Distinguishing dengue fever from other infections on the basis of simple clinical and laboratory features: Application of logistic regression analysis

Background: Dengue fever is a frequent cause of admission to hospital in South East Asia, however many of the clinical characteristics and abnormalities on laboratory investigations at presentation are found in other common infections. Objectives: To describe the clinical and laboratory features of dengue fever and other common febrile illnesses in Singapore. Study design: We performed a prospective study of consecutive adult admissions to an infectious diseases hospital. Logistic regression analysis was used to identify symptoms, physical signs and laboratory features that differentiated dengue fever from other febrile illnesses within the first 2 days of admission. Results: Of the 381 patients studied, 148 had serologically confirmed den-Lie fever. Most of these had uncomplicated dengue fever, and only 9% had dengue haemorrhagic fever. A model based on clinical features alone (including a variety of cutaneous signs, pulse rate and the presence of pharyngeal injection) was able to differentiate dengue fever front other infections with a sensitivity of 74% and specificity of 79%. A model based on clinical features (rash) and laboratory parameters (white cell count, haemoglobin, prothrombin time, creatinine and bilirubin levels), achieved a sensitivity of 84% and specificity of 85%. Conclusions: A combination of simple clinical and laboratory parameters are potentially able to predict dengue fever with a high level of accuracy in adults presenting to hospital with febrile illnesses in Singapore. (C) 2005 Elsevier B.V. All rights reserved.