Objective: Micronutrients deficiencies in hemodialysis patients are due to low dietary intakes and intradialytic losses for hydrophilic micronutrients. Conversely, lipophilic nondialyzable compounds might accumulate because of a lack of elimination through renal metabolism or dialysis. Other compounds have complex metabolism: their concentration is not explained by these phenomenons. The aim of this study was to report plasma concentrations of lipophilic micronutrients in hemodialysis patients and to analyze if these concentrations were predictive of mortality. Design: The design was monocentric observational longitudinal study. Subjects: A total of 123 hemodialysis patients included in this observational study. Main Outcome Measure: Plasma concentration of lipophilic micronutrients retinol and its two co-transporters transthyretin and retinol-binding protein 4, tocopherol, and carotenoids (α-carotene and β-carotene, β-cryptoxanthin, lycopene, lutein, and zeaxanthin), and all factors associated with 1-year mortality. Results: Within the 123 patients of the study, median age (interquartile range) was 77.5 (69.5-84.5) years and 58.5% were male. Median retinol plasma concentration was 4.07 (2.65-5.51) μmol/L, and 91.9% of patient had high plasma retinol concentrations. In monovariate analysis, retinol levels were inversely correlated with mortality (hazard ratio = 0.57 [0.45-0.72]; P < .001). This effect remained significant after adjustment with several parameters. Nevertheless, the correlation between retinol and mortality disappeared as soon as transthyretin was added in the statistical model, suggesting an effect of transthyretin as confusing bias. Median tocopherol plasma concentration was 34.8 (28.3-42.9) μmol/L and 72.4% of patients had high plasma tocopherol concentration. Neither tocopherol plasma levels nor carotenoids concentrations were correlated with death in multivariate analysis. Conclusions: In hemodialysis patients, the correlation between retinol plasma concentration and mortality represents the nutritional status but not a direct biological effect of retinol. Retinol is only a surrogate predictor of mortality. It might not represent vitamin A levels, but likely the transthyretin level. Plasma retinol levels should be interpreted cautiously in hemodialysis patients.