573
Intracranial pressure in African children with cerebral malaria
Opening lumbar cerebrospinal fluid (CSF) pressure was measured with a paediatric spinal fluid manometer in 26 of 61 Kenyan children (mean age 39 months) with cerebral malaria. In all cases pressure was above normal (mean [SD] 22·6 [7·4] cm CSF, range 10·5-36). Clinical features of our patients suggest that intracranial hypertension is important in the pathogenesis of cerebral malaria in children, especially as a cause of death. We suggest that raised intracranial pressure is secondary to increased cerebral blood volume. Lowering intracranial pressure may significantly reduce the mortality and morbidity of cerebral malaria. The potential risks and benefits of lumbar puncture should be considered carefully in patients with suspected cerebral malaria. Introduction
falciparum infection is one of the of acute encephalopathy in children. Cerebral malaria (CM) in children has a 10-40% mortality,i even when effective antimalarial drug concentrations are achieved rapidly. The exact cause of death is often obscure. We have measured intracranial pressure (ICP) in African Plasmodium
commonest causes
children with CM to determine whether intracranial hypertension contributes to the pathophysiology of this condition.
transducer (Hewlett Packard) was connected to the spinal needle for measurement of CSF pressure for up to 30 min. Blood pressure was measured as close to the time of LP as possible so that cerebral perfusion pressure could be estimated from the difference between mean arterial pressure and lumbar CSF pressure. In 10 children, transcranial pulsed doppler ultrasonography (’TC2-64B’, Eden Medizinische Elektronik, Uberlingen, Germany) of the middle cerebral arteries3 was done frequently during admission. Children were given parenterally either a loading dose of 20 mg/kg quinine dihydrochloride followed by 10 mg/kg every 12 h, or a 10 mg/kg loading dose with 5 mg/kg every 12 h thereafter. Parenteral therapy was continued until the child could swallow, when quinine sulphate was given orally. Each patient received 10 doses of quinine. Children had a full neurological examination at least every 6 h until they became conscious and were able to localise pain. These clinical data were analysed retrospectively for combinations of physical signs (table I) compatible with uncal herniation or with one of the four stages of central cerebral hemiation-ie, diencephalic, upper pontine, lower pontine, or medullary. The minimum criteria required to define a herniation syndrome were decided before data were analysed. Uncal herniation could be diagnosed with only two of four criteria present, since partial third-nerve palsy is unlikely to be due to any other cause in an unconscious child. For diagnosis of diencephalic syndrome, it was decided that four of five criteria should be present, since some of the signs also occur in diffuse bilateral hemispheric dysfunction due, for example, to drug or metabolic effects. Diagnosis of upper pontine (three of four criteria), lower pontine (three of five), and medullary (two of three) syndromes was uncomplicated. If a herniation syndrome was documented during hypoglycaemia or within 1 h of a seizure, data were excluded from analysis. Differences between groups were analysed with Fisher’s exact test.
Patients and methods Ethical permission for the study was granted by the Kenya Medical Research Institute (KEMRI), and informed written consent was obtained from childrens’ parents or guardians. Since May, 1989, we have identified children admitted to the Kilifi District Hospital, Kenya, who fulfilled the World Health Organisation definition of CM—ie, they were unable to localise pain, had peripheral parasitaemia, and other causes of coma were excluded. On admission, a full clinical history was taken and a physical examination done. Blood was obtained for a complete blood count, glucose determination, and preparation of thick and thin blood films for peripheral parasite counts. Lumbar puncture (LP) was done in all children, except those with signs pathognomonic of brainstem herniation or raised ICP (eg, papilloedema). (Since July, 1990, we have delayed LP until the return of consciousness, although antimicrobials to cover for meningitis are given from the time of admission.) LP was done with the child in the lateral decubitus position. A 22 G spinal needle was used, and, where possible, cerebrospinal fluid (CSF) pressure was measured with a paediatric spinal fluid manometer. Excessive flexion was avoided and pressure was recorded only if the child was quiet and there was a swing in pressure with respiration. Maximum height above the spine reached by the column of fluid was recorded as the opening pressure in cm CSF. Closing pressures were not measured since this would have required the removal of too large a volume of CSF. Less than 1 ml of CSF was removed for microscopy, culture, glucose, and lactate. In 2 children, a pressure
Results Between May, 1989, and August, 1990, 586 children with primary diagnosis of malaria were admitted, of whom 61 fulfilled the definition of CM (mean [SD] age 39 [21 °6] months, mean duration of disease 38 [2°6] days). Table n shows the clinical and laboratory features of these children on admission. 47 patients had LP on admission. Opening pressure was measured reliably in 26 children. No opening pressure was recorded in the remaining children because manometers were not available during the first 5 months of study (19), the child was struggling (3), or because LP was deferred until 24 h after consciousness had been regained a
(13). The figure shows opening pressure in 26 patients plotted against age, compared with the normal range.4 Mean (SD) opening pressure was 22-6 (7-3) cm CSF [16°7 (5-4) mm Kenya Medical Research Institute, Kilifi Coastal Unit, Kilifi, Kenya (C. R. J. C. Newton, MRCP, P. A. Winstanley, MD, N. Peshu, MB ChB, K. Marsh, MRCP); Department of Neurology, Institute of Child Health, London, UK (F J. Kirkham, MRCP); and Nuffield Department of Clinical Medicine, John Radcliffe Hospital, Oxford, UK (Prof G. Pasvol, DPhil, Prof D. A. Warrell, DM). Correspondence to Dr F. J Kirkham, Department of Neurology, Institute of Child Health, 30 Guilford Street, London WC1 N 2AP, UK.
