|Home | About | Journals | Submit | Contact Us | Français|
The Western diet has been associated with prostate cancer (PC) incidence as well as risk of disease progression after treatment. Conversely, plant-based diets have been associated with decreased risks. A 6-month pilot intervention trial was conducted to determine whether adoption of a plant-based diet, reinforced by stress reduction, could reduce the rate of prostate specific antigen (PSA) rise, a marker of disease progression, in asymptomatic, hormonally untreated patients experiencing consistently rising PSA levels after surgery or radiation.
A pre-post design was employed in which each patient served as his own control. In the current investigation, we examine: (1) the effect of intervention on potential mediators of disease progression, including body composition and weight-related biomarkers (sex steroid hormones and cytokines), and (2) whether changes in these variables were associated with change in rate of PSA rise. The baseline rate of PSA rise (from the time of post-treatment recurrence to the start of intervention) was ascertained by review of patients’ medical records. Body composition and biomarker assessments were performed at Baseline (prior to intervention), during intervention (3 months), and at the end of intervention (6 months). Changes in body composition and biomarkers were determined and compared with rates of PSA rise over the corresponding time intervals.
Over the course of the intervention period, there was a significant reduction in waist-to-hip ratio (WHR) (p=0.03) and increase in circulating sex hormone binding globulin (SHBG) (p=0.04). The rate of PSA rise decreased when comparing the pre-intervention period (PSA slope=0.059) to the period from 0–3 months (PSA slope=−0.002, p<0.01) and increased slightly, though not significantly, when comparing the period from 0–3 months to the period from 3–6 months (0.029, p=0.43).
These results provide preliminary evidence that adoption of a plant-based diet and stress reduction may result in reduction of central adiposity and improvement of the hormonal milieu in patients with recurrent PC. Changes in the rate of rise in PSA were in the same direction as changes in WHR and opposite those of SHBG, raising the possibility that the effect of the intervention may have been mediated, in part, by these variables.
Carcinoma of the prostate is the most commonly occurring cancer (other than skin cancer) among males in Western populations. In the U.S., 1 man in 6 will develop prostate cancer (PC) in his lifetime.1 Most patients who present with PC receive definitive primary treatment consisting of either surgical removal of the prostate (radical prostatectomy, RP), radiation therapy to the prostate (RT), or surgical removal followed by radiation to the prostatic bed or pelvis. In spite of treatment, about one third of patients will have a biochemically defined recurrence, marked by successive increases in prostate-specific antigen (PSA) levels after a post-treatment nadir (the lowest PSA value observed after RP or RT), within the first ten years.2 These individuals are at increased risk of metastasis formation and premature death; over a third of those with rising PSA will go on to develop metastatic disease within the subsequent five years.2
Typically, the PSA tends to rise exponentially after prostatectomy or radiation therapy, reflecting the gradual, inexorable growth of the cancer in the body.2–4 After local treatment, the rate of PSA rise is the single best predictor of both the probability of and time to development of overt metastatic disease 2,3, as well as of overall survival.5–6 Hormonal therapy is sometimes employed at this point, although there is little evidence that early use significantly improves prognosis except in the subset of patients with pelvic lymph node metastasis. Hormonal therapy also frequently produces side effects including hot flashes, loss of libido, gynecomastia, and loss of bone and muscle mass. Therefore, many physicians employ a strategy of active surveillance.7,8 Yet patients, keenly aware of the paucity of curative treatment options, frequently suffer from severe anxiety as they watch their PSA levels continue to rise. This has motivated a search for novel adjunctive strategies that could retard tumor progression while avoiding the side effects of hormonal therapy.
