Intake of Plant Foods and Associated Nutrients in ... - Semantic Scholar

2 downloads 55 Views 101KB Size Report
vs. white flour bread or pasta products), we cannot discern if ..... to be significantly linked to decreased PC risk in this study co- hort. These factors ... Atlanta, GA:.
Nutrition and Cancer, 61(2), 216–224 Copyright © 2009, Taylor & Francis Group, LLC ISSN: 0163-5581 print / 1532-7914 online DOI: 10.1080/01635580802419756

Intake of Plant Foods and Associated Nutrients in Prostate Cancer Risk John E. Lewis Department of Psychiatry and Behavioral Sciences, University of Miami Miller School of Medicine, Miami, Florida, USA

Hosanna Soler-Vil´a Department of Epidemiology and Public Health and Sylvester Comprehensive Cancer Center, University of Miami Miller School of Medicine, Miami, Florida, USA

Peter E. Clark Department of Urologic Surgery, Vanderbilt University Medical Center, Nashville, Tennessee, USA

Laura A. Kresty, Glenn O. Allen, and Jennifer J. Hu Department of Epidemiology and Public Health and Sylvester Comprehensive Cancer Center, University of Miami Miller School of Medicine, Miami, Florida, USA

Plant foods and associated nutrients may impact prostate cancer (PC) risk and survival. Therefore, we compared dietary intake, mainly plant food groups among 382 controls and 478 PC cases (373 incident and 105 prevalent cases). Caucasian controls had significantly higher daily servings of vegetables (3.4 vs. 2.5, P = 0.002) and fruits and/or fruit juices (1.6 vs. 1.3, P = 0.02) compared to African American controls. In Caucasians, incident cases reported lower intake of fiber, vitamin C, vitamin A, α-carotene, β-carotene, cryptoxanthin, folate, genistein, daidzein, and fruits and/or fruit juice than controls and/or prevalent cases. In African Americans, incident cases had lower intake of α-carotene compared to controls and prevalent cases. Reduced PC risk was associated with the highest tertile of cryptoxanthin (OR = 0.51; 95% CI = 0.35–0.75), fiber (OR = 0.56; 95% CI = 0.35–0.89), vitamin C (OR = 0.60; 95% CI = 0.41–0.88), and fruits and/or fruit juices (OR = 0.46; 95% CI = 0.31–0.68), with significant linear trends. Increased risk of PC was associated with the highest tertile of protein (OR = 1.99; 95% CI = 1.05–3.79) and daily servings of grains (OR = 1.99; 95% CI = 1.23–3.22) with significant linear trends. In summary, we demonstrate racial/ethnic differences in dietary intake of plant foods. The significantly higher consumption of protective dietary constituents among prevalent cases compared to incident cases suggests that PC survivors may be amenable to dietary change.

Submitted 2 February 2008; accepted in final form 4 August 2008. Address correspondence to Jennifer J. Hu, PhD, Professor, Department of Epidemiology and Public Health, Associate Director of Cancer Control, Sylvester Comprehensive Cancer Center, University of Miami Miller School of Medicine, 1120 NW 14th Street, CRB 1511, Miami, FL 33136. Phone: 305-243-3356. Fax: 305-243-2997. E-mail: [email protected]

