-years in the control group (RR, 1.50; 95% CI, 1.21 to 1.83; P < .001). Rates of hip/femur fractures were 1.33 per 100 person-years in the study group and 0.98 per 100 person-years in the control group (RR, 1.36; 95% CI, 1.14 to 1.62; P < .001).
In multivariate analyses, age, race, and GnRH agonist treatment were independently associated with greater fracture risk (Table 4). Men 81 to 85 years of age (hazard ratio, 1.42; 95% CI, 1.28 to 1.59; P < .0001) and men older than 85 years (hazard ratio, 1.48; 95% CI, 1.32 to 1.67; P < .0001) had significantly higher fracture risk than men 65 to 70 years of age (Table 2). Black men had significantly lower fracture risk than white men (hazard ratio, 0.64; 95% CI, 0.64 to 0.65; P < .0001). After controlling for other covariates, men treated with GnRH agonists had a significantly higher fracture risk than men who were not exposed to GnRH agonists (hazard ratio, 1.14; 95% CI, 1.07 to 1.23; P < .0001). The hazard ratio for GnRH agonist treatment for less than 1 year was 1.10 (95% CI, 0.97 to 1.21; P = .08). The hazard ratio for GnRH agonist treatment for 1 year was 1.16 (95% CI, 1.08 to 1.26; P < .0001).
The Kaplan-Meier method was used to estimate time to first fracture (Fig 1). The curve for men with GnRH agonist treatment less than 1 year diverged from that of the control group early but crossed the curve for the control group in the later follow-up period (log-rank P = .14). In contrast, the curve for men with GnRH agonist treatment 1 year diverged from that of the control group at approximately 1 year and remained separate throughout the remaining follow-up period (log-rank P = .002).
DISCUSSION
In this claims-based cohort study of men with nonmetastatic prostate cancer, GnRH agonist treatment increased risk of any fracture, hip fractures, and vertebral fractures. After controlling for age, race, geographic location, and comorbidity, GnRH agonist treatment independently predicted fracture risk.
Our results are consistent with earlier reports from uncontrolled retrospective studies that clinical fractures are common in GnRH agonist–treated men with prostate cancer.9-12 Our results are also consistent with a recently reported claims-based study using the linked database of Medicare and the Surveillance, Epidemiology, and End Results program.15
The relative risks of fracture observed in our study are similar to those reported for patients treated with other high-risk medications. For example, a recent study of patients treated with oral corticosteroids using the General Practice Research Database reported RRs for nonvertebral fractures of 1.33 (95% CI, 1.29 to 1.38), hip fractures of 1.61 (95% CI, 1.47 to 1.76), and vertebral fractures of 2.60 (95% CI, 2.31 to 2.92).16 These relative risks are similar to those observed in our study, where men receiving GnRH agonist treatment had relative risks of 1.25 (95% CI, 1.09 to 1.45) for any fracture and 1.36 (95% CI, 1.14 to 1.62) for hip fractures. The relative risk for vertebral fractures was somewhat lower at 1.50 (95% CI, 1.21 to 1.83).
This study was designed to provide conservative estimates of fracture risk in men with prostate cancer. We excluded men who had received GnRH agonist treatment before the study period. Men with incident GnRH agonist treatment were included in the study group regardless of treatment duration, and 35% of men received less than 1 year of GnRH agonist treatment. In addition, we assigned no more than one fracture to any patient. Accordingly, the study design may underestimate the effect of GnRH agonist on fracture risk. Other limitations of this claims-based analysis may contribute to an underestimation of fracture risk, including inability to ascertain a distant history of orchiectomies or GnRH agonist treatment and the inability to exclude patients receiving treatment of osteoporosis.
