- Email: [email protected]
Impact of Family History on Prostate Cancer Mortality in White Men Undergoing Prostate Specific Antigen Based Screening
Oncology: Prostate/Testis/Penis/Urethra
Impact of Family History on Prostate Cancer Mortality in White Men Undergoing Prostate Specific Antigen Based Screening Michael A. Liss,* Haitao Chen, Sij Hemal, Spencer Krane, Christopher J. Kane,† Jianfeng Xu and A. Karim Kader From the Department of Urology, University of California-San Diego Health System (MAL, CJK, AKK), San Diego, California, State Key Laboratory of Genetic Engineering, School of Life Sciences, Fudan University (HC, JX), Shanghai, People’s Republic of China, and Departments of Genomics and Personalized Medicine Research (HC, JX) and Urology (SH, SK), Wake Forest University School of Medicine, Winston-Salem, North Carolina
Purpose: We assessed whether prostate cancer screening would decrease prostate cancer mortality in white men with a family history of prostate cancer. Materials and Methods: Data from the PLCO cancer screening trial were used to compare the screening and usual care arms in the subset of men with and without a family history of prostate cancer. Univariate and multivariate Cox regression analysis, and log rank analysis of Kaplan-Meier curves were done to examine the data for differences in prostate cancer specific survival. Results: A total of 65,179 white subjects were included in the prostate specific antigen screening trial, of whom 7,314 (11.2%) were diagnosed with prostate cancer. Only 4,833 white men (7.4%) had a family history of prostate cancer. Those with a positive family history had a significantly higher incidence of prostate cancer (16.9% vs 10.8%) and higher prostate cancer specific mortality (0.56% vs 0.37%, each p <0.01). On multivariate analysis this trended toward significance (HR 1.47, 95% CI 0.98e2.21, p ¼ 0.06). Screening in men with a positive family history also showed a trend toward decreased prostate cancer specific mortality (HR 0.49, 95% CI 0.22e1.1, p ¼ 0.08) and decreased time to death from prostate cancer (log rank p ¼ 0.05). Conclusions: White men with a family history of prostate cancer are at increased risk for being diagnosed with and subsequently dying of prostate cancer. Yearly digital rectal examination and prostate specific antigen testing may decrease prostate cancer death in these individuals. Key Words: prostatic neoplasms, mortality, prostate-specific antigen, mass screening, European continental ancestry group
PROSTATE cancer is a heterogeneous disease with a variable natural history that accounts for a large proportion of cancer deaths in men.1 The postulated value of screening for PCa is early detection, which provides an opportunity to change the lethal outcome. Two large PCa screening trials demonstrated little to no
Abbreviations and Acronyms BMI ¼ body mass index FH ¼ family history PCa ¼ prostate cancer PLCO ¼ Prostate, Lung, Colorectal and Ovarian PSA ¼ prostate specific antigen Accepted for publication July 14, 2014. Study received University of California-San Diego institutional review board approval. Supported by NIH (National Institutes of Health) Grant U54HL108460 through the NIH Roadmap for Medical Research, NIH Grant CA140262 (JX) and Grant 81130047 from the Key Project of the National Natural Science Foundation of China (JX) * Correspondence: University of CaliforniaSan Diego Moores Cancer Center, 3855 Health Sciences Dr., #0987, La Jolla, California 92093 (telephone: 858-822-7874; FAX: 858-822-6188; e-mail: [email protected]). † Financial interest and/or other relationship with Amgen, Janssen, Dendreon and Intuitive Surgical. Presented at annual meeting of American Urological Association, Orlando, Florida, May 20, 2014.
survival benefit for PSA based screening, although there was contamination bias.2,3 Particularly the American PLCO cancer screening trial showed significant prescreening and contamination bias, making interpretation controversial.4 As a result, USPSTF (United States Preventive Services Task
0022-5347/15/1931-0075/0 THE JOURNAL OF UROLOGY® © 2015 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RESEARCH, INC.
http://dx.doi.org/10.1016/j.juro.2014.07.085 Vol. 193, 75-79, January 2015 Printed in U.S.A.
www.jurology.com
j
75
76
IMPACT OF FAMILY HISTORY ON PROSTATE CANCER MORTALITY IN WHITE MEN
Force) recommended against PCa screening due to concern of the potential harms of over diagnosis and overtreatment, and the relative lack of benefit.5 In response some groups such as the AUA (American Urological Association) and others suggest that screening efforts should not be abandoned completely, given that PCa is the second leading cause of cancer related death in the United States. Rather, screening should focus on men at high risk for the disease, namely men of black race and those with a positive FH.6 However, to our knowledge there is no evidence to date that there is a survival benefit to this approach. We investigated the effect of screening on PCa specific mortality in white men with a FH of PCa using data from the PLCO screening trial.