ADDRESSES:
574
TABLE I-HERNIATION SYNDROMES IN CHILDEN WITH CM
Hg]. Mean cerebral perfusion pressure at LP was 52-6(13-8) Hg (range 30-3-88-3). LP was repeated in 2 children, because of recrudescence of fever. 1 child had opening pressures of 105, 21, and 14 em CSF (8,16, and 10 mmHg) at admission, 54, and 115 h after start of treatment, respectively. The other child had an opening pressure of 31 cm CSF (23 mm Hg) at admission and 29 cm CSF (21 mm Hg) at 84 h after start of treatment. He also had the lowest recorded cerebral perfusion pressure (30-3 mm Hg), and was left with dystonic motor disorder, delayed language, and cortical blindness at 3-months follow-up. CSF pressure fell from 21 to 10 mm Hg and from 20 to 9 mm Hg in the 2 children in whom it was measured with a transducer, 30 min after a 1 g/kg bolus dose of mannitol. Extensor posturing with Cheyne-Stokes respiration and absent gag was seen 12 h later in 1 of these children; hypoglycaemia was excluded. Posturing ceased after a further bolus of mannitol mm
and a full recovery was made. Of those studied with transcranial doppler ultrasound, forward flow was observed in 9 of 10 throughout diastole during every study; all survived. The tenth child had normal middle cerebral artery sonograms at admission; however, after 3 h, when brainstem reflexes were absent and the child was being hand ventilated, there was reverse flow throughout diastole with low velocity and a direction-of-flow index similar to that seen in brainstem death.5 12 of the 61 children (20%) died, all in coma, between 10 min and 24 h after admission. The fontanelle was closed in all, and 10 had had an LP. Herniation was observed in all 12 patients who died but in only 17 of 49 (35%) survivors (p
Intracranial pressure in African children with cerebral malaria
Opening lumbar cerebrospinal fluid (CSF) pressure was measured with a paediatric spinal fluid manometer in 26 of 61 Kenyan children (mean age 39 months) with cerebral malaria. In all cases pressure was above normal (mean [SD] 22·6 [7·4] cm CSF, range 10·5-36). Clinical features of our patients suggest that intracranial hypertension is important in the pathogenesis of cerebral malaria in children, especially as a cause of death. We suggest that raised intracranial pressure is secondary to increased cerebral blood volume. Lowering intracranial pressure may significantly reduce the mortality and morbidity of cerebral malaria. The potential risks and benefits of lumbar puncture should be considered carefully in patients with suspected cerebral malaria. Introduction
falciparum infection is one of the of acute encephalopathy in children. Cerebral malaria (CM) in children has a 10-40% mortality,i even when effective antimalarial drug concentrations are achieved rapidly. The exact cause of death is often obscure. We have measured intracranial pressure (ICP) in African Plasmodium
commonest causes
children with CM to determine whether intracranial hypertension contributes to the pathophysiology of this condition.
transducer (Hewlett Packard) was connected to the spinal needle for measurement of CSF pressure for up to 30 min. Blood pressure was measured as close to the time of LP as possible so that cerebral perfusion pressure could be estimated from the difference between mean arterial pressure and lumbar CSF pressure. In 10 children, transcranial pulsed doppler ultrasonography (’TC2-64B’, Eden Medizinische Elektronik, Uberlingen, Germany) of the middle cerebral arteries3 was done frequently during admission. Children were given parenterally either a loading dose of 20 mg/kg quinine dihydrochloride followed by 10 mg/kg every 12 h, or a 10 mg/kg loading dose with 5 mg/kg every 12 h thereafter. Parenteral therapy was continued until the child could swallow, when quinine sulphate was given orally. Each patient received 10 doses of quinine. Children had a full neurological examination at least every 6 h until they became conscious and were able to localise pain. These clinical data were analysed retrospectively for combinations of physical signs (table I) compatible with uncal herniation or with one of the four stages of central cerebral hemiation-ie, diencephalic, upper pontine, lower pontine, or medullary. The minimum criteria required to define a herniation syndrome were decided before data were analysed. Uncal herniation could be diagnosed with only two of four criteria present, since partial third-nerve palsy is unlikely to be due to any other cause in an unconscious child. For diagnosis of diencephalic syndrome, it was decided that four of five criteria should be present, since some of the signs also occur in diffuse bilateral hemispheric dysfunction due, for example, to drug or metabolic effects. Diagnosis of upper pontine (three of four criteria), lower pontine (three of five), and medullary (two of three) syndromes was uncomplicated. If a herniation syndrome was documented during hypoglycaemia or within 1 h of a seizure, data were excluded from analysis. Differences between groups were analysed with Fisher’s exact test.