Some of the excess fracture risk associated with GnRH agonist treatment may be related to the development of bone metastases. In addition, potential differences in patterns of care between the groups may cause unintended bias in ascertainment of bone metastases. The consistency of results when patients with incident bone metastases were either included, excluded, or censored from the analyses, however, suggests that neither development of bone metastases nor any potential bias in ascertainment of bone metastases are sufficient to account for the observed significant association between GnRH agonists treatment and fracture risk. Moreover, most clinical fractures in men with prostate cancer do not directly result from bone metastases. In retrospective cohort studies, for example, between 9% and 16% of clinical fractures in men receiving hormone therapy for prostate cancer were classified as pathologic fractures.10,17
The risk of treatment-related fractures may have important implications for prostate cancer survivors. Fracture-related mortality is higher in men than in women.18,19 In persons older than 75 years, for example, the 1-year mortality after hip fracture is 20.7% for men compared with 7.5% of women.19 In addition, fractures are associated with decreased quality of life and shorter survival in men with prostate cancer.20,21
Treatment-related increases in fracture risk seem to be restricted to men with at least 1 year of GnRH agonist exposure. In multivariate analyses, men who received less than 1 year of GnRH agonist therapy had no significant increase in fracture risk compared with the men who never received a GnRH agonist. Additionally, time to first fracture was similar for the group who received less than 1 year of GnRH agonist treatment and the comparison group. This observation is biologically plausible because most men recover gonadal function after short-term GnRH agonist therapy,22 and rates of bone loss are attenuated after discontinuation of GnRH agonist treatment.23
There are no consensus guidelines for management of treatment-related osteoporosis in men receiving GnRH agonist treatment. The results of this study suggest that strategies to prevent fractures are warranted for men receiving long-term GnRH agonist treatment for prostate cancer. To date, however, no prospective studies have been reported with fracture end points in this setting. Several studies have reported that intravenous bisphosphonates preserve bone mineral density in men with prostate cancer. In two randomized controlled trials, pamidronate prevented bone loss in men treated with a GnRH agonist.24,25 In another randomized controlled trial, zoledronic acid significantly increased bone mineral density in men treated with a GnRH agonist or bilateral orchiectomy.26 Selective estrogen receptor modulators also preserve bone mineral density in men with prostate cancer treated with GnRH agonists.8,27 In recent reviews, most authors recommend routine screening for osteoporosis by measurement of bone mineral density, supplemental calcium and vitamin D to mitigate bone loss, and bisphosphonate treatment for men who either meet diagnostic criteria for osteoporosis or develop fractures.28,29 Large prospective studies are needed to assess the efficacy, safety, and cost effectiveness of interventions to prevent fractures in GnRH agonist–treated men.
In summary, GnRH agonists are associated with increased risk for any clinical fracture, hip fractures, and vertebral fractures in men with nonmetastatic prostate cancer. GnRH agonist treatment independently predicts fracture risk. Strategies to prevent treatment-related fractures in prostate cancer survivors may be warranted.
Appendix
The Appendix is included in the full-text version of this article, available online at www.JCO.org. It is not included in the PDF (via Adobe? Acrobat Reader?) version.
International Classification of Disease Codes Pathologic fracture (excludes traumatic)733.81-733.82Fracture of vertebral column without mention of spinal cord injury CategoryCodeDescription Prostate cancer185.xPrimary malignant neoplasm of prostate 233.4Carcinoma-in-situ, prostate 236.5Neoplasm, prostate, uncertain behavior Androgen-deprivation therapyJ9217Zoladex J9218Leuprolide acetate J9202Goserelin acetate J1950Leuprolide acetate/3.75 mg Bone metastasis170.2Neoplasm of vertebral column 170.6Neoplasm of pelvic bone and sacrum 198.3Secondary malignant neoplasm of brain and spinal cord 198.5Secondary malignant neoplasm of bone and bone marrow Pathologic fracture (excludes traumatic)733.10Pathologic fracture 733.11-14Pathologic fracture, humerus, radius, vertebra, and femur 733.19Pathologic fracture, other specified site Osteoporosis/osteopenia733.00Osteoporosis, unspecified 733.02Osteoporosis, idiopathic 733.09Osteoporosis, other drug-induced 733.90Disorder of bone and cartilage (osteopenia) Fractures (includes traumatic)800-804Fracture of skull 805.x-809.xFracture of neck and trunk 805.xFracture of vertebral column without mention of spinal cord injury 806.xFracture of vertebral column with spinal cord injury 810.x-819.xFracture of upper limb 820.x-829.xFracture of lower limb 830-839Dislocation 850.x-854.xIntracranial injury, excluding those with skull fracture 767.2Fracture of clavicle 767.3Other injuries to skeleton 767.4Injury to spine and spinal cord 733.81-733.82Malunion and nonunion of fracture 733.2Cyst of bone 733.4Aseptic neurosis of bone
Authors' Disclosures of Potential Conflicts of Interest
Although all authors completed the disclosure declaration, the following author or immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
NOTES
Supported by Novartis Pharmaceutical Corp (East Hanover, NJ) and awards from the Prostate Cancer Foundation (Santa Monica, CA) and the Claire and John Bertucci Center for Genitourinary Cancers (Boston, MA; M.R.S.).
