Malaria contributes up to 40% of paediatric hospital admissions in some parts of sub-Saharan Africa. Ten percent of these admissions are due to cerebral malaria, the most severe neurological complication of infection with Falciparum malaria(8). About 20% of children admitted with cerebral malaria die and up to 24% of those who survive have persistent neuro-cognitive sequelae(11). The risk of poor outcome is increased in patients admitted in deep coma, with repeated seizures, severe acidosis, hypoglycaemia, shock or with intracranial hypertension(5, 6, 9).
A small proportion of children with cerebral malaria have been observed to have a biphasic clinical course in which an initial recovery of consciousness with treatment is followed by a second period of coma. In an earlier description, a biphasic clinical course of cerebral malaria in Gambian children was associated with neurological sequelae - all 5 children with a biphasic clinical course had neurological sequelae at discharge from hospital (2). There have been no other reports describing this phenomenon. Although they constitute a numerically small entity, this appears to be a distinct group of children in whom neurological sequelae commonly develops. A study of this group may contribute to understanding the pathogenesis of cerebral malaria. Here, we present a clinical description of 11 children with a biphasic clinical course in cerebral malaria; we examine the risk factors associated with the biphasic pattern and outcome.
This is a retrospective (chart review) study of children admitted to hospital with cerebral malaria. The study was carried out in Kilifi district hospital at the rural coast of Kenya. The hospital provides paediatric in-patient service to approximately 5000 children annually. Most children are cared for on a 36-bed general paediatric ward but those with life threatening illnesses such as cerebral malaria are cared for in a high dependency ward(8).
In Patient Care
During the course of the admission, the nursing staff recorded the level of consciousness using the Blantyre coma score (BCS), the temperature, blood pressure, oxygen saturations, pulse and respiratory rates every 4 hours on a standard proforma. Laboratory tests included parasite density, full blood count, microbial culture of blood, glucose and acid base status. All patients were treated presumptively with parenteral first line antibiotics and Quinine until full laboratory results were available to guide further management. Supportive therapy was guided by clinical presentation and laboratory results(1, 8).
Patient Selection and Definition of Cases
We examined case notes of children with impaired consciousness and Falciparum malaria parasitaemia admitted to the high dependency ward between January 1994 and December 2004 and identified those with cerebral malaria. A child was said to have cerebral malaria if s/he was unable to localise a painful stimulus (9) at least 30 minutes after treatment of seizures or correction of hypoglycaemia, with falciparum malaria parasites on Giemsa stained blood smears and in whom meningitis had been excluded(15). Children with a biphasic course were those, who having regained full consciousness (BCS = 5), relapsed into a second period of coma. The first and second periods of coma were labelled coma 1 and coma 2 respectively.
Data Management We extracted the following clinical information from the patients’ case notes: age, gender, history and number of seizures prior to admission, the duration of coma before admission, duration of coma 1 and 2, the length of the intervening period of consciousness, events that occurred within an hour of the onset of the second period of unconsciousness and the interventions given.
Characteristics of children with a Biphasic clinical course
Of the 1,507 children with impaired consciousness whose case notes were reviewed, 587 had cerebral malaria. Out of these, 11 (1.9%) children had a biphasic clinical course. They were of median age of 29 months (inter-quartile range (IQR) 20-52) and only two were females. Nine of these children had a history of seizures and six of status epilepticus before admission. The median duration of coma prior to admission was 6 hours (IQR 2-9).
One child presented with hypoglycaemia at admission. The median percentage parasitaemia at admission was 0.22% (IQR 0.48-13.8) and only 3 had parasitaemia greater than five percent. Four children were admitted with severe metabolic acidosis (base deficit > 8), three had raised blood white cell counts (>17x103/µL) and three others had severe anaemia (Haemoglobin <5g/dL). Bacillus sp., probably a contaminant, was isolated from the blood culture of one patient. Lumbar punctures were performed later when the patients were stable. Examination of the cerebrospinal fluid was normal but for two children who had raised white cell counts (18 & 56 cells/mL).
