Severe anaemia in Zambian children with Plasmodium falciparum malaria

Transcription

1 TMIH506 Tropical Medicine and International Health volume 5 no 1 pp 9 16 january 2000 Severe anaemia in Zambian children with Plasmodium falciparum malaria Godfrey Biemba 1, Dennis Dolmans 2, Philip E.Thuma 1,3, Günter Weiss 4 and Victor R. Gordeuk 5 1 Macha Mission Hospital, Choma, Zambia 2 Department of Medicine, Universiteit Utrecht, Utrecht, The Netherlands 3 Department of Natural Sciences, Messiah College, Grantham, PA, USA 4 Department of Medicine, University of Innsbruck, Innsbruck, Austria 5 Department of Medicine, The George Washington University Medical Center, Washington DC, USA Summary background Severe anaemia and cerebral malaria are highly prevalent complications of Plasmodium falciparum malaria among African children. The mechanisms of severe malarial anaemia, and the relative importance of this condition in comparison to cerebral malaria, are not known for many regions of Africa. methods We reviewed the records of 6200 children up to 6 years of age admitted to one rural Zambian hospital between 1994 and Severe malarial anaemia was defined as an haemoglobin concentration 5.0 g/dl in a patient with asexual forms of P. falciparum in the peripheral blood. Cerebral malaria was defined as impaired consciousness (Blantyre coma score 5) not attributable to any other cause in a patient with a positive malaria smear. results Severe malarial anaemia was found in 590 children (9.5% of paediatric admissions) and strictly defined cerebral malaria occurred in 286 children (4.6% of paediatric admissions); 98 of these patients had the combination of both complications. Severe malarial anaemia correlated strongly with the degree of parasitaemia, with malnutrition as indicated by low weight for age, with absence of fever and with presentation late in the malaria season. In comparison, patients with cerebral malaria were more often febrile and presented earlier in the malaria season. The case fatality rate of severe malarial anaemia (0.088) was about half that of cerebral malaria (0.189), but because severe malarial anaemia was more common, these two forms of complicated malaria were implicated in similar numbers of in-hospital paediatric deaths. conclusion Severe anaemia is a more common complication of P. falciparum malaria in hospitalized Zambian children than cerebral malaria and is associated with a similar number of deaths. Malnutrition and changes in immune response patterns due to prolonged exposure to P. falciparum may contribute to the development of this complication. keywords malaria, anaemia, cerebral malaria, Plasmodium falciparum, Zambia correspondence Dr Godfrey Biemba, CMAZ, P.O. Box 34511, Lusaka, Zambia. cmaz@zamnet.zm Introduction Malaria is a major threat to global health with about 40% of the world s population exposed to this infection. Every year million cases of Plasmodium falciparum malaria occur worldwide and of these one to two million die. Most of these deaths occur in African children between the ages of six months and six years (Molyneux et al. 1989). Two major manifestations of complicated malaria in African children are profound anaemia (severe malarial anaemia) and coma (cerebral malaria) (Brewster et al. 1990). As a cause of hospitalization, cerebral malaria seems to be more common than severe malarial anaemia in the Gambia (Brewster et al. 1990; Waller et al. 1995), while severe malarial anaemia seems to be more frequent in Kenya (Marsh et al. 1995; Snow et al. 1997). In Malawi and Zambia, both cerebral malaria and severe malarial anaemia (Molyneux et al. 1989; Slutsker et al. 1994; Mabeza et al. 1995) are major causes of hospitalization, but the relative importance of the two conditions has not been reported. In most areas, the mortality associated with cerebral malaria is higher than that associated with severe malarial anaemia. Our aim was to 2000 Blackwell Science Ltd 9

