Neuraxial anesthesia has advantages compared to general anesthesia including reduced requirements for sedatives, opioid analgesics, and airway instrumentation. However, conflicting data exist regarding the differing perioperative morbidity and mortality outcomes between the anesthetic techniques,1,2 and therefore many anesthesiologists may be reluctant to routinely recommend neuraxial anesthesia to their patients undergoing major surgery.
Health care practice has evolved significantly over the past 2 decades with the development of evidence-based therapies providing prophylaxis against surgical site infection and thromboembolic events. Furthermore, perioperative care has evolved with the introduction of new surgical techniques, expansion of interventions to an ageing population, introduction of anesthetic agents with predictable pharmacokinetics and rapid recovery times, improved antiemetic and multimodal analgesic therapies, and changes in anesthetic techniques. Therefore, there is an ongoing need to evaluate perioperative outcomes using studies that reflect recent and current practice.
Several contemporary randomized control trials (RCTs) have investigated perioperative outcomes associated with neuraxial anesthesia compared with general anesthesia.3–13 RCTs are the gold standard of evidence-based medicine; however, their content forms only a portion of the available knowledge. Recently, large, observational population-based studies have provided unique insights into the impact of neuraxial anesthesia on a surgical population.14–39 In contrast to RCTs, these studies do not have strict inclusion and exclusion criteria, thus their findings reflect routine clinical practice.
In this systematic review and meta-analysis of both RCTs and large-scale population studies, our goal was to evaluate the role of neuraxial anesthesia as a health care intervention in the context of contemporary perioperative care. Specifically our objectives were the following: (1) to summarize findings of contemporary RCTs and large-scale population-based studies, where neuraxial anesthesia was used for major truncal and lower extremity surgery in adults; (2) to quantify the effect size of neuraxial anesthesia when used alone or when combined with general anesthesia, if it exists, compared to general anesthesia with regard to predefined outcomes (30-day mortality, cardiopulmonary morbidity, surgical site infection, thromboembolic events, blood transfusion, and resource use [critical care services, length of hospital stay]); and (3) evaluate the quality of the evidence using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) system.40 The main question to be answered was: is the use of neuraxial anesthesia alone or when combined with general anesthesia associated with superior perioperative outcomes when compared to general anesthesia for major surgery of the trunk and lower extremity.
This systematic review was conducted in adherence to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) guidelines.41 Our health care question was defined a priori using the PICO (Population, Intervention, Comparator and Outcomes) format.
Inclusion criteria were RCTs with no lower limit in sample size and large (n > 500) observational, cohort, case-control, and cross-sectional studies. English language studies were included that involved adult (over 18-years-old), nonpregnant patients, published between January 1, 2010 and May 31, 2016, undergoing major truncal or lower extremity surgery, where combined neuraxial-general anesthesia and/or neuraxial anesthesia was compared to general anesthesia. Studies were included if they reported one or more of the following outcomes: 30-day mortality (primary outcome), cardiopulmonary morbidity (including pneumonia, myocardial infarction), surgical site infection, thromboembolic events, blood transfusion, and resource use (critical care services, length of hospital stay). For this review, neuraxial anesthesia, defined as spinal or epidural anesthesia, were differentiated from regional techniques (eg, peripheral nerve blockade). Previously published meta-analyses and strictly descriptive articles were excluded. Studies were excluded if they did not contain the population of interest; that is, did not compare combined neuraxial-general anesthesia or neuraxial anesthesia with general anesthesia; or did not report the outcomes of interest. Reasons for exclusion of full-text articles are outlined in Appendix 2 and Appendix 3.
Data Sources and Search Strategy
Two authors systematically searched bibliographic databases including the US National Library of Medicine database, MEDLINE, and the Excerpta Medica database, Ovid EMBASE, and retrieved relevant studies fulfilling the inclusion criteria, with the final search being conducted on May 31, 2016. The references of relevant publications were manually reviewed for additional material. Our core search consisted of MeSH terms related to anesthesia type (“anesthesia, epidural,” “anesthesia, spinal,” “anesthesia, general”) combined with specific MeSH terms relating to outcomes (“treatment outcome,” “treatment failure,” “mortality,” “outcome,” “postoperative complications,” “complication,” “morbidity”). Articles with the relevant MeSH terms within the title or abstract of the article were identified, with both “anaesthesia” and “anesthesia” being included in the search. A summary of the search strategy is available in Appendix 1.
The search yielded 1393 results. Bibliographies of identified relevant studies and previously published systematic reviews were reviewed to identify any additional studies that may have been missed in the primary literature search.
Titles and abstracts were assessed for eligibility by 2 investigators (YU and LS) working independently and in duplicate. Studies that fulfilled the eligibility criteria were selected and the full-text articles obtained and reviewed. Authors (CC, LS, MB) conducted a full-text review of all included articles. Any discrepancy in study selection between the reviewers was resolved by consensus agreement between all authors.
Data were manually extracted in duplicate from the full-text versions of eligible studies by authors LS and CC. Data extracted included the total number of events and controls for each outcome for neuraxial anesthesia alone, general anesthesia alone, and, where, applicable combined neuraxial and general anesthesia. Data presented in tabular format were used as the primary source for data extraction. Due to potential for overlap of patient records from the source databases and observational studies, publications were excluded from analysis on an outcome basis if there was an overlap in data collection time and surgical type for the same database. In cases of overlap, studies with larger sample sizes were included preferentially in the analysis if potential overlap in records was identified. Exclusion of studies on an outcome basis due to potential overlap of patient records is outlined in Appendix 4.
