The Association Between Serum Prostate‐Specific Antigen and Glycemic Index, Glycemic Load, and Metformin in Individuals with Diabetes: a Cross‐sectional Analysis

Background Improving glycemic control with Metformin therapy, glycemic index (GI), and glycemic load (GL) have been protectively associated with prostate cancer risk. No studies have yet assessed the relationship between GI, GL, and Metformin on serum prostate‐specific antigen (PSA), a controversial‐yet‐widely used marker of prostate health. Methods A cross‐section analysis of baseline PSA data from men with type 2 diabetes mellitus (T2DM) who previously completed a trial at this research center was conducted. Multivariate linear regression quantified associations between PSA and log‐PSA with GI, GL, and Metformin dose. Results Data were available from 317 men, with a mean PSA of 1.41±0.09ng/mL. Medians exposures (ranges) were 80 (56 – 96) for GI, 164 (62 – 386) for GL, and 1500mg (0 ‐ 3000 mg) for Metformin; 90% of men were under Metformin therapy. Neither PSA nor log‐PSA were significantly related to GI or GL. A negative trend was seen with PSA and metformin dose (p=0.09) in multivariate models. PSA was 24% lower among individuals taking 蠅2000mg of Metformin (n=125) compared to individuals taking <1000mg (n=62) (p=0.06). Log‐PSA significantly decreased by 0.15±0.09 ng/mL for every 1000mg increase in Metformin dose (r=‐0.38, p=0.02). Conclusions Metformin, but not GI or GL, inversely related with log‐PSA in a sample of men with T2DM. Modification of PSA‐test sensitivity as a marker of prostate cancer with diet and Metformin warrants attention. Funding CIHR, Barilla, Loblaw Brands ltd