Compliance with the 2018 World Cancer Research Fund/American Institute for Cancer Research Cancer Prevention Recommendations and Prostate Cancer

The etiology of prostate cancer (PCa) remains largely unknown. Compliance with the 2018 World Cancer Research Fund/American Institute for Cancer Research (WCRC/AICR) cancer prevention recommendations and its relationship to PCa was evaluated. A total of 398 incident PCa cases and 302 controls were included. The selection criteria for both cases and controls were: (i) age between 40–80 years; and (ii) residence in the coverage area of the reference hospitals for 6 months or more prior to recruitment. A score to measure the compliance with the recommendations of 2018 WCRC/AICR criteria was built. The level of compliance was used as a continuous variable and categorized in terciles. The aggressiveness of PCa was determined according to the ISUP classification. Adjusted odds ratios (aOR) and their 95% confidence intervals (95% CI) were estimated using multivariable logistic regression models. A slight protective tendency was observed between the level of compliance with the preventive recommendations and PCa risk, aOR = 0.81 (95% CI 0.69–0.96) for the total cases of PCa. This association also was observed when the aggressiveness was considered. In addition, limiting consumption of “fast foods”, sugar-sweetened drinks, and alcohol were independently associated with lower risk of PCa.


Introduction
Prostate cancer (PCa) is the most common cancer in men, and has the highest incidence and the third highest mortality, after lung and colorectal cancer, in Europe [1]. In recent years, the incidence of PCa has become widespread in all countries [2], with around 450,000 estimated new cases per year [3]. This increase may be related to population aging and new exposures, or higher levels of exposure, to environmental risk factors, together with the massive use of screening for prostate-specific antigen (PSA) and diagnostic techniques [4].
were selected from the Urology Department of the participating hospitals, using the Pathological Anatomy listings and checking the new positive biopsies for PCa. Controls were randomly selected from the population lists assigned to general practitioners of the primary care centers located in the catchment area of the hospitals from which the cases were selected (Granada-Metropolitan Sanitary District). The controls were matched to cases by age, with a maximum age difference of 5 years, based on information from the Granada Cancer Registry, a population-based cancer registry with data quality certified by the International Agency for Research on Cancer (http://cancergranada.org/es/index.cfm).

Data Sources and Variables
Data were collected between 2017 and 2019 by face-to-face interviews conducted by trained interviewers using a structured computerized epidemiological questionnaire. The same interviewers surveyed cases and controls. Information on the following background variables, for both cases and controls, was obtained: sociodemographic data (age, education level, occupation, and marital status), lifestyles (smoking status, alcohol consumption, physical activity, sleeps habits, and diet), and personal/family medical history, including first-degree family history of PCa in father and/or brothers, among other variables. Waist and hip circumference were measured at the interview, while height and weight one year prior to diagnosis were self-reported, and body mass index (BMI) calculated. If there were any missing data, it was recovered by reviewing the clinical record or via telephone contact.
The physical activity information was collected using the International Physical Activity Questionnaire (IPAQ), validated for the Spanish population [34]. In addition, subjects were provided with a semi-quantitative food frequency questionnaire (FFQ); also previously validated for the Spanish population [35] which included 134 foods, specifying the portion size for each of them, referring to the 12 months prior to diagnosis. Nutrient intakes were estimated using Spanish food composition tables [36]. In addition, this study included only subjects with plausible energy intakes. An intake of less than 800 kcal/day and more than 4000 kcal/day were considered as implausible extreme energy intakes [37].