2 determine the importance of severe anaemia, in comparison to cerebral malaria, as a complication of falciparum malaria in children hospitalized at Macha Hospital in southern Zambia. Since the mechanisms that lead to severe malarial anaemia are not completely understood (Abdalla 1990), we also wished to gain insight into the factors that contribute to the development of severe malarial anaemia, and the mortality associated with this complication. Since there are annual and seasonal variations in malaria transmission in the study area, we reviewed paediatric admissions data over a period of three years. Methods The study was conducted at Macha Mission Hospital, a 208- bed major provider of primary and advanced health care in the Southern Province of Zambia. Macha Mission Hospital is located 72 km north of Choma, Zambia, in an area where P. falciparum malaria is endemic. The peak incidence of clinical malaria (January to June) correlates with the rainy season from November to April. Although the exact EIR has never been documented in the area, the pattern of malaria transmission is known. From hospital data and records, the Macha hospital catchment area depicts a stable malaria epidemiological pattern. Transmission is intense for 6 months of the year from November to April/May, and children 5 years are most severely affected, suffering repeated attacks each year. Unpublished community data collected by GB as part of a WHO study on the impact of health reforms on malaria control indicates that under-fives suffer, on average, three attacks of clinical malaria each year. Cerebral malaria is more common in children aged 5 years, and usually starts 2 4 weeks after the onset of rains in November; but there is some variation from year to year in terms of this pattern. The area is holoendemic for malaria with parasite rates of over 75% in children and approximately 60% in adults (inferred from unpublished screening data collected by GB, PT, VRG and Stenford Zulu in asymptomatic parasitaemia studies). True epidemics of malaria do not occur in this area, though there are years with exceptionally high transmission rates. Chloroquine resistance is a problem in Zambia (Cook 1986) and the clinical impression is that resistance to mefloquine, sulfadoxine-pyrimethamine and quinine is also emerging in the area around Macha Hospital. Type of study We conducted a retrospective study of the paediatric admissions to Macha Mission Hospital from 1994 to The hospital files of 6200 children, 72 months of age, were retrieved and reviewed. The following information was extracted: age, sex, weight, admission temperature, the result for the peripheral blood malaria smear on admission, haemoglobin concentration on admission, the lowest haemoglobin concentration measured during hospitalization, whether a blood transfusion was administered, whether the patient died and the final diagnosis or diagnoses given by the doctor who discharged the patient. All admissions to Macha hospital are recorded in an admission book and charts are drawn for each patient which are permanently filed after discharge. We reviewed the admission book records to compile a list of all paediatric admissions, then retrieved each patient s chart. We were able to locate more than 98% of the records. Each patient admitted to Macha hospital is seen by a qualified doctor within 24 h, and one doctor was usually in charge of the only paediatric ward. During the period under review, the first author (GB) was responsible for completing the records of more than 75% of patients, while two other doctors were responsible for the remainder; interobserver differences were therefore minimal. Definition of categories Severe anaemia was taken to be a haemoglobin concentration of 5.0 g/dl. Severe malarial anaemia was defined as a haemoglobin concentration of this level in a patient who had a positive malaria smear. Cerebral malaria was diagnosed if a patient 6 months of age with a positive blood smear for malaria presented with impaired consiousness as measured by a Blantyre coma score of 5. The coma score was determined as described by Molyneux et al. (1989) for all comatose patients as part of a screening process for enrolment into an ongoing cerebral malaria study. If the patient was eligible, the coma score was assessed more than 30 min after the seizure was controlled by arousing the patient as much as possible to exclude sleep. When there was no appreciable response, a painful stimulus (through horizontal pressure with a pencil) was applied on the nail or the sternum (using the blunt end of the pencil). Three parameters were assessed: best motor response (0 no or inappropriate response, 1 withdraws from painful stimulus, 2 localizes painful stimulus); best verbal response (0 no verbal response, 1 moans or cries abnormally with painful stimulus, 2 cries appropriately or speaks); and eye movement, assessed by asking the mother to change position while watching the child s eyes or by showing the child a lit torch from different positions (0 fails to watch or follow, 1 watches or follows mother s face). The scores for the three parameters were added; with a minimum score of 0 and a maximum score of 5. Asymptomatic parasitaemia is common among children in the Macha community, and, in a child with fever and a positive malaria smear, it is not always clear whether Blackwell Science Ltd

