BACKGROUND: Malaria burden in Brazil has reached its lowest levels in 35 years and Plasmodium vivax now accounts for 84% of cases countrywide. Targeting residual malaria transmission entrenched in the Amazon is the next major challenge for ongoing elimination efforts. Better strategies are urgently needed to address the vast reservoir of asymptomatic P. vivax carriers in this and other areas approaching malaria elimination. METHODS: We evaluated a reactive case detection (RCD) strategy tailored for P. vivax transmission in farming settlements in the Amazon Basin of Brazil. Over six months, 41 cases detected by passive surveillance triggered four rounds of RCD (0, 30, 60, and 180 days after index case enrollment), using microscopy- and quantitative real-time polymerase chain reaction (qPCR)-based diagnosis, comprising subjects sharing the household (HH) with the index case (n = 163), those living in the 5 nearest HHs within 3 km (n = 878), and individuals from 5 randomly chosen control HHs located > 5 km away from index cases (n = 841). Correlates of infection were identified with mixed-effects logistic regression models. Molecular genotyping was used to infer local parasite transmission networks. PRINCIPAL FINDINGS/CONCLUSIONS: Subjects in index and neighbor HHs were significantly more likely to be parasitemic than control HH members, after adjusting for potential confounders, and together harbored > 90% of the P. vivax biomass in study subjects. Clustering patterns were temporally stable. Four rounds of microscopy-based RCD would identify only 49.5% of the infections diagnosed by qPCR, but 76.8% of the total parasite biomass circulating in the proximity of index HHs. However, control HHs accounted for 27.6% of qPCR-positive samples, 92.6% of them from asymptomatic carriers beyond the reach of RCD. Molecular genotyping revealed high P. vivax diversity, consistent with complex transmission networks and multiple sources of infection within clusters, potentially complicating malaria elimination efforts.