Use of area under the curve to evaluate the effects of antimalarial drugs on malaria-associated anemia after treatment.

ABSTRACT

To evaluate the effects of antimalarial drugs on Plasmodium falciparum malaria-associated anemia, we use the area under the curve (AUC) of anemia levels after treatment as an approach to combine their duration and magnitude. The method involves numeric estimation, by trapezoidal rule, of AUC from a plot of deficit in hematocrit levels from 30% (the lower threshold of normal) versus time in anemic children. Using the method, we evaluated, in randomized trials, the effects of artesunate-mefloquine versus mefloquine alone and artemether-lumefantrine versus amodiaquine-artesunate on the time course of recovery from malaria-associated anemia in 109 children. Anemia resolution times were similar (10.9 ± 6.2 [standard deviation] versus 13.3 ± 8.9 days, P = 0.2), but mean AUC was significantly lower in artesunate-mefloquine- compared with mefloquine-treated children (35.5 ± 7.1 [standard error of mean] versus 49.8 ± 11.3 %·h, P = 0.02) indicating larger exposure to anemia in mefloquine-treated children. In artemether-lumefantrine- and amodiaquine-artesunate-treated children, both anemia resolution times (8.6 ± 5.3 [standard deviation] versus 8.6 ± 4.8 days, P = 0.98) and mean AUC (57.1 ± 12.9 [standard error of mean] versus 46.3 ± 8.7 %·h, P = 0.74) were similar. Estimation of AUC appears more robust than estimation of anemia resolution time in evaluating antimalarial drug effects and can be used in both observational studies and clinical trials assessing the effects of therapies on malaria-associated anemia.