M.E. Bauer, a J.A. Kountanis, a L.C. Tsen, b M.L. Greenfield, a J.M. Mhyre a ORIGINAL ARTICLE. Introduction

Transcription

1 International Journal of Obstetric Anesthesia (2012) 21, X/$ - see front matter c 2012 Elsevier Ltd. All rights reserved. ORIGINAL ARTICLE Risk factors for failed conversion of labor epidural analgesia to cesarean delivery anesthesia: a systematic review and meta-analysis of observational trials M.E. Bauer, a J.A. Kountanis, a L.C. Tsen, b M.L. Greenfield, a J.M. Mhyre a a Department of Anesthesiology, University of Michigan Health System, Ann Arbor, MI, USA b Department of Anesthesiology, Brigham and Women s Hospital, Boston, MA, USA ABSTRACT Background: This systematic review and meta-analysis evaluates evidence for seven risk factors associated with failed conversion of labor epidural analgesia to cesarean delivery anesthesia. Methods: Online scientific literature databases were searched using a strategy which identified observational trials, published between January 1979 and May 2011, which evaluated risk factors for failed conversion of epidural analgesia to anesthesia or documented a failure rate resulting in general anesthesia. Results: 1450 trials were screened, and 13 trials were included for review (n = 8628). Three factors increase the risk for failed conversion: an increasing number of clinician-administered boluses during labor (OR = 3.2, 95% CI ), greater urgency for cesarean delivery (OR = 40.4, 95% CI ), and a non-obstetric anesthesiologist providing care (OR = 4.6, 95% CI ). Insufficient evidence is available to support combined spinal epidural versus standard epidural techniques, duration of epidural analgesia, cervical dilation at the time of epidural placement, and body mass index or weight as risk factors for failed epidural conversion. Conclusion: The risk of failed conversion of labor epidural analgesia to anesthesia is increased with an increasing number of boluses administered during labor, an enhanced urgency for cesarean delivery, and care being provided by a non-obstetric anesthesiologist. Further high-quality studies are needed to evaluate the many potential risk factors associated with failed conversion of labor epidural analgesia to anesthesia for cesarean delivery. c 2012 Elsevier Ltd. All rights reserved. Keywords: General anesthesia; Cesarean delivery; Epidural; analgesia/anesthesia Introduction Neuraxial anesthesia is commonly preferred over general anesthesia for cesarean delivery because it permits maternal participation in the birth process, limits the potential for difficult airway management or awareness under general anesthesia, avoids the depressant effects of systemic anesthesia medication on the fetus and uterine tone, and facilitates the provision of postoperative analgesia. 1,2 Conversion of labor epidural analgesia to cesarean delivery anesthesia is an important strategy in limiting the use of general anesthesia in obstetrics. Failure of conversion can result in unpredictable hazards associated with managing the obstetric airway, performing a spinal technique in the presence of partial epidural Accepted May 2012 Correspondence to: M.E. Bauer, Department of Anesthesiology, University of Michigan Health System, Ann Arbor, MI, USA. address: mbalun@med.umich.edu blockade, or titration of appropriate analgesic and sedative medications. 3 8 In contrast, high rates of successful conversion may represent a useful quality measure, indicating the prior presence of functional epidural analgesia as well as the avoidance of general anesthesia. 9,10 The Royal College of Anaesthetists has published best practice guidelines for providing anesthesia for cesarean delivery. 11,12 These state that an acceptable rate of general anesthesia in a parturient receiving labor epidural analgesia should be no more than 3%. An understanding of the risk factors for failed conversion creates an opportunity to identify and replace an at-risk epidural catheter. Strategies designed to increase the success rate for conversion of labor analgesia to anesthesia may improve safety and quality. A systematic review of the literature to evaluate the risk factors associated with failed labor epidural analgesia conversion to cesarean delivery anesthesia was undertaken. Several observational studies have indicated that risk factors may include: an increased number

2 M.E. Bauer et al. 295 of top-ups or breakthrough pain during labor; increased urgency for cesarean delivery; higher body mass index (BMI) or weight; care being provided by a non-obstetric anesthesiologist; increased duration of epidural analgesia; use of an epidural versus combined spinal epidural (CSE) technique for analgesia; and lower cervical dilation at the time of epidural placement The study goal was to evaluate the evidence associated with any identified risk factor for a failed conversion of labor epidural analgesia to cesarean delivery anesthesia. Methods Data sources, search strategies, and study selection OvidSP, PubMed, and Excerpta Medical Data (EM- BASE) databases were searched to identify observational studies (cohort, case-control, and cross-sectional studies) published in English between 1979 and May Articles that included failed conversion of labor epidural analgesia to cesarean delivery epidural anesthesia were included according to a written protocol. The final search strategy combined medical subject headings from this preliminary search and applied to OvidSP and PubMed, with corresponding terms used for EMBASE (see web supplementary of search strategies for each database). Final articles included in the systematic review were entered into the Institute for Scientific Information (ISI) Web of Science to identify citing articles that could also be considered for study inclusion. Bibliographies of selected articles were hand-searched for other articles that might be suitable for inclusion. Two reviewers independently evaluated article titles for relevance. The inclusion criterion was any observational study whose title suggested that the study might include incidence of failed conversion, even if the failed conversion was not the primary outcome. Exclusion criteria were: non-english studies, case reports, case series, correspondence, editorials, reviews, systematic reviews, meta-analyses, or articles referring solely to either intrathecal anesthesia, general anesthesia, postpartum management, or neonatal outcomes. The same inclusion and exclusion criteria were used for the title review followed by the abstract review. Evaluators scored titles and abstracts as relevant, possibly relevant, or not relevant to failed conversion. Titles classed as not relevant by both evaluators were excluded. Further reviewers evaluated the remaining study abstracts for potentially relevant articles, resulting in 88 articles. These were reviewed in detail by two investigators, who included observational studies that reported risk factors predicting failed conversion or the incidence of failed conversion of labor epidural analgesia to cesarean delivery anesthesia. Disagreements regarding inclusion were resolved by consensus following independent analysis by a third investigator. Kappa statistics were used to evaluate agreement between the two reviewers. Definitions, data extraction and analyses Data were extracted for all articles that included associated risk factors for conversion of epidural analgesia to anesthesia that appeared in two or more articles. Articles that included a failure rate without associated risk factors were also included for meta-analysis for overall failure rate, second anesthetic utilization, and supplementation. Data were combined when possible in a random-effects meta-analysis to evaluate the evidence for each risk factor. The point estimate and 95% confidence intervals on the summary event rates across studies were obtained using random-effects meta-analysis, assuming a binomial distribution. Analyses were conducted in Stata version 12.0 (StataCorp LP, College Station, TX, USA). Several studies specifically excluded urgent cesarean delivery from analysis. We included these prior exclusions in our meta-analysis for urgency as a risk factor for epidural failure, and to provide a more accurate rate of general anesthesia or second anesthetic with a labor epidural catheter in situ. Authors were contacted to provide data that were not available in their original publications, and specifically to clarify details about failed epidural conversions that were not reported in the published paper. Methodological evaluation To assess the quality of reporting, the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement was used together with the STROBE Explanation and Elaboration tool to guide the investigators in their independent evaluations. 20,21 Specifically, the methodology described by Papathanasiou was adopted, limiting the assessment to only those STROBE items or sub-items related to study methods and reporting of results. 22 For each article, this score was the number of items marked yes divided by the total applicable items. An item was scored yes if it was explicitly described, no if the item was presented incompletely (or not all), or not applicable to the study design. A standardized form was created for the adapted STROBE checklist for each study. Disagreements were resolved by discussion. The frequency of reporting of the STROBE items among the relevant articles was reported for all relevant articles. Results Search results by database are presented in Fig. 1. The kappa statistic for final article selection was 0.91 (95% CI 0.79 to 1.0); all included studies were published between 1994 and. From a total of 1450 trials screened, 13 observational trials were included (n = 8628 women) for reporting failure rate and identification of risk factors for failed conversion of labor epidural analgesia to cesarean delivery anesthesia ,23 28 A summary of included articles and risk factors for

