A Model to Predict Residual Volume from Forced Spirometry Measurements in Chronic Obstructive Pulmonary Disease

Background: Lung hyperinflation with elevated residual volume (RV) is associated with poor prognosis in adults with chronic obstructive pulmonary disease (COPD) and is a critical criterion for lung volume reduction selection. Here, we proposed that patterns within spirometric measures could represent the degree of hyperinflation.

Methods: Fractional polynomial multivariate regression was used to develop a prediction model based on age, biological sex, forced expiratory volume in 1 second (FEV₁), and forced vital capacity (FVC) to estimate plethysmography measured RV in patients in the Pittsburgh Specialized Center for Clinically Oriented Research (SCCOR) cohort (n=450). Receiver operating characteristic area under the curve (ROC-AUC) and optimal cut-points from the model were identified. The model was validated in a separate cohort (n=793).

Results: The best fit model: RV %est=[FVC %predicted] x 3.46-[FEV₁/FVC] x 179.80- [FVC % (sqrt)] x 79.53-[age] x 0.98- [sex] x 10.88 + 737.06, where [sex], m=1. R² of observed versus %predicted RV was 0.71. The optimal cut-point to predict an RV % >175% was 161. At this cut-point, ROC-AUC was 0.95, with a sensitivity 0.95, specificity 0.86, positive predictive value (PPV) of 97%, negative predictive value (NPV) of 76%, positive likelihood ratio (LR) of 6.6, and negative LR of 0.06. In a validation cohort of COPD patients (n=793), the model performed similarly, with a sensitivity of 0.82, specificity of 0.83, PPV of 85%, NPV of 79%, positive LR of 4.7, and negative LR of 0.21.

Conclusion: In patients with COPD, a model using only spirometry, age, and biological sex can estimate elevated RV. This tool could facilitate the identification of candidates for lung volume reduction procedures and can be integrated into existing epidemiologic databases to investigate the clinical impact of hyperinflation.