Predicting mortality after lung volume reduction surgery

Purpose: To identify factors important for predicting death after bilateral lung volume reduction surgery (LVRS). Methods: Prospective data collection from 192 patients. Outcomes included vital status and time to death after LVRS. As potential predictors of death, we evaluated preoperative (preop) demographics, pulmonary function, exercise tolerance, dyspnea, and imaging data. Data were analyzed using logistic (hospital death) and Cox proportional hazards regression (late/all deaths). Results: At baseline (after pulmonary rehab.): 48% females; age 61±8 y (mean±SD); FEV₁ .71 ±.26 L; RV 5.8±1.3 L; PaO₂ 62±9 mm Hg; DLCO 9±4 ml/min/mm Hg; six min. walk distance (6MWD) 1138±285 ft. CT showed moderate to severe emphysema. 71% of patients had an upper lobe predominance of disease by radionuclide perfusion scanning. Actuarial survivals at 1, 2, and 3 y were 93%, 86%, and 81%. The best model for predicting all deaths (26 of 191 patients) indicated a protective effect for increasing 6MWD (Odds Ratio (OR) .998; p=.010) and FEV₁/FVC (OR 0.889; p=.002) and increasing risk with age (OR 1.06; p=.078). The best model predicting hospital deaths (10 of 191 patients) indicated a protective effect for increasing FEV₁/FVC (OR 0.73; p=.003) and DLCO (OR 0.77; p=.062) and increasing risk with age (OR 1.16; p=.037). This model accounted for 30% of the variance in death rate. The best model for predicting late deaths in hospital survivors (16 of 180 patients) indicated a protective effect for increasing PaO₂ (OR .938; p=.020). Conclusions: Age and parameters of pulmonary function and exercise capacity contributed to the prediction of death after LVRS. Imaging studies characterizing emphysema severity and distribution were not predictive of death. Clinical Implications: Physiologic and exercise testing should remain integral to patient evaluation and selection for LVRS.