Autor(es): Veiga Oliveira, Paulo ; Cabral, Daniel ; Antunes, Mariana ; Torres, Carolina ; Alvoeiro, Magda ; Rodrigues, Cristina ; Sousa-Uva, Miguel ; Félix, Francisco
Data: 2022
Origem: Portuguese Journal of Cardiac Thoracic and Vascular Surgery
Assunto(s): Risk score; Major perioperative complications; Non-small-cell lung cancer
Objectives: Identify risk factors for major perioperative complications (MPC) after anatomical lung resection for NonSmall-Cell Lung Cancer (NSCLC) and establish a scoring system. Methods: Single center retrospective study of all consecutive patients diagnosed with NSCLC submitted to anatomical lung resection from 2015 to 2019 (N=564). Exclusion criteria: previous lung surgery, concomitant non-lung cancer related procedures, urgency surgery. Study population: 520 patients. Primary end-point: MPC defined as a composite endpoint including at least one of the in-hospital complications. Univariable and Multivariable analyses were developed to identify predictors of perioperative complications and create a risk score. Discrimination was assessed using the C-statistic. Calibration was evaluated by Hosmer and Lemeshow test and internal validation was obtained by means of bootstrap replication. Results: Mean age of 65 years and 327 (62.9%) were males. Mean hospital stay of 9 days after surgery. Overall MPC rate was 23.3%. Male gender, hypertension, FEV1<75%, thoracotomy, bilobectomy/pneumectomy and additional resection were independent predictors of MPC. A risk score based on the odds ratios was developed - Major Perioperative Complications of Lung Resection (MPCLR) scoring system - and ranged between 0 and 14 points. It was divided in 5 groups: 1-2 points (positive preditive value 15%); 3-4 (PPV 25%); 5-7 (PPV 35%); 8-9 (PPV 60%); >10 points (PPV 88%). The score showed rea- sonable discrimination (C-statistic=0.70), good calibration (P=.643) and it was internally validated (C-statistic=0,70 BCa95% CI,0.65-0.79). Conclusions: This study proposes a simple and daily-life risk score system that was able to predict the incidence of perioperative complications.