WCRF/AICR Score Construction
The 2018 WCRF/AICR Cancer Prevention Recommendations includes a total of ten recommendations: (1) body fatness, (2) physical activity, (3) consumption of whole grains, vegetables, fruit, and beans, (4) fast food and other processed foods, (5) red and processed meat, (6) sugar-sweetened drinks, and (7) alcohol, (8) dietary supplements, (9) breastfeeding, and (10) cancer survivors' recommendations. In this study, we omitted the latest three recommendations. The reasons for avoiding the three last recommendations were the following: (i) The use of dietary supplements for cancer prevention and choice to consume nutrients through food alone is largely addressed through the other five dietary recommendations, as AICR refers in the article on the operationalizing of the score [16]; and (ii) the specific recommendation for cancer survivors and breastfeeding are not applicable to our population. Table 1 shows the goals for the ten recommendations and their operationalization.
Briefly, the method of estimating the score according to the standardized scoring system for 2018 WCRF/AICR cancer prevention recommendations [16] is based on the following criteria: 1 point was assigned when the recommendation was met, 0.5 points when it was partially met, and 0 points when not met. When the recommendation was composed of two subitems, such as body fatness and consumption of whole grains, vegetables, fruit, and beans recommendation, the scoring weight is divided equally between both to retain a total of one point (0.5, 0.25, and 0 points for meeting, partially meeting, and not meeting each subitem, respectively). Eat a diet high in all types of plant foods including at least five portions or servings (at least 400 g or 15 oz in total) of a variety of non-starchy vegetables and fruit every day Fruits and vegetables intake a ≥ 400 g/day 0.5 Fruits and vegetables intake a ≥ 200-< 400 g/day 0.25 Fruits and vegetables intake a < 200 g/day 0 If you eat starchy roots and tubers as staple foods, eat non-starchy vegetables, fruit, and pulses (legumes) regularly too if possible No information for operationalization n.a.
4. Limit consumption of "fast foods" and other processed foods high in fat, starches, or sugars Limit consumption of processed foods high in fat, starches or sugars including "fast foods", many prepared dishes, snacks, bakery foods and desserts, and confectionery (candy) Percent of total kcal from ultra-processed foods (aUPFs): Tercile 1 1 Tercile 2 0.5 Tercile 3 0

Limit consumption of red and processed meat
If you eat red meat, limit consumption to no more than about three portions per week. Three portions are equivalent to about 350 to 500 g (about 12 to 18 oz) cooked weight of red meat. Consume very little, if any, processed meat.
Red meat <500 g/wk and processed meat < 21 g/wk 1 Red meat <500 g/wk and processed meat ≥ 21-<100 g/wk 0.5 Red meat ≥500 g/wk or processed meat ≥ 100g/wk 0 In particular, for the recommendation on "fast foods" we selected all ultra-processed foods-according to the NOVA classification [38]-available in the FFQ. After excluding those items already included in other recommendations (i.e., processed meats, sweetened drinks, and alcoholic beverages), we calculated the proportion of calorie intake (kcal/day) derived from ultra-processed foods (group 4 NOVA classification) with respect to the total caloric intake per day. The ratio of fast food caloric intake divided by total caloric intake was finally broken down into terciles according to its distribution in the control group.
The score of each recommendation was added to obtain the total score, which ranged from a minimum value of 0 to a maximum value of 7 points, with higher values indicating high compliance with the cancer prevention recommendations. From the score obtained for the control group, the cut-off points of the terciles were established and applied to both groups to define the level of compliance: (i) Tercile 1: minimal compliance with the recommendations; (ii) Tercile 2: intermediate compliance and; (iii) Tercile 3: minimum compliance.

Measurement of Tumor Aggressiveness
The Gleason score was collected from the pathological report of each participant. It assigns two grades for each patient: a primary grade is given to describe the cells that make up the largest area of the tumor and a secondary grade is given to describe the degree of differentiation of the cells of the next largest area. Both grades use values from 1 to 5: 1 for a high degree of differentiation and 5 for a minimal degree of differentiation [39]. The aggressiveness of the tumor was determined according to the classification of the International Society of Urological Pathology (ISUP) [40], which establishes the existence of five grade groups that are: (i) ISUP 1 (Gleason 3 + 3); (ii) ISUP 2 (Gleason 3 + 4); (iii) ISUP 3 (Gleason 4 + 3); (iv) ISUP 4 (Gleason 8); and (v) ISUP 5 (Gleason > 8). From this, two categories of aggressiveness were constructed: low aggressiveness (ISUP 1, 2) and high aggressiveness (ISUP 3, 4, 5) [41].

