BackgroundOwing to the increasing interest in public health research of antioxidant micronutrients and the inaccuracy of routine serum concentrations of the fat‐soluble vitamins A (retinol) and E (DL‐α‐tocopherol) measurements, we developed a reliable, highly sensitive, robust and rapid method for the quantification of two clinically important lipophilic antioxidants in serum using a reverse‐phase HPLC/DAD method.MethodSample preparation and analytical conditions that would affect extraction efficiency and quantitative results of vitamins A and E were investigated and optimized. Vitamins A and E were extracted from serum via liquid‐liquid extraction (LLE). After adequate sample preparation, the samples were injected directly into the HPLC system with diode‐array detector (DAD). Chromatographic separation was completed in 7 minutes for vitamins A and E. With vitamin A acetate and vitamin E acetate as internal standards, the method was applied to the measurement of vitamins A and E in human serum.ResultsWe evaluated method linearity, accuracy (recovery rate and trueness), precision, carryover, limit of quantitation and limit of detection, and measurement uncertainty. The method was evaluated for trueness using NIST Standard Reference Material SRM 968f. The serum concentration of the studied compounds had a good linear relationship in the range of 0.05 ~ 3.0 μg/mL concentration (r = 0.9998), with 0.0077 μg/mL detection limit and 0.025 μg/mL quantitative limit for vitamin A, respectively, and 1.0 ~ 60.0 μg/mL concentration (r = 0.9999), with 0.40 μg/mL detection limit and 0.50 μg/mL quantitative limit for vitamin E, respectively. The intra‐ and inter‐assay coefficients of variation were calculated by using three concentrations (1, 2, and 3) of the studied compounds in human serum samples. Intra‐assay and inter‐assay precision were 1.23%‐4.97% and 0.97%‐3.79% for vitamin A, respectively, and 0.64%‐4.07% and 0.81%‐5.96% for vitamin E, respectively. The average recovery rates were 100.98% for vitamin A, and 99.21% for vitamin E, respectively. The carryover rate of vitamins A and E was below 1%. As for the evaluation of accuracy, the biases were <± 5% by comparing with NIST standard reference material SRM 968f.ConclusionThe method is a simple sample treatment procedure for the determination of fat‐soluble vitamins A and E in human serum with high sensitivity and specificity. The proposed method could be recommended as a candidate reference method for the determination of serum concentrations of the fat‐soluble vitamins A and E in human serum.