Outcome of Spinal Versus General Anesthesia in Revision Total Hip Arthroplasty

A Propensity Score-Matched Cohort Analysis

Venkatsaiakhil Tirumala, MS; Georges Bounajem, MD; Christian Klemt, PhD; Stephen P. Maier, MD; Anand Padmanabha, MD; Young-Min Kwon, MD, PhD


J Am Acad Orthop Surg. 2021;29(13):e656-e666. 

In This Article


Because the overall volume of primary THA continues to increase, the incidence of revision THA will likely experience a concomitant increase. It has been well-established that revision THA is associated with a markedly greater risk of complications compared with primary THA, and optimization of modifiable preoperative, intraoperative, and postoperative risk factors are crucial in minimizing this risk.[25,26] One intraoperative factor that has been identified as having potential implications on THA outcome is anesthesia type. Although previous studies have demonstrated a reduction in complication rates with the use of spinal anesthesia compared with general anesthesia in the setting of primary THA, we investigated the potential effects in the setting of revision THA. We found markedly decreased surgical times, blood loss, transfusion rates, hospital LOS, and septic re-revision rates in the propensity-score-matched spinal anesthesia cohort.

The observed reduction in blood loss and rate of transfusion in the spinal anesthesia group is consistent with the existing primary THA and revision TKA literature.[7,13,27,28] This has been attributed to the sympathetic blockade, decreased vascular resistance, and resultant hypotension that spinal anesthesia induces. The reduction in blood loss may be further amplified by the shorter surgical times observed in the spinal anesthesia group because prolonged surgical time has been associated with increased blood loss in TJA.[29] This relative decrease in surgical time in the spinal anesthesia group is also consistent with the existing primary THA and revision TKA literature,[7,13,27] likely because of the reduced amounts of bleeding into the surgical field allowing for enhanced visualization and shorter surgical time. In addition, the surgeon awareness to complete surgery under the effectiveness of spinal anesthesia could have led to the shorter operation times observed in this cohort. Although it should be noted that case complexity may potentially confound surgical time, we used the extent of revision (modular only, single nonmodular, or both nonmodular) as a proxy for case complexity, with both cohorts having similar proportions of these implant revisions (P = 0.185). Furthermore, we controlled for surgical time, among other covariates, as a potential confounding factor in our multivariate regression analysis, and we determined that the general anesthesia cohort still had a demonstrably higher blood loss than the spinal anesthesia cohort.

We observed a markedly increased rate of septic re-revision in the general anesthesia cohort, which was present despite propensity-matching for known risk factors for infection in TJA. The cause of this observation is likely multifactorial. First, neuraxial anesthesia has been shown to increase lower-body oxygenation, likely via sympathetic blockade and enhanced local perfusion.[30,31] Moreover, volatile anesthetics used in general anesthesia have been shown to have suppressive effects on immune cell function.[31,32] Finally, increased surgical times have been shown to increase infection risk after TJA;[29] however, with surgical time controlled for, our study found that general anesthesia was associated with a 120% increased odds for requiring subsequent re-revisions for septic failures.

