Severe malaria in Europe: an 8-year multi-centre observational study

Demographic data are shown in Table 2. The proportion of female patients was comparatively small (29%). The majority of patients (106/185, 57%) were of European origin without history of migration. Tourism was the main purpose of travel in Europeans (54/106, 51%), whereas VFRs were the predominant purposes of travel in patients with history of migration (55/68, 81%). European patients were on average older than patients with history of migration [median age 47 (IQR 33–57) vs 36 (IQR 27–45) years, p < 0.0001]. All malaria infections were acquired in Africa with the exception of two cases from Central America. By far the largest proportion of infections came from West Africa (109/185, 59%), followed by Central Africa (40/185, 22%), where Cameroon was the country with the highest number of imported cases (21/185, 11%, Fig. 1). VFR patients acquired malaria infections almost exclusively in West Africa, whereas European tourists acquired infections also in the tourist destinations of East Africa (Fig. 2). There was no change in age (F = 0.84, p = 0.5), gender (p = 0.11), origin of patients (p = 0.54) or purpose of travel (p = 0.10) during the 8-year course of the study. Almost 9 out of 10 patients (162/185, 88%) had not taken any anti-malarial chemoprophylaxis. Among 23 patients who took anti-malarial chemoprophylaxis, only six fully adhered to the prescribed regimen.

Clinical manifestations and laboratory findings leading to classification as severe malaria are shown in Table 2. Median baseline parasitaemia was 6.5% (IQR 4–11) and hyperparasitaemia (≥5%) was the most common criterion of severe disease, followed by jaundice, which was a criterion for severe malaria in this study according to WHO guidelines as of 2006. Eight patients with jaundice had no other vital organ dysfunction and would not have been classified as severe disease according to WHO guidelines as of 2010 [14].

Underlying chronic conditions were found in 43% of patients (63/185), of which hypertension was the most frequent (9%, 16/185). Seven per cent of patients (13/185) were HIV positive. The majority of patients (119/185, 64%) met one or two criteria of severe malaria, whereas 8% (17/185) met more than four criteria. Patients ≥60 years presented, on average, with more criteria for severe malaria than younger patients (median 3 vs 2, p = 0.02). Table 3 shows the risk of presenting with a particular criterion of severe malaria according to age (≥60 years vs younger patients). There was no difference in type and number of criteria for severe malaria among patients of European versus non-European origin with the exception of a lower median baseline parasitaemia (7 vs 5%, respectively, p = 0.04).

Intravenous quinine was the main first-line treatment in 93/185 patients (50%) whereas intravenous artesunate was used in 63/185 patients (35%). Seven patients (4%) received intravenous quinine and artesunate in combination. Table 4 gives an overview of the drugs and drug combinations used as follow-on treatment after intravenous therapy. Altogether 56 different combinations of intravenous and oral drugs were used across the different centres.

The proportion of patients treated with intravenous artesunate increased steadily during the course of the study from 27% (8/30) in 2006 to 60% (18/30) in 2013. In 22/185 patients (12%) only oral first-line treatments were given such as oral quinine (n = 16, exclusively in Italy), oral atovaquone-proguanil (n = 1, in a patient with HIV), and oral artemether–lumefantrine (n = 5). Patients treated with oral anti-malarials presented exclusively with hyperparasitaemia (n = 19, median parasitaemia 7%, range 5–8%) and/or jaundice (n = 6) as criteria for severe malaria and had no co-morbidities (except the one patient with HIV).

Choice of the first-line treatment was very heterogeneous across different sites and was mainly dependent on the country where the patient was treated. Centres in Norway, The Netherlands and Belgium reported treatment almost exclusively with intravenous artesunate, whereas the participating treatment centres in Spain and France used intravenous quinine in the majority of reported cases (9/12, 75% and 44/50, 88%, respectively).

An overview of supportive treatments is given in Table 5. Antibiotic therapy (44% of patients) and erythrocyte transfusion (21% of patients) were the most common. Erythrocyte exchange transfusion was performed in 8/185 (4%) patients at seven centres in Italy, Spain, Belgium, and The Netherlands. Seven of these patients were treated with intravenous quinine and one with intravenous artesunate. Baseline parasitaemia was >10% in all these patients (median 18%, IQR 10–27). Erythrocyte apheresis was performed in 5/185 patients (3%) at two centres (Vienna, Austria and Leiden, The Netherlands). Parasitaemia was >15% in all these patients (median 19%, IQR 18–28) and all five patients were treated with intravenous artesunate.

Adverse drug reactions were reported in 27/100 patients (27%) treated with intravenous quinine, and in 21/70 (30%) patients treated with intravenous artesunate. None of them was fatal. Cinchonism was the most common adverse drug reaction in patients treated with intravenous quinine (19/100, 19%). It was rated as mild in 17/19 cases and moderate in 2/19 cases by the treating physician. Hypoglycaemia occurred in 4/100 patients (4%) and cardiac arrhythmias in 1/100 patient (1%) treated with intravenous quinine.

