The impact of pulmonary function tests on early postoperative complications in open lung resection surgery: an observational cohort study

Abstract We investigated whether pulmonary function tests (PFTs) can predict pulmonary complications and if they are, to find new cutoff values in current open lung resection surgery. In this observational study, patients underwent open lung resection surgery at a tertiary hospital were analyzed (n = 1544). Various PFTs were tested by area under the receiver-operating characteristic curve (AUCROC) to predict pulmonary complications until 30 days postoperatively. In results, PFTs were generally not effective to predict pulmonary complications (AUCROC: 0.58–0.66). Therefore, we could not determine new cutoff values, and used previously reported cutoffs for post-hoc analysis [predicted postoperative forced expiratory volume in one second (ppoFEV1) < 40%, predicted postoperative diffusing capacity for carbon monoxide (ppoDLCO) < 40%]. In multivariable analysis, old age, male sex, current smoker, intraoperative transfusion and use of inotropes were independent risk factors for pulmonary complications (model 1: AUCROC 0.737). Addition of ppoFEV1 or ppoDLCO < 40% to model 1 did not significantly increase predictive capability (model 2: AUCROC 0.751, P = 0.065). In propensity score-matched subgroups, patients with ppoFEV1 or ppoDLCO < 40% showed higher rates of pulmonary complications [13% (21/160) vs. 24% (38/160), P = 0.014], but no difference in in-hospital mortality [3% (8/241) vs. 6% (14/241), P = 0.210] or mean survival duration [61 (95% CI 57–66) vs. 65 (95% CI 60–70) months, P = 0.830] compared to patients with both > 40%. In conclusion, PFTs themselves were not effective predictors of pulmonary complications. Decision to proceed with surgical resection of lung cancer should be made on an individual basis considering other risk factors and the patient's goals.

Abstract licence: CC BY-NC-ND

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.