Risk of Bias
We utilized the GRADE system to summarize the quality of the evidence.40 The risk of bias was assessed according to limitations in study design inherent to study type.42 Study limitations in RCTs include allocation concealment, blinding, follow-up, selective reporting, and premature trial cessation for early benefit. Observational trials are limited by lack of randomization, failure to develop and apply appropriate eligibility criteria, flawed measurement of exposure or outcome, failure to adjust for prognostic imbalance, and incomplete follow-up. According to GRADE, RCTs and observational studies are considered high- and low-quality evidence, respectively. Each study was evaluated for limitations in study design and risk of bias was established across studies for each outcome.
Criteria for evaluating heterogeneity include similarity of point estimates, extent of overlap of confidence intervals (CIs), and the I2 statistic.43 We predefined a potential source of heterogeneity as whether neuraxial anesthesia was used alone or in combination with general anesthesia. Therefore, we created subgroup comparisons of combined neuraxial-general anesthesia and neuraxial anesthesia, comparing both subgroups with general anesthesia. According to GRADE,43 the level of evidence was downgraded for outcomes where large heterogeneity in results remained, after exploration of reasons that could explain inconsistency.
Direct evidence comes from studies that directly compare the interventions of interest in the target population.44 To ensure there was a high level of directness regarding the impact of neuraxial anesthesia, subgroup analysis was performed in 2 steps to answer the health care question: (1) Was the use of neuraxial anesthesia associated with no increase in adverse outcomes compared with the use of general anesthesia? (2) Was the addition of neuraxial anesthesia to general anesthesia beneficial compared to the sole use of general anesthesia? Due to the low number of randomized controlled trials available in this context, this separation was only feasible for observational trials.
The optimal information size (OIS), meaning the number of patients required for an adequately powered individual trial was calculated for each outcome and downgrading was performed if OIS was not met by outcome.45
Publication bias was assessed by utilizing the funnel plot and assessing whether the outcomes of effect were symmetrical around the best estimate of effect or pooled odds ratios (ORs).46
Evaluating the Quality of Evidence
The GRADE approach was used to upgrade the quality of evidence, for example, where there were large pooled effects and dose-response relations, and when all plausible confounders would decrease the estimate of effect.47
Data were entered into the Cochrane Review Manager software RevMan, version 5.3, The Cochrane Collaboration, Oxford, United Kingdom, 2015, by 1 author (CC) and checked by another (LS). In addition to dichotomous data, RevMan facilitates the pooling of continuous data such as length of hospital stay, when reported as mean and standard deviation. For studies that reported length of stay in median and interquartile ranges, the method utilized by Hozo et al48 was used to estimate mean and standard deviation as required for analysis in RevMan. Outcomes were expressed as ORs or mean differences. To account for multiple comparisons, the level of statistical significance was set at P < .01 and effect sizes presented with 99% CIs.
After screening the 1393 identified results, 51 full-text articles, comprising 32 observational studies and 19 RCTs, were assessed for eligibility. In total, 27 observational studies and 11 RCTs met the inclusion criteria, summarized in the PRISMA flow diagram (Figure 1).
A qualitative summary of observational studies and RCTs included in the meta-analysis are outlined in Tables 1 and 2, respectively. This review included studies that included a total of 1,084,099 patients, 1,082,965 from observational studies, and 1134 from RCTs. These included observational studies investigating hip fracture surgery (7 studies, 229,472 patients),14,17,21,22,25,28,29 total hip arthroplasty (3 studies, 41,842 patients),15,18,31 total hip or knee arthroplasty (4 studies, 415,341 patients),19,20,31,38 and total knee arthroplasty alone (3 studies, 46,294 patients).23,24,26 The remainder of observational studies investigated lower extremity amputation,27 elective surgical procedures (single study investigating procedures on vasculature, bladder, bowel, lung, gastrointestinal),30 high-risk patients having noncardiac surgery,49 multiple surgical types in patients with chronic obstructive pulmonary disease,33 open lung resection,32 open colectomies,34 percutaneous nephrolithotomy,35 lower extremity vascular surgeries,36,39 and endovascular aortic aneurysm repair.37 The RCTs investigated percutaneous nephrolithotomy,7,9,12 total knee arthroplasty,3 total hip arthroplasty,13 hip fracture surgery,5 abdominal hysterectomy,4 laparoscopic colonic resection,11 emergency laparotomy,10 gynecological surgery,8 and urogenital surgery.50 In total, 13 studies investigated truncal surgery (125,838 patients),4,6–12,32,34,35,37,50 22 studies investigated lower limb surgery (923,347 patients),3,5,13–24,26–29,31,36,38,39 and 3 studies investigated both truncal and lower limb surgery (34,914 patients).30,33,49 When categorized by anesthetic type, 14 studies investigated spinal anesthetic alone,3–5,8,11,13–15,17,23,25,27,35,49 5 studies investigated epidural anesthetic alone,6,9,10,39,50 and 19 studies investigated both spinal or neuraxial anesthesia, or did not specify which type of neuraxial anesthesia was used.7,16,18–21,24,26,28–34,36–38,51 The following number of participants (N [percentage]) were included in operative types: total knee or hip arthroplasty (503,597 [46.5]), hip fracture surgery (229,794 [21.2]), mixed types of hip surgery (182,307 [16.8]), truncal surgery (125,838 [11.6]), mixed surgical types (34,914 [3.2]), and nonorthopedic lower extremity surgery (7649 [0.70]). A summary of the databases used for the included studies is outlined in Appendix 4.
