SPECIAL ARTICLE

Effects of Preemptive Epidural Analgesia on Post-thoracotomy Pain

Choon Looi Bong, MBChB, FRCA,* Miny Samuel, MSc, PhD, Ju Mei Ng, MBBS, FANZCA, and Chris Ip-Yam, MBChB, FFARCSI, FRCA, FAMS, MBA
Objective: The purpose of this study was to determine whether preemptive thoracic epidural analgesia (TEA) initiated before surgical incision would reduce the severity of acute post-thoracotomy pain and the incidence of chronic post-thoracotomy pain. Method: Meta-analysis of randomized controlled trials (RCTs) Search Strategy: MEDLINE, the Cochrane Central Register of Controlled Trials (CENTRAL) and EMBASE were searched from 1966 to December 2004 for prospective RCTs published in all languages using the following MeSH terms: post-thoracotomy pain, epidural analgesia, chronic pain, and preemptive analgesia. Selection Criteria: All RCTs that compared thoracic epidural analgesia initiated before surgical incision (preemptive group) versus thoracic epidural analgesia initiated after completion of surgery (control group) in adult patients undergoing unilateral thoracotomy. Measurements and Main Results: Three authors reviewed all citations and simultaneously extracted data on sample size, patient characteristics, surgical and analgesic interventions, methods of pain assessment, and pain scores at 24 hours, 48 hours, and 6 months postoperatively. Six studies were included with a total of 458 patients. Pooled analyses indicated that preemptive TEA was associated with a statistically signicant reduction in the severity of acute pain on coughing at 24 and 48 hours (weighted mean difference 1.17 [95% condence interval (CI) 1.50 to 0.83] and 1.08 [95% CI 1.17 to 0.99]), respectively. Acute pain was a good predictor of chronic pain. However, there was no statistically signicant difference in the overall incidence of chronic pain at 6 months between the preemptive TEA group (39.6%) and the control group (48.6%). Conclusion: Preemptive TEA appeared to reduce the severity of acute pain but had no effect on the incidence of chronic pain. 2005 Elsevier Inc. All rights reserved. KEY WORDS: thoracic epidural, post-thoracotomy pain, preemptive analgesia

HORACOTOMY IS WIDELY recognized as being one of the most painful surgical procedures. Acute post-thoracotomy pain is aggravated by the constant movement of breathing. Adequate relief of acute post-thoracotomy pain is therefore essential to facilitate coughing, deep breathing, and early mobilization, all of which contribute to an enhanced postoperative outcome. Chronic post-thoracotomy pain refers to pain that persists along the thoracotomy scar for more than 2 months after surgery1 and is reported to have an incidence between 50% to 80%.2,3 Chronic postthoracotomy pain is particularly severe and is likely to last several years after surgery.4 Katz et al5 showed that the intensity of early postoperative pain correlates with that of chronic post-thoracotomy pain. This has led to the hypothesis that reducing acute

From the *Department of Paediatric Anesthesia, KK Women and Childrens Hospital, Singapore; Clinical Trials and Epidemiology Research Unit, Ministry of Health, Singapore; and Department of Anesthesia and Surgical Intensive Care, Singapore General Hospital, Singapore. An abstract of this review was presented as a poster at the Society of Cardiovascular Anesthesiologists 26th Annual Meeting and Workshops, April 26-27, 2004, Honolulu, HI. Address reprint requests to Choon Looi Bong, MBChB, FRCA, Department of Paediatric Anesthesia, KK Women and Childrens Hospital, 100 Bukit Timah Road, Singapore 229899. E-mail: cchia@ doctors.org.uk 2005 Elsevier Inc. All rights reserved. 1053-0770/05/1906-0019$30.00/0 doi:10.1053/j.jvca.2005.08.012
786

