Share this article on:

Continuous Thoracic Epidural Anesthesia with 0.2% Ropivacaine Versus General Anesthesia for Perioperative Management of Modified Radical Mastectomy

Doss, Nabil W. MD*,; Ipe, Joseph MD*,; Crimi, Thomas MD*,; Rajpal, Sanjeev MD†,; Cohen, Steven MD†,; Fogler, Richard J. MD†,; Michael, Rafik MD*,; Gintautas, Jonas MD, PhD, MBA

doi: 10.1097/00000539-200106000-00041
Regional Anesthesia: Research Report

We evaluated in this prospective study the effectiveness of continuous thoracic epidural anesthesia (TEA) and postoperative analgesia with ropivacaine and compared it with general anesthesia (GA) and opioids for pain relief, side effects, postanesthesia recovery, and hospital discharge after modified radical mastectomy. Sixty ASA physical status II and III patients undergoing mastectomy were randomly assigned to two study groups of 30 patients each. In the TEA group, an epidural catheter was inserted at T6-7, and 5–10 mL of 0.2% ropivacaine was injected to maintain anesthesia and to continuously administer adequate analgesia for 48 h. GA was induced with IV 1–2 mg of midazolam or 50–100 μg/mL of fentanyl followed by 50–150 mg of propofol and was maintained with sevoflurane and 50% N2O in oxygen. The Aldrete score system was used to evaluate postanesthesia recovery, a verbal rating scale was used for assessment of pain intensity, and a postanesthesia discharge scoring system was used for discharge home. The demographic data and side effects (except for nausea and vomiting) (GA 43%, TEA 10%, P = 0.0074) and discharge home were similar in both groups. However, the number of patients ready for discharge from the recovery room during the first postanesthesia hour (Aldrete score of 10) was significantly larger after TEA (80%) than after GA (33%) (P = 0.0006). GA patients experienced significantly more (P < 0.001) substantial pain than TEA patients on Day 0 (70%), Day 1 (53%), and Day 2 (27%) after the surgery. Patient satisfaction was greater with TEA (70%) than with GA (30%) (P < 0.001). We conclude that TEA with ropivacaine provides better postoperative pain relief and less nausea and vomiting, facilitates postanesthesia recovery, and gives greater patient satisfaction than GA.

Departments of *Anesthesiology and †Surgical Services, The Brookdale University Hospital and Medical Center, Brooklyn, New York

February 2, 2001.

Address correspondence and reprint requests to Nabil W. Doss, MD, The Brookdale University Hospital and Medical Center, 1 Brookdale Plaza, Brooklyn, NY 11212.

Breast cancer is the most common cause of tumors in women. It was estimated in the year 2000 that in the United States, 182,800 new cases of invasive breast cancer were expected to occur among women, and about 1400 new cases of the disease were expected to be diagnosed in men (1). The American Cancer Society estimated that 41,200 patients would die because of breast cancer during 2000.

Until recently, oncologic breast surgeries were typically performed by using inhaled anesthesia followed by inpatient hospitalization. However, general anesthesia (GA) cannot provide postoperative incisional pain control. Even though the clinical importance remains uncertain, several laboratory and clinical studies have demonstrated that general anesthetics (with the exception of large doses of opioids), although they produce the desired state of unconsciousness, do not eliminate the surgical stress response, may aggravate immunosuppression, and may cause undesirable side effects (2). Postoperative nausea and emesis are some of the most unpleasant side effects, which can affect patients undergoing all types of surgical procedures requiring anesthesia or sedation. There is evidence that under GA, 56% of patients suffer nausea and vomiting up to 24 h after breast cancer surgery (3). The complication has been described by patients as being more debilitating than the mastectomy itself (4). The routine use of parenteral narcotics (strong emetogens) in these cases further aggravates the situation. Recovery time is prolonged, hospital stay is lengthened, and hospital costs are increased (3).

In light of this situation, various regional anesthetic techniques for oncologic breast surgery have been suggested, including field block, local anesthetic infiltration, intercostal nerve block, epidural anesthesia, paravertebral block, and brachial plexus block (5,6). However, because of technical difficulties, inherent limitations of anesthetic and analgesic effects, and the increased possibilities of complications (e.g., pneumothorax), the application of some of these approaches is not entirely satisfactory in the anesthetic management of extensive oncologic breast surgeries and in postoperative pain control.

