[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
Purchase Options:
[Skip to Content Landing]
Visual Abstract. Effect of a Behavioral Intervention to Increase Vegetable Consumption on Cancer Progression Among Men With Early-Stage Prostate Cancer
Effect of a Behavioral Intervention to Increase Vegetable Consumption on Cancer Progression Among Men With Early-Stage Prostate Cancer
Figure 1.  Recruitment, Randomization, and Patient Flow
Recruitment, Randomization, and Patient Flow

aThe unmet eligibility criteria were not captured as part of the protocol.

bCentralized pathology review was conducted by a single pathologist on tissue specimens collected at diagnosis to determine tumor volume and Gleason scores.

cThe unmet inclusion criteria included baseline prostate-specific antigen level greater than 10 ng/mL (n = 1) and more than 25% biopsy tissue cores positive for cancer (n = 1).

dThe unmet inclusion criteria included baseline prostate-specific antigen level greater than 10 ng/mL (n = 1), more than 25% biopsy tissue cores positive for cancer (n = 2), Gleason score of 7 for men aged 70 years or younger (n = 1), and consumed 6 or more servings per day of fruits and vegetables (n = 1).

Figure 2.  Time to Progression Where Progression Was Defined as Prostate-Specific Antigen Level ≥10 ng/mL, Prostate-Specific Antigen Doubling Time <3 Years, or Upgrading on Follow-up Prostate Biopsy (A) and Upgrading on Follow-up Prostate Biopsy (B)
Time to Progression Where Progression Was Defined as Prostate-Specific Antigen Level ≥10 ng/mL, Prostate-Specific Antigen Doubling Time <3 Years, or Upgrading on Follow-up Prostate Biopsy (A) and Upgrading on Follow-up Prostate Biopsy (B)

A, The median follow-up time was 24.1 months (interquartile range [IQR], 22.7-24.7) for the MEAL intervention group and 24.0 months (IQR, 22.9-24.6) for the control group. B, The median follow-up time was 23.9 months (IQR, 22.4-24.5) for the MEAL intervention group and 23.9 months (IQR, 22.1-24.6) for the control group.

