A fever study at the Colombo North Teaching Hospital (CNTH)

Abstract

OBJECTIVES: To determine the aetiology of febrile illnesses at CNTH, to identify effective laboratory tests for their confirmation, and to develop clinical prediction rules that will assist diagnosis. DESIGN, SETTING AND METHODS: A prospective cohort of in-patients is being studied during a one-year period. Patients with oral temperatures >38°C are eligible for inclusion unless they are aged <16 years, have been admitted for >24 hours or have received antibiotics in hospital. Written consent is obtained and a structured questionnaire is completed. Blood is taken for cultures, biochemistry assays, serology and PCR tests. Urine is taken for assays to detect antimicrobial activity. RESULTS: During the first 4 months, there were 180 eligible patients of whom 138 (77%) were recruited. The mean age was 36 years, the male:female ratio was 2.1 and 90% were from Gampaha district. There were no significant differences regarding age or sex in comparison to patients not recruited. Infections were unlocalised in 67% (50% unconfirmed, 43% dengue fever, 3% leptospirosis, 3% scrub typhus, 1% malaria). Localised infections were respiratory (9%), urinary tract (8%), neurological (4%), gastrointestinal (3%) and skin (2%). Non-infectious causes accounted for 3% of febrile patients. Bacteraemia was found in only 4% despite every patient having 2 high-quality blood cultures. Leucopenia or neutropcnia were useful early markers of dengue fever. CONCLUSIONS: Non-bacterial agents cause most febrile illnesses in Gampaha district. New laboratory tests and clinical prediction rules are required for their diagnosis.