Abstract
Purpose
Uncertainty exists about the ideal timing for initiation of androgen deprivation therapy for men with metastatic prostate cancer. We assessed factors associated with early or delayed androgen deprivation therapy among men diagnosed with metastatic prostate cancer and assessed the relationship between this therapy and overall survival.
Patients and methods
We studied a population-based cohort of American men aged ≥66 years diagnosed with metastatic prostate cancer during 1992–2002 and followed through 2003. We assessed receipt of androgen deprivation therapy early (within 4 months of diagnosis), delayed (after 4 months), or not at all, used multinomial logistic regression to identify factors associated with treatment, and used Cox proportional hazards models to assess whether treatment was associated with survival.
Results
Overall, 69.5% of men received early hormonal therapy and 7.3% delayed. Adjusted rates of early therapy were lower for black vs. white men (58.3% vs. 71.0%) and adjusted rates of delayed use were higher for black vs. white men (12.7% vs. 6.2%). Receipt of androgen deprivation therapy was associated with improved survival (adjusted HR=.69 95% CI .66–.73). The benefit for early treatment did not differ from delayed treatment (P=.58)
Conclusions
A large minority of men with metastatic prostate cancer, particularly black men, receive delayed or no hormonal therapy. Receipt of early or delayed hormonal therapy was associated with similarly prolonged survival. After controlling for patient and tumor characteristics, survival did not differ by race, and receipt of hormonal therapy did not contribute to racial differences in survival.
Keywords: prostate cancer, androgen deprivation therapy, disparities, outcomes
Prostate cancer is the most frequently diagnosed cancer among men, with over 234,000 estimated new cases diagnosed in the United States in 2006.1 Although only a minority of prostate cancers are metastatic at the time of diagnosis, survival for such men is poor, with only one-third expected to be alive in 5 years.2
Androgen deprivation therapy with a gonadotropin-releasing hormone (GnRH) agonist or bilateral orchiectomy has long been known to inhibit growth of prostate cancers in men with advanced prostate cancer,3, 4 but uncertainty exists about the ideal timing for initiation of androgen deprivation in men with metastatic disease.5, 6 Some data suggest fewer disease-related complications compared to treatment delayed until clinical progression, despite similar survival.7 However, a recent review of published studies of hormone therapy in metastatic or progressive prostate cancer suggested that immediate versus symptom-onset initiation of hormone therapy results in a moderate decrease in risk of prostate cancer survival, a moderate increase in the risk of non-prostate cancer survival, and no overall survival advantage.6
Use of androgen deprivation therapy has increased greatly since the early 1990s among American men with nonmetastatic prostate cancer,8–10 but few data are available about patterns of androgen deprivation therapy in men with metastatic prostate cancer. A recent study of men who died of prostate cancer during 1991–2000 found that approximately one-quarter of men did not receive androgen deprivation therapy before death, and black men were much less likely than white men to receive this treatment.11 We studied a population-based cohort of men diagnosed with metastatic prostate cancer to evaluate use of androgen deprivation therapy among men with advanced prostate cancer in the U.S.. Specifically, we assessed whether men were treated with a GnRH agonist or orchiectomy early (within 4 months of diagnosis), delayed (after 4 months), or not at all and whether receipt of this treatment varied by patient or tumor characteristics or year of diagnosis. We also assessed whether treatment with androgen deprivation therapy was associated with improved survival for men diagnosed with metastatic prostate cancer.
Patients and methods
Data
We used Surveillance, Epidemiology, and End Results (SEER)-Medicare data for this analysis.12 The SEER program of the National Cancer Institute is an authoritative source of information on cancer incidence and survival in the United States. For years, SEER registrars have collected uniformly reported data from 11 population-based cancer registries covering approximately 14% of the U.S. population.13 In 2000, SEER expanded to include 4 additional registries (Kentucky, Louisiana, New Jersey, and Greater California); the SEER registries now represent approximately 26% of the U.S. population. Since 1991, for each incident cancer, SEER data on patient demographics, tumor characteristics, and primary treatment are linked with Medicare administrative data (successfully linking more than 94% of SEER patients diagnosed at age ≥65).14 Medicare is the national health insurance program for Americans aged 65 and older and administrative data can be used to identify services received by patients.
Study cohort
We identified men aged ≥66 years with a first diagnosis of prostate cancer during 1992–1999 who were continuously enrolled in Parts A and B of fee-for-service Medicare as of 1 year before diagnosis (N=150,378). We excluded 3,716 men with no administrative claims from 45 days before diagnosis through 195 days after diagnosis because we considered their claims incomplete. From the remaining 146,662 men, we identified the 8,671 men with metastatic disease at diagnosis.
Androgen deprivation therapy
We used claims data to ascertain receipt of androgen deprivation therapy, including GnRH agonists and bilateral orchiectomy (Appendix), early (within 4 months of diagnosis), delayed (after 4 months from diagnosis), or not at all. Duration of treatment with GnRH agonists was calculated by summing the number of 1-month equivalent doses; for men who underwent orchiectomy, we summed the number of months from the time of surgery onward.
Appendix.
