We examined the roles of diffusion, convection and capillary transporters in solute removal from extracellular space (ECS) of the brain. Radiolabeled solutes (eight with passive distribution and four with capillary or cell transporters) were injected into the brains of rats (n=497) and multiple-time point experiments measured the amount remaining in brain as a function of time. For passively distributed compounds, there was a relationship between lipid:water solubility and total brain efflux:diffusional efflux, which dominated when k(p), the transcapillary efflux rate constant, was >10(0) h(-1); when 10(-1)<k(p)<10(-2) h(-1) both diffusion and convection contributed, and when k(p)<10(-3) h(-1), convective efflux dominated. Para-aminohippuric acid (PAH) experiments (n=112) showed that PAH entered the brain passively, but had efflux transporters. The total efflux rate constant, k(eff), was the sum of a passive component (k(p)=0.0018 h(-1)), a convective component (k(csf)=0.2 h(-1)), and a variable, concentration-dependent component (k(x)=0 to 0.45 h(-1)). Compounds with cell membrane transporters had longer clearance half times as did an oligonucleotide, which interacted with cell surface receptors. Manipulation of physiologic state (n=35) did not affect efflux, but sucrose efflux half time was longer with pentobarbital anesthesia (24 h) than with no anesthesia or ketamine-xylazine anesthesia (2 to 3 h). These results show that solute clearance from normal brain ECS may involve multiple physiologic pathways, may be affected by anesthesia, and suggests that convection-mediated efflux may be manipulated to increase or decrease drug clearance from brain.