The differential diagnosis of fever in a severely ill patient is broad. Coma and fever may result from meningoencephalitis or malaria. Cerebral malaria is not associated with signs of meningeal irritation (neck stiffness, photophobia, Kernig sign) but the patient may be opisthotonic. As untreated bacterial meningitis is almost invariably fatal, a diagnostic lumbar puncture should be performed to exclude this condition. There is also considerable clinical overlap between septicaemia, pneumonia and severe malaria - and these conditions may coexist. In malaria endemic areas particularly, where parasitaemia is common in the young age group, it is often impossible to rule out septicaemia in a shocked or severely ill obtunded child. Where possible, blood should always be taken on admission for culture, and if there is any doubt, empirical antibiotic treatment should be started immediately along with antimalarial treatment.
8.10.2 Other treatments
Many other supportive strategies and interventions have been proposed in severe malaria, but very few are supported by evidence of benefit, and many have proved harmful.
Heparin, prostacyclin, deferoxamine, pentoxifylline, low molecular weight dextran, urea, high-dose corticosteroids, acetylsalicylic acid, deferoxamine, anti-tumour necrosis factor antibody, cyclosporin, dichloroacetate, adrenaline and hyperimmune serum have all been suggested - but none of these is recommended. Evidence on use of corticosteroids is summarized in the box below.
EVIDENCE: trials of routine use of corticosteroids (e.g. dexamethasone) in the treatment of severe falciparum malariaa
Interventions: corticosteroids i.v.
Summary of RCTs: one systematic review and no additional trials. No significant difference in mortality; increased risk of gastrointestinal bleeding.
Expert comment: no additional information.
Basis of decision: systematic review.
Recommendation: do not use corticosteroids.
a See also Annex 9.9.
Severe metabolic acidosis is common but apart from correction of hypovolaemia and anaemia, no specific treatment is of proven value. Significant electrolyte abnormalities are relatively unusual, and potassium supplementation is often not required in the acute phase. The optimum fluid resuscitation regimens, the thresholds for blood transfusion, the role of exchange transfusion, and the management of seizures remain areas of uncertainty, and these are discussed in more detail below. The optimum body positioning in comatose patients, and the timing and type of feeding in patients who remain unconscious for >24 h have not been studied. It is generally agreed that early ventilation for respiratory abnormalities and early management of renal failure or severe metabolic acidosis are beneficial. In acute renal failure, haemofiltration is associated with a lower mortality, and more rapid correction of biochemical abnormalities compared with peritoneal dialysis. There have been no comparative trials of haemodialysis and haemofiltration.
8.10.3 Fluid therapy
Patients, especially children with severe malaria may be dehydrated. However, the degree of fluid depletion varies considerably. As a result, it is not possible to give general recommendations on fluid replacement. Each patient must be individually assessed and fluid resuscitation based on estimated deficit. In high-transmission settings where severe malaria is confined to childhood, children commonly present with severe anaemia and hyperventilation (sometimes termed "respiratory distress"). In the past this was ascribed to "anaemic heart failure" (i.e. pulmonary oedema), and sometimes diuretics were administered. It is now clear that this syndrome is not a result of anaemic heart failure, but results from severe metabolic acidosis and anaemia, and so should be treated by blood transfusion. In general children tolerate rapid fluid resuscitation better than adults, and are less likely to develop pulmonary oedema. In adults, there is a very thin dividing line between overhydration, which may produce pulmonary oedema, and underhydration contributing to shock and worsening acidosis and renal impairment. Careful and frequent evaluations of the jugular venous pressure, peripheral perfusion, venous filling, skin turgor and urine output should be made. Where there is uncertainty over the jugular venous pressure, and if nursing facilities permit, a central venous catheter should be inserted and the central venous pressure measured directly. The optimum rate of resuscitation, the role of colloids compared with crystalloids, and the optimum electrolyte composition of the replacement fluid have not been determined.
8.10.4 Blood transfusion
Severe malaria is associated with rapid development of anaemia as infected and uninfected erythrocytes are removed from the circulation. In areas of high stable transmission, severe anaemia in young children is the principal manifestation of severe falciparum malaria. Ideally fresh blood should be transfused, and the patient's relatives are often willing donors. However, in most settings cross-matched virus-free blood is in short supply. As with fluid resuscitation, there have not been enough studies to provide strong evidence-based recommendations, so the recommendations given here are based on expert opinion. In high-transmission settings, blood transfusion is recommended for children with a haemoglobin level of <5 g/100 ml (haematocrit <15%). Mortality as a direct result of anaemia rises at lower haemoglobin levels. In low-transmission settings, a threshold of 20% (haemoglobin 7 g/100ml) is recommended. These general recommendations still need to be tailored to the individual, as the pathological consequences of rapid development of anaemia are worse than those of acute on chronic anaemia, where there has been adaptation and a compensatory right shift in the oxygen dissociation curve (Annex 9.10).
8.10.5 Exchange blood transfusion (EBT)
There have been many anecdotal reports and several series claiming benefit for EBT in severe malaria but no comparative trials, and there is no consensus on whether it reduces mortality or how it might work. The rationale for EBT has been variously proposed as:
• removing infected red blood cells from the circulation and therefore lowering the parasite burden (although only the circulating relatively non-pathogenic stages are removed - and this is also achieved rapidly with artemisinin derivatives);
• reducing rapidly both the antigen load and the burden of parasite-derived toxins, metabolites and toxic mediators produced by the host;
• replacing the rigid unparasitized red cells by more deformable cells and therefore alleviating microcirculatory obstruction.
EBT requires intensive nursing and a relatively large volume of blood, and carries significant risks. There is no consensus on the indications, benefits and dangers involved, or on practical details such as the volume of blood that should be exchanged. It is therefore not possible to make any recommendation regarding the use of EBT.
8.10.6 Use of anticonvulsants
Seizures are common in cerebral malaria, particularly in children. The treatment of convulsions in cerebral malaria with intravenous (or, if not possible, rectal) benzodiazepines or intramuscular paraldehyde is similar to that for repeated seizures from any cause. In a large double-blind placebo-controlled evaluation of a single intramuscular injection of 20 mg/kg bw of phenobarbital (pheno-barbitone) in children with cerebral malaria there was a reduction in seizures but a significant increase in mortality in phenobarbital recipients. This resulted from respiratory arrest, and was associated with additional benzodiazepine use. Clearly the 20 mg/kg dose of phenobarbital should not be given without respiratory support, but it is not known, whether a lower dose would be effective and safer or whether, if ventilation is given, mortality would not be increased. In the absence of further information, prophylactic anticonvulsants are not recommended.
EVIDENCE: trials on use of phenobarbital for the treatment of convulsions in severe falciparum malariaa
Interventions: phenobarbital i.v.
Summary of RCTs: systematic review of three trials. In the two trials with adequate blinding, death was more common with phenobarbital.
Expert comment: no additional information.
Basis of decision: systematic review.
Recommendation: avoid routine use of phenobarbital.
a See also Annex 9.11.
8.10.7 Concomitant use of antibiotics
The threshold for administering antibiotic treatment should be low in severe malaria. Septicaemia and severe malaria are associated and there is diagnostic overlap, particularly in children. Unexplained deterioration may result from a supervening bacterial infection. Although enteric bacteria (notably Salmonella) have predominated in most trial series, a variety of bacteria have been cultured from the blood of patients diagnosed as having severe malaria, and so broad-spectrum antibiotic treatment should be given initially.