MILITARY MEDICINE, 179, 1:81,2014
Vitamin D and Prostate Cancer Survival in Veterans Tatyana Der, MD*; Beth A. Bailey, PhDf; Dima Youssef, MD*; Todd Manning, BAf; William B. Grant, PhD§; Alan N. Peiris, MD, PhD*¡I ABSTRACT Prostate cancer remains the second most commonly diagnosed cancer among the male population worldwide. Vitamin D deficiency has been linked to prostate cancer and its aggressiveness. Herein, we initiated a retrospective study to evaluate vitamin D status and monitoring in veterans with prostate cancer, and to examine the potential link between vitamin D and survival status and length of survival in this population. We found that veterans who were initially vitamin D deficient were significantly less likely to survive than those who were not initially deficient, and that both initial and follow-up vitamin D deficiency were associated with decreased likelihood of survival after prostate cancer diagtiosis. We recomtnend that vitamin D deficiency be replaced in veterans with prostate cancer.
INTRODUCTION Prostate cancer remains the second most commonly diagnosed cancer among the male population worldwide. In the United States, the estimated cost of prostate cancer screening was calculated to be as high as $1.86 billion per year if 50 million men aged 40 to 74 years were to undergo screening.' Crawford et al" estimated that the average 2-year cost for watchful waiting of prostate cancer was $24,809 and for active treatment over the same period was $59,286. Older age, race, family history, obesity, physical activity, smoking, antioxidants, vitamin D, and calcium have been linked to the incidence of prostate cancer and its aggressiveness.^ The theory that vitamin D could have anticancer benefits was suggested from epidemiological studies. However, there has been a great deal of controversy surrounding the protective role of vitamin D in prostate cancer based on small studies. Previous studies determined that vitamin D deficiency is prevalent in veterans, but monitoring and treatment of this deficiency remains suboptimal.'* Vitamin D deficiency tnay have a role in cancer survival rates in both AfricanAmericans and white Americans after consideration of other risk factors such as socioeconomic status, cancer stage at time of diagnosis and treatment.^ Deficiency also appears closely related to increased health care costs.^""^ Peiris et al'° recently described a relationship between
*Department of Internal Medicine, East Tennessee State University, P.O. Byx 70622, Johnson City, TN 37614. tDepartment Family Medicine, East Tennessee State University, P.O. Box 70621, Johnson City, TN 37614. :i:Mountain Home VAMC, Mountain Home, TN 37684. §Sunlight, Nutrition and Health Research Center, P.O. Box 641603, San Francisco, CA 94164-1603. ¡Division of Endocrinology, Mountain Home VAMC, Medicine Service111, Mountain Home, TN 37684. This article does not represent the position of the Department of Veterans Affairs or the U.S. Government. doi : 10.7205/MtLMED-D-12-00540
MILITARY MEDICINE, Vol. 179, January 2014
vitamin D status and monitoring, and survival in veterans diagnosed with bladder cancer. This study was conducted to evaluate how vitamin D status at diagnosis and monitoring might predict outcomes in veterans with prostate cancer across the six Southeastern U.S. Veteran Medical Centers in Veterans Integrated Service Network 9 (VISN 9). We hypothesized that vitamin D status and monitoring parameters would be linked to survival status and length of survival in veterans with prostate cancer irrespective of other risk factors.
MATERIALS AND METHODS This study was conducted at a Veterans Administration facility. The Research and Development committee at the Veterans Affairs (VA) Medical Center as well as the institutional review board at the affiliated university approved procedures and protocol. Patient data from the six VA Medical Centers in VISN 9 (Southeastern United States) between October 1, 1999 and April 30, 2012 were extracted frotn the electronic health records and included in the present study. Initial and subsequent vitamin D levels were extracted and analyzed. Serum 25(OH)D concentration was determined by immunochemiluminometric assay (LabCorp, Burlington, NC), and vitamin D status was examined as a dichotomous variable with deficiency classified as 25(OH)D < 20 ng/tnL." Also extracted were 5-year survival status and days to death (for nonsurvivors). Einally, several background and medical variables were considered: age, race, body mass index (BMI), tobacco use, and latitude of residence. Descriptive analysis was used to describe the study sample. Data analysis examining the association between vitamin D status variables (deficiency and monitoring) and survival utilized /"-analysis, although examination of status against days to death involved t tests. The association between number of vitamin D tests and survival was examined with a correlation coefficient, although the link between number of tests and survival status was examined with a t test. Follow-up
81
Vitamin D and Prostate Cancer Sun'ival in Veterans TABLE \.
TABLE III.
