This study was performed to evaluate the sensitivity of immunoglobulin (Ig)G and IgA antibodies to pertussis toxin and filamentous hemagglutinin in diagnosing pertussis from a single serum sample. The pertussis group was defined according to the World Health Organization pertussis case definition. The control group coughed for 21 days or more but had no microbiological or serological evidence of Bordetella infection. Both cohorts were divided into infants (< 12 months of age), toddlers (1-4 years) and school children (5-10 years). There were 525 subjects in the pertussis group and 321 in the control group, with an even distribution of genders. IgG and IgA antibodies to pertussis toxin and filamentous hemagglutinin were measured in a standardized enzyme immunoassay. Antibody levels beyond the 95 percentile of the control cohort were regarded as indicative of recent contact, setting the specificity level at 0.95. Acute serum samples drawn between 1 week and 3 weeks after the onset of coughing showed a low sensitivity (2-19%) for diagnosing pertussis. In convalescent samples taken 5-10 weeks after the onset of symptoms, detection of IgG anti-pertussis toxin was the best single test, with a sensitivity of 61%, 65%, and 74% in infants, toddlers and school children, respectively. A combination of IgG anti-pertussis toxin and IgA anti-filamentous hemagglutinin using age-specific reference values had a sensitivity of 81-89% in diagnosing pertussis from a single serum sample taken 5-10 weeks after the beginning of symptoms.