Codes used to define GnRH agonist use and orchiectomy
| Diagnosis or Procedure | HCPCS | CPT | ICD-9 Procedure |
|---|---|---|---|
| Leuprolide injection | J9217, J9218, J9219, J1950 | ||
| Goserelin injection | J9202 | ||
| Triptorelin | J3315 | ||
| Orchiectomy | 54520, 54521, 54522, 54530, 54535, 54690, 49510 | 62.3, 62.4, 62.41, 62.42 |
Survival
Men were observed from the date of diagnosis until December 31, 2003, or sooner if they died or disenrolled from parts A and B of fee-for-service Medicare. We ascertained date of death from Medicare enrollment data.
Patient and tumor characteristics
We documented each man’s age, race, Hispanic ethnicity, marital status, urban residence, SEER region, year of diagnosis, tumor grade, census-tract level income (categorized in quartiles within registries), and comorbid illness based on Diagnostic Cost Groups (DCGs),15 a risk-adjustment tool for predicting future costs and disease burden for Medicare beneficiaries based on 182 clinical diagnostic categories from inpatient and ambulatory claims (compared to 18 for the Charlson score16). DCGs are highly predictive of mortality for Medicare beneficiaries with cancer17 and myocardial infarction.18 We calculated summary DCG scores (categorized in quartiles; excluding prostate cancer codes) based on the 12-months before diagnosis.
Analyses
We used Chi-square tests to assess bivariate associations of patient characteristics with receipt of GnRH agonist or orchiectomy early, delayed, or not at all. We used multinomial logistic regression models to assess characteristics associated with receiving early vs. delayed vs. no androgen deprivation therapy. Independent variables included the patient and tumor characteristics described above (categorized as in Table 1). We calculated adjusted rates of diagnosis at each stage for our patient population using a standardized regression approach.19
Table 1.
Patient characteristics associated with receiving androgen deprivation, unadjusted
| N | % receiving GnRH agonist or orchiectomy within 4 months | % receiving GnRH agonist or orchiectomy after 4 months | % not receiving GnRh agonist or orchiectomy | P value | |
|---|---|---|---|---|---|
| Total | 8,671 | 69 | 7 | 23 | |
| Age in years | <.001 | ||||
| 66–69 | 1,149 | 71 | 9 | 20 | |
| 70–74 | 1,900 | 73 | 8 | 19 | |
| 75–79 | 2,134 | 73 | 7 | 19 | |
| 80–84 | 1,879 | 71 | 6 | 23 | |
| 85 and older | 1,609 | 58 | 6 | 36 | |
| Race | <.001 | ||||
| White | 6,911 | 72 | 6 | 22 | |
| Black | 1,245 | 54 | 12 | 34 | |
| Other | 473 | 73 | 8 | 19 | |
| Unknown | 42 | 81 | 7 | 12 | |
| Hispanic ethnicity | .38 | ||||
| No | 8,170 | 69 | 7 | 23 | |
| Yes | 429 | 69 | 9 | 22 | |
| Unknown | 72 | 76 | 7 | 17 | |
| Marital status | <.001 | ||||
| Unmarried | 3,078 | 64 | 7 | 29 | |
| Married | 5,174 | 73 | 8 | 20 | |
| Unknown | 419 | 71 | 7 | 23 | |
| Residence | <.001 | ||||
| Major metropolitan area | 4,644 | 66 | 8 | 25 | |
| Metropolitan county | 2,341 | 73 | 6 | 21 | |
| Urban | 636 | 73 | 7 | 20 | |
| Less urban | 863 | 72 | 6 | 22 | |
| Rural | 187 | 71 | 7 | 22 | |
| SEER region | <.001 | ||||
| California | 2,331 | 71 | 7 | 23 | |
| Connecticut | 977 | 74 | 6 | 20 | |
| Detroit | 1,193 | 57 | 10 | 33 | |
| Hawaii | 295 | 74 | 7 | 19 | |
| Iowa | 1,261 | 74 | 6 | 19 | |
| New Mexico | 350 | 67 | 8 | 25 | |
| Seattle | 704 | 70 | 8 | 21 | |
| Utah | 374 | 72 | 6 | 21 | |
| Atlanta | 334 | 66 | 7 | 27 | |
| Rural Georgia | 41 | 66 | 0 | 34 | |
| Kentucky | 204 | 74 | 3 | 23 | |
| Louisiana | 223 | 65 | 5 | 30 | |
| New Jersey | 384 | 73 | 7 | 20 | |
| Median household income in Census tract of residence | <.001 | ||||
| Quartile 1 (lowest) | 2,100 | 65 | 8 | 27 | |
| Quartile 2 | 2,125 | 69 | 8 | 23 | |
| Quartile 3 | 2,124 | 71 | 6 | 22 | |
| Quartile 4 (high) | 2,140 | 72 | 7 | 21 | |
| Unknown | 182 | 72 | 7 | 21 | |
| Tumor grade | <.001 | ||||
| Well differentiated | 165 | 67 | 7 | 26 | |
| Moderately differentiated | 2,221 | 73 | 10 | 17 | |
| Poorly differentiated/undifferentiated | 3,846 | 79 | 7 | 15 | |
| Unknown | 2,439 | 52 | 6 | 42 | |
| DCG comorbidity | <.001 | ||||
| Quartile 1 (lowest) | 2,223 | 71 | 9 | 20 | |
| Quartile 2 | 2,152 | 73 | 7 | 21 | |
| Quartile 3 | 2,130 | 70 | 7 | 23 | |
| Quartile 4 (high) | 2,166 | 64 | 7 | 29 | |
| Year of diagnosis | .005 | ||||
| 1992 | 1,244 | 71 | 7 | 21 | |
| 1993 | 967 | 68 | 8 | 24 | |
| 1994 | 808 | 71 | 7 | 23 | |
| 1995 | 734 | 70 | 8 | 22 | |
| 1996 | 611 | 65 | 9 | 26 | |
| 1997 | 599 | 68 | 5 | 27 | |
| 1998 | 562 | 67 | 8 | 25 | |
| 1999 | 510 | 67 | 11 | 22 | |
| 2000 | 876 | 71 | 7 | 22 | |
| 2001 | 871 | 71 | 6 | 24 | |
| 2002 | 889 | 72 | 5 | 23 |
Percentages may not add to 100% due to rounding.
