Ambulatory surgery is rapidly growing worldwide. In the past decade, the number of ambulatory surgical procedures has grown from 3.2 million to more than 21 million annually in the United States alone (1), and 50%–70% of all surgical procedures in North America are now performed on an ambulatory basis (2). A primary goal for ambulatory anesthesia is rapid recovery from anesthesia leading to rapid patient discharge with minimal side effects. Reduced time spent in the postanesthesia care unit (PACU) and ambulatory surgery unit (ASU) may be economically advantageous, especially if the PACU can be bypassed altogether (2,3). In addition to inherent speed of emergence from ambulatory anesthesia, surveys have demonstrated that postoperative pain and nausea are common reasons for delayed patient discharge, unplanned hospital admission, and patient concern (4–6). Regional anesthesia (RA) in the form of single-shot major conduction block has been advocated as an ideal anesthetic for ambulatory surgery. RA would be expected to speed patient recovery because of minimization of systemic anesthetic. RA would provide excellent postoperative analgesia and thus reduce need for opioids and risk of nausea. However, RA also requires additional performance time and technical skill and may introduce its own set of side effects such as urinary retention after central neuraxial blocks (CNB). Newer rapid-acting general anesthetics (GA) (desflurane, propofol) have become popular and are well suited for ambulatory anesthesia (7). Although these drugs hasten patient recovery, postoperative pain and nausea may be factors that delay recovery and discharge from the ASU. However, increasing use of multimodal analgesia (nonsteroidal analgesics, ketamine, local anesthetic infiltration) (4,8) and appropriate use of antiemetic prophylaxis may reduce these potential disadvantages for GA (9,10). Thus, we performed this meta-analysis to compare RA versus GA for ambulatory anesthesia.
The National Library of Medicine’s Medline database and the Cochrane Database of Systematic Reviews were searched for the time period 1966 to April 2005. A previously recommended search strategy using multiple search terms and text words was used without language restrictions (11). The terms “ambulatory surgical procedures,” “ambulatory care,” and the text word “outpatient” were combined with the term “OR.” This search set was then combined with the following search sets with “AND.” The term “spinal anesthesia” and text words “intrathecal” and “subarachnoid” were combined with “OR”. The term “epidural anesthesia” and text words “peridural” and “extradural” were combined with “OR.” The term “nerve block” and text words “axillary,” “brachial,” “interscalene,” “infraclavicular,” “femoral,” “sciatic,” “popliteal,” and “lumbar” were combined with “OR.” The three combination searches were then limited by the terms “adult” and “randomized controlled trial” (RCT). This final search identified 214 potential RCTs.
All of the above abstracts were reviewed for potential inclusion in the meta-analysis. Our intent was to compare single-shot regional anesthesia (major central or peripheral conduction block) versus general anesthesia (GA) as the primary anesthetic techniques for ambulatory surgery. All RCTs that stated the use of hospitalized patients were excluded. All RCTs that used continuous perineural techniques were excluded. All RCTs that stated the use of monitored sedation instead of GA as a control group were excluded. All RCTs that considered intraarticular analgesia or local infiltration to be RA were excluded. No minimum sample sizes were invoked for inclusion of studies in the analysis, and only studies on adults (age ≥19 yr) were included. After selecting the initial articles, the reference list of each of the analyzed articles was checked for any additional studies, as were the authors’ personal files for additional references that met all inclusion criteria.
Each study’s methodology and results were recorded. Data were extrapolated from figures as needed. Wherever possible, data were converted to incidence for dichotomous outcomes and to mean and sd (normal distribution was assumed) for continuous outcomes. Definition of outcomes and criteria for PACU and ASU discharge were recorded as originally defined by the study. Specific outcomes that were extracted included time for induction of anesthesia (beginning of induction until patient ready for surgery), incidences of nausea or vomiting, visual analog scale (VAS) pain scores in the PACU, incidence of rescue with analgesics in the PACU, incidence of ability to bypass PACU, time in PACU (patient arrival in PACU until discharge from PACU), and time until ASU discharge (total time from end of surgery until discharge from ASU). Patient satisfaction was measured with a variety of measures. Thus, we extracted the most prevalent measure, which was incidence of patients rating their satisfaction as excellent. RA was further subdivided into CNB consisting of spinal and epidural anesthesia or peripheral nerve blocks (PNB).
