Three Spirometry-Based At-Risk Phenotypes Predict Incident Airflow Limitation: A Population-Based Study

Incident airflow limitation is frequently diagnosed at advanced stages. Identifying individuals at risk through primary care spirometry may enable earlier intervention, yet validated, pragmatic frameworks remain lacking.

We applied a framework of three mutually exclusive spirometric at-risk phenotypes, termed Three-Phenotype Spirometry-Based Identification (hereafter TriSpi), to 6123 participants from two population-based cohorts: KORA (n = 1973, 3-year Follow-up, derivation) and SHIP (n = 4150, 5-year Follow-up, validation). TriSpi⁺ individuals were defined as meeting criteria for one of the three phenotypes: Early airflow limitation (EAL, FEV₁/FVC > 0.7 and < 10th percentile or < 0.7 and > 5th percentile), small airway dysfunction (SAD) defined using FEF₅₀- or FEF₇₅-based thresholds; and preserved-ratio impaired-spirometry (PRISm). Associations with incident airflow limitation were tested using Firth's regression.

EAL, PRISm, and SAD (defined using either FEF₅₀ or FEF₇₅-based thresholds) were significantly associated with incident airflow limitation across cohorts and follow-ups. TriSpi⁺ individuals accounted for 26%-36% of the population, identifying 74%-93% of future cases, while TriSpi⁺ was associated with a 10-25-fold increase in risk. Negative predictive values exceeded 95% across definitions, and the number needed to screen among TriSpi⁺ individuals ranged from 7 to 13. EAL showed the strongest individual association (OR up to 53.2), while SAD was more common in younger adults.

Combining definitions for EAL, PRISm, and SAD enables robust prediction of incident airflow limitation. TriSpi may serve as a scalable, pragmatic approach for early risk stratification in primary care, in absence of post-BD spirometry.