Nitric oxide uptake in the lungs in smokers with emphysema
Description
Hypothesis Since its introduction in 1983, pulmonary diffusing capacity for nitric oxide (DLNO) has been underutilized clinically, partly due to limited awareness of its potential alongside pulmonary diffusing capacity for carbon monoxide (DLCO). This study evaluates DLNO’s effectiveness in predicting and classifying emphysema compared to DLCO or spirometry alone using standardized z-scores. It was hypothesized that DLNO z-scores would outperform conventional diagnostic metrics, either alone or in combination with DLCO. Data Overview This analysis pooled data from four published studies: (1) Int J Chron Obstruct Pulm Dis (2024), (2) Pneumologie (2021), (3) Eur Respir J (1990), and (4) Respir Med (2009). Studies 2, 3, and 4, which used breath-hold times (BHT: 8–12 s), were analyzed across three European hospitals. Participants were categorized as smokers with CT-diagnosed emphysema or smokers without emphysema. Data Processing Quality control excluded cases with BHTs outside 8–12 s, alveolar volume/total lung capacity (VA/TLC) ≥ 1.00, inspired volume/forced vital capacity (IV/FVC) < 0.85, or forced expiratory volume in one second (FEV1)/FVC ≥ 1.00. Non-emphysema cases with a residual volume/total lung capacity (RV/TLC) ratio < 0.20 were also excluded. Analysis After filtering, 408 patients remained (20% emphysema, 80% no emphysema). LASSO regression identified key variables, including z-scores of FEV1/FVC, FEV1, DLCO, TLC, RV/TLC, DLNO, KNO, KCO, and VA. Binary logistic regression was then applied, and model performance was assessed using Matthews correlation coefficient (MCC) for classification and the area under the receiver operating characteristic curve (AUROC) for discrimination. Notable Findings The best model incorporated z-scores of DLNO, FEV1, and TLC, yielding the lowest Bayesian Information Criterion (BIC). Substituting DLCO for DLNO increased the BIC, indicating a worse fit. The MCC for this model was 0.80 (95% CI: 0.68–0.89), and AUROC was 0.97 (95% CI: 0.95–0.98). A summed z-score model of DLCO, FEV1/FVC, and FEV1 also performed well but was less effective than the DLNO-based model. Implications No single predictor provided optimal diagnostic accuracy. Instead, DLNO combined with FEV1 and TLC significantly improved classification, outperforming traditional spirometry-based models. These findings support integrating DLNO into routine emphysema diagnostics, as it surpasses DLCO in predictive value and enhances clinical assessments.
Files
Steps to reproduce
This study utilized an individual participant data (IPD) meta-analysis, pooling raw participant-level data from prior studies to standardize variables, harmonize analytical methods, and enhance precision. Data were drawn from four published studies: [(1) Int J Chron Obstruct Pulmon Dis 2024, 19: 2123-2133; (2) Pneumologie 2021, 75: S3-S3; (3) Eur Respir J 1990, 3: 318-322; (4) Respir Med 2009, 103: 1892-1897]. The initial dataset included 496 cases, 52 of which used a 5-second breath-hold technique. After excluding cases that did not meet technical acceptability criteria and those with a breath-hold time (BHT) outside 8.0–12.0 seconds, 408 cases remained (323 without emphysema; 85 with CT-confirmed emphysema). Among the final dataset, 86% were smokers, with an interquartile range (IQR) of 14–43 pack-years. Pulmonary function testing adhered to ATS/ERS guidelines, incorporating DLNO, DLCO, VA, KCO, and KNO measured with a standardized 10 ± 2 s BHT (referred to as DLNO10s, DLCO10s, VA10s, KCO10s, and KNO10s). Lung function variables were converted to z-scores to normalize for confounders (age, sex, height, and equipment differences) using Global Lung Function Initiative (GLI) reference equations for spirometry [(Eur Respir J 2012; 40: 1324-1343)], lung volumes [(Eur Respir J 2021; 57(3) DOI: 10.1183/13993003.00289-2020)], and DLCO10s [(Eur Respir J 2017; 50(3) DOI: 10.1183/13993003.00010-2017)]. NO-CO double diffusion technique z-scores were derived from established reference equations [(Respir Med 2007; 101(7): 1579-84)] and those accounting for device variability [(BMJ Open Respir Res 2022; 9(1) DOI: 10.1136/bmjresp-2021-001087)]. Data quality control involved removing cases with poor technical quality, including BHTs outside 8.0–12.0 seconds, VA/TLC ≥ 1.00, inspired volume/FVC < 0.85 during the breath-hold maneuver, or FEV1/FVC ≥ 1.00. Additional exclusions applied to non-emphysema cases with RV/TLC < 0.20. Statistical analyses utilized least absolute shrinkage and selection operator (LASSO) regression to identify which variables (z-scores of FEV1/FVC, FEV1, DLCO, TLC, RV/TLC, DLNO, KNO, KCO, and VA) were most significant in predicting emphysema. Generalized linear models (GLMs) and generalized linear mixed-effects models (GLMMs) were then used with a random intercept for "Study" to account for clustering. Model selection was guided by the Bayesian Information Criterion (BIC) for frequentist models and the Leave-One-Out Information Criterion (LOOIC) for Bayesian models. The lowest BIC indicated the best balance of model fit and complexity, while the lowest LOOIC signified superior predictive accuracy and generalizability. Model performance was further assessed using the area under the receiver operating characteristic curve (AUROC) for discriminatory ability and Matthews Correlation Coefficient (MCC) for classification accuracy, particularly in imbalanced datasets. The pooled dataset files are saved in SPSS (.sav) and .csv formats.