Risk factors for prolonged air leak after pulmonary surgery: A systematic review and meta-analysis

This study aimed to comprehensively identify risk factors for the occurrence of prolonged air leak (PAL) in patients undergoing pulmonary surgery. Studies were retrieved from 3 databases, including PubMed, Web of Science, and EmBase up to 13 May 2020. We performed meta-analysis using Bayesian random effect models through divergence restricting conditional tessellation (DIRECT) algorithm. The effect size was expressed as odds ratio (OR) or mean difference (MD), each with 95% credible interval (CrI). The evidence quality was evaluated. Subgroup analyses and sensitivity analyses were conducted. Thirty-nine studies with 89006 patients were finally included. Pooled PAL incidence was 15%. Of 30 risk factors, 22 were significantly associated with increased PAL incidence. Five risk factors were ultimately selected with high evidence quality: smoking history (OR 1.84, 95%CrI 1.45 to 2.31, P<0.001), preoperative steroid use (OR 1.51, 95%CrI 0.87 to 2.65, P = 0.031), lower ratio of forced expiratory volume in 1 s and forced vital capacity (OR 1.99, 95%CrI 1.22 to 3.33, P = 0.005), non-fissureless technique (OR 2.14, 95%CrI 1.31 to 3.66, P = 0.003), and pathological TNM stage III/IV (OR 1.50, 95%CrI 1.07 to 2.12, P = 0.003). Regarding the negative impact of PAL on the personal cost and postoperative recovery, the verification of previous proposed factors and investigation of recently discovered ones both implied directions for risk stratification and the establishment of an applicable prediction model.