Meta-analysis was performed with a random effects model. The level of significance for all tests was set at a P value of 0.05 and variances were not assumed to be equal. For dichotomous outcomes, study results were pooled and odds ratios calculated with the Mantel-Haenszel method. Thus, odds ratios with 95% confidence intervals are displayed for effect statistic. For continuous outcomes, study results were pooled and means and standard deviations calculated with the inverse variance method. Thus, weighted mean differences and 95% confidence intervals are displayed for effect statistic. All statistical analyses were performed with Review Manager 4.2 (Cochrane Collaboration Information Management System, Nordic Cochrane Centre Rigshospitalet, Copenhagen, Denmark).
Fifteen (1003 patients) and 7 (359 patients) trials for CNB and PNB were included in the meta-analysis (Tables 1 and 2) (12–32). Results are summarized in Tables 3 and 4. Both CNB and PNB were associated with increased anesthesia induction time by 8–9 min as weighted by inverse variance. Both CNB and PNB were associated with decreased VAS pain scores in the PACU (−9 and −24 mm) and decreased need for PACU analgesics (odds ratios, 0.32 and 0.11). However, CNB was not associated with decreased PACU time or reduced nausea despite reduced analgesic need, and CNB was associated with increased ASU time by 35 min. The little difference between CNB and GA probably explains the similar incidences of excellent patient satisfaction. In contrast, PNB was associated with increased ability to bypass PACU (odds ratio, 14), decreased PACU time by 24 min, decreased risk of nausea (odds ratio, 0.17), increased incidence of excellent patient satisfaction (odds ratio, 4.7) but was not associated with decreased time until ASU discharge.
Our meta-analysis identified several potential advantages of RA over GA; however, none of the advantages translated into shortened ASU time. Use of CNB expectedly increased time for anesthetic induction by a small amount (∼8 minutes). This time investment was rewarded with reduced VAS scores and decreased need for postoperative analgesics. These observations are presumably explained by the maintained analgesia from the resolving anesthetic sensory block from CNB. Interestingly, the reduction in analgesic use did not reduce incidence of nausea. This is possibly a result of the ability of CNB itself to cause nausea and vomiting as a result of sympathetic block, hypotension, and the common use of spinal opioids in included trials (33). Use of CNB was not associated with reduced PACU use, and total time until discharge from ASU was prolonged by 35 minutes. Delay in achievement of several common discharge criteria may explain the prolongation of total ASU recovery time. Ability to ambulate without assistance is a common discharge criterion. Although the majority of studies used appropriate anesthetics (Table 1), prolonged bilateral lower extremity motor block from CNB may have delayed achievement of this goal. Recovery of ambulation may be accelerated by alterations in technique of CNB. Use of unilateral spinal anesthesia with small doses of hyperbaric local anesthetic and prolonged positioning in the lateral position may decrease the time needed until ability to ambulate with crutches, albeit at the cost of preoperative delay from the required lateral positioning (34). For epidural anesthesia, consistent use of shorter-acting local anesthetics such as 2-chloroprocaine may hasten recovery of bilateral motor block and ability to ambulate (17), albeit at the cost of potential backache with increasing doses (35). Another common discharge criterion that may have increased ASU time with CNB is urinary voiding. CNB can cause urinary retention, but evidence suggests that low-risk patients undergoing appropriate CNB for ambulatory anesthesia may be safely discharged before urinary voiding (36). Adoption of more lenient discharge criteria may improve the ASU profile of CNB and thereby improve patient satisfaction.
Use of PNB increased anesthetic induction time by a surprisingly small amount (8 minutes). This might reflect increasing proficiency in RA as a result of increased exposure to RA in residency training programs (37). The amount of additional time needed to perform PNB may be minimized by creating a separate RA block room with separate staff to perform the block, albeit at the cost of additional space and staffing (38). Use of PNB was associated with benefits of increased PACU bypass, decreased VAS scores, decreased need for postoperative analgesics, decreased incidence of nausea, shortened PACU time, and increased patient satisfaction. These findings should be cautiously applied, as a large number of patients (100 of 359) were enrolled from a single RCT (25). Nonetheless, examination of individual RCTs confirms the finding of some type of benefit from PNB versus GA. Such findings are consistent with the reduced need for systemic anesthetics, excellent postoperative analgesia, and lack of nausea resulting from the use of PNB. However, none of these benefits translated into decreased time until discharge from ASU. This is a curious finding, as pain and nausea have been implicated in prolonged ASU times (4,39,40) and PACU bypass has been proposed to decrease ASU time and readmission (2,3,41). Thus, it is unclear why the benefits of PNB were not associated with decreased time until discharge from ASU. Perhaps there is a negative bias among medical personnel and patients that prevents early ASU discharge, although recent surveys of patients’ and orthopedic surgeons’ attitudes towards RA do not consistently support such a bias (42,43). Alternatively, common discharge criteria and anesthesia practice may not be suitable to take advantage of the benefits of PNB. A recent survey of the Society of Ambulatory Anesthesia indicated that only 56% of respondents routinely discharged patients home with a long-lasting regional block still in place (44). Certainly, requiring patients to remain in the ASU until resolution of a medium- to long-lasting PNB would both delay patient discharge and nullify the benefit of postoperative analgesia at home. Finally, the familiarity of each individual institution with fast-tracking ambulatory patients may have influenced ability to rapidly discharge patients. The heterogeneous nature of the included RCTs and the small number of patients allows only speculation about these potential confounding factors. Future studies will be needed to determine the optimal environment to translate the potential benefits of RA into shorter ASU stays.