Presented in part at the 40th Annual Meeting of the American Society of Clinical Oncology, June 5, 2004, New Orleans, LA.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Jemal A, Tiwari RC, Murray T, et al: Cancer statistics, 2004. CA Cancer J Clin 54:8-29, 2004
The Medical Research Council Prostate Cancer Working Party Investigators Group: Immediate versus deferred treatment for advanced prostatic cancer: Initial results of the Medical Research Council Trial. Br J Urol 79:235-246, 1997
Bolla M, Gonzalez D, Warde P, et al: Improved survival in patients with locally advanced prostate cancer treated with radiotherapy and goserelin. N Engl J Med 337:295-300, 1997
Messing EM, Manola J, Sarosdy M, et al: Immediate hormonal therapy compared with observation after radical prostatectomy and pelvic lymphadenectomy in men with node-positive prostate cancer. N Engl J Med 341:1781-1788, 1999
Eriksson S, Eriksson A, Stege R, et al: Bone mineral density in patients with prostatic cancer treated with orchidectomy and with estrogens. Calcif Tissue Int 57:97-99, 1995
Diamond T, Campbell J, Bryant C, et al: The effect of combined androgen blockade on bone turnover and bone mineral densities in men treated for prostate carcinoma: Longitudinal evaluation and response to intermittent cyclic etidronate therapy. Cancer 83:1561-1566, 1998
Maillefert JF, Sibilia J, Michel F, et al: Bone mineral density in men treated with synthetic gonadotropin-releasing hormone agonists for prostatic carcinoma. J Urol 161:1219-1222, 1999
Smith MR, Fallon MA, Lee H, et al: Raloxifene to prevent gonadotropin-releasing hormone agonist-induced bone loss in men with prostate cancer: A randomized controlled trial. J Clin Endocrinol Metab 89:3841-3846, 2004
Daniell HW: Osteoporosis after orchiectomy for prostate cancer. J Urol 157:439-444, 1997
Townsend MF, Sanders WH, Northway RO, et al: Bone fractures associated with luteinizing hormone-releasing hormone agonists used in the treatment of prostate carcinoma. Cancer 79:545-550, 1997
Hatano T, Oishi Y, Furuta A, et al: Incidence of bone fracture in patients receiving luteinizing hormone-releasing hormone agonists for prostate cancer. BJU Int 86:449-452, 2000
Oefelein MG, Ricchuiti V, Conrad W, et al: Skeletal fracture associated with androgen suppression induced osteoporosis: The clinical incidence and risk factors for patients with prostate cancer. J Urol 166:1724-1728, 2001
Charlson ME, Pompei P, Ales KL, et al: A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 40:373-383, 1987
Yao SL, Lu-Yao G: Population-based study of relationships between hospital volume of prostatectomies, patient outcomes, and length of hospital stay. J Natl Cancer Inst 91:1950-1956, 1999
Shahinian VB, Kuo YF, Freeman JL, et al: Risk of fracture after androgen deprivation for prostate cancer. N Engl J Med 352:154-164, 2005
Van Staa TP, Leufkens HG, Abenhaim L, et al: Use of oral corticosteroids and risk of fractures. J Bone Miner Res 15:993-1000, 2000
Melton LJ 3rd, Alothman KI, Khosla S, et al: Fracture risk following bilateral orchiectomy. J Urol 169:1747-1750, 2003
Diamond TH, Thornley SW, Sekel R, et al: Hip fracture in elderly men: Prognostic factors and outcomes. Med J Aust 167:412-415, 1997
Cooper C: The crippling consequences of fractures and their impact on quality of life. Am J Med 103:12S-17S, 1997 (suppl)
Weinfert K, Li Y, Castel L, et al: The impact of skeletal-related events on health-related quality of life of patients with metastatic prostate cancer. Ann Oncol 13:180, 2002 (suppl 5, abstr 662)
Oefelein MG, Ricchiuti V, Conrad W, et al: Skeletal fractures negatively correlate with overall survival in men with prostate cancer. J Urol 168:1005-1007, 2002
Nejat RJ, Rashid HH, Bagiella E, et al: A prospective analysis of time to normalization of serum testosterone after withdrawal of androgen deprivation therapy. J Urol 164:1891-1894, 2000
Higano C, Shields A, Wood N, et al: Bone mineral density in patients with prostate cancer without bone metastases treated with intermittent androgen suppression. Urology 64:1182-1186, 2004
Smith MR, McGovern FJ, Zietman AL, et al: Pamidronate to prevent bone loss in men receiving gonadotropin releasing hormone agonist therapy for prostate cancer. N Engl J Med 345:948-955, 2001
Diamond TH, Winters J, Smith A, et al: The antiosteoporotic efficacy of intravenous pamidronate in men with prostate carcinoma receiving combined androgen blockade: A double blind, randomized, placebo-controlled crossover study. Cancer 92:1444-1450, 2001
Smith MR, Eastham J, Gleason D, et al: Randomized controlled trial of zoledronic acid to prevent bone loss in men undergoing androgen deprivation therapy for nonmetastatic prostate cancer. J Urol 169:2008-2012, 2003
Steiner MS, Patterson A, Israeli R, et al: Toremifene citrate versus placebo for treatment of bone loss and other complications of androgen deprivation therapy in patients with prostate cancer. Proc Am Soc Clin Oncol 23:404, 2004 (abstr 4597)
Diamond TH, Higano CS, Smith MR, et al: Osteoporosis in men with prostate carcinoma receiving androgen-deprivation therapy: Recommendations for diagnosis and therapies. Cancer 100:892-899, 2004
Bae DC, Stein BS: The diagnosis and treatment of osteoporosis in men on androgen deprivation therapy for advanced carcinoma of the prostate. J Urol 172:2137-2144, 2004(Matthew R. Smith, Won Cha)
Abt Associates Clinical Trials, Bethesda, MD
Novartis Pharmaceutical Corp, East Hanover, NJ
ABSTRACT
PURPOSE: Gonadotropin-releasing hormone (GnRH) agonists decrease bone mineral density, a surrogate for fracture risk, in men with prostate cancer. We conducted a claims-based cohort study to characterize the relationship between GnRH agonists and risk for clinical fractures in men with nonmetastatic prostate cancer.