3 296 Conversion of epidural analgesia to anesthesia Fig. 1 Results of search strategy and exclusion criteria for observational studies of failed analgesia to anesthesia conversion. failure are reported in Tables 1 and 2. Based on randomeffects meta-analysis of all studies, the percentage of cesarean deliveries performed under general anesthesia with a pre-existing labor epidural catheter in place was 5.0% (95% CI 3.5 to 6.5%; Fig. 2). Requirement of a second anesthetic, including spinal or repeat epidural techniques or general anesthesia at the time of cesarean delivery was 7.7% (95% CI 5.0 to 10.5%; Fig. 3). Four studies reported the requirement for any intravenous or inhalational supplementation. 14,18,19,25 Overall, the estimated proportion of patients with a pre-existing labor epidural catheter who were given supplementation (intravenous, inhalational, or not specified) for cesarean delivery was 10.7% (95% CI 4.2 to 17.3%). In individual studies, the rate ranged from 3.0% to 14.7% (Fig. 4). Evidence for heterogeneity was present in all three measures, with the I 2 > 90%. Number of top-ups during labor Six studies evaluated the number of unscheduled topups required to maintain effective analgesia during labor as a risk factor for failed epidural anesthesia Halpern and Lee identified two or more clinician topups as an independent risk factor for failure, 14,15 while Campbell identified one or more clinician top-ups as an independent risk factor for failure. 13 Using random-effects meta-analysis, these studies suggest that among women who require a greater number of top-ups, the rate for subsequent failed conversion from epidural analgesia to anesthesia increases threefold, 16.4% compared with 4.6%, respectively (OR = 3.2, 95% CI 1.8 to 5.5; Fig. 5). Less statistical heterogeneity was found among groups (I 2 value of 46.6%, P = 0.15). Two additional trials reported that a greater mean number of topups were required to maintain labor analgesia among women who subsequently experienced failed versus successful epidural anesthesia (Table 2). 16,17 Orbach-Zinger reported that mean Visual Analog Score (VAS) on a scale of in the 2 h before cesarean delivery was higher among women who subsequently experienced failed epidural anesthesia (mean ± SD) 34.6 ± 32.4 versus 15.7 ± 26.6, P = 0.03; additional top-ups were given when the VAS score was >30 during labor. 16 The study by Tortosa had insufficient power to identify any relationship between the number of top-ups administered to patients who had a failed or successful conversion of epidural analgesia to anesthesia. 18 Urgency of cesarean delivery Kinsella and Halpern both studied an association between urgency of cesarean delivery and rates of failed epidural conversion. 14,19 Using Lucas s classification of

4 Table 1 Characteristics of articles included for systematic review Population and primary outcome Albright (1999) 23 Patients having labor epidural (n = 554) * and CSE analgesia; complications and failure rates of CSE Bamgbade () 24 (n = 94) * CD in obese and non-obese patients; perioperative complications Campbell () 13 (n = 895) * Patients with labor epidural analgesia requiring CD; successful rate of conversion and identification of risk factors for failure Gaiser (1994) 25 (n = 109) Patients undergoing emergent CD; time from drug administration to surgical incision Garry (2002) 26 (n = 827) * Patients having CD under neuraxial anesthesia; number of conversions to GA Halpern () 14 (n = 501) Patients having CD after labor epidural analgesia; incidence of GA and identification of risk factors for failure Kan (2004) 27 (n = 850) * Patients having CD; rate of GA for CD Kinsella (2008) 19 (n = 1392) * Patients having CD; rate of GA conversion from neuraxial technique for CD and to identify risk factors for failure Analgesic maintenance regimen Epidural anesthesia bolus regimen Definition of failure Quality score Bolus: sufentanil ± bupivacaine for CSE. Maintenance: 0.1% bupivacaine + fentanyl 2 lg/ml at 8 12 ml/h or PCEA C8 12 ml/h; B5 6 ml; L5; M3 doses/h 2% lidocaine with epinephrine 1: or 3% 2-chlorprocaine Repeat block or conversion to GA Not defined Not defined Not explicitly defined 50% 0.08% ropivacaine with fentanyl 2 lg/ml, PCEA or CIEA, settings not given Multiple: 2% carbonated lidocaine ± fentanyl, 3% 2- chlorprocaine, 0.5% ropivacaine, 0.5% bupivacaine Not defined Comparison between 3% 2- chlorprocaine and 1.5% lidocaine Not defined Bolus: ml of bupivacaine %. Maintenance: 0.08% bupivacaine and fentanyl 2 lg/ml, PCEA C5 10; B5 7; L 10; M not given Not defined Not defined ml of either 2% lidocaine or 0.5% bupivacaine, (or a mixture of the two) or 0.75% ropivacaine. Addition of fentanyl 100 lg was recommended Drugs and dosages according to the judgment of the attending anesthesiologist 1.5% lidocaine ml with epinephrine 1:200000, 8.4% sodium bicarbonate 2 ml and fentanyl lg Not standardized: bupivacaine and lidocaine and levobupivacaine, ± epinephrine 1:200000, 8.4% sodium bicarbonate, fentanyl or diamorphine Unsuccessful conversion from labor analgesia to surgical anesthesia or no attempted conversion for CD 50% 72.2% Conversion to GA 61.1% Conversion to GA 27.8% Conversion to another form of anesthesia or to replace the epidural catheter in the operating room 61.1% Conversion to GA 60% Pre-operative failure: conversion to another anesthetic pre-surgery or failure to achieve a satisfactory block. Intra-operative failure: anesthesia that required analgesia 84.2% (continued on next page) M.E. Bauer et al. 297

