Physician practices of bone density testing and drug prescribing to prevent or treat osteoporosis during androgen deprivation therapy [An erratum to this Article has been published in Cancer 2006;106(11):2530] †
The opinions expressed herein do not necessarily reflect the views of the U.S. Department of Veterans Affairs.
Androgen deprivation therapy (ADT) is a strong risk factor for osteoporosis. The current study identified physician practices in preventing or treating osteoporosis during ADT. The practices of interest are the uses of dual-energy X-ray absorptiometry (DXA) scans, bisphosphonates, calcium or vitamin D supplement, calcitonin, or estrogen.
A retrospective medical record review was conducted. Patients were included if they had received ADT with goserelin injection for ≥ 1 year. Multivariable logistic regression analysis was performed to identify independent predictors of receiving at least one intervention.
Analyses included 184 patients. Most were the elderly with multiple risk factors for osteoporosis. Only 8.7% (95% confidence interval [CI], 4.6–13.0%) of patients received a DXA scan at least once during the past 3 years. Oral and intravenous bisphosphonates were prescribed in 4.9% (95%CI, 1.8–8.0%) and 0.5% (95%CI, 0–2.0%) of patients, respectively, during the past year. Overall, 14.7% of patients (95%CI, 9.5–20.0%) received at least one intervention. Concurrent risk factors for osteoporosis, including smoking, alcoholism, advanced age, low body mass index, long duration of ADT, multiple comorbidities, history of fractures, and steroid use, were not independent predictors of having received interventions. However, bone metastasis was, with a hazard ratio of 5.6 (95%CI, 1.99–15.6%). Primary care physicians provided the greatest number of interventions and cancer-related specialists provided the fewest.
Androgen deprivation therapy (ADT), a standard treatment for high-risk patients with prostate carcinoma after prostatectomy and for those with metastatic or recurrent disease, can lead to accelerated osteoporosis and fracture.1, 2 In general, osteoporosis is defined as a bone mineral density (BMD) measurement of ≥ 2.5 standard deviations below the young adult mean.3 Data from retrospective studies suggest that the incidence of osteoporotic fracture, typically hip or spinal fracture, ranges from 5% after 22 months of ADT to 40% after 15 years.4, 5 A longer duration of ADT is associated with a higher incidence of osteoporosis.6 In fact, many of these men already have osteoporosis before initiation of therapy.7 Osteoporosis fracture impairs quality of life, leads to nursing home placement, and increases the cost of care.8, 9 In addition, it is often fatal in older patients with multiple comorbid diseases.10
Several measures may prevent osteoporotic fracture during ADT although no established practice guidelines exist. Detection of osteoporosis is feasible by dual-energy X-ray absorptiometry (DXA) scan. The BMD at the femoral neck is particularly predictive of future hip fracture.11 Prospective studies with a short follow-up also suggest that intravenous bisphosphonates such as pamidronate and zoledronic acid may prevent bone loss associated with ADT.12, 13 In addition, alendronate, an oral bisphosphonate, has shown efficacy in the treatment of primary male osteoporosis.14 Moreover, calcium and vitamin D supplements maybe beneficial, especially for men with inadequate intake.15 Other potential pharmacologic interventions include calcitonin and estrogens. Finally, modification of life-styles that are not conducive to bone health can also be accomplished by behavioral counseling.
Despite the vast array of available interventions, prevention and treatment of osteoporosis are often inadequate. Examples include those among patients entering menopause, developing hip fracture, or undergoing long-term steroid therapy.16–21 In the current report, physician practices of ordering DXA scan and prescribing drugs to prevent or treat osteoporosis for patients undergoing ADT are investigated. Furthermore, other concurrent risk factors for osteoporosis, including age, smoking, alcoholism, medications, comorbid diseases, duration of ADT, body mass index, and history of previous fractures, were also obtained. More interventions may be prescribed in the presence of other concurrent risk factors for osteoporosis. The following analyses identify whether these concurrent risk factors independently predict receipt of at least one type of intervention to prevent or treat osteoporosis.
