E to be the analgesic method of choice for painful

Thoracic Versus Lumbar Epidural Fentanyl for Postthoracotomy Pain Corey W. T. Sawchuk, MD, Bill Ong, MD, Helmut W. Unruh, MD, Thomas A. Horan, MD, and Roy Greengrass, MD Departments of Anesthesia and Surgery, The University of Manitoba, Winnipeg, Manitoba, Canada Thirty patients were prospectively randomized to receive either thoracic or lumbar epidural fentanyl infusion for postthoracotomy pain. Epidural catheters were inserted, and placement was confirmed with local anesthetic testing before operation. General anesthesia consisted of nitrous oxide, oxygen, isoflurane, intravenous fentanyl citrate (5 pgkg), and vecuronium bromide. Pain was measured by a visual analogue scale (0 = no pain to 10 = worst pain ever). Postoperatively, patients received epidural fentanyl in titrated doses every 15 minutes until the visual analogue scale score was less than 4 or until a maximum fentanyl dose of 150 pg by bolus and an infusion rate of 150 pgh was reached. The visual analogue scale score of patients who received thoracic infusion decreased from 8.8 f 0.5 to 5.5 f 0.7 (p 5 0.05) by 15 minutes and to 3.5 f 0.4 (p c: 0.05) by 45 minutes. The corresponding values in the lumbar group were 8.8 f 0.6 to 7.8 f 0.7 at 15 minutes and 5.3 f 0.9 at 45 minutes (p 5 0.05). The infusion rate needed to maintain a visual analogue scale score of less than 4 was lower in the thoracic group (1.55 f 0.13 pg - kg-' * h-') than in the lumbar group (2.06 f 0.19 fig - kg-' * h-') during the first 4 hours after operation (p 5 0.05). The epidural fentanyl infusion rates could be reduced at 4, 24, and 48 hours after operation without compromising pain relief. Four patients in the lumbar group required naloxone hydrochloride intravenously. Three of these patients had respiratory rates of less than 6/min and were difficult to arouse. The fourth patient was difficult to arouse and had an arterial carbon dioxide tension of 83 mm Hg. We conclude that thoracic epidural fentanyl infusion is better than lumbar infusion for postthoracotomy pain control because of more rapid onset, smaller dose requirements, and less respiratory depression. (Ann Thorac Surg 1993;55:1472-6) pidural analgesia is extremely effective and may prove E to be the analgesic method of choice for painful procedures such as thoracotomy. Shulman and colleagues [ 11 compared epidural with intravenous administration of morphine in postthoracotomy patients and demonstrated that postthoracotomy patients receiving epidural morphine have better acute indicators of respiratory function than patients receiving intravenous morphine. Both thoracic and lumbar epidural fentanyl citrate infusions have been reported to provide good pain relief after thoracic operations [2-51, but uncertainty remains about the best site and dose. Chamberlain and co-workers [6] reported that thoracic epidural fentanyl infusion provided better pain scores at smaller doses than lumbar epidural fentanyl infusion in a group of 20 patients. The patients in this study controlled the infusion rates themselves, and the study design did not include assessment of postoperative pulmonary function. Lumbar epidural anesthesia is often used because it is technically easier to perform and because the risk of spinal cord injury is lower if inadvertent dural puncture occurs. The spinal cord ends and the cauda equina begins at about L1-2, and a lumbar epidural catheter is usually placed below L2. Accepted for publication Sep 21, 1992. Address reprint requests to Dr Unruh, Department of Surgery, The University of Manitoba, Health Sciences Centre, 107-671 William Ave, Winnipeg, Man, Canada R3E 022. The onset of epidural analgesia varies and depends primarily on drug concentration and bolus volume. Badner and associates [2] reported a delay of 3 to 4 hours with continuous lumbar epidural infusion, whereas Melendez and co-workers [3] achieved adequate pain relief within 15 minutes by administering 200,ug of fentanyl in a 20-mL volume into the lumbar epidural space. A smaller dose or a smaller bolus volume might have produced an equally rapid onset of action in the latter study, whereas the small volume probably accounted for the delayed onset in the former. We performed a prospective, randomized study using a combination of incremental epidural boluses and continuous infusion to compare thoracic and lumbar epidural fentanyl infusion. Data collected included the time required for the onset of pain relief, the degree of pain relief, and the drug dosages required to achieve a satisfactory level of analgesia. In addition, postoperative pulmonary function and side effects were monitored and compared in the two groups. From our data, we tried to answer the question as to which epidural site is better. Material and Methods With prior institutional approval of both The University of Manitoba Faculty Committee on the Use of Human Subjects in Research and the Health Sciences Centre Research Advisory Committee, 31 patients were enrolled in the 0 1993 by The Society of Thoracic Surgeons