INTRODUCTION Prostate cancer (PC) is the most common non-skin cancer and the second leading cause of cancer deaths among U.S. men, with 186,320 new cases and over 28,000 deaths estimated in 2008 (1). Despite rapid advances in PC research and its treatment, critical questions about etiology remain unanswered. This is particularly true for African Americans, an understudied population with a 56% higher incident rate, higher proportion of advanced stage at diagnosis, and more than twice the mortality compared to Caucasians (2). Several PC risk factors, such as aging, race, family history (FH) of PC, and germline mutations and polymorphisms, are simply unavoidable (3–8). Diet represents 1 of the few potentially modifiable risk factors following a diagnosis of prostate cancer (9,10). Previous studies have suggested that specific foods, macronutrients, and micronutrients are related to PC risk (11,12). Promising results from cohort studies with selenium and vitamin E regarding PC risk led to the Selenium and Vitamin E Cancer Prevention Trial, an ongoing phase III PC prevention study (13,14). The impact of eating fruits and vegetables on PC risk is inconclusive, partially due to dietary assessment error (15). A review of 17 cohort and case-control studies concluded that the evidence had not supported a protective effect of fruits and vegetables against PC (16). However, other research has associated high consumption of all vegetables (17), vitamin C-rich vegetables (18,19), and tomatoes (20) with lower PC risk. Finally, consumption of soy products has been inversely related to prostatespecific antigen (PSA) measures and the risk of PC development and progression (21,22). Moreover, plant-based diets are rich in potentially thousands of food constituents, which either

216

PLANT FOODS AND ASSOCIATED NUTRIENTS IN PROSTATE CANCER RISK

individually or in combination may inhibit cancer-associated pathways (23,24). These seemingly inconclusive findings regarding some of the most promising foods/nutrients, for example, carotenoids, vitamin E, and selenium and PC risk or prognosis, are not surprising (25–27) given the diversity of the human diet. Between 2000 and 2003, the average annual PC incident rate was 60% higher, and the death rate was 140% higher for African Americans compared to Caucasians (2). Because the prevalence of FH of PC is similar between the two populations (28), and excess mortality is only partially explained by stage at diagnosis (1,2), differences in diet may play a role in racial and ethnic disparities in PC outcomes (29). In addition, recent studies have found that fat, calcium, and vegetable intake had different impacts on insulin-like growth factor-I (IGF-I) or IGF binding protein-3, both greater PC risk factors (30–33) for African Americans than Caucasians (33), which may further explain PC health disparities. Thus, the aims of the present study are threefold: 1) identifying potential nutritional factors that contribute to PC risk by comparing dietary intake of healthy controls and recently diagnosed cases (incident cases from this point), 2) comparing dietary intake patterns between African Americans and Caucasians, and 3) exploring dietary intake differences between incident and prevalent cases by race as a proxy indicator of potential change in dietary behavior between the two phases of survivorship. MATERIALS AND METHODS Study Population All participants (N = 860) were recruited from the Departments of Urology and Internal Medicine of the Wake Forest University School of Medicine during 1998–2004 in WinstonSalem, North Carolina. Sequential patient populations were recruited as part of a case-control study of the etiology of PC with a focus on genetic and nongenetic regulation of DNA damage/repair as described previously (4,5,34). Approximately 87% (n = 747) of the subjects were Caucasians, and the remainder were African Americans (n = 113). Two groups of cases were recruited from the Urology Clinic: 1) incident, newly diagnosed, untreated cases (n = 373), and 2) prevalent cases diagnosed with PC within 5 yr and free of cancer or treatments for at least 6 mo prior to study entry (n = 105). Case status was confirmed by medical records and pathology reports. Controls were frequency matched to cases on age within ±5 yr and were cancer-free men who attended the clinic for: 1) routine screening with normal PSA and digital rectal examination (DRE), or 2) follow-up visit for benign urological problems with abnormal PSA or DRE. African American controls were oversampled to increase statistical power. The response rate for cases (both incident and prevalent) and controls was 94% and 83%, respectively. About 83% of incident cases (within 6 mo post-diagnosis) were recruited into the study less than a month after diagnosis. All prevalent cases were recruited between 6 mo to 5 yr

