To investigate whether progesterone metabolites' antinociceptive effects correlate with their previously established binding efficacies at the GABA receptor complex (GBR), seven progestin metabolites were administered to ovariectomized Long-Evans rats s.c. (Expt. 1), via i.c.v. implantation (Expt. 2) and then i.c.v. infusion (Expt. 3). Progestins, listed from most to least efficacious at the GBR, were THP [5 alpha-pregnan-3 alpha-ol-20-one], THDOC [5 alpha-pregnan-3 alpha,21-diol-20-one], P [4-pregnen-3,20-dione], DHP [5 alpha-pregnan-3,20-dione],17-OH-P [17-hydroxyprogesterone], DHEAS [5-androsten-3 beta-ol-17-one sulfate] and PS [5-pregnen-3 beta-ol-20-one sulfate]. Pain sensitivity was measured via the radiant heat tailflick method 0, 5, 20, 40, 60, 80, 100 and 120 min after weekly progestin administration. Peripheral administration of 0.0, 0.1, 0.4, 1.6, 3.2 or 6.4 mg/kg of potent to moderate agonists of the GBR (THP, THDOC, P and DHP) tended to elevate tailflick latencies above baseline, whereas administration of the non-5 alpha-reduced metabolite (17-OH-P) and GBR antagonists (DHEAS and PS) did not. Intracerebroventricular implantation and infusion (0.0, 0.5, 1.0, 2.0 micrograms/rat) of THP, THDOC, P and DHP all significantly increased tailflick latencies above baseline and vehicle control, consistent with their GBR efficacies. Central 17-OH-P, DHEAS and PS did not elevate tailflick latencies. These rapid differences were unlikely confounded by stress given that corticosterone levels were not elevated (Expt. 4). As pain sensitivity was attenuated rapidly (0-5 min post-i.c.v.) and consistent with GBR efficacies, this suggests that progestins' modulation of pain may occur via GBR action.