post-thoracotomy pain with preemptive analgesia would result in a lower incidence of chronic post-thoracotomy pain.6,7 Previous studies have shown that thoracic epidural analgesia (TEA) is superior to conventional modes of analgesia in reducing early post-thoracotomy pain.8,9 In practice, many centers, including this one, routinely provide TEA for thoracotomy surgery. In the authors center, the anesthesia technique involves insertion of a thoracic epidural catheter at the T5-T8 level before induction of general anesthesia. A bolus of bupivacaine or ropivacaine is given via the catheter, usually before surgical incision to provide intraoperative anesthesia and analgesia. Subsequently, an infusion of ropivacaine and fentanyl is commenced before the end of surgery to provide postoperative analgesia. This analgesic regimen is effective in preventing acute post-thoracotomy pain in the vast majority of patients. However, it is not known whether TEA initiated before surgical incision (preemptive analgesia) is any more effective than that initiated after surgery, in terms of preventing acute and chronic post-thoracotomy pain. The aim of this systematic review was to determine whether preemptive TEA initiated before surgical incision reduced the severity of acute post-thoracotomy pain, as well as the incidence of chronic post-thoracotomy pain, when compared with TEA initiated after completion of surgery.
METHODS A thorough literature search was conducted in the following electronic databases: MEDLINE using Ovid (1966-December 2004), MEDLINE using PubMed (1966-December 2004), EMBASE using Dialog Datastar (1974-December 2004), the Cochrane Central Register of Controlled Tri-

Journal of Cardiothoracic and Vascular Anesthesia, Vol 19, No 6 (December), 2005: pp 786-793

PREEMPTIVE EPIDURAL ANALGESIA

787

Table 1. Appraisal of the Quality of the Studies

Study No. of Subjects Priori Power Analysis Randomized? Allocation concealment Blinding Intention-to-treat Analysis Follow-up

Aguilar et al11

65

Yes

Yes

Yes

Patients Physicians Assessors

No

Obata et al6

70

Yes

Yes

Not stated

Sentrk et al7

85

No

Yes

Not stated

Patients Physicians Assessors Physicians Assessors Patients Physicians Assessors Patients Physicians

No

No

Ochroch et al12

157

Yes

Yes

Yes

Yes

Neustein et al13

40

Yes

Yes

No

Yes

Yegin et al14

61

No

Yes

Not stated

Assessors

Yes

8 of 45 patients (17.8%) withdrawn from study. Patients replaced until number 45 12 of 70 patients (17.1%) excluded from analysis 16 of 85 patients (18.8%) excluded from analysis 37 of 157 patients (23.6%) excluded from analysis 8 of 40 patients (20%) randomized did not undergo thoracotomy so excluded from analysis. Complete

als, Issue 3, 2004 for prospective randomized controlled trials comparing the analgesic effects of TEA initiated before surgical incision with that of TEA initiated after completion of surgery. The MeSH terms used were as follows: post-thoracotomy pain, epidural analgesia, chronic pain, and preemptive analgesia. The search included articles published in all languages. After this, a careful hand-search was done from the reference list of all relevant articles. Randomized controlled trials (RCTs) were selected that compared TEA initiated before surgical incision with TEA after completion of surgery in adult patients undergoing unilateral thoracotomy. Primary outcomes sought were acute pain scores in the immediate postoperative period up to 48 hours after surgery and the incidence of chronic pain 6 months after surgery. Studies that only compared the effects of TEA with nonepidural modes of analgesia were excluded, as well as studies investigating the effects of lumbar epidural analgesia for upper abdominal surgery. Two of the authors (BCL, NJM) reviewed all citations independently and in duplicate. Any disagreements on the inclusion criteria were resolved by consensus after discussion with the third author (CI). All 3 authors then independently extracted the following data from the primary studies using a standard data collection form: author, quality of the trials, sample size, patient characteristics, details of the surgical and analgesic interventions, methods of assessing pain, acute pain scores in the immediate postoperative period, and chronic pain scores at 6 months postoperatively. The quality of a trial was determined by whether there was randomization and allocation concealment. Other criteria used were blinding, intention-to-treat analysis, and loss to follow-up (Table 1). Where relevant data were unavailable in the published articles, the authors obtained the information directly from the corresponding authors via electronic mail. Outcome data on acute pain were expressed as continuous variables (mean and standard deviation of visual analog scale [VAS] or numeric rating scale [NRS] scores) and data on chronic pain as dichotomous variables (presence or absence of pain at 6 months). To estimate the treatment effects, continuous outcomes were calculated as weighted mean difference (WMD) and dichotomous outcomes as relative risk, with their respective 95% condence intervals. Pooled effect estimates

