Secondary Logo

Journal Logo

Systematic Review

Effects of regional anaesthesia on mortality in patients undergoing lower extremity amputation: A retrospective pooled analysis

Quak, Su M.,; Pillay, Nanthini1; Wong, Suei N.2; Karthekeyan, Ranjith B.3; Chan, Diana X.H.4,5; Liu, Christopher W. Y.4

Author Information
Indian Journal of Anaesthesia: June 2022 - Volume 66 - Issue 6 - p 419-430
doi: 10.4103/ija.ija_917_21
  • Open



The leading indication for lower extremity amputation (LEA) in the United States is peripheral vascular disease, primarily in the older person with diabetes mellitus complicated by peripheral neuropathy, ulcers, gangrene or osteomyelitis.[1] Generally, these patients are at a high-risk for poor postoperative outcomes because of their old age, sepsis and multiple medical comorbidities (such as ischaemic heart disease, renal impairment and stroke).[2] Consequently, LEA has been associated with 30-day postoperative mortality rates that are as high as 7-32%.[34]

In theory, regional anaesthesia (RA) may confer some advantages over general anaesthesia (GA). For example, RA may result in less hypotension, bleeding, venous thromboembolism, surgical stress and pulmonary complications.[5] For this reason, in the recent years, a number of studies have examined whether RA is associated with lower 30-day postoperative mortality when compared to GA for patients undergoing LEA.

This pooled analysis was performed to determine if RA reduces the 30-day mortality of patients undergoing LEA surgery, when compared to GA.


This review followed the recommendations outlined in the Cochrane handbook for systematic reviews and was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines.[67] The protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO) (registration number CRD42021231265).

Search strategy

An electronic search was conducted using the following databases from January 2010 to August 2021: PubMed, Embase, Scopus and Cochrane Central Register of Controlled Trials. No language restrictions were applied. All publications including conference abstracts were included in the initial search. The search used key terms such as “anesthesia”, “lower extremity” and “amputation”. The details of the search strategy are available in the Appendix 1. The reference lists of all included studies were manually searched to identify other studies which were eligible for inclusion.


Two reviewers (NP and SQ) independently reviewed the titles and abstracts of all search entries to exclude irrelevant studies. The full text of the remaining studies was further examined for inclusion based on the inclusion and exclusion criteria. Disagreements about the study eligibility were arbitrated by another author (CL).

Selection criteria

The inclusion criteria were studies involving:

  • Population: adult patients (>18 years old) undergoing non-traumatic LEA
  • Intervention: RA (central neuraxial and/or peripheral nerve block)
  • Comparator: GA
  • Outcome (primary): 30-day mortality
  • Studies that involved any ongoing trials, children, animals, combined GA/RA technique and LEA for trauma/malignancy were excluded.

Quality assessment

Two reviewers (NP, SQ) independently assessed the quality of each included study. The Newcastle-Ottawa Scale (NOS),[8] an instrument for assessing the quality of non-randomised studies, was used to assess for bias. This scale contains eight items within three domains (selection, comparability and outcomes). NOS scores of 7-9, 4-6 and 0-3 indicate that the studies are of high quality, high risk of bias and very high risk of bias, respectively.[8] The level of evidence was assessed in accordance to the Grading of Recommendations, Assessment, Development and Evaluations (GRADE) system.[9] As all of the included studies were retrospective in nature, the evidence for each outcome was initially graded as low. The grade was then upgraded or downgraded based on the risk of bias, imprecision, inconsistency, indirectness and publication bias. Any disagreement was resolved through consultation with another author (CL). The NOS and GRADE tables may be found in the Appendix 2 and 3.


Data extraction

Two reviewers (NP, SQ) independently extracted data into a standardised template created using Microsoft Excel version 2016 (Redmond: Microsoft Corporation, 2015). The following data was collected from each study: author names, publication year, study design, sample size, characteristics of study participants, type of intervention, 30-day mortality, 90-day mortality, intensive care unit (ICU) admission, length of stay, cardiovascular outcomes, respiratory outcomes, stroke, renal failure, vasopressor use, wound complications, bleeding and blood transfusion. Any disagreements were resolved through consultation with another reviewer (CL).

