Background
Approximately 10,000 to 12,000 malaria cases imported to Europe are reported to the World Health Organization (WHO) annually[
1]. In Germany, 617 cases of imported malaria were reported to the Robert Koch Institute in 2010, with an overall mortality rate of 0.3% (two cases)[
2]. In a recent retrospective French study on patients with severe imported malaria – 97.8% of which were treated with intravenous quinine - mortality was determined to be 10.5%[
3].
Parenteral artesunate has been shown to be a superior treatment option compared to parenteral quinine in adults and children with severe malaria in endemic countries[
4,
5]. In a large meta-analysis a significant reduction in mortality could be confirmed for adults (relative risk (RR) 0.61) and children (RR 0.76) treated with parenteral artesunate compared to parenteral quinine[
6]. Artesunate showed a superior parasite clearance time compared to quinine. Artesunate is now the first-line treatment for severe malaria according to WHO guidelines[
7].
Regarding short-term safety, hypoglycaemia occurred significantly less frequently in patients treated with artesunate. In children but not in adults, however, the incidence of neurological sequelae at the time of discharge was higher after treatment with artesunate compared to quinine which might be due to the fact that more severely diseased children survived. There was no difference discernible on long-term neurologic sequelae between artesunate and quinine arms 28 days after discharge in the African multicentre trial by Dondorp
et al.[
5,
6]. Little evidence is available in particular on long-term safety due to trial designs and/or lacking infrastructure in endemic regions[
8]. Very recently, cases of late-onset haemolysis have been reported retrospectively in European travellers with imported severe malaria[
9]. The aetiology of this complication is unknown.
An extensive follow-up of patients treated for severe malaria to detect any uncommon findings and complications after parenteral artesunate was implemented at the University Medical Centre Hamburg-Eppendorf. From the beginning of this follow-up programme in August 2011 to January 2012, three patients with hyperparasitaemia were treated with parenteral artesunate, all of which developed late-onset haemolysis.
Discussion
This case series describes three cases of late-onset haemolysis after parenteral artesunate for severe malaria in non-immune travellers returning from sub-Saharan Africa. In all three patients, an initial stabilization of Hb- and a decline of elevated LDH-levels in the first week after treatment initiation was seen. In the second week, Hb levels continued to decline while around day 14 a second rise in LDH levels was observed. At this time point haptoglobin was undetectable and total bilirubin rose again. Re-occurrence of biochemical markers compatible with delayed haemolytic anaemia 1–2 weeks after starting parenteral artesunate contrasted the rapid clinical improvement seen early after administering the first dose. Blood stage parasites were cleared around day 4–5 and all patients remained parasite-free during follow-up. One patient required blood transfusions. In all cases, haemolysis resolved slowly and Hb levels had not returned to normal levels 30 days after treatment initiation.
These findings are in line with a recent report by Zoller
et al.[
9]. Of 25 patients in seven centres for whom data were available, six developed a self-limiting episode of haemolytic anaemia, which was passively detected between days 15 and 32 (median: 15.5). Five of the patients required blood transfusions, all patients recovered fully.
As this case report is only descriptive and no control group (e.g. treated with quinine) exists, it is not possible to state with absolute certainty that the observed haemolysis is drug-related rather than disease-specific. The literature on longer-term follow-up of haematological parameters after severe malaria in general as well as in relation to specific antimalarial treatment is scarce. One report of 192 children treated with quinine showed a (clinically nonsignificant) drop in Hb after treatment and stabilization between days five and 21[
11]. The lack of a standardized follow-up precludes any definite statement on post-treatment haemolysis from this report. In decades of intensive use of quinine, no reports of post-treatment haemolysis have been published, which renders the existence of a comparable complication after quinine unlikely but not impossible.
Haemolytic anaemia is a characteristic finding in acute malaria. The aetiology of malarial anaemia is multifactorial and includes destruction of infected and uninfected erythrocytes as well as impaired erythropoiesis. Both, parasite toxicity as well as host immune mechanisms are causally involved[
12]. Impaired erythropoiesis is reflected by a low reticulocyte production index (RPI) – which was also seen initially in the three patients described in this series (Figure
1). Usually, however, RPI rises concurrently with parasite clearance[
13,
14]. An adequate rise in RPI was delayed until the second week – long after parasites had been cleared and approximately at the time when haemolytic anaemia was diagnosed (Figure
1). Artemisinins have been reported to target erythropoiesis and to reduce reticulocyte counts several days after treatment[
15,
16] possibly explaining the late rise of RPI. Additionally, in the first patient the development of pneumonia could have prolonged inflammation-mediated impairment of erythropoiesis. No secondary infection can however explain the finding in the other two patients.
An explanation of a second peak of haemolysis might lie in the reduced life span of “pitted” erythrocytes. The number of these once-infected erythrocytes rises significantly after treatment with artesunate but not with quinine[
17,
18]. This mechanism would explain the correlation between delayed haemolysis and hyperparasitaemia, as the number of “pitted” erythrocytes is related to parasitaemia. This is supported by the facts, that i) all three patients had parasite levels of >10% infected red blood cells and that ii) all cases with post-treatment haemolysis described by Zoller
et al. also had high parasite levels upon presentation (4 - 30%)[
9].
Another potential explanation of late-onset haemolysis includes a delayed (re-)activation of pro-inflammatory reactions possibly triggered by the rapid and massive destruction of malaria parasite by artesunate and an increased presentation of parasitic antigens.
The fact that the patients described in this report received lower cumulative doses of artesunate than those in the publication by Zoller
et al.[
9] argues against a dose-dependent effect.
One (patient 2) of the two patients receiving immuno-haematological testing developed a positive Coombs’ test with IgG of anti-E specificity. This patient was the only one to receive packed red blood cell transfusions repeatedly after the first dose of artesunate. A delayed haemolytic transfusion reaction cannot be completely excluded in this patient, although it seems very unlikely. Delayed haemolytic transfusion reactions generally occur following alloimmunization after a previous transfusion. The time of onset is usually two to eleven days after the transfusion. The extent of haemolysis is mostly mild without clinical implications as only the transfused erythrocytes are being destroyed[
19,
20]. The titer of the detected antibody in our patient was extremely low and would not explain the amount of haemolysis seen in this patient. More important seems to be the fact that this patient showed the most severe malaria symptoms as well as the highest parasitaemia upon hospitalization (21%). In the case series by Zoller
et al., Coombs’ test was negative in all three patients in whom this test was performed[
9]. All in all, this does not rule out immune-mediated haemolytic anaemia, but this has to be investigated in greater detail.
Two of the patients received drugs with anti-malarial activity prior to the first dose of artesunate. Patient 1 received 750 mg of mefloquine while patient 2 received a seven-day course of doxycycline for suspected bacterial infection. It is unclear whether these medications may have contributed to delayed haemolysis.
Conclusions
After treating three hyperparasitaemic patients with parenteral artesunate, malaria parasites were cleared within a few days and the patients’ clinical condition improved rapidly. Post-treatment haemolysis, however, seems to be a relevant complication in non-immune travellers with imported malaria. Risk factors and pathophysiology are unknown. To gain statistically significant results for patients with imported severe malaria, data from cases at multiple centres will have to be accumulated in a standardized manner. Whether this complication also occurs in children with severe malaria in endemic regions is currently unknown. A regular follow-up of at least one month after treatment with parenteral artesunate including controls of haematological parameters seems to be indicated.
Competing interests
The authors declare having no competing interests.
Authors' contributions
TR, JPC, DW and SS took part in the patients’ care. TR drafted the manuscript with contributions of JPC and GDB. All authors read and approved the final manuscript.