Table 1.  Baseline Characteristics of Participants
Baseline Characteristics of Participants
Table 2.  Dietary Patterns of Study Participantsa
Dietary Patterns of Study Participantsa
1.
Resnick  MJ, Lacchetti  C, Bergman  J,  et al.  Prostate cancer survivorship care guideline: American Society of Clinical Oncology Clinical Practice Guideline endorsement.  J Clin Oncol. 2015;33(9):1078-1085. doi:10.1200/JCO.2014.60.2557PubMedGoogle ScholarCrossref
2.
Brooks  JD, Paton  VG, Vidanes  G.  Potent induction of phase 2 enzymes in human prostate cells by sulforaphane.  Cancer Epidemiol Biomarkers Prev. 2001;10(9):949-954.PubMedGoogle Scholar
3.
Nordström  T, Van Blarigan  EL, Ngo  V,  et al.  Associations between circulating carotenoids, genomic instability and the risk of high-grade prostate cancer.  Prostate. 2016;76(4):339-348. doi:10.1002/pros.23125PubMedGoogle ScholarCrossref
4.
Zu  K, Mucci  L, Rosner  BA,  et al.  Dietary lycopene, angiogenesis, and prostate cancer: a prospective study in the prostate-specific antigen era.  J Natl Cancer Inst. 2014;106(2):djt430. doi:10.1093/jnci/djt430PubMedGoogle Scholar
5.
Frattaroli  J, Weidner  G, Dnistrian  AM,  et al.  Clinical events in prostate cancer lifestyle trial: results from two years of follow-up.  Urology. 2008;72(6):1319-1323. doi:10.1016/j.urology.2008.04.050PubMedGoogle ScholarCrossref
6.
Ornish  D, Lin  J, Chan  JM,  et al.  Effect of comprehensive lifestyle changes on telomerase activity and telomere length in men with biopsy-proven low-risk prostate cancer: 5-year follow-up of a descriptive pilot study.  Lancet Oncol. 2013;14(11):1112-1120. doi:10.1016/S1470-2045(13)70366-8PubMedGoogle ScholarCrossref
7.
Cooperberg  MR, Carroll  PR.  Trends in management for patients with localized prostate cancer, 1990-2013.  JAMA. 2015;314(1):80-82. doi:10.1001/jama.2015.6036PubMedGoogle ScholarCrossref
8.
Bul  M, Zhu  X, Valdagni  R,  et al.  Active surveillance for low-risk prostate cancer worldwide: the PRIAS study.  Eur Urol. 2013;63(4):597-603. doi:10.1016/j.eururo.2012.11.005PubMedGoogle ScholarCrossref
9.
Klotz  L, Vesprini  D, Sethukavalan  P,  et al.  Long-term follow-up of a large active surveillance cohort of patients with prostate cancer.  J Clin Oncol. 2015;33(3):272-277. doi:10.1200/JCO.2014.55.1192PubMedGoogle ScholarCrossref
10.
Tosoian  JJ, Trock  BJ, Landis  P,  et al.  Active surveillance program for prostate cancer: an update of the Johns Hopkins experience.  J Clin Oncol. 2011;29(16):2185-2190. doi:10.1200/JCO.2010.32.8112PubMedGoogle ScholarCrossref
11.
Welty  CJ, Cowan  JE, Nguyen  H,  et al.  Extended followup and risk factors for disease reclassification in a large active surveillance cohort for localized prostate cancer.  J Urol. 2015;193(3):807-811. doi:10.1016/j.juro.2014.09.094PubMedGoogle ScholarCrossref
12.
Hamdy  FC, Donovan  JL, Lane  JA,  et al; ProtecT Study Group.  10-Year outcomes after monitoring, surgery, or radiotherapy for localized prostate cancer.  N Engl J Med. 2016;375(15):1415-1424. doi:10.1056/NEJMoa1606220PubMedGoogle ScholarCrossref
13.
Parsons  JK, Pierce  JP, Mohler  J,  et al.  A randomized trial of diet in men with early stage prostate cancer on active surveillance: rationale and design of the Men’s Eating and Living (MEAL) Study (CALGB 70807 [Alliance]).  Contemp Clin Trials. 2014;38(2):198-203. doi:10.1016/j.cct.2014.05.002PubMedGoogle ScholarCrossref
14.
Parsons  JK, Newman  VA, Mohler  JL, Pierce  JP, Flatt  S, Marshall  J.  Dietary modification in patients with prostate cancer on active surveillance: a randomized, multicentre feasibility study.  BJU Int. 2008;101(10):1227-1231. doi:10.1111/j.1464-410X.2007.07365.xPubMedGoogle ScholarCrossref
15.
National Cancer Institute. Cancer therapy evaluation program instructions and guidelines version 3.0, release 5. https://ctep.cancer.gov/protocolDevelopment/electronic_applications/docs/cdus_ig_3r5.pdf. Published January 11, 2011. Accessed September 12, 2019.
16.
Humphrey  PA, Moch  H, Cubilla  AL, Ulbright  TM, Reuter  VE.  The 2016 WHO Classification of Tumours of the Urinary System and Male Genital Organs-Part B: prostate and bladder tumours.  Eur Urol. 2016;70(1):106-119. doi:10.1016/j.eururo.2016.02.028PubMedGoogle ScholarCrossref
17.
Epstein  JI, Amin  MB, Reuter  VE, Humphrey  PA.  Contemporary Gleason grading of prostatic carcinoma: an update with discussion on practical issues to implement the 2014 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason Grading of Prostatic Carcinoma.  Am J Surg Pathol. 2017;41(4):e1-e7. doi:10.1097/PAS.0000000000000820PubMedGoogle ScholarCrossref
18.
Pierce  JP, Natarajan  L, Caan  BJ,  et al.  Influence of a diet very high in vegetables, fruit, and fiber and low in fat on prognosis following treatment for breast cancer: the Women’s Healthy Eating and Living (WHEL) randomized trial.  JAMA. 2007;298(3):289-298. doi:10.1001/jama.298.3.289PubMedGoogle ScholarCrossref
19.
Prostate Cancer Foundation. Prostate cancer diet. https://www.pcf.org/patient-resources/living-prostate-cancer/prostate-cancer-diet/. Accessed September 12, 2019.
20.
Freedman  LS.  Tables of the number of patients required in clinical trials using the logrank test.  Stat Med. 1982;1(2):121-129. doi:10.1002/sim.4780010204PubMedGoogle ScholarCrossref
21.
Klotz  L, Zhang  L, Lam  A, Nam  R, Mamedov  A, Loblaw  A.  Clinical results of long-term follow-up of a large, active surveillance cohort with localized prostate cancer.  J Clin Oncol. 2010;28(1):126-131. doi:10.1200/JCO.2009.24.2180PubMedGoogle ScholarCrossref
22.
Bang  H, Jung  SH, George  SL.  Sample size calculation for simulation-based multiple-testing procedures.  J Biopharm Stat. 2005;15(6):957-967. doi:10.1080/10543400500265710PubMedGoogle ScholarCrossref
23.
Diggle  PJ, Heagerty  PJ, Liang  KY, Zeger  SL.  Analysis of Longitudinal Data. Oxford, UK: Clarendon Press; 1996.
24.
Thomson  CA, Ravia  J.  A systematic review of behavioral interventions to promote intake of fruit and vegetables.  J Am Diet Assoc. 2011;111(10):1523-1535. doi:10.1016/j.jada.2011.07.013PubMedGoogle ScholarCrossref
25.
Sfanos  KS, Markowski  MC, Peiffer  LB,  et al.  Compositional differences in gastrointestinal microbiota in prostate cancer patients treated with androgen axis-targeted therapies.  Prostate Cancer Prostatic Dis. 2018;21(4):539-548. doi:10.1038/s41391-018-0061-xPubMedGoogle ScholarCrossref
26.
Newton  RU, Kenfield  SA, Hart  NH,  et al.  Intense Exercise for Survival Among Men With Metastatic Castrate-Resistant Prostate Cancer (INTERVAL-GAP4): a multicentre, randomised, controlled phase III study protocol.  BMJ Open. 2018;8(5):e022899. doi:10.1136/bmjopen-2018-022899PubMedGoogle Scholar
27.
Gardner  JR, Livingston  PM, Fraser  SF.  Effects of exercise on treatment-related adverse effects for patients with prostate cancer receiving androgen-deprivation therapy: a systematic review.  J Clin Oncol. 2014;32(4):335-346. doi:10.1200/JCO.2013.49.5523PubMedGoogle ScholarCrossref
28.
Klein  EA, Thompson  IM  Jr, Tangen  CM,  et al.  Vitamin E and the risk of prostate cancer: the Selenium and Vitamin E Cancer Prevention Trial (SELECT).  JAMA. 2011;306(14):1549-1556. doi:10.1001/jama.2011.1437PubMedGoogle ScholarCrossref
29.
Gaziano  JM, Glynn  RJ, Christen  WG,  et al.  Vitamins E and C in the prevention of prostate and total cancer in men: the Physicians’ Health Study II randomized controlled trial.  JAMA. 2009;301(1):52-62. doi:10.1001/jama.2008.862PubMedGoogle ScholarCrossref
30.
Manson  JE, Cook  NR, Lee  IM,  et al; VITAL Research Group.  Marine n-3 fatty acids and prevention of cardiovascular disease and cancer.  N Engl J Med. 2019;380(1):23-32. doi:10.1056/NEJMoa1811403PubMedGoogle ScholarCrossref
31.
Manson  JE, Cook  NR, Lee  IM,  et al; VITAL Research Group.  Vitamin D supplements and prevention of cancer and cardiovascular disease.  N Engl J Med. 2019;380(1):33-44. doi:10.1056/NEJMoa1809944PubMedGoogle ScholarCrossref
32.
Meyskens  FL  Jr, Szabo  E.  Diet and cancer: the disconnect between epidemiology and randomized clinical trials.  Cancer Epidemiol Biomarkers Prev. 2005;14(6):1366-1369. doi:10.1158/1055-9965.EPI-04-0666PubMedGoogle ScholarCrossref
33.
Ioannidis  JPA.  The challenge of reforming nutritional epidemiologic research.  JAMA. 2018;320(10):969-970. doi:10.1001/jama.2018.11025PubMedGoogle ScholarCrossref
34.
Trepanowski  JF, Ioannidis  JPA.  Perspective: limiting dependence on nonrandomized studies and improving randomized trials in human nutrition research: why and how.  Adv Nutr. 2018;9(4):367-377. doi:10.1093/advances/nmy014PubMedGoogle ScholarCrossref
35.
Schoenfeld  JD, Ioannidis  JP.  Is everything we eat associated with cancer? a systematic cookbook review.  Am J Clin Nutr. 2013;97(1):127-134. doi:10.3945/ajcn.112.047142PubMedGoogle ScholarCrossref
Views 11,854
Citations 0
Original Investigation
January 14, 2020

Effect of a Behavioral Intervention to Increase Vegetable Consumption on Cancer Progression Among Men With Early-Stage Prostate Cancer: The MEAL Randomized Clinical Trial

