Factors associated with malaria and arboviral disease in patients with imported febrile syndrome: a retrospective cohort study

Abstract

Objective. To identify predictors of malaria and arboviral disease in patients with febrile syndrome who seek care after traveling from tropical or subtropical locations. Methods. Observational retrospective cohort study. We collected demographic, epidemiologic, and clinical data; laboratory findings; and the clinical and final microbiologic diagnoses. Multivariate analysis was used to calculate indices of diagnostic accuracy (sensitivity, specificity, and predictive values) and coefficients of probability of combinations of variables. Results. Data for 291 patients with febrile syndrome were included; 108 had malaria (37.1%), 28 had an arboviral disease (9.6%), and 155 had other causes of fever (53.3%). Multivariate analysis showed patients most likely to have malaria were those from sub-Saharan Africa, adjusted odds ratio (aOR) of 45.85 (95% CI, 9.45-222.49); immigrants who returned to visit friends and relatives (VFR), aOR of 3.55 (95% CI, 1.21-10.46); or had platelet concentrations <150 000/mm3, aORa of 16.47 (95% CI, 5.46-49.70) or headache, aOR of 10.62 (95% CI, 3.20-35.28). The combination of these 4 variables gave a positive probability coefficient (PPC) of 23.72 (95% CI, 5.76-97.62). Patients with febrile syndrome most likely to have an arboviral disease were those from Central or South America, OR 5.07 (95% CI, 1.73-14.92), and those who had exanthems, OR 5.10 (95% CI, 1.72-17.02) or joint pain, OR 14.50 (95% CI, 3.05-68.80). The combination of these 3 variables gave a PPC of 20.66 (95% CI, 7.74-55.21). Conclusions. Patients with febrile syndrome with the greatest probability of having malaria are those from sub-Saharan Africa, those who are VFR, and those with platelet concentrations under 150.000/μL or headache. Arboviral disease was more likely in patients from Central and South America who had exanthems or joint pain.