Association of Preoperative Lung Function with Complications after Lobectomy Using Race-Neutral and Race-Specific Normative Equations

Abstract

Rationale: Pulmonary function testing (PFT) is performed to aid patient selection before surgical resection for non-small cell lung cancer (NSCLC). The interpretation of PFT data relies on normative equations, which vary by race, but the relative strength of association of lung function using race-specific or race-neutral normative equations with postoperative pulmonary complications is unknown. Objectives: To compare the strength of association of lung function, using race-neutral or race-specific equations, with surgical complications after lobectomy for NSCLC. Methods: We studied 3,311 patients who underwent lobectomy for NSCLC and underwent preoperative PFT from 2001 to 2021. We used Global Lung Function Initiative equations to generate race-specific and race-neutral normative equations to calculate percentage predicted forced expiratory volume in 1 second (FEV₁%). The primary outcome of interest was the occurrence of postoperative pulmonary complications within 30 days of surgery. We used unadjusted and race-adjusted logistic regression models and least absolute shrinkage and selection operator analyses adjusted for relevant comorbidities to measure the association of race-specific and race-neutral FEV₁% with pulmonary complications. Results: Thirty-one percent of patients who underwent surgery experienced pulmonary complications. Higher FEV₁, whether measured with race-neutral (odds ratio [OR], 0.98 per 1% change in FEV₁% [95% confidence interval (CI), 0.98-0.99]; P < 0.001) or race-specific (OR, 0.98 per 1% change in FEV₁% [95% CI, 0.98-0.98]; P < 0.001) normative equations, was associated with fewer postoperative pulmonary complications. The area under the receiver operator curve for pulmonary complications was similar for race-adjusted race-neutral (0.60) and race-specific (0.60) models. Using least absolute shrinkage and selection operator regression, higher FEV₁% was similarly associated with a lower rate of pulmonary complications in race-neutral (OR, 0.99 per 1% [95% CI, 0.98-0.99]) and race-specific (OR, 0.99 per 1%; 95% CI, 0.98-0.99) models. The marginal effect of race on pulmonary complications was attenuated in all race-specific models compared with all race-neutral models. Conclusions: The choice of race-specific or race-neutral normative PFT equations does not meaningfully affect the association of lung function with pulmonary complications after lobectomy for NSCLC, but the use of race-neutral equations unmasks additional effects of self-identified race on pulmonary complications.

Keywords: lobectomy; non–small cell lung cancer; pulmonary complications; pulmonary function testing; race-based interpretation.

Figure 1.

Flow diagram for inclusion in the study cohort. NSCLC = non–small cell lung cancer.

Figure 2.

(A and B) Area under the ROC curve (blue) for unadjusted (A) race-neutral and (B) race-specific logistic regression models in all patients, with 95% confidence intervals (red) at every decile of false positive rate (1 − specificity). FEV₁ = forced expiratory volume in 1 second; ROC = receiver operating characteristic.

Figure 3.

(A and B) Association of forced expiratory volume in 1 second (FEV₁) with pulmonary complications when FEV₁ is modeled as a nonlinear spline function using generalized additive models without covariate adjustment, using (A) race-neutral and (B) race-specific equations. The risk for complications rises more rapidly as FEV₁ progressively decreases.