MATERIALS AND METHODS
After receiving approval from the institutional review board, the pharmacy registry was searched for patients who had received ADT with goserelin for ≥ 12 months before April 1, 2003 at a large, suburban teaching Veteran Administration hospital. Because goserelin injection was scheduled every 3 months, patients receiving < 3 injections were excluded, along with those who had received primary care outside of the hospital, became androgen independent, or received intermittent ADT. To reflect current practice, patients who discontinued ADT or were deceased before January 1, 2002 were also excluded.
Definition and Data Extraction
The date of the last visit for goserelin injection in each patient was used as the ending date of the observation period for that particular patient. Study variables were the use of the DXA scan during the past 3 years, as well as receipt of bisphosphonates, calcium or vitamin D supplements, calcitonin, or estrogen during the past year. Cancer-related specialists included oncologists, radiation oncologists, and urologists. Primary care physicians also included geriatricians and internists. Bone metastasis was confirmed by radiologic reports. The number of medications received was recorded at the last visit, but excluded topical agents and the medications considered to be study variables. The presence of hip and spinal fractures was confirmed by available radiologic reports involving the hip and spine conducted during the past 3 years. Visits were categorized into visits to nurses, primary care physicians, cancer-related specialists, cancer-unrelated specialists, and hospital admission. As part of institutional preventive programs, assessment of cigarette smoking and alcohol dependency using the CAGE questionnaire [“Have you ever felt you should cut down on your drinking?” “Have people annoyed you by criticizing your drinking?” “Have you ever felt bad or guilty about drinking?” “Have you ever taken a drink first thing in the moring (eye-opener) to steady your nerves or get rid of a hangover?”]22 was required and was documented routinely in the medical records yearly.
The main outcomes were the presence of the interventions of interest. Among patients who received these interventions, BMD results, if applicable, and specialty of the associated prescribing physicians were obtained. The chi-square test was used to detect a significant difference between proportions. The Fisher exact test was performed when necessary. P values < 0.0 were considered statistically significant. Hazard ratios were calculated with a 95% confidence interval (CI). To determine variables associated with the use of interventions, the dependent variables in the regression model were the interventions received. The independent variables were patient characteristics and other risk factors for osteoporosis or fractures. The relationships were examined in unadjusted models first. Variables with P < 0.25 were then examined using multivariable backward stepwise logistic models. Analyses were performed using SPSS, Version 10.1 (Chicago, IL).
Analyses included 184 patients (Table 1). Most were elderly Caucasians. The median duration of ADT was 32 months, but many patients had previously received ADT from other facilities. Patients often had multiple comorbid diseases including other cancers, stroke, arrhythmia, hypertension, diabetes mellitus, congestive heart failure, chronic renal insufficiency, ischemic heart disease, and chronic obstructive pulmonary disease. An average patient was taking seven medications at last visit and making nine visits to physicians yearly.
Approximately 15% of patients received ≥ 1 type of interventions (Table 2). The most common interventions were DXA scans and supplementation with calcium and vitamin D. DXA scans were ordered at the same rate as calcium and vitamin D supplementation, although not by the same physicians. Of the 16 patients who received DXA scans, 8 (50.0%) had osteoporosis and 4 (25.0%) had osteopenia (defined as a BMD t score between −1 and −2.5). Four patients who received DXA scans also had bone metastasis. The mean BMD t scores of the femoral neck and lumbar spine of patients who received DXA scans were −1.76 ± 1.52 and −0.12 ± 1.92, respectively.
Table 2. Use of Osteoporosis Prevention or Treatment
Percent (95% confidence interval)
DXA: dual-energy X-ray absorptiometry.
Receive DXA scan at least once in the past 3 yrs
Receive calcium supplement in the past year
Receive vitamin D supplement in the past year
Receive oral bisphosphonate in the past year
Receive intravenous bisphosphonate in the past year
Receive estrogen or calcitonin in the past year
Any of the above
A few patients received an oral bisphosphonate and one patient received pamidronate. No calcitonin or estrogen was prescribed. The most commonly prescribed drug was the combined calcium carbonate and vitamin D tablet. Of the 10 patients who received bisphosphonates, 6 also received a DXA scan, which showed osteoporosis in 5 patients and osteopenia in 1 patient.
