Background Guidelines recommend that for patients uncontrolled on inhaled corticosteroids (ICSs), step-up options include an increase in ICS dosage or addition of a long-acting β-agonist (LABA). Controversy persists about the best option in routine practice. Objective To compare asthma outcomes in patients whose first step-up from ICS monotherapy was by addition of LABA (LABA cohort) or increase in ICS dosage or formulation (ICS cohort). Methods Observational study using the General Practice Research Database, comparing outcomes in the following 12 months with regression modeling allowing for baseline cohort differences: age, sex, socioeconomic status, body mass index, comorbidity (rhinitis, heart disease), smoking status, short-acting β-agonist (SABA) use, oral corticosteroid use, and use of asthma complicating medication. Results We found 46,930 patients in the ICS and 17,418 in the LABA cohort. In adjusted analysis, the odds ratio (95% CI) of successful treatment (no hospitalization, no oral corticosteroid use, average daily SABA use <1 dose/d) was lower in the ICS cohort (0.75; 0.72-0.79). The adjusted odds ratio of needing rescue SABA prescriptions was higher in the ICS cohort (1.67; 1.59-1.76). However, the adjusted odds of using any oral corticosteroids were lower (0.75; 0.71-0.78), particularly of using 3 or more courses (0.50, 0.46-0.55), and the adjusted odds of respiratory hospitalization were lower (0.69; 0.59-0.81). Conclusion Although symptomatic control and rescue bronchodilator use may be improved by the addition of a LABA to ICS, there may be a lower risk of severe exacerbations and hospitalizations from ICS dose increase.