7 mg/kg halved P. vivax relapse across Indonesia: what a 1,797-patient meta-analysis just said about higher-dose primaquine

The malaria that comes back from inside the patient's own liver is the harder malaria to control. Plasmodium vivax has the ability, unique among the major human malaria species, to lie dormant in liver cells as hypnozoites and reactivate weeks to months after an initial infection has been treated. The standard 14-day course of primaquine is what kills those dormant forms. The dose has been the policy tension for decades. A new individual-patient-data meta-analysis from Indonesia, published on 18 June in The Lancet Regional Health Western Pacific, has now put a number on the question that national malaria programmes in South-East Asia and the Western Pacific have been quietly arguing about for years: at a total dose of 7 mg/kg, primaquine roughly halves the rate of recurrence over six months, with a manageable safety profile in patients whose G6PD status has been screened.

The lead investigator is Ihsan Fadilah of the Oxford University Clinical Research Unit Indonesia, with senior authors across Oxford's Centre for Tropical Medicine and Global Health, the Mahidol-Oxford Tropical Medicine Research Unit in Bangkok, Menzies School of Health Research in Darwin, and the University of Melbourne. The funding line is the NDM Tropical Network Fund and the Bill and Melinda Gates Foundation. The sample is 1,797 patients from seven eligible studies across Indonesia.

What the meta-analysis actually compared

The headline comparison is total-dose, body-weight-adjusted primaquine at 7 mg/kg versus 3.5 mg/kg. Both arms received standard blood-stage treatment, and the higher-dose regimen is the one under test. The systematic search covered studies published between 1 January 2000 and 23 July 2024 that prospectively enrolled patients with acute uncomplicated P. vivax malaria and treated at least some of them with primaquine. Of ten eligible studies, seven had individual patient data available for pooling and harmonisation. The analysis then fitted one-stage individual-patient-data multivariable regression models to estimate the causal relationship between body-weight-adjusted primaquine dose and three separate primary outcomes.

The first outcome was time to first P. vivax recurrence, measured between days 7 and 180 after treatment. The second was any gastrointestinal discomfort, measured between days 5 and 7. The third was a clinically meaningful haemoglobin drop, defined as a reduction of at least 25% from baseline combined with a fall below 7 g/dL, measured between days 1 and 14.

The numbers, in plain English

Patients treated with a 7 mg/kg total dose had a 47% lower rate of P. vivax recurrence over six months compared with those treated with a 3.5 mg/kg total dose. The adjusted hazard ratio was 0.53, with a 95% confidence interval of 0.45 to 0.63, on 1,797 patients. The relative benefit was consistent across Indonesian transmission settings, even though the absolute benefit varied.

The gastrointestinal discomfort signal went the other way, as expected. Every 0.25 mg/kg increase in the daily primaquine dose was associated with a 32% relative increase in the risk of gastrointestinal discomfort between days 5 and 7. The adjusted risk ratio was 1.32 per 0.25 mg/kg daily dose, with a 95% CI of 1.15 to 1.51, on 952 patients. Higher dose, more nausea and abdominal discomfort. Manageable, not absent.

The haemolysis signal, which is the safety concern that has historically held back the higher-dose regimens, was reassuring. Of 822 patients assessed for haematological safety, 788 (96%) had G6PD activity of 70% or higher, and 34 (4%) had activity in the 30%-to-less-than-70% range. Across the entire assessed population, only one patient developed clinically relevant haemolysis.

Why this question has been sitting open

Primaquine's radical-cure dose has been the recurring policy tension in P. vivax malaria for a generation. The World Health Organization's current recommendations, in the third-edition malaria treatment guidelines and the 2021 update, reflect the underlying uncertainty: lower total doses are safer but less effective at killing hypnozoites, higher total doses reduce relapse but raise the gastrointestinal-tolerability and (in unscreened G6PD-deficient patients) haemolysis burden. National malaria programmes in South-East Asia, the Western Pacific and South Asia have ended up running different regimens in part because the comparative data across transmission settings has been thin. The case for a single, well-supported dose has been building for years but has not until now been matched by an individual-patient-data meta-analysis on this scale.

Indonesia is the right place to ask. P. vivax is the dominant malaria species in much of the country, transmission settings range from low in Java and Bali to high in Papua, and the country's malaria-elimination commitments make a well-tolerated, high-efficacy radical cure operationally important. The Fadilah analysis is, in effect, the Indonesian answer to a question that national malaria programmes in Myanmar, Thailand, India and the Pacific have been quietly waiting for.