There are several limitations to this meta-analysis relating to the evolving nature of ambulatory anesthesia and to the general use of meta-analysis. There is continued refinement in the practice of ambulatory anesthesia. Targeted use of antiemetic prophylaxis can decrease risk of postoperative nausea and was used somewhat in 4 RCTs. Use of opioid-sparing multimodal analgesia with nonsteroidal analgesics, local anesthetic infiltration, or small-dose ketamine has been reported to reduce pain and side effects and decrease ASU stay (45) and was used somewhat in 15 RCTs. Use of bispectral index monitoring may allow more precise titration of GA with resultant reduction in side effects and PACU stay (46).
There are also limitations in the underlying data and technique available for meta-analysis. Data from RCTs were weighted based on trial size and not by quality. We chose not to assign quality scores, as use of these has been problematic (47). Ideally, a meta-analysis would include only a completely homogenous group of patients to approximate effects of a single large RCT. Examination of Tables 1 and 2 illustrates the current difficulties in finding completely homogenous data. To maximize the number of subjects and applicability of our findings to different environments, we chose to combine several different ambulatory surgical procedures, various patients, and various clinical practices of different institutions. Post hoc power analysis based on the GA patients from the included RCTs indicates that if our data came from a single RCT, then there are sufficient patients to detect a 13-minute difference in ASU time associated with CNB (n = 839) and a 20-minute difference associated with PNB (n = 328). This power analysis is not truly applicable because of the previously mentioned heterogeneity in RCTs in this meta-analysis but it does provide some perspective. We attempted to minimize heterogeneity by separating RA patients into the clinically distinct CNB and PNB groups. We also used a random effects model for meta-analysis, which assumes some heterogeneity between studies and is thus more stringent in assigning statistical significance than a fixed effect model. Nonetheless, the summary estimates of effect (odds ratio and weighted mean difference) should be interpreted with caution. The exact values of these estimates from this meta-analysis may not agree with a future single large RCT as a result of the inherent heterogeneity of pooled RCTs. The PNB versus GA group was particularly heterogeneous, and the summary effect estimates for this subanalysis may be more illustrative than predictive. The relative merits and agreement between meta-analyses and large RCTs are controversial. Several studies have reported varying amounts of agreement in magnitude of difference or statistical significance (65%–90%) between meta-analyses and large RCTs while still concluding that that meta-analyses and large RCTs usually point in the same direction (48,49).
In conclusion, this meta-analysis associated several potential advantages for RA versus GA for ambulatory anesthesia. Curiously, none of the benefits were associated with decreased ASU time and use of CNB is associated with a 35-minute delay until patient discharge from the ASU. As all included RCTs were relatively small (26 to 162 subjects), we hope this investigation stimulates further large RCTs examining RA blocks that incorporate optimal fast-tracking pathways, multimodal analgesia, efficient patient discharge criteria, and postoperative follow-up.
The authors thank Dr. Admir Hadzic and Dr. Colin McCartney for providing additional data.
1. Hall MJ, Lawrence L. Ambulatory surgery in the United States, 1996. Adv Data 1998;300:1–16.
2. Apfelbaum JL, Walawander CA, Grasela TH, et al. Eliminating intensive postoperative care in same-day surgery patients using short-acting anesthetics. Anesthesiology 2002;97:66–74.
3. Williams BA, Kentor ML, Vogt MT, et al. Economics of nerve block pain management after anterior cruciate ligament reconstruction: potential hospital cost savings via associated postanesthesia care unit bypass and same-day discharge. Anesthesiology 2004;100:697–706.
4. Pavlin DJ, Chen C, Penaloza DA, et al. Pain as a factor complicating recovery and discharge after ambulatory surgery. Anesth Analg 2002;95:627–34.