Statistical Analysis
Characteristics of the participants were examined using means and standard deviations (SD) for continuous variables and percentages for categorical variables. These characteristics were described for case and control groups as well as 2018 WCRF/AICR terciles for the control group. Chi-squared tests were used to evaluate the level of significance of the differences observed in categorical variables, and Student's t-tests or one-way ANOVA for continuous variables.
Multivariable logistic regression models were used to estimate odds ratios (OR) and 95% confidence intervals (95% CI) for the association between the 2018 WCRF/AICR recommendations and PCa. The 2018 WCRF/AICR score was analyzed as a continuous variable (one-unit increment) and as a categorical variable through the terciles. The first tercile was used as the reference category (minimum compliance with the cancer prevention recommendations). We used the information from previous studies and directed acyclic graph (DAG) to identify potential confounders; thus, epidemiological and statistical criteria were used to construct the models. An adjusted model was constructed including as covariates: age, educational level, smoking status, primary family history of PCa, and total energy intake. In addition, analyses were also conducted stratifying by aggressiveness: low aggressiveness and high aggressiveness. We estimated the individual association of each component of the 2018 WCRF/AICR score with PCa risk, after adjustment for all other components of the score and the aforementioned potential confounders. For this analysis, participants without information on anthropometric measures, physical activity, or dietary information were excluded. People with implausible energy intake values were also excluded for the analysis. All statistical tests were two-sided and statistical significance was set at p < 0.05. Statistical analyses were performed using statistical program Stata v.15 (Stata Corp., College Station, TX, USA, 2017). Figure 1 shows the flow-chart diagram for the participants in the CAPLIFE study until December 2019. A total of 398 PCa cases and 302 controls had complete information for the 2018 WCRF/AICR score and a plausible total energy intake. There were no statistically significant differences between controls with and without information on the score and PCa cases with and without the score, except for the educational level in PCa cases (Supplementary Table S1). component of the 2018 WCRF/AICR score with PCa risk, after adjustment for all other components of the score and the aforementioned potential confounders. For this analysis, participants without information on anthropometric measures, physical activity, or dietary information were excluded. People with implausible energy intake values were also excluded for the analysis. All statistical tests were two-sided and statistical significance was set at p < 0.05. Statistical analyses were performed using statistical program Stata v.15 (Stata Corp., 2017). Figure 1 shows the flow-chart diagram for the participants in the CAPLIFE study until December 2019. A total of 398 PCa cases and 302 controls had complete information for the 2018 WCRF/AICR score and a plausible total energy intake. There were no statistically significant differences between controls with and without information on the score and PCa cases with and without the score, except for the educational level in PCa cases (Supplementary Table 1). Distribution characteristics between cases and controls are shown in Table 2. Compared with controls, PCa cases were slightly older, 67.7 (SD 7.4) vs. 65.3 (8.2) years, they had a higher alcohol consumption (16.8 for cases vs. 10.3% for controls) and a higher energy intake, 2511.8 (SD 705.7) vs. 2438.1 (SD 617.6) kcal/day. More than three quarters of the PCa cases were of low aggressiveness (77.1%) cases with an ISUP of 1 or 2. In terms of compliance with the 2018 WCRF/AICR score, this was slightly lower in the PCa cases than in controls, 3.27 (DS 0.93) vs. 3.42 (DS 1.01) points, respectively. The highest percentage of PCa cases (40.2%) had a score >2 and ≤ 3 points, while 34.8% controls had a score >3 and ≤ 4 points. The distribution of controls' characteristics according to terciles of compliance with the score is provided in Table 3. Those participants with the highest compliance vs. subjects with the lowest compliance with the 2018 WCRF/AICR recommendations were older, had a lower energy intake, and had lower alcohol consumption (3.3% with high alcohol consumption in controls with the highest compliance of the recommendations vs. 19.8% in controls with the lowest compliance; p < 0.01). When we analyzed the association with recommendation compliance, a slight protective relationship was observed between a higher compliance with the 2018 WCRF/AICR recommendations and PCa when this variable was analyzed as continuous (Table 4). For each unit of increase in the score, the risk of PCa was reduced by 19%, aOR = 0.81 (95% CI 0.69-0.96) (p-value = 0.02). Its protective association is maintained when the 2018 WCRF/AICR score is considered as an ordinal variable in terciles, although without reaching statistical significance. The results are similar when we stratify by aggressiveness, aOR = 0.79 (95% CI 0.66-0.95) and aOR = 0.86 (95% CI 0.69-1.06) for each unit of increase in the score for low and high aggressiveness, respectively.

