Statistical model for postoperative apnea-hypopnea index after multilevel surgery for sleep-disordered breathing

The objective of the study was to formulate a statistical model for postoperative apnea-hypopnea index (AHI) 3 and 12 months after multilevel surgery using the predictors preoperative AHI, body mass index (BMI) and age. The study design was a prospective cohort study. Data of 144 patients were collected prospectively 3 and 12 months after multilevel surgery for obstructive sleep apnea syndrome (OSAS) or upper airway resistance syndrome with excessive daytime sleepiness. The primary endpoint postoperative AHI and the secondary endpoint success according to the Sher criteria (postoperative AHI < 20 h and > 50% reduction of preoperative AHI) were modeled with multiple linear and logistic regression using the predictors preoperative AHI, BMI, age and the indicator whether the patient had undergone a tonsillectomy. Preoperative AHI and tonsillectomy had a highly significant positive influence on postoperative AHI after 3 months, whereas the influence of preoperative BMI was only marginally significant but numerically rather large. Age was not a significant decisive factor. The success according to the Sher criteria was highly significantly determined by the circumstance whether the patient had undergone a tonsillectomy, but not by the other predictors preoperative BMI or age. The responder rate with and without tonsillectomy was 58 and 19%, respectively. The odds ratio to be a responder if a tonsillectomy was conducted was 5.7. This study provides statistical models predicting postoperative AHI and success according to the Sher criteria after multilevel surgery for OSAS.