Comparison of six disease severity scores for allergic rhinitis against pollen counts a prospective analysis at population and individual level

Background Many different symptom (medication) scores are nowadays used as measures of allergic rhinoconjunctivitis severity in individual patients and in clinical trials. Their differences contribute to the heterogeneity of the primary end-point in meta-analyses, so that calls for symptom (medication) score harmonization have been launched. Objective To prospectively compare six different severity scores for allergic rhinitis (AR) against pollen counts at both population and individual levels. Methods Two groups of children with seasonal AR and grass pollen sensitization were recruited in Ascoli, Italy (n = 76) and Berlin, Germany (n = 29). Symptoms and drug intake were monitored daily for 40 and 30 days of the grass pollen season in 2011 (Ascoli) and 2013 (Berlin), respectively, through an Internet-based platform (AllergyMonitor™, TPS Production srl, Rome, Italy). From the gathered data, the informatics platform automatically generated one symptom score (RTSS) and five symptom–medication scores (RC-ACS©, ACS, RTSS[LOCF], RTSS[WC] and AdSS). Values were then statistically normalized for reciprocal comparison and matched against the daily variations of local grass pollen counts (Spearman's rank correlation). Results The grass pollen counts were higher in Ascoli than in Berlin (peak values 194 vs. 59 grains/m3). At population level, the trajectories of the normalized average values of the six scores differed only slightly in both studies and correlated well with the pollen counts (ranges r2: 0.38–0.50 in Ascoli, 0.41–0.56 in Berlin). By contrast, in individual patients, trajectories of different scores were often quite heterogeneous. The RTSS[WC] had a very low discriminatory power and generated in many patients long, flat horizontal segments. Conclusions Disease severity scores for seasonal AR, as evaluated via an Internet-based platform, tend to provide similar results at population level but can often produce heterogeneous slopes in individual patients. The choice of the disease severity score might have only a low impact on the outcome of a very large clinical trial, but it may be crucial in the management of individual patients.