Morphine has been predicted to show nonlinear blood-brain barrier transport at lower concentrations. In this study, we investigated the possibility of separating active influx of morphine from its efflux by using very low morphine concentrations and compared accelerator mass spectrometry (AMS) with liquid chromatography-tandem mass spectrometry (LC-MS/MS) as a method for analyzing microdialysis samples. A 10-min bolus infusion of morphine, followed by a constant-rate infusion, was given to male rats (n = 6) to achieve high (250 ng/ml), medium (50 ng/ml), and low (10 ng/ml) steady-state plasma concentrations. An additional rat received infusions to achieve low (10 ng/ml), very low (2 ng/ml), and ultralow (0.4 ng/ml) concentrations. Unbound morphine concentrations from brain extracellular fluid and blood were sampled by microdialysis and analyzed by LC-MS/MS and AMS. The average partition coefficient for unbound drug (K(p,uu)) values for the low and medium steady-state levels were 0.22 ± 0.08 and 0.21 ± 0.05, respectively, when measured by AMS [not significant (NS); p = 0.5]. For the medium and high steady-state levels, K(p,uu) values were 0.24 ± 0.05 and 0.26 ± 0.05, respectively, when measured by LC-MS/MS (NS; p = 0.2). For the low, very low, and ultralow steady-state levels, K(p,uu) values were 0.16 ± 0.01, 0.16 ± 0.02, and 0.18 ± 0.03, respectively, when measured by AMS. The medium-concentration K(p,uu) values were, on average, 16% lower when measured by AMS than by LC-MS/MS. There were no significant changes in K(p,uu) over a 625-fold concentration range (0.4-250 ng/ml). It was not possible to separate active uptake transport from active efflux using these low concentrations. The two analytical methods provided indistinguishable results for plasma concentrations but differed by up to 38% for microdialysis samples; however, this difference did not affect our conclusions.