The risk of death due to cardiorespiratory causes increases with time after right pneumonectomy: a propensity score-matched analysis.

OBJECTIVES The study aimed to compare in-hospital, 30-day and non-cancer-related 6-month death rates in a series of right and left pneumonectomy cases matched according to functional parameters. METHODS A retrospective study was conducted on a series of 263 non-small cell lung cancer patients who underwent pneumonectomy. Left and right pneumonectomy cases were matched according to propensity scores using the following variables: age, coronary artery disease, any other cardiac comorbidity and predicted postoperative forced expiratory volume in the 1st second (ppoFEV1). After matching, 89 pairs of cases were selected. In-hospital, 30-day and 6-month crude and risk-adjusted death rates not related to cancer relapse or distant metastases were calculated for right and left pneumonectomy and compared on 2-by-2 tables using odds ratios. Death hazards were estimated by Cox regression, introducing the following independent variables in the model: age, cardiac comorbidity, ppoFEV1 and occurrence of any postoperative cardiorespiratory complication or bronchial fistula. RESULTS Non-cancer-related in-hospital, 30-day and 6-month death rates were, respectively, 8.4 (3.4 in left and 13.5 in right cases; P = 0.015), 11.8 (7.8 in left and 15.7 in right cases; P = 0.10) and 18.5% (12.4 in left and 24.7 in right cases; P = 0.033). On Cox regression, age, right pneumonectomy and the occurrence of postoperative cardiorespiratory complications (but not bronchial fistula) were related to the risk of death at 6 months. CONCLUSIONS The risk of death after pneumonectomy increases with time and strongly depends on the side of the operation (it is higher after right pneumonectomy) and on the occurrence of any postoperative cardiorespiratory complication. Neither hospital nor 30-day mortality should be reported as a valid outcome after pneumonectomy since they do not represent the real risk of the operation.