Risk of Bias Analysis
Due to the intervention being neuraxial anesthesia, none of the RCTs masked participants or personnel; however, we did not exclude or downgrade the quality of evidence on these grounds. Serious risk of bias in specific outcome analysis was identified in 9 studies as follows. One study was excluded because of a significantly different baseline risk in cardiac complications.49 Three studies were excluded because of in-hospital mortality instead of 30-day mortality were reported.16,22,28 Five studies reporting LOS were excluded because of significant variation in LOS thought to be related to the procedure itself (lower extremity amputation,27 transurethral procedure,52 inappropriate outcome measurement,16 method of reporting LOS,53 and large number of included surgical procedure types ).30
The outcomes of myocardial infarction (in the combined neuraxial and general anesthesia group), thromboembolic events (in the combined neuraxial and general anesthesia group), and the outcome LOS (in the RCT group) were downgraded by 1 level because of variation in point estimates of treatment effect.
Three studies were excluded for indirectness or noncompliance with the health care question. One study was excluded because the intervention was linked to surgical type,54 one used a surrogate marker of LOS (duration of postanesthesia care unit stay),55 and the third administered epidural opioid only at wound closure.56
The outcomes of myocardial infarction (combined neuraxial and general anesthesia versus general anesthesia), thromboembolic events (neuraxial versus general anesthesia), and blood transfusion (RCTs) were downgraded for imprecision.
The outcome cardiac composite (neuraxial versus general anesthesia) was downgraded for suspected publication bias.
Rating Up the Quality of Evidence
In our analysis of population-based studies, a large effect size was established for the outcomes of intensive care unit (ICU) admission (neuraxial versus general anesthesia) and pulmonary composite complications (neuraxial versus general anesthesia), and therefore level of evidence was upgraded by one level.
Overall, the level of evidence from available studies, regarding the health care question ranged from moderate to very low by outcomes. A summary of findings with GRADE analysis is outlined in Table 3.
Results From Meta-analysis by Outcome
Six observational studies20,29,30,32,49,53 compared combined neuraxial-general anesthesia with general anesthesia with a total of 396,869 patients. There was no difference in the odds of 30-day mortality (OR 0.88; 99% CI, 0.77–1.01). Thirteen observational studies17,18,20,21,23,25,27,29,36,37,39,49,51 compared neuraxial with general anesthesia with a total of 491,611 patients. There was no difference in the odds of 30-day mortality (OR 0.98; 99% CI, 0.92–1.04) (Figure 2). The quality of the evidence for 30-day mortality by GRADE analysis was determined as low. Subgroup analysis was performed for 30-day mortality: studies investigating hip fracture surgery comparing neuraxial with general anesthesia included 4 studies with a total of 90,037 patients.17,21,25,29 This subgroup analysis resulted in no difference in odds of 30-day mortality (OR 1.00; 99% CI, 0.93–1.07). Further subgroup analysis of studies investigating hip and/or knee arthroplasty that compared neuraxial with general anesthesia included 5 studies with a total of 389,355 patients.18,20,23,31,51 This subgroup analysis resulted in no difference in odds of 30-day mortality (OR 0.72; 99% CI, 0.51–1.01).
Four observational studies20,30,32,34 compared combined neuraxial-general anesthesia with general anesthesia with a total of 471,812 patients. There was an increase in the odds of myocardial infarction (OR 1.18; 99% CI, 1.01–1.37), that is general anesthesia was protective. The quality of this evidence by GRADE analysis was rated as very low. Nine observational studies14,16,17,20,22,23,25,36,37 compared neuraxial with general anesthesia with a total of 520,052 patients. This analysis resulted in no difference in the odds of myocardial infarction (OR 0.91; 99% CI, 0.81–1.02) (see Supplemental Digital Content 1, Figure 1, https://links.lww.com/AA/B726). The quality of this evidence by GRADE analysis was rated as low.
Cardiac Composite Complications
Four observational studies19,20,24,32 compared combined neuraxial-general anesthesia with general anesthesia with a total of 361,143 patients. There was no difference in the odds of cardiac composite complication (OR 1.04; 99% CI, 1.00–1.09). The quality of this evidence by GRADE analysis was rated as low. Eleven observational studies17,18,20,22–25,36,37,39,51 compared neuraxial with general anesthesia with a total of 447,748 patients. There was no significant difference in the odds of cardiac composite complications (OR 0.99; 99% CI, 0.94–1.03) (see Supplemental Digital Content 2, Figure 2, https://links.lww.com/AA/B727). The quality of this evidence by GRADE analysis was rated as very low.
