Objectives: We assessed the value of the clinical symptoms included in the case definition of pertussis in household contacts of laboratory-confirmed cases.
Methods: A prospective epidemiological study was made in two Spanish regions. Household contacts were identified for each confirmed case reported during 2012 and 2013. Two clinical samples were taken to determine the presence or absence of Bordetella pertussis by culture or real-time PCR. Clinical variables, age and vaccination status were recorded. Positive and negative likelihood ratios (PLR, NLR) were estimated for each symptom.
Results: 2852 household contacts of 688 confirmed cases were reported. 178 household contacts with clinical symptoms were analyzed: 150 were laboratory confirmed and 28 were not. The clinical symptom with the highest PLR in comparison with the NLR was paroxysmal cough(PLR 4.76; 95% CI 1.91-11.87 and NLR 0.37; 95% CI 0.28-0.49). The contrast between the PLR and NLR was especially important for persons aged <18 years (PLR 7.08; 95% CI 1.10-45.74 and NLR 0.32; 95% CI 0.21-0.49).
Conclusions: The clinical symptoms of pertussis are poor predictors of pertussis disease, independently of the vaccination status. Differences were observed between persons aged <18 years and adults. To adopt the appropriate treatment and control measures, rapid laboratory confirmation by PCR of all household contacts of confirmed cases who present any clinical symptoms compatible with pertussis should be recommended.
Keywords: Clinical symptoms; Household contacts; Pertussis; Sensitivity; Specificity.
Copyright © 2017 The British Infection Association. Published by Elsevier Ltd. All rights reserved.