Urine is currently the matrix of choice for the detection of exogenous substances but also for the application of the steroidal module of the Athlete Biological Passport (ABP) consisting in a longitudinal monitoring of steroid biomarkers. To fill the limitations related to urine, the longitudinal monitoring of serum steroids concentration in the so-called 'blood steroid profile' has recently been proposed. Although serum samples are collected much less than urine samples, plasma derived from ABP whole blood samples used for the full blood count could be exploited for the quantification of endogenous steroids. Alternatively, dried blood spots (DBS) that are much easier to collect could also serve as matrix for the steroid profile. In this study, we compared the concentration levels of several endogenous steroids measured in three different blood matrices (serum, plasma and DBS) collected from 100 elite athletes participating in the 2019 Doha World Athletics Championships using UHPLC-MS/MS. Plasma and serum samples were collected by venipuncture, whereas DBS were generated from whole blood samples. Although steroids demonstrated a good agreement between the three matrices, a slight but acceptable underestimation (10%-20%) was observed in plasma compared with serum. The difference between DBS and the two other matrices was dependent of the bias between serum and plasma. We also showed that a generic HCT correction for DBS could be a valuable approach for quantitative measurements. This study demonstrates the possibility to use three different matrices for the quantification of endogenous steroids although the slight discrepancies should be considered for longitudinal evaluation.