We next used Cox proportional hazards regression models to assess the association between patient and tumor characteristics among men diagnosed with metastatic prostate cancer, including the same covariates as above. We then added a time-varying treatment variable reflecting receipt of androgen deprivationto assess the association between this therapy and overall survival and to test if this explained any differences in survival based on patient characteristics. The time-varying treatment variables allowed men to contribute information to the control group when not on treatment and to the treatment groups (GnRH agonist or orchiectomy) when on treatment (nearly all men were treated continuously once treatment began). Men treated with GnRH agonists were considered continuously treated for 6 months after each dose of GnRH agonist (because hypogonadism may persist for prolonged periods after discontinuation of a GnRH agonist).20, 21
We next tested for a survival benefit from early vs. late hormonal therapy using a term for the interaction of hormone therapy and an indicator variable for whether the patient was treated with early vs. late hormonal therapy(if a patient had not yet received any therapy then to ensure that only information from the past was used for predictions in the next time period, both the early and the late hormone therapy variables were coded as “none”).
In sensitivity analyses, we restricted the cohort to the 7,234 men who survived and were continuously enrolled in fee-for-service Medicare Parts A and B for at least 4 months to be sure the cohort survived long enough to obtain outpatient care for their prostate cancer, and thus have the opportunity to be treated with androgen deprivation therapy. All tests of statistical significance were two-sided. We conducted analyses using SAS statistical software, version 8.2 (SAS Institute, Inc., Cary, North Carolina). Multinomial regression models were conducted using Stata, version 9 (StataCorp LP, College Station, Texas). The Harvard Medical School Committee on Human Studies approved the study protocol.
Results
The mean age of the cohort was 77.9 (SD=7.0) years. Most men (60%) were married at the time of diagnosis, 14% were black, 5% were of Hispanic ethnicity, and 44% had poorly differentiated or undifferentiated cancers.
Receipt of androgen deprivation
Overall, 69.5% of men with metastatic prostate cancer at diagnosis were treated with a GnRH agonist or orchiectomy within 4 months of diagnosis (early treatment), 7.3% received their first dose of GnRH agonist or orchiectomy after 4 months from diagnosis (delayed treatment), and 23.3% were not treated with androgen deprivation therapy before death. Men treated with early GnRH agonist therapy were treated for a mean (SD) of 20.8 (20.7) months, similar to the duration of treatment for men who received delayed treatment (mean (SD)=20.7 (24.2) months).
Unadjusted rates of early, late, or no GnRH agonist treatment or orchiectomy by patient and tumor characteristics are presented in Table 1. Rates of early hormonal therapy were lowest among men aged ≥85 compared to men of other ages, black men (compared to non-black men), and unmarried men (compared to married men). Black men had higher rates of delayed hormonal therapy than others. The median number of days to first hormonal therapy was 23 days for white men (mean (SD)=64 (203) days) and 28 days for black men (mean (SD)=103 (241) days). Rates of hormonal therapy varied by SEER region, and rates of early use were higher for men in the highest area-level income quartiles (compared to lower quartiles). Men with less-well differentiated cancers were more likely to be treated with androgen deprivation therapy early. Unadjusted rates of treatment varied over time without a clear pattern.
In adjusted analyses (Table 2), race was most strongly related to receipt of androgen deprivation therapy, with black men substantially less likely than white men to receive early androgen deprivation therapy (adjusted rate=58.3% vs. 71.0% of white men), and more likely to receive delayed hormonal therapy (12.7% vs. 6.2%) or no hormonal therapy (29.0% vs. 22.4%). Of those patients that did not receive early androgen deprivation therapy, 21.0% of whites and 30% of blacks subsequently received delayed hormonal therapy. Men aged 85 and older were less likely than other men to be treated with hormonal therapy, as were men who were unmarried (compared to married men). Rates of hormonal therapy varied by SEER region, with highest rates in New Jersey, Connecticut, and Georgia registries and lowest rates in Detroit. Men with poorly-or undifferentiated tumors were more likely to be treated with androgen deprivation therapy than men with well or moderately differentiated tumors and men with moderate levels of comorbid illnesses were also more likely than those with the fewest or the most comorbid illnesses to receive this therapy. Rates of use of androgen deprivation therapy varied slightly over time after adjustment for patient and tumor characteristics, although this difference was not statistically significant (P=.054). We found no association between use of androgen deprivation therapy and year of diagnosis when we considered this as a continuous variable (P=.70).