In searching for new therapeutic options, it is important to begin by examining the possible underlying causes of PC. Dramatic international variations in age-adjusted incidence and mortality rates provide clues to the etiology of PC. For example, Qidong County in China has an incidence rate of only 0.8 per 100,000 men, whereas the rate for African-American men in Atlanta is 102.1 per 100,000, a relative risk of 127.5.9,10 Some of these differences may be accounted for by the much higher level of PSA screening and detection in the U.S. compared with China. Yet Japanese men, like Chinese, also have much lower incidence and mortality rates than Americans. The large difference in rates between the U.S. and Japan has been observed for decades, long before the advent of PSA testing. Upon migration to the U.S., PC rates in Japanese men increase 4–9 fold within the first generation and approximate U.S. rates by the second generation.11–13
Epidemiologic and laboratory investigations suggest that diet may constitute an important set of environmental factors impacting development and progression of PC. As reviewed by Kolonel and colleagues, 16 of 22 studies (14 case-control and 8 cohort) found a positive association of meat intake with PC risk, with 15 showing odds ratios or relative risks of 1.3 or more.14 Similarly, in a review by Chan et al, 12 of 23 studies (14 case-control and nine cohort) found positive associations of dairy foods with PC risk.15 Arachidonic acid, synthesized endogenously from omega-6 fatty acids and also found preformed (in cell membranes) in foods of animal origin, has been shown to stimulate the growth of both LNCaP (hormone-sensitive) and PC3 (hormone-insensitive) cell lines and is as effective as testosterone (T) in stimulating growth of LNCaP cells.16
Conversely, plant foods including whole grains, vegetables, legumes, and fruits, appear to be protective. As summarized in a recent review by Chan and colleagues, 8 of 16 studies (13 case-control and 3 cohort) reported an inverse association of specific or total vegetable intake with PC risk, whereas eight reported no association. None reported increased risk. The strongest protective effects were seen for beans and legumes, nuts, carrots, leafy greens, cruciferous (cabbage family) vegetables , and tomatoes.17 Cruciferous vegetables have been found in two population-based studies to be associated with a reduction in PC incidence.18 Indole-3-carbinol, derived from diets rich in cruciferous vegetables, inhibits the growth of PC3 human PC cells by inducing G1 cell cycle arrest leading to apoptosis, and regulates the expression of apoptosis-related genes.14
Growing evidence further suggests that dietary modification emphasizing, in particular, increased whole grain and vegetable intake (as stressed in macrobiotic and many vegan diets), may influence the course of PC after diagnosis. Adoption of whole grain and vegetable based macrobiotic and vegan diets has been associated with prolonged survival and documented remissions of bone and visceral metastases in men with advanced PC.20 Plant-based diet intervention trials were found to significantly reduce the rate of PSA rise in patients with recurrent disease21 and to lower overall PSA, in comparison with randomized controls, in patients with untreated disease.22
However, the precise biological mechanism mediating the apparent cancer-inhibitory effects of dietary and lifestyle modification is uncertain. One possibility is that these effects resulted from a reduction in body weight. Several recent reviews of the link between obesity and PC have reported that while the relationship is complex, the evidence points to an association of obesity with more aggressive disease, worsened prognosis and higher mortality from PC.23,24 In addition, African Americans, who have as a group been reported to have the highest body weight at time of diagnosis, also have the highest disease related mortality.25 Body mass index has been found in several studies to be associated with risk of recurrence after prostatectomy 26–28 and adoption of plant-based diets, which involve a reduction in intake of total fat and energy, have been marked by weight loss in patients with PC.21,22
In addition, studies have shown that the level of inflammation-related adipokines generally rises with increased adipose tissue mass.23 Therefore, weight reduction or improvements in body composition (such as a reduction in body mass index) may have impacted immune function by altering levels of circulating adipokines such as tumor necrosis factor-alpha (TNF-α) or interleukin-6 (IL-6). Serum levels of IL-6 have been reported to be higher in patients with PC, and it has been argued that serum IL-6 levels could be used as a prognostic indicator (with higher levels corresponding to poorer prognosis).29 Michalaki et al reported higher levels of both IL-6 and TNF-α in patients with PC compared to normal controls. In addition, serum TNF-α levels were significantly higher in subjects with metastatic disease when compared to subjects with local disease and TNF-α levels were also correlated with rising PSA.30
While the literature is highly variable on the topic, it is also clear that sex steroid hormones play a significant role in the progression of PC and PC recurrence. Since the groundbreaking discovery of Charles Huggins in 1941 that hormone ablation therapy can be used to treat metastatic PC, androgen deprivation therapy (ADT) has been a mainstay of PC therapy.31,32 Under normal conditions, prostatic tissue that expresses the androgen receptor requires T or dihydroxytestosterone (DHT), a more potent metabolite of T, to stimulate growth and differentiation. In vivo, androgen stimulation of PC cell growth has been demonstrated in the LNCaP (human PC) cell line.33 Without androgen stimulation, the secretory epithelial cells of the prostate undergo apoptosis, leaving behind only basal epithelial cells (which do not express the androgen receptor).34 It is not surprising, therefore, that cancerous cells of the prostate will also undergo apoptosis when deprived of androgen.