Ann Thorac Surg 1993;55:1472-6 SAWCHUK ET AL 1473 study. A detailed written and verbal explanation of the study was given to each potential patient, and informed consent to participate in the study was thereby obtained. Patients were prospectively randomized to receive either a lumbar epidural catheter at the L2-3 or L M level (n = 15) or a thoracic epidural catheter at the T7-8 level (n = 15). One patient had to be excluded from the study because of a change in the surgical procedure. Study Protocol Before operation, forced expiratory volume in 1 second (FEV,), forced vital capacity (FVC), and arterial blood gases were measured in all patients. On the day before operation, patients were instructed on the use of a visual analogue scale (VAS) (0 = no pain to 10 = worst pain ever) to express the level of pain. The use of a VAS for pain assessment is a widely used and extensively validated psychometric technique [7]. One hour before they were taken to the operating room, the patients were premedicated with oral diazepam (0.1 mgkg). On arrival in the operating room, intravenous and intraarterial catheters were inserted under local anesthesia. The thoracic or lumbar catheter was also inserted before the induction of anesthesia, and catheter placement was verified by a 3-mL test dose of lidocaine hydrochloride 2%. All patients demonstrated a band of analgesia at appropriate dermatomes within 15 minutes of the test dose injection. Intraoperative monitoring consisted of continuous electrocardiography, pulse oximetry, end-tidal carbon dioxide measurement, inhalation agent analysis, and blood pressure measurement. Anesthesia was induced with fentanyl (5 pglkg) and sodium thiopental (3 to 5 mg/kg). Vecuronium bromide (0.1 mgkg) was used to facilitate tracheal intubation. Anesthesia was maintained by isoflurane, nitrous oxide, and vecuronium, and mechanical ventilation was adjusted to maintain an end-tidal carbon dioxide between 32 and 36 mm Hg. No other intravenous opiates or sedatives were given during the perioperative period. At the end of the operation, muscle relaxation was reversed with glycopyrrolate (10 pg/kg) and neostigmine methylsulfate (50 pgkg), and the endotracheal tube was removed in the operating room. All patients received supplemental oxygen for the full duration of this study. After the patient had recovered sufficiently from anesthesia in the recovery room or intensive care unit to report a VAS score greater than 4, a 50-pg bolus of epidural fentanyl (10 pg/ml) was administered and fentanyl infusion at 50 pg/h was started. There was no delay in initiating the epidural infusions. As soon as the patient was sufficiently awake, he or she was asked to indicate the level of pain on the VAS, and the fentanyl infusion was begun. Repeat 25-pg fentanyl boluses were subsequently administered every 15 minutes, and the hourly infusion rate was equally increased until the VAS score was less than 4 or a predetermined maximum of 150 pg had been given or the respiratory rate fell to less than lo/min. This hourly infusion rate of epidural fentanyl was then maintained for 4 hours after the end of the operation. Patients were reassessed at 4, 24, 48, and 72 hours, and infusion rates were reduced by 25% at 4, 24, and 48 hours Table I. Summary of Patient Data" Thoracic Epidural Lumbar Epidural Variable Group (n = 15) Group (n = 15) Age (Y) 61 2 3 Sex (M:F) 78 Height (cm) 166 f 3 Weight (kg) 74 2 5 60 2 5 10:5 160 k 12 83 2 10 a Data are shown as the mean 5 the standard error of the mean. if the VAS score was 4 or less. If breakthrough pain occurred, the infusion rate was returned to the previous level. Epidural fentanyl infusion was maintained for up to 72 hours after operation. Most patients were cared for on a cardiothoracic surgical nursing ward, with the minority in the intensive care unit. All the nurses caring for patients in this study received special instructions on continuous epidural fentanyl infusion. Patients were monitored according to standardized nursing protocols. Vital signs were determined every hour for the first 12 hours after operation and then every 3 hours thereafter until completion of infusion. The nurses were advised to discontinue the infusion and