The association between diet-exercise patterns and self-reported prostate cancer: A cross-sectional NHANES study

The association between diet-exercise patterns and self-reported prostate cancer remains unclear, which was explored based on the National Health and Nutrition Examination Survey (NHANES) database in this study. A cross-sectional study design was adopted using data from the NHANES 2007-2010. The diet-exercise components (37 food variables and 6 exercise variables) were derived through multiple factor analysis, and diet-exercise patterns were derived by K-means clustering analysis. Given the complex sampling design of NHANES, this study accounted for stratification (SDMVSTRA), cluster sampling (SDMVPSU), and sample weights. We applied the Mobile Examination Center examination weights (WTMEC2YR) and adjusted them for the pooled survey cycles. Weighted multivariate logistic regression analysis, weighted subgroup analysis and weighted sensitivity analysis were applied to explore the association between different diet-exercise clusters and prostate cancer. This study enrolled 2825 participants: prostate cancer group (N = 105) and non-prostate cancer group (N = 2720). A total of 43 principal components were extracted via multiple factor analysis. The k-means clustering algorithm identified 2 distinct clusters with divergent diet and physical activity profiles. Weighted logistic regression models revealed no significant association between Cluster 2 and lower odds of prevalent self-reported prostate cancer relative to Cluster 1 (Model 1: odds ratio [OR] = 0.61, 95% CI: 0.37-1.00, P = .057; Model 2: OR = 1.34, 95% CI: 0.70-2.57, P = .379; Model 3: OR = 1.46, 95% CI: 0.73-2.92, P = .300). A significant interaction by age was observed (P for interaction < .001). Compared with Cluster 1, Cluster 2 was associated with lower prevalent self-reported prostate cancer among participants younger than 65 years (OR = 0.20, 95% CI: 0.04-0.86, P = .044), whereas this pattern associated with higher prevalent self-reported prostate cancer among older adults aged ≥65 years (OR = 2.61, 95% CI: 1.37-4.99, P = .009). Sensitivity analyses yielded consistent primary findings. No significant association between diet-exercise patterns and the prevalent self-reported prostate cancer was detected in the overall population, whereas age exerted a statistically significant effect modification on this relationship. Among participants younger than 65 years, Cluster 2, characterized by high physical activity combined with an unhealthy diet, was correlated with reduced prevalence of prostate cancer; conversely, Cluster 2 was significantly associated with elevated prevalence in individuals aged 65 years and older.