Following a marked decline in the efficacy in vivo of mefloquine between 1990 and 1994, a combination of artesunate (4 mg/kg/d for 3 d) and mefloquine (25 mg/kg) has been used as first line treatment of uncomplicated falciparum malaria in camps for displaced persons located along the north-western border of Thailand. Antimalarial drug susceptibility of fresh isolates of Plasmodium falciparum from this population was evaluated using a radioisotope microdilution assay between 1995 and 1999. In total, 268 isolates were collected, of which 189 were from primary infections and 79 from recrudescent infections. The geometric mean 50% inhibitory concentration (IC50) values from primary infections were: dihydroartemisinin 1.2 ng/mL, artesunate 1.6 ng/mL, artemether 4.8 ng/mL, atovaquone 0.4 ng/mL, lumefantrine 32 ng/mL, chloroquine 149 ng/mL, quinine 354 ng/mL, mefloquine 27 ng/mL and halofantrine 4.1 ng/mL. A significant positive correlation was found between the susceptibility in vitro to artesunate and quinine (r = 0.43, P < 0.001), mefloquine (r = 0.46, P < 0.001), and halofantrine (r = 0.51, P < 0.001). These levels of resistance in vitro are among the highest reported and confirm continuing high level multidrug resistance in this area. Despite intensive use of the combination between 1995 and 1999 there has been a significant improvement in mefloquine sensitivity (P < 0.001) and artesunate sensitivity (P < 0.001). This supports observations in vivo that the combination of artesunate and mefloquine has reversed the previous decline in mefloquine sensitivity.