The relationship between body composition, sex steroid hormones, and PC is even more complex. Obese patients tend to have lower serum T levels at time of diagnosis than non-obese patients due to increased peripheral conversion of T to estradiol in adipocytes.35 Free T levels also decrease because estradiol leads to feedback inhibition of the pituitary-hypothalamic axis. While it has not been clearly demonstrated that low androgen levels are associated with increased PC risk, recent studies have demonstrated an association between low prediagnostic serum androgen levels and risk of more aggressive disease.35 Obesity has also been inversely associated with sex hormone binding globulin (SHBG), lower levels of which have been linked with metabolic syndrome as well as with in vitro growth of androgen dependent PC cell lines.36 Thus, the lower circulating T or SHBG levels observed in obese patients may constitute part of the biochemical basis for the associations of obesity with more aggressive presentation at diagnosis, increased risk of recurrence, and higher disease-specific mortality.25,37–39
Previously, in the University of California, San Diego (UCSD) Healthy Men Study – a pilot intervention trial of patients with recurrent PC – we examined the effect of adoption of a plant-based diet and stress reduction on the rate of disease progression. In that study, we found a significant change in the rate of rise in PSA, comparing pre- and post-intervention PSA log slopes and doubling times both within individuals and for the group as a whole. Four of 10 evaluable patients experienced an absolute reduction in their PSA levels over the entire 6-month study. Nine of 10 had a reduction in their rates of PSA rise and an improvement of their PSA doubling times. Median PSA doubling time increased from 11.9 months (pre-intervention) to 112.3 months (end of intervention). For the group as a whole, there was a significant decrease in the rate of PSA rise from pre-intervention to end of intervention (P < .01). These results provide preliminary evidence that adoption of a plant-based diet, in combination with stress reduction, may attenuate disease progression and have therapeutic potential for clinical management of recurrent PC.40
The purpose of the present study is to elucidate the mechanism(s) by which the earlier diet and stress reduction intervention may have led to a change in the rate of disease progression. Given the growing scientific evidence implicating body weight in PC progression, we have chosen to make this the central focus of our investigation. Using secondary data from this trial, we herein describe changes observed over the course of the intervention period in body composition and weight-related biomarkers (sex steroid hormone and cytokine levels). We then proceed to relate these changes to changes observed in PSA doubling time during the same period.
The UCSD Healthy Men Study was a pre-post pilot clinical trial in which each patient served as his own control. Its purpose was to determine whether a plant-based dietary intervention, reinforced by stress reduction, could effect a major dietary change and influence the progression of recurrent PC. In this paper we examine the body composition and levels of several weight-related biomarkers in patients at each of the 3 study timepoints: Baseline, 3 months, and 6 months, and we explore whether the changes in the rates of PSA rise from 0–3 months and 3–6 months track changes in these markers during the corresponding intervals.
Fourteen study-eligible patients were recruited with the assistance of urologists at the UCSD and San Diego Veterans Affairs Medical Centers and community hospitals. All patients provided informed consent before being enrolled in the study. Patients were eligible if they had biopsy-confirmed, operable, invasive PC that was treated by radical prostatectomy or radiation therapy; had rising PSA documented on a minimum of 3 serial tests, each at least one month apart from the others, after achieving post-treatment nadir; had no radiological or pathological evidence of overt metastatic disease since completion of initial local treatment; and had not used any form of hormonal therapy for at least 12 months prior to the last nadir PSA.
