Sequestration of Plasmodium falciparum-infected erythrocytes (PfIE) in the capillaries of the central nervous system (CNS) is the pathognomonic feature of cerebral malaria, a condition frequently leading to death. Sequestration of PfIE in the placental intervillous spaces is the characteristic feature of malaria in pregnancy and is associated with low birthweight and prematurity. Although both patterns of sequestration are thought to result from the expression of different parasite proteins involved in cytoadhesion to human receptors, scant information exists on whether both conditions can coexist and whether this can lead to death. We conducted a prospective autopsy study including all consecutive pregnancy-related deaths in a tertiary-level referral hospital in Maputo, Mozambique, between October 2002 and December 2006. Extensive sampling of all major viscera was performed. All cases showing parasites in any of the viscera were included in the analysis. From 317 complete autopsies PfIEs were identified in ten women (3.2%). All cases showed massive accumulation of PfIE in small capillaries of the CNS but also in most visceral capillaries (heart, lung, kidney, uterus). Placental tissue, available in four cases, showed a massive accumulation of maternal PfIE in the intervillous space. Coma (six women) and dyspnoea (five women) were the most frequent presenting clinical symptoms. In conclusion, massive visceral sequestration of PfIE with significant involvement of the CNS is an infrequent but definite direct cause of maternal death in endemic areas of Africa. The PfIE sequestered in cerebral capillaries and the placenta coexist in these fatal cases.
Pregnant women are more susceptible to Plasmodium falciparum malaria than non-pregnant women or men [1, 2]. The ability of P. falciparum infected erythrocytes (PfIE) to accumulate in the placental intervillous spaces, a condition described as placental malaria, is considered to contribute to this phenomenon . In areas of stable transmission, where women are semi-immune and often asymptomatic during infection, malaria in pregnancy is associated with maternal anaemia, low birthweight and prematurity, but rarely with severe disease [1, 3].
Cerebral malaria (CM) is a serious complication of P. falciparum infection, contributing to over three-quarters of the 655 000 malaria-related deaths per year worldwide . Ninety percent of CM-associated deaths occur in children and, although recent estimations suggest that the malaria mortality burden in adults is larger that previously thought , the general consensus is that CM in adulthood almost exclusively affects non-immune individuals living in non-endemic areas or where endemicity is low .
Sequestration of PfIE in the micro-vasculature of the central nervous system (CNS) and the intervillous spaces of the placenta is considered to be a key feature in the pathogenesis of CM and placental malaria, respectively . The capacity of P. falciparum to sequester is related to its ability to adhere to a variety of host endothelial receptors through parasite antigens expressed on the surface of erythrocytes, the best known of which is P. falciparum erythrocyte membrane protein 1 (PfEMP1) . Despite immense diversity in global var genomic repertoires, genes can be classified into different groups (A–E) . Transcription of A and B groups has been associated with symptomatic and severe malaria [9, 10], whereas C genes have been linked to asymptomatic infections and CM . The single and relatively conserved member of group E, var2csa, is transcribed by placental isolates and provides high-affinity binding to chondroitin sulphate A (CSA) present in placental trophoblasts . The extent to which different subsets of parasites may coexist and cause disease in a patient remains unknown.
In a previous study on the causes of maternal mortality conducted in Mozambique between 2002 and 2004 we identified a series of women who died of malaria [12, 13]. The current study aimed to analyse the pattern of parasite sequestration in women showing parasites in the autopsy study.
All women dying in the Maputo Central Hospital between October 2002 and December 2006 and fulfilling the standard definition of the WHO for a maternal death were initially included in the study. Verbal informed consent was asked of the closest relatives. The study protocol was approved by the National Mozambican Ethics Committee and the Hospital Clinic of Barcelona Ethics Review Committee. A detailed description of the study design and the causes of maternal mortality has been published elsewhere [12, 13]. Three hundred and seventeen complete autopsies were performed, representing 86.8% of the deaths occurring during the study period.