5. Coley KC, Williams BA, DaPos SV, et al. Retrospective evaluation of unanticipated admissions and readmissions after same day surgery and associated costs. J Clin Anesth 2002;14:349–53.
6. Eberhart LH, Morin AM, Wulf H, Geldner G. Patient preferences for immediate postoperative recovery. Br J Anaesth 2002;89:760–1.
7. Gupta A, Stierer T, Zuckerman R, et al. Comparison of recovery profile after ambulatory anesthesia with propofol, isoflurane, sevoflurane and desflurane: a systematic review. Anesth Analg 2004;98:632–4.
8. Gan TJ, Joshi GP, Viscusi E, et al. Preoperative parenteral parecoxib and follow-up oral valdecoxib reduce length of stay and improve quality of patient recovery after laparoscopic cholecystectomy surgery. Anesth Analg 2004;98:1665–73.
9. Apfel CC, Korttila K, Abdalla M, et al. A factorial trial of six interventions for the prevention of postoperative nausea and vomiting. N Engl J Med 2004;350:2441–51.
10. Gan TJ, Meyer T, Apfel CC, et al. Consensus guidelines for managing postoperative nausea and vomiting. Anesth Analg 2003;97:62–71.
11. Haynes RB, Wilczynski N, McKibbon KA, et al. Developing optimal search strategies for detecting clinically sound studies in MEDLINE. J Am Med Inform Assoc 1994;1:447–58.
12. Dahl V, Gierloff C, Omland E, Raeder JC. Spinal, epidural or propofol anaesthesia for out-patient knee arthroscopy? Acta Anaesthesiol Scand 1997;41:1341–5.
13. Danelli G, Berti M, Casati A, et al. Spinal block or total intravenous anaesthesia with propofol and remifentanil for gynaecological outpatient procedures. Eur J Anaesth 2002;19:594–9.
14. Erhan E, Ugur G, Anadolu O, et al. General anaesthesia or spinal anaesthesia for outpatient urological surgery. Eur J Anaesth 2003;20:647–52.
15. Lennox PH, Vaghadia H, Henderson C, et al. Small-dose selective spinal anesthesia for short-duration outpatient laparoscopy: recovery characteristics compared with desflurane anesthesia. Anesth Analg 2002;94:346–50.
16. Li S, Coloma M, White PF, et al. Comparison of the costs and recovery profiles of three anesthetic techniques for ambulatory anorectal surgery. Anesthesiology 2000;93:1225–30.
17. Mulroy MF, Larkin KL, Hodgson PS, et al. A comparison of spinal, epidural, and general anesthesia for outpatient knee arthroscopy. Anesth Analg 2000;91:860–4.
18. Richardson MG, Dooley JW. The effects of general versus epidural anesthesia for outpatient extracorporeal shock wave lithotripsy. Anesth Analg 1998;86:1214–8.
19. Song D, Greilich NB, White PF, et al. Recovery profiles and costs of anesthesia for outpatient unilateral inguinal herniorrhaphy. Anesth Analg 2000;91:876–81.
20. Toivonen J, Permi J, Rosenberg PH. Analgesia and discharge following preincisional ilioinguinal and iliohypogastric nerve block combined with general or spinal anaesthesia for inguinal herniorrhaphy. Acta Anaesthesiol Scand 2004;48:480–5.
21. Wong J, Marshall S, Chung F, et al. Spinal anesthesia improves the early recovery profile of patients undergoing ambulatory knee arthroscopy. Can J Anaesth 2001;48:369–74.
22. Casati A, Cappelleri G, Berti M, et al. Randomized comparison of remifentanil-propofol with a sciatic-femoral nerve block for out-patient knee arthroscopy. Eur J Anaesth 2002;19:109–14.
23. Hadzic A, Arliss J, Kerimoglu B, et al. A comparison of infraclavicular nerve block versus general anesthesia for hand and wrist day-case surgeries. Anesthesiology 2004;101:127–32.
24. Jankowski CJ, Hebl JR, Stuart MJ, et al. A comparison of psoas compartment block and spinal and general anesthesia for outpatient knee arthroscopy. Anesth Analg 2003;97:1003–9.
25. McCartney CJ, Brull R, Chan VW, et al. Early but no long-term benefit of regional compared with general anesthesia for ambulatory hand surgery. Anesthesiology 2004;101:461–7.
26. Terheggen MA, Wille F, Borel Rinkes IH, et al. Paravertebral blockade for minor breast surgery. Anesth Analg 2002;94:355–9.
