Methaemoglobin as a surrogate marker of primaquine antihypnozoite activity in Plasmodium vivax malaria: A systematic review and individual patient data meta-analysis

Abstract

Background: The 8-aminoquinolines, primaquine and tafenoquine, are the only available drugs for the radical cure of Plasmodium vivax hypnozoites. Previous evidence suggests that there is dose-dependent 8-aminoquinoline induced methaemoglobinaemia and that higher methaemoglobin concentrations are associated with a lower risk of P. vivax recurrence. We undertook a systematic review and individual patient data meta-analysis to examine the utility of methaemoglobin as a population-level surrogate endpoint for 8-aminoquinoline antihypnozoite activity to prevent P. vivax recurrence.

Methods and findings: We conducted a systematic search of Medline, Embase, Web of Science, and the Cochrane Library, from 1 January 2000 to 29 September 2022, inclusive, of prospective clinical efficacy studies of acute, uncomplicated P. vivax malaria mono-infections treated with radical curative doses of primaquine. The day 7 methaemoglobin concentration was the primary surrogate outcome of interest. The primary clinical outcome was the time to first P. vivax recurrence between day 7 and day 120 after enrolment. We used multivariable Cox proportional-hazards regression with site random-effects to characterise the time to first recurrence as a function of the day 7 methaemoglobin percentage (log base 2 transformed), adjusted for the partner schizonticidal drug, the primaquine regimen duration as a proxy for the total primaquine dose (mg base/kg), the daily primaquine dose (mg/kg), and other factors. The systematic review protocol was registered with PROSPERO (CRD42023345956). We identified 219 P. vivax efficacy studies, of which 8 provided relevant individual-level data from patients treated with primaquine; all were randomised, parallel arm clinical trials assessed as having low or moderate risk of bias. In the primary analysis data set, there were 1,747 patients with normal glucose-6-phosphate dehydrogenase (G6PD) activity enrolled from 24 study sites across 8 different countries (Indonesia, Brazil, Vietnam, Thailand, Peru, Colombia, Ethiopia, and India). We observed an increasing dose-response relationship between the daily weight-adjusted primaquine dose and day 7 methaemoglobin level. For a given primaquine dose regimen, an observed doubling in day 7 methaemoglobin percentage was associated with an estimated 30% reduction in the risk of P. vivax recurrence (adjusted hazard ratio = 0.70; 95% confidence interval [CI] [0.57, 0.86]; p = 0.0005). These pooled estimates were largely consistent across the study sites. Using day 7 methaemoglobin as a surrogate endpoint for recurrence would reduce required sample sizes by approximately 40%. Study limitations include the inability to distinguish between recrudescence, reinfection, and relapse in P. vivax recurrences.

Conclusions: For a given primaquine regimen, higher methaemoglobin on day 7 was associated with a reduced risk of P. vivax recurrence. Under our proposed causal model, this justifies the use of methaemoglobin as a population-level surrogate endpoint for primaquine antihypnozoite activity in patients with P. vivax malaria who have normal G6PD activity.

Fig 1. Directed acyclic graph showing our hypothesised causal relationships between primaquine-induced changes in blood methaemoglobin concentrations and P. vivax relapse.

Red nodes represent the outcomes of interest: relapse and blood methaemoglobin (measured on day 7, for example), between which the association was estimated. Under this causal model, blood methaemoglobin is a proxy measurement for the hypnozontocidal activity of primaquine (but not on the causal pathway mediating the effect of primaquine on relapse). Host factors include but not limited to patient’s genetics (e.g., those related to CYP2D6 and G6PD), behaviours, age, immunity to P. vivax, and geographical location.

Fig 2. Study and patient selection.

Databases systematically searched were from Medline, Embase, Web of Science, and the Cochrane Library. Patients included in the secondary analyses were subsets of the patients in the primary, time-to-event analysis.

Fig 3. Dynamics of primaquine-induced increases in blood methaemoglobin over time, stratified by primaquine regimen and schizonticidal drug.

Methaemoglobin levels increased after starting primaquine in both regimens, usually reaching a maximum after about a week. Methaemoglobin increased at a faster rate among the 7-day regimen reflecting the higher daily dose taken and generally methaemoglobin started to decrease during the second week, when primaquine was no longer administered. Meanwhile, after peaking at also day 7 for the 14-day regimen, methaemoglobin appears to be at a more constant level during the second week. Boxplot represents the distribution of methaemoglobin levels on a particular day following primaquine. In a boxplot, the box shows the interquartile range with a line for the median, and the whiskers extend to data points within 1.5 times the interquartile range, highlighting potential outliers. Solid curve summarises the data points by partner drug over time. Box width is proportional to the square root of the number of patients. ACT artemisinin-based combination therapy (artesunate/amodiaquine, artesunate/pyronaridine, dihydroartemisinin/piperaquine).

Fig 4. Day 7 blood methaemoglobin (%) as a function of daily mg/kg primaquine dose and P. vivax recurrence status.

There was an increasing trend of day 7 methaemoglobin as daily primaquine dose increased among patients with at least 120 days of follow up. Patients who developed P. vivax recurrences typically had lower day 7 methaemoglobin levels. Top panel highlights comparisons between the dose groups and recurrence status using percentile summaries for day 7 methaemoglobin levels. Bottom panel provides a more detailed view of individual variability in day 7 methaemoglobin levels across different daily primaquine doses received and recurrence statuses. Solid line on the bottom panel denotes a regression line (with a 95% CI) of day-7 methaemoglobin levels on daily primaquine dose. Vertical axis is shown on the logarithmic scale. Box width is proportional to the square root of the number of patients. In a boxplot, the box shows the interquartile range with a line for the median, and the whiskers extend to data points within 1.5 times the interquartile range, highlighting potential outliers (dots).

Fig 5. Forest plot of the relationship between day 7 methaemoglobin levels and the risk of vivax recurrence.

Dashed line represents no association (i.e., hazard ratio = 1). Estimates to the left of the dashed line indicate a lower risk of recurrence with higher day 7 methaemoglobin levels (blue arrow). Conversely, estimates to the right indicate a higher risk of recurrence with higher day 7 methaemoglobin levels (orange arrow). Both the common-effect and random-effects models yielded comparable estimates. Horizontal axis represents dose-adjusted hazard ratios for vivax recurrence associated with a doubling in day 7 methaemoglobin, and is displayed on a logarithmic scale. CI confidence interval, N number of patients.