Brief Report: Antimalarial Benefit of HIV Antiretroviral Therapy in Areas of Low to Moderate Malaria Transmission Intensity

Abstract

Background: We previously used mathematical modeling to predict reduced malaria incidence in children with protease inhibitor (PI)-, compared with nonnucleoside reverse transcriptase inhibitor-, based highly active antiretroviral therapy (HAART), in moderate to high malaria transmission areas. These effects were accounted for, in part, by pharmacokinetic (PK) interactions between PIs and artemether-lumefantrine (AL).

Objective: Because of potentially reduced malaria transmission reservoirs in HIV-infected children due to PI/AL PK interactions impacting non-HIV-infected children, we estimate the antimalarial benefit of PI-based HAART in all children, and in HIV-infected children only residing in low to moderate malaria transmission areas.

Design: A dynamic model of malaria transmission was developed to evaluate the PK interaction of PI-based HAART with the antimalarial, AL for preventing malaria.

Methods: To evaluate the benefit of HIV PI-based HAART on malaria incidence, a malaria transmission model with varying degrees of HIV newborn prevalence was developed using recent pediatric clinical trial data in Lilongwe, Malawi.

Results: Comparing situations of low to high HIV newborn prevalence, and low to moderate malaria transmission intensities, our model predicts the combination of PI-based HAART with AL-treated malaria prevents 0.04-24.8 and 0.05-34.5 annual incidences of malaria overall per 1000 children, and saves 0.003-1.66 and 0.003-2.30 disability-adjusted life years per 1000 children, respectively. When incorporating seasonality, 0.01-7.3 and 0.01-5.9 annual incidences of malaria overall per 1000 children, and 0.0-0.5 and 0.001-0.41 disability-adjusted life years per 100 children, are prevented, respectively.

Conclusions: In low to moderate malaria transmission intensity areas, PI-based HAART may reduce malaria events in children when AL is used.

Figure 1.

a) Time until next malaria infection (black dot) and corresponding fit of weibull distribution (green) for HIV infected children treated with PIs, and the time until next malaria infection (magenta dash-dot) and corresponding fit of weibull distribution (blue) for HIV infected children treated with NNRTIs. b) Distribution of the duration of the PI-extended post-treatment prophylactic effect of AL in HIV-infected children on study c) Monthly malaria incidence in all children (HIV-infected and uninfected) and sinusoidal fit (red). d) First-order sensitivity indices for the mean adult mosquito life span (1/d⁰), extrinsic incubation rate (1/τ_M), mean duration of malaria infection recovery (1/γ), mosquito-to-human transmission (β_M), human-to-mosquito transmission (β_H), proportion reporting for treatment to a health care setting (ω), and duration of PI-extended post-treatment prophylactic effect of AL as extrapolated for HIV-infected children (1/δ_PI).

Figure 1.

a) Time until next malaria infection (black dot) and corresponding fit of weibull distribution (green) for HIV infected children treated with PIs, and the time until next malaria infection (magenta dash-dot) and corresponding fit of weibull distribution (blue) for HIV infected children treated with NNRTIs. b) Distribution of the duration of the PI-extended post-treatment prophylactic effect of AL in HIV-infected children on study c) Monthly malaria incidence in all children (HIV-infected and uninfected) and sinusoidal fit (red). d) First-order sensitivity indices for the mean adult mosquito life span (1/d⁰), extrinsic incubation rate (1/τ_M), mean duration of malaria infection recovery (1/γ), mosquito-to-human transmission (β_M), human-to-mosquito transmission (β_H), proportion reporting for treatment to a health care setting (ω), and duration of PI-extended post-treatment prophylactic effect of AL as extrapolated for HIV-infected children (1/δ_PI).

Figure 1.

a) Time until next malaria infection (black dot) and corresponding fit of weibull distribution (green) for HIV infected children treated with PIs, and the time until next malaria infection (magenta dash-dot) and corresponding fit of weibull distribution (blue) for HIV infected children treated with NNRTIs. b) Distribution of the duration of the PI-extended post-treatment prophylactic effect of AL in HIV-infected children on study c) Monthly malaria incidence in all children (HIV-infected and uninfected) and sinusoidal fit (red). d) First-order sensitivity indices for the mean adult mosquito life span (1/d⁰), extrinsic incubation rate (1/τ_M), mean duration of malaria infection recovery (1/γ), mosquito-to-human transmission (β_M), human-to-mosquito transmission (β_H), proportion reporting for treatment to a health care setting (ω), and duration of PI-extended post-treatment prophylactic effect of AL as extrapolated for HIV-infected children (1/δ_PI).

Figure 1.

a) Time until next malaria infection (black dot) and corresponding fit of weibull distribution (green) for HIV infected children treated with PIs, and the time until next malaria infection (magenta dash-dot) and corresponding fit of weibull distribution (blue) for HIV infected children treated with NNRTIs. b) Distribution of the duration of the PI-extended post-treatment prophylactic effect of AL in HIV-infected children on study c) Monthly malaria incidence in all children (HIV-infected and uninfected) and sinusoidal fit (red). d) First-order sensitivity indices for the mean adult mosquito life span (1/d⁰), extrinsic incubation rate (1/τ_M), mean duration of malaria infection recovery (1/γ), mosquito-to-human transmission (β_M), human-to-mosquito transmission (β_H), proportion reporting for treatment to a health care setting (ω), and duration of PI-extended post-treatment prophylactic effect of AL as extrapolated for HIV-infected children (1/δ_PI).