Abstract
This study was designed to compare length of hospital stay, and 30-day major and minor complications between patients undergoing total knee arthroplasty (TKA) with general anesthesia, to those undergoing TKA with spinal or epidural anesthesia with or without regional nerve blocks.
Patients 18 years and older undergoing TKA between the years of 2005 and 2016 were identified from the American College of Surgeons National Surgical Quality Improvement Program (NSQIP). Patient demographics, anesthesia type, length of operation and hospital stay, as well as 30-day major and minor complications were collected from the database. Patients with “primary anesthesia technique” codes for either spinal or epidural anesthesia along with “other anesthesia technique” codes for regional anesthesia were assumed to have been given a regional nerve block. Chi square tests, and analysis of variance were utilized to evaluate unadjusted differences in demographics and outcomes between anesthesia types. Multivariable regression was utilized to compare outcomes (length of stay and complications) between anesthesia types, while adjusting for age, American Society of Anesthesiologist (ASA) class, comorbidities, sex, steroid/immunosuppressant use, body mass index (BMI), diabetes, length of operation and smoking status.
A total of 214,665 TKA patients were identified (average age 67 ± 10 years). Of these, 257 (0.12%) underwent epidural anesthesia with a nerve block (EB), while 2,318 (1.08%) underwent epidural anesthesia with no block (E), 14,468 (1.08%) underwent spinal anesthesia with a block (SB), and 85,243 (39.7%) underwent spinal anesthesia with no block (S), and 112,377 (52.4%) underwent general anesthesia (G).
The unadjusted length of stay (LOS) was significantly longer in the E group (3.67 ± 5 days) compared to the G group (3.1 ± 3.9 days), while the unadjusted LOS was significantly shorter in the EB group (2.6 ± 1.2), and both SB and S groups (2.6 ± 3 and 2.9 ± 3, respectively), compared to the G group p < 0 .001. Following covariable adjustment, anesthesia type remained an independent predictor of length of stay. Compared to the G group, patients in the E group stayed 0.56 days longer (95% Confidence interval [95%CI] 0.42 – 0.71 days), while patients in the SB were discharged 0.28 days (95%CI 0.21 – 0.35 days) earlier, and those in the S group were discharged 0.06 days earlier (95%CI 0.02–0.09), (p < 0 .0001).
While the unadjusted rates of major complications were not significantly different between groups, the unadjusted rates of minor complications were higher in the E, EB, and G groups compared to the S and SB groups. Following covariable adjustment, there were no differences between groups in the risk of minor complications.
In conclusion, these data indicate that anesthesia type following TKA is associated with length of hospital stay, but not with 30-day complications. After adjusting for relevant covariables, patients who received epidural anesthesia without a nerve block for TKA were discharged later, while patients who received spinal anesthesia, both with and without a nerve block for TKA were discharged earlier, compared to patients who received general anesthesia for TKA.
