Abstract

Background The Plasmodium vivax that was once prevalent in temperate climatic zones typically had an interval between primary infection and first relapse of 7–10 months, whereas in tropical areas P.vivax infections relapse frequently at intervals of 3–6 weeks. Defining the epidemiology of these two phenotypes from temporal patterns of illness in endemic areas is difficult or impossible, particularly if they overlap. Methods A prospective open label comparison of chloroquine (CQ) alone versus CQ plus unobserved primaquine for either 5 days or 14 days was conducted in patients presenting with acute vivax malaria in Kolkata. Patients were followed for 15 months and primary and recurrent infections were genotyped using three polymorphic antigen and up to 8 microsatellite markers. Results 151 patients were enrolled of whom 47 (31%) had subsequent recurrent infections. Recurrence proportions were similar in the three treatment groups. Parasite genotyping revealed discrete temporal patterns of recurrence allowing differentiation of probable relapse from newly acquired infections. This suggested that 32 of the 47 recurrences were probable relapses of which 22 (69%) were genetically homologous. The majority (81%) of probable relapses occurred within three months (16 homologous, 10 heterologous) and six genetically homologous relapses (19%) were of the long latency (8–10 month interval) phenotype. Conclusions With long follow-up to assess temporal patterns of vivax malaria recurrence, genotyping of P.vivax can be used to assess relapse rates. A 14 day unobserved course of primaquine did not prevent relapse. Genotyping indicates that long latency P.vivax is prevalent in West Bengal, and that the first relapses after long latent periods are genetically homologous. Trial Registration Controlled-Trials.com ISRCTN14027467