Pulmonary Complications Including Mechanical Ventilation
Three observational studies20,32,34 compared combined neuraxial-general anesthesia with general anesthesia with a total of 459,433 patients. The odds of pulmonary complications (OR 0.84; 99% CI, 0.79–0.88) was reduced in favor of combined general and neuraxial anesthesia. The quality of this evidence by GRADE analysis was rated as low. Nine observational studies16–18,20,22,23,36,37,39 compared neuraxial with general anesthesia with a total of 498,229 patients. The odds of pulmonary complications (OR 0.38; 99% CI, 0.36–0.40) was reduced in favor of neuraxial anesthesia (Figure 3). The quality of this evidence by GRADE analysis was rated as moderate.
Four observational studies20,30,32,34 compared combined neuraxial-general anesthesia with general anesthesia with a total of 471,812 patients. There was no difference in the odds of developing pneumonia (OR 0.94; 99% CI, 0.87–1.02). The quality of this evidence by GRADE analysis was rated as low. Eight observational studies14,17,20,22,25,33,36,37 compared neuraxial with general anesthesia with a total of 396,106 patients. There was no difference in the odds of developing pneumonia (OR 0.92; 95% CI, 0.84–1.01) (Figure 4). The quality of this evidence by GRADE analysis was rated as low.
Surgical Site Infection
Three observational studies20,32,34 compared combined neuraxial-general anesthesia with general anesthesia with a total of 459,433 patients. There was a reduced odds of surgical site infection (OR 0.93; 99% CI, 0.88–0.98), in favor of combined general and neuraxial anesthesia. The quality of this evidence by GRADE analysis was rated as low. Seven observational studies17,18,20,26,36–38 compared neuraxial with general anesthesia with a total of 380,682 patients. There was a reduced odds of surgical site infection (OR 0.76; 99% CI, 0.71–0.82) in favor of neuraxial anesthesia (see Supplemental Digital Content 3, Figure 3, https://links.lww.com/AA/B728). The quality of this evidence by GRADE analysis was rated as low.
Requirement for Blood Transfusion
Three observational studies20,32,34 compared combined neuraxial-general anesthesia with general anesthesia with a total of 459,433 patients. There was a reduced odds of requiring blood transfusion (OR 0.90; 99% CI, 0.87–0.93), in favor of combined general and neuraxial anesthesia. The quality of this evidence by GRADE analysis was rated as low. Six observational studies14,17,20,23,35,37 compared neuraxial anesthesia with general anesthesia with a total of 369,653 patients. There was a reduced odds of requiring blood transfusion (OR 0.85; 99% CI, 0.82–0.88), in favor of neuraxial anesthesia (see Supplemental Digital Content 4, Figure 4, https://links.lww.com/AA/B729). The quality of this evidence by GRADE analysis was rated as low. Five RCTs5,7,9,11,12 compared neuraxial anesthesia with general anesthesia with a total of 585 patients. There was no difference in the relative risk of blood transfusion requirement (RR 1.05; 99% CI, 0.65–1.71) (see Supplemental Digital Content 5, Figure 5, https://links.lww.com/AA/B730). The quality of this evidence by GRADE analysis was rated as moderate.
Three observational studies20,32,34 compared combined neuraxial-general anesthesia with general anesthesia with a total of 459,433 patients. There was a reduced odds of thromboembolic events (OR 0.84; 99% CI, 0.73–0.98), in favor of combined general and neuraxial anesthesia. The quality of this evidence by GRADE analysis was rated as low. Seven observational studies14,17,20,23,25,36,37 compared neuraxial with general anesthesia with a total of 397,806 patients. This analysis resulted in a reduction in the odds of thromboembolic events (OR 0.79; 99% CI, 0.68–0.91) (see Supplemental Digital Content 6, Figure 6, https://links.lww.com/AA/B731), in favor of neuraxial anesthesia. The quality of this evidence by GRADE analysis was rated as very low.
Length of Stay
Three observational studies19,32,34 compared combined neuraxial-general anesthesia with general anesthesia with a total of 142,998 patients. There was a reduced length of stay (mean difference −0.16 days; 99% CI, −0.17 to −0.15), in favor of combined general and neuraxial anesthesia. The quality of this evidence by GRADE analysis was rated as low. Nine observational studies18,19,21,23,25,31,36,37,39 compared neuraxial with general anesthesia with a total of 120,755 patients. This analysis resulted in a reduced length of stay (mean difference −0.29 days; 99% CI, −0.29 to −0.28), in favor of neuraxial anesthesia. The quality of this evidence by GRADE analysis was rated as low. Eleven RCTs compared neuraxial anesthesia with general anesthesia,4–13,50 with a total of 1131 patients. This analysis resulted in a reduced length of stay (mean difference −0.15; 99% CI, −0.27 to −0.04) (see Supplemental Digital Content 7, Figure 7, https://links.lww.com/AA/B732), in favor of neuraxial anesthesia. The quality of this evidence by GRADE analysis was rated as moderate.
Three observational studies20,32,34 compared combined neuraxial-general anesthesia with general anesthesia with a total of 459,433 patients. There was a reduced odds for ICU admission (OR 0.77; 99% CI, 0.73–0.81), in favor of combined general and neuraxial anesthesia. The quality of this evidence by GRADE analysis was rated as moderate. Four observational studies16,20,25,51 compared neuraxial versus general anesthesia with a total of 474,156 patients. There was a reduced odds for ICU admission (OR 0.50; 99% CI, 0.48–0.53) (Figure 5), in favor of neuraxial anesthesia. The quality of this evidence by GRADE analysis was rated as moderate.
