Severe falciparum malaria in pregnancy in Southeast Asia: a multi-centre retrospective cohort study

Plasmodium falciparum infection remains the main cause of death from malaria worldwide [1‐3]. Although other malaria species can occasionally cause severe malaria and contribute to mortality [3], untreated severe falciparum malaria is commonly fatal [2]. If treated promptly with intravenous artesunate, however, the mortality of severe malaria is reduced by one-third compared with the previously recommended quinine [3‐5]. Since 2006, intravenous artesunate has been recommended by the World Health Organization (WHO) as the first-line treatment for all patients with severe malaria, including pregnant women in all trimesters [6]. In pregnancy, the risk of developing severe malaria is higher than in non-pregnant women in both high- and low-transmission areas [7] (2–3 times higher in low-transmission areas [8]), and the mortality of severe malaria is usually reported as higher in pregnancy [2, 9, 10]. However, the Southeast Asian Quinine Artesunate Malaria trial (SEAQUAMAT), a cornerstone multi-centre randomised controlled trial (RCT) on the treatment of severe malaria, reported that mortality was lower in pregnancy (10%, 5/49) than in non-pregnant adults (19%, 65/337) [4]. Other studies have reported that mortality from severe malaria was not significantly different between pregnant and non-pregnant groups [8, 11]. These differences could be explained by the different inclusion criteria, within the definition of severe malaria, of sub-groups with a relatively good prognosis.

In contrast to the large prospective series of severe malaria in children and non-pregnant adults that have been reported [4, 12], there are a few studies on severe malaria in pregnancy [13]. Certain clinical or laboratory findings (e.g. coma, acidosis) are associated with an increased risk of death within the broader definition of severe malaria [5, 9, 14] but this has not been studied specifically in pregnancy. Although the priority in treating severe malaria in pregnancy is to save the life of the mother, severe malaria carries a substantial risk of adverse foetal and neonatal outcomes: a study in Rwanda showed the risks of stillbirth, preterm birth and low birth weight were higher in mothers who had severe malaria compared with those who had uncomplicated malaria [15]. In lower-transmission areas, these risks of adverse outcomes from severe malaria are regarded as higher [11]. Few studies have described foetal outcomes following severe malaria in pregnancy particularly beyond the acute phase. For these reasons, we retrospectively reviewed the records of severe malaria in pregnancy managed by the Shoklo Malaria Research Unit (SMRU) on the Thailand-Myanmar border, where pregnant women are followed prospectively in antenatal clinics until delivery, together with the data collected in prospective severe malaria studies of hospitalised patients in six Southeast Asian countries conducted or coordinated by the Mahidol-Oxford Tropical Medicine Research Unit (MORU) and the Oxford University Clinical Research Unit (OUCRU). Our aim was to describe maternal and foetal mortality and morbidity associated with severe falciparum malaria in pregnancy (as defined by the WHO severe malaria criteria) and to identify the factors predicting higher maternal mortality.

In total, 213 episodes of severe falciparum malaria in pregnancy were included in this pooled analysis (Fig. 1). These were in six Southeast Asian countries: Bangladesh (n = 18), India (n = 1), Indonesia (n = 10), Myanmar (n = 21), Thailand (n = 13) / Thailand-Myanmar border (n = 123) and Vietnam (n = 27).

Severe falciparum malaria is an important cause of maternal death in malaria-endemic areas, particularly in lower-transmission areas such as those in most of Asia [2, 11]. The overall maternal mortality in severe malaria in pregnancy in Asia including the current study was estimated to be 25.1% (19 cohorts, range 0–70.6%, 95% CI 15.2–36.3%, I² 87.1%). This study included two different series of pregnant women with severe falciparum malaria, one community-based and the other hospital-based, showed clearly that, whereas falciparum malaria causing vital organ dysfunction carries a high mortality, severe anaemia or hyperparasitaemia only can present in ambulant women and carry a good prognosis if they are treated properly. Indeed, oral treatment under careful supervision was sufficient in 70% of the groups with either hyperparasitaemia only or severe anaemia only. This marked difference in mortality between the different WHO severe malaria criteria probably explains the variation in reported malaria-related mortality in pregnancy in the international literature, and it raises the question of whether these should be included in the criteria for a diagnosis of severe malaria in pregnancy.