notify the anesthesiologist on call if the respiratory rate was less than lo/min. If the patient could not be roused by voice while the respiratory rate was less than lo/min, the nurse was advised to administer naloxone hydrochloride (0.4 mg intravenously), discontinue the infusion, and notify the anesthesiologist on call. Patients who did not respond to verbal commands but had a satisfactory respiratory rate and good vital signs were considered somnolent, and their infusions were either stopped or reduced by 25% depending on the assessment of the anesthesiologist. Arterial blood gases were routinely measured 2 hours after operation and at other times at the discretion of the medical staff. The FEV, and FVC were measured at 14,48, and 72 hours after operation. S tatis tical Analysis All results are presented as the mean * the standard error of the mean. The patient demographics, VAS scores, dosage data, arterial blood gases, and pulmonary function were analyzed using analysis of variance. Multiple comparisons were done using Duncan's multiple range test. Fisher's exact test was used for comparisons of side effects and complications. Values of p of 0.05 or less were taken to indicate significant differences in all instances. Results The patients in the thoracic and lumbar groups were similar in age, weight, height, and sex distribution (Table 1). All patients had excellent pain relief as defined by a VAS score less than 4 except 1 patient in the lumbar group, who was withdrawn from the study at 4 hours because of poor analgesia. Catheter placement was

1474 SAWCHUK ET AL Ann Thorac Surg 1993:55:1472-6 Table 2. Arterial Blood Gas Before and 2 Hours After Operation' Variable Thoracic Group Lumbar Group Pa02 (mm Hg) Preop 79 f 3 74 f 2 Postop (2 h) 111 f llb 92 -t 6b PaCO, (mm Hg) Preop 37 f 1 39 f 2 Postop (2 h) 46 f lb 48 f Zb PH Preop 7.42 2 0.01 7.42 f 0.01 Postop (2 h) 7.35 f 0.01b 7.33 f O.OZb 0 15 30 45 60 4 24 48 72 m i nu tes hours Time Fig 7. Within 75 minutes after the start of epidural fentanyl infusion, the thoracic group had a significant reduction in pain scores. In the lumbar group, 45 minutes elapsed before the reduction in pain scores became significant. Pain relief was sustained throughout the entire study period of 72 hours. WAS = visual analogue scale; * p 5 0.05.) retested with local anesthesia and found to be correct. One of the patients in the thoracic group also had to be removed from the study at 4 hours because of bladder catheter discomfort, which required sedation. There was a significant reduction in the VAS score in the thoracic group by 15 minutes after the first injection compared with 45 minutes for the lumbar group (p I 0.05) (Fig 1). All the patients had adequate pain relief by 60 minutes. Both groups continued to have excellent pain relief even after the fentanyl infusion rates were reduced at 4, 24, and 48 hours after operation. The thoracic group required significantly less fentanyl during the first 4 postoperative hours to achieve similar levels of pain relief, and the dosage of fentanyl required to maintain adequate analgesia could be significantly reduced after 4 hours in both groups (p 5 0.05) (Fig 2). 2.5 7 + Thoracic ----o-- Lumbar I I I I I I I 0 12 24 36 48 60 72 Fig 2. The dosage of fentanyl required at 4 hours was significantly less in the thoracic group. The dosage requirement could be progressively reduced in both groups at 4, 24, and 48 hours without loss of pain control. (* p 5 0.05.) a Data are shown as the mean 2 the standard error of the mean. Significance: p 5 0.05 compared with preop value. PaCO, = arterial carbon dioxide tension; PaO, = arterial oxygen tension. The arterial blood gases were similar for the two groups before operation and at 2 hours after operation (Table 2). At 2 hours postoperatively, the arterial carbon dioxide tension was increased and the ph was decreased to a similar extent in both groups. All patients received supplemental oxygen at 3 Wmin by nasal prongs, and blood gas measurements were done while they were receiving oxygen. The two groups had a similar FEV, and FVC before operation (Fig 3). All patients had a decreased FEV, and FVC in the postoperative period despite good analgesia. The incidence of somnolence and naloxone administration was greater in the lumbar group (Table 3). Three of the 4 patients who received naloxone had respiratory rates of less than 6/min and did not respond to verbal commands and shaking of the shoulders. One of these 3 patients had a carbon dioxide tension of 57 mm Hg at the time naloxone was given. The other 2 patients received naloxone and became alert before arterial blood