METHODS Study Population After receiving institutional review board approval and project approval (PAR Study ID 2013-00046) we obtained de-identified data from men enrolled in the prostate arm of the PLCO cancer screening trial. The data were loaded in iDASH (Integrating Data for Analysis, Anonymization and Sharing), a secure, web based storage facility. PLCO is a NCI (National Cancer Institute) sponsored, randomized trial designed to determine the effects of screening on cancer outcomes in subjects 55 to 74 years old from 1993 to 2001 at 10 centers in the United States. The study enrolled 73,200 men and followed them for 13 years with followup analysis anticipated at 15 years.2,7 A block randomization protocol was used, stratified by age, gender and institution. Patients were excluded from study if they already had prostate, lung, colon or rectal cancer, or were undergoing cancer therapy. In 1995 the study added the exclusion of men with more than 1 serum PSA measurement in the previous 3 years to decrease prescreening bias. Screening was defined as yearly digital rectal examination and serum PSA measurement. Abnormal prostate examination or PSA greater than 4.0 ng/ml was considered a positive screening test. Such men were encouraged to seek diagnostic evaluation and their primary provider was notified. The comparator was usual care, meaning that men were screened as the general population would be screened. In men in the control arm almost 3 PSA tests were performed during the 6-year study.4 PCa specific mortality was the primary outcome of PCLO study analysis. PCa death was determined from the National Death Index, review of death certificates and panel review of the data to determine cause of death. A decision was made to focus on white race due to the already known high mortality rate of PCa in the black American population. There were also many fewer black men in this trial (3,370 or 4.4%) and the FH rate was much lower (234 or 4.4%). Additionally, 30 deaths occurred and none of these patients had a FH of PCa. Like black men, there were too few men of other races with a
positive FH to perform accurate analysis. White race was based on self-declaration at trial entry.
Outcomes The primary study outcome was PCa specific mortality as indicated by death review. PCa mortality was compared in men with vs without a FH to determine a difference in survival if men in this high risk group were targeted for PSA based PCa screening. The primary predictor variable was a FH of PCa obtained from the baseline questionnaire for male participants. The answer was provided along with the relationship to the family member. FH information was dichotomized in the data set as “Family history of prostate cancer?” with a yes response indicating a PCa diagnosis in an immediate family member.
Statistical Analysis Initial analysis was performed to determine the risk of PCa death in men with a FH vs those with no FH to determine PCa risk factors in the PLCO. Subsequently we investigated only men with a FH positive for PCa to determine the impact of screening vs usual care in this specific population. We analyzed the effect of screening in this select group using intent to screen analysis. Initially a rate ratio was performed as in the initial study of the PCa results of the PLCO trial.4 Univariate and multivariate Cox regression analysis, and log rank analysis of Kaplan-Meier curves were used to examine the data. The limited data set obtained from PLCO included the nonPCa associated variables of age, BMI, marital status, smoking and education. Due to the potential that these factors might have impacting PCa treatment after diagnosis we included all of these variables on multivariate analysis.
RESULTS Overall A total of 65,179 white subjects were included in the PLCO screening trial, of whom 7,314 (11.22%) were diagnosed with PCa. Only 4,833 white men (7.4%) had a FH of PCa. Median age was 62 years (range 49 to 78) and median followup was 11.6 years (range 11.3 to 13.3) (supplementary table, http://jurology. com/). Overall 248 men died of PCa. Positive vs Negative FH Men with a FH of PCa were less likely to smoke (nonsmoker 39.6% vs 35.5%, p <0.001) and more likely to be married (85.4% vs 83.6%, p <0.001). Those with a positive FH were also more likely to have undergone prostate biopsy before study enrollment than men with a negative FH (6.81% or 344 of 4,833 vs 4.83% or 2,764 of 58,767, p <0.001). Compared to men with a negative FH those with a positive FH had a significantly higher PCa incidence (16.9% vs 10.8%) and higher PCa specific mortality (0.56% vs 0.37%, each p <0.01). However, age at diagnosis, PSA at diagnosis and Gleason score were similar in the negative and positive FH
IMPACT OF FAMILY HISTORY ON PROSTATE CANCER MORTALITY IN WHITE MEN
groups (supplementary table, http://jurology.com/). PCa death was significantly higher in men with a positive vs a negative FH (0.56% or 27 of 4,833 vs 0.37% or 216 of 58,767, p ¼ 0.039). This trended toward significance on multivariate analysis (HR 1.47, 95% CI 0.98e2.21, p ¼ 0.06, see table). On additional multivariate analyses advanced patient age (HR 1.14, 95% CI 1.12e1.17, p <0.001) and BMI (HR 1.05, 95% CI 1.01e1.07, p ¼ 0.004) were significantly associated with PCa mortality (see table). PCa Screening When investigating only the subset of men with a FH of PCa in the PLCO trial, 18 of the 2,350 in the usual care arm died of PCa and 9 of the 2,483 in the screening arm died (rate ratio 0.5, 95% CI e0.05e0.8). Therefore, lower PCa specific mortality was found in the screening arm than in the usual care arm (0.36% vs 0.77%, p ¼ 0.06, see table). On multivariate Cox regression analysis controlling for age, BMI, education level and smoking screened men with a positive FH also showed a trend toward decreased PCa specific mortality (HR 0.49, 95% CI 0.22e1.1, p ¼ 0.08). Kaplan-Meier curves were generated to show the potential benefit of PSA based screening in positive FH cases (log rank p ¼ 0.05, fig. 1). This was not seen in negative FH cases (fig. 2).