Author Affiliations
  • 1Department of Urology, UC San Diego Moores Comprehensive Cancer Center and VA San Diego Healthcare System, La Jolla, California
  • 2Alliance Statistics and Data Center, Mayo Clinic, Rochester, Minnesota
  • 3Department of Urology, Roswell Park Comprehensive Cancer Center, Buffalo, New York
  • 4Comprehensive Cancer Center, Department of Medicine, The Ohio State University College of Medicine, Columbus
  • 5Department of Pathology, University of California, San Diego, La Jolla
  • 6Division of Urology, Dana-Farber/Brigham and Women’s Cancer Center, Harvard Medical School, Boston, Massachusetts
  • 7Division of Biostatistics and Bioinformatics, Department of Family Medicine and Public Health and UC San Diego Moores Comprehensive Cancer Center, La Jolla, California
  • 8Alliance Central Protocol Operations, University of Chicago, Chicago, Illinois
  • 9Moores Cancer Center, Department of Family Medicine and Public Health, University of California, San Diego, La Jolla
  • 10Department of Urology, Naval Medical Center San Diego, San Diego, California
  • 11Menorah Medical Center, Midwest Oncology Associates, Overland Park, Kansas
  • 12Siteman Cancer Center, Department of Radiation Oncology, Washington University, St Peters, Missouri
  • 13UCSF Helen Diller Family Comprehensive Cancer Center, San Francisco, California
  • 14Memorial Sloan Kettering Cancer Center, New York, New York
  • 15Department of Prevention and Population Sciences, Roswell Park Comprehensive Cancer Center, Buffalo, New York
JAMA. 2020;323(2):140-148. doi:10.1001/jama.2019.20207
Visual Abstract. Effect of a Behavioral Intervention to Increase Vegetable Consumption on Cancer Progression Among Men With Early-Stage Prostate Cancer
Effect of a Behavioral Intervention to Increase Vegetable Consumption on Cancer Progression Among Men With Early-Stage Prostate Cancer
Key Points

Question  Does a behavioral intervention promoting vegetable consumption decrease cancer progression in men with early-stage prostate cancer on active surveillance?

Findings  In this randomized clinical trial that included 478 patients, there was no significant difference in prostate cancer progression over 2 years among men who participated in a counseling program that encouraged consumption of leafy green, carotenoid, and cruciferous vegetables compared with controls (hazard ratio, 0.96).

Meaning  A behavioral intervention that increased vegetable consumption did not significantly reduce the risk of prostate cancer progression among men with early-stage disease.

Abstract

Importance  Guidelines endorsing vegetable-enriched diets to improve outcomes for prostate cancer survivors are based on expert opinion, preclinical studies, and observational data.

Objective  To determine the effect of a behavioral intervention that increased vegetable intake on cancer progression in men with early-stage prostate cancer.

Design, Setting, and Participants  The Men’s Eating and Living (MEAL) Study (CALGB 70807 [Alliance]) was a randomized clinical trial conducted at 91 US urology and medical oncology clinics that enrolled 478 men aged 50 to 80 years with biopsy-proven prostate adenocarcinoma (International Society of Urological Pathology grade group = 1 in those <70 years and ≤2 in those ≥70 years), stage cT2a or less, and serum prostate-specific antigen (PSA) level less than 10 ng/mL. Enrollment occurred from January 2011 to August 2015; 24-month follow-up occurred from January 2013 to August 2017.

Interventions  Patients were randomized to a counseling behavioral intervention by telephone promoting consumption of 7 or more daily vegetable servings (MEAL intervention; n = 237) or a control group, which received written information about diet and prostate cancer (n = 241).

Main Outcomes and Measures  The primary outcome was time to progression; progression was defined as PSA level of 10 ng/mL or greater, PSA doubling time of less than 3 years, or upgrading (defined as increase in tumor volume or grade) on follow-up prostate biopsy.

Results  Among 478 patients randomized (mean [SD] age, 64 [7] years; mean [SD] PSA level, 4.9 [2.1] ng/mL), 443 eligible patients (93%) were included in the primary analysis. There were 245 progression events (intervention: 124; control: 121). There were no significant differences in time to progression (unadjusted hazards ratio, 0.96 [95% CI, 0.75 to 1.24]; adjusted hazard ratio, 0.97 [95% CI, 0.76 to 1.25]). The 24-month Kaplan-Meier progression-free percentages were 43.5% [95% CI, 36.5% to 50.6%] and 41.4% [95% CI, 34.3% to 48.7%] for the intervention and control groups, respectively (difference, 2.1% [95% CI, −8.1% to 12.2%]).

Conclusions and Relevance  Among men with early-stage prostate cancer managed with active surveillance, a behavioral intervention that increased vegetable consumption did not significantly reduce the risk of prostate cancer progression. The findings do not support use of this intervention to decrease prostate cancer progression in this population, although the study may have been underpowered to identify a clinically important difference.

Trial Registration  ClinicalTrials.gov Identifier: NCT01238172

Introduction

Clinical guidelines for prostate cancer, circulated widely in the public domain, endorse the consumption of diets high in micronutrient-enriched vegetables. Drawing on expert opinion, epidemiological studies, and small preclinical experiments,1 these recommendations propose that vegetable-enriched diets may decrease cancer progression and death among prostate cancer survivors.

Indirect evidence suggests that consumption of micronutrient-enriched plants may promote genomic stability, induce expression of cytoprotective enzymes, and decrease the risk of lethal prostate cancer,2-6 but data from randomized clinical trials (RCTs) focused on actionable clinical end points are lacking. RCTs demonstrating the efficacy of vegetable-enriched diets to diminish cancer progression would justify the development of interventions to promote adoption of high-vegetable diets for prostate cancer survivors.

Most patients with prostate cancer present with early-stage disease.7 Active surveillance—in which patients with early-stage disease defer immediate treatment with surgery or radiation and instead pursue monitoring with prostate-specific antigen (PSA) and follow-up prostate biopsies—provides an alternative for many. However, at least 30% of active surveillance patients will progress or undergo definitive treatment within 2 years of follow-up.8-12 Prevention of disease progression in these patients would decrease treatment incidence, improve health-related quality of life, and reduce health care costs. Diet modification is one potential intervention to prevent progression of prostate cancer.

A behavioral intervention that promotes micronutrient-enriched plant intake among patients with prostate cancer was piloted and validated. This intervention uses motivational interviewing techniques and a centralized, telephone-based counseling system to remotely deliver effective behavioral change with high fidelity.13,14

The Men’s Eating and Living (MEAL) Study (CALGB 70807 [Alliance]), a phase 3 study, tested the efficacy of this intervention to prevent disease progression in prostate cancer by increasing micronutrient-enriched vegetable consumption.13 The hypothesis was that this intervention would decrease the incidence of progression among patients with early-stage prostate cancer on active surveillance.