Definitions of severe malaria are valuable in triage and in prompting emergency treatment. In this study, coma, respiratory failure and hypotension were each independently associated with mortality in pregnant women. Similar findings have been reported previously in both children and adults showing that coma, increased blood urea, metabolic acidosis, shock and pulmonary oedema were associated with increased mortality [40]. In areas of high transmission, severe anaemia is the most common reason for hospitalising young children with malaria [1, 41], who bear the brunt of global severe malaria mortality [2, 3]. Anaemia is also very common in pregnant women and is often multifactorial [41]. Pregnant women who are already anaemic may have a further reduction in haemoglobin when infected with malaria, and thereby meet the severe anaemia criterion for severe malaria. Malaria itself can cumulatively increase the risk of anaemia [41] despite routine ANC screening for anaemia and provision of haematinics for mild anaemia (haematocrit < 30%). In the current study, women presenting with severe anaemia alone commonly had multiple episodes of malaria before developing severe anaemia. In a study of 418 pregnant women with malaria at SMRU, only 2% did not have anaemia (haematocrit < 30%) detected between day 0 and day 42 [42]. Severe anaemia (haematocrit < 20% or haemoglobin < 7 g/dL) is often reported as the most common presenting sign of severe malaria in pregnant women [8, 10]: while this has adverse consequences for the foetus increasing the risk of foetal loss and foetal growth restriction, the anaemia is often well compensated and, if maintained in the 5–7 g/dL range, may not pose a risk for the mother—particularly when blood transfusion is available. Life-threatening anaemia (< Hb 4 g/dL) [7] was uncommon in this study (2.5%, 5/198). A reconsideration of haemoglobin thresholds, particularly in pregnancy, may be warranted.

Hyperparasitaemia reflects low immunity against falciparum malaria, such as in younger children or pregnant women [1, 3, 43]. Although higher parasitaemia is generally associated with severe clinical manifestations, the relationship is not linear [3, 41, 44]. Previous studies have shown that uncomplicated hyperparasitaemia can be treated successfully with oral ACTs without mortality [19, 45], although longer courses of artemisinin derivatives (5–7 days) [19] and close monitoring are required. The relationship between parasite density and mortality is complex and depends on background immunity, stage of parasite development at presentation and access to artemisinin-containing antimalarials, which prevent circulating young ring stage parasites maturing and resulting in red cell sequestration [2, 3]. The proportion of patients with severe malaria with hyperparasitaemia alone therefore affects the overall severe malaria mortality (particularly in the context of artemisinin use). The optimum cut-off point for the definition of hyperparasitaemia in pregnancy is unclear.

All three maternal deaths after parenteral artesunate/artemether in the SMRU cohort occurred within 24 h [4]. Patients who die rapidly may have presented too late in the disease process—emphasising the importance of providing parenteral artesunate promptly to suspected severe cases. In this cohort, the majority (61.8%) of severe episodes occurred in women who came to ANC for the first time or after more than 2 weeks of absence. Encouraging women to come to ANC early in the first trimester and improving regular attendance may reduce the incidence of severe malaria.

Increased risks of foetal loss and SGA after severe malaria in pregnancy, particularly among those with organ dysfunction, were reported previously in a small case series [30]. The majority of foetal losses (84.0%) occurred within the first week of diagnosis; 44.7% of the SMRU severe malaria with organ dysfunction group had either maternal death, foetal loss or preterm birth within 1 week. A higher risk of foetal loss and preterm birth in severe malaria compared with uncomplicated malaria has also been reported in sub-Saharan Africa [15].