77
DISCUSSION Screening white men with a FH of PCa resulted in almost a 50% decrease in PCa mortality. However, this did not attain statistical significance on multivariate analysis, likely due to the low event rate despite the more than 75,000 men enrolled in PLCO. A positive FH was only noted in 7% of the overall population with a total of 27 deaths in this group. Early divergence of the Kaplan-Meier curves seems to provide evidence of the long held but to our knowledge unproven belief that screening these men at high risk would result in decreased PCa specific mortality at longer followup (fig. 1). With these promising data it may be more prudent to focus screening efforts on high risk groups such as those with a positive FH, and possibly even black men and others with a genetic predisposition rather than abandoning them in general as suggested by USPSTF. There are several criticisms of the PLCO trial, including the facts that a proportion of men were screened before trial enrollment and screening contamination existed in the usual care arm.8 Prescreening and contamination would bias our results to the null hypothesis of no benefit of screening. Moreover, men with a FH of PCa were more likely to have undergone prostate biopsy before enrollment (6.81% vs 4.83%, p <0.001), further biasing results to the null. Therefore, the modest benefit of
Risk factors for PCa death overall and in men with PCa FH in PLCO
Overall/PCa Death
Median Days to Death
Univariate Cox Analysis
Multivariate Cox Analysis
HR (95% CI)
p Value
HR (95% CI)
p Value
Overall PCa death Randomization arm: Usual care Intervention Age FH: Neg Pos BMI College: No Yes Smoking: Never Ever Randomization arm: Usual care Intervention Age BMI College: No Yes Smoking: Never Ever * Mean SD.
32,136/124 Pts (0.39%) 33,043/124 Pts (0.38%) 62.73 5.32/65.95 5.32*
3,377 3,365.5 3,375.5
1 0.966 (0.753e1.239) 1.134 (1.107e1.161)
0.785 <0.001
1 0.97 (0.753e1.25) 1.144 (1.116e1.172)
0.815 <0.001
58,767/216 Pts (0.37%) 4,833/27 Pts (0.56%) 27.62 4.14/28.05 4.50 kg/m2*
3,362 3,382 3,375.5
1 1.503 (1.007e2.242) 1.026 (0.996e1.056)
0.046 0.090
1 1.473 (0.98e2.214) 1.045 (1.014e1.077)
0.062 0.004
3,165.5 3,407
1 1.069 (0.824e1.386)
0.617
1 1.141 (0.874e1.489)
0.333
1 0.957 (0.741e1.235)
0.735
1 0.985 (0.759e1.28)
0.912
3,370 3,527 3,382 3,382
1 0.463 (0.208e1.031) 1.108 (1.033e1.188) 1.082 (0.997e1.174)
0.059 0.004 0.059
1 0.488 (0.217e1.095) 1.131 (1.051e1.216) 1.108 (1.023e1.201)
0.082 <0.001 0.012
1,770/8 Pts (0.45%) 3,051/19 Pts (0.62%)
4,025 3,306
1 1.323 (0.579e3.021)
0.507
1 1.438 (0.622e3.325)
0.395
1,917/14 Pts (0.73%) 2,914/13 Pts (0.45%)
3,960 2,858
1 0.645 (0.303e1.373)
0.256
1 0.611 (0.282e1.324)
0.212
24,606/88 Pts (0.36%) 40,412/160 Pts (0.40%) 23,864/97 Pts (0.41%) 41,299/151 Pts (0.37%) 2,350/18 Pts (0.77%) 2,483/9 Pts (0.36%) 62.62 5.34/65.30 5.38* 27.39 4.00/28.82 5.65 kg/m2*
3,363 3,379 Pos FH PCa death
78
IMPACT OF FAMILY HISTORY ON PROSTATE CANCER MORTALITY IN WHITE MEN
Figure 1. Kaplan-Meier curve of PCa mortality in white men with positive FH.
decreasing PCa specific mortality by PSA screening in white men with a positive FH is likely to be much greater than shown on this subanalysis. Additionally, our findings are in sharp contrast to the observation that PSA screening in the entire PLCO cohort increased PCa mortality by approximately 9%.7 We report that men with a FH of PCa are more likely to be diagnosed with PCa, as previously described in this trial and in men undergoing study mandated biopsies in REDUCE (Reduction by Dutasteride of Prostate Cancer Events)9 and PCPT (Prostate Cancer Prevention Trial).10 Additionally, we noted a significant increase in death from PCa despite a low event rate (0.56% vs 0.37%, p ¼ 0.039, HR 1.47, 95% CI 0.98e2.21, p ¼ 0.06).