Our rationale for the inclusion of thoracic epidural anesthesia (TEA) in this study was based on clinical evidence suggesting that extradural anesthesia has been associated with fewer postsurgical recovery complications, shorter hospital stays, and, consequently, decreased health care costs (7). Furthermore, epidural anesthesia and analgesia have been associated with a reduction in the incidence and severity of many perioperative physiologic trepidations. TEA selectively blocks cardiac sympathetic fibers, and this offers potential patient benefits: attenuation of the surgical stress response, improvement of myocardial oxygen balance, and stabilization of intraoperative hemodynamics (8,9). This is particularly relevant for the patient with coexisting cardiac morbidity (10). Finally, although it is not yet scientifically proved, many clinical studies have embraced the concept that epidural anesthesia offers a preemptive analgesic effect (11). A limited number of studies have found that TEA with local anesthetics coadministered with opioids provide better outcomes after breast surgery than GA (3,12).

Ropivacaine is a relatively new member of the long-acting amino amide class of local anesthetics. It is exclusively the S-(−)-enantiomer, which is closely related to the chemical structure of bupivacaine and mepivacaine, which are racemic mixtures (13). Initial clinical studies of epidural anesthesia have demonstrated that pharmacokinetic properties for ropivacaine are similar to those seen with bupivacaine (14,15). The two drugs were also comparable as far as the onset and duration of sensory block; however, the frequency, degree, and the duration of motor block are less affected by ropivacaine than bupivacaine when compared in the same concentration (16). From both animal and human studies, there is some evidence that ropivacaine has a decreased central nervous system toxicity and cardiotoxic potential than bupivacaine (17).

Anesthetic-analgesic effects of small-dose epidural ropivacaine on oncologic breast surgery have not been yet described, despite its widespread clinical use. Therefore, we conducted a prospective, randomized clinical study to assess the efficacy of continuous TEA and postoperative analgesia with 0.2% ropivacaine as compared with GA and opioids for pain relief and postsurgical recovery in the perioperative management of oncologic mastectomy.

Back to Top | Article Outline


After obtaining our IRB approval and written, informed consent from the patients, we prospectively enrolled 60 adult women, ASA physical status II or III, who were scheduled for elective oncologic unilateral breast surgery. Before the surgery, the surgeon and anesthesiologist instructed all patients regarding the use of either a TEA technique or GA for the operation. Half of the patients (n = 30) were assigned to the TEA group and the other half (n = 30) to the GA group by using a predetermined random 1:1 sequence. An anesthesia research fellow who was blinded to the treatment group conducted postoperative data collection. (A mock epidural catheter was placed in the GA patients to simulate an epidural infusion.) An anesthesiologist unaware of the patient’s group assignment evaluated and treated postoperative pain and nausea.

In the TEA group, exclusion criteria were any contraindications to TEA, infection at the site of the planned epidural placement, any coagulation disorders, or known allergy to ropivacaine or opioids. An 18-gauge IV catheter was inserted, and lactated Ringer’s solution was infused before anesthesia at the rate of 3 mL · kg−1 · h−1. Mild sedation was induced with 1–2 mg midazolam or 50–100 μg fentanyl, depending on the age of the patient. TEA was performed with aseptic technique, with the patient in the lateral decubitus position. A 16-gauge Tuohy needle was inserted in the posterior midline at the level of T6-7; the thoracic epidural space was identified by means of a loss-of-resistance technique. An epidural catheter was measured and inserted 3 to 5 cm into the epidural space through the Tuohy needle. A test dose of 2 to 3 mL of 0.2% ropivacaine with epinephrine 1:200,000, with repeated aspiration, guided by the sensory effect, was given to exclude intravascular or intrathecal injection. The catheter was secured and left in place for 48 h. It was used for continuous postoperative analgesia with the same small dose of ropivacaine. The initial titrated dose of 5–10 mL of 0.2% ropivacaine injected through the catheter resulted in full, bilateral anesthesia of the skin, subcutaneous tissue, mammary glands, and thoracic wall in the area 2–3 cm below the clavicle superiorly and the costal arch inferiorly. The targeted anesthetic dermatomal levels achieved were determined by pinprick and measured at 2-min intervals for the first 10 min after injection of the drug. Subsequent 3–5-mL doses of ropivacaine, if sensory or motor block (Bromage scale) was not adequate, were titrated individually, and supplemental oxygen (3–6 L/min) was administered through a face mask for the duration of the surgery. In those patients in whom axillary node dissection required supplemental analgesia, the surgeon did local infiltration of the area with 5–10 mL of 0.2% ropivacaine. Postoperative pain in the TEA group was controlled with the continuous infusion of 4–6 mL/h of 0.2% ropivacaine through the epidural catheter for the first 24 h. For the second 24 h, the epidural infusion was reduced to 3–4 mL/h. Then oral analgesics (acetaminophen, total dose 1200–1800 mg/24 h, with codeine, total dose 120–180 mg/24 h) or nonsteroidal antiinflammatories (ibuprofen, total dose 400–600 mg/24 h) were administered.