There are several limitations to this study. First, some of the WHO severe malaria criteria were unavailable. Important determinants of outcome (e.g. renal failure and metabolic acidosis) could not be always assessed and blood glucose was not monitored frequently particularly when the patient was conscious. These missing criteria require laboratory investigation and are frequently unavailable at the primary healthcare level in malaria-endemic countries. The results of this study have practical utility for settings where most malaria is diagnosed and treated, i.e. in rural and remote settings. The number of fulfilled criteria [14] can be another useful indicator to be assessed if information is complete. Although co-morbidities that are known to be associated with maternal mortality were not measurable except for eclampsia/pre-eclampsia, HIV infection was an unlikely comorbidity on the Thailand-Myanmar border where the seroprevalence has been reported as < 0.5% [46]. Second, much of the analysis was based on a single cohort at SMRU where birth outcomes could be compared, whereas the hospital-based studies provided data only on maternal outcomes. At SMRU, early and regular screening and prompt effective treatment has been provided for both malaria and anaemia, which could have resulted in a lower incidence of severe malaria and lower mortality from severe anaemia compared with other studies. On the other hand, the substantial prevalence of vivax malaria in this area and the consequent frequent recurrences are significant contributors to anaemia in pregnancy [41]. Furthermore, the intensive care and specialist support, management and characteristics were different between two series. Pregnant women with hyperparasitaemia or severe anaemia only were systematically treated with oral antimalarials at SMRU consistent with treatment evidenced in non-pregnant patients in the same setting [19, 45], but even without intravenous treatment, these two groups had better prognosis than those with other severe malaria criteria. These differences reflect the heterogeneity and variability of reports of severe malaria in pregnancy in the past literature, which make comparisons with these previous studies difficult. Third, data of severe malaria in the postpartum period were limited in our study and also in the literature. Although the placenta, the site of sequestration of falciparum malaria parasites, no longer exists after delivery, hormonal changes take place over the several weeks of postpartum period, and little is known about the recovery of immunity which has been altered by pregnancy [47]. Finally, the cohorts included in this study were in Southeast Asian countries, where the transmission of malaria is low, as are levels of falciparum premunition in women of childbearing age. The observed lack of association between gravidity and maternal mortality among women who developed severe malaria might be explained by the failure to develop pregnancy-specific acquired immunity against parasites harbouring VAR2CSA, which causes placental malaria, in this low-transmission area [11, 43, 48]. This does, however, ensure that the probability of misdiagnosis of severe malaria (e.g. sepsis with coincidental parasitaemia) is less likely [3, 49]. The signs and symptoms associated with increased risks of maternal death by severe malaria we identified are also commonly seen in maternal sepsis [50]. The three risk criteria identified in this study coincide with those used in the quick SOFA score for screening for sepsis [51], highlighting the importance of these three signs in life-threatening infections. Identification of the causal pathogen is thus important. Although this pooled analysis included the largest numbers of pregnant women with severe malaria in Asia, the numbers were still small, increasing the beta error. Further studies will be needed to assess whether any other WHO severe criteria (e.g. renal failure) are independently associated with maternal mortality.

Anaemia and malaria are common in pregnant women in tropical settings. Hyperparasitaemia is more common in low than high-transmission settings. However, anaemia and hyperparasitaemia often are not associated with concomitant vital organ dysfunction requiring parenteral malaria treatment. On the contrary, they carry a good prognosis for the mother with ready access to oral artemisinin treatments and if necessary, blood transfusion. The reported variability in the mortality of severe malaria in pregnancy is probably explained largely by the different prevalence of these two severe malaria criteria in different series. Maternal mortality is consistently high in severe malaria with vital organ dysfunction, notably cerebral malaria, requiring immediate life-saving intravenous artesunate administration. Severe malaria in pregnancy increases the immediate risk of foetal loss and preterm delivery, and the risk that the baby is born small-for-gestational-age.

Greater malaria control efforts in the non-pregnant population along the Thailand-Myanmar border has resulted in a significant decline in malaria in pregnancy and hence mortality in all patients from severe malaria, including pregnant women [17]. Waning immunity achieved through elimination efforts in this area makes pregnant women vulnerable to unpredictable programmatic break down due to war and pandemic. Intensive efforts to maintain access to ANC to minimise barriers to early diagnosis and treatment of malaria and anaemia are required.