27. Jacobson E, Forssblad M, Rosenberg J, et al. Can local anesthesia be recommended for routine use in elective knee arthroscopy? A comparison between local, spinal, and general anesthesia. Arthroscopy 2000;16:183–90.
28. Stewart AVG, Vaghadia H, Collins L, Mitchell GWE. Small-dose selective spinal anaesthesia for short-duration outpatient gynaecological laparoscopy: recovery characteristics compared with propofol anaesthesia. Br J Anaesth 2001;86:570–2.
29. Korhonen AM, Valanne JV, Jokela RM, et al. A comparison of selective spinal anesthesia with hyperbaric bupivacaine and general anesthesia with desflurane for outpatient knee arthroscopy. Anesth Analg 2004;99:1668–73.
30. Forssblad M, Jacobson E, Weidenhielm L. Knee arthroscopy with different anesthesia methods: a comparison of efficacy and cost. Knee Surg Sports Traumatol Arthrosc 2004;12:344–9.
31. Hadzic A, Karaca PE, Hobeika P, et al. Peripheral nerve blocks result in superior recovery profile compared with general anesthesia in outpatient knee arthroscopy. Anesth Analg 2005;100:976–81.
32. Hadzic A, Williams BA, Karaca PE, et al. For outpatient rotator cuff surgery, nerve block anesthesia provides superior same-day recovery over general anesthesia. Anesthesiology 2005;102:1001–7.
33. Borgeat A, Ekatodramis G, Schenker CA. Postoperative nausea and vomiting in regional anesthesia: a review. Anesthesiology 2003;98:530–47.
34. Borghi B, Stagni F, Bugamelli S, et al. Unilateral spinal block for outpatient knee arthroscopy: a dose-finding study. J Clin Anesth 2003;15:351–6.
35. Stevens RA, Urmey WF, Urquhart BL, Kao TC. Back pain after epidural anesthesia with chloroprocaine. Anesthesiology 1993;78:492–7.
36. Mulroy MF, Salinas FV, Larkin KL, Polissar NL. Ambulatory surgery patients may be discharged before voiding after short-acting spinal and epidural anesthesia. Anesthesiology 2002;97:315–9.
37. Kopacz DJ, Neal JM. Regional anesthesia and pain medicine: residency training: the year 2000. Reg Anesth Pain Med 2002;27:9–14.
38. Williams BA, Kentor ML, Williams JP, et al. Process analysis in outpatient knee surgery: effects of regional and general anesthesia on anesthesia-controlled time. Anesthesiology 2000;93:529–38.
39. Williams BA, Kentor ML, Vogt MT, et al. Femoral-sciatic nerve blocks for complex outpatient knee surgery are associated with less postoperative pain before same-day discharge: a review of 1,200 consecutive cases from the period 1996–1999. Anesthesiology 2003;98:1206–13.
40. Chung F, Mezei G. Factors contributing to a prolonged stay after ambulatory surgery. Anesth Analg 1999;89:1352–9.
41. Williams BA, Kentor ML, Williams JP, et al. PACU bypass after outpatient knee surgery is associated with fewer unplanned hospital admissions but more phase II nursing interventions. Anesthesiology 2002;97:981–8.
42. Matthey PW, Finegan BA, Finucane BT. The public’s fears about and perceptions of regional anesthesia. Reg Anesth Pain Med 2004;29:96–101.
43. Oldman M, McCartney CJ, Leung A, et al. A survey of orthopedic surgeons’ attitudes and knowledge regarding regional anesthesia. Anesth Analg 2004;98:1486–90, table of contents.
44. Klein SM, Pietrobon R, Nielsen KC, et al. Peripheral nerve blockade with long-acting local anesthetics: a survey of the Society for Ambulatory Anesthesia. Anesth Analg 2002;94:71–6.
45. Menigaux C, Guignard B, Fletcher D, et al. Intraoperative small-dose ketamine enhances analgesia after outpatient knee arthroscopy. Anesth Analg 2001;93:606–12.
46. Liu SS. Effects of Bispectral Index monitoring on ambulatory anesthesia: a meta-analysis of randomized controlled trials and a cost analysis. Anesthesiology 2004;101:311–5.
47. Balk EM, Bonis PA, Moskowitz H, et al. Correlation of quality measures with estimates of treatment effect in meta-analyses of randomized controlled trials. JAMA 2002;287:2973–82.
48. Bailar JC 3rd. The promise and problems of meta-analysis. N Engl J Med 1997;337:559–61.
49. Ioannidis JP, Cappelleri JC, Lau J. Issues in comparisons between meta-analyses and large trials. JAMA 1998;279:1089–93.