All patients in the GA group were induced with IV midazolam 1–2 mg or fentanyl 50–150 μg (depending on the patient’s age and general condition), followed by propofol 50–150 mg IV. Tracheal intubation was facilitated with succinylcholine 1–1.5 mg/kg IV. Anesthesia was maintained with sevoflurane in combination with N2O 50% in oxygen. A nondepolarizing muscle relaxant, rocuronium 0.5 mg/kg or cisatracurium 0.1 mg/kg, was administered IV as clinically indicated. Sevoflurane and propofol were discontinued when the skin closure began, and N2O was discontinued after the last suture was placed. Neostigmine 0.05 mg/kg with atropine 0.02 mg/kg or glycopyrrolate 0.01 mg/kg were injected IV for neuromuscular blockade reversal. Postoperative pain management in the GA group was provided initially with scheduled IV opioids (hydromorphone 1–2 mg or meperidine 50–75 mg) every 4 h for the first 24 h and every 4–8 h for the next 24 h. After that, the same oral analgesics (including nonsteroidal antiinflammatories) as in the TEA group (the total dose was also similar) were given.

Patient monitoring included noninvasive blood pressure measurement, heart rate, end-tidal CO2, and pulse oximetry. In all patients, intraoperative hypotension and hypertension (±20% deviation from baseline), bradycardia (heart rate <50 bpm), and tachycardia (heart rate >100 bpm) were recorded. The incidence of postoperative side effects (e.g., nausea, vomiting, dizziness), as well as the duration of the postanesthesia care unit and hospital stays, was documented. Pain intensity was assessed with a simple, categorical verbal rating scale consisting of four adjectives describing levels of pain (no pain, mild pain, moderate pain, and severe pain).

We used the same postanesthesia recovery and discharge criteria for TEA and GA patients. Postanesthesia recovery was evaluated by using the modified Aldrete score system involving the level of consciousness, motor activity, respiration, and circulation (18). An Aldrete score of 10 (patient fully awake, oriented, and comfortable, with stable cardiovascular and respiratory signs) was used for discharge from the postanesthesia care unit. Subsequently, patients were evaluated every 6 h during the first 24 h and at least twice daily until discharge from the hospital. A standardized criterion for home readiness was applied (19), and the day of actual discharge was recorded by using a modified postanesthesia discharge scoring system. The score of 9 (vital signs within 20% of the preoperative value, patient well oriented and manifesting a steady gait, no dizziness, no appreciable pain, no nausea or vomiting, and the absence of surgical bleeding) was considered evidence for discharge home. Patient satisfaction with the anesthetic experience was noted. All were asked to rate their overall experience with the anesthetic technique as “satisfactory” or “unsatisfactory.”

Statistical analysis was performed with the SAS system (SAS, Cary, NC). Fisher’s exact test for 2 × 2 contingency tables was applied, and continuous data were analyzed with the unpaired Student’s t-test. Power analysis was conducted to determine whether nonsignificant findings could plausibly become significant with larger sample size. Results are presented as mean ± sd unless otherwise stated. Probability values of <0.05 were considered to be statistically significant.

