The Air-Test score for prediction of postoperative pulmonary complications after general anaesthesia: analysis of harmonised trial data and a prospective observational cohort.

Abstract

The Air-Test score, which measures preoperative and postoperative Spo₂ in room air, is associated with postoperative pulmonary complications (PPCs), a major cause of morbidity and mortality. We evaluated the utility of the Air-Test score in a heterogeneous population of surgical patients. This secondary analysis of a harmonised dataset comprised 3268 patients included in three multicentre RCTs and a prospective observational study of surgical patients undergoing general anaesthesia. The patients were randomly split into training (n=2288, 70%) and validation (n=980, 30%) cohorts. The primary outcome was a composite of severe PPCs developing within the first 7 postoperative days. The exposure of interest was the Air-Test, evaluated by assessment of discriminative performance and area under the curve (AUC) calculation. We also assessed sensitivity, specificity, positive, and negative predictive values. Severe PPCs within 7 days of surgery were recorded in 164/3268 (5.0%) participants (mean age: 63.5 (49.7-77.3) yr; 39% female). Participants with a positive Air-Test were more likely to have severe PPCs in both training (odds ratio [OR] 1.85, 95% confidence interval [CI] 1.21-3.15; P=0.007) and validation (OR 2.27, 95% CI 1.27-4.61; P=0.011) cohorts. The AUCs for detecting PPCs in both cohorts were similar (OR 0.57, 95% CI 0.50-0.64). In each cohort, sensitivity was 53-54%, specificity 59-60%, positive predictive value 6-8%, and negative predictive value 95-96%. The Air-Test is a simple noninvasive bedside score with a good performance in identifying patients after major surgery who will not develop severe PPCs.