Pre-diagnostic lipid metabolites are enriched in men who develop advanced prostate cancer: a nested case-control study
Graff, R. E.; Fuller, H.; Wilson, K. M.; Dickerman, B. A.; Chan, J. M.; Kantoff, P. W.; Feng, X.; Clish, C. B.; Vander Heiden, M. G.; Darst, B. F.; Ebot, E. M.; Mucci, L. A.
Show abstract
Few studies with pre-diagnostic samples have estimated associations between circulating metabolites and risk of advanced prostate cancer. We performed untargeted metabolomic profiling of pre-diagnostic blood samples from 212 advanced prostate cancer cases (stage [≥]T3b or lethal during follow-up) and 212 matched controls from the Health Professionals Follow-up Study. 243 metabolites were assayed using liquid chromatography-tandem mass spectrometry (Broad Institute) and met quality control standards. We used multivariable conditional logistic regression to generate odds ratios (OR) and 95% confidence intervals (95%CI) for associations between individual metabolites and risk of advanced prostate cancer, and conducted metabolite set enrichment tests to identify metabolite classes enriched in advanced prostate cancer. Subgroup analyses were conducted by body mass index (BMI) and time between blood draw and diagnosis. Levels of 16 lipid species were nominally associated with advanced prostate cancer at p<0.05, though none were statistically significant after multiple testing correction. The strongest signals were for C56:1 triacylglycerol (TAG; OR: 1.34, 95%CI: 1.07-1.67) and C38:4 diacylglycerol (DAG; OR: 1.27, 95%CI: 1.04-1.55). Enrichment analyses revealed six metabolite classes associated with advanced prostate cancer after multiple testing adjustment, the top four of which were DAGs and TAGs: DAGs overall (P=3.4E-07), unsaturated DAGs (P=5.9E-07), unsaturated TAGs (P=2.3E-06), and TAGs overall (P=2.4E-06). 43 metabolites were nominally associated with advanced prostate cancer among individuals with BMI <25 kg/m2; only three demonstrated nominal associations in individuals with BMI [≥]25 kg/m2. These findings suggest associations between circulating pre-diagnostic lipid levels and aggressive prostate cancer risk, particularly in lean individuals.
Matching journals
The top 7 journals account for 50% of the predicted probability mass.