Table 2.
Adjusted rates of receiving early, delayed, or no androgen deprivation
| Adjusted % receiving androgen deprivation therapy within 4 months | Adjusted % receiving androgen deprivation therapy after 4 months | Adjusted % not receiving androgen deprivation therapy | P value | |
|---|---|---|---|---|
| Age in years | <.001 | |||
| 66–69 | 71.1 | 8.6 | 20.3 | |
| 70–74 | 72.5 | 7.1 | 20.4 | |
| 75–79 | 72.2 | 7.3 | 20.5 | |
| 80–84 | 69.5 | 6.9 | 23.6 | |
| 85 and older | 61.6 | 6.4 | 31.9 | |
| Race | <.001 | |||
| White | 71.3 | 6.3 | 22.4 | |
| Black | 58.3 | 12.7 | 29.0 | |
| Other | 70.6 | 8.2 | 21.1 | |
| Unknown | 76.4 | 6.3 | 17.3 | |
| Hispanic ethnicity | .34 | |||
| No | 69.6 | 7.0 | 23.4 | |
| Yes | 66.3 | 9.9 | 23.8 | |
| Unknown | 71.0 | 9.4 | 19.5 | |
| Marital status | <.001 | |||
| Unmarried | 70.8 | 6.1 | 23.1 | |
| Married | 71.3 | 7.6 | 21.1 | |
| Unknown | 66.3 | 6.8 | 26.9 | |
| Residence | .18 | |||
| Major metropolitan area | 67.4 | 7.6 | 25.0 | |
| Metropolitan county | 71.7 | 6.3 | 22.0 | |
| Urban | 73.0 | 7.2 | 19.8 | |
| Less urban | 71.9 | 6.7 | 21.4 | |
| Rural | 69.2 | 9.1 | 21.7 | |
| SEER region | <.001 | |||
| California | 71.5 | 6.4 | 22.1 | |
| Connecticut | 73.1 | 7.3 | 19.7 | |
| Detroit | 61.7 | 8.6 | 29.7 | |
| Hawaii | 66.9 | 7.5 | 25.6 | |
| Iowa | 69.9 | 7.6 | 22.5 | |
| New Mexico | 64.6 | 8.2 | 27.2 | |
| Seattle | 70.4 | 8.7 | 20.9 | |
| Utah | 67.5 | 7.9 | 24.7 | |
| Atlanta | 72.2 | 5.4 | 22.4 | |
| Rural Georgia | 76.4 | 1.0 | 23.6 | |
| Kentucky | 71.1 | 4.3 | 24.6 | |
| Louisiana | 65.9 | 5.3 | 28.9 | |
| New Jersey | 74.1 | 7.6 | 18.3 | |
| Median household income in Census tract of residence | <.09 | |||
| Quartile 1 (lowest) | 67.9 | 6.7 | 25.4 | |
| Quartile 2 | 68.9 | 7.8 | 23.3 | |
| Quartile 3 | 70.0 | 6.5 | 23.5 | |
| Quartile 4 (highest) | 70.6 | 8.0 | 21.4 | |
| Unknown | 71.8 | 6.9 | 21.3 | |
| Tumor grade | <.001 | |||
| Well differentiated | 67.4 | 6.5 | 26.0 | |
| Moderately differentiated | 72.2 | 9.6 | 18.2 | |
| Poorly differentiated/undifferentiated | 78.5 | 6.3 | 15.2 | |
| Unknown | 53.4 | 6.6 | 40.0 | |
| DCG comorbidity | <.001 | |||
| Quartile 1 (lowest) | 65.8 | 8.1 | 26.1 | |
| Quartile 2 | 71.6 | 6.8 | 21.6 | |
| Quartile 3 | 71.8 | 7.1 | 21.1 | |
| Quartile 4 (high) | 68.5 | 6.6 | 24.9 | |
| Year of diagnosis | .054 | |||
| 1992 | 70.1 | 7.0 | 22.9 | |
| 1993 | 67.6 | 7.9 | 24.6 | |
| 1994 | 70.4 | 6.6 | 22.9 | |
| 1995 | 69.7 | 7.6 | 22.7 | |
| 1996 | 65.9 | 8.5 | 25.6 | |
| 1997 | 68.3 | 5.0 | 26.7 | |
| 1998 | 68.7 | 7.7 | 23.5 | |
| 1999 | 67.7 | 10.7 | 21.6 | |
| 2000 | 71.2 | 7.5 | 21.3 | |
| 2001 | 69.9 | 6.2 | 23.9 | |
| 2002 | 71.9 | 5.6 | 22.5 |
Using multinomial logistic regression to control for all variables in the table.
Survival
When we examined survival among men diagnosed with metastatic prostate cancer, we found that men who were older (vs. younger), unmarried (vs. married), and men with higher comorbidity scores had worse survival (Table 3). Men in the highest income quartile had better survival then men in the lowest quartile. Survival also varied by tumor grade as well as region and residence in a major metropolitan area vs. other areas. We found no association between black (vs. white) race and survival after controlling for patient demographic and clinical characteristics (Table 3). Men diagnosed more recently were at higher risk of death (HR for each year=1.01, 95% CI 1.00 to 1.02).
Table 3.