While the relative risk reduction can be viewed as minimal (0.2%), we believe that these data are relatively early signs of benefit due to the prolonged natural history of PCa. PCa continues to be a great contributor to death in American men. Therefore, screening continues to be a controversial topic to decrease the morbidity and mortality of this prevalent cancer. Screening men at high risk such as those with a FH of PCa seems to be a rational compromise until improved screening methods are developed. We also cannot ignore the results of European trials that show significant benefits to screening.3,11 In parallel men in the PLCO trial with a FH showed increased PCa mortality, similar to that in a Swedish study that also indicated higher PCa mortality rates in men with a positive FH (nonfatal vs fatal positive FH HR 1.59 vs 2.03).12 Our definition of positive FH as an immediate family member with PCa was shown to be reliable13 but it was only noted in 7.4% of the white men in our analysis. FH is subject to various challenges, including but not limited to recall bias, adoption, education and the number of male family members.14 Given the mentioned challenges, negative FH does not exclude men from genetic risk. Given the fact that FH is a surrogate for genetic predisposition, our findings could provide the basis for a genomic targeted approach to PSA screening.15 A more individualized approach may provide additional risk profile information, which may differentiate family members and provide a marker of risk that would remain constant throughout the life of the man.16 A previous study indicated that using genetic variants and FH to provide information similar to that of PSA could potentially be proposed as an initial screen to determine the degree of future screening events.16 These results should be interpreted with caution since this was a secondary analysis of a randomized trial with a small number of events. However, PLCO is the only such randomized trial performed in the United States with regimented protocols and followup. Results may not be extrapolated to other races.
CONCLUSIONS White men with a FH of PCa are at increased risk for being diagnosed with PCa and subsequently dying of PCa. However, yearly digital rectal examination and PSA screening may decrease the number of PCa deaths in these individuals.
ACKNOWLEDGMENTS Figure 2. Kaplan-Meier curve of PCa mortality in white men with negative FH.
NCI granted access to PLCO data. iDASH is housed at the San Diego Supercomputer Center, University of California-San Diego.
IMPACT OF FAMILY HISTORY ON PROSTATE CANCER MORTALITY IN WHITE MEN
79
REFERENCES 1. Siegel R, Naishadham D and Jemal A: Cancer statistics, 2013. CA Cancer J Clin 2013; 63: 11.
6. Carter HB, Albertsen PC, Barry MJ et al: Early detection of prostate cancer: AUA guideline. J Urol 2013; 190: 419.
2. Andriole GL, Crawford ED, Grubb RL 3rd et al: Mortality results from a randomized prostatecancer screening trial. N Engl J Med 2009; 360: 1310.
7. Andriole GL, Crawford ED, Grubb RL 3rd et al: Prostate cancer screening in the randomized Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial: mortality results after 13 years of follow-up. J Natl Cancer Inst 2012; 104: 125.
3. Schroder FH, Hugosson J, Roobol MJ et al: Screening and prostate-cancer mortality in a randomized European study. N Engl J Med 2009; 360: 1320. 4. Pinsky PF, Blacka A, Kramer BS et al: Assessing contamination and compliance in the prostate component of the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. Clin Trials 2010; 7: 303. 5. Moyer VA: Screening for prostate cancer: U.S. Preventive Services Task Force recommendation statement. U.S. Preventive Services Task Force. Ann Intern Med 2012; 157: 120.
8. Andriole GL: Update of the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial. Recent results. Cancer Res 2014; 202: 53. 9. Elshafei A, Moussa AS, Hatem A et al: Does positive family history of prostate cancer increase the risk of prostate cancer on initial prostate biopsy? Urology 2013; 81: 826. 10. Thompson IM, Ankerst DP, Chi C et al: Assessing prostate cancer risk: results from the Prostate Cancer Prevention Trial. J Natl Cancer Inst 2006; 98: 529. 11. Hugosson J, Carlsson S, Aus G et al: Mortality results from the Goteborg randomised
population-based prostate-cancer screening trial. Lancet Oncol 2010; 11: 725. 12. Roudgari H, Hemminki K, Brandt A et al: Prostate cancer risk assessment model: a scoring model based on the Swedish Family-Cancer Database. J Med Genet 2012; 49: 345. 13. Ziogas A and Anton-Culver H: Validation of family history data in cancer family registries. Am J Prev Med 2002; 24: 190. 14. Murff HJ, Spigel DR and Syngal S: Does this patient have a family history of cancer? An evidence-based analysis of the accuracy of family cancer history. JAMA 2004; 292: 1480. 15. Zheng SL, Sun J, Wiklund F et al: Cumulative association of five genetic variants with prostate cancer. N Engl J Med 2008; 358: 910. 16. Sun J, Na R, Hsu FC et al: Genetic score is an objective and better measurement of inherited risk of prostate cancer than family history. Eur Urol 2013; 63: 585.