3 Table 2 Demographic and clinical features on admission according to P. falciparum parasitaemia in 6200 children admitted to Macha Hospital from 1994 to 1996 Malaria parasitaemia on admission P-value Not done for trend Characteristic (n 1057) (n 1386) (n 1814) (n 624) (n 567) (n 752) % female* Median age* (months) Weight* in kg (mean SD) % with low weight for age* % with fever* ( 37.8 C) % with splenomegaly** % with hepatomegaly** Haemoglobin** in g/dl (mean SD) *Total n varies from 6096 to 6190 because of missing information in some patient records. **Total n varies from 3609 to 3943 because of observation not being made or haemoglobin concentration not being determined in some patients. P. falciparum or some other infectious agent caused the symptoms leading to admission. For the purpose of this retrospective study, children without severe anaemia or cerebral malaria were classified as having malaria if the malaria smear was 2 (see below) or greater. This cut-off point almost certainly underestimates the prevalence of clinical malaria, but we wished to be conservative in our data analysis since in an area with high asymptomatic parasitaemia, including all positive smears would most likely encompass a significant number of patients with other fevers. Also, as shown in Table 2, the prevalence of splenomegaly and hepatomegaly increased markedly from parasitaemia 1 to parasitaemia 2. Children with temperatures 37.8 C were considered afebrile while those with temperatures 37.8 C were considered febrile. Children whose weights fell below the 5th percentile for age (Vaughan 1983) were classified as having low weight for age while those whose weight exceeded this level were considered to have normal weight for age. Management of patients Due to possible chloroquine resistance, most children admitted to hospital with the diagnosis of malaria were treated with a combination of chloroquine and sulfadoxinepyrimethamine. Any patient who did not respond to this regimen and those with cerebral malaria were treated with quinine. A course of sulfadoxine-pyrimethamine was routinely given along with the quinine therapy. In cases of severe anaemia or if anaemia was accompanied by signs of heart failure, blood transfusions (20 ml of whole blood/kg body weight) were administered as a matter of policy. The blood was donated by relatives and screened for HIV before administration. Laboratory procedures The haemoglobin concentration was determined with a photometric method (Sanford et al. 1933) or by dividing the haematocrit, measured by capillary tube centrifugation, by three. Asexual malaria parasites were counted in a thick blood smear stained with Giemsa under a light microscope at 1000x as follows: 0 no parasites; parasites per 100 microscopic fields; parasites per 100 fields; parasites per field; 4 more than 10 parasites per field (World Health Organization 1991). Of 3642 admitted patients with low Hb, 3609 (99.1%) had Hb concentration tests on admission. Of 2011 children 12 months, 1678 (83.4%) had a malaria smear. Statistical analysis Continuous variables that followed a normal distribution were compared with the Student t-test (two categories) or analysis of variance (several categories). Skewed continous variables were compared with the Mann Whitney U-test (two categories) or the Kruskall Wallis test (several categories). Proportions were compared with Pearson s 2. The relationships of the finding of severe anaemia (a haemoglobin concentration 5.0 g/dl) during hospitalization with patient characteristics on admission were examined using logistic regression. In the subgroups of patients with severe malarial anaemia and cerebral malaria the relationship of the outcome of death to patient characteristics was also examined using logistic regression Blackwell Science Ltd 11

