The antifolate combination of pyrimethamine (PM) and sulfadoxine (SD) is the last affordable drug combination available for wide-scale treatment of falciparum malaria in Africa. Wherever this combination has been used, drug-resistant parasites have been selected rapidly. A study of PM-SD effectiveness carried out between 1997 and 1999 at Kilifi on the Kenyan coast has shown the emergence of RI and RII resistance to PM-SD (residual parasitemia 7 days after treatment) in 39 out of 240 (16.25%) patients. To understand the mechanism that underlies resistance to PM-SD, we have analyzed the dihydrofolate reductase (DHFR) and dihydropteroate synthase (DHPS) genotypes of 81 patients. Fifty-one samples were obtained, before treatment, from patients who remained parasite free for at least 7 days after treatment. For a further 20 patients, samples were obtained before treatment and again when they returned to the clinic with parasites 7 days after PM-SD treatment. Ten additional isolates were obtained from patients who were parasitemic 7 days after treatment but who were not sampled before treatment. More than 65% of the isolates (30 of 46) in the initial group had wild-type or double mutant DHFR alleles, and all but 7 of the 47 (85%) had wild-type DHPS alleles. In the paired (before and after treatment) samples, the predominant combinations of DHFR and DHPS alleles before treatment were of triple mutant DHFR and double mutant DHPS (41% [7 of 17]) and of double mutant DHFR and double mutant DHPS (29% [5 of 17]). All except one of the posttreatment isolates had triple mutations in DHFR, and most of these were "pure" triple mutants. In these isolates, the combination of a triple mutant DHFR and wild-type DHPS was detected in 6 of 29 cases (20.7%), the combination of a triple mutant DHFR and a single mutant (A437G) DHPS was detected in 4 of 29 cases (13.8%), and the combination of a triple mutant DHFR and a double mutant (A437G, L540E) DHPS was detected in 16 of 29 cases (55.2%). These results demonstrate that the triply mutated allele of DHFR with or without mutant DHPS alleles is associated with RI and RII resistance to PM-SD. The prevalence of the triple mutant DHFR-double mutant DHPS combination may be an operationally useful marker for predicting the effectiveness of PM-SD as a new malaria treatment.