Study selection and population characteristics
The systematic search returned 289 articles, 83 of which were selected for full text screening. Fourteen articles were considered eligible for data extraction and meta-analysis according to our criteria of eligibility; two articles8, 9 were parts of another of our included articles10. The publications of Bez et al., 2004 and Rolinck-Werninghaus et al., 2004 were found to be post hoc analyses of the study of Kuehr et al., 2002. The study of Bez et al., 2004 was excluded from further analyses as it did not contain any outcome of interest, while the study of Rolinck-Werninghaus et al., 2004 was used for the outcomes of interest which were not available in the study of Kuehr et al., 2002. Eventually, twelve studies were included in quantitative synthesis and meta-analysis. Figure 1 shows the flow chart of the study selection process. A summary of the characteristics of the included studies are presented in Table 1. Briefly, 4 trials were conducted in Europe, 4 in USA and 3 in Japan, assessing a total of 3,211 patients. Two trials included only pediatric patients, while 5 included only adult patients (>17 years of age). In 3 trials AR was indicated as birch/grass Seasonal Allergic Rhinitis (SAR) / Seasonal Allergic Rhino-Conjunctivitis (SARC) in 3 trials the AR indication was ragweed SAR and in 3 trials the AR indication was cedar SAR. In 5 trials the symptom severity range was from moderate to severe. Omalizumab was administered subcutaneously in all the trials every 2 or 4 weeks to provide either a fixed dose or a dose dependent on body weight and serum IgE levels (Table 1). The methodologic quality for the majority of studies found to be good (Supplementary Table 1). All studies were double blinded and almost all of them provided a sufficient description regarding the follow-up of the patients; 5 studies performed an intention-to-treat (ITT) analysis. Nevertheless, most of the studies did not report adequately any methodological measurements to ensure allocation concealment and blinding of outcome assessment.