Predictive model of hypoxemia after shoulder arthroscopy: A retrospective observational study

Abstract

The study is aimed to establish a predictive model of hypoxemia after shoulder arthroscopy. The predictive model was based on a retrospective study with 756 patients who underwent shoulder arthroscopic surgery in Sichuan Orthopaedic Hospital from June 2019 to December 2020. Independent risk factors of hypoxemia in the post-anesthesia care unit (PACU) were screened out by the binary logistics regression and the primary predictive model was completed, which was evaluated by the receiver operating characteristic (ROC) curve and Hosmer-Lemeshow goodness-of-fit test. A separate cohort of 324 patients in the PACU from January 2021 to June 2021 was enrolled to validate the predictive model. Seven hundred fifty-six patients and 19 variables were enrolled in the binary logistics regression and 324 patients were validated by the primary predictive model. Logistics regression showed that application of irrigating solution ≥20 L, age, body mass index, and number of B-lines were independent risk factors of hypoxemia in the PACU (P < .05). The risk predictive model of hypoxemia in the PACU was established according to those factors. The model was validated by the Hosmer-Lemeshow test and the area under the curve of ROC was 0.823. The model area under the curve of external effect subject ROC was 0.870. The risk predictive model established in our study can predict the risk of hypoxemia in the PACU well and have good efficacy.Abbreviations: AUC = area under the receiver operating characteristic curve, BMI = body mass index, BP = blood pressure, CI = confidence interval, NB = the number of B-lines, PACU = post-anesthesia care unit, ROC = receiver operating characteristic.