Back to Top | Article Outline


All 60 patients completed the study. There were no significant differences between the TEA and GA groups in age (64 ± 17 vs 66 ± 6 yr), weight (78 ± 16 vs 76 ± 12 kg), height (162 ± 9 vs 160 ± 9 cm) or ASA physical status. Overall, our patients with breast cancer were older women, the majority of whom (90%) had coexisting morbid diseases (hypertension, coronary disease, chronic obstructive pulmonary disease, obesity, and diabetes mellitus). All patients underwent a unilateral modified radical mastectomy with some degree of axillary node dissection. There were no failures in the placement of the epidural catheter. The same surgeon performed all the surgeries, which lasted 1–3 h. There were no significant differences in the duration of the surgery between the TEA and the GA groups (103 ± 36 vs 109 ± 41 min). Estimated blood loss was minimal and similar in the two groups; none of the patients required blood transfusion. The most frequently reported complication of the surgery was nausea and vomiting. Statistical analysis indicated that GA produced a significantly more frequent incidence of nausea and vomiting, reaching 43%, as compared with 10% in the TEA group (P = 0.0074). Hypertension (>20% from the baseline) was noted during surgery in 16.6% of patients in the GA group only. It was correlated with surgical stimulation and was corrected by an increasing depth of anesthesia. Hypotension, bradycardia, and tachycardia were mild problems (≤10% of the baseline) in the two groups of patients during the entire perioperative period. One patient from the GA group developed cellulitis after surgery.

Figure 1 illustrates the percentages of patients from both groups achieving a postanesthesia recovery score (Aldrete score) of 10 at 1-, 2-, and 3-h intervals after discontinuation of anesthesia (at 1 h, P = 0.0006; at 2 h, P = 0.146; and at 3 h, P = 1.000). Thus, the difference in the postanesthesia recovery score reached statistical significance between the groups only at the 1-h interval. The results of four-point scales of pain intensity, as recorded by verbal rating scale, for statistical purposes were dichotomized as a “substantial” (moderate pain and severe pain) versus a “nonsubstantial” (no pain and mild pain) pain for postoperative Days 0, 1, 2, and 3. The analysis showed (Table 1) that the GA patients experienced significantly more substantial pain than the TEA patients during Day 0 (the operative day) (n = 21, 70%), Day 1 (n = 16, 53%), and Day 2 (n = 8, 33%) (P < 0.001). Nine patients (30%) during TEA required axillary local anesthetic supplementation. At the surgeon’s discretion, eight patients (27%) in the GA group received the axillary infiltration at the completion of the surgery. The total ropivacaine consumption in the TEA group for postoperative pain control ranged between 336 and 480 mg (mean, 408 ± 36 mg/48 h). The total hydromorphone dose consumed in the GA group for pain control after the surgery was between 9 and 18 mg (mean, 13.5 ± 2.0 mg/48 h); the total meperidine dose used in the same group was 450–675 mg (mean, 562.5 ± 42/48 h). The amounts of the oral analgesics in the two groups were similar: acetaminophen (1200–1800 mg/24 h) with codeine (120–180 mg/24 h) and ibuprofen (400–600 mg/24 h) used as required.

Figure 1

Figure 1

Table 1

Table 1

The data from the postanesthesia discharge scoring system (a score of 9 indicated home ready) show similar times for home readiness for both groups of patients. Although at least 16%–20% of patients from either group were discharged on Day 1 after the surgery, the majority of the TEA (23 patients, 77%) and the GA (21 patients, 70%) went home on postoperative Day 2. Statistical analysis was applied for Days 1 and 2 only. Because only one patient (TEA group) was discharged on Day 0, she was included in the contingency table of Day 1. The results yielded no significant differences in the number of patients being home ready between the TEA and GA groups on either Day 1 or Day 2. A significantly larger percentage of patients (P < 0.001) were satisfied with TEA (n = 21, 70%) than with GA (n = 9, 30%) after oncologic mastectomy.

