Rationale: Chronic obstructive pulmonary disease (COPD) exacerbations accelerate lung function decline, reduce quality of life, and increase mortality. A subset of patients (n=457) from the FLAME study used the EXAcerbations of COPD Tool (EXACT®) to capture symptom-defined exacerbations. Objectives: To evaluate the effect of indacaterol/glycopyrronium (IND/GLY) versus salmeterol/fluticasone (SFC) on symptom-defined exacerbations measured using EXACT, and to assess differences between these events and exacerbations requiring healthcare resource use (HCRU). Methods: All patients in FLAME used an e-Diary to record and detect symptom deteriorations; HCRU-related exacerbations were confirmed by investigators. In patients using the EXACT questionnaire, the onset, recovery and magnitude of symptom-defined exacerbations were identified by changes in total scores relative to baseline. We analyzed the annualized rate and time-to-first symptom-defined (EXACT) exacerbation, and assessed differences between symptom-defined and HCRU events in terms of number, severity and concordance. Measurements and Main Results: A non-significant 17% reduction in the annualized rate of symptom-defined (EXACT) exacerbations (rate ratio: 0.83, 95% CI: 0.60, 1.14, P = 0.242) and a numerically longer time to first symptom-defined exacerbation were observed with IND/GLY versus SFC (HR 0.76, 95% CI: 0.56, 1.03, P = 0.075). These results were consistent with data from the overall FLAME population. 23.5% of symptom-defined (EXACT) events corresponded to HCRU events, and 22.2% of HRCU events were captured by EXACT (kappa index 0.24, 95% CI: 0.15, 0.33). Conclusions: Regardless of the exacerbation definition used, our findings support the use of LABA/LAMAs as the preferred treatment option for patients at risk of future exacerbations.