This meta-analysis of studies reflective of contemporary clinical care indicates that, compared to general anesthesia, neuraxial anesthesia when used alone or when combined with general anesthesia is not likely to decrease 30-day mortality, our primary outcome. This analysis indicates that general anesthesia may protect against myocardial infarction compared to neuraxial anesthesia. Although neuraxial anesthesia was not associated with a reduced risk of pneumonia when compared to general anesthesia, neuraxial anesthesia alone or when combined with general anesthesia was associated with a reduced risk of pulmonary complications and requirement for mechanical ventilation (60% risk reduction, upper CI 0.40, Figure 3) and requirement for ICU admission (19% risk reduction, upper CI 0.81, Figure 5). Neuraxial anesthesia was also associated with reduced risk of surgical site infection and thromboembolic events, albeit with smaller effect sizes. Neuraxial anesthesia was associated with a reduced length of stay in our analysis of both RCTs and observational studies. Our analysis comparing neuraxial and general anesthesia yielded conflicting results for risk of blood transfusion. The RCTs indicated no difference in the relative risk, whereas the observational studies indicated that neuraxial anesthesia was associated with a reduced risk of blood transfusion. This apparent disconnect between results for the observational and RCTs for this outcome may indeed be consistent with confounding bias present in observational trials. However, the difference could also be accounted for by the poor quality of the included RCTs, which even when pooled did not meet the OIS required to adequately power the result for this outcome. Because observational studies with their inherent risks of bias are included in this analysis, these results need to be interpreted with caution. Furthermore, the reductions in OR for many outcomes were small and the GRADE analysis evaluated the data as being low quality.
The increased risk of myocardial infarction in patients receiving combined neuraxial-general anesthesia compared to general anesthesia deserves comment. A plausible mechanism for this association is the potential for sympathetic blockade and resultant hypotension associated with the combined anesthetic technique; however, due to the observational nature of the dataset, no causal mechanism for this result can be established.
Previous systematic reviews and meta-analyses of RCTs have compared neuraxial with general anesthesia for morbidity and mortality outcomes, and include older studies which are no longer representative of current clinical practice.1,2,57,58 The purpose of this review was to provide a review and analysis of studies which are likely to be representative of outcomes associated with contemporary anesthetic and surgical techniques.
To the authors’ knowledge, this systematic review and meta-analysis is the only analysis which includes both contemporary large-scale observational studies and RCTs. Therefore the results of this review are more reflective of current practice compared to studies published in previous decades. Large-scale population studies provide clinicians with an opportunity to pool the results of an entire population undergoing a procedure or surgery and gain unique insights into the quality of clinical care. Outcomes reported from large-scale observational studies are likely representative of actual current clinical practice, because they do not use rigid filters in the form of inclusion and exclusion criteria before sampling. Moreover, for this review there was insufficient information available from RCTs for many of our predefined outcomes of interest, and as a result, only 0.1% of patient data utilized in this analysis was from RCTs. The RCTs identified during our search were, even when pooled for meta-analysis, inadequately powered to assess our outcomes of interest, with the overall quality of the individual RCTs being low. The large sample sizes used in observational studies allow reporting of infrequently occurring yet important outcomes and thus have the ability to detect a difference between groups.
However, there are clearly limitations to including observational studies in a systematic review and meta-analysis. Due to the lack of randomization, observational studies bear the risk of bias relating to residual confounding of known or unknown factors that could impact on the primary outcome. One purpose of randomization is to increase the likelihood that the experimental and the control group have similar baseline risks for the primary outcome. However, a recent study identified an association between both age and American Society of Anesthesiologists (ASA) status and choice of anesthesia for total knee arthroplasty.59 Patients who were in older age and higher ASA categories were more likely to receive neuraxial anesthesia compared to general anesthesia. Therefore, for some surgical types, patients receiving neuraxial anesthesia may be at a higher baseline risk compared to those receiving general anesthesia.
Given the high risk of bias inherent in observational studies, meticulous analysis of risk of bias is required to determine the validity of any effect demonstrated by our meta-analysis. We utilized the methodological approach of GRADE because it provides a systematic process of presenting evidence summaries and the quality of the body of evidence. Notably, in GRADE the quality of a body of evidence is evaluated for each individual outcome across studies.42 In contrast, many systematic reviews rate the quality of evidence across outcomes, rather than separately for each outcome, which limits the usefulness for guideline development.42 Regardless of the methods used to assess the quality of the evidence, it is likely to be subjective.
Heterogeneity between studies can be measured using the I2 statistic, which assesses the percentage of variability in treatment estimates; however, sole reliance on a single measure of heterogeneity can be misleading. Heterogeneity can be evaluated based on the similarity of point estimates, extent of overlap of confidence intervals, and statistical tests.43 For example, in our analysis of pulmonary complications, including mechanical ventilation, the I2 statistic for the neuraxial versus general anesthesia group was 99% (Figure 3), which if interpreted alone would indicate very high heterogeneity between included studies. However, as shown by Rucker,60 the heterogeneity statistic I2 greatly depends on precision, meaning the size of the studies included. Therefore, as precision increases, estimates of I2 increase toward 100%. We observed this limitation of I2 in our analysis of large sample size studies. As such, for the outcome of pulmonary complications, the point estimates show largely a consistent effect favoring neuraxial anesthesia and large overlap of confidence intervals, which despite the apparently large heterogeneity when assessed by the I2 statistic, should not reduce the confidence in the result favoring neuraxial anesthesia, given the consistency of the point estimates.
