Clinical and laboratory features that discriminate dengue from other febrile illnesses: a diagnostic accuracy study in Rio de Janeiro, Brazil
1 Germano Sinval Faria Teaching Primary Care Center/National School of Public Health, Oswaldo Cruz Foundation – Fiocruz, Rio de Janeiro, 21041-210, Brazil
2 Laboratory of Clinical Epidemiology/Evandro Chagas Clinical Research Institute, Oswaldo Cruz Foundation – Fiocruz, Rio de Janeiro, 21040-361, Brazil
3 Flavivirus Laboratory, Department of Virology, Instituto Oswaldo Cruz/FIOCRUZ, 21040-190, Rio de Janeiro, Brazil
4 Immunology Service/Evandro Chagas Clinical Research Institute, Oswaldo Cruz Foundation – Fiocruz, Rio de Janeiro, 21040-361, Brazil
5 Laboratory of Acute Febrile Illnesses/Evandro Chagas Clinical Research Institute, Oswaldo Cruz Foundation – Fiocruz, Rio de Janeiro, Brazil
BMC Infectious Diseases 2013, 13:77 doi:10.1186/1471-2334-13-77Published: 8 February 2013
Dengue is an acute febrile illness caused by an arbovirus that is endemic in more than 100 countries. Early diagnosis and adequate management are critical to reduce mortality. This study aims to identify clinical and hematological features that could be useful to discriminate dengue from other febrile illnesses (OFI) up to the third day of disease.
We conducted a sectional diagnostic study with patients aged 12 years or older who reported fever lasting up to three days, without any evident focus of infection, attending an outpatient clinic in the city of Rio de Janeiro, Brazil, between the years 2005 and 2008. Logistic regression analysis was used to identify symptoms, physical signs, and hematological features valid for dengue diagnosis. Receiver-operating characteristic (ROC) curve analyses were used to define the best cut-off and to compare the accuracy of generated models with the World Health Organization (WHO) criteria for probable dengue.
Based on serological tests and virus genome detection by polymerase chain reaction (PCR), 69 patients were classified as dengue and 73 as non-dengue. Among clinical features, conjunctival redness and history of rash were independent predictors of dengue infection. A model including clinical and laboratory features (conjunctival redness and leukocyte counts) achieved a sensitivity of 81% and specificity of 71% and showed greater accuracy than the WHO criteria for probable dengue.
We constructed a predictive model for early dengue diagnosis that was moderately accurate and performed better than the current WHO criteria for suspected dengue. Validation of this model in larger samples and in other sites should be attempted before it can be applied in endemic areas.