Impact of Family History on Prostate Cancer Mortality in White Men Undergoing Prostate Specific Antigen Based Screening Michael A. Liss,* Haitao Chen, Sij Hemal, Spencer Krane, Christopher J. Kane,† Jianfeng Xu and A. Karim Kader From the Department of Urology, University of California-San Diego Health System (MAL, CJK, AKK), San Diego, California, State Key Laboratory of Genetic Engineering, School of Life Sciences, Fudan University (HC, JX), Shanghai, People’s Republic of China, and Departments of Genomics and Personalized Medicine Research (HC, JX) and Urology (SH, SK), Wake Forest University School of Medicine, Winston-Salem, North Carolina
Purpose: We assessed whether prostate cancer screening would decrease prostate cancer mortality in white men with a family history of prostate cancer. Materials and Methods: Data from the PLCO cancer screening trial were used to compare the screening and usual care arms in the subset of men with and without a family history of prostate cancer. Univariate and multivariate Cox regression analysis, and log rank analysis of Kaplan-Meier curves were done to examine the data for differences in prostate cancer specific survival. Results: A total of 65,179 white subjects were included in the prostate specific antigen screening trial, of whom 7,314 (11.2%) were diagnosed with prostate cancer. Only 4,833 white men (7.4%) had a family history of prostate cancer. Those with a positive family history had a significantly higher incidence of prostate cancer (16.9% vs 10.8%) and higher prostate cancer specific mortality (0.56% vs 0.37%, each p <0.01). On multivariate analysis this trended toward significance (HR 1.47, 95% CI 0.98e2.21, p ¼ 0.06). Screening in men with a positive family history also showed a trend toward decreased prostate cancer specific mortality (HR 0.49, 95% CI 0.22e1.1, p ¼ 0.08) and decreased time to death from prostate cancer (log rank p ¼ 0.05). Conclusions: White men with a family history of prostate cancer are at increased risk for being diagnosed with and subsequently dying of prostate cancer. Yearly digital rectal examination and prostate specific antigen testing may decrease prostate cancer death in these individuals. Key Words: prostatic neoplasms, mortality, prostate-specific antigen, mass screening, European continental ancestry group
PROSTATE cancer is a heterogeneous disease with a variable natural history that accounts for a large proportion of cancer deaths in men.1 The postulated value of screening for PCa is early detection, which provides an opportunity to change the lethal outcome. Two large PCa screening trials demonstrated little to no
Abbreviations and Acronyms BMI ¼ body mass index FH ¼ family history PCa ¼ prostate cancer PLCO ¼ Prostate, Lung, Colorectal and Ovarian PSA ¼ prostate specific antigen Accepted for publication July 14, 2014. Study received University of California-San Diego institutional review board approval. Supported by NIH (National Institutes of Health) Grant U54HL108460 through the NIH Roadmap for Medical Research, NIH Grant CA140262 (JX) and Grant 81130047 from the Key Project of the National Natural Science Foundation of China (JX) * Correspondence: University of CaliforniaSan Diego Moores Cancer Center, 3855 Health Sciences Dr., #0987, La Jolla, California 92093 (telephone: 858-822-7874; FAX: 858-822-6188; e-mail: [email protected]). † Financial interest and/or other relationship with Amgen, Janssen, Dendreon and Intuitive Surgical. Presented at annual meeting of American Urological Association, Orlando, Florida, May 20, 2014.
survival benefit for PSA based screening, although there was contamination bias.2,3 Particularly the American PLCO cancer screening trial showed significant prescreening and contamination bias, making interpretation controversial.4 As a result, USPSTF (United States Preventive Services Task
0022-5347/15/1931-0075/0 THE JOURNAL OF UROLOGY® © 2015 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RESEARCH, INC.
http://dx.doi.org/10.1016/j.juro.2014.07.085 Vol. 193, 75-79, January 2015 Printed in U.S.A.
www.jurology.com
j
75
76
IMPACT OF FAMILY HISTORY ON PROSTATE CANCER MORTALITY IN WHITE MEN
Force) recommended against PCa screening due to concern of the potential harms of over diagnosis and overtreatment, and the relative lack of benefit.5 In response some groups such as the AUA (American Urological Association) and others suggest that screening efforts should not be abandoned completely, given that PCa is the second leading cause of cancer related death in the United States. Rather, screening should focus on men at high risk for the disease, namely men of black race and those with a positive FH.6 However, to our knowledge there is no evidence to date that there is a survival benefit to this approach. We investigated the effect of screening on PCa specific mortality in white men with a FH of PCa using data from the PLCO screening trial.
