Background: The STOP-BANG questionnaire is a validated, eight-point dichotomized scale used to screen preoperative patients for obstructive sleep apnoea. Sleep apnoea causes hypoxaemia, and nocturnal oxygen desaturation is diagnostic in these patients. We tested the hypothesis that STOP-BANG score is associated with hypoxaemia after noncardiac surgery.
Methods: This analysis was a sub-study of VISION, a prospective cohort study of perioperative cardiovascular events. With institutional review board approval, we included 630 patients in the final analysis. We assessed the association between the STOP-BANG score and postoperative hypoxaemia, defined as integrated area under the curve of [Formula: see text] saturation of 90% per h using median quantile regression. Secondarily, we selected a subset of STOP-BANG questions that best predicts postoperative hypoxaemia using 'Least Absolute Shrinkage and Selection Operator' method, and then assessed the association between the new score based on the selected questions and the primary outcome using quantile regression.
Results: The median [q1, q3] area under [Formula: see text] of 90% per h was 0.09 [0.02, 0.39] %-h. The STOP-BANG score was not associated with hypoxaemia, with a multivariable slope coefficient of 0.002 (95% CI: -0.01, 0.01) %-h for a unit increase in the score (P=0.76). Secondarily, no association was found between the new score based on the two retained STOP-BANG questions, treatment for hypertension and neck circumference >40 cm, and the primary outcome with a multivariable slope coefficient of 0.03 (98.3% CI: -0.01, 0.06) %-h/score (P=0.07).
Conclusions: The STOP-BANG score does not predict hypoxaemia in adults recovering from noncardiac surgery.
Clinical trial registration: NCT00512109.
Keywords: anaesthesia; apnoea; care; hypoxaemia; obstructive sleep; oximetry; postoperative; pulse.
© The Author 2016. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved. For Permissions, please email: email@example.com.