Methods

This study was conducted through the Cancer and Leukemia Group B, now part of the Alliance for Clinical Trials in Oncology, a Clinical Trials Network group of the National Cancer Institute.13 The trial protocol and statistical plan are available in Supplement 1 and Supplement 2, respectively. Each participant signed an institutional review board–approved, protocol-specific informed consent document in accordance with federal and institutional guidelines. This study was monitored twice annually by the Alliance Data and Safety Monitoring Board, a standing committee with members drawn from both within and outside of the Alliance. Data collection was conducted by the Alliance Statistics and Data Management Center (SDMC) and by UC San Diego Moores Comprehensive Cancer Center staff. Data quality was ensured by Alliance SDMC and by the study chairperson following Alliance policies. Patient-reported race/ethnicity, selected from fixed categories, was collected in compliance with National Cancer Institute reporting guidelines.15

From January 2011 to August 2015, patients were recruited from participating urology and medical oncology clinics. Eligible patients were randomized (1:1) to either the telephone-based counseling intervention (MEAL intervention) or a control condition. Instead of the counseling intervention, control patients received a publication on diet for prostate cancer.

Dynamic allocation (minimization method) was used for stratified randomization. Randomization was stratified by age (<70 years vs ≥70 years), race/ethnicity (African American vs other) and time since diagnostic prostate biopsy (0-12 months prior to registration vs >12 and ≤24 months prior to registration).

Study Participants

Eligible patients were 50 to 80 years of age and diagnosed as having stage cT2a or less prostate adenocarcinoma within 24 months of baseline with prostate biopsy (≥10 cores). Patients were required to have International Society of Urological Pathology (ISUP)16,17 grade group 1 (Gleason score 3 + 3 = 6) cancer in those younger than 70 years and group 2 or less (Gleason score ≤3 + 4 = 7) cancer in those aged 70 years and older and serum PSA level less than 10 ng/mL. Additional biopsy volume criteria included less than 25% of total cores and 50% or less of any single core involving adenocarcinoma. A single genitourinary pathologist (D.E.H.) performed centralized pathology review in a blinded manner to confirm eligibility.

Exclusion criteria included prior prostate cancer treatment with surgery, radiation, local ablation, or androgen deprivation therapy; metastatic disease; and baseline consumption of 6 or more servings per day of fruits and vegetables. Consumption of dietary supplements was permitted. Men taking 5-α reductase inhibitors for symptomatic benign prostatic hyperplasia were eligible after discontinuation of drug followed by a 90-day washout period.

Interventions

The telephone counseling intervention, provided through the UC San Diego Moores Comprehensive Cancer Center and previously described, used an approach adopted from social cognitive theory to achieve sustained behavior change.13,14 A similar program had successfully affected behavior change in breast cancer survivors.18 After randomization, each intervention participant was assigned to a counselor, who encouraged consumption of at least 7 daily vegetable-fruit servings (defined as a half-cup of raw or cooked vegetables or fruits or 100% vegetable juice), including at least 2 servings each of cruciferous vegetables and tomatoes.

Details of the intervention structure and content were previously reported13 and are provided in the trial protocol (Supplement 1). Briefly, the intervention was divided into 4 phases. The first phase (6 counseling telephone calls over 1 month) focused on building self-efficacy; the second (4 calls over 2 months) on consolidating the new dietary pattern; the third (4 calls over 4 months) on relapse prevention; and the fourth (8 calls over 16 months) on providing positive feedback and monitoring for declining interest. To ensure fidelity and minimize bias, counselors completed an intensive 80-hour training program and used a computer-assisted coaching protocol.

Control participants did not receive counseling calls and were instead provided printed materials from the Prostate Cancer Foundation encouraging consumption of a vegetable-rich diet. The latest version of these materials is available online.19

Primary Outcome

The protocol-specified primary outcome was time to progression (TTP), defined as the length of time from the date of random assignment to PSA level of 10 ng/mL or greater, PSA doubling time (PSADT) less than 3 years, or pathological progression on follow-up biopsy defined as any of the following: more than 25% of cores positive for cancer; more than 50% of any 1 core positive for cancer; or ISUP grade group above 1 in those younger than 70 years and ISUP grade group greater than 2 in those aged 70 years and older. Participants who died of any cause without experiencing progression were censored at the time of death. Additionally, participants who elected to pursue prostate cancer treatment (ie, surgery, radiation, tissue ablation, or androgen deprivation) during the study despite not meeting the criteria for progression were censored at the time of withdrawal for treatment.

Secondary Outcomes

A prespecified secondary outcome was the incidence of prostate cancer treatment among participants who otherwise did not meet the protocol-defined criteria for progression. Other secondary outcomes not reported in this article included the Personal Habits Questionnaire, the Functional Assessment of Cancer Therapy Scale–Prostate, the Memorial Anxiety Scale for Prostate Cancer, the International Prostate Symptom Score, the Expanded Prostate Cancer Index Composite 26, the Nutrition Self-Efficacy Scale, and Satisfaction with the MEAL Program (Supplement 1). These results will be reported in subsequent articles.

Preplanned Correlative Science Outcomes

Preplanned correlative science outcomes collected to assess fidelity to the intervention were diet composition and plasma carotenoid concentrations. For diet composition assessment, each participant was scheduled for a series of 4 separate diet assessment interviews (randomly selected and distributed evenly between weekends and weekdays) at baseline, 12-month follow-up, and 24-month follow-up. Each set of interviews was conducted over a 3-week period using the Nutrition Data System for Research software and nutrient database (NDS-R, version 2010, University of Minnesota Nutrition Coordinating Center, University of Minnesota, Minneapolis). At least 3 complete interviews at each time point were required for a valid diet assessment. Plasma carotenoids were measured from fasting blood samples collected at baseline and 12-month follow-up using high-performance liquid chromatography.13,14

Primary Outcome Evaluation

PSA was evaluated every 3 months starting from baseline. PSA assays were performed in the clinical laboratory of each study site using standard platforms; results were entered into the patient registration system. PSADT was calculated at the Alliance SDMC as the natural logarithm of 2 divided by the slope (the least squares estimator) of log (PSA) observations over time. The first 3 PSA measurements were used at the 6-month assessment (ie, at baseline and at months 3 and 6). From the 9-month assessment onward, all available PSA measurements were used to calculate PSADT, as long as the patients had at least 3 measurements.

Participants who did not receive definitive treatment underwent an end-of-study biopsy 24 months after baseline. Additional for-cause biopsies were performed as clinically indicated at the discretion of the treating physician.

