We investigated prospectively the cause of fever in patients requiring
hospitalization after returning from the tropics. All consecutive adm
issions (n = 195) with oral temperature >37.0 degrees C at the time of
admission were enrolled. Final diagnosis as recorded on the discharge
summary by the attending physician and results of any relevant labora
tory or radiological investigations were recorded on standard proforma
, Malaria accounted for 42% of admissions; two patients had returned t
o Britain more than 6 months before presentation. The second largest g
roup was assumed to have a non-specific viral infection (25%). Cosmopo
litan infections (urinary tract infection, community-acquired pneumoni
a, streptococcal sore throat, etc.) accounted for 9%.Coincidental infe
ctions (schistosomiasis, filariasis, intestinal helminths) were found
in 16%. Serology was positive for HIV infection in 3%. The most useful
investigation was a malaria film, which was positive in 45% of cases
in which it was performed. The combination of thrombocytopaenia (plate
let count <100 x 10(9)) and hyperbilirubinaemia (bilirubin >18 IU/ml)
were useful predictive markers of malaria: all 23 patients with both a
bnormalities had positive malaria films. Malaria must be excluded in a
ny febrile patient returning from the tropics. In the absence of a pos
itive malaria film, the combination of a low platelet count and raised
bilirubin may suggest the need for an empirical course of therapy.