5 Table 1 (continued) Population and primary outcome Lee () 15 (n = 1033) * Patients with labor epidural analgesia who required CD; failure rate of epidural anesthesia conversion and identification of risk factors for failure Orbach-Zinger (2006) 16 (n = 103) * Patients undergoing CD with a functioning epidural catheter in place; failure rate of epidural anesthesia conversion and identification of risk factors for failure Pan (2004) 28 (n = 1830) * Labor epidural and CSE techniques for patients having vaginal and CD; epidural and CSE technique failure rates Riley (2002) 17 (n = 246) * Patients for whom an anesthetic record was present for labor analgesia and CD; rate of failed epidural anesthesia conversion and identification of risk factors Tortosa (2003) 18 (n = 194) Patients requiring non-elective CD with prior labor epidural analgesia; rate of failed epidural anesthesia conversion and identification of risk factors for failure Analgesic maintenance regimen Epidural anesthesia bolus regimen Definition of failure Quality score Bolus % bupivacaine or ropivacaine 8 15 ml Maintenance: CIEA % ropivacaine or bupivacaine plus 2 lg/ml fentanyl at 6 12 ml/h or PCEA of the same solution C 0 10/B 5/L10/M20 Bolus: 0.1% bupivacaine 10 ml + 0.1% bupivacaine 5 ml plus fentanyl 100 lg Maintenance: 0.1% bupivacaine and fentanyl 2 lg/ml at ml/h 0.11% bupivacaine, with or without fentanyl 2 lg/ml, PCEA settings: C 6 12 ml/h; B 5 ml; L 8 10 min: M 35 ml Bolus 0.125% bupivacaine with sufentanil 10 lg ml Maintenance: CIEA % bupivacaine plus sufentanil 0.33 lg/ml at ml/h Bupivacaine 0.08% and sufentanil 0.5 lg/ml infusion after a bolus of 15 ml of the same solution 1.5% lidocaine 20 ml, + bicarbonate (1 meq/10 ml), epinephrine (5 lg/ml) & fentanyl lg/ml 2% lidocaine 16 ml with 8.4% bicarbonate 1 ml and fentanyl 100 lg Inadequate neuraxial blockade for CD in the presence of adequate time for onset of epidural anesthesia Conversion to GA at any time after surgery commenced 3% 2-chlorprocaine Inadequate anesthesia requiring catheter replacement or conversion to alternative anesthetic technique 95% of cases: 2% lidocaine with 1: epinephrine and bicarbonate (1 meq/10 ml) 5% of cases:0.5% bupivacaine or 3% 2- chlorprocaine Lidocaine with 1: epinephrine with possible addition of sufentanil and/or clonidine Another anesthetic was used for the surgery 66.7% 61.1% 61.1% 38.9% Conversion to GA 55.6% B = bolus; C = continuous; CD = cesarean delivery; CIEA = continuous infusion epidural analgesia; CSE = combined spinal epidural; GA = general anesthesia; L = lockout; M = max hourly; PCEA = patient controlled epidural analgesia. * Denotes number of patients meeting criteria for inclusion in this study, may also include patients previously excluded from original publication. 298 Conversion of epidural analgesia to anesthesia

6 Table 2 Albright (1999) 23 (n = 554) Bamgbade () 24 (n = 94) Campbell () 13 (n = 895) Garry (2002) 26 (n = 827) Halpern () 14 (n = 501) Kinsella (2008) 19 (n = 1286) Lee () 15 (n = 1025) Orbach-Zinger (2006) 16 (n = 101) Risk factors for failed epidural anesthesia Number of top ups & breakthrough pain episodes Boluses P1, F = 20.6%(60/290) Bolus = 0, F = 9.9%(60/605) #boluses F = 0.98 ± 1.2 #boluses S = 0.56 ± 0.96 P < Boluses P2, F = 13.2% Boluses 61, F = 4.7% P = Boluses 61, F = 1.1%, Boluses = 2 5, F = 6.8% OR = % CI P < # Boluses S = 0.6 ± 1.1 F = 1.7 ± 1.6 P = Urgency of CD 2 Patients received GA for urgency 30 Patients received GA for urgency F = 30% (9/30) S = 18% (85/471) Category 1 CD OR = 2.45 (95% CI ) P = * 6 Patients excluded for urgency 2 Patients excluded for urgency Body mass index/weight Non-obese: F = 0% (0/34) Obese: F = 6.6%(4/60) BMI > 40 F = 16.4% (19/ 116) S = 13.3% (99/ 743) BMI > 35 & BMI for epidural failure, BMI > 30, P=0.039 * S =73±31 F =74±14 BMI S = 27.7 ± 3.7 F = 31.5 ± 3.8 P = Non-obstetric anesthesiologist OB anesthesiologist F = 1.2% (3/246) Non-OB anesthesiologist F = 5.9% (36/ 649) P = Cervical dilation OR = 0.84 ( ) S = 3 [2 4] F = 3 [2 4] S = 3.6 ± 1.4 F = 3.6 ± 0.8 Duration of epidural analgesia (h) S = 8.3 ± 4.4 F = 8.4 ± 4.1, S = 10.0 ± 5.3 F = 6.9 ± 4.3 P = S =8±6 F =12±6 OR = % CI P = 0.02 S = 7.8 ± 5.3 F = 8.6 ± 3.9 CSE technique compared with epidural technique CSE technique F = 4.5% (25/554) CSE S = 5.7% (27/471) F = 13.3% (4/30) CSE, F = 1.1% (10/905) Epidural, F = 5.8% (7/120) OR = % CI P = (continued on next page) M.E. Bauer et al. 299