Sample Size Calculation

The log-rank test with a 2-sided α = 5% and a sample size of 418 had at least 80% power to detect a difference in progression rate of 20% in the control vs 10% in the experimental group during the 24-month follow-up period. Fifty-seven events in the 2 groups combined were required to achieve the specified power for the test comparing TTP. Freedman’s formula for the required number of events was used.20

A 24-month progression rate of 20% in the control group was conservatively estimated. Because most patients who progress undergo treatment, it was determined that a 10% decrease in progression represented a clinically meaningful difference.21 Under the exponential distribution assumption for TTP, a 24-month progression rate of 20% vs 10% in the intervention group corresponds to a hazard ratio (HR) of 2.1. Assuming a 10% dropout rate, targeted enrollment was 464 participants. A conservative assumption was made that there would be a 10% dropout rate. In the pilot study, the dropout rate was 2%.14

Statistical Methods

Statistical analyses were conducted by the Alliance SDMC on a database frozen on April 9, 2018. The primary analysis set included all randomized participants; participants later deemed ineligible by review of eligibility criteria or centralized pathology of baseline prostate biopsy specimens were excluded from the primary analysis. Participants who did not have progression documented during the study period were censored at the date of the last follow-up. Participants who died of any cause without experiencing progression were censored at the time of death. Participants who elected to pursue treatment despite not meeting progression criteria were censored at the time of withdrawal for treatment.

TTP was analyzed using the Kaplan-Meier method. The log-rank test was used to determine superiority of the intervention group compared with the control group. A Cox proportional hazards model was used to estimate the HR and 95% CIs for the group comparison. As supportive analyses, Cox proportional hazards regression modeling was applied to adjust for the 3 stratification factors, and the primary analysis was repeated using the intention-to-treat analysis population, defined as all randomized participants.13,22 The Kolmogorov-type supremum test was applied to test the proportional hazards assumption for the treatment variable as well as graphical checks; based on these methods, the assumption of proportional hazards was satisfied (P = .46). As a protocol-defined sensitivity analysis, the primary analysis was repeated based on the Kaplan-Meier method in which progression was defined only as the Gleason score on repeat biopsy. In a post hoc analysis, a shared frailty model extension of the Cox proportional hazards regression model was fit to account for within-site homogeneity.

Time to treatment was defined as the length of time from the date of random assignment to receipt of alternative therapy; participants who died of any cause without receiving alternative therapy were censored at the time of death. Treatment-free percentages and intervention effect comparison were obtained from the Kaplan-Meier method and log-rank test, respectively. Additionally, the numbers and percentages of participants who pursued prostate cancer treatment were determined. A Fisher exact test of independence was conducted to assess whether the proportions were different between the study groups.

For the correlative end points of diet composition, mean (95% CI) daily intake of target dietary variables and dietary carotenoids at each study visit (baseline, 12 months, and 24 months) was calculated by study group. Linear mixed-effects models, which allowed for missing outcome data under the reasonable assumption of missing at random, were used to assess longitudinal changes for each targeted dietary variable.23 A major advantage of this approach is that partially complete records can be included in the analysis, thus avoiding the biases of completers-only analyses. In the models, repeated measures for each participant’s dietary intake at baseline, 12-month follow-up, and 24-month follow-up comprised the dependent variable. Study group, time, and group-by-time interactions were included as fixed-effect independent variables. A participant-specific intercept was included to account for between-person variation. Appropriate contrast terms were used to compare dietary changes within and between groups at each time point. Likelihood ratio tests and Wald statistics were used to test statistical significance of fixed effects. A statistically significant group-by-time interaction indicated that dietary change profiles differed between study groups. A similar mixed model was developed for the plasma carotenoid data. Residual plots were used to check model assumptions, with a log-transformation applied, as needed, to better approximate gaussian residuals. To further assess the robustness of the results, in sensitivity analysis, a series of alternative imputation models for the missing diet data (using the SAS PROC MI function), including multiple imputation, were considered. The differences in the results were negligible and, therefore, the reported results are based on the linear mixed-effects model.

Because of the potential for type I error due to multiple comparisons, findings for analyses of secondary, correlative, and other end points should be interpreted as exploratory. P values are 2-sided and P ≤ .05 was deemed statistically significant.

Results

From 2011 to 2015, 478 participants were randomized at 91 study sites: 237 to the intervention and 241 to the control group (Figure 1). The number of participants deemed ineligible in the intervention and control groups was 11 (2 based on eligibility review and 9 based on central pathology review) and 24 (5 based on eligibility review and 19 based on central pathology review), respectively. Therefore, the primary analysis set included 443 participants: 226 intervention and 217 control (Figure 1). Of these 443 participants, 183 (81.7%) and 171 (79.5%) in the intervention and control groups, respectively, met the study’s per-protocol criteria. The reasons for not completing the protocol were withdrawal (intervention group: 36 [15.9%]; control group: 34 [15.8%]), alternative therapy (intervention group: 4 [1.8%]; control group: 4 [1.9%]), death (intervention group: 1 [0.4%]; control group: 3 [1.4%]); and other, including missing (intervention group: 2 [0.9%]; control group: 5 [2.3%]).

Baseline characteristics were balanced across the 2 groups (Table 1). The mean (SD) age was 64 (7) years and the mean (SD) PSA level was 4.9 (2.1) ng/mL. Fifty participants (11%) were African American. There were no clinically significant differences between groups for age, race/ethnicity, region, time elapsed since prostate cancer diagnosis, PSA level, clinical stage, or tumor grade. A total of 9 participants (2%) had ISUP grade group 2 disease. The median number of PSA values collected for each participant was 7 in the intervention and 7 in the control group; 48.2% (95% CI, 41.7%-54.7%) in the intervention group and 46.5% (95% CI, 39.9%-53.1%) in the control group underwent the 24-month follow-up biopsy.

Primary Outcome

There were 245 progression events (intervention: 124; control: 121). There were 4 deaths (0.9%) during the study (intervention group: 1 [0.4%]; control group: 3 [1.4%]), and these participants were censored at the time of death. Five participants (1.1%) (intervention group: 3 [1.3%]; control group: 2 [0.9%]) withdrew from the study to pursue alternative treatment without experiencing clinical progression and, therefore, were censored at the time of study withdrawal. There was no significant difference in TTP comparing intervention vs control (unadjusted HR, 0.96 [95% CI, 0.75 to 1.24], P = .76; adjusted HR, 0.97 [95% CI, 0.76 to 1.25], P = .84) (Figure 2A), including the shared frailty model (unadjusted HR, 0.96 [95% CI, 0.75 to 1.24]; P = .76). The progression-free Kaplan-Meier percentages for the intervention and control groups were 43.5% (95% CI, 36.5% to 50.6%) and 41.4% (95% CI, 34.3% to 48.7%), respectively (difference, 2.1% [95% CI, −8.1% to 12.2%]).

There was no significant between-group difference for TTP by tumor grade on repeat biopsy (HR, 1.40 [95% CI, 0.79 to 2.46]; P = .24) (Figure 2B). There were 49 such events, including 28 in the intervention group and 21 in the control group. The progression-free Kaplan-Meier percentages for the intervention and control groups were 89.9% (95% CI, 84.4% to 94.1%) and 90.2% (95% CI, 84.4% to 94.8%), respectively (difference, −0.3% [95% CI, −7.3% to 6.7%]).

