BACKGROUND: Growing evidence suggests intergenerational effects of paternal pre-conceptional smoking through the germ line, but its specific impact on offspring semen quality remains uncertain because of challenges in isolating paternal exposure from maternal passive smoking or underreporting.

METHODS: We reran previous analyses estimating differences in semen parameters and testicular size according to paternal smoking in 867 young adult men, adding first-trimester maternal plasma cotinine to the original adjustment for maternal self-reported smoking. We also estimated differences in sperm DNA fragmentation. Paternal smoking was reported by the pregnant women around gestational week 16. Analyses were additionally adjusted for household occupational status, parental ages at birth, maternal pre-pregnancy body mass index and alcohol consumption, and abstinence time, and accounted for spillage, minutes from ejaculation to analysis, and son's own smoking.

RESULTS: We found no association between paternal preconceptional smoking and any of the semen parameters or testicular size. Adjustment for son's own smoking did not change results.

DISCUSSION: While maternal plasma cotinine offers an objective measure of tobacco exposure and allows for a more thorough adjustment of maternal smoking, the high correlation between paternal pre-conceptional smoking and maternal cotinine exposure may, have resulted in overadjustment removing some paternal effect. Inability to distinguish between paternal never smokers and former smokers, may have led to misclassification of paternal pre-conceptional smoking and underestimation of associations.

CONCLUSION: We found no support for an independent association between paternal pre-conceptional smoking and semen quality in young adult sons, but studies with more detailed paternal smoking history are needed before firm conclusions can be drawn.