The best quality of outcome variables is attained when included studies investigate the outcome of interest as their primary outcome. For the included RCTs, the outcome of length of stay was the only outcome that was listed as a primary outcome of interest for a number of individual studies. The remainder of the outcomes of interest investigated was secondary outcomes, and, as such, this has been identified as a limitation within our analysis. For the included observational studies, the majority investigated our outcomes of interest as primary outcomes.
This review provides a summary of the pool of knowledge obtained from contemporary observational and RCTs that compare neuraxial with general anesthesia. It is relevant to anesthesiologists who care for patients undergoing major truncal or lower extremity surgery. Because this review included nonrandomized trials and the quality of the evidence ranged between very low and moderate, these current results should be interpreted with caution. No causal relationship can be established; however, this analysis adds to the existing body of evidence that suggest that neuraxial anesthesia may improve perioperative outcomes for major surgery. Large multicenter RCTs are required to more definitively investigate the potential impact that neuraxial anesthesia has on perioperative outcomes.
Compared to general anesthesia, neuraxial anesthesia when used alone or when combined with general anesthesia was not associated with decreased 30-day mortality. General anesthesia may protect against myocardial infarction compared to neuraxial anesthesia. Neuraxial anesthesia when used alone or when combined with general anesthesia may improve important perioperative outcomes when compared with general anesthesia. Neuraxial anesthesia (alone) was associated with a significantly reduced incidence of pulmonary complications including the requirement for mechanical ventilation. Neuraxial anesthesia was associated with superior treatment effects for nonpulmonary outcomes; however, these effects were reduced in magnitude. However, due to the inclusion of observational studies and database records within this analysis, results should be interpreted with caution.
Name: Lauren Smith, MD.
Contribution: This author helped design the study, collect data, and prepare the manuscript.
Name: Crispiana Cozowicz, MD.
Contribution: This author helped design the study, collect data, analyze the data, and contribute to the manuscript.
Name: Yoshiaki Uda, MBBS, FANZCA.
Contribution: This author helped design the study, collect data, and contribute to the manuscript.
Name: Stavros G. Memtsoudis, MD, PhD.
Contribution: This author helped design the study, analyze the data, and contribute to the manuscript.
Name: Michael J. Barrington, MBBS, PhD.
Contribution: This author helped design the study, contribute to the manuscript, and approve the final manuscript.
This manuscript was handled by: Richard C. Prielipp, MD.
1. Guay J, Choi PT, Suresh S, Albert N, Kopp S, Pace NLNeuraxial anesthesia for the prevention of postoperative mortality and major morbidity: an overview of cochrane systematic reviews. Anesth Analg. 2014;119:716–725.
2. Johnson RL, Kopp SL, Burkle CM, et al.Neuraxial vs general anaesthesia for total hip and total knee arthroplasty: a systematic review of comparative-effectiveness research. Br J Anaesth. 2016;116:163–176.
3. Harsten AA, Kehlet HA, Toksvig-Larsen SA, et al.Recovery after total intravenous general anaesthesia or spinal anaesthesia for total knee arthroplasty: a randomized trial. Br J Anaesth. 2013: 391.
4. Kjølhede P, Langström P, Nilsson P, Wodlin NB, Nilsson LThe impact of quality of sleep on recovery from fast-track abdominal hysterectomy. J Clin Sleep Med. 2012;8:395–402.
5. Parker MJ, Griffiths RGeneral versus regional anaesthesia for hip fractures. A pilot randomised controlled trial of 322 patients. Injury 2015;46:1562–1566.
6. Salman N, Durukan AB, Gurbuz HA, et al.Comparison of effects of epidural bupivacaine and intravenous meperidine analgesia on patient recovery following elective abdominal aortic surgery. Med Sci Monit. 2013;19:347–352.
7. Singh V, Sinha RJ, Sankhwar SN, Malik AA prospective randomized study comparing percutaneous nephrolithotomy under combined spinal-epidural anesthesia with percutaneous nephrolithotomy under general anesthesia. Urol Int. 2011;87:293–298.
8. Surico D, Mencaglia L, Riboni F, Vigone A, Leo L, Surico NMinilaparotomy in spinal anaesthesia: a surgical choice in treatment of benign gynaecologic disease. Arch Gynecol Obstet. 2010;281:461–465.
9. Tangpaitoon T, Nisoog C, Lojanapiwat BEfficacy and safety of percutaneous nephrolithotomy (PCNL): a prospective and randomized study comparing regional epidural anesthesia with general anesthesia. Int Braz J Urol. 2012;38:504–511.
10. Tyagi A, Seelan S, Sethi AK, Mohta MRole of thoracic epidural block in improving post-operative outcome for septic patients: a preliminary report. Eur J Anaesthesiol. 2011;28:291–297.
