For vivax, the schizontocide buys the time that primaquine needs

Design

open-label, single-centre, parallel-group randomised trial, 1:1:1:1, per-protocol primary analysis

Population

418 patients over 6 months old with microscopically confirmed P vivax malaria in Manaus, G6PD activity above 30% of normal

Primary outcome

recurrence rate at day 42

Effect

with delayed primaquine: 2.4% (95% CI 0.3-8.5) vs 27.3% (18.3-37.8), HR 0.08 (0.00-0.22), p<0.0001; with immediate primaquine 1.0% vs 2.0%

Curavivax randomised 418 patients older than six months with microscopically confirmed Plasmodium vivax malaria in Manaus, Brazil, to four arms: dihydroartemisinin-piperaquine or chloroquine, each with primaquine started on day 0 or delayed to day 42. G6PD activity had to be above 30% of normal. The primary outcome was recurrence at day 42.

With primaquine started immediately, the two schizontocides were indistinguishable: 2.0% recurrence with chloroquine against 1.0% with dihydroartemisinin-piperaquine. The separation appeared when primaquine was delayed. Chloroquine alone recurred in 27.3% (95% CI 18.3-37.8); dihydroartemisinin-piperaquine alone in 2.4% (0.3-8.5), hazard ratio 0.08 (0.00-0.22). All treatments were well tolerated.

That gap is the clinically useful part. Primaquine often cannot start on day 0 — G6PD testing is pending, the patient is pregnant, adherence over 14 days is doubtful. With chloroquine, that delay costs one patient in four a recurrence within six weeks. With dihydroartemisinin-piperaquine, the long piperaquine tail covers the interval. For Indian practice, where vivax dominates the malaria burden and G6PD testing is inconsistently available, that is the argument worth taking to a treatment protocol — though the data are Brazilian, and chloroquine resistance patterns are not the same here.

• Where primaquine must be delayed, prefer a long-acting artemisinin combination as the schizontocide
• Test G6PD before primaquine rather than delaying the schizontocide to wait for it
• A recurrence at six weeks is not necessarily treatment failure — relapse from hypnozoites and reinfection look the same
• Record whether the 14-day primaquine course was completed; that is the variable most trials cannot control and most clinics cannot either
• These are Brazilian data with chloroquine-sensitive comparison; confirm local resistance patterns before extrapolating

Why it matters

It turns a logistical delay — waiting for G6PD, doubting adherence — from a clinical cost into a manageable one.

Don't overread it

Single-centre, open-label, Brazilian Amazon; recurrence at 42 days, not radical cure.

The statistics, in plain English

A hazard ratio of 0.08 sounds extreme, and the absolute numbers say why: 27.3% recurrence against 2.4%, or roughly one in four against one in forty. The confidence interval on the immediate-primaquine comparison runs from 0.00 to over 40,000, which means with only three or four events in those arms the trial could not compare them at all — the honest reading is that they looked the same, not that they were shown equivalent.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.