Patients participated, along with their spouses or another designated support person, in an intensive 6-month, individual and group-based diet and stress reduction intervention conducted at the Moores UCSD Cancer Center. They were taught to increase intake of whole grains, vegetables, fruit, and legumes, and to decrease meat, dairy, and refined carbohydrates. The intervention included individual dietary counseling as well as a series of 10 three-hour group meetings over the 6-month period (once per week during the first month, once per month during months 2–5, and twice during month 6). At these meetings, patients and spouses/support persons received a hands-on cooking demonstration, were served a healthy meal, and participated in supportive group discussion. During the group meetings, patients were also taught how to practice meditation as well as how to perform several basic yoga and t’ai chi movements. Patients also received telephone calls from the dietitian on a weekly basis throughout the intervention. During these calls, patients were guided in dietary goal setting, problem solving, and self-monitoring. The intervention has been described in detail previously.40
Data collection and assessments, as well as phlebotomy, were performed at the UCSD General Clinical Research Center (GCRC) ambulatory clinic. Medical records were collected and reviewed prior to Baseline to confirm study eligibility and to obtain pre-study PSA and treatment histories, tumor characteristics, and other clinical information. Demographic and identifying data were collected at Baseline only. Assessments of diet, body composition (anthropometric status), physical activity, practice of stress reduction, symptoms, and disease-related quality of life were performed at each study time point. Phlebotomy was also performed at Baseline, 3, and 6 months and was used for determination of sex-steroid hormone, cytokine, and study-period PSA levels.
Body weight and waist and hip circumference were assessed on all study participants at the time of their phlebotomy at Baseline, 3, and 6 months. Height was assessed only once, at Baseline. A trained nutritionist at the UCSD General Clinical Research Center performed these assessments.
Weight was measured to the nearest 0.1 kg and height to the nearest 0.1 cm. BMI was calculated as weight in kilograms divided by the square of height in meters. Waist circumference was measured twice at the midpoint between the lower margin of the ribs and the iliac crest to the nearest 0.5 cm. A third measurement was taken if the initial two readings were more than 2 cm apart. The mean of the closest two measurements was then used to calculate waist circumference. Hip circumference was measured as the maximum circumference, when viewed from the side, around the buttocks posteriorly and symphysis pubis anteriorly.
A licensed GCRC phlebotomist drew blood at each of the 3 study timepoints: Baseline, 3 months, and 6 months. The GCRC Core Lab then performed all assays. Sex steroid hormones, using a kit from RIA Diagnostics System Laboratories DSL (now part of Beckman-Coulter), included: (1) T, (2) DHT, (3) estradiol, and (4) SHBG. Cytokines, using an ELISA test from RND Systems, included: (1) IL-6 and (2) TNF-α.
Pre-study PSA readings were obtained by reviewing patients’ medical records. The rate of PSA rise for the period prior to intervention (covering the period from the end of the post-treatment PSA nadir up to, but not including, Baseline) was derived from pre-study PSA readings. The complete methodology used to ascertain pre-study PSA readings has been described previously.40 Linear regression was used to calculate rates of PSA rise for each patient for the following periods: Pre-study, 0–3 months, and 3–6 months.
The rates of PSA rise at 3 months (reflecting the change in PSA from 0–3 months) and 6 months (reflecting the change in PSA from 3–6 months) were derived from PSA readings (one reading obtained for each timepoint) that were performed at the main UCSD Medical Center Chemistry Lab on serum samples obtained at each timepoint. Intervention period PSA tests were all performed using the Immulite 2000 PSA test kit, a completely automated, ultrasensitive chemiluminescence assay with a sensitivity limit of 0.04 ng/ml (Diagnostic Products Corp., Los Angeles, California).
