OBJECTIVES: Formoterol/budesonide (BUD/FORM) and salmeterol/fluticasone propionate (SAL/FLU) have been shown to be effective in the treatment of asthma. This retrospective, observational study compared characteristics of patients initiating treatment with BUD/FORM or SAL/FLU in routine clinical practice, subsequent treatment outcomes and resource utilisation. METHODS: A cohort of German patients diagnosed with asthma, followed from 12 months before to 12 months after treatment initiation with BUD/FORM or SAL/FLU, was extracted from a longitudinal, primary care database of records collected from June 2000 to June 2006. The primary outcome was the proportion of successfully treated patients defined according to utilisation of short-acting beta-agonists (SABA) and switches or addition of controller medications during the post-index year. Secondary outcomes included resource utilisation and acute exacerbations, defined as at least one oral corticosteroid (OCS) prescription and/or hospitalisation related to asthma. Regression models were used to adjust for patient characteristics, including treatment history. RESULTS: The BUD/FORM and SAL/FLU groups included 1,456 and 982 patients, respectively. Prior to treatment initiation, patients in the BUD/FORM group received less asthma-related OCS prescriptions (mean difference: -0.049, p=0.0328) but utilisation of SABA, ICS and LABA was similar. In the year following treatment initiation, patients initiating on BUD/FORM had a greater probability of treatment success (OR=1.54, p=0.0001), fewer acute exacerbations (-40.2%, p=0.0007) and fewer physician visits (-4.4%, p=0.0259), than those on SAL/FLU, after controlling for baseline characteristics. The BUD/FORM patients were less likely to switch to an alternative ICS+LABA combination (OR=0.58, p=0.0067). Numbers of hospitalisations, referrals and work absences did not differ between groups. CONCLUSION: BUD/FORM was associated with a lower probability of treatment switches, fewer acute exacerbations, and with similar or lower resource utilisation compared with SAL/FLU. Although the groups appeared well matched at treatment initiation, these results should be interpreted with caution given the observational nature of this study.