Association between serum cotinine levels and prostate-related clinical endpoints
TL;DR
The pack-years analysis supported partial agreement between recent biomarker exposure and cumulative smoking history, but also showed that cotinine and pack-years capture different exposure dimensions.
Abstract
This study examined the association between serum cotinine, an objective biomarker of tobacco exposure, and prostate-related clinical endpoints in men participating in the National Health and Nutrition Examination Survey. Cross-sectional data from five NHANES cycles between 2001 and 2010 were analyzed. Eligible participants were men with available serum cotinine, total prostate-specific antigen (PSA), and free PSA measurements. The free-to-total PSA ratio was calculated from measured free and total PSA. Multivariable linear regression was used to examine associations of serum cotinine with total PSA and PSA ratio. To address the concern that a single cotinine measurement mainly reflects recent exposure, cumulative smoking exposure was estimated from self-reported cigarette history as pack-years. Correlation analyses and sensitivity models were then performed to compare serum cotinine with pack-years. The final analytic population included 7,174 men. Pack-years were available for 6,132 participants, including 3,433 current or former smokers. Serum cotinine was positively but modestly correlated with pack-years (Pearson r = 0.209; Spearman r = 0.270), and the correlation between log-transformed cotinine and log-transformed pack-years was stronger (Pearson r = 0.337). In fully adjusted models, serum cotinine was not associated with total PSA (β=-0.00006, 95% CI: -0.00054 to 0.00042; P = 0.801), but was independently associated with a lower PSA ratio (β=-0.00494, 95% CI: -0.00705 to -0.00283; P = 0.000005). Pack-years was not significantly associated with PSA ratio after adjustment (β=-0.00750, 95% CI: -0.02010 to 0.00509; P = 0.243). Serum cotinine was associated with a lower free-to-total PSA ratio in this nationally representative sample. The pack-years analysis supported partial agreement between recent biomarker exposure and cumulative smoking history, but also showed that cotinine and pack-years capture different exposure dimensions. Longitudinal studies with repeated exposure assessment and detailed prostate disease history are warranted.