P. vivax was detected by PCR in 45% of children aged 5-10 years from our study area in Papua New Guinea (PNG). 504 children were randomized into 2 arms according to Primaquine (PQ) treatment or not at baseline and actively and passively followed for 9 months. We genotyped all P. vivax infections, the majority of these being multi-clone infections. All blood samples positive for P. vivax by qPCR were tested for gametocyte carriage by targeting pvs25 transcripts. Primaquine reduced the risk of P. vivax infections by 80%. The multiplicity of infection and the density of asexual P. vivax stages were not significantly different in both treatment arms. The number of new clones (force of blood-stage infection) was 2.38 ± 0.17 per person per year-at-risk in the PQ-arm compared to 8.04 ± 0.41 in the Placebo arm (P < 0.05). The duration of infections did not differ between the treatment arms, with 73 days [95% CI: 33-849] and 68 days [95% CI: 40-247] in the PQ or Placebo arm, respectively. Detectability of P. vivax clones was low with 0.26 ± 0.06 and 0.24 ± 0.04 in the PQ and Placebo arms. PQ-treated children had a 75% lower risk of carrying gametocytes compared to Placebo recipients. P. vivax positive children in both arms were equally likely to show gametocyte positivity. We conclude that P. vivax relapses contribute significantly to the high burden of P. vivax infection and transmission in PNG. All other infection dynamics parameters were consistent between treatment arms and apparent relapses behave like new infections.