Secondary Outcomes

Eight participants withdrew from the study to pursue active treatment (intervention group: 4 [1.8%]; control group: 4 [1.9%]). Two additional participants in the intervention group pursued active treatment, although they completed the intervention per-protocol criteria. The type of active treatment pursued was not recorded. There was no significant difference in the total number of patients who pursued active treatment during the 24-month follow-up (intervention group: 6 [2.7%]; control group: 4 [1.8%]; P = .75). The HR for time to treatment was 1.38 (95% CI, 0.39 to 4.90; P = .61).

Correlative Outcomes

Table 2 shows the results of the diet composition analyses. At baseline, there were no significant between-group differences. At 12-month follow-up, intervention participants (n = 208) reported significant increases compared with controls (n = 199) in daily total vegetable servings (mean change, 2.43 [95% CI, 2.17 to 2.69] vs 0.45 [95% CI, 0.19 to 0.70]; P < .001), cruciferous servings (mean change, 43.10 g/d [95% CI, 35.21 to 50.99] vs 6.44 g/d [95% CI, −1.39 to 14.26]); P < .001), and total carotenoids (mean change, 13 839.31 µg/d [95% CI, 11 960.15 to 15 718.48] vs 2030.79 µg/d [167.32 to 3894.25]; P < .001). At 24-month follow-up, significant between-group differences (intervention group: 190; control group: 185) persisted for total vegetable servings (mean change, 2.01 [95% CI, 1.75 to 2.28] vs 0.37 [95% CI, 0.11 to 0.62]; P < .001), cruciferous servings (mean change, 0.50 [95% CI, 0.38 to 0.61] vs 0.01 [95% CI, −0.1 to 0.12]; P < .001), and total carotenoids (9687.87 µg/d [95% CI, 7801.18 to 11574.57] vs 324.73 µg/d [95% CI, −1543.61 to 2193.07]; P < .001). Additional diet composition data are presented in the eTable and eFigure in Supplement 3.

Plasma carotenoids, measured at 12-month follow-up in a sample of participants as a biomarker of vegetable intake, validated the self-reported dietary data. There were no significant baseline differences in mean levels of the log-transformed carotenoid values between groups: 0.41 log-μmol/L (95% CI, 0.34-0.47) in the intervention group (n = 152) vs 0.40 log-μmol/L (95% CI, 0.34-0.47) in the control group (n = 142). At 12 months, the mean levels of the log-transformed carotenoid values between groups were 0.64 (95% CI, 0.56-0.72) in the intervention group vs 0.52 (95% CI, 0.45-0.60) in the control group, with a significant between-group difference in baseline to 12-month changes (intervention vs control difference in mean change scores, 0.1 [95% CI, 0.02-0.18]; P = .01).

Discussion

The behavioral intervention in this study produced robust, sustained increases in carotenoid, cruciferous-rich, and leafy green vegetable intake for 2 years, but did not significantly reduce the risk of clinical progression compared with control in patients with early-stage prostate cancer on active surveillance. These data fail to support prevailing assertions in evidence-based clinical guidelines and the popular media that diets high in micronutrient-enriched vegetables improve cancer-specific outcomes among prostate cancer survivors.

To our knowledge, this study is the first RCT to test the effect of a dietary intervention for prostate cancer and demonstrates that scalable, robust behavior modification is feasible in this population. The magnitudes of the dietary changes, which persisted through 24 months of follow-up, were substantial and suggest a clinically meaningful effect.24 While behavior modifications of diet or activity may potentially benefit patients with more advanced stages of disease,25-27 RCTs in this population would be needed to test this hypothesis.

These results are consistent with 2 RCTs, involving nearly 50 000 men, that observed that selenium and vitamin C supplements did not alter, and vitamin E supplements modestly increased, the risk of incident prostate cancer,28,29 as well as with an RCT of more than 12 000 men that demonstrated lack of efficacy of vitamin D or marine n-3 (omega-3) fatty acids to prevent prostate cancer.30,31 These results also failed to support the hypothesis that dietary interventions of macronutrients may have greater efficacy than those of micronutrients.32

Nevertheless, enthusiasm for diet-based cancer interventions remains high, driven by assumptions of causality made from epidemiological data.33 Inherent limitations in these data include small effect sizes and substantial confounding.33-35 The overdependence of prostate cancer nutrition guidelines on observational studies with uncertain clinical validity suggests a need to shift nutritional research toward definitive RCTs.

Limitations

This study has several limitations. First, biopsy undersampling of tumors at baseline may have led to misclassification of baseline Gleason grade. Second, somatic gene expression testing of biopsy tissue was not used as an additional method for determining suitability for active surveillance. Third, magnetic resonance imaging–guided biopsy was not routinely used. However, randomization would have minimized the potential for differential bias, and protocol-defined biopsy methodology was consistent with standard-of-care practices.

Conclusions

Among men with early-stage prostate cancer managed with active surveillance, a behavioral intervention to increase vegetable consumption did not significantly reduce the risk of prostate cancer progression. The findings do not support use of this intervention to decrease prostate cancer progression in this population, although the study may have been underpowered to identify a clinically important difference.

Back to top
Article Information

Accepted for Publication: November 19, 2019.

Corresponding Author: J. Kellogg Parsons, MD, MHS, UC San Diego Moores Comprehensive Cancer Center, 9400 Campus Point Dr, MC7987, La Jolla, CA 92093 (jkparsons@ucsd.edu).

Author Contributions: Dr Parsons had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Pierce and Marshall contributed equally to the study.

Concept and design: Parsons, Mohler, Paskett, Kibel, Hahn, Taylor, Stroup, Small, Pierce, Marshall.

Acquisition, analysis, or interpretation of data: Parsons, Zahrieh, Hansel, Kibel, Liu, Seisler, Natarajan, White, Hartman, Stroup, Van Veldhuizen, Hall, Morris, Pierce, Marshall.

Drafting of the manuscript: Parsons, Zahrieh, Hansel, Seisler, Pierce, Marshall.

Critical revision of the manuscript for important intellectual content: Parsons, Mohler, Paskett, Kibel, Liu, Natarajan, White, Hahn, Taylor, Hartman, Stroup, Van Veldhuizen, Hall, Small, Morris, Pierce, Marshall.

Statistical analysis: Parsons, Zahrieh, Liu, Seisler, Natarajan, White, Marshall.

Obtained funding: Parsons, Mohler, Hahn, Marshall.

Administrative, technical, or material support: Mohler, Paskett, Hansel, Kibel, Hahn, Taylor, Hartman, Stroup, Van Veldhuizen, Morris, Pierce, Marshall.

Supervision: Parsons, Mohler, Kibel, Hahn, Small, Pierce.