7 300 Conversion of epidural analgesia to anesthesia Table 2 (continued) CSE technique compared with epidural technique Duration of epidural analgesia (h) Cervical dilation Non-obstetric anesthesiologist Body mass index/weight Urgency of CD Number of top ups & breakthrough pain episodes CSE technique F = 7.5% (3/41) Epidural technique F = 8.5% (17/199) S = 9.0 ± 6.0 F = 10.0 ± 6.2 OB anesthesiologist F =3% Non-OB anesthesiologist F = 11% S =77±14 F =81±15 3 Patients excluded for urgency # Boluses S = 1 [0 8] F = 3 [0 10] P < Riley (n = 240) S = 6.1 ± 3.0 F = 4.8 ± 2.4 P < hour OB anesthesiologist present S =74±13 F =71±8 3 Patients received GA for urgency # Boluses S = 0 [0 4] F = 0 [0 4] Tortosa (2003) 18 (n = 194) Data are mean ± SD or median [range] unless stated otherwise; CD = cesarean delivery; CSE = combined spinal epidural; F = failed conversion; S = successful conversion; BMI = body mass index; OB = obstetric; = not significant; OR = odds ratio; CI = confidence interval. * Includes all neuraxial anesthesia failures including spinal, de novo epidural technique, epidural anesthesia conversion. urgency of cesarean delivery, 29 Kinsella reported that 25% (18/72) of attempts to convert epidural analgesia to anesthesia ultimately required general anesthesia for cesarean delivery classified as category 1 (immediate threat to life of woman or fetus), compared with 7% (35/505) for category 2 (maternal or fetal compromise that is not immediately life-threatening), and 2.4% (11/ 452) for category 3 (needing delivery but no maternal or fetal compromise), P < It was reported that 106 (7.6%) women had epidural analgesia established during labor that was not topped-up for cesarean delivery, primarily due to poor analgesic quality. Although Kinsella demonstrated an association between urgency of cesarean delivery and the proportion requiring general anesthesia, a number of studies specifically excluded urgent cesarean delivery from analysis. We included these in our meta-analysis as urgency as a risk factor for epidural failure, (OR = 40.4, 95% CI ; Fig. 6). Halpern observed that fetal heart rate abnormalities were not an independent risk factor for failed conversion of epidural anesthesia. 14 Non-obstetric anesthesiologist Two studies found that the rate of conversion from epidural analgesia to general anesthesia was significantly increased when a non-obstetric anesthesiologist was managing the anesthetic. There was a 7.2% failure rate for non-obstetric anesthesiologists versus a 1.6% failure rate for obstetric anesthesiologists on average; OR = % CI ,17 Results were consistent between studies, with an I 2 value of 0%, P = (Fig. 7). Campbell reported that ineffective conversion to epidural surgical anesthesia could be remedied in 84.6% of cases by withdrawing the epidural catheter 1 cm and administering further doses of local anesthetic. This technique was applied after failed surgical anesthesia in 21/36 (58.3%) of cases performed by obstetric anesthesiologists, and in 5/84 (5.9%) by general anesthesiologists. 13 Body mass index or weight Eight studies reported BMI or weight as a factor possibly associated with failed epidural conversion ,24 Of these, six reported the mean weight for both groups Random-effects meta-analysis suggests that the standardized mean difference in weight was not different between women with successful or failed conversion of epidural analgesia to anesthesia (standardized mean difference 0.14 kg greater in women with failed conversion, 95% CI = ; Fig. 8. Statistical heterogeneity was present, although not statistically significant, I 2 value = 53.4%, P = In addition, only Orbach- Zinger reported a statistically significant association with weight and an increase in failed epidural anesthesia, although mean weights for each group were not reported. 16 Bamgbade found no association between failed epidural conversion and BMI. 24

8 M.E. Bauer et al. 301 Proportion Converted to General Anesthesia Year General Total Event Rate in % Study Published n N percent (95% CI) Gaiser Albright Garry Riley Tortosa Pan Kan Orbach-Zinger Kinsella Campbell Lee Halpern Bamgbade 2 94 Overall (I-squared = 91.7%, p = 0.000) 0.00 (0.00, 2.75) (1.63, 4.50) (8.43, 12.61) (5.71, 12.99) (0.35, 4.81) (3.14, 4.95) (5.23, 8.65) (13.44, 29.27) (5.43, 8.07) 4.36 (3.02, 5.70) 2.23 (1.33, 3.13) 4.19 (2.44, 5.95) 2.13 (0.00, 5.04) 5.00 (3.47, 6.53) NOTE: s are from random effects analysis Fig. 2 Pooled results of the number of patients converted to general anesthesia for cesarean delivery with a labor epidural catheter in situ. This includes patients previously excluded from analysis in the original publication. CI = confidence interval. Proportion who Received a Second Anesthetic Year Total Event Rate in % Study Published secondanes N percent (95% CI) Gaiser (0.00, 3.33) 8.12 Albright (2.78, 6.24) 8.10 Garry (8.43, 12.61) 7.98 Riley (6.73, 14.41) 7.21 Tortosa (0.35, 4.81) 7.93 Pan (5.52, 7.81) 8.24 Kan (5.23, 8.65) 8.10 Orbach-Zinger (13.44, 29.27) 4.98 Kinsella (17.39, 21.55) 7.99 Campbell (8.90, 13.00) 8.00 Lee (1.33, 3.13) 8.29 Halpern (3.91, 8.07) 7.99 Bamgbade (0.17, 8.34) 7.08 Overall (I-squared = 96.6%, p = 0.000) 7.74 (4.96, 10.52) NOTE: s are from random effects analysis Fig. 3 Pooled results of the number of patients receiving a second anesthetic (including general anesthesia, spinal, epidural catheter replacement) for cesarean delivery with a pre-existing labor epidural catheter in situ. CI = confidence interval.

