Postartesunate delayed hemolysis is a predictable event related to the lifesaving effect of artemisinins

Abstract

Patients with severe malaria treated with artesunate sometimes experience a delayed hemolytic episode. Artesunate (AS) induces pitting, a splenic process whereby dead parasites are expelled from their host erythrocytes. These once-infected erythrocytes then return to the circulation. We analyzed hematologic parameters in 123 travelers treated with AS for severe malaria. Among 60 nontransfused patients observed for more than 8 days, 13 (22%) had delayed hemolysis. The peak concentration of circulating once-infected erythrocytes was measured during the first week in 21 patients and was significantly higher in 9 patients with delayed hemolysis than in 12 with other patterns of anemia (0.30 vs 0.07; P = .0001). The threshold of 180 million once-infected erythrocytes per liter discriminated patients with delayed hemolysis with 89% sensitivity and 83% specificity. Once-infected erythrocyte morphology analyzed by using ImageStream in 4 patients showed an 8.9% reduction in their projected area, an alteration likely contributing to their shorter lifespan. Delayed clearance of infected erythrocytes spared by pitting during AS treatment is an original mechanism of hemolytic anemia. Our findings consolidate a disease framework for posttreatment anemia in malaria in which delayed hemolysis is a new entity. The early concentration of once-infected erythrocytes is a solid candidate marker to predict post-AS delayed hemolysis.

Figure 1

Program flowchart. Named-patient program 2011-2013; 123 patients with severe malaria treated with artesunate. D, day.

Figure 2

Three typical patterns of posttreatment anemia in 42 nontransfused patients with severe malaria treated with AS and monitored for more than 8 days. (A1-A3). Mean and standard error of the mean (SEM; vertical bars) of (solid red circles) blood hemoglobin level and (empty red circles) blood reticulocyte count per 10 g/L on day 0, day 2 ± 1, day 7 ± 2, day 14 ± 3, day 21 ± 3, and day 28 ± 3 in subgroups of patients categorized according to their evolution pattern as (A1) rising, (A2) persistent, or (A3) PADH, as defined in the “Materials and methods” section. At least 2 samples were collected before and after day 8 in each patient. (B1-B3) Mean and SEM of (solid black diamonds) plasma LDH 10 IU/L and (solid grey squares) haptoglobin levels (×15 g/L) in the same subgroups. The blood hemoglobin level was normalized against the hemoglobin level on day 0. The delayed drop in hemoglobin levels in patients with PADH is indicated by a black arrow.

Figure 3

Pitting and surface area loss. (A) Quinine- vs AS-induced parasite clearance and pitting rates. Mean and SEM for (solid circles) parasitemia and concentration of (open circles) once-infected erythrocytes (both normalized against parasitemia at day 0) in patients with severe malaria treated either with quinine (n = 11) or AS (n = 16). (B) Concentration of circulating once-infected erythrocytes (billions of cells per liter) in patients with severe malaria treated with AS for whom posttreatment evolution of anemia could be categorized as (solid red circles) PADH (9 patients) and (solid black squares) non-PADH (12 patients). (C) Population frequency of erythrocytes—uninfected erythrocytes (uEs, yellow line; C2-C3), infected erythrocytes (iEs, blue line; C2), or once-infected erythrocytes (o-iEs, orange line; C3)—according to the projected surface areas (μm²) estimated on the bright field picture after identifying erythrocyte subtype by using a combination of immunofluorescent (IF) and Hoechst staining for parasite DNA. Uninfected erythrocytes were IF- and DNA-negative, infected erythrocytes were both IF- and DNA-positive, and once-infected erythrocytes were IF-positive and DNA-negative. Typical aspects are shown in panel C1. Samples were collected from the same patient at (C2) day 0 or (C3) day 4. At least 5000 infected or once-infected erythrocytes and 100 000 never-infected erythrocytes were analyzed. (D) (open circles) Individual values and (open bars) mean projected surface of uninfected and infected erythrocytes in samples from 4 patients on day 0 and of uninfected and once-infected erythrocytes from 4 patients on day 3 ± 1. The lines connect 2 subpopulations analyzed simultaneously in the same sample. The percentages indicate the mean decrease in projected surface area between uninfected and infected erythrocytes at day 0 or between uninfected and once-infected erythrocytes at day 3 ± 1.

Figure 4

Parasitemia, pitting rate, concentration of once-infected erythrocytes, and the risk of evolution toward PADH or other patterns of post-AS anemia in patients with severe malaria. (A) Individual values of parasitemia (as a percentage of infected erythrocytes) in patients with severe malaria treated with AS in whom posttreatment evolution of anemia could be categorized as rising (19 patients), persistent (10 patients), or PADH (13 patients). The dotted line corresponds to the threshold of 4% parasitemia, which provides the best discrimination between PADH and other patterns of post-AS anemia. Percentages at the top of the panel are the proportion of patients above this 4% threshold for each pattern of anemia. (B-C) Individual values of pitting rate (normalized against initial parasitemia [B]), and concentration of once-infected erythrocytes (g/L [C]) in 35 samples from 21 patients with severe malaria treated with AS in whom posttreatment evolution of anemia could be categorized as rising, persistent, or PADH. The dotted lines correspond to thresholds providing the best discrimination between PADH (9 patients) and other patterns of post-AS anemia (12 patients). Percentages at the top of the panel are the proportion of samples above this threshold of 45% pitting rate and above 0.18 billion once-infected erythrocytes per liter for each pattern of anemia. (D) Graphic representation of the relative influences of parasitemia on admission, peak pitting rate, and peak concentration of once-infected erythrocytes on the risk of evolution toward the different patterns of post-AS anemia. The concentration of circulating once-infected erythrocytes was computed by multiplying parasitemia by the pitting rate. Patients with high initial parasitemia and a high pitting rate harbor a high concentration of once-infected erythrocytes (upper right zone of the panel) and are at high risk of intense PADH in the following days or weeks. On the basis of these markers, the risk of subsequent PADH can be predicted before the end of the first week. The risk becomes significant when parasitemia on admission is above 4% and the peak pitting rate on days 2 through 8 is above 45%. On the basis of this data set (panels A-C), a concentration of circulating once-infected erythrocytes on days 2 to 8 greater than 0.18 g/L discriminates patients with PADH from patients with other patterns of anemia more effectively than do parasitemia or the peak pitting rate taken separately. This model also explains why delayed hemolysis has not been observed in severe malaria patients treated with quinine, a compound that induces low pitting rates.