Descriptive statistics were calculated for patient characteristics at Baseline; body composition, sex-steroid hormone, and cytokine levels (at Baseline, 3 and 6 months), and the rate of PSA rise (Pre-study, 0–3 months, and 3–6 months). Indicators of body composition, as well as levels of each of the sex steroid hormones and cytokines, at each study timepoint were compared (Baseline vs. 3 months, 3 months vs. 6 months, and Baseline vs. 6 months) using the Wilcoxon signed-rank test for paired data. The rate of PSA rise was compared across periods (Pre-study vs. 0–3 months, 0–3 months vs. 3–6 months, and Pre-study vs. 0–6 months), also using the Wilcoxon signed-rank test. Comparisons were conducted at the alpha = 0.05 level of significance. All analyses were conducted using SAS® (version 8.01, SAS Institute Inc., 2000).
Fourteen recurrent PC patients were enrolled; 1 patient withdrew and 2 initiated hormone therapy prior to the start of the intervention and another withdrew after three months. Table 1 shows the baseline characteristics for the 10 patients who were: hormonally naïve (and therefore whose PSA data was not influenced by the use of hormonal therapy), completed the intervention, and were included in all subsequent analyses. Study participants were predominantly older (median age=70 years), of normal or slightly overweight (median weight=79.5 kg; median height=179.3 cm). Pathological characteristics of their cancers (median Gleason score=7; median tumor stage=T2) were consistent with those of other patients with recurrent disease. Individual-level clinical data for each patient were presented previously.40
Table 2 shows the change, over the course of the intervention, in indicators of body weight and adiposity as well as in weight-related biomarkers. At baseline, the median BMI was 24.3 kg/m2 and the median waist-to-hip ratio (WHR) was 0.91. By 6 months, median BMI had dropped to 23.9 and WHR to 0.87. The decline in WHR was statistically significant (p=0.027) although the decrease in BMI was not.
At Baseline, the median serum level of T was 3.7 pg/ml. By 6 months, it had decreased slightly to 3.6 pg/mL. At Baseline, DHT was 900.4 pg/mL and estradiol was 32.1 ng/mL. At 6 months, their levels had increased to 1004.7 pg/mL and 33.9 ng/mL, respectively. Neither of these changes was significant. Finally, SHBG increased significantly from 138.5 nmol/L to 155.4 nmol/L (p=0.037).
Finally, also at Baseline, the median serum levels of Interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-a) were 2.6 pg/mL and 1.7 pg/mL, respectively. At 6 months, median IL-6 had decreased by 53% to 1.8 pg/mL. However, TNF-a increased to 1.9 pg/mL. Neither of these changes was significant.
As seen in Figure 1, median levels of 3 key variables, weight, WHR, and SHBG all changed over the course of the intervention. Median body weight decreased significantly during the period 0–3 months from 79.5 kg to 76.1 kg (p<0.01). However, it increased slightly, though non-significantly, to 76.5 kg, during the period 3–6 months (p=0.57)(data not shown). WHR continued to decrease throughout the intervention period, from 0.91 at Baseline to 0.89 at 3 months and then to 0.87 by 6 months. The change from 0–3 months was significant (p=0.01) but the change from 3–6 months was not (p=0.70)(data not shown). Finally, SHBG continued to increase throughout the intervention period, from 138.5 nmol/L at Baseline to 143.3 nmol/L at 3 months and then to 155.3 nmol/L by 6 months. Similarly to what was observed for body weight and WHR, the change in SHBG from 0–3 months was significant (p=0.02) but the change from 3–6 months was not (p=0.43)(data not shown).
Figure 1 also shows the rates of PSA rise at Baseline (representing the period, the length of which varied by patient, from which the time of PSA rise was first detected until the start of the study), 0–3 months, and 3– months. Individual-level data on patients’ rates of PSA rise is provided elsewhere (36). The median (range) rate of PSA rise for Pre-study was 0.059 (0.014 – 0.129). From 0–3 months, the median (range) rate of PSA rise was −0.002 (−0.096 – 0.079), representing a significant decrease from the rate during the pre-study period (p < 0.01). The negative value indicates a median reduction in absolute PSA. From 3–6 months, the median (range) rate of PSA rise was 0.029 (−0.067 – 0.136), representing a non-significant increase when compared to the 0–3 month period (p=0.4316).