9 302 Conversion of epidural analgesia to anesthesia Proportion who Received Supplementation Study Year Published Received Not Received Event Rate in percent (95% CI) % Gaiser (8.04, 21.32) Tortosa (9.30, 19.27) Kinsella (10.30, 14.14) Halpern (1.44, 4.51) Overall (I squared = 95.6%, p = 0.000) (4.17, 17.26) Fig. 4 Pooled results of the patients receiving supplementation, when reported, for cesarean delivery. This includes intravenous narcotics, benzodiazepines, ketamine, nitrous oxide, inhalational agent, and not otherwise specified supplementation. CI = confidence interval. Analgesic Top ups study Number of Top ups OR (95% CI) % Campbell > (1.60, 3.50) Halpern > (1.38, 6.85) Lee > (2.47, 17.87) Overall (I squared = 46.6%, p = 0.154) 3.17 (1.83, 5.46) Fig. 5 Pooled results of patients receiving analgesic top-ups during labor and the effect on failed anesthesia conversion. CI = confidence interval; OR = odds ratio.

10 M.E. Bauer et al. 303 Emergency versus Urgency Study Year Published Emergency GA Emergency NAB Urgency GA Urgency NAB OR (95% CI) % Albright (7.98, ) Garry (47.42, ) Riley (3.81, ) Orbach Zinger (0.92, ) Kinsella (3.59, 12.15) Lee (36.58, ) Halpern (1.42, 8.53) Overall (I squared = 82.0%, p = 0.000) (8.81, ) Higher GA with Emergency Fig. 6 Pooled results of patients requiring emergent cesarean delivery with a labor epidural catheter in situ and the effect on general anesthesia for cesarean delivery. All unplanned cesarean deliveries were considered urgent and emergent cesarean deliveries were classified as such based on author s description in the original publication. CI = confidence interval; OR = odds ratio. Non specialist versus Obstetric Anesthesiologist Study OR (95% CI) % Campbell 4.76 (1.45, 15.59) Riley (0.97, 18.89) Overall (I squared = 0.0%, p = 0.913) 4.56 (1.81, 11.54) Fig. 7 Pooled effect of a non-specialist anesthesiologist compared with an obstetric anesthesiologist managing labor and delivery on the rate of failed epidural anesthesia conversion. CI = confidence interval; OR = odds ratio.

11 304 Conversion of epidural analgesia to anesthesia Fig. 8 Pooled results of the weight in kilograms at delivery and the effect on failed epidural anesthesia conversion. SMD = standardized mean difference; CI = confidence interval. Duration of Epidural Analgesia Study Year Published SMD in hours (95% CI) % Riley ( 0.29, 0.62) Tortosa ( 0.83, 0.06) Orbach Zinger ( 0.27, 0.71) Lee 0.67 (0.19, 1.15) Campbell 0.02 ( 0.17, 0.22) Halpern 0.59 ( 0.96, 0.22) Overall (I squared = 78.6%, p = 0.000) 0.02 ( 0.34, 0.31) Shorter duration increases risk Longer duration increases risk Fig. 9 Pooled effects of the duration of labor analgesia and the effect on failed epidural anesthesia conversion. SMD = standardized mean difference; CI = confidence interval.

12 M.E. Bauer et al. 305 Duration of epidural analgesia Six studies reported the duration of epidural analgesia before attempted conversion to anesthesia as a possible risk factor for failed conversion Random-effects meta-analysis suggests no statistically significant difference in the duration of epidural analgesia among those with successful or failed conversion of epidural analgesia to anesthesia (overall standardized mean difference 0.02 h, 95% CI 0.34 to 0.31; considerable statistical heterogeneity was present, I 2 = 78.6%, P < ; Fig. 9). Combined spinal epidural analgesia versus epidural analgesia Two studies compared the proportion of patients with successful conversion of CSE analgesia versus standard epidural analgesia with epidural anesthesia for cesarean delivery. 15,17 Lee reported a large increase in the rate of failed conversion with epidural versus CSE analgesia (OR = 5.54, 95% CI , P = 0.001). 15 Riley was not able to replicate a similar effect size. 17 Random-effects meta-analysis of these two studies yields insufficient evidence to suggest that either technique is more successful in converting epidural analgesia to anesthesia for cesarean delivery; statistical heterogeneity was present with I 2 value = 73.2%, P = (Fig. 10). 17 Cervical dilation Orbach-Zinger and Lee both reported cervical dilation at the start of labor epidural analgesia. Neither group found this to be associated with failed epidural conversion. 15,16 Halpern did not find an association between a smaller cervical dilation measurement before surgery and increased rate of failure of conversion (OR = 0.84, 95% CI , P = 0.055). 14 Quality evaluation As described in the Methods, two reviewers independently applied a quality of reporting score (STROBE Statement) for the articles chosen for review (Table 3). The kappa statistic for quality of reporting was % CI Discussion Of the seven investigated risk factors associated with failed conversion of epidural analgesia to anesthesia, an increased risk was found with an increased number of clinician-administered boluses during labor, an increased urgency of cesarean delivery, and provision of anesthetic care by a non-obstetric anesthesiologist. There is insufficient evidence to support an association between failed conversion and CSE as opposed to epidural analgesia, duration of labor epidural analgesia, cervical dilation at the time of initiation of labor analgesia, or BMI or weight at the time of delivery. In patients in whom an epidural catheter had been inserted for labor analgesia, general anesthesia was used for cesarean delivery in 5.0% of patients reported 95% CI %, a second anesthetic was used in 7.7% CSE versus Standard Epidural Technique Study Year Published OR (95% CI) % Lee (0.07, 0.48) Riley 0.85 (0.24, 3.03) Overall (I squared = 73.2%, p = 0.054) 0.37 (0.08, 1.75) lower GA with CSE higher GA with CSE Fig. 10 Pooled effects of combined spinal epidural technique compared with epidural technique on the rate of failed epidural anesthesia conversion. CI = confidence interval; CSE = combined spinal epidural; GA = general anesthesia; OR = odds ratio.