Risk Factors for Osteoporosis and the Receipt of Interventions
Concurrent risk factors for osteoporosis were prevalent, but they were not independent predictors of receiving interventions (Table 3). The only predictor, however, was the presence of bone metastasis. Among 19 patients who had bone metastasis, 8 (42.1%) received ≥ 1 intervention: 5 received a DXA scan, 3 received calcium, 3 received bisphosphonates, and 2 received vitamin D. The most frequent fractures were compression fractures of the spine. Although a few patients currently smoked cigarettes, most were previous smokers.
Table 3. Risk Factors for Osteoporosis or Fractures and Interventions
Hazard ratio for receiving any intervention
N/A: not available; ADT: androgen deprivation therapy; BMI: body mass index; CI: confidence interval.
No patient in the category received an intervention.
Statistically significant (P < 0.001).
BMI was calculated by weight in kilograms divided by the square of height in meters.
Thirty-three physicians provided 58 interventions to 27 patients (Fig. 1). In the current study, 51.5% of the physicians were primary care physicians (internal medicine, 42.4%; geriatrics, 9.1%), 27.3% of the physicians had cancer-unrelated specialties (endocrinology, 12.1%; nephrology, 6.1%; cardiology, 6.1%; pulmonology, 6.1%), and 21.2% had cancer-related specialties (medical oncology, 18.2%; radiation oncology, 3%). Overall, primary care physicians provided the greatest number of interventions and cancer-related specialists provided the fewest.
The majority of patients undergoing ADT did not receive osteoporosis prevention. Many of them had additional risk factors for osteoporosis or fracture, particularly advanced age and multiple comorbid diseases. The presence of these additional risk factors, however, was not predictive of having received osteoporosis prevention. Although cancer-related specialists prescribed ADT, they prescribed the fewest interventions.
What leads to the underuse of osteoporosis prevention? Perhaps, it is due to the lack of established guidelines. Currently, the optimal evaluation methods and appropriate interventions to prevent osteoporosis-associated fractures for men undergoing ADT remain unclear. Although practice guidelines for screening and prevention of osteoporosis associated with postmenopausal women, gastrointestinal disease, and chronic steroid therapy are available,23–25 none have been established for men undergoing ADT.
In addition, because osteoporosis is a multidisciplinary problem, uncertainty regarding who should prescribe interventions may arise among multiple specialists who are involved in patient care. Primary care physicians have been the main providers of osteoporosis prevention and treatment. However, cancer-related specialists should also be responsible for therapy-related complications. In fact, among cancer-related specialists themselves, there may be a confusion regarding who should be responsible for prescribing interventions.
It is worrisome that other concurrent risk factors for osteoporosis did not lead to increased osteoporosis prevention. Theoretically, these patients have higher risks of osteoporotic fracture and should be deemed a priority, although ADT alone already may have markedly elevated the risk. The presence of bone metastasis, although associated with increased preventive measures, was, in fact, addressed inadequately by the current standard of care. During the study period, however, zoledronic acid had not yet been indicated for such patients to prevent skeletal complications.26
The limitations of the current study involve the definition of outcome variables, which represents an intention of the physician, rather than an effective means to prevent osteoporosis. In fact, calcium and vitamin D alone are unlikely to be adequate to prevent osteoporosis during ADT.12, 13 In addition, calcium, vitamin D, or bisphosphonates could have been prescibed for other clinical purposes with no intention to prevent or treat osteoporosis. Moreover, DXA scans may become unreliable among patients with osteoblastic metastasis or degenerative joint disease because osteophytes, especially in the spine, may falsely elevate the BMD if included in the measurement.27 In the current study, however, only 4 of the 16 patients who received DXA scans had bone metastasis.
A great opportunity to improve osteoporosis prevention among men receiving ADT exists. Their frequent visits, especially nursing visits for goserelin injection, though reflect many missed opportunity, can be used for future interventions. Established practice guidelines for BMD measurement and pharmacologic intervention for patients undergoing ADT are necessary. Ideally, interventions should be safe, effective, and convenient for a typical patient population, e.g., the elderly with multiple chronic illnesses who take several medications.