|
|
Diffusing Capacity Predicts Pulmonary Complications After Lung Resection in Patients With and Without Obstructive Pulmonary Disease
M. K. Ferguson1, H. A. Gaissert2, J. D. Grab3. 1The University of Chicago, Chicago, IL, 2Massachusetts General Hospital, Boston, MA, 3Duke Clinical Research Institute, Durham, NC,
BACKGROUND:
Despite being an independent predictor of pulmonary complications after lung resection, diffusing capacity (DLCO) is not routinely used in assessing operative risk. This is due to uncertainty as to whether or not patients with normal spirometry require additional evaluation. We sought to determine whether DLCO is predictive of pulmonary complications after lung resection in patients with normal spirometry.
METHODS: We reviewed outcomes of major lung resection in The Society of Thoracic Surgeons General Thoracic Surgery Database from 2002 to 2006 to determine the relationship of DLCO (expressed as a percent of predicted; DLCO%) to postoperative pulmonary complications. Obstructive pulmonary disease (COPD) was defined as FEV1/FVC <70%. Pulmonary complications included air leak >5 days, atelectasis, pneumonia, ARDS, bronchopleural fistula, pulmonary embolism, initial ventilation >48 hrs, reintubation, tracheostomy, or other pulmonary complications. Analyses employed logistic regression using generalized estimating equations.
RESULTS: DLCO% was measured in just over half of the patients (3267 of 5739; 57%). Patient groups with and without measured DLCO% were similar. There were 1632 women and 1635 men with a mean age of 66.4 ± 10.4 years who underwent lobectomy (2859; 87.5%), bilobectomy (153; 4.7%), and pneumonectomy (255; 7.8%). COPD was identified in 1671 patients (51%). Pulmonary complications occurred in 11.7%, and the 30 day mortality was 1.5%. DLCO% was the preoperative clinical variable most strongly associated with the development of pulmonary complications. Decreasing DLCO% was incrementally related to an increased incidence of pulmonary complications (table) regardless of whether or not patients had COPD. There was no apparent interaction between DLCO% and COPD status in the model. For all patients the odds ratio for pulmonary complications was 1.23 for every 10 point decrement in DLCO% (95% confidence interval 1.15 to1.32; p<0.0001). The odds ratios were similar for patients with COPD (1.20) and those without COPD (1.24).
CONCLUSIONS: DLCO% predicts pulmonary complications after lung resection in patients with COPD and those with normal spirometry. In order to appropriately counsel patients and anticipate postoperative complications, all patients undergoing lung resection should undergo measurement of DLCO% regardless of COPD status.
Pulmonary complications grouped by DLCO% category| DLCO% category | <60 | 60-74 | 75-89 | ≥90 | | Complications in COPD group | 96 / 527(18.2%) | 67 / 488 (13.7%) | 30 / 339 (8.9%) | 24 / 317 (7.6%) | | Complications in non-COPD group | 43 / 259(16.6%) | 54 / 420 (12.9%) | 40 / 429 (9.3%) | 27 / 488 (5.5%) |
Back to 2008 Annual Meeting
Back to Program Outline
|