PATIENTS AND METHODS: Using medical claims data from a 5% national random sample of Medicare beneficiaries, we identified a study group of men with nonmetastatic prostate cancer who initiated GnRH agonist treatment from 1992 to 1994 (n = 3,887). A comparison group of men with nonmetastatic prostate cancer who did not receive GnRH agonist treatment during the study period (n = 7,774) was matched for age, race, geographic location, and comorbidity. Clinical fractures were identified using inpatient, outpatient, and physician claims during 7 years of follow-up.
RESULTS: In men with nonmetastatic prostate cancer, GnRH agonists significantly increased fracture risk. The rate of any clinical fracture was 7.88 per 100 person-years at risk in men receiving a GnRH agonist compared with 6.51 per 100 person-years in matched controls (relative risk, 1.21; 95% CI, 1.14 to 1.29; P < .001). Rates of vertebral fractures (relative risk, 1.45; 95% CI, 1.19 to 1.75; P < .001) and hip/femur fractures (relative risk, 1.30; 95% CI, 1.10 to 1.53; P = .002) were also significantly higher in men who received a GnRH agonist. GnRH agonist treatment independently predicted fracture risk in multivariate analyses. Longer duration of treatment conferred greater fracture risk.
CONCLUSION: GnRH agonists significantly increase risk for any clinical fracture, hip fractures, and vertebral fractures in men with prostate cancer.
INTRODUCTION
Prostate cancer is the most common noncutaneous malignancy and the second leading cause of cancer death among men in the United States. Approximately 230,110 new cases of prostate cancer were diagnosed in the United States in 2004.1 Gonadotropin-releasing hormone (GnRH) agonists are the mainstay of treatment for metastatic prostate cancer and a routine part of management for many men with nonmetastatic prostate cancer. Early primary androgen-deprivation therapy improves survival for men with locally advanced nonmetastatic prostate cancer.2 Adjuvant androgen-deprivation therapy improves survival for men with locally advanced prostate cancer treated with radiation therapy3 and men with lymph node–positive prostate cancer treated with radical prostatectomy and pelvic lymphadenectomy.4 In addition, GnRH agonists are commonly administered in clinical settings where the effects of GnRH treatment on disease-specific and overall survival are unknown. These clinical settings include primary therapy for low or intermediate risk disease and salvage therapy for men with an increasing prostate-specific antigen level as the only indication of disease recurrence after surgery or radiation therapy for early-stage prostate cancer.
Osteoporosis has emerged as an important adverse effect of GnRH agonist treatment in men with prostate cancer. Bone mineral density, a surrogate for fracture risk, decreases during both initial GnRH agonist therapy5-7 and long-term GnRH agonist therapy.8 Several uncontrolled retrospective studies have reported that clinical fractures are common in GnRH agonist–treated men with prostate cancer.9-12 Potential limitations of these studies include small sample size, short follow-up period, and ascertainment bias. We conducted a large, claims-based cohort study to characterize more precisely the relationship between GnRH agonist therapy and risk for clinical fracture in men with prostate cancer.
PATIENTS AND METHODS
Data Source
The study compared fracture rates in men with nonmetastatic prostate cancer who initiated GnRH agonist treatment with rates in matched (for age, race, and comorbidity) patients with nonmetastatic prostate cancer who did not initiate GnRH agonist treatment during the follow-up period. Men with prostate cancer were identified from Medicare Public Use File databases that comprise the medical care claims of a 5% national random sample of Medicare beneficiaries. Men were included in the study cohort if they had at least two medical claims between 1992 and 1994 with a diagnosis code for prostate cancer (codes 185.0, 233.4, or 236.0 from International Classification of Disease, ninth revision [ICD9]). From this sample, codes J9217, J9218, J9202, and J1950 were used to identify men who initiated GnRH agonist treatment during the years 1992 to 1994 (Fig 1). Men with bone metastases at baseline, history of orchiectomy, or treatment with a GnRH agonist before the study period were excluded.