4 Diagnosis n % Table 1 Most common final diagnoses among 6200 children admitted to Macha Hospital in southern Zambia Clinical P. falciparum malaria, combined cases* Blood smear 1, not severely anaemic, noncerebral All severe malarial anaemia (haemoglobin 5.0 g/dl) All cerebral malaria Severe anaemia and cerebral malaria Pneumonia Gastrointestinal disease Protein calorie malnutrition Surgery, trauma and burns Bronchitis or upper respiratory tract infection Tuberculosis *Children with peripheral blood malaria smears of 1 who did not have severe anaemia or cerebral malaria are not included. Results Patient characteristics 6200 children aged 72 months were admitted to Macha Hospital from 1994 to A malaria smear was performed for 5143 (83.0%) upon admission and the haemoglobin concentration was determined for 3642 (58.7%) either upon admission, subsequently during the hospitalization, or both. Positive malaria smears were found in 3757 (60.0%) children, but not all children with positive smears were classified as having clinical malaria. The most common final diagnoses are shown in Table 1. Clinical malaria without severe anaemia or cerebral complications was the most common diagnosis, followed by pneumonia, gastrointestinal disease, protein calorie malnutrition and severe malarial anaemia. Demographic features, clinical characteristics, and haemoglobin concentrations at the time of admission are shown in Table 2 according to degree of parasitaemia. Admission haemoglobin concentrations were lower in children with positive malaria smears than those with negative smears, and there was a significant trend to declining haemoglobin concentrations with progressively greater degrees of parasitaemia (Table 2). Relationship of severe anaemia to malaria A total of 659 children (10.6% of all admissions) had severe anaemia, defined as a lowest haemoglobin concentration of 5.0 g/dl. Of these, 590 had positive and 69 negative malaria smears. In a logistic regression model that adjusted for year of presentation, season of the year, weight for age, splenomegaly, hepatomegaly and cerebral malaria, the odds of having severe anaemia with a positive malaria smear were 5.6 times the odds with a negative malaria smear (95% confidence interval of ; P 0.001). The significant relationship between severe anaemia and increasing malarial parasite counts in peripheral blood is depicted graphically in Figure 1: 1.9% of children with negative malaria smears had lowest haemoglobin concentrations 5.0 g/dl compared to 8.2% of children with 1 malaria smears and 28.5% of those with 4 smears. Severe malarial anaemia A total of 590 children (9.5% of paediatric hospital admissions; 89.5% of severely anaemic children; 15.7% of all children with positive malaria smears) were classified as having severe malarial anaemia on the basis of a haemoglobin concentration 5.0 g/dl during hospitalization in combination with a positive malaria smear. The relationship between the Number with severe anaemia (%) Malaria parasitaemia on admission ND Figure 1 Percentage of paediatric patients with severe anaemia (lowest haemoglobin concentrations 5.0 g/dl) according to admission malaria parasitaemia. The total number for each category is as shown in Table 1. ND signifies that the parasite count was not determined. Parasitaemia as defined in laboratory procedures Blackwell Science Ltd

5 Table 3 Admission characteristics that in logistic regression modelling correlated significantly and independently with the finding of severe anaemia among patients with positive malaria smears Adjusted odds of 95% confidence Condition severe anaemia interval P Year Part of malaria season Early (Jan-Mar) 1.0 Late (Apr-Jun) Weight for age Normal 1.0 Low Temperature on admission Afebrile ( 37.8 C) 1.0 Febrile ( 37.8 C) Splenomegaly Absent 1.0 Present Hepatomegaly Absent 1.0 Present Blood smear for parasites Cerebral malaria Absent 1.0 Present finding of severe malarial anaemia and patient characteristics is shown in Table 3. Severe malarial anaemia was associated with higher parasite counts, cerebral malaria, hepatosplenomegaly, low weight for age, the absence of fever on admission to hospital, and presentation later in the malaria season. In the setting of 4 parasitaemia, severe anaemia occurred in 33.0% of children with low weight for age compared to 26.9% in children with normal weight for age (P 0.096), in 38.9% of afebrile children compared to 20.7% of those with fever (P 0.001), and in 33.9% of children who presented late in the season vs. 24.6% (P 0.006) who presented early. Severe malarial anaemia compared to cerebral malaria Of