Results
The mutually aOR for the individual components of the 2018 WCRF/AICR score and PCa are shown in Table 5. Three of the components of the score were independently associated with a lower risk of PCa: fast food and other processed foods, sugar-sweetened drinks, and alcohol consumption. Lower consumption of "fast foods' and other processed foods high in fat, starches or sugars, and sugar-sweetened drinks were associated with a lower risk of PCa, aOR = 0.63 (95% CI 0.42-0.64) and aOR = 0.30 (95% CI 0.12-0.77), respectively (p-trend < 0.05). Also, not consuming alcohol showed a protective association, aOR = 0.49 (95% CI 0.27-0.91) for men who met with the recommendations (p-trend < 0.05). Results according to aggressiveness are detailed in Supplementary Table S2 and similar results were obtained for all of the individual items; although these results should be taken with caution because of the sample size.

Discussion
This case-control study explores the association between the 2018 WCRF/AICR Cancer Prevention Recommendations and the risk of PCa, being one of the first studies in exploring this issue. A slight protective association was found between each one-unit increase of the 2018 WCRF/AICR score and the risk of PCa, especially for low aggressiveness PCa. In addition, three components of the 2018 WCRF/AICR score were independently associated with a lower risk of PCa: (i) "Fast foods" and other processed foods high in fat, starches or sugars; (ii) sugar-sweetened drinks; and (iii) alcohol.
If we compare the 2018 WCRF/AICR score with the previous recommendations, six of them partially coincide, while the following four recommendations have changed substantially: physical activity, alcohol intake, "fast foods" and other processed foods, and sugar-sweetened drinks [15]. In particular, we have found an association for three of the four changed recommendations. These changes permit a better definition of the recommendations and their characteristics, reducing the chances of misclassification. The physical activity recommendation of ≥30 min/day without indicating intensity was changed and subsequently considered only moderate-vigorous physical activity. In terms of alcohol consumption, the new recommendation supports a consumption equal to zero, instead of a daily intake of no more than two drinks for men in the 2007 WCRF/AICR recommendations. This recommendation is similar to the advice included in the European Code against Cancer [42]. Regarding foods and drinks that promote weight gain, the previous recommendation of 2007 is divided in two new recommendations: consumption of "fast foods" and other processed foods, and sugar-sweetened drinks.
To date the association between compliance with the 2018 WCRF/AICR recommendations and PCa has not been evaluated, but it has already been explored for the 2007 WCRF/AICR recommendations. The studies that use the previous recommendations do not find an association between these recommendations and PCa when the aggressiveness of the tumor is not considered [21][22][23]29,31]. The absence of predefined cut-off points for the different items of the 2007 recommendations might also have contributed to these negative results.
By stratifying for aggressiveness, a protective association between the cancer prevention recommendations and low aggressiveness PCa was observed; a finding not previously found by other researchers. For cases with a tumor with high aggressiveness, we also observed a protective effect, although the results do not reach the statistical significance. This may be due to the small sample size for high aggressiveness PCa in our study. In fact, other studies find similar results, MCC-Spain [22], where an inverse association was found for Gleason score ≥ 7 cases (aOR = 0.85, 95% CI 0.76-0.96, per 1-point increment in the WCRF/AICR score). A similar association (aOR = 0.87, 95% CI 0.79-0.96, per 1-point increment in the WCRF/AICR score) was observed in the North Carolina-Louisiana Prostate Cancer Project [30].
