Background
Malaria remains a major cause of mortality and morbidity worldwide, with 229 million cases and 409,000 reported deaths in 2019 [
1]. Furthermore, it is likely that the disruption of services due to the current COVID-19 pandemic may significantly increase the number of malaria cases and malaria deaths [
2]. Malaria-associated anaemia is one of the most important complications of malaria [
3‐
9]. Hospital-based studies have demonstrated that severe anaemia can occur in patients infected with
Plasmodium falciparum,
Plasmodium vivax and
Plasmodium malariae, and in all three species increases the risk of death [
3]. More recently, cross-sectional community surveys have demonstrated that asymptomatic and sub-microscopic infections are also associated with a high risk of anaemia, contributing significantly to the overall burden of malaria-associated anaemia [
10,
11].
Malaria-associated anaemia is multi-factorial. Although occurring in part due to the rupture of parasitized red blood cells (RBCs), the major contributor to malaria is the loss of unparasitized RBCs, with previous studies estimating that the ratio of the loss of parasitized to unparasitized RBCs is 1:8 in
P. falciparum [
12,
13] and 1:34 in
P. vivax [
14]
. Potential contributors to this loss of unparasitized cells include increased free radical damage [
15], the loss of erythrocyte surface complement regulatory proteins [
16] and the production of anti-phosphatidylserine antibodies [
17]. Dyserythropoiesis caused by bone marrow suppression secondary to the direct effects of the parasites as well as the effect of cytokines [
18‐
21], and inflammation-induced iron deficiency [
22], are other key contributors to malaria-associated anaemia. Shortened red cell lifespan following artesunate therapy [
23] may also be contributory.
The induced blood-stage malaria (IBSM) volunteer infection model developed at QIMR Berghofer in 1995 has recently assumed a key role in anti-malarial drug development [
24]. To date, over 400 participants have been enrolled in these studies at QIMR Berghofer, including 342 inoculated with the fully sensitive
P. falciparum 3D7 strain, 15 with the K13 artemisinin resistant stain, and 46 with
P. vivax. In these studies, parasitaemia is closely monitored with a highly sensitive quantitative PCR targeting the 18S rRNA gene, and frequent blood sampling occurs to monitor haematological parameters throughout the course of infection. Therefore, these studies provide a unique opportunity to investigate the haematological response that occurs in early
P. falciparum and
P. vivax blood-stage infection.
In this study, haematology data from 26 IBSM volunteer infection studies undertaken at QIMR Berghofer over the last ten years were analysed, in order to describe the haematological response to P. falciparum and P. vivax infection, and to evaluate factors associated with the fractional fall in haemoglobin. The haematological and parasitaemia data were also used to estimate the relative contribution of the loss of parasitized and unparasitized cells to the total malaria-attributable haemoglobin loss.
Discussion
In this study, the haematological response to early experimental P. falciparum and P. vivax infection is described. The analysis demonstrates that in both P. falciparum and P. vivax, experimental infection results in an ~ 11% fractional fall in haemoglobin, approximately half of which occurs prior to treatment. The haemoglobin nadir occurred ~ 12 days after treatment in participants inoculated with P. falciparum and 8 days in participants inoculated with P. vivax, returning to normal by 28 days in P. falciparum and 20 days in P. vivax.
Volunteer infection studies are associated with frequent blood sampling, and it was estimated that phlebotomy-related losses accounted for ~ 60% of the total fractional fall in haemoglobin. Nonetheless, the malaria-attributable loss from early experimental malaria, with parasitaemias only just reaching the limit of detection by microscopy, still accounted for a fractional fall in haemoglobin of ~ 4% in Pf3D7, ~ 7% in PfK13 and ~ 5% in
P. vivax, after accounting for phlebotomy related losses. It should be noted that in experimental malaria the duration of parasitaemia prior to treatment is only a few days, in contrast to endemic regions where low-level asymptomatic parasitaemia may persist in partially immune individuals for prolonged periods, thus the haemoglobin loss may be expected to be substantially greater. This is supported by recent studies from endemic regions, which have demonstrated a significant burden of anaemia associated with submicroscopic parasitaemia [
10,
11].
In participants inoculated with
P. falciparum, the fractional fall in haemoglobin was greater in those inoculated with an artemisinin-resistant PfK13 strain, compared to those inoculated with the artemisinin-sensitive Pf3D7 strain. Contributing factors that may have accounted for this greater fractional fall include a later day of treatment and higher parasitaemias in participants inoculated with PfK13 compared to Pf3D7, as well as the slower parasite reduction ratio, and the recrudescence that occurred in all of the PfK13 participants following treatment with artesunate. The day of haemoglobin nadir also occurred later in the PfK13 participants, and no participant recovered their haemoglobin prior to the end of study, again likely reflecting delayed parasite clearance and recrudescent parasitaemia. In the PfK13 participants the parasite reduction ratio was significantly and inversely correlated with the fall in haemoglobin, with slower parasite clearance associated with greater haemoglobin fall. Although this study was not designed to directly compare haemoglobin losses in participants inoculated with PfK13 vs Pf3D7, the results reported here are consistent with data from clinical studies reporting a greater incidence of anaemia in patients with drug-resistant
P. falciparum [
54,
55] and those given drugs with slower parasite clearance effects [
56‐
59]. Thus, while this study did not allow direct comparison of the effect of different drugs on haemoglobin loss, the impact of parasite clearance on rates of anaemia should be considered in clinical trials evaluating antimalarials.
