Febrile illness is the primary cause of childhood outpatient attendance, admission to
hospital and death in Africa. This series of studies were aimed at ascertaining the
treatable causes of infection in children admitted to a district hospital typical of those
found throughout East Africa, in an area of high transmission of malaria. The studies
were also designed to determine the clinical correlates of infection and predictors of
mortality, looking in particular at malaria, invasive bacterial disease and HIV
infection. These studies also explored to what extent clinical examination by one
group of staff was replicable by another.
After informed consent a detailed history and structured examination was performed
on all children admitted to the hospital. Blood was drawn for culture, microscopy for
malaria, HIV testing, full blood count, bedside haemoglobin, blood glucose and
lactate measurement and HRP-2 based rapid diagnostic test for falciparum malaria.
Outcomes were recorded at death or discharge.
Sufficient data was available on 3,639 children including 184 deaths (5.1%).
Invasive bacterial disease was detected in 341 children (9.4%) and HIV in 142
(3.9%). Children with HIV and those with evidence of recent malaria were
significantly more likely to have invasive bacterial disease. The most common
organisms isolated were non-typhi Salmonella (46.9%), Strep, pneumoniae (16.4%)
and Haemophilus influenzae b (11.4%). The most frequently encountered pathogen
was P. falciparum, with 2,195 children found to have asexual parasitaemia (60.3%).
Falciparum parasitaemia was detected in 100 children with invasive bacterial disease (29.3%). Falciparum malaria was detected in over half (51.6%) of childhood deaths,
invasive bacterial disease was documented in 31.5%.
In children with a positive blood slide for malaria, WHO severe malaria criteria
identified 91.6% of the children that died. A multivariate analysis showed that signs
of malnutrition, respiratory distress, altered consciousness, hypoxia according to
pulse oximetry, hypoglycaemia, raised blood lactate, invasive bacterial disease and
female sex were all associated with an increased risk of death. In children with
negative blood slides signs of malnutrition, respiratory distress, altered
consciousness, hypoglycaemia, raised blood lactate and invasive bacterial disease
were all independently associated with mortality by multivariate analysis.
WHO defined criteria of syndromes which would warrant antibiotics predicted 56%
of cases of coinfection with invasive bacterial disease and malaria and 69.7% of
cases of invasive bacterial disease in slide negative children. Treating all children
with severe malaria for bacterial disease would result in 71% of children with
coinfection being treated. In children with negative slides including severe anaemia
or prostration as syndromes requiring antibiotic therapy would have resulted in
74.7% of children with invasive bacterial disease receiving antibiotic therapy. There
was moderate agreement between staff over the presence of clinical signs in children,
with hospital nurses performing as well as hospital clinical officers. Agreement was
better in children over 18 months of age and in children who were not crying during
Current WHO guidelines on antibiotic use performed poorly in this setting. Gram
negative infections were the most common cause of invasive infection and many of
these are likely to be resistant to penicillin and other commonly used antibiotics.
Consideration should be given to expanding the indications for antibiotic use and
using more broad-spectrum antibiotics in severely ill children.