New findings on pathophysiology, epidemiology, and outcome have raised concerns on the relevance of the apnoea-hypopnoea index (AHI) in the classification of obstructive sleep apnoea (OSA) severity. Recently, a multicomponent grading system decision integrating symptomatology and comorbidities (Baveno classification), was proposed to characterise OSA and to guide therapeutic decisions. We evaluated whether this system reflects the OSA population, whether it translates into differences in outcomes, and whether the addition of AHI improves the scheme. A total of 14 499 OSA patients from the European Sleep Apnoea Database cohort were analysed. The groups were homogeneously distributed and were found to clearly stratify the population with respect to baseline parameters. Differences in sleepiness and blood pressure between the groups were analysed in a subgroup of patients after 24-36 months of treatment. Group A (minor symptoms and comorbidities) did not demonstrate any effect of treatment on outcome. However, groups B (severe symptoms, minor comorbidities), C (minor symptoms, severe comorbidities) and D (severe symptoms and comorbidities) were associated with improvement in either or both parameters with treatment. The AHI is an essential prerequisite of the diagnosis; however, adding the AHI did not improve the classification. Rather, it was inferior with respect to guiding the treatment decision. Thus, the Baveno classification allows a better stratification of the OSA population and may provide a better guidance for therapeutic decisions in OSA.