Events before the onset of the second period of coma (table 1)
The median duration of coma 1 was 8 hours (IQR 4-20) while the median duration of the intervening period of consciousness between coma 1 and coma 2 was 16 hours (IQR 8-40). In eight of the 11 patients, a recurrence of seizures heralded the onset of coma 2. One of these patients received blood transfusion shortly after lapsing into coma 2. The temperature was raised (>37.8ºC) in six patients but information on the parasite density at this time was often lacking. The onset of coma 2 occurred after a median duration of 32 hours (IQR 19-52) from admission and lasted for a median duration of 38 hours (IQR 12-64). Features of raised intracranial pressure were not consistently examined for.
The median duration of hospital stay was 6 days (IQR 4-8). Expectedly, those who had a longer duration of coma 2 stayed longest in hospital. The two children who had raised white cell counts in their CSF had long periods of hospitalisation. One of them developed multiple neurological deficits (recurrent seizures, impaired hearing and vision) at discharge. Neurological sequelae were observed in three other children and included ataxia, diplegia, paraplegia and impaired speech. No death was reported. Three illustrative case histories are summarised below:
K.C., a 2 year 6 month old male child was admitted to the high dependency unit on 15/11/98 at 1300hrs. He had a history of fever and cough for 3 days, had suffered 2 episodes of prolonged right sided partial seizures within the previous 24 hours and had been unconscious for 2 hours prior to admission. He had received an unknown dose of intramuscular diazepam at a peripheral health unit. He had never had a seizure in the past, his growth and development had been normal and he had no other co-morbid condition.
On examination, he had no features of respiratory or circulatory impairment; he was febrile with a rectal temperature of 40.6 ºC and a pulse rate of 174 beats per minute (bpm). He was comatose with BCS of 2. His pupils were midpoint but sluggishly reacting to light and he was noted to have twitching of the mouth and the right hand. No other neurological deficits were noted. He had a high malaria parasitaemia of 69/500 RBCs (14% parasitaemia), a white cell count (WCC, X 103/µL) of 7.5, haemoglobin (Hb) of 5.2g/dL, low platelets of 28x103/µL, a blood gas profile of compensated metabolic acidosis (Base Excess (BE) -6.5) and normal blood sugar and electrolytes. He was started on Parenteral Quinine, Benzyl Penicillin, Chloramphenicol and maintenance dextrosaline (4% dextrose and 0.18 Saline). By the time of review six hours later, he had suffered 4 seizures and had received Diazepam, Paraldehyde and a loading dose of Phenytoin. He was still comatose (BCS=2) with sluggishly reacting pupils. He remained seizure free thereafter and by the time of review the next morning (16/11/2007), he was fully conscious (BCS =5) and was attempting to take feeds orally. Repeat blood examination showed a parasitaemia of 128/500RBCs (26%), raised white cell count of 12.2, Hb of 4.8g/dl and normal electrolytes. About 33 hours after admission, the child relapsed into unconsciousness (BCS=1). He was observed to have twitches of the right hand associated with noisy breathing. The haemoglobin had dropped to 3.8g/dl. Whole blood transfusion (20mls/Kg) was started 1 hour later. This second period of coma lasted 12 hours. During this period, he experienced five episodes of short partial seizures. A day after the transfusion, the malaria parasitaemia had dropped to 6/100WBC ( 0.012%) and the Hb had risen to 7.9g/dL. Thereafter, he made an uneventful recovery and he was discharged from hospital 8 days later without any neurological deficits.
