Delayed haemolysis following artesunate in a child with profound anaemia and Coca-Cola-coloured urine

Abstract

An 8-year-old girl of African descent presented to the hospital with a headache, lethargy, pallor and ‘Coca-Cola’-coloured urine. She had been admitted 11 days before with Plasmodium falciparum malaria, which was successfully treated with 48 hours of parenteral artesunate. Investigations revealed signs of severe haemolytic anaemia, with a haemoglobin level of 52 g/L that reached a nadir of 10 g/L within 4 hours, in addition to haemoglobinuria, hyperbilirubinaemia and raised lactate dehydrogenase levels. She was diagnosed with post-artemisinin delayed haemolysis, which is usually self-limiting but has the potential to cause severe, life-threatening anaemia 7–21 days following malaria treatment with artesunate. There was excellent response to blood transfusion, and the child made a full recovery. This case highlights the importance of providing safety netting advice regarding signs and symptoms of anaemia to patients receiving artesunate, in addition to monitoring of haemoglobin levels in the weeks after treatment.

Background

Malaria is a notifiable disease in the UK with 1683 cases reported in 2018.¹ Malaria is caused by a parasite, most commonly Plasmodium falciparum, that is transmitted to humans by the bite of an infected female Anopheles mosquito.² Malaria can be difficult to detect, given its lack of endemicity in the UK and the vague presentation, meaning it is impossible to diagnose based on clinical picture alone.³ An artemisinin class drug called artesunate has replaced quinine as the gold-standard treatment for falciparum malaria due to improved efficacy and a better side-effect profile; however, one of the observed adverse effects is post-artemisinin delayed haemolysis (PADH), which typically occurs 14–21 days after treatment.^{4 5}

We report a case of PADH in a child who presented 11 days after receiving artesunate with profound haemolytic anaemia and a haemoglobin nadir of 10 g/L. Rapid action with blood transfusion was needed to prevent circulatory collapse. This side effect of artesunate may not be well known, given the relatively low prevalence of malaria in the UK; however, it is an important topic with potentially severe consequences if adequate discharge plans are not formulated. This case is presented to highlight the importance of follow-up in patients receiving artesunate, including safety netting advice and repeat haematological investigations to monitor for anaemia.

Case presentation

A UK-born 8-year-old girl of African descent presented to the emergency department with a 3-day history of fever, cough, headache and myalgia. She had returned from a holiday in Sierra Leone 3 weeks before and no antimalarial medications had been taken. There was a medical history of asthma, but she was otherwise usually fit and well, and there was no family history of any medical conditions. The child was diagnosed with Plasmodium falciparum malaria and was admitted to the paediatric ward. There was 1.5% parasitaemia, which cleared after 48 hours of parenteral artesunate 2.4 mg/kg at 12-hour intervals, and she was discharged home with 3 days of oral artemether/lumefantrine (Riamet).

Eight days after discharge (and 11 days after the first dose of artesunate), the child was readmitted overnight with a 4-day history of malaise, weakness, loss of appetite and headache. On the day of presentation, she had noticed that her urine had become ‘Coca-Cola coloured’. On examination, she was lethargic and quiet, with conjunctival pallor and icteric sclerae, but the remainder of the systemic examination was unremarkable. She had a low-grade fever of 37.9°C and tachycardia of 118 beats/min, with other observations normal. Her urine was deep red/brown in colour with blood and protein detected on dipstick.

Investigations

During the first admission with malaria, blood smear detected 1.5% parasitaemia with P. falciparum on day 1, <1% parasitaemia on day 2 and no parasitaemia on day 3. The haemoglobin level at discharge was 93 g/L, and glucose-6-phosphate dehydrogenase (G6PD) levels were low at 5.2 U/g haemoglobin (normal range 6.0–12.8 U/g haemoglobin).

