Table 2

Logistic regression models of postoperative outcomes by OSA status and anesthesia type after multiple imputation

OSA statusAnesthesia type
Yes (reference=no)Neuraxial (reference=general)
OR (95% CI)P valueOR (95% CI)P value
Cardiac complications1.12 (0.88 to 1.43)>0.9990.74 (0.49 to 1.10)0.938
Pulmonary complications1.87 (1.51 to 2.30)<0.0070.50 (0.36 to 0.69)<0.007
Gastrointestinal complications1.52 (1.13 to 2.04)0.0420.68 (0.40 to 1.17)>0.999
Renal/genitourinary complications1.19 (0.95 to 1.48)0.9310.50 (0.36 to 0.68)<0.007
Thromboembolic complications1.17 (0.79 to 1.72)>0.9990.76 (0.39 to 1.50)>0.999
Delirium1.33 (1.03 to 1.72)0.1960.45 (0.32 to 0.63)<0.007
Prolonged LOS1.44 (1.31 to 1.57)<0.0070.51 (0.44 to 0.59)<0.007
  • Model adjusted for OSA status, type of surgery (TKA or THA), age at surgery (in years), time period in which the surgery occurred (categorized every 2 years from 2005 to 2014), sex (male or female), baseline laboratory values (hemoglobin, creatinine, INR, and platelets), primary type of anesthesia (neuraxial or general), dichotomized perioperative medications (benzodiazepines and ketamine), and ASA status (dichotomized as ASA status I–II or III–IV).

  • ASA, American Society of Anesthesiologists; INR, International normalized ratio; LOS, Length of stay; OSA, Obstructive sleep apnea; THA, Total hip arthroplasty; TKA, Total knee arthroplasty.