What the analysis is, and is not

The study is an individual-patient-data meta-analysis. That is a higher-evidence design than a literature-based meta-analysis because the harmonisation of covariates across patients is tighter than the harmonisation of effect estimates across studies, and the regression models can adjust at the patient level. It is, however, still an observational pooling rather than a head-to-head randomised trial, and the dose comparisons depend on between-study variation as well as within-study variation. The authors are careful to describe their estimates as "causal" in the sense that the design supports causal inference, but the residual confounding surfaces are real.

The analysis pools seven of ten eligible studies. The three studies that did not contribute individual patient data are a meaningful limitation, because their inclusion could shift the pooled estimates in either direction. The G6PD safety analysis covers 822 patients, which is a meaningful sample but a small one for ruling out rare events. One clinically significant haemolysis event across the sample is reassuring; it is not the same as saying the higher-dose regimen is safe in unscreened populations, and the authors do not say that. The risk-management principle that holds for any 8-aminoquinoline holds here: the higher-dose regimen is for patients whose G6PD status has been measured and confirmed to be in the normal range (and ideally above 30% activity, which is the cutoff the safety analysis used).

The other caveat is geography. This is an Indonesian analysis. The relative benefit was consistent across Indonesian transmission settings. The generalisability to Myanmar, Thailand, Cambodia or the Pacific Islands is plausible but is not the same proposition. National programmes looking at this finding will need to consider their own P. vivax relapse patterns and their own G6PD screening coverage before they adopt the higher-dose regimen.

What to watch next

The realistic next signals on the higher-dose primaquine question are: (i) any movement by the WHO Guidelines Development Group for Malaria Control and Elimination (on which two of the senior authors serve) towards updating the primaquine dose recommendations; (ii) the publication of the three studies whose individual patient data was not available for this analysis, which could shift the pooled estimates; (iii) any parallel individual-patient-data analyses from other South-East Asian or Western Pacific countries that confirm or qualify the Indonesian finding. The structural signals matter more than the headlines. The headline this finding wants is "new hope for vivax radical cure". The structural signal is whether the WHO guidance shifts.

For clinicians and national malaria programmes in the region, the operative position has been effectively this: continue to use the dose regimen recommended by national guidelines, screen G6PD status before initiating primaquine, and watch the evidence base for dose refinement. The Fadilah analysis is the strongest single piece of evidence so far in favour of a 7 mg/kg total dose across Indonesian settings, with adjustment for tolerability in patients with intermediate G6PD activity.

What we know

• An individual-patient-data meta-analysis of 1,797 patients from seven eligible Indonesian studies finds that a total primaquine dose of 7 mg/kg reduces the rate of P. vivax recurrence over six months by approximately 47% compared with the 3.5 mg/kg regimen, with adjusted hazard ratio 0.53 (95% CI 0.45-0.63). The relative benefit was consistent across Indonesian transmission settings. [Fadilah I et al. Lancet Reg Health West Pac 2026; PMID 42375875]
• Every 0.25 mg/kg increase in the daily primaquine dose was associated with a 32% relative increase in the risk of gastrointestinal discomfort between days 5 and 7 after treatment, adjusted risk ratio 1.32 per 0.25 mg/kg daily dose (95% CI 1.15-1.51), measured on 952 patients. [Fadilah I et al. Lancet Reg Health West Pac 2026; PMID 42375875]
• Of 822 patients assessed for haematological safety (788 [96%] with G6PD activity ≥70% and 34 [4%] with activity 30% to less than 70%), only one patient developed clinically relevant haemolysis. The analysis is consistent with a manageable haemolysis risk in patients whose G6PD activity has been confirmed to be above 30%. [Fadilah I et al. Lancet Reg Health West Pac 2026; PMID 42375875]

Sources cited

1. Fadilah I, Watson JA, Pasaribu AP, Sutanto I, Nelwan EJ, Lidia K, Rajasekhar M, Elyazar IR, Taylor WR, Thriemer K, Day NP, Poespoprodjo JR, Simpson JA, Price RN, Baird JK, Commons RJ. Effect of higher dose primaquine for the radical cure of Plasmodium vivax malaria in Indonesia: a systematic review and individual patient data meta-analysis. Lancet Reg Health West Pac. 2026 Jul;72:101908. doi:10.1016/j.lanwpc.2026.101908. PMID 42375875; PMCID PMC13310647. https://pubmed.ncbi.nlm.nih.gov/42375875/
2. World Health Organization. Guidelines for the treatment of malaria, 3rd edition (2015, with 2021 update). https://www.who.int/publications/i/item/guidelines-for-the-treatment-of-malaria

Published 2026-07-01 · Mosticare Editorial