In patients treated with intravenous artesunate, PADH was reported in 19/70 patients (27%), a finding which first became known during the study period in the year 2011. Onset of PADH was reported during days 10 to 14 (median 14) and median duration of haemolysis was reported to be 14 (IQR 8–18) days. Three patients (15%) with PADH received blood transfusions, with 2 patients (10%) re-hospitalized (for 3–5 days, respectively). In 1 patient, PADH was reported after therapy with only oral artemether–lumefantrine. This patient and some of the other patients with delayed PADH have already been reported elsewhere [15‐17].

In one patient, an acute cerebellar syndrome (ataxia, dysarthria, dysmetria, adiadochokinesis) was described beginning 3 days after the end of anti-malarial treatment (day 10). The patient had not shown any neurological symptoms during the acute phase of malaria and had been treated with a loading dose of intravenous quinine on day 1, intravenous artesunate from day 1 to 4 and oral artemether–lumefantrine from day 5 to 7. An MRI scan and lumbar puncture showed no abnormalities. Due to persistence of symptoms, the patient received physiotherapy until ten weeks after anti-malarial treatment, where neurological symptoms steadily improved. Complete recovery was reported 7 months after the malaria episode.

Three patients died, two of European origin and one with history of migration, resulting in a 28-day survival rate of 98.4%. All three patients had been treated with intravenous quinine and one of them also with intravenous artesunate simultaneously. All deaths occurred within the first 3 days after admission. All three patients had presented with hyperparasitaemia (9, 10 and 40%) and respiratory distress requiring mechanical ventilation. Respiratory distress with the need for mechanical ventilation was significantly associated with the risk of death in the study population (13 vs 0%, p = 0.001). Of note, two of the patients who died were 22–34 years of age, respectively, and had no underlying co-morbidities. One of them had a history of migration. The third patient was 70 years of age, suffered from a pre-existing chronic cardiomyopathy and died from therapy-refractory shock.

In 76% of patients (117 of 153 patients with available data) treatment took place in an ICU, where the median length of stay was three (IQR 2–5) days. The median length of inpatient treatment was 7 days (IQR 5–9). Median time to 99% parasite clearance was 48 h (IQR 24–72, n = 126) and median time to complete parasite clearance, was 72 h (IQR 60–120, n = 104). Data showing shorter parasite clearance time and shorter ICU and inpatient treatment in patients treated with intravenous artesunate compared to intravenous quinine were reported elsewhere [8]. The 22 patients who received oral anti-malarial treatment had comparatively long median 99% parasite clearance and complete parasite clearance times (72 h, IQR 48–72, and 120 h, IQR 84–144, respectively). There were no documented early or late parasitological failures.

Among 46 patients with cerebral malaria, six had neurological sequelae at discharge such as confusion, dysphasia, ataxia, and imbalance. In four patients (2/25 treated with artesunate vs 2/21 treated with quinine) confusion and ataxia persisted at day 28. Among 36 patients with acute renal failure at presentation, 13 (36%) patients had elevated creatinine levels at discharge, which had not been reported by the patient or documented in medical charts before malarial infection. In nine of these patients (4/17 treated with artesunate vs 5/16 treated with quinine) elevated creatinine levels persisted after day 28. Two patients suffered from necroses of fingers and toes as sequelae of vasopressor therapy during malaria.

Among eight patients with jaundice as only criterion for severe malaria, who would not have been classified as severe according to WHO 2010 criteria, none experienced documented complications of treatment or sequelae, seven were not treated at an ICU and patients had a comparably short median length of stay in hospital of 4 days (IQR 4–6).

Intravenous artesunate has been shown to improve survival in patients with severe malaria in endemic areas, with particular benefit for patients with high parasitaemia [>10% infected red blood cells (RBCs)] [7, 9, 10]. A randomized, controlled trial to confirm superiority of artesunate over quinine in non-endemic areas would be unethical. Other benefits of treatment with intravenous artesunate such as shorter ICU and hospital treatment were clearly demonstrated in European patients [8].

Quinine is still widely used in Europe, but the rate of patients receiving intravenous artesunate almost doubled in the 8-year study period. In the final year 2014, every second patient received intravenous artesunate. Current surveillance data from national reference centres indicate that the proportion of patients treated with artesunate is further increasing, particularly in countries where participating centres still reported frequent treatment with quinine during this study [24]. Although prospectively collected safety data from Europe is not available, this study adds to the evidence that artesunate is effective and can safely be used to treat patients in Europe. Artesunate is used in Europe despite considerable legal problems: The manufacturer of intravenous artesunate has been prequalified for good manufacturing practice-standard (GMP) by WHO [6], but the drug is not available in a European GMP-standard quality. It has an orphan designation for Europe by the European Medicines Agency since 2007, but no marketing authorization in Europe or in USA. Only in France, USA, Belgium, Denmark, and The Netherlands, named-patient programmes or similar protocols are in place, providing a legal basis for treating patients with this lifesaving drug [25, 26]. Until intravenous artesunate receives approval from the European Medicines Agency and the US Food and Drug Administration, the legal context for physicians procuring and applying this drug in Europe will remain unsatisfactory. This study also highlights the differences in treatment practices and guidelines across Europe: from occasional use in some countries to exclusive use of artesunate to treat severe malaria in The Netherlands or Norway. This finding likely reflects the respective national legal framework for using non-licensed drugs as well as national treatment recommendations [27]. The treatment of patients with intravenous artesunate and quinine in parallel, reported in seven patients, might also be based on judicial reasons: physicians might want to combine the most effective but unapproved drug (artesunate) with the approved and recommended standard first-line treatment (quinine) in order to avoid a legal risk [28]. No benefit of this combination has been shown compared to treatment with artesunate alone, whereas the frequency of adverse events increased [29]. A small number of patients were treated exclusively with oral anti-malarials in this study. These patients were less severely ill, including three patients with jaundice, as the only criterion for severe disease, who would not have been classified as severe according to the current WHO classification. No treatment complications were reported for these patients, yet time to parasite clearance was comparatively long. Overall, there was a remarkable variety of altogether 56 different combinations of intravenous and oral anti-malarial drugs reported. Harmonized, evidence-based European treatment guidelines would be useful to support clinicians in their choice of anti-malarial treatments.