The Maputo urban area has low malaria transmission, although some of the peripheral suburbs are considered to have moderately stable transmission with seasonal variations and the surrounding rural areas have moderate and stable transmission. Malaria incidence is higher during the rainy season (October to May) than during the dry season (June to September) . The reported human immunodeficiency virus (HIV) seroprevalence in pregnant women for 2004 was 20.7% for the Maputo area (Ministry of Health, Mozambique, 2004).
Pathological and laboratory methods
A complete dissection with macroscopic evaluation of each organ by a pathologist using a standardized macroscopic protocol was performed in all cases. At least the following samples were taken from each woman for histological study: CNS (three blocks including frontal and parietal lobes and cerebellum), right and left lung, liver, spleen, right and left kidney, heart and placenta when available. Tissue specimens were fixed in 10% buffered formalin and embedded in paraffin wax using standard procedures. A blood sample (100 µL) was obtained from the inferior vena cava and stored on filter paper. Four-micrometre sections were stained with haematoxylin & eosin. Detection of integrated HIV provirus was determined by qualitative DNA PCR using the standard Amplicor HIV-1 kit (Roche, Johannesburg, South Africa) on blood collected on filter paper.
Selection of cases
In all cases the presence of PfIE was investigated in CNS, liver, lung, spleen, kidney, heart, pancreas, uterus and placenta, by scanning all slides at low magnification (100×) with polarized light. In at least one of the slides of each organ, 50 high-power fields (1000×) were analysed to completely exclude the presence of PfIE. We selected as study group all cases showing unequivocally the presence of any PfIE in any of the organs of the autopsy. The medical records of all patients included in the study were reviewed to retrieve the clinical and epidemiological data.
Microscopy and quantification of sequestered parasites
Under immersion oil (1000×) 100 perpendicularly cross-sectioned capillaries were evaluated and the contents were counted by a single observer in a section of the following organs: CNS, lung, kidney, heart, pancreas and uterus. A capillary was defined as a blood vessel, circular or oval in profile, with a maximum to minimum diameter of <2:1 and with at most one visible endothelial cell nucleus in the wall. All parasites were counted. The number of vessels showing sequestration was quantified and calculated as a percentage of the total vessel count. The number of PfIE in each single capillary was recorded. The presence or absence of intravascular macrophages, and the number per cross-sectioned capillary were also registered, as well as the presence of microthrombi, haemorrhages and any other histological abnormalities. These analyses in placenta, liver and spleen were performed using a high-power field (1000×) as reference.
The analysis was performed using the statistical software Stata, version 12.0 (StataCorp. 2011, College Station, TX, USA). We conducted nonparametric analysis using the Skillings–Mack statistic , a general Friedman-type statistic that can be used in almost any block design with an arbitrary missing-data structure. To provide accurate p-values, 5000 simulations were performed to approximate the distribution of Skillings–Mack values under the null hypothesis, taking into account the particular missing-data structure and tied rankings. The level of significance was set at p <0.05.
PfIE were identified in ten of the 317 autopsies (3.2%). No differences were recorded in the proportion of cases showing PfIE during the period of the study. Median age of the patients was 24.5 (range 17–34) years. All patients were black. The epidemiological and clinical features are shown in Table 1. All but one death occurred during the period of high malaria transmission. Coma and dyspnoea were the most common clinical symptoms (six patients and five patients, respectively). Five patients were HIV positive.