Adjusted hazard ratio of death after diagnosis of metastatic prostate cancer
| Adjusted HR (95% CI) of dying during follow-up in model not including androgen deprivation variable | Adjusted HR (95% CI) of dying during follow-up in model including androgen deprivation variable | |
|---|---|---|
| Androgen deprivation therapy | - | .69 (.66 to .73) |
| Age in years | ||
| 66–69 | Reference | Reference |
| 70–74 | 1.09 (.99 to 1.19) | 1.09 (.99 to 1.19) |
| 75–79 | 1.28 (1.17 to 1.46) | 1.3 (1.19 to 1.42) |
| 80–84 | 1.47 (1.34 to 1.62) | 1.50 (1.36 to 1.65) |
| 85 and older | 1.76 (1.60 to 1.94) | 1.75 (1.59 to 1.93) |
| Race | ||
| White | Reference | Reference |
| Black | .99 (.91 to 1.07) | .95 (.87 to 1.03) |
| Other | .81 (.71 to .94) | .82 (.71 to .94) |
| Unknown | .52 (.29 to .94) | .51 (.28 to .93) |
| Hispanic ethnicity | ||
| No | Reference | Reference |
| Yes | 1.00 (.88 to 1.13) | .98 (.86 to 1.10) |
| Unknown | .80 (.53 to 1.20) | .80 (.53 to 1.19) |
| Marital status | ||
| Unmarried | Reference | Reference |
| Married | .87 (.83 to .92) | .88 (.84 to .93) |
| Unknown | .74 (.65 to .84) | .73 (.64 to .83) |
| Residence | ||
| Major metropolitan area | Reference | Reference |
| Metropolitan county | .92 (.84 to 1.00) | .93 (.85 to 1.02) |
| Urban | .85 (.75 to .95) | .86 (.76 to .97) |
| Less urban | .81 (.72 to .92) | .83 (.73 to .94) |
| Rural | .94 (.78 to 1.15) | .96 (.79 to 1.16) |
| SEER region | ||
| California | Reference | Reference |
| Connecticut | 1.24 (1.12 to 1.36) | 1.24 (1.13 to 1.37) |
| Detroit | 1.21 (1.11 to 1.31) | 1.17 (1.07 to 1.27) |
| Hawaii | .94 (.77 to 1.13) | .91 (.75 to 1.10) |
| Iowa | 1.33 (1.19 to 1.50) | 1.33 (1.19 to 1.49) |
| New Mexico | 1.37 (1.18 to 1.60) | 1.36 (1.16 to 1.58) |
| Seattle | 1.14 (1.03 to 1.26) | 1.14 (1.03 to 1.26) |
| Utah | 1.13 (.97 to 1.30) | 1.12 (.97 to 1.30) |
| Atlanta | 1.32 (1.15 to 1.50) | 1.34 (1.18 to 1.53) |
| Rural Georgia | 1.36 (.96 to 1.93) | 1.30 (.92 to 1.85) |
| Kentucky | 1.33 (1.08 to 1.63) | 1.36 (1.10 to 1.66) |
| Louisiana | 1.32 (1.08 to 1.61) | 1.31 (1.08 to 1.60) |
| New Jersey | 1.25 (1.08 to 1.43) | 1.26 (1.10 to 1.45) |
| Median household income in Census tract of residence | ||
| Quartile 1 (lowest) | Reference | Reference |
| Quartile 2 | .97 (.91 to 1.04) | .97 (.91 to 1.04) |
| Quartile 3 | 1.03 (.96 to 1.11) | 1.04 (.97 to 1.12) |
| Quartile 4 (high) | .90 (.83 to .97) | .90 (.84 to .97) |
| Unknown | .89 (.75 to 1.05) | .90 (.76 to 1.07) |
| Tumor grade | ||
| Well differentiated | .59 (.49 to .71) | .56 (.47 to .67) |
| Moderately differentiated | .69 (.65 to .73) | .67 (.63 to .71) |
| Poorly differentiated/undifferentiated | Reference | Reference |
| Unknown | 1.44 (1.36 to 1.52) | 1.37 (1.29 to 1.45) |
| DCG comorbidity | ||
| Quartile 1 (lowest) | Reference | Reference |
| Quartile 2 | .94 (.87 to 1.01) | .95 (.88 to 1.02) |
| Quartile 3 | 1.05 (.97 to 1.14) | 1.08 (.99 to 1.16) |
| Quartile 4 (high) | 1.41 (1.30 to 1.52) | 1.43 (1.32 to 1.54) |
| Year of diagnosis | 1.01 (1.00 to 1.02) | 1.01 (1.00 to 1.02) |
In a second survival model that also included a time-varying variable for receipt of androgen deprivation therapy within 4 months of diagnosis, we found that receipt of androgen deprivation therapy was associated with improved survival (adjusted hazard ratio .69, 95% CI .66 to .73; Table 3, right column). Inclusion of this variable in the model did not influence the association between other factors and survival. In the model that also included a term for the interaction of androgen deprivation therapy and initiation of this therapy early (vs. delayed), we found no difference in the benefit of hormone therapy based on timing of initiation (regression coefficient=.06, P=.58).
We repeated all analyses restricting the cohort to the 7,234 men who survived and were continuously enrolled in fee-for-service Medicare Parts A and B for at least 4 months, and results were similar (data not shown).