Our study investigated two related issues. First, we examined the degree to which patients with recurrent PC, enrolled in an intensive diet and stress reduction intervention, experienced changes in body composition as well as circulating sex steroid hormones and cytokines. Second, we explored whether the changes observed in these biomarkers tracked changes observed in the rate of PSA rise, a marker of disease progression in recurrent PC.
The intervention resulted in a statistically significant decrease in median WHR, suggesting a corresponding decrease in central adiposity. Median weight also decreased significantly during the first half of the intervention (months 0–3) but increased slightly during the second half (months 3–6). In a related manner, we observed a significant increase in median circulating SHBG levels over the course of the intervention. SHBG levels have been consistently observed to be lower in obese individuals and to increase as a result of weight loss interventions.36,41 In fact, it has been suggested that among markers of the sex steroid hormone profile, SHBG concentration may be the most predictably responsive to changes in weight.42
We also found that as adiposity decreased and SHBG increased, the rate of rise in PSA – an indicator of disease progression – decreased, raising the interesting possibility that PSA may have tracked changes in these biomarkers. During the first 3 months of the intervention, as both median WHR and body weight declined significantly, the median rate of PSA rise not only declined but became negative, reflecting a slight reduction in absolute PSA and possibly disease regression in patients with absolute reductions. Conversely, during the second 3 months of the intervention, when median body weight increased (though not significantly), median PSA began to rise again, albeit more slowly than during the period prior to Baseline.
These observations may offer some interesting clues to the ways the intervention influenced disease progression. For example, they tend to underscore the proposition 36,41–43 that PC may constitute a manifestation of metabolic syndrome, a common condition strongly associated with obesity and central adiposity. Metabolic syndrome is also marked by a reduction in SHBG, lower levels of which have also been associated with PC incidence.44 Hyperinsulinemia, resulting from insulin resistance, suppresses liver production of SHBG and also results in a decrease in gonadal T production.42 Most circulating T is bound by either SHBG or albumin while only about 2% of T is unbound or free (non-protein-bound) and available to enter cells and bind with androgen receptors (AR’s).44 While albumin is relatively constant, there can be considerable short-term variation in SHBG levels, particularly in response to changes in weight or adiposity. Thus, SHBG is the single most important factor in determining the relative proportions of bound vs. unbound T; as SHBG increases, the proportion of unbound T declines.45,46 Upon entering cells, unbound T is converted, via the enzyme 5-α-reductase, into DHT.44 DHT, in turn, is the bioactive moiety that binds AR’s with the greatest affinity. The end result of an increase in SHBG, at least in theory, should be a decrease in DHT binding by AR’s and a reduction in AR-mediated PC cell proliferation.
However, we did not observe significant changes during the intervention in either total serum T or DHT levels. One possible reason for this may have to do with the means by which circulating T levels are regulated. When SHBG levels fall, the relative proportion of unbound T rises, leading to feedback inhibition of gonadal T production.45 When SHBG levels rise, the reverse is the case; in order to maintain a constant concentration of unbound T, total T production rises. The net result is that small changes in total T may be obscured by this mechanism. Of course, it should also be noted that with only 10 patients, the study was likely insufficiently powered to detect all but very large changes in T, even if they had occurred.
Body weight is also known to impact circulating levels of adipokines (cytokines produced in adipose tissue). Two of the most widely studied adipokines, and the two that were examined in this study, are TNF-α and IL-6. These are both pro-inflammatory and respond to changes in adiposity.47,48 Growing evidence supports the suggestion that chronic inflammation may play a role in prostate carcinogenesis.49,50 Not surprisingly, both of these markers have been found to correlate with risk of recurrence, extent of metastatic disease, and poorer survival.30 Again, our inability to detect significant changes in TNF-α and IL-6 is largely due to our small sample size.