Conflict of Interest Disclosures: Dr Parsons reported receiving grants from the National Cancer Institute (NCI) and personal fees from Endocare and INSIGHTEC outside the submitted work and owning stock in Pfizer and Johnson & Johnson. Dr Mohler reported receiving grants from the NCI, US Department of Defense, and the Prostate Cancer Foundation during the conduct of the study and grants from the NCI, Department of Defense, and the National Institute of Diabetes and Digestive and Kidney Diseases and personal fees from the National Comprehensive Cancer Network outside the submitted work. Dr Paskett reported receiving grants from the NCI during the conduct of the study and grants from the Merck Foundation outside the submitted work, and owning stock in Pfizer. Dr Kibel reported receiving personal fees from Janssen, Bristol-Myers Squibb, Merck, INSIGHTEC, and Profound outside the submitted work. Dr Hahn reported receiving personal fees from Via Oncology outside the submitted work. Dr Stroup reported receiving CALGB 70807 funding paid through the Henry Jackson Foundation. Dr Small reported receiving personal fees from Janssen Pharmaceuticals, Harpoon Therapeutics, Fortis Therapeutics, and BeiGene Therapeutics outside the submitted work. Dr Morris reported receiving personal fees from Blue Earth Diagnostics, Advanced Accelerator Applications, Tokai Pharmaceuticals, and ORIC Pharmaceuticals outside the submitted work and travel funding from Bayer, Endocyte, and Fujifilm, as well as funding paid to his institution for his participation in clinical trials for Bayer, Endocyte, Progenics Pharmaceuticals, Corcept Therapeutics, Roche, and Janssen. Dr Pierce reported receiving grants from the NCI and Alliance Cooperative Group during the conduct of the study. No other disclosures were reported.

Data Sharing Statement: See Supplement 4.

Funding/Support: This study was supported by the NCI of the National Institutes of Health under awards UG1CA189823 (Alliance for Clinical Trials in Oncology NCORP grant; https://acknowledgments.alliancefound.org), U24CA196171, U10CA037447, U10CA011789, U10CA041287, U10CA059518, U10CA077651, U10CA077658, U10CA138561, U10CA180791, U10CA180850, U10CA180866, U10CA180830, and U10CA180888 (SWOG); NCI grant 1R01 CA132951-01A1; Department of Defense grant PC073412; the Prostate Cancer Foundation; and William Hamilton Fund.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views or official policy of the National Institutes of Health, Department of the Navy, Department of Defense, or the US government.

Additional Contributions: We acknowledge the contributions of Vicky Newman, MS, RDN, and Leslie Barbier, BS, for their management of the MEAL UC San Diego Counseling Team and Shirley Flatt, MS, for statistical and data management support for the dietary measurements. Mss Newman, Barbier, and Flatt received compensation for their roles in the study as staff of the University of California, San Diego. We also acknowledge the contributions of Janet M. Koball, BS, and Elizabeth M. Storrick, BA, for data management and statistical programming support, respectively; Mss Kobal and Storrick received compensation for their roles as staff within the Alliance Statistics and Data Management Center. We also acknowledge the MEAL UC San Diego Counseling Team, Alliance Central Protocol Operations, and the patients who participated in the trial.