S.M., a 2 year 5 month old male child presented on 01/08/2000 at 2300 hrs with a one day history of illness characterized by diarrhoea, vomiting and seizures but without a history of fever. He had had 5 episodes of generalized tonic clonic seizures, some of which were prolonged (>30 minutes), starting 12 hours before admission. During this period, he had also been unconscious without any recovery between the seizures. He had received unknown treatment at a peripheral health unit before being referred. His growth and development had been normal. On admission, he had deep (acidotic) breathing and was comatose (BCS=1). His pupils were constricted and he was globally hypotonic and hypo-reflexic. No other neurological deficits were observed. Admission blood results showed a parasitaemia of 18/500 RBCs (4%), a raised WCC of 19.5, Hb of 10.3g/dl, hypoglycaemia of 1.1mmol/L and a blood gas picture of compensated metabolic acidosis (BE -13.7). A bolus infusion of Normal saline (10mls/Kg) was started after correction of hypoglycaemia and he was also put on Benzyl penicillin, Chloramphenicol and Quinine. On review 5 hours later, he was still comatose (BCS=1) and episodes of abnormal motor posturing were reported in the preceding period. This continued until the following morning (02/08/00) when he was also noted to have dysconjugate gaze, bruxism and global hypertonia. On this account, a single intravenous infusion of 20% Mannitol (2mls/Kg) was administered. Thereafter, several short generalized tonic clonic seizures were observed. Phenytoin (18mg/Kg) was administered for status epilepticus and this terminated the seizures. The child regained consciousness about 20 hours after admission and was able to take oral feeds. A lumbar puncture was performed and showed clear CSF with a white cell count of 56/mm3 and a CSF/blood sugar ratio of 0.7. Shortly, the seizures recurred, short and multiple. In between the fits, the level of consciousness would improve. However, 40 hours after the resolution of coma 1, he developed prolonged seizures and lapsed into coma (BCS=2). For these seizures, he received Diazepam (0.3mg/Kg), Paraldehyde (0.2ml/Kg), Phenobarbital (15mg/Kg) and lastly Thiopental which terminated the seizures. Four hours later, the child had achieved a BCS score of 4. Short lived seizure events with short periods of impaired consciousness were observed during the rest of the child’s hospital stay. He was maintained on Phenobarbital (6mg/Kg/day) and Phenytoin (5mg/Kg/day). At the time of discharge 10 days later, he was observed to have developed multiple neurological deficits: impaired vision, hearing and speech and was only able to sit up without support. He was discharged on Phenobarbitone but was lost to follow up.
C.M, a 3 year 9 month old male child, presented on 12/07/99 at 2130hrs with a 3 day history of fever. He had a one-day history of poor feeding and agitation characterized by shouting, teeth grinding and stiffening of the body. No obvious seizures had been observed although by the time of admission to hospital, he was reported to have been unconscious for about 8 hours. There were no other preceding events to account for the neurological state. He had been seen in a private clinic where he received unknown oral and injectable drugs. On admission, he was tachypnoeic (Respiratory rate = 44/min) with normal oxygen saturations and no features of circulatory impairment. The rectal temperature was 38.5ºC. He was unconscious (BCS 2) and had intermittent episodes of abnormal motor posturing. He was globally hypertonic and hyper-reflexic but no other neurological deficits were observed. Initial laboratory examination showed a malaria parasitaemia of 110/100 WBCs, a WCC of 8.5x103/mm3, low Hb of 4.5g/dL and normal blood sugar. He was started on Benzyl Penicillin, Chloramphenicol and Quinine with maintenance infusion of 4.3% dextrose-saline. He regained consciousness 4 hours later (BCS= 4). He maintained this state for another 4 hours after which he again lapsed back into coma (BCS=0), with repeated episodes of decerebrate posturing. This state lasted for 54 hours during which period no seizures were observed. About 32 hours after admission, the malaria parasitaemia had cleared and the Hb had dropped to 3.8g/dL. He was transfused with whole blood (20mls/Kg). The CSF examination was normal. At discharge, 4 days after admission, he was fully alert and had no gross neurological deficits.