and heterogeneity between studies were tested with The Rev Man 4.2.3 statistical package (The Cochrane Collaboration, Oxford, UK). Where signicant result heterogeneity was found with the xed-effects model, a random-effects model was used to estimate the overall treatment effect. The electronic search identied 7 randomized controlled trials6,7,10-14 that were published between 1996 to 2004. Only 6 of these studies fullled the inclusion criteria. The study by Subramanian et al10 was a well-designed prospective randomized study looking at preemptive analgesia with epidural morphine or morphine and bupivacaine. The authors studied 80 patients undergoing both upper abdominal and thoracic surgery who had epidural catheters inserted at the lumbar level and advanced to thoracic levels. However, only 9 of these 80 patients underwent thoracic procedures, and there were no specic data on either acute or chronic post-thoracotomy pain. Hence, this study was excluded from the analysis. The 6 remaining studies6,7,11-14 all included patients undergoing thoracotomy who had epidurals inserted at the upper thoracic levels, and follow-up assessment of post-thoracotomy pain was conducted for at least 48 hours postoperatively. Assessment of acute pain was done on a 10-point NRS where 0 no pain and 10 maximum pain experienced in 1 study7 and a 10-point VAS from 0 cm no pain to 10 cm worst pain imaginable in all other studies.6,11-14 In the study by Aguilar et al,11 45 patients undergoing thoracotomy for lung resection were randomized into 1 of 3 groups and studied prospectively in a double-blind, placebo-controlled, cross-over design. In the preemptive analgesia group, 15 patients received 8 mL of 0.5% bupivacaine with epinephrine 1/200,000 through a thoracic epidural catheter (tip at T3-5) 30 minutes before skin incision and saline 8 mL 15 minutes after skin incision. In the control group, 15 patients received 8mL of epidural saline 30 minutes before skin incision and 8 mL of 0.5% bupivacaine with epinephrine 1/200,000 through the epidural catheter 15 minutes after skin incision. In the placebo group, 15 patients received epidural saline both before and after skin incision. The intraoperative anesthesia technique for all 3 groups included an alfentanil infusion at 1 g/kg/min and alfentanil boluses of 5 g/kg every 5 minutes if arterial pressure or heart rate exceeded 125% of

788

BONG ET AL

Table 2. Summary of Individual Studies

Study Subjects Intervention/Control Epidural Level Epidural Drugs Opioids Used? Outcomes Assessed

Aguilar et al11

45 patients randomized to 3 groups

Obata et al6

70 patients randomized to 2 groups

Sentrk et al7

112 patients randomized into 3 groups to evaluate 3 different analgesia techniques

Ochroch et al12

157 patients randomly assigned into groups

Intervention group: TEA initiated with bolus of bupivacaine 30 min before skin incision. saline bolus 15 min after incision Control group: saline bolus 30 min before skin incision. TEA initiated with bolus of bupivacaine 15 min after incision. Placebo group: Epidural saline only before and after surgical incision Intervention group: TEA initiated with mepivacaine 20 min before the start of surgery Control group: TEA initiated with mepivacaine only at completion of surgery Intervention group (preTEA): TEA bolus 30 min preinduction continued intraoperatively by infusion and postoperatively by PCEA Control group (post-TEA): TEA bolus given only after extubation postoperative PCEA IV-PCA group: 23 patients had no TEA but only IV morphine PCA after surgery Intervention group: TEA initiated after induction of anesthesia. Further dose at time of rib approximation Control Group: TEA initiated only at time of rib approximation

T3-T5

Intraoperative: bolus 8 mL 0.5% bupivacaine with adrenaline 5 g/mL and intravenous infusion and boluses of alfentanil. Postoperative: infusion of 0.125% bupivacaine and fentanyl 5 g mL

Yes

1. VAS scores at rest, mobilization and coughing during rst 48 hours 2. No. of successful PCEA demands 3. Complications

T5-T6

Intraoperative: 4 mL 1.5% mepivacaine followed by continuous infusion Postoperative: 1.5% mepivacaine infusion only, no opioids used.

No

1. VAS scores at 4 hours 1-7 days postoperative 2. Pain at 3 and 6 months measured by NRS

T7-T8

Intraoperative: Bolus 10 ml 0.1% bupivacaine with morphine 1 mg followed by continuous infusion Postoperative: PCEA with 0.1% bupivacaine and morphine 0.05 mg/mL

Yes

1. VAS scores on movement and cough at 0, 4, 8, 12, 24, and 48 hours postoperative 2. Pain lasting at least 2 months and pain at 6 months measured using NRS

T6-T8

Intraoperative: Bolus 6 mL of 0.375% bupivacaine with 3 g/mL fentanyl followed by infusion Postoperative: 5 mL of 0.5% bupivacaine with fentanyl 50 g followed by PCEA