the 590 children with severe malarial anaemia, 492 had severe anaemia without cerebral complications (7.9% of paediatric admissions), while 98 patients suffered combined severe anaemia and cerebral malaria (1.6% of admissions). An additional 188 children had strictly defined cerebral malaria alone (3.0% of admissions). The characteristics of 492 patients with severe malarial anaemia (but not coma) are compared to those of 188 patients with cerebral malaria (but not severe anaemia) in Table 4. The presentation of severe malarial anaemia peaked later in the malaria season than the presentation of cerebral malaria. Compared to children with cerebral malaria, children with severe malarial anaemia were younger and weighed less. They also had lower admission temperatures, shorter hospital stays and lower mortality. Most of the children with severe anaemia received at least one blood transfusion during hospitalization. The case fatality rate was 7.3% in children with severe malarial anaemia not associated with cerebral malaria, and 8.8% in all children with severe malarial anaemia. The case fatality rate was 20.2% in cerebral malaria cases who did not have associated severe anaemia, and 18.9% in all cerebral malaria cases. Mortality Table 5 shows that 487 of 6200 patients died (7.9%). Many children had more than one final diagnosis recorded at the time of death, so actual cause of death was difficult to determine. Nevertheless, protein calorie malnutrition was the most common diagnosis in those that died followed by falciparum 2000 Blackwell Science Ltd 13

6 Table 4 Admission characteristics and clinical course in children with severe malarial anaemia compared to children with cerebral malaria Severe malarial anaemia Cerebral malaria Variable (n 492*) (n 188 ) P Yearly incidence Seasonal incidence Jan-Mar (%) 193 (39.2) 109 (58.0) Apr-Jun (%) 279 (56.7) 070 (37.2) Jul-Dec (%) 020 (4.1) 009 (4.8) Age (months; median) 020 (n 491) 031 (n 187) Female sex(%) (n 491) Weight (kg; mean SD) (n 488) (n 187) Low weight for age (%) (n 488) (n 187) 0.62 Splenomegaly (%) (n 362) (n 170) Hepatomegaly (%) (n 351) (n 169) Fever ( 37.8 C) (n 483) (n 185) Admission haemoglobin (g/dl; mean SD) (n 490) (n 175) Admission parasitaemia (%) Blood transfusion given (%) Duration of hospitalization 004 (1 77) 005 (1 28) (days; median and range) Mortality (%) * number of patients of a total of 590 with severe malarial anaemia who did not have cerebral complications. number of patients of a total of 286 with cerebral malaria who did not have severe anaemia. mortality was 8.8% in all 590 children with severe malarial anaemia. Mortality was 18.9% in all 286 patients with cerebral malaria Table 5 Most common final diagnoses in 487 hospitalized children who died % of total number Case fatality Diagnosis n* of children who died* rate Protein calorie malnutrition Falciparum malaria, combined Blood smear 1, not severely anaemic, noncerebral All severe malarial anaemia (haemoglobin 5.0 g/dl) All cerebral malaria Severe anaemia and cerebral malaria Pneumonia Gastrointestinal disease Tuberculosis Sepsis Congestive heart failure and other cardiac disease Meningitis * totals do not equal 487 or 100% because many patients had multiple final diagnosis listed Blackwell Science Ltd

7 malaria (combined categories), pneumonia and gastrointestinal disease. Severe malarial anaemia and cerebral malaria were implicated in 10.7% and 11.1% of hospital paediatric deaths, respectively. In a logistic regression model, blood transfusion significantly reduced the odds of death in children with severe malarial anaemia. Discussion We found severe malarial anaemia in 9.5% of all paediatric admissions to a rural Zambian hospital over 3 years and in 10.7% of paediatric deaths. For comparison, cerebral malaria was diagnosed in 4.6% of paediatric admissions and implicated in 11.1% of paediatric deaths. It should be noted that cerebral malaria was diagnosed on the basis of a Blantyre coma score of 4 rather than the cut-off of 2 used in more recent studies, and that it may be difficult to compare recent data relying on this parameter with ours. In spite of this, however, these results suggest that in southern Zambia severe malarial anaemia is twice as common in hospitalized children as cerebral malaria and is associated with a similar proportion of in-hospital deaths. It is not certain how these hospital data reflect the incidence of severe complications