A total of 51,612 men 65 years of age and older living in the 50 U.S. states and the District of Columbia were identified with at least two medical claims with a diagnosis code for prostate cancer. From this group of men with prostate cancer, 5,298 men were identified as having received GnRH agonist treatment between 1992 and 1994. The study group (n = 3,887) was identified after excluding men with any claim for GnRH agonist treatment, bone metastases, or orchiectomy within 1 year before the index date. From the group of 51,612 men with prostate cancer, a cohort of 41,250 men was identified with no claims for GnRH agonist treatment, bone metastases, or orchiectomy within 1 year before the index date or during the 7-year study period. From this cohort, a control group of 7,774 men were identified by matching with the study group in a 2:1 ratio for age, race, and comorbidity as assessed by the Charlson comorbidity index.13 Patients in both groups were longitudinally tracked until death or for 7 years after the index date. A patient was considered deceased if no medical claims were recorded for a 1-year period. Inclusion and exclusion criteria were determined before any data analyses.
A compendium of all pertinent ICD9 codes and Current Procedural Terminology (fourth edition) codes was created to assess the outcomes of bone-related complications according to the protocol described by Yao et al.14 The outcome of interest was any fracture. Fractures were identified based on the following fracture codes: ICD9 733.1 through 733.19, 22325 through 22327, 805, 806, and 820 through 829). Each patient was assigned no more than one fracture, even if multiple events were observed during the follow-up period.
The primary analyses of fracture incidence included all men in the study and control groups. Additional analyses were performed to evaluate potential effect(s) of incident bone metastases on fracture risk. In these analyses, men who developed bone metastases (identified via diagnosis codes 170.2, 170.6, 198.3, and 198.5) before a fracture during follow-up were excluded from the analyses. After excluding men with incident bone metastases, the total sample size decreased from 11,661 (study group, n = 3,887; control group, n = 7,774) to 10,617 (study group, n = 3,346; control group, n = 7,271).
Statistical Analyses
Person-year fracture incidence rates were calculated to account for varying follow-up duration and to decrease the potential bias associated with differences in median follow-up between the control and study groups. The numerator for the incidence rate was the number of fracture events in the population. The denominator was the sum of each individual's time at risk. Relative risk was defined as the incidence of fracture in the study group divided by the corresponding incidence in the control group.
Cox proportional hazards analyses were also used to evaluate fracture risk. Age, race, geographic location, cardiovascular disease, diabetes mellitus, GnRH agonist treatment, and duration of GnRH agonist treatment were included as covariates in the model. The duration of GnRH agonist treatment was specified as less than 1 year versus 1 year to distinguish between short-term neoadjuvant therapy and longer-term treatment. The multivariate analyses included all men in the study group (n = 3,887) and control group (n = 7,774). Men with incident bone metastases were censored at time of diagnosis with bone metastases.
The Kaplan-Meier method was used to estimate time to first fracture. In these analyses, men who either died or had any claim for bone metastases before fracture were censored from the analyses.
RESULTS
A total of 11,661 eligible patients were included in the analyses (study group, n = 3,887; comparison group, n = 7,774). Demographics are shown in Table 1. Thirty-five percent of the patients in the study group received less than 1 year of GnRH agonist treatment. The median follow-up was 4.7 years for the study group and 5.4 years for the comparison group.
GnRH agonist treatment was associated with significantly increased fracture risk (Table 2). Rates of any fracture were 7.88 per 100 person-years at risk in the study group and 6.51 per 100 person-years at risk in the control group (relative risk [RR], 1.21; 95% CI, 1.14 to 1.29; P < .001). Rates of vertebral fractures were 0.98 per 100 person-years in the study group and 0.68 per 100 person-years in the control group (RR, 1.45; 95% CI, 1.19 to 1.75; P < .001). Rates of hip/femur fractures were 1.26 per 100 person-years in the study group and 0.97 per 100 person-years in the control group (RR, 1.30; 95% CI, 1.10 to 1.53; P = .002)
Additional analyses were performed to control for the potential effect of incident bone metastases on fracture risk. In these analyses, all patients who developed bone metastases during the observation period were excluded. After controlling for incident bone metastases, GnRH agonist treatment was associated with increased fracture risk (Table 3). Rates of any fracture were 8.29 per 100 person-years at risk in the study group and 6.64 per 100 person-years at risk in the control group (RR, 1.25; 95% CI, 1.09 to 1.45; P < .001). Rates of vertebral fractures were 1.02 per 100 person-years in the study group and 0.68 per 100 person-years in the control group (RR, 1.50; 95% CI, 1.21 to 1.83; P < .001). Rates of hip/femur fractures were 1.33 per 100 person-years in the study group and 0.98 per 100 person-years in the control group (RR, 1.36; 95% CI, 1.14 to 1.62; P < .001).