Yes

Neustein et al13

40 patients randomly assigned to 2 groups

Yegin et al14

61 patients randomized into 2 groups

Intervention group: TEA initiated before skin incision. Further dose at chest closure Control group: TEA initiated only at time of skin closure Intervention group: TEA initiated 30 min before induction of anesthesia. Further dose at chest closure. Control group: TEA initiated at chest closure

Not stated

Intraoperative: bolus 8 mL 0.25% bupivacaine with fentanyl 100 mg Postoperative: Infusion of 0.1% bupivacaine with fentanyl 10 mg/mL Intraoperative: 8 mL 0.25% bupivacaine and fentanyl 50 g Postoperative: 8 mL 0.25% bupivacaine and fentanyl 50 g followed by PCEA with bupivacaine and fentanyl

Yes

1. Total analgesic use in acute postoperative period 2. VAS scores at 0, 1, 2, 3, 4, and 5 days postoperative 3. Pain and physical activity scores at 1, 2, 3, 4, and 6 months postoperative 1. VAS scores for rst 48 hours postoperative

T6-T8

Yes

1. VAS scores at rest and coughing at 2, 4, 8, 12, 24 and 48 hours postoperative 2. PCEA fentanyl consumption over rst 24 hours

PREEMPTIVE EPIDURAL ANALGESIA

789

baseline. The alfentanil infusion was stopped before chest closure and fentanyl, 50 g, was given through the epidural catheter. Postoperative analgesia was maintained with patient-controlled epidural analgesia (PCEA) using 0.125% bupivacaine with epinephrine 1/400,000. Eight patients (20%) were withdrawn from the study because of inadequate analgesia despite PCEA because of ipsilateral shoulder pain (n 5), loss of epidural catheter function before the end of the study (n 1), and incomplete data collection (n 2). The patients withdrawn were replaced until 45 patients had completed the study. This constituted a signicant source of bias for such a small study. In the study by Obata et al,6 70 patients scheduled for thoracic surgery were randomized to 2 groups. In the preemptive analgesia group, 28 patients received 4 mL of 1.5% mepivacaine initiated 20 minutes before the start of surgery, followed by continuous epidural analgesia until 72 hours after surgery. In the control group, 30 patients received 4 mL of 1.5% mepivacaine only at completion of surgery. Indomethacin was the only supplementary analgesic, and no opioids were allowed in either group throughout the perioperative period. A total of 12 of 70 patients were excluded from the study because of catheter dislodgement (n 2), loss to follow-up (n 5), death (n 6), or recurrence of malignant tumor (n 2), leading to a dropout rate of 17.1%. In the study by Sentrk et al,7 112 ASA 2-3 patients undergoing unilateral thoracotomy were randomly divided into 3 groups to evaluate 3 different analgesia techniques. In the preemptive analgesia group (pre-TEA), 28 patients each had an epidural catheter inserted through the T7-8 intervertebral space and received an epidural bolus of 10 mL of 0.1% bupivacaine with 0.1 mg/mL of morphine initiated more than 30 minutes before anesthetic induction. This was followed by an epidural infusion of 7 mL/h of the same solution intraoperatively. Analgesia was continued postoperatively by PCEA. In the control group (post-TEA), 29 patients each had an epidural catheter inserted through the T7-8 intervertebral space but received the initial bolus of 10 mL of 0.1% bupivacaine with 0.1 mg/mL of morphine only after extubation. Postoperative analgesia was similar to group pre-TEA. In the intravenous PCA group, 28 patients received no TEA but had intravenous PCA with morphine after surgery. Fentanyl, 2 g/kg, was used for induction of anesthesia in all 3 groups. No further fentanyl was required for the pre-TEA group, whereas fentanyl boluses of 5 to 10 g/kg were given to the other 2 groups intraoperatively. Acute pain scores as assessed by NRS showed that postoperative pain on movement and cough in the pre-TEA group were signicantly less than the post-TEA or intravenous PCA groups at 0, 4, 8, 12, 24, and 48 hours postoperatively. In the study by Ochroch et al,12 157 patients scheduled for thoracic surgery through a posterolateral or muscle-sparing incision each had an epidural catheter inserted through the T6-7 or T7-8 intervertebral space and were randomly assigned into 2 groups. In the preemptive analgesia group, 78 patients received intraoperative and postoperative TEA in the form of a 6-mL mixture of 0.375% bupivacaine and 3 g/mL of fentanyl before surgical incision, followed by an epidural infusion at 8 mL/h of the same mixture. In the control group, 79 patients received 6 mL of saline followed by an infusion of epidural saline at 8 mL/h. Both groups received fentanyl, 3 g/kg, at induction and an epidural bolus dose of 5 mL of 0.5% bupivacaine and 50 g of fentanyl at the time of rib approximation. Postoperatively, both groups received PCEA with 0.05% bupivacaine and 10 g/mL of fentanyl. In the study by Neustein et al,13 40 patients scheduled for elective thoracic surgery (thoracotomy or thoracoscopy with possible thoracotomy) were randomized to 2 groups. In the preemptive analgesia group, 20 patients received 8 mL of bupivacaine 0.25% and 2 mL of fentanyl (50 g/mL) via the epidural catheter after induction of anesthesia and before skin incision. This was followed by an infusion of bupivacaine 0.1% and fentanyl 10 g/mL at 6 mL/h until chest closure. In the control group, 20 patients received 10 mL of normal saline injected via

