Native circulating oxytocin (OT) levels in non-pregnant/non-lactating/non-medicated humans are very low (≤ 8 pg/mL). The lower limit of detection (LLOD) of our previous liquid chromatography mass spectrometry (LC-MS) method (10-25 pg/mL) precluded their quantification in serum and urine. Thus, we sought to improve the LC-MS sensitivity of OT measurements in these matrices by hydrophobic tagging and solid phase extraction (SPE). In the former approach, OT was reduced then alkylated with N-alkyl acetamide (C12, C14, C16, and C18) tags or derivatized using sulfonyl chloride-based reagents. In the latter approach, native OT in serum and urine was concentrated by offline SPE using gradient acetonitrile washings after first crashing with acetonitrile. Peak urinary eluate fractions were further concentrated online then analyzed by orbitrap-based LC-MS with electrospray ionization. All hydrophobic OT derivatives had lower sensitivity than native OT. Washing with a water-acetonitrile gradient during SPE improved the LLOD of OT in spiked serum to 2.5 pg/mL, while adding a subsequent online-concentration step improved the LLOD in spiked urine to 1-5 pg/mL and allowed us to detect OT in urine from lactating women. We were unable to improve the sensitivity of OT measurements by hydrophobic tagging or by derivatization using sulfonyl chloride-based reagents. However, we were successful in improving the sensitivity of native OT measurements in serum and urine 2- and 5-fold, respectively, from our previous orbitrap-based LC-MS method. Offline SPE was mandatory for both matrices and a subsequent online-concentration step was required for urine.