METHODS AND FINDINGS: Clinical efficacy studies of uncomplicated P. vivax treated with DP or AL and published between January 1, 2000, and January 31, 2018, were identified by conducting a systematic review registered with the International Prospective Register of Systematic Reviews (PROSPERO): CRD42016053310. Investigators of eligible studies were invited to contribute individual patient data that were pooled using standardised methodology. The effect of mg/kg dose of piperaquine/lumefantrine, ACT administered, and PQ on the rate of P. vivax recurrence between days 7 and 42 after starting treatment were investigated by Cox regression analyses according to an a priori analysis plan. Secondary outcomes were the risk of recurrence assessed on days 28 and 63. Nineteen studies enrolling 2,017 patients were included in the analysis. The risk of recurrent P. vivax at day 42 was significantly higher in the 384 patients treated with AL alone (44.0%, 95% confidence interval [CI] 38.7-49.8) compared with the 812 patients treated with DP alone (9.3%, 95% CI 7.1-12.2): adjusted hazard ratio (AHR) 12.63 (95% CI 6.40-24.92), p < 0.001. The rates of recurrence assessed at days 42 and 63 were associated inversely with the dose of piperaquine: AHRs (95% CI) for every 5-mg/kg increase 0.63 (0.48-0.84), p = 0.0013 and 0.83 (0.73-0.94), p = 0.0033, respectively. The dose of lumefantrine was not significantly associated with the rate of recurrence (1.07 for every 5-mg/kg increase, 95% CI 0.99-1.16, p = 0.0869). In a post hoc analysis, in patients with symptomatic recurrence after AL, the mean haemoglobin increased 0.13 g/dL (95% CI 0.01-0.26) for every 5 days that recurrence was delayed, p = 0.0407. Coadministration of PQ reduced substantially the rate of recurrence assessed at day 42 after AL (AHR = 0.20, 95% CI 0.10-0.41, p < 0.001) and at day 63 after DP (AHR = 0.08, 95% CI 0.01-0.70, p = 0.0233). Results were limited by follow-up of patients to 63 days or less and nonrandomised treatment groups.

METHODS: A systematic review identified P. vivax efficacy studies of chloroquine with or without primaquine published between January 2000 and March 2017. Individual patient data were pooled using standardised methodology, and the haematological response versus time was quantified using a multivariable linear mixed effects model with non-linear terms for time. Mean differences in haemoglobin between treatment groups at day of nadir and day 42 were estimated from this model.

RESULTS: In total, 3421 patients from 29 studies were included: 1692 (49.5%) with normal G6PD status, 1701 (49.7%) with unknown status and 28 (0.8%) deficient or borderline individuals. Of 1975 patients treated with chloroquine alone, the mean haemoglobin fell from 12.22 g/dL [95% CI 11.93, 12.50] on day 0 to a nadir of 11.64 g/dL [11.36, 11.93] on day 2, before rising to 12.88 g/dL [12.60, 13.17] on day 42. In comparison to chloroquine alone, the mean haemoglobin in 1446 patients treated with chloroquine plus primaquine was - 0.13 g/dL [- 0.27, 0.01] lower at day of nadir (p = 0.072), but 0.49 g/dL [0.28, 0.69] higher by day 42 (p  25% to  5 g/dL.

CONCLUSIONS: Primaquine has the potential to reduce malaria-related anaemia at day 42 and beyond by preventing recurrent parasitaemia. Its widespread implementation will require accurate diagnosis of G6PD deficiency to reduce the risk of drug-induced haemolysis in vulnerable individuals.