Hypoxic burden and respiratory support escalation within 72 hours after video-assisted thoracoscopic surgery pulmonary resection in lung cancer patients with combined fibrotic interstitial lung disease: a retrospective cohort study.

BACKGROUND: Patients with fibrotic interstitial lung disease (fILD) undergoing video-assisted thoracoscopic surgery (VATS) for lung cancer are at high risk of postoperative hypoxemia and respiratory deterioration. However, conventional assessment relies mainly on single-point peripheral oxygen saturation (SpO2) measurements, which fail to capture the cumulative burden of hypoxic exposure. This study aimed to evaluate whether a time-based postoperative hypoxic burden metric is associated with respiratory support escalation and postoperative outcomes. METHODS: A total of 120 patients with fILD and lung cancer who underwent VATS pulmonary resection were retrospectively enrolled. Hypoxic burden proportion was defined as cumulative hours with SpO2 <90% divided by total monitoring hours within 0-72 h postoperatively. Patients were classified as no hypoxic burden (G0, n=36), mild hypoxic burden (G1, n=44), or severe hypoxic burden (G2, n=40). The primary outcome was respiratory support escalation within 72 h. Progressively adjusted logistic regression, sensitivity analysis excluding preoperative SpO2, and Firth penalized logistic regression were performed. RESULTS: Respiratory support escalation occurred in 24/120 patients (20.0%), including 2/36 (5.6%), 6/44 (13.6%), and 16/40 (40.0%) in the G0, G1, and G2 groups, respectively (P<0.001). Although severe hypoxic burden and continuous hypoxic burden were associated with escalation in unadjusted analysis, these associations were attenuated after adjustment. In the full model, neither severe hypoxic burden [G2 vs. G0: adjusted odds ratio (aOR) =1.21, 95% confidence interval (CI): 0.11-13.01, P=0.87] nor continuous hypoxic burden (per 5-percentage-point increase: aOR =1.29, 95% CI: 0.89-1.89, P=0.18) remained significant. Sensitivity analysis excluding preoperative SpO2 showed a persistent association for continuous hypoxic burden, whereas Firth regression confirmed lobectomy and lower preoperative SpO2 as robust predictors. Severe hypoxic burden was associated with more postoperative pulmonary complications, prolonged oxygen therapy, and longer hospital/intensive care unit (ICU) stay. CONCLUSIONS: Hypoxic burden was associated with postoperative complications and resource consumption, but its independent association with escalation was attenuated after adjustment. It may be better interpreted as a postoperative process marker rather than a standalone predictor.