Discussion
In a cohort of older men diagnosed with metastatic prostate cancer in the U.S., approximately two-thirds of men were treated with early hormone therapy, another 7% with delayed hormone therapy, and nearly one-quarter with no hormonal therapy. Although other studies have demonstrated marked increases in use of hormone therapy for men with non-metastatic disease since the early 1990s8–10 for both white and black men,22 we did not observe similar increases in use of androgen deprivation therapy for metastatic disease during this time. Receipt of hormonal therapy was associated with improved survival, an effect that did not vary by timing of initiation of therapy. Although black men compared to white men were particularly likely to receive no hormonal therapy, the racial differences in use of this therapy did not contribute to racial differences in survival.
The primary benefits of androgen deprivation therapy are palliative: less pain and fewer pathological fractures, spinal cord compressions, and episodes of ureteral obstruction.23 Whether hormonal therapy for advanced prostate cancer influences survival has been less clear.6 We observed that men treated with hormonal therapy had better survival than men who did not receive this treatment, similar to another recent report.24 Importantly, the survival benefit did not vary by the timing of initiation of therapy.
The observational nature of our study limits our ability to conclude causality. Men who died soon after diagnosis may not have had the same opportunity as other men to receive hormonal therapy. Nevertheless, we accounted for censored observations using survival analysis methods, and we still observed a survival benefit associated with hormonal therapy (with no benefit of early over late therapy) when we restricted the cohort to men surviving at least 4 months. The lack of benefit for early vs. late hormone therapy suggests that delaying hormonal therapy, particularly for men who are asymptomatic, is unlikely to be harmful. This is particularly important in light of data suggesting that hormonal therapy may be associated with an increased risk of diabetes and cardiovascular disease.25 Despite this potential for increased non-prostate morbidity, we did not observe an increase in non-prostate cancer deaths.
Prior studies suggest that white and black men derive equivalent benefits from androgen deprivation therapy.26, 27 Thus, our finding of disparities in treatment, which is consistent with other data showing less aggressive treatment for minority patients with prostate cancer,28 raises the possibility that use of this therapy might impact outcomes for these men. Yet, despite survival statistics suggesting that black men with prostate cancer have worse survival than white men,2 we found no difference between race or ethnicity and survival in our population-based cohort of men with metastatic cancer at diagnosis after adjusting for patient and tumor characteristics. This finding is consistent with other studies demonstrating similar survival for black and white men with prostate cancer once adjusting for stage, grade, and age29–31 or socioeconomic factors32 or access to medical care.33 Despite the survival advantage we observed with use of hormonal therapy, the racial differences in use did not translate into survival differences, suggesting that physicians may be appropriately selecting patients for delayed or no therapy.
We also observed that unmarried men and men living in certain regions of the U.S., such as Detroit, were less likely than other men to receive androgen deprivation therapy. These findings are consistent with other studies demonstrating treatment differences for individuals without a spouse or partner34, 35 and substantial geographic variations in treatment.36 Future efforts to better understand care for men with prostate cancer should target these nonclinical factors, in addition to race, that contribute to variations in care. Nevertheless, although androgen deprivation was associated with greater survival among men in this cohort, receipt of androgen deprivation therapy did not explain survival differences based on age, marital status, comorbid illness, or area of residence.
Our study has some limitations. First, we had no information about patients’ preferences for hormone therapy, nor did we have information about the presence of symptoms, which is likely to be a strong driver of receipt of early versus late hormonal therapy. In addition, we had no information about PSA levels, although PSA values were not widely used to guide therapy for metastatic disease during the period of study. Second, our analysis is observational and we therefore cannot be sure whether the survival advantages among men who received hormonal therapy were a result of the hormonal therapy or because men who died quickly did not have the same opportunity as other men to be treated with hormonal therapy. However, because we used appropriate statistical methods to account for censored observations and we found similar results when we restricted the cohort to men surviving at least 4 months, we believe that this is unlikely to be a major problem. Second, we studied older men living in regions of the U.S. with SEER registries, so the generalizability of our findings to younger men and those living in different areas requires further study. Older men account for the majority of prostate cancer diagnoses; however, and this population-based sample included cancer patients from areas representing 14% of the U.S. population. Third, SEER-Medicare data do not include information about oral medications, and therefore, we could not ascertain use of oral antiandrogens or estrogens. However, treatment with antiandrogen monotherapy is not approved for prostate cancer in the U.S., so few men likely received such therapy. Estrogens are an acceptable alternative medical therapy for metastatic prostate cancer, but because of their greater toxicity, they are not recommended over GnRH agonists in this setting.
In conclusion, a large minority of older American men diagnosed with metastatic prostate cancer receive delayed or no hormonal therapy, particularly black men. Despite substantial increases in use of hormonal therapy for non-metastatic prostate cancer during the same study period, we did not observe similar increases in use among men with metastatic cancer for whom the benefits of treatment have been better studied. Receipt of androgen deprivation therapy was associated with prolonged survival; however racial differences in use of this treatment did not contribute to racial differences in survival, which were not evident after controlling for patient and tumor characteristics.
Acknowledgments
This study used the linked SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors. The authors acknowledge the efforts of the Applied Research Program, NCI; the Office of Research, Development and Information, CMS; Information Management Services (IMS), Inc.; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database.