of malaria in the community, because some children die at home and some are taken to traditional healers. Since we found that blood transfusions significantly reduce mortality in hospitalized patients with severe malarial anaemia, it is almost certain that the case fatality rate of this complication is much higher among children who do not present at hospital. Taken together, these considerations suggest that severe malarial anaemia is a major cause of paediatric mortality in southern Zambia and that this complication may cause similar or larger numbers of deaths as cerebral malaria. The strong independent associations between increasing parasite counts and lower haemoglobin concentrations (Table 2) and between increasing parasite counts and severe anaemia (Figure 1) argue strongly for a cause-and-effect relationship between P. falciparum malaria and severe anaemia. In fact, 89.5% of severely anaemic patients had positive malaria smears. Therefore, haemolysis of erythrocytes by the parasites most likely is one factor that lowers the haemoglobin in these children. At the same time, several aspects of our study are consistent with the possiblity that mechanisms other than haemolysis contribute to the development of severe malarial anaemia. Further diagnostic tests which may be helpful to characterize anaemia, such as reticulocyte count, serum iron studies, folate and vitamin B12 levels, haptoglobin, lactate dehydrogenase and Coomb s test were not available. It was possible, however, to examine the relationship of several more general clinical features of the patients with the development of severe anaemia, namely, the nutritional status as determined by weight for age, febrile response as measured by admission temperature, and whether children presented early or late in the malaria season. As shown in Table 3, children who were malnourished (as indicated by a low weight for age) had a higher incidence of severe malarial anaemia compared to children with normal weight. A number of nutritional deficits can contribute to anaemia, and further investigations would be needed to identify the causes in the children in this study. Possibilities to be considered include lack of folic acid or iron. Children without fever on admission and those who presented later in the malaria season had a higher rate of development of severe anaemia compared to children who presented earlier in the malaria season or who were febrile (Table 3). These observations are consistent with the possibility that children with severe anaemia tend to have a chronic malarial infection, as has been proposed by other investigators (Abdalla et al. 1980), and they suggest that the development of severe anaemia may be partially mediated by a chronic interaction between the malaria organism and the body s immune system. Th-1 related cytokines such as interferon-gamma, interleukin-1 (IL-1) and tumour necrosis factor- (TNF- ) are recognized as major mediators of the anaemia of inflammation (Weiss et al. 1997), and also the chronic anaemia that may persist for weeks after successful treatment of malaria (Biemba et al. 1998; Camacho et al. 1998). A recent study from Ghana suggested that insufficient production of the Th-2 cytokine, IL-10, in response to high production of TNF-alpha may contribute to the development of severe malarial anaemia (Kurtzhals et al. 1998). The contrasting features of cerebral malaria and severe malarial anaemia that we observed (Table 4) might reflect different Th-1 and Th-2 immune response patterns in these two complications. Future studies will be required to investigate these issues. There are several limitations to this study. First, the investigation was conducted retrospectively. Second, the research was conducted in an area endemic for malaria where many children have asymptomatic parasitaemia. In such a region it may be difficult to make a definitive diagnosis of symptomatic malaria in a patient with fever and a positive malaria smear, for a positive smear may just be a coincidental finding in a patient whose symptoms and signs are actually due to a disease other than malaria. Third, not all patients had a blood sample taken to perform a malaria smear or to determine haemoglobin concentration. Fourth, infection with HIV is a substantial problem among children admitted to Macha Hospital, but tests for HIV were not routinely performed. We are therefore unable to say whether HIV infection may have contributed to the development of severe malarial anaemia, although several studies suggest that HIV-positive patients are not at a higher risk for P. falciparum infection and its complications than HIV-negative ones (Greenberg et al. 1991; Lucas 2000 Blackwell Science Ltd 15