217

post-diagnosis. All subjects consented to participate following review and approval of the study by the institutional review board for the protection of human subjects. Baseline Questionnaire and Dietary Data Collection A self-administered questionnaire collected demographic information such as age, race, height, and weight [used to calculate body mass index (BMI)]; medical history and medication use; smoking history (i.e., “ever smoked more than 100 cigarettes in your entire life”); smoking status (current or former smoker); alcohol consumption; and FH of PC. Men who had at least one first-degree relative with PC were considered to have a positive FH. The 100-item Block food frequency questionnaire (FFQ; Block 98) was used to evaluate food intake. The Block FFQ is a self-administered questionnaire that takes approximately 30 min to complete and has been widely utilized and validated among several populations (35–37). Participants were asked to estimate their frequency of eating various foods and normal portion sizes during the prior year. Each item had 9 frequency options (i.e., “never or less than once per month” to “2 or more times per day”) and 3 choices for portion size. For this study, we examined the dietary intake of total energy, protein, fat, carbohydrates, fiber, daily servings of fruits and vegetables, and various micronutrients. Statistical Analyses Characteristics of PC cases and controls were compared using chi-square or analysis of variance (ANOVA). We also used ANOVA to test differences of nutrient variables among controls, incident, and prevalent cases for the total sample and for African Americans and Caucasians separately. Significant effects were further examined with post hoc pair-wise comparisons to determine between-group differences. Prevalent cases were not included in diet and PC risk association analysis because we anticipated that some prevalent cases might have undergone dietary changes post-diagnosis. Macronutrients, micronutrients, and various plant constituents were either categorized into tertiles or 2 groups (genistein, daidzein) using the cutoff from all controls. Odds ratio (OR) and 95% confidence intervals (CI) were calculated using logistic regression and adjusted for potential confounders including age, education, body mass index, smoking history, FH of PC, and total caloric intake. The significance of the linear trend was also estimated using logistic regression by classifying the categorical dietary variables as continuous variables. We used the likelihood ratio test to determine the fit of the model. Data were analyzed using SPSS 15.0 for Windows (SPSS, Chicago, IL) and P values less than 0.05 (two-sided) were considered statistically significant. RESULTS Table 1 presents a summary of participant characteristics. Prevalent cases were significantly older than incident cases or

218

J. E. LEWIS ET AL.

TABLE 1 Characteristics of study subjectsa Variable

Category

Age (yr) Age (yr)

Mean ± SD ≤ 50 51–60 61–70 ≥ 71 African Americans Caucasians Mean ± SD No Yes No Yes Missing No Yes Missing No Yes Less than high school At least high school Missing ≤6 7 ≥8 Missing

Race BMI Smoking historyc Current smoking status

Benign prostatic hyperplasia

FHd Education

Gleason score

Controls (n = 382)

Incident Cases (n = 373)

Prevalent Cases (n = 105)

62.0 ± 10.7 63 (16.5%) 110 (28.8%) 113 (29.6%) 96 (25.1%) 75 (19.6%) 307 (80.4%) 27.4 ± 4.4 141 (36.9%) 241 (63.1%) 328 (87.2%) 48 (12.8%) 6 203 (53.3%) 178 (46.7%) 1 317 (83.0%) 65 (17.0%) 35 (9.3) 341 (90.7) 6

63.3 ± 8.2 19 (5.1%) 122 (32.7%) 162 (43.4%) 70 (18.8%) 31 (8.3%) 342 (91.7%) 27.2 ± 3.9 117 (31.4%) 256 (68.6%) 304 (84.0%) 58 (16.0%) 11 198 (53.1%) 175 (46.9%) 0 273 (73.2%) 100 (26.8%) 40 (10.9) 326 (89.1) 7 211 (63.4) 86 (25.8) 36 (10.8) 49

66.9 ± 8.1 4 (3.8%) 19 (18.1%) 46 (43.8%) 36 (34.3%) 7 (6.7%) 98 (93.3%) 26.6 ± 4.0 39 (37.1%) 66 (62.9%) 95 (90.5%) 10 (9.5%) 0 60 (57.7%) 44 (42.3%) 1 77 (73.3%) 28 (26.7%) 10 (9.5) 95 (90.5) 0 64 (72.7) 19 (21.6) 5 (5.5) 17

NA

P Valueb