the epidural catheter before skin incision followed by an infusion of normal saline at 6 mL/h after chest closure. At the beginning of chest closure, patients in both groups received 8 mL of bupivacaine 0.25% and 2 mL of fentanyl (50 g/mL) thorough the epidural catheter. Postoperatively, an epidural infusion of bupivacaine 0.1% and fentanyl (10 g/mL) was begun at 2 mL/h and titrated according to the needs of the patient. Eight patients were withdrawn from the study when a thoracotomy was not performed. The remaining patients all completed the protocol. In the study by Yegin et al,14 61 patients undergoing posterolateral thoracotomy were randomized into 2 groups. In the preemptive analgesia group (pre-op TEA), 30 patients received 8 mL of 0.25% bupivacaine and fentanyl (50 g/2 mL) via the epidural route at least 30 minutes before induction of anesthesia, followed by a further dose of the same solution at the time of pleural closure. In the control group, (post-op TEA), 31 patients received intravenous fentanyl intraoperatively and 8 mL of 0.25% bupivacaine and fentanyl (50 g in 2 mL) via the epidural route only at the time of pleural closure. Postoperatively, both groups received PCEA using an identical protocol. RESULTS

The 6 studies included a total of 458 (American Society of Anesthesiologists 2 to 3) patients presenting for elective unilateral open thoracotomies (Table 2). The study by Aguilar et al11 included 30 patients, but because of inadequate reporting of pain scores (no standard deviation available), data could not be pooled with the other studies for statistical analysis. Because of various exclusion criteria specic to each individual study, as well as loss to follow-up, eventually only 355 patients who had adequate data were subjected to statistical analysis. Acute postoperative pain at rest in the rst 24 and 48 hours after surgery was assessed using 10-point scores (0 no pain to 10 worst pain) by all studies, and these were subject to a metaanalysis (Fig 1). At 24 hours, the pooled mean difference showed a trend toward a reduction in acute postoperative pain in the preemptive group (WMD 0.27 [95% CI, 0.91 to 0.37]), although this was not statistically signicant. Aguilar et als study,11 which could not be pooled with the other studies, reported the mean pain score at rest at 24 hours to be slightly less in the control group (2.0) compared with the preemptive group (2.4), but again this difference was not statistically signicant. However, at 48 hours, the pooled estimate showed a signicant reduction in acute postoperative pain in the preemptive group (WMD 0.59 [95% CI, 1.14 to 0.04]). The authors noted signicant result heterogeneity between studies for pain at rest at 24 and 48 hours (I2 98.6%, I2 98.2%, respectively) and performed a sensitivity analysis by excluding Obatas study6 because it was the only study in which no opioid was used. After excluding this study, there was no difference in the mean pain score reported in both preemptive and the control groups at 24 and at 48 hours (at 24 hours WMD 0.04 [95% CI, 0.66 to 0.73]; at 48 hours WMD 0.08 [95% CI, 0.30 to 0.14]). Only the studies by Sentrk et al7 and Yegin et al14 included scores on acute postoperative pain on coughing in the rst 24 and 48 hours. The overall pooled analysis showed a statistically signicant reduction in pain on coughing in the intervention group at 24 hours (WMD 1.17 [95% CI 1.50 to 0.83]) and at 48 hours (WMD 1.08 [95% CI, 1.17 to 0.99]) (Fig 2). There was no signicant result heterogeneity observed.

790

BONG ET AL

Fig 1.

Acute pain at rest.