Data analysis

Statistical analyses were performed using the Review Manager (RevMan) software version 5.3. (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) and R Studio software Version 1.2.1335 (Boston: R Studio Inc, 2018). Only studies that utilised propensity-score matched outcomes were pooled. Continuous data were compared using mean differences and 95% confidence intervals (CIs). If studies reported the variables in median and interquartile range, the Box-Cox method was used to estimate the sample mean and standard variance.[10] Dichotomous data were pooled and analysed using the Mantel-Haenszel odds ratio with 95% CIs. Random effects model was applied to all pooled data. The Higgin’s I2 test was used to estimate the degree of statistical heterogeneity. I2 of <40%, 30-50%, 50-75% and 75-100% were considered to represent non-important, moderate, substantial and considerable heterogeneity, respectively.[6] Funnel plots for the primary outcome were constructed and visually inspected to evaluate for the risk of publication bias.


Study selection and characteristics

A total of 1082 studies were identified during the initial search. After the removal of 435 duplicate studies, 649 titles and abstracts were screened. Thirty-two studies were identified for full-text review, of which 19 studies were excluded as they did not fulfil the inclusion criteria. Another three studies were excluded because their study population were derived from the same National Surgical Quality Improvement Program database. To avoid double-counting, the study with the larger sample size was used. Ten studies were included in the review [Figure 1].

Figure 1:
PRISMA flow diagram. NSQIP: National Surgical Quality Improvement Program; GA: General anaesthesia; RA: Regional anaesthesia; vs: versus; n: Number

All of the studies that were included in this review were retrospective observational studies that examined the impact of anaesthetic technique on 30-day mortality and perioperative outcomes in adult patients undergoing LEA. These studies were conducted between 2013 and 2018. Eight studies were restricted to patients undergoing major LEA, while two studies involved patients undergoing major and minor LEA. Six studies compared GA with either peripheral nerve blocks or central neuraxial anaesthesia. Three studies compared GA with central neuraxial anaesthesia alone, and one study compared GA with peripheral nerve blocks alone. Eight out of 10 studies were assessed to be of high quality on the NOS [Table 1].

Table 1:
Characteristics of studies

1.30-day Mortality after LEA

The overall incidence (GA and RA groups) of 30-day mortality following LEA ranged from 3.7% to 15%.[1112] Out of the 10 studies, two did not control for confounders.[1314] The rest of the studies utilised either multivariate logistic regression[12151617] or propensity score matching[4111819] to adjust for confounders. With the exception of one study which was reported by SA Khan et al.,[4] all of the eight studies that controlled for confounders reported no significant benefit of RA over GA in reducing the incidence of 30-day mortality after LEA [Table 2].

Table 2:
Primary outcome (30-day mortality)

A meta-analysis was performed using the results of the four studies[4111819] which utilised propensity score matching. A funnel plot constructed to investigate publication bias showed a symmetrical distribution [Figure 2] and no important heterogeneity was observed among the studies. Based on these four studies which involved 6647 patients (of whom, 1933 received RA), the use of RA did not result in a significant reduction in incidence of 30-day mortality following LEA (odds ratio (OR) 0.83, 95% CI: 0.65, 1.05, I2 20%, P = 0.12) [Figure 3]. This evidence was graded as low.

Figure 2:
Funnel plot – 30-day mortality rate, regional anaesthesia versus general anaesthesia (primary outcome)
Figure 3:
Forest plot –30-day mortality rate, RA versus GA (primary outcome). CI: Confidence interval; RA:Regional anaesthesia; GA:General anaesthesia

It was postulated that the impact of RA may be greater in patients undergoing major LEA (defined as amputation above the ankle joint[20]) as major LEA is associated with higher mortality rates when compared to minor LEA.[21] As such, a subgroup analysis was performed with studies that looked only at patients undergoing major LEA. Two studies[418] involving 2570 patients (of whom, 1015 received RA) were included in this subgroup meta-analysis which found that RA significantly reduced 30-day mortality compared to GA (OR 0.73; 95% CI (0.57, 0.94), I2 0%, P = 0.01). See Figure 3 (forest plot).