11. Wongyingsinn M, Baldini G, Stein B, Charlebois P, Liberman S, Carli FSpinal analgesia for laparoscopic colonic resection using an enhanced recovery after surgery programme: better analgesia, but no benefits on postoperative recovery: a randomized controlled trial. Br J Anaesth. 2012;108:850–856.
12. Nouralizadeh A, Ziaee SA, Hosseini Sharifi SH, et al.Comparison of percutaneous nephrolithotomy under spinal versus general anesthesia: a randomized clinical trial. J Endourol.2013;27:974–978.
13. Harsten A, Kehlet H, Ljung P, Toksvig-Larsen STotal intravenous general anaesthesia vs. spinal anaesthesia for total hip arthroplasty: a randomised, controlled trial. Acta Anaesthesiol Scand. 2015;59:298–309.
14. Basques BA, Bohl DD, Golinvaux NS, Samuel AM, Grauer JGGeneral versus spinal anaesthesia for patients aged 70 years and older with a fracture of the hip. Bone Joint J. 2015;97-B:689–695.
15. Basques BA, Toy JO, Bohl DD, Golinvaux NS, Grauer JNGeneral compared with spinal anesthesia for total hip arthroplasty. J Bone Joint Surg Am. 2015;97:455–461.
16. Chu CC, Weng SF, Chen KT, et al.Propensity score-matched comparison of postoperative adverse outcomes between geriatric patients given a general or a neuraxial anesthetic for hip surgery: a population-based study. Anesthesiology2015;123:136–147.
17. Fields AC, Dieterich JD, Buterbaugh K, Moucha CSShort-term complications in hip fracture surgery using spinal versus general anaesthesia. Injury 2015;46:719–723.
18. Helwani MA, Avidan MS, Ben Abdallah A, et al.Effects of regional versus general anesthesia on outcomes after total hip arthroplasty: a retrospective propensity-matched cohort study. J Bone Joint Surg Am. 2015;97:186–193.
19. Memtsoudis SG, Stundner O, Rasul R, et al.Sleep apnea and total joint arthroplasty under various types of anesthesia: a population-based study of perioperative outcomes. Reg Anesth Pain Med. 2013;38:274–281.
20. Memtsoudis SG, Sun X, Chiu YL, et al.Perioperative comparative effectiveness of anesthetic technique in orthopedic patients. Anesthesiology2013;118:1046–1058.
21. Neuman MD, Rosenbaum PR, Ludwig JM, Zubizarreta JR, Silber JHAnesthesia technique, mortality, and length of stay after hip fracture surgery. JAMA 2014;311:2508–2517.
22. Neuman MD, Silber JH, Elkassabany NM, Ludwig JM, Fleisher LAComparative effectiveness of regional versus general anesthesia for hip fracture surgery in adults. Anesthesiology 2012;117:72–92.
23. Pugely AJ, Martin CT, Gao Y, Mendoza-Lattes S, Callaghan JJDifferences in short-term complications between spinal and general anesthesia for primary total knee arthroplasty. J Bone Joint Surg Am. 2013;95:193–199.
24. Stundner O, Chiu YL, Sun X, et al.Comparative perioperative outcomes associated with neuraxial versus general anesthesia for simultaneous bilateral total knee arthroplasty. Reg Anesth Pain Med. 2012;37:638–644.
25. Seitz DP, Gill SS, Bell CM, et al.Postoperative medical complications associated with anesthesia in older adults with dementia. J Am Geriatr Soc. 2014;62:2102–2109.
26. Liu J, Ma C, Elkassabany N, Fleisher LA, Neuman MDNeuraxial anesthesia decreases postoperative systemic infection risk compared with general anesthesia in knee arthroplasty. Anesth Analg. 2013;117:1010–1016.
27. Khan SA, Qianyi RL, Liu C, Ng EL, Fook-Chong S, Tan MGEffect of anaesthetic technique on mortality following major lower extremity amputation: a propensity score-matched observational study. Anaesthesia 2013;68:612–620.
28. Patorno E, Neuman MD, Schneeweiss S, Mogun H, Bateman BTComparative safety of anesthetic type for hip fracture surgery in adults: retrospective cohort study. BMJ 2014;348:g4022.
29. White SM, Moppett IK, Griffiths ROutcome by mode of anaesthesia for hip fracture surgery. An observational audit of 65 535 patients in a national dataset. Anaesthesia 2014;69:224–230.
30. Nash DM, Mustafa RA, McArthur E, et al.Combined general and neuraxial anesthesia versus general anesthesia: a population-based cohort study. Can J Anesth. 2015;62:356–368.
31. Chen WH, Hung KC, Tan PH, Shi HYNeuraxial anesthesia improves long-term survival after total joint replacement: a retrospective nationwide population-based study in Taiwan. Can J Anesth. 2015;62:369–376.
32. Özbek U, Poeran J, Mazumdar M, Memtsoudis SGPatient safety and comparative effectiveness of anesthetic technique in open lung resections. Chest 2015;148:722–730.
33. Hausman MS Jr, Jewell ES, Engoren MRegional versus general anesthesia in surgical patients with chronic obstructive pulmonary disease: does avoiding general anesthesia reduce the risk of postoperative complications? Anesth Analg. 2015;120:1405–1412.
34. Poeran J, Yeo H, Rasul R, Opperer M, Memtsoudis SG, Mazumdar MAnesthesia type and perioperative outcome: open colectomies in the United States. J Surg Res. 2015;193:684–692.