13 306 Conversion of epidural analgesia to anesthesia Table 3 Frequency of article reporting of items in the STROBE statement 21 STROBE items Number of studies of total possible Methods 1 Key elements of study design 12 of 13 2 Details of the setting of the study 13 of 13 3 Participants 12 of 13 4 Definition of all variables considered for and included in the analysis 12 of 13 5 Data sources/measurement 12 of 13 6 Efforts to address bias 3 of 13 7 Study size 11 of 13 8 Handling of quantitative variables 9 of 13 9 Statistical methods a. Details of all statistical methods used 12 of 13 b. For subgroups and interactions 2 of 13 c. For missing data 0 of 13 d. For loss to follow-up/matching/sampling strategy * e. Sensitivity analysis 1 of 13 Results 10 Participants a. Reporting of the number of individuals at each stage of the study 10 of 13 b. Reasons for non-participation at each stage 1 of 13 c. Use of a flow diagram 2 of 2 11 Descriptive data a. Characteristics of study participants 9 of 13 b. Indicating the number of participants with missing data 0 of 13 c. Summarize follow-up time * 12 Outcome data 10 of Main results a. Giving estimates and their precision 3 of 13 b. Reporting of category boundaries * c. absolute risk of a meaningful time period * 14 Other analysis done 3of13 STROBE = Strengthening the Reporting of Observational Studies in Epidemiology. * These items were not applicable to any of the 13 studies. 95% CI %, and intravenous or inhalational supplementation was used in 10.7% 95% CI %. There was substantial heterogeneity of all variables with an I 2 of >90%, the result of variation in practice patterns in labor analgesia management and conversion of labor epidural analgesia to anesthesia for cesarean delivery. The rate of general anesthesia for cesarean delivery has been proposed as a quality and safety measure for obstetric anesthesia. Reports that indicate the rate of general anesthesia without reporting the need for a second neuraxial anesthetic procedure, inhalational or intravenous supplementation may obscure the quality of anesthesia and analgesia delivered. Only reliable labor epidural analgesia administered via a functioning epidural catheter can be converted successfully to anesthesia for cesarean delivery. Substantial intravenous supplementation increases the risk of airway compromise. 3 Successful conversion without the need for supplementation or a second anesthetic is therefore considered a useful quality measure for analgesia and a safety measure for anesthesia. The Royal College of Anaesthetists guideline, suggesting that no more than 3% of women receiving labor epidural analgesia should require general anesthesia for caesarean delivery may be overly ambitious. Increased maternal pain during labor appears to be a risk factor for failed conversion to epidural anesthesia. Patients receiving unscheduled epidural bolus doses for breakthrough pain during labor, and who have high visual analog scores for pain within 2 h before cesarean delivery, have a higher rate of failed conversion to epidural anesthesia in five of the six studies that evaluated this outcome Breakthrough pain may be a marker for a poorly functioning epidural catheter. Pain may also signify dysfunctional labor and the need for obstetric intervention. 30 Increased pain with an epidural in place should prompt the provider to evaluate the patient. The authors are unaware of studies that have shown a decreased risk of epidural failure when epidural analgesia is replaced during labor. Lee reported that 21 of the 1025 catheters were replaced intrapartum before presentation for cesarean delivery; all catheters replaced were successfully converted to cesarean delivery anesthesia. 15 This finding, although small, represents an area for future study. Kinsella found that the greater the urgency for cesarean delivery, the higher the risk of failed conversion to

