BACKGROUND Rapid diagnostic tests (RDTs) for malaria provide a practical alternative to light microscopy for malaria diagnosis in resource-limited settings. Three-band RDTs incorporating two parasite antigens may have enhanced diagnostic specificity, relative to two-band RDTs with a single parasite antigen (typically histidine-rich protein 2 [HRP2]). METHODS Phase 1: 2,000 children, two months to five years of age, admitted to a referral hospital in Jinja, Uganda, with acute febrile illness were enrolled. A WHO highly rated three-band RDT was compared to light microscopy of thick peripheral blood films read by local expert microscopists.Phase 2: the three-band RDT was used as a screening tool for inclusion of patients in a clinical trial, and subjects with three positive RDT bands were tested by microscopy using blood samples drawn in parallel. Discordant results were adjudicated by PCR. RESULTS Phase 1: 1,648 children had both a RDT and peripheral blood smear performed. The specificity of a RDT with all three bands positive was 82% (95% CI: 79-85%) compared to 62% (95% CI: 59-66%) for HRP2 alone. The sensitivity was 88% (95% CI: 85-89%) and 94% (95% CI: 92-95%) for three-band positive RDT and HRP2 antigen, respectively. 119 patients (7.2%) had a positive HRP2 band, but negative parasite lactate dehydrogenase (pLHD) band and negative peripheral smear, and 72 (61%) of these had received pre-treatment with anti-malarials, suggesting a false positive HRP2 result (p = 0.002).Phase 2: the positive predictive value (PPV) of the three-band RDT was 94% (95% CI 89%-97%) using microscopy as the reference standard. However, microscopy-discordant results were shown to be positive for P. falciparum by PCR in all cases, suggesting that the PPV was in fact higher. CONCLUSIONS The pLDH antigen on three-band RDTs, used in combination with HRP2, provides added diagnostic specificity for malaria parasitaemia and may be useful to distinguish acute infection from recently treated infection. In situations where diagnostic specificity is desirable (e.g., for selection of malaria-infected participants in clinical trials), a three-band RDT should be considered in a sub-Saharan African setting.