Patients and methods Ethical permission for the study was granted by the Kenya Medical Research Institute (KEMRI), and informed written consent was obtained from childrens’ parents or guardians. Since May, 1989, we have identified children admitted to the Kilifi District Hospital, Kenya, who fulfilled the World Health Organisation definition of CM—ie, they were unable to localise pain, had peripheral parasitaemia, and other causes of coma were excluded. On admission, a full clinical history was taken and a physical examination done. Blood was obtained for a complete blood count, glucose determination, and preparation of thick and thin blood films for peripheral parasite counts. Lumbar puncture (LP) was done in all children, except those with signs pathognomonic of brainstem herniation or raised ICP (eg, papilloedema). (Since July, 1990, we have delayed LP until the return of consciousness, although antimicrobials to cover for meningitis are given from the time of admission.) LP was done with the child in the lateral decubitus position. A 22 G spinal needle was used, and, where possible, cerebrospinal fluid (CSF) pressure was measured with a paediatric spinal fluid manometer. Excessive flexion was avoided and pressure was recorded only if the child was quiet and there was a swing in pressure with respiration. Maximum height above the spine reached by the column of fluid was recorded as the opening pressure in cm CSF. Closing pressures were not measured since this would have required the removal of too large a volume of CSF. Less than 1 ml of CSF was removed for microscopy, culture, glucose, and lactate. In 2 children, a pressure
Results Between May, 1989, and August, 1990, 586 children with primary diagnosis of malaria were admitted, of whom 61 fulfilled the definition of CM (mean [SD] age 39 [21 °6] months, mean duration of disease 38 [2°6] days). Table n shows the clinical and laboratory features of these children on admission. 47 patients had LP on admission. Opening pressure was measured reliably in 26 children. No opening pressure was recorded in the remaining children because manometers were not available during the first 5 months of study (19), the child was struggling (3), or because LP was deferred until 24 h after consciousness had been regained a
(13). The figure shows opening pressure in 26 patients plotted against age, compared with the normal range.4 Mean (SD) opening pressure was 22-6 (7-3) cm CSF [16°7 (5-4) mm Kenya Medical Research Institute, Kilifi Coastal Unit, Kilifi, Kenya (C. R. J. C. Newton, MRCP, P. A. Winstanley, MD, N. Peshu, MB ChB, K. Marsh, MRCP); Department of Neurology, Institute of Child Health, London, UK (F J. Kirkham, MRCP); and Nuffield Department of Clinical Medicine, John Radcliffe Hospital, Oxford, UK (Prof G. Pasvol, DPhil, Prof D. A. Warrell, DM). Correspondence to Dr F. J Kirkham, Department of Neurology, Institute of Child Health, 30 Guilford Street, London WC1 N 2AP, UK.
ADDRESSES:
574
TABLE I-HERNIATION SYNDROMES IN CHILDEN WITH CM
Hg]. Mean cerebral perfusion pressure at LP was 52-6(13-8) Hg (range 30-3-88-3). LP was repeated in 2 children, because of recrudescence of fever. 1 child had opening pressures of 105, 21, and 14 em CSF (8,16, and 10 mmHg) at admission, 54, and 115 h after start of treatment, respectively. The other child had an opening pressure of 31 cm CSF (23 mm Hg) at admission and 29 cm CSF (21 mm Hg) at 84 h after start of treatment. He also had the lowest recorded cerebral perfusion pressure (30-3 mm Hg), and was left with dystonic motor disorder, delayed language, and cortical blindness at 3-months follow-up. CSF pressure fell from 21 to 10 mm Hg and from 20 to 9 mm Hg in the 2 children in whom it was measured with a transducer, 30 min after a 1 g/kg bolus dose of mannitol. Extensor posturing with Cheyne-Stokes respiration and absent gag was seen 12 h later in 1 of these children; hypoglycaemia was excluded. Posturing ceased after a further bolus of mannitol mm
and a full recovery was made. Of those studied with transcranial doppler ultrasound, forward flow was observed in 9 of 10 throughout diastole during every study; all survived. The tenth child had normal middle cerebral artery sonograms at admission; however, after 3 h, when brainstem reflexes were absent and the child was being hand ventilated, there was reverse flow throughout diastole with low velocity and a direction-of-flow index similar to that seen in brainstem death.5 12 of the 61 children (20%) died, all in coma, between 10 min and 24 h after admission. The fontanelle was closed in all, and 10 had had an LP. Herniation was observed in all 12 patients who died but in only 17 of 49 (35%) survivors (p