References
1.
Resnick  MJ, Lacchetti  C, Bergman  J,  et al.  Prostate cancer survivorship care guideline: American Society of Clinical Oncology Clinical Practice Guideline endorsement.  J Clin Oncol. 2015;33(9):1078-1085. doi:10.1200/JCO.2014.60.2557PubMedGoogle ScholarCrossref
2.
Brooks  JD, Paton  VG, Vidanes  G.  Potent induction of phase 2 enzymes in human prostate cells by sulforaphane.  Cancer Epidemiol Biomarkers Prev. 2001;10(9):949-954.PubMedGoogle Scholar
3.
Nordström  T, Van Blarigan  EL, Ngo  V,  et al.  Associations between circulating carotenoids, genomic instability and the risk of high-grade prostate cancer.  Prostate. 2016;76(4):339-348. doi:10.1002/pros.23125PubMedGoogle ScholarCrossref
4.
Zu  K, Mucci  L, Rosner  BA,  et al.  Dietary lycopene, angiogenesis, and prostate cancer: a prospective study in the prostate-specific antigen era.  J Natl Cancer Inst. 2014;106(2):djt430. doi:10.1093/jnci/djt430PubMedGoogle Scholar
5.
Frattaroli  J, Weidner  G, Dnistrian  AM,  et al.  Clinical events in prostate cancer lifestyle trial: results from two years of follow-up.  Urology. 2008;72(6):1319-1323. doi:10.1016/j.urology.2008.04.050PubMedGoogle ScholarCrossref
6.
Ornish  D, Lin  J, Chan  JM,  et al.  Effect of comprehensive lifestyle changes on telomerase activity and telomere length in men with biopsy-proven low-risk prostate cancer: 5-year follow-up of a descriptive pilot study.  Lancet Oncol. 2013;14(11):1112-1120. doi:10.1016/S1470-2045(13)70366-8PubMedGoogle ScholarCrossref
7.
Cooperberg  MR, Carroll  PR.  Trends in management for patients with localized prostate cancer, 1990-2013.  JAMA. 2015;314(1):80-82. doi:10.1001/jama.2015.6036PubMedGoogle ScholarCrossref
8.
Bul  M, Zhu  X, Valdagni  R,  et al.  Active surveillance for low-risk prostate cancer worldwide: the PRIAS study.  Eur Urol. 2013;63(4):597-603. doi:10.1016/j.eururo.2012.11.005PubMedGoogle ScholarCrossref
9.
Klotz  L, Vesprini  D, Sethukavalan  P,  et al.  Long-term follow-up of a large active surveillance cohort of patients with prostate cancer.  J Clin Oncol. 2015;33(3):272-277. doi:10.1200/JCO.2014.55.1192PubMedGoogle ScholarCrossref
10.
Tosoian  JJ, Trock  BJ, Landis  P,  et al.  Active surveillance program for prostate cancer: an update of the Johns Hopkins experience.  J Clin Oncol. 2011;29(16):2185-2190. doi:10.1200/JCO.2010.32.8112PubMedGoogle ScholarCrossref
11.
Welty  CJ, Cowan  JE, Nguyen  H,  et al.  Extended followup and risk factors for disease reclassification in a large active surveillance cohort for localized prostate cancer.  J Urol. 2015;193(3):807-811. doi:10.1016/j.juro.2014.09.094PubMedGoogle ScholarCrossref
12.
Hamdy  FC, Donovan  JL, Lane  JA,  et al; ProtecT Study Group.  10-Year outcomes after monitoring, surgery, or radiotherapy for localized prostate cancer.  N Engl J Med. 2016;375(15):1415-1424. doi:10.1056/NEJMoa1606220PubMedGoogle ScholarCrossref
13.
Parsons  JK, Pierce  JP, Mohler  J,  et al.  A randomized trial of diet in men with early stage prostate cancer on active surveillance: rationale and design of the Men’s Eating and Living (MEAL) Study (CALGB 70807 [Alliance]).  Contemp Clin Trials. 2014;38(2):198-203. doi:10.1016/j.cct.2014.05.002PubMedGoogle ScholarCrossref
14.
Parsons  JK, Newman  VA, Mohler  JL, Pierce  JP, Flatt  S, Marshall  J.  Dietary modification in patients with prostate cancer on active surveillance: a randomized, multicentre feasibility study.  BJU Int. 2008;101(10):1227-1231. doi:10.1111/j.1464-410X.2007.07365.xPubMedGoogle ScholarCrossref
15.
National Cancer Institute. Cancer therapy evaluation program instructions and guidelines version 3.0, release 5. https://ctep.cancer.gov/protocolDevelopment/electronic_applications/docs/cdus_ig_3r5.pdf. Published January 11, 2011. Accessed September 12, 2019.
16.
Humphrey  PA, Moch  H, Cubilla  AL, Ulbright  TM, Reuter  VE.  The 2016 WHO Classification of Tumours of the Urinary System and Male Genital Organs-Part B: prostate and bladder tumours.  Eur Urol. 2016;70(1):106-119. doi:10.1016/j.eururo.2016.02.028PubMedGoogle ScholarCrossref
17.
Epstein  JI, Amin  MB, Reuter  VE, Humphrey  PA.  Contemporary Gleason grading of prostatic carcinoma: an update with discussion on practical issues to implement the 2014 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason Grading of Prostatic Carcinoma.  Am J Surg Pathol. 2017;41(4):e1-e7. doi:10.1097/PAS.0000000000000820PubMedGoogle ScholarCrossref
18.
Pierce  JP, Natarajan  L, Caan  BJ,  et al.  Influence of a diet very high in vegetables, fruit, and fiber and low in fat on prognosis following treatment for breast cancer: the Women’s Healthy Eating and Living (WHEL) randomized trial.  JAMA. 2007;298(3):289-298. doi:10.1001/jama.298.3.289PubMedGoogle ScholarCrossref
19.
Prostate Cancer Foundation. Prostate cancer diet. https://www.pcf.org/patient-resources/living-prostate-cancer/prostate-cancer-diet/. Accessed September 12, 2019.
20.
Freedman  LS.  Tables of the number of patients required in clinical trials using the logrank test.  Stat Med. 1982;1(2):121-129. doi:10.1002/sim.4780010204PubMedGoogle ScholarCrossref
21.
Klotz  L, Zhang  L, Lam  A, Nam  R, Mamedov  A, Loblaw  A.  Clinical results of long-term follow-up of a large, active surveillance cohort with localized prostate cancer.  J Clin Oncol. 2010;28(1):126-131. doi:10.1200/JCO.2009.24.2180PubMedGoogle ScholarCrossref
22.
Bang  H, Jung  SH, George  SL.  Sample size calculation for simulation-based multiple-testing procedures.  J Biopharm Stat. 2005;15(6):957-967. doi:10.1080/10543400500265710PubMedGoogle ScholarCrossref
23.
Diggle  PJ, Heagerty  PJ, Liang  KY, Zeger  SL.  Analysis of Longitudinal Data. Oxford, UK: Clarendon Press; 1996.
24.
Thomson  CA, Ravia  J.  A systematic review of behavioral interventions to promote intake of fruit and vegetables.  J Am Diet Assoc. 2011;111(10):1523-1535. doi:10.1016/j.jada.2011.07.013PubMedGoogle ScholarCrossref
25.
Sfanos  KS, Markowski  MC, Peiffer  LB,  et al.  Compositional differences in gastrointestinal microbiota in prostate cancer patients treated with androgen axis-targeted therapies.  Prostate Cancer Prostatic Dis. 2018;21(4):539-548. doi:10.1038/s41391-018-0061-xPubMedGoogle ScholarCrossref
26.
Newton  RU, Kenfield  SA, Hart  NH,  et al.  Intense Exercise for Survival Among Men With Metastatic Castrate-Resistant Prostate Cancer (INTERVAL-GAP4): a multicentre, randomised, controlled phase III study protocol.  BMJ Open. 2018;8(5):e022899. doi:10.1136/bmjopen-2018-022899PubMedGoogle Scholar
27.
Gardner  JR, Livingston  PM, Fraser  SF.  Effects of exercise on treatment-related adverse effects for patients with prostate cancer receiving androgen-deprivation therapy: a systematic review.  J Clin Oncol. 2014;32(4):335-346. doi:10.1200/JCO.2013.49.5523PubMedGoogle ScholarCrossref
28.
Klein  EA, Thompson  IM  Jr, Tangen  CM,  et al.  Vitamin E and the risk of prostate cancer: the Selenium and Vitamin E Cancer Prevention Trial (SELECT).  JAMA. 2011;306(14):1549-1556. doi:10.1001/jama.2011.1437PubMedGoogle ScholarCrossref
29.
Gaziano  JM, Glynn  RJ, Christen  WG,  et al.  Vitamins E and C in the prevention of prostate and total cancer in men: the Physicians’ Health Study II randomized controlled trial.  JAMA. 2009;301(1):52-62. doi:10.1001/jama.2008.862PubMedGoogle ScholarCrossref
30.
Manson  JE, Cook  NR, Lee  IM,  et al; VITAL Research Group.  Marine n-3 fatty acids and prevention of cardiovascular disease and cancer.  N Engl J Med. 2019;380(1):23-32. doi:10.1056/NEJMoa1811403PubMedGoogle ScholarCrossref
31.
Manson  JE, Cook  NR, Lee  IM,  et al; VITAL Research Group.  Vitamin D supplements and prevention of cancer and cardiovascular disease.  N Engl J Med. 2019;380(1):33-44. doi:10.1056/NEJMoa1809944PubMedGoogle ScholarCrossref
32.
Meyskens  FL  Jr, Szabo  E.  Diet and cancer: the disconnect between epidemiology and randomized clinical trials.  Cancer Epidemiol Biomarkers Prev. 2005;14(6):1366-1369. doi:10.1158/1055-9965.EPI-04-0666PubMedGoogle ScholarCrossref
33.
Ioannidis  JPA.  The challenge of reforming nutritional epidemiologic research.  JAMA. 2018;320(10):969-970. doi:10.1001/jama.2018.11025PubMedGoogle ScholarCrossref
34.
Trepanowski  JF, Ioannidis  JPA.  Perspective: limiting dependence on nonrandomized studies and improving randomized trials in human nutrition research: why and how.  Adv Nutr. 2018;9(4):367-377. doi:10.1093/advances/nmy014PubMedGoogle ScholarCrossref
35.
Schoenfeld  JD, Ioannidis  JP.  Is everything we eat associated with cancer? a systematic cookbook review.  Am J Clin Nutr. 2013;97(1):127-134. doi:10.3945/ajcn.112.047142PubMedGoogle ScholarCrossref
×