Some large ecological studies have noted a significant association of testicular cancer (TC) with maternal smoking during pregnancy, while several more controlled studies have been negative. It has been difficult to obtain reliable data on exposure because of the long lag time to cancer diagnosis. We performed a case-control study nested within Finnish, Swedish and Icelandic maternity cohorts exploiting early pregnancy serum samples to evaluate the role of maternal smoking in the risk of TC in the offspring. After reviewing the literature, we also performed a meta-analysis of published studies. For each index mother of the TC patient, three to nine matched control mothers with a cancer-free son born at the same time as the TC case were identified within each cohort. First trimester sera were retrieved from the 70 index mothers and 519 control mothers and were tested for cotinine level by a novel HPLC-MS-MS method developed. No statistically significant association between maternal cotinine level and risk of TC in the offspring was found (OR 0.68; 95% CI 0.35, 1.34). This is the first study based on individual exposure measurements. Its results agree with our meta-analysis of seven previous epidemiological studies (total number of 2149 cases, 2762 controls) using indirect exposure assessment (OR 1.0; 95% CI 0.88, 1.12).