Regarding limiting consumption of "fast foods" and other processed foods high in fat, starches or sugars, a recent cohort study, published in 2018, studied the association between "fast foods" and processed food and PCa risk [43]. This study found associations between ultra-processed foods and cancer in general, and breast cancer in particular; for PCa, no relation with consumption of fast food was observed, however only 281 PCa cases were analyzed.
For sugar-sweetened drinks and PCa, the results are also limited. According to a meta-analysis carried out with four observational studies conducted before 2012, the results are highly heterogeneous and they did not find an association between carbonated beverages and PCa [44]. In our study, we have considered both carbonated beverages and industrial sugar juices, and this difference could explain the inconsistent results. The French NutriNet-Santé prospective cohort (2009-2017) suggests a relationship between drink consumption and the total risk of cancer, producing an increase in the total risk of cancer of 18% per 100 mL per day [45]. In this cohort study, a higher risk is also observed for PCa, although it did not reach statistical significance because of a low sample size (291 PCa cases). Data on alcohol consumption and PCa is far more consistent, and a recent meta-analysis points to positive association for PCa; this risk starts even with a low volume consumption of alcohol (from 1.3 g per day) [46].
We found that PCa cases without a WCRF/AICR score had a lower educational level (Supplementary Table S1), as subjects with a lower educational level tend to have a lower compliance with recommendations, this would be underestimating the findings found. We may have been limited by the small sample size and lack of statistical power to detect significant associations, especially among high aggressiveness PCa cases. Finally, although we adjusted for a range of potential confounders which were associated with both the score and PCa risk, we cannot rule out the possibility of residual confounding.
Regarding advantages of our study, we have to point out the use of a previously validated FFQ for the Spanish population, including regional products [35]. In order to avoid possible changes in dietary habits after cancer diagnosis, we have only included incident cases before starting any type of treatment for PCa. In addition, the approach using the 2018 WCRF/AICR recommendations permits the evaluation of diet as a whole, accounting for possible synergistic effects between nutrients and foods on cancer risk, and it also incorporates physical activity and body fatness; hence being an indicator of an overall healthy lifestyle. As an additional strength of our study, the 2018 WCRF/AICR recommendations have been based on a standardized scoring system [16], allowing the construction of comparable and solid cut-off points for each recommendation, a problem found in the previous applications of this score. It is also noteworthy, that most of the participants of the CAPLIFE study had detailed information about the diet and plausible energy intake which allowed the construction of the score for 89.4% of PCa cases and 91.2% of controls.

Conclusions
In conclusion, in this Spanish population-based case-control study, a slight protective association was found between compliance with the 2018 World Cancer Research Fund/American Institute for Cancer Research (WCRC/AICR) cancer prevention recommendations on diet, physical activity, and body fatness and overall PCa, especially in low aggressiveness PCa. In addition, limiting consumption of "fast foods" and other processed foods high in fat, starches or sugars, sugar-sweetened drinks, and alcohol were independently associated with lower risk of PCa. Altogether, it is advisable to comply with healthy lifestyles in the prevention of PCa.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6643/12/3/768/s1. Table S1: Characteristics of controls and PCa cases with and without WCRF/AICR score in CAPLIFE study. Table S2: Mutually adjusted odds ratios and 95% confidence intervals for low and high aggressiveness PCa associated with the components of the 2018 WCRF/AICR score in CAPLIFE study.