Following the initial description of an episode of severe prolonged haemolysis after treatment of a patient with severe malaria with intravenous artesunate in Japan in 2002 [30], late haemolytic reactions 2–6 weeks after treatment with intravenous artesunate were described in a case series in European patients in 2011 [12], and then confirmed in studies in European [15‐17] and African [31] patients. Removal of parasites from RBCs in the spleen, leaving behind a once-infected ‘pitted’ erythrocyte with a shorter life span has been shown to be a potential mechanism causing late haemolysis [11], but the pathophysiology is not fully understood. The rate of haemolytic reactions and transfusions reported in the literature is variable depending on size, context, type, and setting of a study as well as on definitions of post-treatment haemolysis; the results of the present study are generally in line with previous observations [12, 16, 32] and late haemolysis can be expected to occur in approximately 20–30% of non-immune patients treated. As shown by the present data, a considerable proportion of them also receive blood transfusions. The results show once more that patients receiving intravenous artesunate for treatment of severe malaria should be routinely observed for signs of haemolysis at least on days 7 and 14 after treatment. An acute cerebellar syndrome 3 days after the end of anti-malarial treatment with intravenous artesunate and oral artemether lumefantrine was reported in one patient. Although the reported time until complete resolution (7 months) is longer than in most cases reported in the literature; symptoms and time of onset are suggestive of post-malaria neurological syndrome [33].

Mortality in this patient population was very low, reflecting the high standard of intensive care in Europe. Previous single-centre and national studies reported mortality rates between 4 and 15% [3, 18, 19, 22, 34]. The proportion of patients who had criteria of severe malaria associated with adverse outcome and death was comparable to other studies (Table 2) [2, 21, 34]. Mortality might have been biased by the fact that most reporting centres are tertiary care institutions with long experience in treating severe malaria. The study did not capture cases of severe malaria in smaller remote hospitals, where mortality might be higher. The increasing use of intravenous artesunate as main first-line treatment may also have contributed to reduce mortality, e.g., through rapid parasite clearance and shorter length of ICU and inpatient treatment [7‐9]. Age as risk factor for adverse outcome of imported severe malaria has been shown by numerous studies [2]. In the present study, population patients ≥60 years were more likely to suffer from acute renal failure or from cerebral malaria, yet there was no increased case fatality among older patients.

Only 10% of patients with severe malaria had taken anti-malarial chemoprophylaxis and very few of them had been fully compliant. These data suggest that correct anti-malarial prophylaxis can effectively prevent severe malaria in European travellers. Counselling of travellers on malaria prevention should be improved and coverage increased, particularly for travellers going to West Africa, where 60% of infections in this study were acquired. Little is known about the proportion of European travellers who take prophylaxis. In a recent study only 60% of travellers from the UK to endemic areas used anti-malarial chemoprophylaxis [35]. VFR travellers are a large traveller population to Africa with different perceptions of malaria and its prevention [4, 5]. This may also influence the longer delay between onset of symptoms of malaria and presentation in hospital. The fact that 37% of patients in the present study had a history of migration clearly demonstrates the risk of these patients to suffer from severe malaria. Moreover, one of the patients who died was a 34 years old, otherwise healthy patient with history of migration. She presented with hyperparasitaemia, acute renal failure, jaundice, and respiratory failure. There was altogether no difference in symptoms and clinical presentation between patients with or without history of migration, suggesting a waning of semi-immunity in migrants who left endemic areas [36].

Supportive treatments such as exchange transfusions and erythrocyte apheresis are a matter of controversy and their use is guided by national or local practices. European single-centre studies recently failed to demonstrate improved parasite clearance through whole blood or erythrocyte exchange, compared to patients treated with quinine or artesunate alone [37, 38]. Only 4% of patients received exchange transfusions in this study. It was performed in only 7 out of 28 participating centres, mainly in patients treated with intravenous quinine. Likewise, only five patients in two centres received automated erythrocyte apheresis. With the increased use of intravenous artesunate and its potential to rapidly reduce high parasite loads it needs to be determined whether particular patient populations might still benefit from these adjunctive treatments [39].