In multivariate analyses, age, race, and GnRH agonist treatment were independently associated with greater fracture risk (Table 4). Men 81 to 85 years of age (hazard ratio, 1.42; 95% CI, 1.28 to 1.59; P < .0001) and men older than 85 years (hazard ratio, 1.48; 95% CI, 1.32 to 1.67; P < .0001) had significantly higher fracture risk than men 65 to 70 years of age (Table 2). Black men had significantly lower fracture risk than white men (hazard ratio, 0.64; 95% CI, 0.64 to 0.65; P < .0001). After controlling for other covariates, men treated with GnRH agonists had a significantly higher fracture risk than men who were not exposed to GnRH agonists (hazard ratio, 1.14; 95% CI, 1.07 to 1.23; P < .0001). The hazard ratio for GnRH agonist treatment for less than 1 year was 1.10 (95% CI, 0.97 to 1.21; P = .08). The hazard ratio for GnRH agonist treatment for 1 year was 1.16 (95% CI, 1.08 to 1.26; P < .0001).
The Kaplan-Meier method was used to estimate time to first fracture (Fig 1). The curve for men with GnRH agonist treatment less than 1 year diverged from that of the control group early but crossed the curve for the control group in the later follow-up period (log-rank P = .14). In contrast, the curve for men with GnRH agonist treatment 1 year diverged from that of the control group at approximately 1 year and remained separate throughout the remaining follow-up period (log-rank P = .002).
DISCUSSION
In this claims-based cohort study of men with nonmetastatic prostate cancer, GnRH agonist treatment increased risk of any fracture, hip fractures, and vertebral fractures. After controlling for age, race, geographic location, and comorbidity, GnRH agonist treatment independently predicted fracture risk.
Our results are consistent with earlier reports from uncontrolled retrospective studies that clinical fractures are common in GnRH agonist–treated men with prostate cancer.9-12 Our results are also consistent with a recently reported claims-based study using the linked database of Medicare and the Surveillance, Epidemiology, and End Results program.15
The relative risks of fracture observed in our study are similar to those reported for patients treated with other high-risk medications. For example, a recent study of patients treated with oral corticosteroids using the General Practice Research Database reported RRs for nonvertebral fractures of 1.33 (95% CI, 1.29 to 1.38), hip fractures of 1.61 (95% CI, 1.47 to 1.76), and vertebral fractures of 2.60 (95% CI, 2.31 to 2.92).16 These relative risks are similar to those observed in our study, where men receiving GnRH agonist treatment had relative risks of 1.25 (95% CI, 1.09 to 1.45) for any fracture and 1.36 (95% CI, 1.14 to 1.62) for hip fractures. The relative risk for vertebral fractures was somewhat lower at 1.50 (95% CI, 1.21 to 1.83).
This study was designed to provide conservative estimates of fracture risk in men with prostate cancer. We excluded men who had received GnRH agonist treatment before the study period. Men with incident GnRH agonist treatment were included in the study group regardless of treatment duration, and 35% of men received less than 1 year of GnRH agonist treatment. In addition, we assigned no more than one fracture to any patient. Accordingly, the study design may underestimate the effect of GnRH agonist on fracture risk. Other limitations of this claims-based analysis may contribute to an underestimation of fracture risk, including inability to ascertain a distant history of orchiectomies or GnRH agonist treatment and the inability to exclude patients receiving treatment of osteoporosis.
Some of the excess fracture risk associated with GnRH agonist treatment may be related to the development of bone metastases. In addition, potential differences in patterns of care between the groups may cause unintended bias in ascertainment of bone metastases. The consistency of results when patients with incident bone metastases were either included, excluded, or censored from the analyses, however, suggests that neither development of bone metastases nor any potential bias in ascertainment of bone metastases are sufficient to account for the observed significant association between GnRH agonists treatment and fracture risk. Moreover, most clinical fractures in men with prostate cancer do not directly result from bone metastases. In retrospective cohort studies, for example, between 9% and 16% of clinical fractures in men receiving hormone therapy for prostate cancer were classified as pathologic fractures.10,17
The risk of treatment-related fractures may have important implications for prostate cancer survivors. Fracture-related mortality is higher in men than in women.18,19 In persons older than 75 years, for example, the 1-year mortality after hip fracture is 20.7% for men compared with 7.5% of women.19 In addition, fractures are associated with decreased quality of life and shorter survival in men with prostate cancer.20,21
Treatment-related increases in fracture risk seem to be restricted to men with at least 1 year of GnRH agonist exposure. In multivariate analyses, men who received less than 1 year of GnRH agonist therapy had no significant increase in fracture risk compared with the men who never received a GnRH agonist. Additionally, time to first fracture was similar for the group who received less than 1 year of GnRH agonist treatment and the comparison group. This observation is biologically plausible because most men recover gonadal function after short-term GnRH agonist therapy,22 and rates of bone loss are attenuated after discontinuation of GnRH agonist treatment.23
There are no consensus guidelines for management of treatment-related osteoporosis in men receiving GnRH agonist treatment. The results of this study suggest that strategies to prevent fractures are warranted for men receiving long-term GnRH agonist treatment for prostate cancer. To date, however, no prospective studies have been reported with fracture end points in this setting. Several studies have reported that intravenous bisphosphonates preserve bone mineral density in men with prostate cancer. In two randomized controlled trials, pamidronate prevented bone loss in men treated with a GnRH agonist.24,25 In another randomized controlled trial, zoledronic acid significantly increased bone mineral density in men treated with a GnRH agonist or bilateral orchiectomy.26 Selective estrogen receptor modulators also preserve bone mineral density in men with prostate cancer treated with GnRH agonists.8,27 In recent reviews, most authors recommend routine screening for osteoporosis by measurement of bone mineral density, supplemental calcium and vitamin D to mitigate bone loss, and bisphosphonate treatment for men who either meet diagnostic criteria for osteoporosis or develop fractures.28,29 Large prospective studies are needed to assess the efficacy, safety, and cost effectiveness of interventions to prevent fractures in GnRH agonist–treated men.