8 et al. 1996; Kalyesubula et al. 1997). In conclusion, severe anaemia is a common complication of P. falciparum infections among Zambian children, is associated with high mortality and probably rivals cerebral malaria as a cause of death. The mechanisms of severe malarial anaemia are incompletely understood, but the balance in the Th-1 and Th-2 immune responses might play a role. Further studies to investigate the mechanisms and prevention of this condition are indicated. Acknowledgements We thank Brian Camus, Bill Green, Winnifred Shenouda, and Mr and Mrs H. Chizyuka for their help in this study. The work was supported by NIH grant no. UO1 A , by an FDA Orphan Products Development Grant (FD-R ), by a grant from the Office of Minority Health to the Cell Biology and Metabolism Branch, National Institute of Child Health and Human Development, and by a grant from the Austrian National Bank, Project References Abdalla S, Weatherall DJ, Wickramasinghe SN & Hughes M (1980) The anaemia of P. falciparum malaria. British Journal of Haematology 46, Abdalla SH (1990) Hematopoiesis in human malaria. Blood Cells 16, Biemba G, Gordeuk VR, Thuma PE, Mabeza GF & Weiss G (1998) Prolonged macrophage activation and persistent anemia in children with complicated malaria. Tropical Medicine and International Health 3, Brewster DR, Kwiakowski D & White NJ (1990) Neurological sequelae of cerebral malaria in children. Lancet 336, Camacho LH, Gordeuk VR, Wilairatana P, Pootrakul P, Brittenham GM & Looareesuwan S (1998) The course of anaemia after the treatment of acute falciparum malaria. Annals of Tropical Medicine and Parasitology 92, Cook GC (1986) Serious problems with antimalarial drugs. Journal of Infection 13, 1 4. Greenberg AE, Nsa W, Ryder RW et al. (1991) Plasmodium falciparum malaria and perinatally acquired human immunodeficiency virus type 1 infection in Kinshasa, Zaire. A prospective, longitudinal cohort study of 587 children. New England Journal of Medicine 325, Kalyesubula I, Musoke-Mudido P, Marum L et al. (1997) Effects of malaria infection in human immunodeficiency virus type 1-infected Ugandan children. Pediatric Infectious Disease Journal 16, Kurtzhals JAL, Adabayeri V, Goka BQ et al. (1998) Low plasma concentrations of interleukin 10 in severe malarial anaemia compared with cerebral and uncomplicated malaria. Lancet 351, Lucas SB, Peacock CS, Hounnou A et al. (1996) Disease in children infected with HIV in Abidjan, Cote d Ivoire. British Medical Journal 312, Mabeza GF, Moyo VM, Thuma PE et al. (1995) Predictors of severity of illness on presentation in children with cerebral malaria. Annals of Tropical Medicine and Parasitology 89, Marsh K, Foster D, Waruiru C et al. (1995) Indicators of lifethreatening malaria in African children. New England Journal of Medicine 332, Molyneux ME, Taylor TE, Wirima JJ & Borgstein A (1989) Clinical features and prognostic indicators in paediatric cerebral malaria: a study of 131 comatose Malawian children. Quarterly Journal of Medicine 71, Sanford AH, Sheard C & Osterberg AE (1933) The photelometer and its use in the clinical laboratory. American Journal of Clinical Pathology 3, Slutsker L, Taylor TE, Wirima JJ & Steketee RW (1994) In-hospital morbidity and mortality due to malaria-associated severe anaemia in two areas of Malawi with different patterns of malaria infection. Transactions of the Royal Society of Tropical Medicine and Hygiene 88, Snow RW, Omumbo JA, Lowe B et al. (1997) Relation between severe malaria morbidity in children and level of Plasmodium falciparum transmission in Africa. Lancet 349, Vaughan VC (1983) Growth and development. In: Nelson Texbook of Pediatrics (eds RE Behrman & VC Vaughan) W.B. Saunders, Philadelphia, pp Waller D, Krishna S, Crawley J et al. (1995) Clinical features and outcome of severe malaria in Gambian children. Clinical Infectious Diseases 21, Weiss G, Bogdan C & Hentze MW (1997) Pathways for the regulation of macrophage iron metabolism by the anti-inflammatory cytokines IL-4 and IL-13. Journal of Immunology 158, World Health Organization (1991). Basic Laboratory Methods in Medical Parasitology. WHO, Geneva Blackwell Science Ltd