In May 2003, we studied amodiaquine (AQ) efficacy in children < 5 years of age with uncomplicated falciparum malaria in Magoda and Mpapayu (with insecticide treated nets [ITNs]) and Mgome (without ITNs) in Muheza, Tanzania. The trial involved 28 days of follow-up, and data were adjusted by polymerase chain reaction (PCR) genotyping of msp1 and msp2 genes. Additionally, Pfcrt codon 72-76 polymorphisms were studied by PCR and sequence-specific oligonucleotide probe (SSOP) ELISA. In 54 cases with complete follow-up, a significant difference in late treatment failure (LTF) rates was seen (60.7% in ITN versus 88.5% in non-ITN villages, P = 0.02) before PCR correction. However, after PCR correction, 23 cases (60.5%) were confirmed as reinfections, giving a true LTF rate of 21.4% (6/28) and 34.6% (9/26) in the above settings, respectively. Frequency of Pfcrt CVIET haplotype mutation pretreatment was high (97.0%); the remaining samples were CVMNK. We conclude that AQ alone is no longer effective in the study area.