In summary, GnRH agonists are associated with increased risk for any clinical fracture, hip fractures, and vertebral fractures in men with nonmetastatic prostate cancer. GnRH agonist treatment independently predicts fracture risk. Strategies to prevent treatment-related fractures in prostate cancer survivors may be warranted.
Appendix
The Appendix is included in the full-text version of this article, available online at www.JCO.org. It is not included in the PDF (via Adobe? Acrobat Reader?) version.
International Classification of Disease Codes Pathologic fracture (excludes traumatic)733.81-733.82Fracture of vertebral column without mention of spinal cord injury CategoryCodeDescription Prostate cancer185.xPrimary malignant neoplasm of prostate 233.4Carcinoma-in-situ, prostate 236.5Neoplasm, prostate, uncertain behavior Androgen-deprivation therapyJ9217Zoladex J9218Leuprolide acetate J9202Goserelin acetate J1950Leuprolide acetate/3.75 mg Bone metastasis170.2Neoplasm of vertebral column 170.6Neoplasm of pelvic bone and sacrum 198.3Secondary malignant neoplasm of brain and spinal cord 198.5Secondary malignant neoplasm of bone and bone marrow Pathologic fracture (excludes traumatic)733.10Pathologic fracture 733.11-14Pathologic fracture, humerus, radius, vertebra, and femur 733.19Pathologic fracture, other specified site Osteoporosis/osteopenia733.00Osteoporosis, unspecified 733.02Osteoporosis, idiopathic 733.09Osteoporosis, other drug-induced 733.90Disorder of bone and cartilage (osteopenia) Fractures (includes traumatic)800-804Fracture of skull 805.x-809.xFracture of neck and trunk 805.xFracture of vertebral column without mention of spinal cord injury 806.xFracture of vertebral column with spinal cord injury 810.x-819.xFracture of upper limb 820.x-829.xFracture of lower limb 830-839Dislocation 850.x-854.xIntracranial injury, excluding those with skull fracture 767.2Fracture of clavicle 767.3Other injuries to skeleton 767.4Injury to spine and spinal cord 733.81-733.82Malunion and nonunion of fracture 733.2Cyst of bone 733.4Aseptic neurosis of bone
Authors' Disclosures of Potential Conflicts of Interest
Although all authors completed the disclosure declaration, the following author or immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
NOTES
Supported by Novartis Pharmaceutical Corp (East Hanover, NJ) and awards from the Prostate Cancer Foundation (Santa Monica, CA) and the Claire and John Bertucci Center for Genitourinary Cancers (Boston, MA; M.R.S.).