Haemoglobin levels had declined to 52 g/L at readmission, dropping to 24 g/L within 2 hours and reaching a nadir of 10 g/L within 4 hours (figure 1). The blood film showed red cell agglutination and spherocytes but no malarial parasites. There was an elevated reticulocyte count 105.6×10⁹/L (normal range 20–100×10⁹/L), elevated bilirubin 42 μmol/L (normal range 0–17 μmol/L), elevated lactate dehydrogenase (LDH) 1786 units/L (normal range 125–243 units/L) and negative Coomb’s test. Haemoglobinopathy screen was normal, and G6PD levels had normalised to 14.2 U/g haemoglobin. Tests for viruses were negative, and blood culture and Mycoplasma serology were normal.

Figure 1.

Timeline of haemoglobin levels from first admission with malaria to final discharge from hospital. FBC, full blood count; h, hour.

Differential diagnosis

In view of the timing of presentation and features indicative of a haemolytic process, including haemoglobinuria, hyperbilirubinaemia, LDH of 1786 units/L, reticulocytosis and significant decline in haemoglobin levels, a diagnosis of PADH was made after discussion with the paediatric infectious diseases team. The other differential diagnoses considered were acute intravascular haemolysis secondary to G6PD deficiency and blackwater fever (BWF).

PADH was concluded to be the final diagnosis as the timing of presentation was delayed, whereas in cases of BWF, haemolysis often occurs much more rapidly, as well as the G6PD levels returning to normal during the haemolytic episode.

Treatment

The child was initially admitted for observation on the paediatric ward, as the diagnosis remained unclear and she was clinically stable. She was noted to have one spike in temperature to 38.6°C shortly after admission, but this soon resolved. After notification from the laboratory that the haemoglobin level had declined to 24 g/L, she was given an urgent transfusion of 20 mL/kg packed red blood cells over 4 hours and started on high-flow oxygen therapy. The transfusion was commenced just as the haemoglobin level reached a nadir of 10 g/L. She remained an inpatient on the paediatric ward for 3 days and was commenced on regular folic acid treatment.

Outcome and follow-up

There was marked clinical improvement following blood transfusion and the child required no other treatment or further transfusions. The post-transfusion haemoglobin level was 115 g/L, and she was discharged home after 3 days. She was reviewed as an outpatient after 1 month, at which time her Hb was 134 g/L and all other haematological parameters had normalised, including the G6PD level (9.1 U/g haemoglobin). She was subsequently discharged from follow-up.

Discussion

Malaria infection can commonly cause anaemia, particularly in endemic regions and with children the most affected.⁶ Despite anaemia being a sequela of malarial disease itself, it can also occur secondary to various antimalarial treatments, and so it can be challenging in some cases to make this differentiation. PADH can occur in any adult or child that receives artesunate treatment for malaria. A documented risk factor is hyperparasitaemia in travellers, but in endemic regions, children are most at risk.⁵ Artesunate results in death of malarial parasites, which are then expelled from infected erythrocytes by the spleen, a process called ‘pitting’.⁷ The delayed haemolysis that occurs at least 7 days after treatment with artesunate is due to the reduced lifespan and subsequent clearance of just these pitted erythrocytes.⁷ Hyperparasitaemia results in higher numbers of previously infected erythrocytes; however, as the rate of pitting has been shown to be variable, not all patients with hyperparasitaemia will develop PADH, nor can the level of parasitaemia be used to predict its occurrence.⁷

In our index case, the profound anaemia was attributed to artesunate treatment in view of the delayed presentation after clearance of parasites and features of haemolysis. The case was initially diagnostically challenging, as there were a few differentials that could present with similar signs and symptoms. Perhaps the most well-known cause of haemoglobinuria in malaria is BWF, which often occurs within 24 hours after treatment with quinine; however, in our index case, there was a longer delay and quinine-based therapy was never used.⁸ Antimalarial drugs can also trigger episodes of acute intravascular haemolysis in patients with G6PD deficiency.⁹ In our case patient, not only were crisis-precipitating antimalarial drugs not used, but also the G6PD level was normal during the episode of haemolysis and at follow-up, making G6PD deficiency unlikely. Furthermore, the initial management was just for observation as the diagnosis at that stage was unknown; there was no familiarity with the phenomenon of PADH, and there was no expectation for haemolysis to be so rapid and profound. The repeat haemoglobin at 2 hours was only performed as peripheral intravenous access was being sited, and it was this that prompted urgent transfusion, highlighting the human factors that played a part in this case.