Back to Top | Article Outline


The goal of this study was to assess whether continuous TEA with 0.2% of ropivacaine provides better postoperative pain relief and recovery than GA and parenteral opioids for women undergoing major surgery because carcinoma of the breast. Our findings indicate that TEA has some advantages over GA in patients undergoing modified radical mastectomy. The main results of this study are that patients in the TEA group had significantly shorter postanesthesia recovery, had significantly less nausea and vomiting, and experienced significantly less pain after surgery than patients in the GA group. Our results confirm similar observations reported by other investigators (3,12). However, our study uniquely demonstrated that continuous TEA monotherapy with small-dose (0.2%) ropivacaine in limited volume produces adequate anesthetic sensory block and provides prolonged postsurgical pain relief without opioid supplementation for older women with coexisting morbid diseases who underwent oncologic modified radical mastectomy. Because surgery of the breast does not require motor blockade, it allowed us to use this small concentration of the local anesthetic to produce a full sensory anesthesia sufficient for the surgery with minimal hemodynamic or respiratory effect or motor embarrassment usually associated with TEA (hypotension or diaphragmatic or intercostal weakness). For postoperative pain management, the TEA patients received on the average only about one-fourth of the suggested cumulative dose of ropivacaine. No opioids were required for postoperative pain in the TEA group. Local infiltration for axillary dissection may account for less pain immediately after surgery. However, only one-third of the patients in the TEA group required axillary infiltration. There were no differences in the postoperative analgesic requirements between those patients in the TEA group who needed the axillary supplementation and those who did not. The effect of axillary infiltrations and TEA on postanesthesia pain management requires further investigation.

The small concentration of ropivacaine used in our TEA patients produced no respiratory, cardiovascular, or neurologic complications. Ropivacaine is described as being less potent, cardiotoxic, and neurotoxic than bupivacaine (13–16). Ropivacaine may be the drug of choice in TEA for patients with concomitant cardiopulmonary diseases who require anesthesia for breast surgery. The value of TEA in a patient with severe cardiac disease undergoing a simple mastectomy has been reported (20). The incidences of nausea and emesis in our patients were slightly less pronounced than those reported in the literature (3,4,12). This may be explained by the fact that in the TEA patients we administered no opioids and used only a continuous infusion of ropivacaine for analgesia. Postoperative nausea and vomiting were evident in the GA patients who felt dizzy. None of our patients developed respiratory problems; mild hypotension (≤10% of the baseline) was observed in two TEA patients and in one GA patient and was controlled with ephedrine and fluid administration. Hypertension during GA in five patients, as well as mild changes in heart rate (one patient in each group) throughout the perioperative period, was corrected pharmacologically. These cardiovascular deviations were within clinical expectations because most of our elderly women were admitted with several preexisting illnesses. Future studies with a larger patient population may answer whether the relative frequency of the side effects is the result of anesthetic technique or the surgically enhanced continuation of a patient’s premorbid condition. The power analysis applied to detect a difference in hypertension rates in our study plausibly suggests that a somewhat larger patient sample would have found a significant improvement caused by TEA. The analysis also indicated that inadequate sample size is an unlikely cause of our failure to detect a difference in other measures. It would also be interesting to study the incidence and duration of persistence of pain in the scar and persistent phantom pain after modified radical mastectomy after the discharge from the hospital. Phantom breast syndrome after mastectomy has already been reported (21).

Duration of surgery in our study was not significantly different between the groups. Our study supports a report that the mean durations of mastectomy times either under general or thoracic epidural techniques are similar (12). Our findings that patients in the TEA group showed significantly faster postanesthesia recovery (Aldrete score) in the first hour as compared with the GA group are confirmed by other observations (3,8,12). In the current climate of managed care delivery and focus on cost containment, the ability to perform oncologic breast surgery under regional anesthesia has the potential for major cost saving. Although the scope of our study did not include the cost evaluation, it could be assumed that TEA has a potential for cost saving because shorter postanesthesia recovery does not increase recovery room charges (US$200 per hour in our institution) or pharmacy and supply charges (about US$400 for mastectomy). It was reported that younger and healthier patients who had received TEA were discharged from the hospital earlier than those who had received GA after breast surgery (3). Advanced age, poor health, and persistent symptoms after GA (nausea and vomiting, pain, cardiovascular instability) delayed home discharge in three of our patients. In one patient, the discharge delay was caused by the lack of availability of an escort. This type of delay has been recognized previously (19).

Breast carcinoma is the most common malignancy among women. Along with lung cancer, it shares the highest fatality rate of all cancers affecting women. Breast cancer is a malignant proliferation of epithelial cells lining the ducts or lobules of the breast. Surgical treatment of breast carcinoma is becoming more and more conservative, particularly for small lesions. However, for more advanced, fungating carcinomas (Stage IIIb) or for the repair of necrotic dermal and chest wall defects caused by radiation, mastectomy may still be needed to resect, debride, or reconstruct the area. It is obvious that the type and extent of breast cancer, as well as the age and general condition of the patient, dictate the selection of surgical and anesthetic management. Economic restraints imposed on clinical practice by recent changes in health care management and scrutiny regarding cost-effectiveness, influenced by both internal and external regulators, force our specialty to search not only for the best method of perioperative anesthesia and analgesia, but also for ways to deliver an effective yet economical analgesic technique.