Table 1. Date of death, age, gestational status, clinical symptoms, presence or absence of parasitaemia in peripheral blood and human immunodeficiency virus (HIV) status of the patients included in the study
Date of death
Puerperium (3 days)
Post-abortum (4 days)
Gestation (28 weeks)
Gestation (20 weeks)
Gestation (26 weeks)
Puerperium (2 day)
Gestation (18 weeks)
Puerperium (3 days)
Post-abortum (2 days)
Puerperium (2 days)
In all cases the visceral capillaries showed a massive accumulation of PfIE. No other pathological lesions were identified in the autopsy in nine out of the ten women. Only one patient positive for HIV showed a disseminated Kaposi's sarcoma with massive involvement of the liver, pancreas and abdominal lymph nodes. In all ten women PfIE were observed throughout the whole CNS, including the cerebellum (Fig. 1a,a′), although some minor variations in the percentage of parasitized capillaries and in the number of parasites per capillary were observed from area to area. PfIE were identified in both the white and the grey matter. In addition to parasites in the CNS, all cases showed PfIE in the liver, spleen, uterus and placenta, 9/10 in the lung (Fig. 1b), 8/10 in the kidney and 5/7 in the heart (Fig. 1b′). Most PfIE sequestered in the microvasculature of both the CNS and other viscera were mature stages characterized by the presence of pigment dots in the cytoplasm. Fig. 2 shows the percentage of capillaries containing PfIE (Fig. 2a), mean number of parasites per capillary cross-section (Fig. 2b), and percentage of capillaries with macrophages containing haemozoin (Fig. 2c), in the CNS, kidney, lung, heart and uterus. All these counts tended to be higher in the CNS than in the other organs, but reached statistically significant values only for percentage of capillaries showing pigmented macrophages (p 0.044) and were borderline for mean number of parasites per capillary (p 0.063). The liver and spleen showed sequestration of parasites in the hepatic sinusoids and spleen cords and also a markedly increased number of haemozoin-laden enlarged Kupffer cells and spleen histiocytes. In the four cases in which the placenta was available for study a high density of PfIE in the intervillous space was observed (Fig. 1c,c′). In all cases massive accumulation of PfIE in the placenta coexisted with massive sequestration in the CNS.
Associated pathological findings were frequently identified in the CNS and in the lung, and occasionally in the kidney. Lesions in the CNS included ring haemorrhages in the white matter in four cases (Fig. 3a,a′) and microthrombi in capillaries in five cases (Fig. 3b). Monocytes with phagocytosed haemozoin accumulated within microvessels containing PfIE, but were not present in the adjacent neuropil (Fig. 3b′). Lung lesions included oedema in nine cases, haemorrhage in four cases and congestion and pneumonia in one case each (Fig. 3c,c′). Microthrombi in the glomerular capillaries were found in one case.
In this autopsy study massive visceral sequestration of PfIE was an infrequent but relevant cause of maternal mortality in women living in a malaria-endemic area of sub-Saharan Africa (3.2% of all maternal deaths). All cases were characterized by massive sequestration of parasites not only in the CNS, but also in other organs (lung, kidney, heart, liver, spleen, pancreas, uterus). Interestingly, in all cases in which the placenta was available for study, massive accumulation of parasites was also identified in the placenta.
Massive sequestration of PfIE in microvessels of the CNS was a major histological finding observed in the women from this study. This finding has been observed in all previous pathological studies of brain tissue from patients dying of CM [16-18], and it is established that the definitive diagnosis of CM can only be made on post-mortem examination [12, 19]. Strikingly, only six of the women from this series presented with coma, whereas nine women presented with other symptoms such as dyspnoea, haemorrhage or severe anaemia. Although our results are limited because of the small number of cases, no differences in terms of the severity of the sequestration or the presence or absence of associated lesions in the CNS were observed between women with or without neurological symptoms. In keeping with these findings no differences in the clinical presentation (coma duration or Blantyre coma score) have been found between children showing sequestration in the CNS as the only finding, or those with sequestration associated with intravascular or perivascular lesions .
