Background: Selenium is an essential trace mineral with potential interest for cardiovascular disease (CVD) prevention owing to its antioxidant properties. Epidemiological data on selenium status and CVD remain inconsistent. The objective of this study was to ascertain whether low serum selenium (SSe) concentrations are related to an increased risk of a first CVD event in a population at high cardiovascular risk. Methods: We undertook a case-control study nested within the “PREvención con DIeta MEDiterránea” (PREDIMED) trial. A total of 207 participants diagnosed with CVD (myocardial infarction, stroke, or cardiovascular death) during the follow-up period (2003–2010) were matched by sex, age, and intervention group to 436 controls by incidence density sampling. Median time between serum sample collection and subsequent CVD event occurrence was 0.94 years. SSe levels were determined using inductively coupled plasma mass spectrometry analysis. Covariates were assessed through validated questionnaires, in-person interviews, and medical record reviews. Conditional logistic regression was used to calculate multivariable-adjusted odds ratios (ORs). Results: Among women, the mean SSe concentration was lower in cases than in controls (98.5 μg/L vs. 103.8 μg/L; p = 0.016). In controls, SSe levels were directly associated with percentage of total energy intake from proteins and fish intake (p for linear trend < 0.001 and 0.049, respectively), whereas SSe concentrations were inversely associated with age, body mass index, and percentage of total energy intake from carbohydrates (p for linear trend < 0.001, 0.008 and 0.016 respectively). In the total group, we observed an inverse dose–response gradient between SSe levels and risk of CVD in the fully-adjusted model (highest vs. lowest quartile: OR = 0.47, 95% CI: 0.27–0.81; ptrend = 0.003). (...)