14 M.E. Bauer et al. 307 epidural anesthesia. 19 Halpern found no difference in failed epidural anesthesia in cesarean deliveries due to fetal heart rate abnormalities. 14 Many authors have described the use of general anesthesia with no attempt made to convert epidural analgesia to anesthesia due to the urgency of cesarean delivery. 13,14,17 19,23 This management may originate from the perception that it takes less time to induce general anesthesia than it does to convert epidural analgesia to anesthesia. This assumption may not be true when an epidural catheter providing highly satisfactory analgesia is combined with strategies to facilitate rapid anesthetic dosing and local anesthetic onset. In an audit by Popham of 444 category 1 cesarean deliveries, the time from decision to delivery for general anesthesia (17 ± 6 min) was not significantly different than the time for an epidural conversion (19 ± 9 min). 31 In a meta-analysis by Hillyard, the solution that offered the fastest onset for labor epidural conversion was 2% lidocaine with epinephrine. There were insufficient studies to assess the effect of bicarbonate. The addition of fentanyl lg to any solution accelerated the onset of surgical anesthesia. Ropivacaine 0.75% resulted in less intraoperative supplementation than bupivacaine or levobupivacaine, but there were no differences when compared with lidocaine. 32 The optimal solution that provides the fastest onset and is associated with the least intraoperative supplementation is unknown. Improved communication between obstetricians, nurses, and anesthesiologists for patients at high-risk of cesarean delivery, the placement of prophylactic epidural catheters (e.g., catheters placed before being requested or required), the initiation of epidural anesthetic administration in the labor room, and the use of appropriate anesthetic solutions may all facilitate successful conversion of epidural analgesia to anesthesia for even category 1 cesarean deliveries. 32,33 Anesthetic solutions administered during transfer to the operating room introduce the remote risk of inadvertent high block or seizure en route; signs and symptoms of these events must be assessed and balanced against the risk of complications resulting from emergency general anesthesia. 34 Both studies that evaluated a general versus obstetric anesthesiologist management found the rate of failed conversion to be higher with a general anesthesiologist. 13,17 Obstetric anesthesiologists may be more aware of the quality of labor epidural analgesia, may be more likely to replace dysfunctional catheters before cesarean delivery, and may use other maneuvers to avoid general anesthesia, such as manipulating the epidural catheter before drug administration, or performing another neuraxial technique. Campbell reported a success rate of 84.6% (22/26) by withdrawing the catheter 1 cm before further drug administration. 13 This may be due to the time this method takes, which allows development of an adequate sensory level, rather than the act of pulling back the catheter. Riley reported that obstetric anesthesiologists had more success than general anesthesiologists in conversion; he postulated that the former may be more likely to allow time for the development of an appropriate sensory level rather than opting for early use of general anesthesia. 17 Tortosa reported a conversion failure rate of 2.6% (5/194), which is less than the rate of 3% recommended by the Royal College of Anaesthetists. 11,12 This was attributed, in part, to obstetric anesthesiologists managing labor and delivery. 18 Obstetric anesthesiologists may be more aware of obstetric strategies to facilitate intrauterine resuscitation, may anticipate rapid changes in the planned mode of delivery, and may be able to capitalize on familiarity with their obstetric colleagues to determine the urgency of delivery. Only one study found BMI or weight to be a risk factor for failed conversion to epidural anesthesia. 16 The articles in this systematic review are limited by inconsistent definitions of obesity (Table 4). Obese women have higher rates of cesarean delivery, unfavorable airway examination, and difficult neuraxial block placement that may prompt more careful monitoring and management of epidural analgesia. 35 A reduced threshold to replace imperfectly functioning epidural catheters in laboring obese women may mitigate any increase in failed conversion of epidural analgesia to anesthesia. However, greater thickness of soft tissue between the skin surface and the ligamentum flavum increases the likelihood of catheter dislodgement with patient movement. Asking the parturient to assume a lateral decubitus position before taping the catheter appears to reduce this risk. 17,36 Table 4 Variation in reporting of body mass index or weight within included studies BMI or weight reported Definition Bamgbade () 24 BMI BMI at term Campbell () 13 BMI and weight not specified Halpern () 14 BMI and weight BMI and at time of delivery Kinsella (2008) 19 BMI and weight BMI and at booking Lee () 15 not specified Orbach-Zinger (2006) 16 BMI and weight Pre-pregnancy weight, weight at end of pregnancy, BMI at end of pregnancy Riley (2002) 17 not specified Tortosa (2003) 18 before delivery BMI = body mass index.

15 308 Conversion of epidural analgesia to anesthesia The relationship between the duration of epidural analgesia and the rate of successful conversion to epidural anesthesia was variable, but not statistically significant overall Heterogeneity of results exceeded that expected from statistical chance alone. For example, a prolonged labor with epidural analgesia may allow more time for the catheter to become dislodged. Conversely, the need for cesarean delivery following the induction of labor analgesia may allow insufficient time to determine if effective operative anesthesia can be achieved using the epidural route. There is insufficient evidence to suggest that either short or long duration of epidural analgesia is a risk factor for failed conversion to epidural anesthesia. In studies examining the effect of a CSE versus conventional epidural technique on successful conversion to epidural anesthesia, Lee indicated that the conventional epidural technique had a statistically higher risk for a failed conversion. 15 Riley reported comparable rates between groups. 17 Pan reported greater labor epidural analgesia failures with the conventional epidural technique compared to the CSE technique, but did not report the actual rates of failure of conversion to anesthesia. 28 The CSE technique may be a confirmatory test allowing for improved epidural space identification and subsequent catheter placement, or may allow a conduit for epidural medications to cross into the dural sac and augment the quality of blockade. 37 The findings of this study are limited by a number of considerations. First, the studies were observational trials, with some of retrospective design. Second, the definition of failed epidural anesthesia varied in each of the studies, and included being undefined, requiring intravenous or inhalational supplementation, or necessitating conversion to general anesthesia. Moreover, epidural catheter management was variable, with some reports involving routine removal of the epidural catheter, without first attempting to administer further drugs through it. 13,14,17,19 Campbell reported that no attempt at labor epidural conversion was made in 60/120 patients, and 59/60 of those remaining received a spinal. 13 Lastly, although statistical heterogeneity can sometimes be controlled by random effects analyses, this does not necessarily apply to true heterogeneity of different measures, end-points, and populations that may occur in these types of reports. We were unable to identify additional articles that evaluated the rate of successful conversion of epidural analgesia to anesthesia due to taping in the lateral as opposed to the sitting position, the use of the Trendelenberg versus flat supine position during anesthetic conversion, or the initiation of local anesthetic administration in the labor and delivery room versus the operating room for emergent cesarean delivery. We were also unable to find articles concerning the success of epidural analgesia conversion with the presence of dedicated staffing of the labor and delivery unit, the workload of the anesthesia providers, the use of scheduled visits to determine epidural analgesia function and quality, or the presence of team meetings with obstetricians to identify patients at risk for unplanned cesarean delivery. We conclude that an increased risk for failed conversion of epidural labor analgesia to cesarean delivery anesthesia was observed with an increasing number of clinician-administered boluses during labor, an enhanced urgency for cesarean delivery, and the presence of a non-obstetric anesthesiologist providing anesthetic care. Insufficient evidence is available to evaluate the influence of type of neuraxial (epidural versus CSE) technique, duration of labor analgesia, cervical dilation at the time of epidural placement, and BMI or weight on failed epidural conversion. Further high quality studies are needed to evaluate the many potential risk factors associated with failed conversion of labor epidural analgesia to cesarean delivery anesthesia. Disclosure This study was investigator-initiated, independently conducted, analyzed, and reported by the study investigators who have no conflicts of interest to disclose. Internal funding only, from the University of Michigan, Department of Anesthesiology, Ann Arbor, MI, USA. Acknowledgements The authors are grateful to Bright Kim BS, Khalil Mroue BS, and Lance Patak MD for their invaluable support for this project and to Mark MacEarnern for his expertise and instruction in the use of EMBASE, PubMed, and OvidSP. References 1. Algert CS, Bowen JR, Giles WB, Knoblanche GE, Lain SJ, Roberts CL. Regional block versus general anaesthesia for caesarean section and neonatal outcomes: a population-based study. BMC Med ;7: Hodgson CA, Wauchob TD. A comparison of spinal and general anaesthesia for elective caesarean section: effect on neonatal condition at birth. Int J Obstet Anesth 1994;3: Mhyre JM, Riesner MN, Polley LS, Naughton NN. A series of anesthesia-related maternal deaths in Michigan, Anesthesiology 2007;106: Djabatey EA, Barclay PM. Difficult and failed intubation in 3430 obstetric general anaesthetics. Anaesthesia ;64: McDonnell NJ, Paech MJ, Clavisi OM, Scott KL. Difficult and failed intubation in obstetric anaesthesia: an observational study of airway management and complications associated with general anaesthesia for caesarean section. Int J Obstet Anesth 2008;17: Mhyre JM, Healy D. The unanticipated difficult intubation in obstetrics. Anesth Analg 2011;112:

