Identifying allergic rhinitis cases from a screening questionnaire: the GEIRD project

Background: Both diagnostic tests and medical history are necessary to provide a good phenotypization of Allergic Rhinitis (AR) since diagnosis using one method alone can produce a lot of false positives and false negatives. Objective: To investigate the validity of the process of selection in identifying AR cases and to evaluate AR phenotype in the frame of GEIRD (Gene-Environment Interaction in Respiratory Diseases), a multi-center nested multi-case-control study. Methods: Subjects reporting nasal allergies or having ever had a problem with sneezing, or a runny or a blocked nose in the absence of a cold or flu (probable cases of AR; n=136) in a screening questionnaire and probable controls (n=262) aged 20-64 years were sent to clinic for phenotypization. Subjects were then labeled as controls, AR subjects and non allergic rhinitis (NAR) subjects depending on patient history and skin prick test. Results: Probable controls and probable cases of AR were phenotyped as 205 controls, 112 AR cases and 65 NAR cases. A good validity of the process of selection was found: the positive predictive value of the definition of probable case of AR was 68.5%. The repeatability was good for nasal allergies between the screening questionnaire and the clinical interview (K=0.73; C.I. 95%=0.61-0.85) while a lower agreement was for “runny/blocked nose” symptom (K=0.52; C.I. 95%=0.42-0.61). AR subjects presented a different profile from NAR subjects both in the screening questionnaire and in the clinical interview: in particular, they reported a higher proportion of “Hay fever” symptoms both in the screening questionnaire (50.9% vs 16.9%, p<0.001) and in the clinical interview (53.6% vs 6.15%). Conclusions: The process of selection of AR cases was efficient and adequate for epidemiological purpose. Questions used to select AR cases were valid and repeatable. AR phenotype was different from NAR phenotype.