The pathological role of extra-cerebral parasite sequestration has largely been neglected. However, recent findings indicate that this phenomenon is observed in a significant number of children with CM . These findings suggest that this sequestration may have important consequences for the patients. Indeed, the pathogenesis of CM is poorly understood. The sequestration of PfIE in visceral deep vessels is a normal step in the life cycle of the P. falciparum parasite but does not always trigger severe disease. Some observations in both human and animal models indicate that the disease severity, the clinical outcome and the degree of sequestration of parasites in vital organs including the brain is strongly associated with the total parasite burden [21, 22]. A particularly striking finding of the present study is the coexistence of massive sequestration of PfIE both the CNS and the placenta. This suggests that infections may consist of multiple parasite populations, each one expressing a different var gene, which results in binding to receptors in different organs of the host. In keeping with this hypothesis are previous studies showing that PfIE from pregnant women have been shown to adhere to other receptors , to simultaneously up-regulate other genes , and to express non-VAR2CSA PfEMP1 in the membrane , suggesting that VAR2CSA may not be the only parasite variant infecting pregnant women, and that a proportion of parasites may express other var genes such as those containing domain cassette 8, which have recently been shown to be over-expressed in CM, hyperparasitaemia or severe malarial anaemia [10, 26, 27]. Alternatively, high parasite density infections might be able to accumulate in all different organs without having a specific tropism towards a single organ, either because the PfEMP1 expressed by these parasites are able to bind to a ligand expressed by all human tissues, or because of mechanical alterations of the blood flow or the deformability of the membrane of infected erythrocytes that facilitate the sequestration of the parasite.
Although 88% of all malaria deaths are thought to occur in children under the age of 5 years , the findings of this study show that severe malaria with massive visceral sequestration of PfIE can also occur in pregnant women living in endemic areas of sub-Saharan Africa. In the current study the deaths occurred in women who lived in an urban area where the endemicity of malaria is likely to be low . However, these findings also indicate that the widely accepted assumption that maternal deaths directly attributable to malaria occur only in areas of unstable malaria transmission and do not occur in sub-Saharan Africa needs to be revised. The percentage of women positive for HIV in this series (50%) was slightly higher than the 29.4% reported in women attending a rural antenatal clinic in Mozambique , but similar to the overall percentage of HIV positivity observed in the general series of maternal mortality .
This study has several limitations. Clinical data were retrieved retrospectively from the clinical records, precluding an adequate correlation between the pathological findings and the clinical symptoms. Particularly, the absence of clinical data resulted in a significant decrease of the number of deaths attributed to malaria from the 10.1% reported in our original report in the period 2002–04 to the 3.2% in this study, as only cases with unequivocal PfIE in the histological analysis could be included in the series. Plasmodium genotypic and phenotypic characterization was not possible because of the suboptimal preservation of tissues, a consequence of the delay in performing routine autopsies. To guide research on effective vaccines in pregnancy, further studies are needed to characterize malaria parasites sequestering in different organs at a molecular and immunological level. This will require performing autopsies within a few hours of death and the adequate freezing and storage of biological material, all of which is usually difficult to achieve in many developing settings.
In conclusion, we have shown that massive sequestration of parasites, although infrequent, is a definite and direct cause of maternal mortality in areas of Africa where malaria is endemic. We have also shown that sequestration of P. falciparum parasites in cerebral capillaries and the placenta coexist in these patients.
We thank all the families of the women included in the study. We also thank Dr Elias Walle and the staff of the Maternity Ward and Pathology Departments of the Hospital Central de Maputo, whose support made this study possible. The authors are grateful to the staff of the ‘Centro de Investigaçao de Manhiça’ for their logistical assistance and to Laura Puyol for her help with the PCR detection of HIV. The study was supported by the ‘Fondo de InvestigacionesSanitarias del Instituto de Salud Carlos III’ (FIS) through career development fellowship to Cleofé Romagosa (number CM03/00125), CP-04/00220 to Alfredo Mayor and a ‘Premi fi de residencia Emili Letang 2011’ from the Hospital Clínic to Paola Castillo. This work was partly supported by a grant from The European Commission Research Directorates General, Fifth Framework (Contract PREMA-EU- ICA 4 CT-2001-1110012) and by a grant from the FIS (number PI060207). The ‘Centro de Investigaçao em Saude de Manhiça’ receives core funding from the Spanish Agency for International Cooperation. Results presented in part at the United States & Canadian Academy of Pathology 99th Annual Meetings in Washington, DC in March 2010.
None of the authors have any conflict of interest.