16 M.E. Bauer et al Furst SR, Reisner LS. Risk of high spinal anesthesia following failed epidural block for cesarean delivery. J Clin Anesth 1995;7: Portnoy D, Vadhera RB. Mechanisms and management of an incomplete epidural block for cesarean section. Anesthesiol Clin North America 2003;21: The Australian Council on Healthcare Standards. Australasian Clinical Indicator Report 2001 : Determining the Potential to Improve. 11th ed. [accessed April 2012]. 10. Haller G, Stoelwinder J, Myles PS, McNeil J. Quality and safety indicators in anesthesia: a systematic review. Anesthesiology ;110: Russell I. Technique of anaesthesia for cesarean section. In: Colvin J, editor. Raising the standard: a compedium of audit recipes. London: The Royal College of Anaesthetists; p Russell I. Technique of anaesthesia for caesarean section. In: Colvin J, editor. Raising the standard: a compedium of audit recipes for continuous quality improvement in anaesthesia. 2nd ed. London: The Royal College of Anaesthetists; p Campbell DC, Tran T. Conversion of epidural labour analgesia to epidural anesthesia for intrapartum cesarean delivery. Can J Anesth ;56: Halpern SH, Soliman A, Yee J, Angle P, Ioscovich A. Conversion of epidural labour analgesia to anaesthesia for caesarean section: a prospective study of the incidence and determinants of failure. Br J Anaesth ;102: Lee S, Lew E, Lim Y, Sia AT. Failure of augmentation of labor epidural analgesia for intrapartum cesarean delivery: a retrospective review. Anesth Analg ;108: Orbach-Zinger S, Friedman L, Avramovich A, et al. Risk factors for failure to extend labor epidural analgesia to epidural anesthesia for cesarean section. Acta Anaesthesiol Scand 2006;50: Riley ET, Papasin J. Epidural catheter function during labor predicts anesthetic efficacy for subsequent cesarean delivery. Int J Obstet Anesth 2002;11: Tortosa JC, Parry, Mercier FJ, Mazoit JX, Benhamou D. Efficacy of augmentation of epidural analgesia for caesarean section. Br J Anaesth 2003;91: Kinsella SM. A prospective audit of regional anaesthesia failure in 5080 caesarean sections. Anaesthesia 2008;63: Vandenbroucke JP, von Elm E, Altman DG, et al. Strengthening the reporting of observational studies in epidemiology (STROBE): explanation and elaboration. PLoS Med 2007;4:e von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP. The Strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Med 2007;4:e Papathanasiou AA, Zintzaras E. Assessing the quality of reporting of observational studies in cancer. Ann Epidemiol 2010;20: Albright GA, Forster RM. The safety and efficacy of combined spinal and epidural analgesia/anesthesia (6002 blocks) in a community hospital. Reg Anesth Pain Med 1999;24: Bamgbade OA, Khalaf WM, Ajai O, Sharma R, Chidambaram V, Madhavan G. Obstetric anaesthesia outcome in obese and nonobese parturients undergoing caesarean delivery: an observational study. Int J Obstet Anesth ;18: Gaiser RR, Cheek TG, Gutsche BB. Epidural lidocaine versus 2-chloroprocaine for fetal distress requiring urgent cesarean section. Int J Obstet Anesth 1994;3: Garry M, Davies S. Failure of regional blockade for caesarean section. Int J Obstet Anesth 2002;11: Kan RK, Lew E, Yeo SW, Thomas E. General anesthesia for cesarean section in a Singapore maternity hospital: a retrospective survey. Int J Obstet Anesth 2004;13: Pan PH, Bogard TD, Owen MD. Incidence and characteristics of failures in obstetric neuraxial analgesia and anesthesia: a retrospective analysis of deliveries. Int J Obstet Anesth 2004;13: Lucas DN, Yentis SM, Kinsella SM, et al. Urgency of caesarean section: a new classification. J R Soc Med 2000;93: Panni MK, Segal S. Local anesthetic requirements are greater in dystocia than in normal labor. Anesthesiology 2003;98: Popham P, Buettner A, Mendola M. Anaesthesia for emergency caesarean section, , at the Royal Women s Hospital, Melbourne. Anaesth Intensive Care 2007;35: Hillyard SG, Bate TE, Corcoran TB, Paech MJ, O Sullivan G. Extending epidural analgesia for emergency caesarean section: a meta-analysis. Br J Anaesth 2011;107: Palanisamy A, Mitani AA, Tsen LC. General anesthesia for cesarean delivery at a tertiary care hospital from 2000 to 2005: a retrospective analysis and 10-year update. Int J Obstet Anesth 2011;20: Rahman K, Jenkins JG. Failed tracheal intubation in obstetrics: no more frequent but still managed badly. Anaesthesia 2005;60: Mhyre JM. Anesthetic management for the morbidly obese pregnant woman. Int Anesthesiol Clin 2007;45: Hamilton CL, Riley ET, Cohen SE. Changes in the position of epidural catheters associated with patient movement. Anesthesiology 1997;86: Cappiello E, O Rourke N, Segal S, Tsen LC. A randomized trial of dural puncture epidural technique compared with the standard epidural technique for labor analgesia. Anesth Analg 2008;107: Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at /j.ijoa