Only the studies by Obata et al,6 Sentrk et al,7 and Ochroch et al12 included data for chronic pain at 6 months after surgery (Fig 3). These were expressed in the form of dichotomous data (the presence or absence of pain at 6 months), and it was possible to calculate the relative benet (RB) of preemptive analgesia on the incidence of pain at 6 months. Of 206 patients, the overall incidence of chronic pain at 6 months was 39.6% in the preemptive analgesia group versus 48.6% in the control group. Of the 3 studies, only Obata et al6 showed that initiation of epidural analgesia before surgical incision signicantly reduced the incidence of chronic post-thoracotomy pain. A metaanalysis of the 3 studies merely showed a trend toward a benet in the preemptive analgesia group, with a relative benet of 1.32 (95% CI, 0.76-2.3) in the preemptive group, which was not statistically signicant. Signicant result heterogeneity was observed among the studies (I2 73.6%), and therefore a random-effects model was used to estimate the overall pooled effect. A sensitivity analysis was performed to see if this heterogeneity was caused by Obata et als study because this was the only study in which opioids were not used. After exclusion of Obata et als study, there was no difference in the incidence of pain between the preemptive analgesia group and the control group (RB 1.05 [95% CI, 0.66-1.67]). Another interesting nding was that both the studies by Sentrk et al7 and Ochroch et al12 revealed that acute pain was a good predictor of chronic pain. Sentrk et al7 found that of the 29

patients who had an NRS 2 on the second postoperative day, 24 (83%) reported having chronic pain at 6 months, whereas only 19 (48%) of the 40 patients who had an NRS 2 had pain at 6 months (2 0.746, p 0.0063). Ochroch et al12 found that VAS scores on the rst postoperative day were predictive of nonzero average pain scores 24 weeks after surgery. Only 4 studies7,11,12,14 reported on adverse events. In general, there were very few adverse events and no epidural catheter related complications were reported (Table 3).
DISCUSSION

There is evidence that peripheral tissue injury leads to central nervous system changes that create hypersensitivity to pain postoperatively. This hypersensitivity is thought to be caused by a lower threshold to pain in the peripheral nociceptors, as well as an activity-dependent increase in excitability of dorsal horn neurons in the spinal cord, a process termed central sensitization. Preemptive analgesia is a concept that originated from basic research indicating that an analgesic intervention is more effective if given before, rather than after, a noxious stimulus.15,16 The aim is to prevent central sensitization by blocking afferent C-ber input from the periphery before the onset of a noxious stimulus. The benet of preemptive analgesia has been supported by some clinical studies using local anesthetics, opioids, and nonsteroidal anti-inammatory drugs.17,18 However, the clinical usefulness of preemptive analgesia has remained controversial,19-21 probably

PREEMPTIVE EPIDURAL ANALGESIA

791

Fig 2.

Acute pain on coughing.

because in part of the wide variation in study conditions such as surgery, drugs, doses, routes of administration, treatment duration, and pain assessment methods used in different studies. All these factors are of crucial importance in the evaluation of preemptive analgesia. Previous studies comparing the effects of preoperative and postoperative epidural analgesia in abdominal surgery have failed to show any benet of preemptive analgesia.20,21 This lack of benet was partly attributed to the less discrete, visceral nature of pain after abdominal surgery. Thoracotomy is known to produce high-intensity noxious stimuli sufcient to cause central sensitization,5,22,23 and the area of post-thoracotomy pain is more discrete and largely restricted to the site of surgery. Hence, any benet of preemptive epidural analgesia should, theoretically, be more apparent in thoracic surgery than in abdominal surgery. There are limited data available regarding the use of preemptive analgesia for thoracic surgery. The present extensive search only revealed 6 prospective RCTs specically comparing the analgesic effects of TEA with local anesthetic (with or without opioids) initiated before surgical incision to that initiated after completion of surgery. Five of 6 of these studies included opioids in their analgesic regimen, both intraoperatively and postoperatively. Obata et al6 performed the only study in which no opioid was used throughout the entire study period. Interestingly, this was the only study that showed that continuous epidural block with local anesthesia alone initiated before surgical incision reduced both the intensity of acute post-thoracotomy pain and the incidence of chronic post-thoracotomy pain.