In conclusion, our study shows that TEA with a single local anesthetic is a safe and reliable alternative to GA with an inhaled anesthetic in women undergoing modified radical oncologic breast surgery. A small dose (0.2%) of ropivacaine administered through a continuous thoracic epidural technique not only provides intraoperative anesthesia but also significantly enhances postoperative analgesia and improves overall satisfaction for patients undergoing modified radical mastectomy.

The authors thank Dr. Jeremy Weedon for statistical consultation.

Back to Top | Article Outline


1. American Cancer Society. Cancer facts and figures 2000. Atlanta, GA: American Cancer Society, 2000.
2. Stevenson GW, Hall SC, Rudnick S, et al. The effect of anesthetic agents on the human immune response. Anesthesiology 1990; 72: 542–52.
3. Lynch EP, Welch KJ, Carabuena TM, et al. Thoracic epidural anesthesia improves outcome after breast surgery. Ann Surg 1995; 222: 663–9.
4. Watcha MF, White PF. Postoperative nausea and vomiting. Anesthesiology 1992; 77: 162–84.
5. Oakley N, Dennison A, Shorthouse A. A prospective audit of simple mastectomy under local anesthesia. Eur J Surg Oncol 1996; 2: 131–6.
6. Klein SM, Bergh A, Steel SM, et al. Thoracic paravertebral block for breast surgery. Anesth Analg 2000; 90: 1402–5.
7. Edwards M, Broadwater J, Bell J, et al. Economic impact of reducing hospitalization for mastectomy patients. Ann Surg 1988; 208: 330–6.
8. Liu S, Carpenter RL, Neal JM. Epidural anesthesia and analgesia. Anesthesiology 1995; 82: 1474–506.
9. Davis RF, DeBoer LW, Maroko PR. Thoracic epidural anesthesia reduces myocardial infarct size after coronary artery occlusion in dogs. Anesth Analg 1986; 65: 711–7.
10. Blomberg S, Emanuelsson H, Kvist H, et al. Effects of thoracic epidural anesthesia on coronary arteries and arterioles in patients with coronary artery disease. Anesthesiology 1990; 73: 840–7.
11. Abram SE. Importance of preincisional analgesia. Reg Anesth 1996; 21 (Suppl 6): 117–21.
12. Yeh CC, Yu JC, Wu CT, et al. Thoracic epidural anesthesia for pain relief and postoperation recovery with modified radical mastectomy. World J Surg 1999; 23: 256–61.
13. Akerman B, Hellberg I, Trossvik C. Primary evaluation of the local anesthetic properties of the amino amide agent ropivacaine (LEA 103). Acta Anaesthesiol Scand 1988; 32: 517–8.
14. Concepcion M, Arthur GR, Steele SM, et al. A new local anesthetic, ropivacaine: its epidural effects in humans. Anesth Analg 1990; 70: 80–5.
15. Whitehand E, Arrigoni B, Bannister J. An open study of ropivacaine in extradural anaesthesia. Br J Anaesth 1990; 64: 67–71.
16. Feldman HS, Covino BG. Comparative motor-blocking effects of bupivacaine and ropivacaine, a new amino amide local anesthetic, in the rat and dog. Anesth Analg 1988; 67: 1047–52.
17. Crosby E, Sandler A, Finucane B, et al. Comparison of epidural anaesthesia with ropivacaine 0.5% and bupivacaine 0.5% for caesarean section. Can J Anaesth 1998; 95: 66–71.
18. Aldrete JA. The post-anesthesia recovery score revisited. J Clin Anesth 1995; 7: 89–91.
19. Chung F. Are discharge criteria changing? J Clin Anesth 1993; 5: S64–8.
20. Nguyen DV, Kokoszka JS. Use of thoracic epidural anesthesia for breast surgery in a patient with severe cardiac disease. Am J Anesth 2000; 27: 98–100.
21. Kroner K, Knudsen UB, Lundby L, et al. Long-term phantom breast syndrome after mastectomy. Clin J Pain 1992; 8: 346–50.
© 2001 International Anesthesia Research Society