2.90-day Mortality after LEA

90-day mortality was reported in two propensity-matched studies.[411] Individually, neither study found a significant difference in the 90-day mortality with the use of either RA or GA. A meta-analysis including the two studies also showed no difference in the 90-day mortality with the use of RA (OR 0.78; 95% CI (0.57,1.07), I2 0%, P = 0.13) [Figure 4]. This evidence was graded as low. A summary of all relevant secondary outcomes, including 90-day mortality, may be found in Table 3.

Figure 4:
Forest plot – secondary outcome measures, regional anaesthesia vs general anaesthesia; (a) 90-day mortality, (b) Stroke, (c) Renal failure. vs: versus
Table 3:
Secondary outcomes

3.Hospital length of stay

Six studies[41213141517] reported the length of hospital stay (H-LOS). There is some evidence that H-LOS is shorter in patients who received RA compared to GA. Three studies[131415] reported a statistically significant shorter H-LOS with the use of RA with mean differences ranging between 3.7 to 5 days.[1315]

4.Intensive care unit admission

There is limited evidence that RA is associated with reduced ICU admission. The admission rate to ICU following LEA was about 8%.[14] Two studies reported ICU admission as an outcome measure. Both these studies reported a significantly lower ICU admission (by 52.1%[11] and 84%[13]) in the RA group compared to the GA group. Among the patients who were admitted to the ICU, RA and GA were associated with similar durations of ICU stay.[1114]

5.Cardiac outcomes

Cardiac outcomes appear to be similar between the RA and GA groups. Six[111314151622] out of seven[11121314151622] studies reported no differences in the incidences of myocardial infarction between the RA and GA groups. A study by Dittman et al. found a small but statistical increase in the incidence of myocardial infarction in patients who received RA for above-knee amputation.[12] The reason for this is unclear and may represent a chance error. Two further studies[1718] examined composite cardiac outcomes and were unable to detect a statistically significant benefit of RA over GA.

6.Pulmonary outcomes

There is evidence that RA does not result in reduced incidences of pneumonia,[11131415131822] pulmonary embolism,[12131516] prolonged ventilation,[1131519] hypoxia[11] and unplanned intubation.[1119] A study utilising a composite pulmonary outcome yielded similar results.[18]

7.Neurological outcomes

There is evidence that RA does not result in increase in cerebrovascular accidents. Five studies reported no association between the choice of anaesthesia and cerebrovascular accidents.[1112161822] A meta-analysis performed on two propensity-matched studies[1118] involving a total of 1,962 patients showed that RA and GA were associated with similar incidences of cerebrovascular accidents. (OR 1.45; 95% CI (0.57, 3.70), I2 0%, P = 0.44) [Figure 4]. This evidence was graded as very low.

8.Renal outcomes

There is very low evidence that RA does not reduce renal failure. Five studies examined the association between RA and renal failure.[1112151618] None of them found an association between RA and decrease in incidence of renal failure. A meta-analysis of two propensity-matched studies[1118] showed that RA and GA were associated with similar incidences of renal failure (OR 1.27; 95%CI (0.73, 2.23), I2 0%, P = 0.4) [Figure 4]. This evidence was graded as low.

9.Blood loss and blood transfusion

Five studies[1112131516] reported blood transfusion and one reported blood loss.[15] None of these studies show a difference between the RA and GA groups.


Consistent with the published literature, this study found that LEA is associated with a high mortality.[23] The main finding of this study is that RA, when compared to GA, does not appear to reduce the incidence of 30-day and 90-day mortality following LEA. However, one study pointed out that in the high risk patients, there might be a benefit of RA over GA in terms of reducing the 30-day mortality.[4]

There is also evidence that RA is not associated with reduced cardiac, neurological, renal and bleeding complications after LEA. However, some studies have found that RA is associated with a lower H-LOS and ICU admission rate. But since these studies were not designed to assess H-LOS and ICU admission rates, these findings should be considered exploratory rather than definitive. Further studies are required to understand the impact of RA on H-LOS and ICU admission.