Participattt Characteristics
Regressiott Results Predicting Patient Survival From Vitamin D Status
Mean (SD; Range) or % Background Age (Years) Race (% White) Latitude of Residence (Degrees) Tobacco Products (% Used) Vitamin D Any Testing (%) At Least I Follow-up Test for Those Initially Tested (%) At Least 2 Follow-up Tests for Those Initially Tested (%) Initial Status (% Deficient) Follow-up Status (% Deficient) Patient Outlook Following Diagnosis 5-year Survival Rate (%) Days to Death (Nonsurvivors Qnly)
Predictor 78(9:36-102) 84.6 36.7(1.8; 18-61) 27.5 16.0 56.4 34.2 34.4 19.4 72.1 843(263:0-1,826)
N= 16,535.
testing controlled for background and medical variables that were significantly associated with patient outlook outcomes (age, race, BMI, tobacco use, and latitude of residence), and used logistic regression analysis. RESULTS During the study target period, there were 24,050 patients with a prostate cancer diagnosis. For those who did not survive, the average time to death was 3.5 years (standard deviation = 1.5 years), meaning that if a patient was going to die, he had a 66% chance of dying within 5 years. Thus, the sample was restricted to males who had been diagnosed at least 5 years ago (before April 30, 2007) to accurately examine survival as an outcome. This reduced the sample to 16,535 patients. Characteristics of the sample are presented in Table I. The average participant was in his 70s, Caucasian, and overweight. Over one quarter of participants used tobacco. Vitamin D testing was not common in this sample, with fewer than one in five prostate cancer patients ever tested, and only just over half of those receiving any follow-up testing. Of those tested, over one third were initially deficient, with only one in five deficient at follow-up. Indeed, vitamin D level increased significantly from initial testing to first follow-up test (24.7 vs. 30.5, t = 15.9, p < 0.001) for
TABLE II.
Qutlook Variable
Initially Deficient" (n = 906)
Initially Nondeficient («= 1,731)
Survival (%) Days to Death*
89% 1.084
95% 1.165
Regression Equation Number 1 Background Factors Race Age BMi Tobacco Use Latitude of Residence Initial Vitamin D Status Regression Equation Number 2 Background Factors Race Age BMI Tobacco Use Latitude of Residence Follow-up Vitamin D Status
R
/?' Change
/'
0.18
0.03
0.24
0.03
Vitamin D and Prostate Cancer Survival in Veterans Tatyana Der, MD*; Beth A. Bailey, PhDf; Dima Youssef, MD*; Todd Manning, BAf; William B. Grant, PhD§; Alan N. Peiris, MD, PhD*¡I ABSTRACT Prostate cancer remains the second most commonly diagnosed cancer among the male population worldwide. Vitamin D deficiency has been linked to prostate cancer and its aggressiveness. Herein, we initiated a retrospective study to evaluate vitamin D status and monitoring in veterans with prostate cancer, and to examine the potential link between vitamin D and survival status and length of survival in this population. We found that veterans who were initially vitamin D deficient were significantly less likely to survive than those who were not initially deficient, and that both initial and follow-up vitamin D deficiency were associated with decreased likelihood of survival after prostate cancer diagtiosis. We recomtnend that vitamin D deficiency be replaced in veterans with prostate cancer.
INTRODUCTION Prostate cancer remains the second most commonly diagnosed cancer among the male population worldwide. In the United States, the estimated cost of prostate cancer screening was calculated to be as high as $1.86 billion per year if 50 million men aged 40 to 74 years were to undergo screening.' Crawford et al" estimated that the average 2-year cost for watchful waiting of prostate cancer was $24,809 and for active treatment over the same period was $59,286. Older age, race, family history, obesity, physical activity, smoking, antioxidants, vitamin D, and calcium have been linked to the incidence of prostate cancer and its aggressiveness.^ The theory that vitamin D could have anticancer benefits was suggested from epidemiological studies. However, there has been a great deal of controversy surrounding the protective role of vitamin D in prostate cancer based on small studies. Previous studies determined that vitamin D deficiency is prevalent in veterans, but monitoring and treatment of this deficiency remains suboptimal.'* Vitamin D deficiency tnay have a role in cancer survival rates in both AfricanAmericans and white Americans after consideration of other risk factors such as socioeconomic status, cancer stage at time of diagnosis and treatment.^ Deficiency also appears closely related to increased health care costs.^""^ Peiris et al'° recently described a relationship between
*Department of Internal Medicine, East Tennessee State University, P.O. Byx 70622, Johnson City, TN 37614. tDepartment Family Medicine, East Tennessee State University, P.O. Box 70621, Johnson City, TN 37614. :i:Mountain Home VAMC, Mountain Home, TN 37684. §Sunlight, Nutrition and Health Research Center, P.O. Box 641603, San Francisco, CA 94164-1603. ¡Division of Endocrinology, Mountain Home VAMC, Medicine Service111, Mountain Home, TN 37684. This article does not represent the position of the Department of Veterans Affairs or the U.S. Government. doi : 10.7205/MtLMED-D-12-00540
MILITARY MEDICINE, Vol. 179, January 2014
vitamin D status and monitoring, and survival in veterans diagnosed with bladder cancer. This study was conducted to evaluate how vitamin D status at diagnosis and monitoring might predict outcomes in veterans with prostate cancer across the six Southeastern U.S. Veteran Medical Centers in Veterans Integrated Service Network 9 (VISN 9). We hypothesized that vitamin D status and monitoring parameters would be linked to survival status and length of survival in veterans with prostate cancer irrespective of other risk factors.
