Linked-evidence modelling of qualitative G6PD testing to inform low- and intermediate-dose primaquine treatment for radical cure of Plasmodium vivax

Abstract

Background: Radical cure of Plasmodium vivax infections is key to the control of vivax malaria. However, the standard doses of 8-aminoquinoline drugs used for radical cure can cause severe haemolysis in G6PD-deficient patients. The availability of near-patient G6PD tests could increase use of primaquine (PQ), however direct evidence of the impacts that G6PD testing has on downstream patient outcomes, such as haemolysis and recurrence is lacking.

Methodology/principle findings: A linked-evidence model was created to investigate changes in the number of severe haemolysis events and P. vivax recurrences within 6 months of treatment when qualitative G6PD testing was used to guide PQ treatment (0.25mg/kg/day for 14 days and 0.5mg/kg/day for 7 days), compared to prescribing 14-day PQ with no G6PD testing. In the model patients identified as G6PD-deficient received 8-week PQ (0.75mg/kg/week). The model was used to simulate scenarios with 1%, 5% and 10% prevalence of G6PD-deficiency (G6PDd) in theoretical populations of 10,000 male and female P. vivax patients and initially assumed 100% adherence to the prescribed PQ regiment. Results illustrate that G6PD testing to guide the 14-day PQ regiment reduced severe haemolysis by 21-80% and increased recurrences by 3-6%, compared to applying the 14-day PQ regiment without G6PD testing. Results for the 7-day PQ regiment informed by G6PD testing were mixed, dependent on G6PDd prevalence and sex. When adherence to the PQ regiments was less than perfect the model predicted reductions in the number of recurrences at all prevalence levels, provided adherence to 7-day PQ was 5-10% higher than adherence to the 14-day regiment.

Conclusions/significance: Introduction of G6PD testing to guide PQ treatment reduces severe haemolysis events for the 14-day regiment, and the 7-day regiment in higher G6PDd prevalence settings, compared to use of 14-day PQ without G6PD testing when all patients adhere to the prescribed PQ treatment. At a population level, there were increases in recurrences, but this could be resolved when the 7-day regiment was used and had superior adherence compared to the 14-day regiment.

Fig 1. Analytical framework for using a qualitative G6PD test to guide PQ treatment decisions.

[1] Is there direct evidence that using a qualitative G6PD test reduces P. vivax infections in the 6 months following treatment? [2] What is the prevalence of G6PD activity <30% of adjusted-male median in the target group? [3] Can G6PD tests be used to accurately classify G6PD activity? [4] Does treatment with PQ result in fewer future P. vivax infections? [5] What are the consequences of not having access to near patient G6PD testing? [6] Does treatment result in adverse effects, specifically severe haemolysis?

Fig 2. Model structure.

Probabilities are defined in Table 1. ID: intermediate-dose.

Fig 3

Median number of severe haemolysis events and P. vivax recurrences per 10,000 male (left) and female (right) P. vivax patients as access to G6PD testing transitions from 0% (baseline) to 100% for low-dose (top panel) or intermediate-dose PQ scenarios (bottom panel), in 20% increments. Note the different axis scaling for males and females.

Fig 4. One-way sensitivity analysis of low-dose PQ treatment guided by G6PD testing (100% access), compared to low-dose PQ without G6PD testing.

Values for each variable were set at the minimum (blue) and maximum (red) values in independent simulations to investigate the impact on model predictions for number of severe haemolysis events (top) and P. vivax recurrences within 6 months of treatment (bottom) in male patients with 1% G6PDd prevalence. Only variables impacting each outcome are shown.

Fig 5. One-way sensitivity analysis of intermediate-dose PQ treatment guided by G6PD testing(100% access), compared to low-dose PQ without G6PD testing.

Values for each variable were set at the minimum (blue) and maximum (red) values in independent simulations to investigate the impact on model predictions for number of severe haemolysis events (top) and P. vivax recurrences within 6 months of treatment (bottom) in male patients with 1% G6PDd prevalence. Only variables impacting each outcome are shown.

Fig 6. Percentage change in the number of recurrences within 6 months of treatment for low-dose PQ guided by G6PD testing in 100% of patients versus baseline (low-dose PQ without G6PD testing) in a population with 1% G6PDd among P. vivax patients.

Shaded regions represent 10^th– 90^th percentile from PCA.

Fig 7

Contour map showing percentage change in median number of P. vivax recurrences within 6 months of treatment for intermediate-dose PQ guided by G6PD testing (100% access) versus baseline (low-dose PQ without G6PD testing) for varying levels of adherence to low-dose and intermediate-dose PQ regiments in a patient population with 1% G6PDd. It is assumed that adherence to the intermediate-dose PQ regiment is always the same or better than adherence to low-dose PQ.