METHODS Study Population After receiving institutional review board approval and project approval (PAR Study ID 2013-00046) we obtained de-identified data from men enrolled in the prostate arm of the PLCO cancer screening trial. The data were loaded in iDASH (Integrating Data for Analysis, Anonymization and Sharing), a secure, web based storage facility. PLCO is a NCI (National Cancer Institute) sponsored, randomized trial designed to determine the effects of screening on cancer outcomes in subjects 55 to 74 years old from 1993 to 2001 at 10 centers in the United States. The study enrolled 73,200 men and followed them for 13 years with followup analysis anticipated at 15 years.2,7 A block randomization protocol was used, stratified by age, gender and institution. Patients were excluded from study if they already had prostate, lung, colon or rectal cancer, or were undergoing cancer therapy. In 1995 the study added the exclusion of men with more than 1 serum PSA measurement in the previous 3 years to decrease prescreening bias. Screening was defined as yearly digital rectal examination and serum PSA measurement. Abnormal prostate examination or PSA greater than 4.0 ng/ml was considered a positive screening test. Such men were encouraged to seek diagnostic evaluation and their primary provider was notified. The comparator was usual care, meaning that men were screened as the general population would be screened. In men in the control arm almost 3 PSA tests were performed during the 6-year study.4 PCa specific mortality was the primary outcome of PCLO study analysis. PCa death was determined from the National Death Index, review of death certificates and panel review of the data to determine cause of death. A decision was made to focus on white race due to the already known high mortality rate of PCa in the black American population. There were also many fewer black men in this trial (3,370 or 4.4%) and the FH rate was much lower (234 or 4.4%). Additionally, 30 deaths occurred and none of these patients had a FH of PCa. Like black men, there were too few men of other races with a
positive FH to perform accurate analysis. White race was based on self-declaration at trial entry.
Outcomes The primary study outcome was PCa specific mortality as indicated by death review. PCa mortality was compared in men with vs without a FH to determine a difference in survival if men in this high risk group were targeted for PSA based PCa screening. The primary predictor variable was a FH of PCa obtained from the baseline questionnaire for male participants. The answer was provided along with the relationship to the family member. FH information was dichotomized in the data set as “Family history of prostate cancer?” with a yes response indicating a PCa diagnosis in an immediate family member.
Statistical Analysis Initial analysis was performed to determine the risk of PCa death in men with a FH vs those with no FH to determine PCa risk factors in the PLCO. Subsequently we investigated only men with a FH positive for PCa to determine the impact of screening vs usual care in this specific population. We analyzed the effect of screening in this select group using intent to screen analysis. Initially a rate ratio was performed as in the initial study of the PCa results of the PLCO trial.4 Univariate and multivariate Cox regression analysis, and log rank analysis of Kaplan-Meier curves were used to examine the data. The limited data set obtained from PLCO included the nonPCa associated variables of age, BMI, marital status, smoking and education. Due to the potential that these factors might have impacting PCa treatment after diagnosis we included all of these variables on multivariate analysis.
RESULTS Overall A total of 65,179 white subjects were included in the PLCO screening trial, of whom 7,314 (11.22%) were diagnosed with PCa. Only 4,833 white men (7.4%) had a FH of PCa. Median age was 62 years (range 49 to 78) and median followup was 11.6 years (range 11.3 to 13.3) (supplementary table, http://jurology. com/). Overall 248 men died of PCa. Positive vs Negative FH Men with a FH of PCa were less likely to smoke (nonsmoker 39.6% vs 35.5%, p <0.001) and more likely to be married (85.4% vs 83.6%, p <0.001). Those with a positive FH were also more likely to have undergone prostate biopsy before study enrollment than men with a negative FH (6.81% or 344 of 4,833 vs 4.83% or 2,764 of 58,767, p <0.001). Compared to men with a negative FH those with a positive FH had a significantly higher PCa incidence (16.9% vs 10.8%) and higher PCa specific mortality (0.56% vs 0.37%, each p <0.01). However, age at diagnosis, PSA at diagnosis and Gleason score were similar in the negative and positive FH
IMPACT OF FAMILY HISTORY ON PROSTATE CANCER MORTALITY IN WHITE MEN
groups (supplementary table, http://jurology.com/). PCa death was significantly higher in men with a positive vs a negative FH (0.56% or 27 of 4,833 vs 0.37% or 216 of 58,767, p ¼ 0.039). This trended toward significance on multivariate analysis (HR 1.47, 95% CI 0.98e2.21, p ¼ 0.06, see table). On additional multivariate analyses advanced patient age (HR 1.14, 95% CI 1.12e1.17, p <0.001) and BMI (HR 1.05, 95% CI 1.01e1.07, p ¼ 0.004) were significantly associated with PCa mortality (see table). PCa Screening When investigating only the subset of men with a FH of PCa in the PLCO trial, 18 of the 2,350 in the usual care arm died of PCa and 9 of the 2,483 in the screening arm died (rate ratio 0.5, 95% CI e0.05e0.8). Therefore, lower PCa specific mortality was found in the screening arm than in the usual care arm (0.36% vs 0.77%, p ¼ 0.06, see table). On multivariate Cox regression analysis controlling for age, BMI, education level and smoking screened men with a positive FH also showed a trend toward decreased PCa specific mortality (HR 0.49, 95% CI 0.22e1.1, p ¼ 0.08). Kaplan-Meier curves were generated to show the potential benefit of PSA based screening in positive FH cases (log rank p ¼ 0.05, fig. 1). This was not seen in negative FH cases (fig. 2).