The effects of stress on cancer may be mediated by the complex interactions of the nervous, endocrine, and immune systems. More specifically, stress has been observed to inhibit activity of the Leydig cells, resulting in a lower circulating T levels.51 It has also been shown to influence levels of cytokines including TNF-α and IL-6.52 Unfortunately, we were unable to evaluate the effects of stress reduction on levels of presumptive stress-responsive biomarkers (T, TNF-α, or IL-6), again most likely because of insufficient study power.
Assuming that the attenuation of PC progression was mediated by weight-related metabolic changes, a question arises as to what aspect of intervention brought about the observed reduction in adiposity. Earlier 53, we described large increases during months 0–3 in intake of whole grains and vegetables, food groups which are fiber and water-rich, very low in fat, and therefore of low energy density. However, intake of these foods declined slightly during months 3–6. Weight loss during the first three months may possibly have resulted from replacing energy-rich foods with energy-poor foods, and the slight increase in body weight during the second three months may have resulted from a small degree of dietary recidivism.
A second question that naturally arises regarding the reduction in adiposity is whether it matters, in terms of effects on prostate cancer progression, how it is achieved. One aspect of this question has to do with the preferred dietary strategy for reducing energy intake. Another facet regards whether it is more desirable to increase energy expenditure or decrease intake to achieve this end. Although our study and its findings did not address these issues, they remain important ones that warrant consideration in the planning and design of future behavioral approaches to the management of progressive PC. What can be said is that while both a plant-based diet and a high-protein, low-carbohydrate diet high in foods of animal origin (such as the popular Atkins diet) may both result in weight loss, the former is far more consistent with the dietary cancer prevention guidelines of various agencies (69).54
Our study has some important strengths. For example, by focusing on recurrent disease, we may have selected a stage of disease that is particularly diet-sensitive, a supposition that, if correct, made study of this population resource-efficient and may have important potential biological and clinical ramifications. Also, because men in this study population had few unambiguous clinical options, they may have been highly motivated to engage in dietary change. In addition, our intervention obtained biomarker information and assays at multiple timepoints, concurrent with assessments of diet and PSA. Thus, we had the ability to examine how changes in these biomarkers varied with other changes brought about by the intervention. However, there were also several limitations. The small sample size limited our statistical power and prevented stratum-specific analyses or meaningful control of covariates. Similarly, the lack of a randomized control group made it harder to be certain of the validity of our findings. Because our intervention incorporated both diet modification as well as stress management training, we cannot easily separate out effects specifically due to either one of these components. Also, while we measured an array of indicators, there are many more (e.g. genomic or proteomic markers), that were unmeasured and could have played mediating roles. Finally, because we selected a study population with a medical condition that likely instilled a high level of motivation to participate in our intervention, we may be limited in our ability to generalize findings regarding their biomarker changes to the larger prostate cancer patient population.
Our findings suggest that with a motivated patient population, i.e., men with recurrent prostate cancer, and a multifaceted intervention, reductions in body weight and adiposity can be achieved and sustained, at least for the short term. Changes in the rate of PSA rise, an indicator of disease progression, were in the same direction as changes in weight and adiposity and opposite those of changes in SHBG, raising the intriguing possibility that these biomarkers may have played a mediating role in terms of the effects of the diet and stress reduction intervention on PC progression. Further, they raise the possibility that the effects of such mediation could be observable within the relatively short timeframe of 6 months or less. In the future, larger randomized trials will be needed to determine whether changes in body weight or related biomarkers can be maintained for a longer period and whether these changes are related to additional clinical outcomes (such as time to development of metastases or overall survival). Also needed is a study to identify precisely how adhering to a plant-based diet, and reducing stress may affect prostate cancer progression.
This study was supported by NIH Grant 1 K23 AT002965-03, American Cancer Society Grant CRTG-03-073-01-PBP, and UCSD GCRC # 1989. We also thank Ms. Eva Brzezinski, Ms. Cindy Knott, Ms. Mina Marjanovic, Ms. Aisha Menon, Ms. Jean Richardson, Dr. Carol Salem-Hand, and Mr. Paul Shragg for their contributions to this study.