The authors thank Laurie Meneades, M.S., and Yang Xu, M.S., for expert programming assistance.
This study was funded by the Prostate Cancer Specialized Program of Research Excellence (SPORE) of the National Cancer Institute (P50CA90381). The funder had no role in design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, or approval of the manuscript.
References
- 1.Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2006. CA Cancer J Clin. 2006;56:106–30. doi: 10.3322/canjclin.56.2.106. [DOI] [PubMed] [Google Scholar]
- 2.Ries LAG, Harkins D, Krapcho M, et al. SEER Cancer Statistics Review, 1975–2003. Bethesda, MD: National Cancer Institute; 2006. [Google Scholar]
- 3.Huggins C, Hodges CV. Studies on prostate cancer. I. The effects of castration, of estrogen, and of androgen injection on serum phosphatases in metastatic carcinoma of the prostate. Cancer Res. 1941;1:293–7. [Google Scholar]
- 4.Huggins C, Stevens RE, Hodges CV. Studies on prostate cancer. II. The effects of castration on advanced carcinoma of the prostate gland. Arch Surg. 1941;43:209–33. [Google Scholar]
- 5.Loblaw DA, Mendelson DS, Talcott JA, et al. American Society of Clinical Oncology recommendations for the initial hormonal management of androgen-sensitive metastatic, recurrent, or progressive prostate cancer. J Clin Oncol. 2004;22:2927–41. doi: 10.1200/JCO.2004.04.579. [DOI] [PubMed] [Google Scholar]
- 6.Loblaw DA, Virgo KS, Nam R, et al. Initial hormonal management of androgen-sensitive metastatic, recurrent, or progressive prostate cancer: 2006 update of an American Society of Clinical Oncology practice guideline. J Clin Oncol. 2007;25:1596–605. doi: 10.1200/JCO.2006.10.1949. [DOI] [PubMed] [Google Scholar]
- 7.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. 1997;79:235–46. doi: 10.1046/j.1464-410x.1997.d01-6840.x. [DOI] [PubMed] [Google Scholar]
- 8.Cooperberg MR, Grossfeld GD, Lubeck DP, Carroll PR. National practice patterns and time trends in androgen ablation for localized prostate cancer. J Natl Cancer Inst. 2003;95:981–9. doi: 10.1093/jnci/95.13.981. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Shahinian VB, Kuo YF, Freeman JL, Orihuela E, Goodwin JS. Increasing use of gonadotropin-releasing hormone agonists for the treatment of localized prostate carcinoma. Cancer. 2005;103:1615–24. doi: 10.1002/cncr.20955. [DOI] [PubMed] [Google Scholar]
- 10.Barry MJ, Delorenzo MA, Walker-Corkery ES, Lucas FL, Wennberg DC. The rising prevalence of androgen deprivation among older American men since the advent of prostate-specific antigen testing: a population-based cohort study. BJU Int. 2006;98:973–8. doi: 10.1111/j.1464-410X.2006.06416.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lu-Yao G, Moore DF, Oleynick J, Dipaola RS, Yao SL. Use of hormonal therapy in men with metastatic prostate cancer. J Urol. 2006;176:526–31. doi: 10.1016/j.juro.2006.03.098. [DOI] [PubMed] [Google Scholar]
- 12.Potosky AL, Riley GF, Lubitz JD, Mentnech RM, Kessler LG. Potential for cancer related health services research using a linked Medicare-tumor registry database. Med Care. 1993;31:732–48. [PubMed] [Google Scholar]
- 13.Warren JL, Klabunde CN, Schrag D, Bach PB, Riley GF. Overview of the SEER-Medicare data-Content, research applications, and generalizability to the United States Elderly Population. Med Care. 2002;40:IV-3–IV-18. doi: 10.1097/01.MLR.0000020942.47004.03. [DOI] [PubMed] [Google Scholar]
- 14.Klabunde CN, Potosky AL, Harlan LC, Kramer BS. Trends and black/white differences in treatment for nonmetastatic prostate cancer. Med Care. 1998;36:1337–48. doi: 10.1097/00005650-199809000-00006. [DOI] [PubMed] [Google Scholar]
- 15.Ellis RP, Pope GC, Iezzoni LI, et al. Diagnosis-based risk adjustment for Medicare capitation payments. Health Care Financing Rev. 1996;17:101–28. [PMC free article] [PubMed] [Google Scholar]
- 16.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–83. doi: 10.1016/0021-9681(87)90171-8. [DOI] [PubMed] [Google Scholar]
- 17.Baldwin LM, Klabunde CN, Green P, Barlow W, Wright G. In search of the perfect comorbidity measure for use with administrative claims data: does it exist? Med Care. 2006;44:745–53. doi: 10.1097/01.mlr.0000223475.70440.07. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ash AS, Posner MA, Speckman J, Franco S, Yacht AC, Bramwell L. Using risk adjustment to examine mortality trends following hospitalization for heart attack in Medicare. Health Serv Res. 2003;38:1253–62. doi: 10.1111/1475-6773.00175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Little RJ. Direct standardization: a tool for teaching linear models for unbalanced data. Am Statistician. 1982;36:38–43. [Google Scholar]