Presented in part at the 40th Annual Meeting of the American Society of Clinical Oncology, June 5, 2004, New Orleans, LA.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Jemal A, Tiwari RC, Murray T, et al: Cancer statistics, 2004. CA Cancer J Clin 54:8-29, 2004
The Medical Research Council Prostate Cancer Working Party Investigators Group: Immediate versus deferred treatment for advanced prostatic cancer: Initial results of the Medical Research Council Trial. Br J Urol 79:235-246, 1997
Bolla M, Gonzalez D, Warde P, et al: Improved survival in patients with locally advanced prostate cancer treated with radiotherapy and goserelin. N Engl J Med 337:295-300, 1997
Messing EM, Manola J, Sarosdy M, et al: Immediate hormonal therapy compared with observation after radical prostatectomy and pelvic lymphadenectomy in men with node-positive prostate cancer. N Engl J Med 341:1781-1788, 1999
Eriksson S, Eriksson A, Stege R, et al: Bone mineral density in patients with prostatic cancer treated with orchidectomy and with estrogens. Calcif Tissue Int 57:97-99, 1995
Diamond T, Campbell J, Bryant C, et al: The effect of combined androgen blockade on bone turnover and bone mineral densities in men treated for prostate carcinoma: Longitudinal evaluation and response to intermittent cyclic etidronate therapy. Cancer 83:1561-1566, 1998
Maillefert JF, Sibilia J, Michel F, et al: Bone mineral density in men treated with synthetic gonadotropin-releasing hormone agonists for prostatic carcinoma. J Urol 161:1219-1222, 1999
Smith MR, Fallon MA, Lee H, et al: Raloxifene to prevent gonadotropin-releasing hormone agonist-induced bone loss in men with prostate cancer: A randomized controlled trial. J Clin Endocrinol Metab 89:3841-3846, 2004
Daniell HW: Osteoporosis after orchiectomy for prostate cancer. J Urol 157:439-444, 1997
Townsend MF, Sanders WH, Northway RO, et al: Bone fractures associated with luteinizing hormone-releasing hormone agonists used in the treatment of prostate carcinoma. Cancer 79:545-550, 1997
Hatano T, Oishi Y, Furuta A, et al: Incidence of bone fracture in patients receiving luteinizing hormone-releasing hormone agonists for prostate cancer. BJU Int 86:449-452, 2000
Oefelein MG, Ricchuiti V, Conrad W, et al: Skeletal fracture associated with androgen suppression induced osteoporosis: The clinical incidence and risk factors for patients with prostate cancer. J Urol 166:1724-1728, 2001
Charlson ME, Pompei P, Ales KL, et al: A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 40:373-383, 1987
Yao SL, Lu-Yao G: Population-based study of relationships between hospital volume of prostatectomies, patient outcomes, and length of hospital stay. J Natl Cancer Inst 91:1950-1956, 1999
Shahinian VB, Kuo YF, Freeman JL, et al: Risk of fracture after androgen deprivation for prostate cancer. N Engl J Med 352:154-164, 2005
Van Staa TP, Leufkens HG, Abenhaim L, et al: Use of oral corticosteroids and risk of fractures. J Bone Miner Res 15:993-1000, 2000
Melton LJ 3rd, Alothman KI, Khosla S, et al: Fracture risk following bilateral orchiectomy. J Urol 169:1747-1750, 2003
Diamond TH, Thornley SW, Sekel R, et al: Hip fracture in elderly men: Prognostic factors and outcomes. Med J Aust 167:412-415, 1997
Cooper C: The crippling consequences of fractures and their impact on quality of life. Am J Med 103:12S-17S, 1997 (suppl)
Weinfert K, Li Y, Castel L, et al: The impact of skeletal-related events on health-related quality of life of patients with metastatic prostate cancer. Ann Oncol 13:180, 2002 (suppl 5, abstr 662)
Oefelein MG, Ricchiuti V, Conrad W, et al: Skeletal fractures negatively correlate with overall survival in men with prostate cancer. J Urol 168:1005-1007, 2002
Nejat RJ, Rashid HH, Bagiella E, et al: A prospective analysis of time to normalization of serum testosterone after withdrawal of androgen deprivation therapy. J Urol 164:1891-1894, 2000
Higano C, Shields A, Wood N, et al: Bone mineral density in patients with prostate cancer without bone metastases treated with intermittent androgen suppression. Urology 64:1182-1186, 2004
Smith MR, McGovern FJ, Zietman AL, et al: Pamidronate to prevent bone loss in men receiving gonadotropin releasing hormone agonist therapy for prostate cancer. N Engl J Med 345:948-955, 2001
Diamond TH, Winters J, Smith A, et al: The antiosteoporotic efficacy of intravenous pamidronate in men with prostate carcinoma receiving combined androgen blockade: A double blind, randomized, placebo-controlled crossover study. Cancer 92:1444-1450, 2001
Smith MR, Eastham J, Gleason D, et al: Randomized controlled trial of zoledronic acid to prevent bone loss in men undergoing androgen deprivation therapy for nonmetastatic prostate cancer. J Urol 169:2008-2012, 2003
Steiner MS, Patterson A, Israeli R, et al: Toremifene citrate versus placebo for treatment of bone loss and other complications of androgen deprivation therapy in patients with prostate cancer. Proc Am Soc Clin Oncol 23:404, 2004 (abstr 4597)
Diamond TH, Higano CS, Smith MR, et al: Osteoporosis in men with prostate carcinoma receiving androgen-deprivation therapy: Recommendations for diagnosis and therapies. Cancer 100:892-899, 2004
Bae DC, Stein BS: The diagnosis and treatment of osteoporosis in men on androgen deprivation therapy for advanced carcinoma of the prostate. J Urol 172:2137-2144, 2004(Matthew R. Smith, Won Cha)