This was in contrast to the studies by Sentrk et al,7 Aguilar,11 Ochroch,12 Neustein et al,13 and Yegin et al,14 which used opioids along with local anesthetics as their analgesic regimen. Conversely, all these studies failed to show a benet of preemptive epidural analgesia in preventing acute or chronic postthoracotomy pain, both when taken individually or collectively in the form of a meta-analysis. The authors postulate that the reason for this disparity is likely to be because of the use of opioids in the analgesic regimen. It has been suggested that administration of opioids as premedication or during induction and maintenance of anesthesia may prevent the development of central sensitization.19,24 One possible explanation is that the combination of opioids and local anesthesia in the studies by Sentrk et al,7 Aguilar et al,11 Ochroch et al,12 Neustein et al,13 and Yegin et al14 could have been so effective in pain relief that any potential benet of preemptive epidural analgesia could have been masked. This would also explain the ndings that patients in these studies reported generally lower pain scores compared with patients in the study by Obata et al.6 This systematic review shows the difculty in comparing different studies on pain management in which many conditions are not standardized, despite using very narrow search criteria. Care must be exercised when trying to draw reasonable conclusions from such a group of heterogenous studies, particularly when the patient numbers were relatively small and the dropout rates were high. In this review, evidence from the pooled analysis of 5 RCTs indicated that preemptive epidural analgesia was associated

792

BONG ET AL

Fig 3.

Number of pain-free patients at 6 months.

with a statistically signicant reduction in the severity of acute pain at rest at 48 hours (WMD 0.59 [95% CI, 1.14 to 0.04]) and on coughing at 24 and 48 hours (WMD 1.17 [95% CI, 1.50 to 0.83] and 1.08 [95% CI, 1.17 to 0.99]), respectively. Although there were only 3 studies with data on acute pain during coughing, the benecial effect of TEA was an important nding and may have a signicant impact on the incidence of postoperative respiratory complications as suggested in a recent meta-analysis by Ballantyne et al.25 In their meta-analysis of 24 studies, the authors found that compared with systemic opioids, epidural opioids decreased the incidence of atelectasis, and epidural local anesthetics in-

creased arterial oxygen tension and decreased the incidence of pulmonary infections and pulmonary complications overall. In conclusion, thoracotomy was associated with a frequent incidence of chronic post-thoracotomy pain (39.2%-48.6%). Pooled analysis of 3 RCTs showed that preemptive TEA was associated with a trend toward a reduction in the incidence of chronic post-thoracotomy pain, RB 1.32 (95% CI, 0.762.3), although this was not statistically signicant. Acute postoperative pain appeared to be a good predictor of chronic post-thoracotomy pain.7,12 By inference, prevention of acute pain could also be effective in prevention of chronic post-thoracotomy pain, although this was not conclusively

Table 3. Adverse Effects

Study Number of Subjects Adverse Effects

Aguilar et al11 Obata et al6 Sentrk et al7 Ochroch et al12

Pre-TEA n 15 Post-TEA n 15 Pre-TEA n 28 Post-TEA n 20 Pre-TEA n 22 Post-TEA n 24 Pre-TEA n 57 Post-TEA n 63

Authors reported no complications or adverse effects Not reported Pruritus: pre-TEA: n 2 (9%), post-TEA: n 2 (8%); IV-PCA: n 4 (17%) PONV: pre-TEA: n 1 (5%), post-TEA: n 1 (4%); IV-PCA: n 3 (13%) Reported on complications not specically related to TEA (eg, uid and phenylephrine use, atrial brillation, SVT, nonlethal events, and death). All these showed no signicant differences between the preemptive and control groups Not reported Pruritus: pre-TEA: n 3 (10%), post-TEA: n 2 (6.7%) Nausea: pre-TEA: n 2 (6.5%), post-TEA: n 2 (6.5%)

Neustein et al13 Yegin et al14

Pre-TEA n 9 Post-TEA n 13 Pre-TEA n 30 Post-TEA n 31

PREEMPTIVE EPIDURAL ANALGESIA

793

proven in this review. To do this, a larger RCT designed specically to evaluate the effects of preemptive TEA would be needed. It would need to be sufciently powered to detect

a signicant difference in the severity of acute pain, as well as having an adequate follow-up period to detect any difference in the incidence of chronic post-thoracotomy pain.