Cerebral malaria remains a major cause of death and neurological damage in African children. The syndrome is characterised by variable manifestations and outcomes possibly due to differences in pathogenesis(7). Brewster et al described a biphasic course in five Gambian children with cerebral malaria and found this to be associated with neurological sequelae(2). We report on twice the number of children they identified - over a longer period - but these patients still form a small proportion (1.9%) of children with cerebral malaria. Although the proportion of children with neurological sequelae at discharge (36%) in our study is not as high as that reported by Brewster (100%), the incidence is still higher than that of the general paediatric cerebral malaria population (10.9%)(11). None of these children died, adding credit to the suggestion that death and neurological sequelae after cerebral malaria may arise from different pathological processes(14). The reason for deterioration after initial recovery is unclear. It could either be a reversible but recurrent insult from the initial episode of coma such as a metabolic, toxic or immunogenic insult to the brain, or the two periods of coma may be mechanistically unrelated. In support of the former theory is the fact that seizures accompanied the onset of the second period of coma in the majority of patients, a finding also reported by Brewster. It is possible that the children who did not have obvious convulsive seizures at the onset of their second coma (n=3) had subtle or electrographic seizure activity, a phenomenon common in children with cerebral malaria(4). Recurrent seizures are thought to play an important role in the pathogenesis of coma (3, 11, 12). They aggravate raised intracranial pressure and may cause neuronal loss. The cause of these recurrent seizures is not certain. They may have been precipitated by metabolically active parasites, a possibility that is difficult to investigate. The minimum duration of coma 2 was 4 hours suggesting that the impaired consciousness was not a simple post-ictal phenomenon. Co-morbidity or misdiagnosis may also play a role in the pathogenesis of a biphasic course. Taylor et al demonstrated misdiagnosis at autopsy in 23.8% of children who died following a clinical diagnosis of cerebral malaria (13). The encephalopathy in these children was due to a viral or bacterial infection or other events such a ruptured arterio-venous malformation rather than malaria. Malaria in this case may either be a co-infection contributing to other features of the illness or just an incidental finding (asymptomatic parasitaemia). In support of this possibility, two children had a markedly raised CSF white cell counts and very high parasite densities. One of them went on to have the longest hospital stay in the group and had multiple neurological deficits at the time of discharge from hospital. Re-perfusion of cerebral areas previously clogged with sequestered parasites may be another mechanism by which a biphasic course may arise. The clearance of parasites from the brain during treatment resulting in reperfusion of the brain may distribute toxic products previously localised within the sequestered mass or precipitate the production of reactive oxygen species resulting in secondary deterioration(11). This effect may be aggravated by blood transfusion. Only one patient received blood transfusion within the period of the start of coma 2 and even so, he made a quick improvement after transfusion. Levels of malaria parasitaemia performed at these times of secondary deterioration would have been enlightening but this was not regularly done. Metabolic disturbances are unlikely to explain the biphasic course in our patients. Only one of the biphasic patients had hypoglycaemia at admission and only 3 had hypoglycaemia observed during the course of their admission. No episode of hypoglycaemia was observed during the onset or course of coma 2. Significant electrolyte derangements were also not observed during these periods and can therefore not explain the secondary deterioration.
The clinical data on deterioration in both studies needs to be viewed with caution since there is a considerable inter-observer difference amongst clinicians in assessing the Blantyre coma scale, particularly the verbal component (10). Although this is a small group of cerebral malaria patients, they suffer a high incidence of neurological sequelae and a prospective examination is warranted to shed more light on causality and by extension, on the pathogenesis of cerebral malaria.
Table 1: Summary of coma profile and outcome
|Patients||Age (Months)||Duration of coma at admission||Duration of coma 1||Intervening duration||event within 1hr of coma 2||Onset of coma 2||Duration of Coma 2||Duration of hospital stay in days||Outcome|
|1||9||9||28||16||SEIZURE||44||64||7||Paraplegia, Impaired speech|
|4||30||2||8||24||SEIZURE & BLOOD TRANSFUSION||32||12||8||A|
|9||29||12||20||40||SEIZURE||60||136||10||Impaired hearing and vision and Persistent fits|
NB. Coma durations are in hours
NE, No significant event
A, Alive with no sequelae