Association Between Antidiabetic Medications and Prostate-Specific Antigen Levels and Biopsy Results

IMPORTANCE Diabetic men appear to have a lower risk of prostate cancer. Whether antidiabetic medications are protective or potentially mask prostate cancer by lowering prostate-specific antigen (PSA) levels is unclear. OBJECTIVE To examine the associations of antidiabetic medication use with (1) PSA levels, (2) frequency of PSA testing, (3) receipt of biopsy following elevated PSA results, and (4) prostate cancer detection at biopsy. DESIGN, SETTING, AND PARTICIPANTS Population-based cohort study using data from the Stockholm PSA and Biopsy Register. Participants were all prostate cancer-free men aged 40 to 79 years residing in Stockholm County, Sweden, between January 1, 2006, and December 31, 2015. Data were analyzed from November 2018 to March 2019. EXPOSURES One or more prescription formetformin, sulfonylurea, or insulin, as recorded in Sweden's National Prescribed Drug Register. MAIN OUTCOMES AND MEASURES Levels of PSA following first exposure to antidiabetic medications were assessed using multivariable linear regression. Frequency of PSA testing was assessed via multivariable Poisson regression. Biopsy following elevated PSA (>= 3.0 ng/mL) and prostate cancer detection at biopsy were assessed via multivariable logistic regression. RESULTS The cohort of 564 666 men (median [range] age, 65 [40-79] years) consisted of 4583 men initially exposed to metformin, 1104 exposed to sulfonylurea, and 978 exposed to insulin who were age matched with unexposed men (1:5). Exposed men had lower median (interquartile range) PSA levels before starting antidiabetic medications compared with unexposed men (1.2 [0.7-2.5] vs 1.6 [0.8-3.2] ng/mL). After accounting for baseline differences, PSA levels did not vary from those of unexposed men following exposure to antidiabetic medications. Frequency of PSA testing was higher for those receivingmetformin (rate ratio, 1.07; 95% CI, 1.06-1.09) and sulfonylurea (rate ratio, 1.06; 95% CI, 1.03-1.08) butwas lower for those receiving insulin (rate ratio, 0.79; 95% CI, 0.77- 0.81). Likelihood of biopsy after elevated PSA was lower among men receivingmetformin (odds ratio, 0.87; 95% CI, 0.80-0.96) and insulin (odds ratio, 0.83; 95% CI, 0.74-0.93). There were no differences in prostate cancer detection at biopsy, regardless of PSA levels that triggered the biopsy. CONCLUSIONS AND RELEVANCE This study's findings do not support the hypothesis that the inverse association between diabetes and prostate cancer is mediated through antidiabetic medications lowering PSA levels to mask prostate cancer. They do suggest potential detection bias due to fewer biopsies among men receiving antidiabetic medications, which may explain the lower prostate cancer risk in men with diabetes.