A review of PADH published in 2015 concluded that this phenomenon has been detected in 7%–21% of patients undergoing artesunate treatment.⁵ Eight studies were described to report cases of PADH in travellers, and in the seven where the treatment was documented, 12/19 patients (63%) required blood transfusions. Although hyperparasitaemia was seen in the majority of patients with PADH reported in these studies, this was not the case with our index patient, who had parasitaemia of just 1.5% at presentation.

This case documents the lowest nadir in haemoglobin (10 g/L) reported in the literature that can be attributed to PADH. Across the 2015 review by Rolling et al,⁵ a literature search published in 2013,¹⁰ and other more recent case series,^11–18 the lowest haemoglobin nadir thus far has been 28 g/L. Children in endemic regions have been noted to be particularly at risk of PADH, and observations from published case series show that they may have lower haemoglobin levels compared with their adult counterparts.^{8 19} For example, a report of delayed haemolysis in 23 patients had only 5 patients with Hb declining at <50 g/L and they were all children¹⁹; meanwhile, the lowest reported haemoglobin thus far (28 g/L) occurred in a child.⁸ Therefore, despite PADH often being reported as self-limiting, there is potential for severe anaemia even in travellers with low parasitaemia, and children may be more at risk, although this is not yet proven for those living in non-endemic regions. Children presenting with severe haemolytic anaemia and tachycardia should be promptly considered for blood transfusion, as tachycardia is a proven early indicator for critical illness, and development of hypotension would be a late sign of disease.²⁰

The index patient had two febrile episodes and all infectious causes were ruled out. Since the febrile episodes resolved with the decline of haemolysis, they were attributed to the haemolytic process, a feature of PADH that has been well documented in other case reports.^{18 21–23} However, it remains important to exclude other causes of fever in patients presenting with similar symptoms.

The current treatment for PADH is blood transfusion, with no recommendation for corticosteroid therapy due to lack of an interventional study analysing their use in this context.⁵ The UK malaria treatment guidelines report that as haemolysis typically occurs 7–21 days after treatment with artesunate, a repeat full blood count should occur after 14 days.²⁴ This recommendation is also documented in the local malaria guideline; however, adherence has not been audited. Rolling et al⁵ recommend a more rigorous follow-up approach in their review of PADH, with weekly clinical examinations and haematological investigations for up to 4 weeks, and the Centres for Disease Control and Prevention also support this recommendation.¹⁰

Following this case, the importance of adhering to the local and national malaria guidelines was emphasised to members of the team. Many clinicians stated that they had no knowledge of this side effect of artesunate, and so this case incited an increase in awareness locally. It is important to remember that despite hyperparasitaemia being a risk factor for the development of PADH in travellers, a low parasitaemia should not falsely reassure clinicians against this adverse effect as a possibility. This case ultimately serves to highlight the importance of providing safety netting advice to patients treated with artesunate and ensuring a clear follow-up plan is in place to monitor haemoglobin levels after discharge.

Learning points.

• Post-artemisinin delayed haemolysis is usually a self-limiting side effect of artesunate but has the potential to cause profound anaemia in both adults and children.

• Patients who receive artesunate treatment for malaria should be followed up for a repeat full blood count (FBC) at 14 days, with strong consideration given to a further FBC at 21 days to monitor for this complication.

• Safety netting advice on the symptoms of anaemia should be given to all patients receiving artesunate.

• Blood transfusion is an effective treatment for post-artemisinin delayed haemolytic anaemia.