The strength of this study is that, it is the first review that has examined this clinical question. The studies included in this review had large sample sizes and looked at an objective outcome measure. Most of the studies scored well on the NOS, an instrument to assess the risk of bias in non-randomised trials. Across most of the outcome measures, the results were fairly homogenous. Although this review included only observational studies, the role of observational studies should not be downplayed when studying a harmful outcome. Randomised controlled trials, while useful for the examination of efficacy, may not be as useful for determining the rates of adverse events. This is due to the low frequency of events, small number of participants, restrictive inclusion/exclusion criteria and short follow-up periods.[24] Furthermore, it is considered unethical to randomise patients to a study that is evaluating a harmful outcome.[25] In this context, this review presents the best level of evidence till date on this topic.

The major limitation of this review is that while many studies adjusted for confounders, not all utilised a propensity score-matching technique. Hence, some of the studies were not pooled in the meta-analysis. Moreover, the results of secondary outcomes such as 90-day mortality, H-LOS and ICU admission should be interpreted with caution. Nevertheless, the results from these papers appear to have fairly uniform results in the primary and secondary outcome measures. This strengthens our confidence that the assessment is accurate. Another limitation is that most studies reported both peripheral and central neuraxial anaesthesia as RA. Since, peripheral nerve blocks result in less physiological derangement than central neuraxial anaesthesia, future studies that compare peripheral nerve blocks to GA may possibly give rise to opposing results from this review.[2627]

Taken together, during the informed consent process, patients should not be routinely counselled that RA reduces the risk of mortality. The choice of the anaesthesia technique should be individualised based on factors such as patient comorbidities, presence of a bleeding diathesis, haemodynamic stability, patient preference etc. It is nevertheless, still reasonable to offer RA to reduce the risk of ICU admission, reduce opioid consumption, reduce pain scores and prevent chronic pain.[111428]


LEA is a commonly performed surgery that carries significant mortality and morbidity. In this study, RA was not found to reduce the 30-day mortality after LEA when compared to GA. There is very low level of evidence that RA may decrease hospital length of stay and ICU admission rates