REFERENCES 1. International Association for the Study of Pain: Subcommittee of Taxonomy. Classication of chronic pain. Pain S138-S139, 1986 (suppl 3) 2. Kalso E, Perttunen K, Kaasinen S: Pain after thoracic surgery. Acta Anaesthesiol Scand 36:96-100, 1992 3. Perttunen K, Tasmuth T, Kalso E: Chronic pain after thoracic surgery, a follow-up study. Acta Anaesthesiol Scand 43:563-567, 1999 4. Dajczman E, Gordan A, Kreisman H, et al: Long-term postthoracotomy pain. Chest 99:270-274, 1991 5. Katz K, Kavanagh BP, Sandler AN: Acute pain after thoracic surgery predicts long-term post-thoracotomy pain. Clin J Pain 12:5055, 1996 6. Obata H, Saito S, Fujita N, et al: Epidural block with mepivacaine before surgery reduces long-term post-thoracotomy pain. Can J Anesth 46:1127-1132, 1999 7. Sentrk M, zcan P, Talu GK, et al: The effects of three different analgesia techniques on long-term post-thoracotomy pain. Anesth Analg 94:11-15, 2002 8. Salomaki TE, Laitinen JO, Nuutinen LS: A randomized doubleblind comparison of epidural versus intravenous fentanyl infusion after thoracotomy. Anesthesiology 75:790-795, 1991 9. Debreceni G, Molnar Z, Szelig L, et al: Continuous epidural or intercostal analgesia following thoracotomy: A prospective randomized double-blind clinical trial. Acta Anaesthesiol Scand 47:10911095, 2003 10. Subramanian B, Pawar DK, Kashyap L: Pre-emptive analgesia with epidural morphine or morphine and bupivacaine. Anaesth Intensive Care 28:392-398, 2000 11. Aguilar JL, Rincon R, Domingo V, et al: Absence of an early pre-emptive effect after thoracic extradural bupivacaine in thoracic surgery. Br J Anaesth 76:72-76, 1996 12. Ochroch EA, Gottschalk A, Augostides J: Long-term pain and activity during recovery from major thoracotomy using thoracic epidural analgesia. Anesthesiology 97:1234-1244, 2002 13. Neustein S, Kreitzer J, Krellenstein D, et al: Preemptive epidural analgesia for thoracic surgery. Mt Sinai J Med 69:101-104, 2002 14. Yegin A, Erdogan A, Kayacan N, et al: Early postoperative pain management after thoracic surgery; pre- and postoperative versus postoperative epidural analgesia: A randomized study. Eur J Cardiothorac Surg 24:420-424, 2003 15. Dickenson AH, Sullivan AF: Subcutaneous formalin-induced activity of dorsal horn neurons in the rat: Differential response to an intrathecal opiate administered pre- and post- formalin. Pain 30:349-360, 1987 16. Yashpal K, Katz J, Coderre TJ: Effects of pre-emptive or postinjury intrathecal local anesthesia on persistent nociceptive responses in rats. Anesthesiology 84:1119-1128, 1996 17. Woolf CJ, Chong MS: Pre-emptive analgesiaTreating postoperative pain by preventing the establishment of central sensitization. Anesth Analg 77:362-379, 1993 18. Dahl JB, Kehlet H: The value of pre-emptive analgesia in the treatment of postoperative pain. Br J Anaesth 70:434-439, 1993 19. Kissin I: Pre-emptive analgesia. Why its effect is not always obvious (editorial). Anesthesiology 84:1015-1019, 1996 20. Dahl JB, Hansen BL, Hjorts NC, et al: Inuence of timing on the effect of continuous extradural analgesia with bupivacaine and morphine after major abdominal surgery. Br J Anaesth 69:4-8, 1992 21. Pryle BJ, Vanner RG, Enriquez N, et al: Can pre-emptive lumbar epidural blockade reduce postoperative pain following abdominal surgery? Anaesthesia 48:120-123, 1993 22. Katz J, Kavanagh BP, Sandler AN, et al: Pre-emptive analgesia: Clinical evidence of neuroplasticity contributing to postoperative pain. Anesthesiology 77:439-446, 1992 23. Benedetti F, Amanzio M, Casadio C, et al: Postoperative pain and supercial abdominal reexes after posterolateral thoracotomy. Ann Thorac Surg 64:207-210, 1997 24. Camu F, Debucquoy F: Alfentanil infusion for postoperative pain: A comparison of epidural and intravenous routes. Anesthesiology 75:171-178, 1991 25. Ballantyne JC, Carr DB, deFerranti S, et al: The comparative effects of postoperative analgesic therapies on pulmonary outcome: Cumulative meta-analyses of randomized, controlled trials. Anesth Analg 86:598-612, 1998