MATERIALS AND METHODS This study was conducted at a Veterans Administration facility. The Research and Development committee at the Veterans Affairs (VA) Medical Center as well as the institutional review board at the affiliated university approved procedures and protocol. Patient data from the six VA Medical Centers in VISN 9 (Southeastern United States) between October 1, 1999 and April 30, 2012 were extracted frotn the electronic health records and included in the present study. Initial and subsequent vitamin D levels were extracted and analyzed. Serum 25(OH)D concentration was determined by immunochemiluminometric assay (LabCorp, Burlington, NC), and vitamin D status was examined as a dichotomous variable with deficiency classified as 25(OH)D < 20 ng/tnL." Also extracted were 5-year survival status and days to death (for nonsurvivors). Einally, several background and medical variables were considered: age, race, body mass index (BMI), tobacco use, and latitude of residence. Descriptive analysis was used to describe the study sample. Data analysis examining the association between vitamin D status variables (deficiency and monitoring) and survival utilized /"-analysis, although examination of status against days to death involved t tests. The association between number of vitamin D tests and survival was examined with a correlation coefficient, although the link between number of tests and survival status was examined with a t test. Follow-up
81
Vitamin D and Prostate Cancer Sun'ival in Veterans TABLE \.
TABLE III.
Participattt Characteristics
Regressiott Results Predicting Patient Survival From Vitamin D Status
Mean (SD; Range) or % Background Age (Years) Race (% White) Latitude of Residence (Degrees) Tobacco Products (% Used) Vitamin D Any Testing (%) At Least I Follow-up Test for Those Initially Tested (%) At Least 2 Follow-up Tests for Those Initially Tested (%) Initial Status (% Deficient) Follow-up Status (% Deficient) Patient Outlook Following Diagnosis 5-year Survival Rate (%) Days to Death (Nonsurvivors Qnly)
Predictor 78(9:36-102) 84.6 36.7(1.8; 18-61) 27.5 16.0 56.4 34.2 34.4 19.4 72.1 843(263:0-1,826)
N= 16,535.
testing controlled for background and medical variables that were significantly associated with patient outlook outcomes (age, race, BMI, tobacco use, and latitude of residence), and used logistic regression analysis. RESULTS During the study target period, there were 24,050 patients with a prostate cancer diagnosis. For those who did not survive, the average time to death was 3.5 years (standard deviation = 1.5 years), meaning that if a patient was going to die, he had a 66% chance of dying within 5 years. Thus, the sample was restricted to males who had been diagnosed at least 5 years ago (before April 30, 2007) to accurately examine survival as an outcome. This reduced the sample to 16,535 patients. Characteristics of the sample are presented in Table I. The average participant was in his 70s, Caucasian, and overweight. Over one quarter of participants used tobacco. Vitamin D testing was not common in this sample, with fewer than one in five prostate cancer patients ever tested, and only just over half of those receiving any follow-up testing. Of those tested, over one third were initially deficient, with only one in five deficient at follow-up. Indeed, vitamin D level increased significantly from initial testing to first follow-up test (24.7 vs. 30.5, t = 15.9, p < 0.001) for
TABLE II.
Qutlook Variable
Initially Deficient" (n = 906)
Initially Nondeficient («= 1,731)
Survival (%) Days to Death*
89% 1.084
95% 1.165
Regression Equation Number 1 Background Factors Race Age BMi Tobacco Use Latitude of Residence Initial Vitamin D Status Regression Equation Number 2 Background Factors Race Age BMI Tobacco Use Latitude of Residence Follow-up Vitamin D Status
R
/?' Change
/'
0.18
0.03
0.24
0.03