Owing to the paucity of anesthesia literature in the specific context of revision THA, most studies evaluated primary THA. Helwani et al[9] performed a retrospective propensity-matched cohort study using the American College of Surgeons–National Surgical Quality Improvement Program (ACS-NSQIP) database to compare 30-day outcome data of primary and revision THA patients who had undergone either regional or general anesthesia. Spinal anesthesia comprised 85.8% of the regional anesthesia cohort. They found a reduction in deep surgical site infection rates and hospital LOS, consistent with the results of our study. In addition, they found a notable reduction in 30-day cardiovascular and pulmonary complications in the regional anesthesia group. Basques et al[13] likewise used ACS-NSQIP data to compare 30-day outcomes in patients who had undergone primary THA with either spinal anesthesia or general anesthesia. Although the study did not observe a difference in hospital LOS, they found that general anesthesia was associated with increased surgical time (mean, +12 minutes per case), prolonged postoperative ventilator use, unplanned intubation, stroke, cardiac arrest, and blood transfusion. Mauermann et al[27] performed a meta-analysis of prospective studies comparing neuraxial and general anesthesia in patients undergoing elective primary THA. Although 3 of the 10 included studies used spinal anesthesia and the remainder used epidural anesthesia, the results of the meta-analysis revealed lower rates of deep vein thrombosis, pulmonary embolism, and blood transfusion in the neuraxial group and reduced surgical times (mean, 7.1 minutes per case) and blood loss. Likewise, a meta-analysis by Hu et al[33] showed that primary THA patients undergoing neuraxial anesthesia had reduced surgical times, blood loss, and need for blood transfusion. Our findings were in accordance with that of Mauermann et al and Hu et al; however, the overall incidence rates of deep vein thrombosis and pulmonary embolism were not markedly different between anesthesia cohorts.

To our knowledge, the only study comparing spinal anesthesia to general anesthesia in revision TJA is the revision TKA study by Wilson et al[7] They used ACS-NSQIP data to compare 30-day complication rates after aseptic revision TKA in patients receiving either general or spinal anesthesia. The general anesthesia cohort showed higher rates of unplanned readmission, nonhome discharge, blood transfusion, deep surgical site infection, and extended LOS. The surgical time was also markedly longer in the general anesthesia group (mean, 17 minutes per case). Although serious neurologic injury can occur from spinal anesthesia, it is rare. Symptoms lasting greater than 6 months have been reported to occur between approximately 2 to 4 incidences per 100,000 neuraxial anesthetics, whereas most major injuries resolve.[34] Transient neurologic symptoms may also occur.[35] In our study, we observed a zero incidence rate of such neurologic injuries in both cohorts within the 90-day postdischarge period.

Many of the aforementioned studies are multiinstitution database studies using the ACS-NSQIP data. The ACS-NSQIP database consists of prospectively collected data obtained from medical charts by trained personnel. Approximately 400 hospitals currently participate in the ACS-NSQIP, and it has been well-established as a database with validated methodology.[36,37] One clear advantage of these studies is the large sample sizes afforded by the database, with each study containing several thousands of patients who meet the inclusion criteria. However, one notable limitation of these studies is the fact that the ACS-NSQIP database is limited to 30-day postoperative data. Major complications such as wound dehiscence, infection, cardiopulmonary events, and mortality may occur beyond the 30-day postoperative period. Data from the Dutch Arthroplasty Register and the National Joint Registry for England, Wales, Northern Ireland, and the Isle of Man have shown that cumulative risk of infection after revision THA continues to increase well beyond the first postoperative month.[38,39] The ACS-NSQIP studies are therefore unable to detect many of these events. Our study, which uses our single-institution database, allows for markedly longer follow-up and monitoring of complications (mean, 5.79 years in spinal anesthesia group, 6.10 years in general anesthesia group). Moreover, although the multi-institutional nature of the ACS-NSQIP database strengthens the generalizability of studies that use it, it also confers potential disadvantages: although ACS-NSQIP uses a risk-adjusted approach, it is very challenging for studies to control for all aspects of interinstitutional heterogeneity.

Our study has potential limitations. Although the cohorts are propensity-matched, the retrospective nature introduces opportunity for biases associated with this study design. One limitation is the inability to control for case complexity between groups. Although the proportions of modular, single, and both-implant revisions are similar between groups, this may not be a perfect proxy for case complexity. Moreover, the single-institution nature of our study may affect the generalizability of our findings.

In conclusion, our study shows that propensity-matched patients who received spinal anesthesia for revision THA exhibited notbale reduction in surgical time, perioperative blood loss and transfusions, hospital LOS, and septic re-revisions compared with patients who received general anesthesia. These findings suggest that spinal anesthesia is a viable alternative to general anesthesia in revision THA for selected patients.