- 20.Hall MC, Fritzsch RJ, Sagalowsky AI, Ahrens A, Petty B, Roehrborn CG. Prospective determination of the hormonal response after cessation of luteinizing hormone-releasing hormone agonist treatment in patients with prostate cancer. Urology. 1999;53:898–902. doi: 10.1016/s0090-4295(99)00061-8. discussion -3. [DOI] [PubMed] [Google Scholar]
- 21.Oefelein MG. Time to normalization of serum testosterone after 3-month luteinizing hormone-releasing hormone agonist administered in the neoadjuvant setting: implications for dosing schedule and neoadjuvant study consideration. J Urol. 1998;160:1685–8. [PubMed] [Google Scholar]
- 22.Zeliadt SB, Potosky AL, Etzioni R, Ramsey SD, Penson DF. Racial disparity in primary and adjuvant treatment for nonmetastatic prostate cancer: SEER-Medicare trends 1991 to 1999. Urology. 2004;64:1171–6. doi: 10.1016/j.urology.2004.07.037. [DOI] [PubMed] [Google Scholar]
- 23.Sharifi N, Gulley JL, Dahut WL. Androgen deprivation therapy for prostate cancer. JAMA. 2005;294:238–44. doi: 10.1001/jama.294.2.238. [DOI] [PubMed] [Google Scholar]
- 24.Lu-Yao G, Moore DF, Oleynick JU, DiPaola RS, Yao SL. Population based study of hormonal therapy and survival in men with metastatic prostate cancer. J Urol. 2007;177:535–9. doi: 10.1016/j.juro.2006.09.049. [DOI] [PubMed] [Google Scholar]
- 25.Keating NL, O’Malley AJ, Smith MR. Diabetes and cardiovascular disease during androgen deprivation therapy for prostate cancer. J Clin Oncol. 2006;24:4448–56. doi: 10.1200/JCO.2006.06.2497. [DOI] [PubMed] [Google Scholar]
- 26.McLeod DG, Schellhammer PF, Vogelzang NJ, et al. Exploratory analysis on the effect of race on clinical outcome in patients with advanced prostate cancer receiving bicalutamide or flutamide, each in combination with LHRH analogues. The Casodex Combination Study Group. Prostate. 1999;40:218–24. doi: 10.1002/(sici)1097-0045(19990901)40:4<218::aid-pros2>3.0.co;2-6. [DOI] [PubMed] [Google Scholar]
- 27.Fowler JE, Bigler SA, Renfroe DL, Dabagia MD. Prostate specific antigen in black and white men after hormonal therapies for prostate cancer. J Urol. 1997;158:150–4. doi: 10.1097/00005392-199707000-00047. [DOI] [PubMed] [Google Scholar]
- 28.Shavers VL, Brown ML. Racial and ethnic disparities in the receipt of cancer treatment. J Natl Cancer Inst. 2002;94:334–57. doi: 10.1093/jnci/94.5.334. [DOI] [PubMed] [Google Scholar]
- 29.Ragland KE, Selvin S, Merrill DW. Black-white differences in stage-specific cancer survival: analysis of seven selected sites. Am J Epidemiol. 1991;133:672–82. doi: 10.1093/oxfordjournals.aje.a115942. [DOI] [PubMed] [Google Scholar]
- 30.Brawn PN, Johnson EH, Kuhl DL, et al. Stage at presentation and survival of white and black patients with prostate carcinoma. Cancer. 1993;71:2569–73. doi: 10.1002/1097-0142(19930415)71:8<2569::aid-cncr2820710822>3.0.co;2-r. [DOI] [PubMed] [Google Scholar]
- 31.Fowler JE, Jr, Bigler SA, Bowman G, Kilambi NK. Race and cause specific survival with prostate cancer: influence of clinical stage, Gleason score, age and treatment. J Urol. 2000;163:137–42. doi: 10.1016/s0022-5347(05)67989-x. [DOI] [PubMed] [Google Scholar]
- 32.Du XL, Fang S, Coker AL, et al. Racial disparity and socioeconomic status in association with survival in older men with local/regional stage prostate carcinoma: findings from a large community-based cohort. Cancer. 2006;106:1276–85. doi: 10.1002/cncr.21732. [DOI] [PubMed] [Google Scholar]
- 33.Optenberg SA, Thompson IM, Friedrichs P, Wojcik B, Stein CR, Kramer B. Race, treatment, and long-term survival from prostate cancer in an equal-access medical care delivery system. JAMA. 1995;274:1599–605. [PubMed] [Google Scholar]
- 34.Iwashyna TJ, Christakis NA. Marriage, widowhood, and health-care use. Soc Sci Med. 2003;57:2137–47. doi: 10.1016/s0277-9536(02)00546-4. [DOI] [PubMed] [Google Scholar]
- 35.Keating NL, Landrum MB, Guadagnoli E, Winer EP, Ayanian JZ. Factors related to underuse of surveillance mammography among breast cancer survivors. J Clin Oncol. 2006;24:85–94. doi: 10.1200/JCO.2005.02.4174. [DOI] [PubMed] [Google Scholar]
- 36.The Center for the Evaluative Clinical Sciences. The quality of medical care in the United States: a report on the Medicare program. Health Forum, Inc; 1999. Last accessed at http://www.dartmouthatlas.org/atlases/99Atlas.pdf on October 30, 2006. [PubMed] [Google Scholar]