77
DISCUSSION Screening white men with a FH of PCa resulted in almost a 50% decrease in PCa mortality. However, this did not attain statistical significance on multivariate analysis, likely due to the low event rate despite the more than 75,000 men enrolled in PLCO. A positive FH was only noted in 7% of the overall population with a total of 27 deaths in this group. Early divergence of the Kaplan-Meier curves seems to provide evidence of the long held but to our knowledge unproven belief that screening these men at high risk would result in decreased PCa specific mortality at longer followup (fig. 1). With these promising data it may be more prudent to focus screening efforts on high risk groups such as those with a positive FH, and possibly even black men and others with a genetic predisposition rather than abandoning them in general as suggested by USPSTF. There are several criticisms of the PLCO trial, including the facts that a proportion of men were screened before trial enrollment and screening contamination existed in the usual care arm.8 Prescreening and contamination would bias our results to the null hypothesis of no benefit of screening. Moreover, men with a FH of PCa were more likely to have undergone prostate biopsy before enrollment (6.81% vs 4.83%, p <0.001), further biasing results to the null. Therefore, the modest benefit of
Risk factors for PCa death overall and in men with PCa FH in PLCO
Overall/PCa Death
Median Days to Death
Univariate Cox Analysis
Multivariate Cox Analysis
HR (95% CI)
p Value
HR (95% CI)
p Value
Overall PCa death Randomization arm: Usual care Intervention Age FH: Neg Pos BMI College: No Yes Smoking: Never Ever Randomization arm: Usual care Intervention Age BMI College: No Yes Smoking: Never Ever * Mean SD.
32,136/124 Pts (0.39%) 33,043/124 Pts (0.38%) 62.73 5.32/65.95 5.32*
3,377 3,365.5 3,375.5
1 0.966 (0.753e1.239) 1.134 (1.107e1.161)
0.785 <0.001
1 0.97 (0.753e1.25) 1.144 (1.116e1.172)
0.815 <0.001
58,767/216 Pts (0.37%) 4,833/27 Pts (0.56%) 27.62 4.14/28.05 4.50 kg/m2*
3,362 3,382 3,375.5
1 1.503 (1.007e2.242) 1.026 (0.996e1.056)
0.046 0.090
1 1.473 (0.98e2.214) 1.045 (1.014e1.077)
0.062 0.004
3,165.5 3,407
1 1.069 (0.824e1.386)
0.617
1 1.141 (0.874e1.489)
0.333
1 0.957 (0.741e1.235)
0.735
1 0.985 (0.759e1.28)
0.912
3,370 3,527 3,382 3,382
1 0.463 (0.208e1.031) 1.108 (1.033e1.188) 1.082 (0.997e1.174)
0.059 0.004 0.059
1 0.488 (0.217e1.095) 1.131 (1.051e1.216) 1.108 (1.023e1.201)
0.082 <0.001 0.012
1,770/8 Pts (0.45%) 3,051/19 Pts (0.62%)
4,025 3,306
1 1.323 (0.579e3.021)
0.507
1 1.438 (0.622e3.325)
0.395
1,917/14 Pts (0.73%) 2,914/13 Pts (0.45%)
3,960 2,858
1 0.645 (0.303e1.373)
0.256
1 0.611 (0.282e1.324)
0.212
24,606/88 Pts (0.36%) 40,412/160 Pts (0.40%) 23,864/97 Pts (0.41%) 41,299/151 Pts (0.37%) 2,350/18 Pts (0.77%) 2,483/9 Pts (0.36%) 62.62 5.34/65.30 5.38* 27.39 4.00/28.82 5.65 kg/m2*
3,363 3,379 Pos FH PCa death
78
IMPACT OF FAMILY HISTORY ON PROSTATE CANCER MORTALITY IN WHITE MEN
Figure 1. Kaplan-Meier curve of PCa mortality in white men with positive FH.
decreasing PCa specific mortality by PSA screening in white men with a positive FH is likely to be much greater than shown on this subanalysis. Additionally, our findings are in sharp contrast to the observation that PSA screening in the entire PLCO cohort increased PCa mortality by approximately 9%.7 We report that men with a FH of PCa are more likely to be diagnosed with PCa, as previously described in this trial and in men undergoing study mandated biopsies in REDUCE (Reduction by Dutasteride of Prostate Cancer Events)9 and PCPT (Prostate Cancer Prevention Trial).10 Additionally, we noted a significant increase in death from PCa despite a low event rate (0.56% vs 0.37%, p ¼ 0.039, HR 1.47, 95% CI 0.98e2.21, p ¼ 0.06).