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1. Kalbaugh CA, Strassle PD, Paul NJ, McGinigle KL, Kibbe MR, Marston WA Trends in surgical indications for major lower limb amputation in the USA from 2000 |mto |d2016 Eur J Vasc Endovasc Surg 2020 60 88 96
2. Gabel J, Jabo B, Patel S, Kiang S, Bianchi C, Chiriano J, et al. Analysis of patients undergoing major lower extremity amputation in the vascular quality initiative Ann Vasc Surg 2018 46 75 82
3. van Netten JJ, Fortington LV, Hinchliffe RJ, Hijmans JM Early post-operative mortality after major lower limb amputation:A systematic review of population and regional based studies Eur J Vasc Endovasc Surg 2016 51 248 57
4. Khan SA, Qianyi RL, Liu C, Ng EL, Fook-Chong S, Tan MGE Effect of anaesthetic technique on mortality following major lower extremity amputation:A propensity score-matched observational study Anaesthesia 2013 68 612 20
5. Srilata M, Durga P, Ramachandran G Influence of changing trends in anaesthetic practice on morbidity and mortality in elderly patients undergoing lower limb surgery Indian J Anaesth 2014 58 132 7
6. Higgins JPT TJ, Chandler J, Cumpston M, Li T, Page MJ, Welch VA Cochrane Handbook for Systematic Reviews of Interventions Chicester (UK) John Wiley &Sons 2019
7. Moher D, Liberati A, Tetzlaff J, Altman DG Preferred reporting items for systematic reviews and meta-analyses:The PRISMA statement PLoS Med 2009 6 e1000097 doi:10.1371/journal.pmed. 1000097
8. Wells G, Shea B, O'Connell D, Robertson J, Peterson J, Welch V, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses Available from:
9. Langendam MW, Akl EA, Dahm P, Glasziou P, Guyatt G, Schunemann HJ Assessing and presenting summaries of evidence in cochrane reviews Syst Rev 2013 2 81 doi:10.1186/2046-4053-2-81
10. McGrath S, Zhao X, Steele R, Thombs BD, Benedetti A Collaboration DESD Estimating the sample mean and standard deviation from commonly reported quantiles in meta-analysis Stat Methods Med Res 2020 962280219889080 doi:10.1177/0962280219889080
11. Kim SJ, Kim N, Kim EH, Roh YH, Song J, Park KH, et al. Use of regional anesthesia for lower extremity amputation may reduce the need for perioperative vasopressors:A propensity score-matched observational study Ther Clin Risk Manage 2019 15 1163 71
12. Dittman J, Tse W, Lavingia KS, Wolfe L, Amendola MF General anesthesia associated with increased blood transfusions and multiple complications following above knee amputation in a large veterans administration cohort J Vasc Surg 2020 72 e134
13. Bilgili B, Haliloglu M, Edipoglu E, Cetingok H, Pektas Y, Bilgili MG The type of anesthesia technique does not have an impact on 30-day mortality in geriatric patients undergoing lower extremity amputation Int Surg 2017 102 178 83
14. Niskakangas M, Dahlbacka S, Liisanantti J, Vakkala M, Kaakinen T Spinal or general anaesthesia for lower-limb amputation in peripheral artery disease –a retrospective cohort study Acta Anaesthesiol Scand 2018 62 226 33
15. Chery J, Semaan E, Darji S, Briggs WT, Yarmush J, D'Ayala M Impact of regional versus general anesthesia on the clinical outcomes of patients undergoing major lower extremity amputation Ann Vasc Surg 2014 28 1149 56
16. Dittman J, Tse W, Lavingia KS, Wolfe L, Amendola MF Anesthesia type does not influence mortality in below the knee amputation in a national veteran cohort J Vasc Surg 2020 72 e251 e2
17. Hall MR, Kalbaugh CA, Tsujimoto THM, McGinigle KL Regional anaesthesia alone is reasonable for major lower extremity amputation in high risk patients and may initiate a more efficacious enhanced recovery programme Eur J Vasc Endovasc Surg 2020 60 747 51
18. Moreira CC, Farber A, Kalish JA, Eslami MH, Didato S, Rybin D, et al. The effect of anesthesia type on major lower extremity amputation in functionally impaired elderly patients J Vasc Surg 2016 63 696 701
19. Abe H, Sumitani M, Matsui H, Aso S, Inoue R, Fushimi K, et al. Comparing outcomes after peripheral nerve block versus general anesthesia for lower extremity amputation:A nationwide exploratory retrospective cohort study in Japan Reg Anesth Pain Med 2020 45 399 404
20. The Global Lower Extremity Amputation Study Group Epidemiology of lower extremity amputation in centres in Europe, North America and East Asia Br J Surg 2000 87 328 37
21. Chung PM, Chau BKH, Chow ECS, Lam KH, Wong NMR Factors that affect the one-year mortality rate after lower limb amputation in the Hong Kong Chinese population J Orthopaedics Trauma Rehabil 2021 28 22104917211056949 doi:10.1177/22104917211056949
22. Mann RA, Bisset WI Anaesthesia for lower limb amputation. A comparison of spinal analgesia and general anaesthesia in the elderly Anaesthesia 1983 38 1185 91
23. Kristensen MT, Holm G, Kirketerp-Moller K, Krasheninnikoff M, Gebuhr P Very low survival rates after non-traumatic lower limb amputation in a consecutive series:What to do? Interact Cardiovasc Thorac Surg 2012 14 543 7
24. Vandenbroucke JP What is the best evidence for determining harms of medical treatment? Can Med Assoc J 2006 174 645 6
25. Mellis CM How to choose your study design J Paediatr Child Health 2020 56 1018 22
26. Lai HY, Foo LL, Lim SM, Yong CF, Loh PS, Chaw SH, et al. The hemodynamic and pain impact of peripheral nerve block versus spinal anesthesia in diabetic patients undergoing diabetic foot surgery Clin Auton Res 2020 30 53 60
27. Kim NY, Lee KY, Bai SJ, Hong JH, Lee J, Park JM, et al. Comparison of the effects of remifentanil-based general anesthesia and popliteal nerve block on postoperative pain and hemodynamic stability in diabetic patients undergoing distal foot amputation:A retrospective observational study Medicine (Baltimore) 2016 95 e4302
28. Karanikolas M, Aretha D, Tsolakis I, Monantera G, Kiekkas P, Papadoulas S, et al. Optimized perioperative analgesia reduces chronic phantom limb pain intensity, prevalence, and frequency:A prospective, randomized, clinical trial Anesthesiology 2011 114 1144 54

Amputation; anaesthesia; conduction; lower extremity

Copyright: © 2022 Indian Journal of Anaesthesia