35. Cicek T, Gonulalan U, Dogan R, et al.Spinal anesthesia is an efficient and safe anesthetic method for percutaneous nephrolithotomy. Urology2014;83:50–55.
36. Ghanami RJ, Hurie J, Andrews JS, et al.Anesthesia-based evaluation of outcomes of lower-extremity vascular bypass procedures. Ann Vasc Surg. 2013;27:199–207.
37. Edwards MS, Andrews JS, Edwards AF, et al.Results of endovascular aortic aneurysm repair with general, regional, and local/monitored anesthesia care in the American College of Surgeons National Surgical Quality Improvement Program database. J Vasc Surg. 2011;54:1273–1282.
38. Chang CC, Lin HC, Lin HW, Lin HCAnesthetic management and surgical site infections in total hip or knee replacement: a population-based study. Anesthesiology 2010;113:279–284.
39. Wiis JT, Jensen-Gadegaard P, Altintas Ü, Seidelin C, Martusevicius R, Mantoni TOne-week postoperative patency of lower extremity in situ bypass graft comparing epidural and general anesthesia: retrospective study of 822 patients. Ann Vasc Surg. 2014;28:295–300.
40. Guyatt G, Oxman AD, Akl EA, et al.GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin Epidemiol. 2011;64:383–394.
41. Moher D, Liberati A, Tetzlaff J, Altman DGPRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg. 2010;8:336–341.
42. Guyatt GH, Oxman AD, Vist G, et al.GRADE guidelines: 4. Rating the quality of evidence–study limitations (risk of bias). J Clin Epidemiol. 2011;64:407–415.
43. Guyatt GH, Oxman AD, Kunz RGRADE Working Group. GRADE guidelines: 7. Rating the quality of evidence–inconsistency. J Clin Epidemiol. 2011;64:1294–1302.
44. Guyatt GH, Oxman AD, Kunz R, et al.GRADE guidelines: 8. Rating the quality of evidence–indirectness. J Clin Epidemiol. 2011;64:1303–1310.
45. Guyatt GH, Oxman AD, Kunz R, et al.GRADE guidelines 6. Rating the quality of evidence–imprecision. J Clin Epidemiol. 2011;64:1283–1293.
46. Guyatt GH, Oxman AD, Montori V, et al.GRADE guidelines: 5. Rating the quality of evidence–publication bias. J Clin Epidemiol. 2011;64:1277–82.
47. Guyatt GH, Oxman AD, Sultan S, et alGRADE Working Group. GRADE guidelines: 9. Rating up the quality of evidence. J Clin Epidemiol.2011;64:1311–1316.
48. Hozo SP, Djulbegovic B, Hozo IEstimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol. 2005;5:13.
49. Leslie K, Myles P, Devereaux P, et al.Neuraxial block, death and serious cardiovascular morbidity in the POISE trial. Br J Anaesth. 2013;111:382–390.
50. Orsolya M, Attila-Zoltan M, Gherman V, et al.The effect of anaesthetic management on neutrophil gelatinase associated lipocalin (NGAL) levels after robotic surgical oncology. J BUON2015;20:317–324.
51. Gabriel RA, Kaye AD, Jones MR, Dutton RP, Urman RDPractice variations in anesthetic care and its effect on clinical outcomes for primary total hip arthroplasties. J Arthroplasty 2016;31:918–922.
52. Ornek D, Metin S, Deren S, et al.The influence of various anesthesia techniques on postoperative recovery and discharge criteria among geriatric patients. Clinics (Sao Paulo)2010;65:941–946.
53. Wax DB, Garcia C, Campbell N, Marin ML, Neustein SAnesthetic experience with endovascular aortic aneurysm repair. Vasc Endovascular Surg. 2010;44:279–281.
54. Pompeo E, Rogliani P, Tacconi F, et al.Randomized comparison of awake nonresectional versus nonawake resectional lung volume reduction surgery. J Thorac Cardiovasc Surg. 2012;143:47–54, e1.
55. Ross SB, Mangar D, Karlnoski R, et al.Laparo-endoscopic single-site (LESS) cholecystectomy with epidural vs. general anesthesia. Surg Endosc. 2013;27:1810–1819.
56. Guilfoyle MR, Mannion RJ, Mitchell P, Thomson SEpidural fentanyl for postoperative analgesia after lumbar canal decompression: a randomized controlled trial. Spine J.2012;12:646–651.
57. Rodgers A, Walker N, Schug S, et al.Reduction of postoperative mortality and morbidity with epidural or spinal anaesthesia: results from overview of randomised trials. BMJ.2000;321:1493.
58. Mauermann WJ, Shilling AM, Zuo ZA comparison of neuraxial block versus general anesthesia for elective total hip replacement: a meta-analysis. Anesth Analg. 2006;103:1018–1025.
59. Fleischut PM, Eskreis-Winkler JM, Gaber-Baylis LK, et al.Variability in anesthetic care for total knee arthroplasty: an analysis from the anesthesia quality institute. Am J Med Qual. 2015;30:172–179.
60. Rücker G, Schwarzer G, Carpenter JR, Schumacher MUndue reliance on I(2) in assessing heterogeneity may mislead. BMC Med Res Methodol. 2008;8:79.