While the relative risk reduction can be viewed as minimal (0.2%), we believe that these data are relatively early signs of benefit due to the prolonged natural history of PCa. PCa continues to be a great contributor to death in American men. Therefore, screening continues to be a controversial topic to decrease the morbidity and mortality of this prevalent cancer. Screening men at high risk such as those with a FH of PCa seems to be a rational compromise until improved screening methods are developed. We also cannot ignore the results of European trials that show significant benefits to screening.3,11 In parallel men in the PLCO trial with a FH showed increased PCa mortality, similar to that in a Swedish study that also indicated higher PCa mortality rates in men with a positive FH (nonfatal vs fatal positive FH HR 1.59 vs 2.03).12 Our definition of positive FH as an immediate family member with PCa was shown to be reliable13 but it was only noted in 7.4% of the white men in our analysis. FH is subject to various challenges, including but not limited to recall bias, adoption, education and the number of male family members.14 Given the mentioned challenges, negative FH does not exclude men from genetic risk. Given the fact that FH is a surrogate for genetic predisposition, our findings could provide the basis for a genomic targeted approach to PSA screening.15 A more individualized approach may provide additional risk profile information, which may differentiate family members and provide a marker of risk that would remain constant throughout the life of the man.16 A previous study indicated that using genetic variants and FH to provide information similar to that of PSA could potentially be proposed as an initial screen to determine the degree of future screening events.16 These results should be interpreted with caution since this was a secondary analysis of a randomized trial with a small number of events. However, PLCO is the only such randomized trial performed in the United States with regimented protocols and followup. Results may not be extrapolated to other races.
CONCLUSIONS White men with a FH of PCa are at increased risk for being diagnosed with PCa and subsequently dying of PCa. However, yearly digital rectal examination and PSA screening may decrease the number of PCa deaths in these individuals.
ACKNOWLEDGMENTS Figure 2. Kaplan-Meier curve of PCa mortality in white men with negative FH.
NCI granted access to PLCO data. iDASH is housed at the San Diego Supercomputer Center, University of California-San Diego.
IMPACT OF FAMILY HISTORY ON PROSTATE CANCER MORTALITY IN WHITE MEN
79
REFERENCES 1. Siegel R, Naishadham D and Jemal A: Cancer statistics, 2013. CA Cancer J Clin 2013; 63: 11.
6. Carter HB, Albertsen PC, Barry MJ et al: Early detection of prostate cancer: AUA guideline. J Urol 2013; 190: 419.
2. Andriole GL, Crawford ED, Grubb RL 3rd et al: Mortality results from a randomized prostatecancer screening trial. N Engl J Med 2009; 360: 1310.
7. Andriole GL, Crawford ED, Grubb RL 3rd et al: Prostate cancer screening in the randomized Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial: mortality results after 13 years of follow-up. J Natl Cancer Inst 2012; 104: 125.
3. Schroder FH, Hugosson J, Roobol MJ et al: Screening and prostate-cancer mortality in a randomized European study. N Engl J Med 2009; 360: 1320. 4. Pinsky PF, Blacka A, Kramer BS et al: Assessing contamination and compliance in the prostate component of the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. Clin Trials 2010; 7: 303. 5. Moyer VA: Screening for prostate cancer: U.S. Preventive Services Task Force recommendation statement. U.S. Preventive Services Task Force. Ann Intern Med 2012; 157: 120.
8. Andriole GL: Update of the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial. Recent results. Cancer Res 2014; 202: 53. 9. Elshafei A, Moussa AS, Hatem A et al: Does positive family history of prostate cancer increase the risk of prostate cancer on initial prostate biopsy? Urology 2013; 81: 826. 10. Thompson IM, Ankerst DP, Chi C et al: Assessing prostate cancer risk: results from the Prostate Cancer Prevention Trial. J Natl Cancer Inst 2006; 98: 529. 11. Hugosson J, Carlsson S, Aus G et al: Mortality results from the Goteborg randomised
population-based prostate-cancer screening trial. Lancet Oncol 2010; 11: 725. 12. Roudgari H, Hemminki K, Brandt A et al: Prostate cancer risk assessment model: a scoring model based on the Swedish Family-Cancer Database. J Med Genet 2012; 49: 345. 13. Ziogas A and Anton-Culver H: Validation of family history data in cancer family registries. Am J Prev Med 2002; 24: 190. 14. Murff HJ, Spigel DR and Syngal S: Does this patient have a family history of cancer? An evidence-based analysis of the accuracy of family cancer history. JAMA 2004; 292: 1480. 15. Zheng SL, Sun J, Wiklund F et al: Cumulative association of five genetic variants with prostate cancer. N Engl J Med 2008; 358: 910. 16. Sun J, Na R, Hsu FC et al: Genetic score is an objective and better measurement of inherited risk of prostate cancer than family history. Eur Urol 2013; 63: 585.