Consistent with the greater fall in haemoglobin in participants inoculated with PfK13, all of whom recrudesced, within the Pf3D7 group those who recrudesced also experienced a significantly greater fall in haemoglobin. These data are consistent with previous studies which have also demonstrated higher rates of anaemia associated with parasite recrudescence [
13,
60].
It is also possible that the greater fractional fall in haemoglobin observed in participants inoculated with PfK13 may have been due in part to the fact that these participants were treated with artesunate. Post-artesunate anaemia is well described, and results from delayed clearance of once-infected erythrocytes [
23]. However, it is unlikely that this would be a significant contributor to erythrocyte loss at such low parasitaemias, and no participant inoculated with PfK13 had elevated markers of haemolysis [
44].
In participants inoculated with
P. vivax, we found that delaying day of treatment was associated with a greater fall in haemoglobin. This is also consistent with clinical studies reporting correlations between the time since symptom onset and degree of anaemia [
61‐
64], and highlights the importance of early initiation of treatment.
In participants inoculated with
P. vivax, a correlation between parasitaemia (whether measured as peak parasitaemia, or pre-treatment total parasite burden) and the fractional fall in haemoglobin was observed. This is as expected, and consistent with clinical data demonstrating a correlation between parasitaemia and anaemia [
63,
64]. Unexpectedly, this association was not observed in the much larger group of participants inoculated with Pf3D7. It is possible the heterogeneity of the Pf3D7 studies (with different treatment days and different drugs evaluated), together with the low parasitaemias, may have obscured any association between parasitaemia and fractional fall in haemoglobin.
In this study we attempted to quantify the loss of parasitized cells as a proportion of the overall malaria-attributable erythrocyte loss. Similar analyses have been conducted previously, including in clinical malaria in endemic regions [
13], and in neurosyphilis patients receiving malariotherapy [
12,
14]. In the former study, Price et al. [
13] evaluated the haematological response to
P. falciparum malaria in over 4000 children and adults in Thailand during 1990–1995, and estimated that parasitized cells accounted for 7.9% of the total malaria-attributable loss. A similar estimate was obtained by Jakeman et al. [
12], who used data from neurosyphilis patients undergoing malariotherapy to estimate that
P. falciparum parasitized cells accounted for ~ 10.5% of erythrocytes lost. In neurosyphilis patients inoculated with
P. vivax, Collins et al. [
14] estimated that only 2.9% of the reduction in haemoglobin was due to destruction of parasitized erythrocytes. In the current study involving very low parasitaemias, our analyses demonstrated that parasitized cells accounted for only 0.015% of erythrocytes lost in volunteers inoculated with Pf3D7, and 0.022% in those inoculated with
P. vivax. This suggests that in submicroscopic infections, the relative contribution of the loss of unparasitized cells to malarial anaemia is likely much greater than that seen in higher parasitaemia infections.
The mechanisms mediating the loss of unparasitized cells at such low parasitaemias remain incompletely understood. In malaria volunteer infection studies, despite the low parasite counts, participants still experience a significant inflammatory response, with elevated levels of IFN-γ and IL-6 [
65‐
67] potentially contributing to inhibition of erythropoiesis [
68]. Additional contributors to the loss of unparasitized cells in acute clinical malaria include haemolysis, decreased red blood cell deformability, antibody and complement binding to erythrocytes, loss of complement regulatory proteins on the surface of unparasitized erythrocytes [
69] and increase in splenic size [
70] with associated splenic clearance of uninfected erythrocytes [
71]. However, the role of these processes in low-parasitaemia infections such as volunteer infection studies has been more difficult to define [
16].
This study had a number of limitations. First, there was substantial heterogeneity between and within the individual malaria volunteer infection studies, including the drug used, the day of treatment, and the study duration, making it difficult to account for factors associated with haemoglobin loss. Second, in the
P. vivax infection studies, the calculation of total parasite burden prior to antimalarial treatment did not account for parasite sequestration. Recent studies have suggested that in chronic
P. vivax infection, a very high proportion of parasitized erythrocytes are sequestered in the spleen [
52], and it has also been shown that splenic accumulation may occur even in early
P. vivax infection [
70]. It is possible that this may in part explain the low contribution of peripheral parasitized cells to the malaria-attributable erythrocyte loss, in this study and in others [
14]. Finally, this study was conducted in malaria-naïve healthy adults, and these data may not be applicable to children, or to adults in malaria-endemic regions where immunity maybe present.
In summary, this study demonstrates that a small but statistically significant fall in haemoglobin occurs in experimental malaria infection, despite parasitaemias that are only just at the level of microscopic detection. This study adds to studies from endemic regions reporting a significant burden of anaemia from asymptomatic and submicroscopic infections, highlighting the importance of treating these groups to reduce the overall burden of anaemia. Finally, this detailed description of the expected haemoglobin loss in malaria volunteer infection studies can be used as a baseline against which to compare the haemoglobin losses that may occur when new antimalarial drugs are evaluated in this model.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.