gases could be measured. The fourth patient was somnolent and had an arterial carbon dioxide tension of 83 mm Hg despite a respiratory rate of 18/min. The 4 patients who received naloxone were not different from the other patients in terms of preoperative blood gases, FEV.,, FVC, age, weight, or total dose of fentanyl received. Five patients in each group had one episode of breakthrough pain in the postoperative period. All responded to a 25% increase in infusion rate. Five patients in the thoracic group did not complete the full study because the epidural catheter fell out perioperatively. One catheter was found to be displaced soon after the end of the operation, one at 36 hours postoperatively, and three catheters after more than 60 hours of service. The data collected on each of these patients up to the time they were pulled from the study were included in calculating the results. Comment In this randomized study, we found that both thoracic and lumbar epidural fentanyl infusions provided good

Ann Thorac Surg 19!93;551472-6 SAWCHUK ET AL 1475 3.5 2.5 1.5 -+- Thoracic -..-o... Lumbar _.._... Preop 24 48 72 3.5 - -0- Thoracic 3- -..-.O.- Lumbar 2.5-2- FVC(1) 1.5 '1 1 O5 0 Preop 24 48 72 Fig 3. Both groups showed significant and sustained drops in forced expiratory volume in 1 second (FEV,) and forced vital capacity (FVC) despite adequate pain relief. ("p 5 0.05.) pain relief after thoracotomy. The patients in the thoracic group obtained significant pain relief within 15 minutes compared with 45 minutes for the lumbar patients. The lumbar patients required more fentanyl during the first 4 hours and had more somnolence and respiratory depression requiring naloxone treatment. We also demonstrated that epidural fentanyl infusion rates can be gradually reduced after operation without compromising pain relief. Table 3. Side Efects and Complications Thoracic Group Lumbar Group Variable (n = 15) (n = 15) p Value Respiratory rate 4 4 NS 10/min Somnolence 1 6 0.04 Naloxone 0 4 0.048 Nausea 5 3 NS Breakthrough pain 5 5 NS Catheter dislodged 5 0 0.03 NS = not significant. The faster onset of pain relief and smaller dose requirements in the thoracic epidural group during the first 4 hours may be explained by the proximity of the site of injection to the site of action. This may produce a higher level of fentanyl in the cerebrospinal fluid at the thoracic levels of the spinal cord. After lumbar administration, a larger dose may be required to produce a similar concentration of fentanyl at the site of desired action in the thoracic region. There is some support for this from the work of Gourlay and associates [8], who measured plasma fentanyl levels and cerebrospinal fluid fentanyl concentrations at the cervical and lumbar levels after lumbar epidural and intravenous injection. The higher concentrations of fentanyl in the cerebrospinal fluid at both levels after a single lumbar epidural administration were similar to plasma fentanyl levels after a similar dose intravenously. The mean maximum cervical cerebrospinal fluid fentanyl concentrations were 10% of the lumbar concentrations. This suggests that there is indeed a cephalad spread of fentanyl after administration in the lumbar region and that systemic absorption and redistribution cannot account for the cerebrospinal fluid level in the cervical region. The thoracic group had fewer episodes of somnolence. However, the incidence of slow respiratory rate (<10/min) was similar in both groups. Of the patients who had lumbar epidural infusions, 3 required naloxone to treat somnolence and slow respiratory rates, and the fourth received naloxone for somnolence and arterial carbon dioxide tension of 83 mm Hg. There was no difference in terms of preoperative breathing capacity and blood gases or fentanyl dosages between the patients who received naloxone and those who did not. Three of the 4 patients were male, a sex distribution similar to that in the lumbar group as a whole. All the patients who received naloxone recovered from the respiratory depression without any permanent sequelae. The epidural fentanyl infusion dosages and rates used in our study are comparable to those established by Badner and associates [2] to be appropriate for patients after thoracotomy. Whereas 2 patients required naloxone 3 to 4 hours after the start of the fentanyl infusion, the other 2 had development of symptoms later at 34 and 40 hours. The epidural fentanyl infusion rates were reduced by 25% at 4 and 24 hours after operation according to our study protocol. Because other patients who received similar and even higher doses of fentanyl did not have these symptoms, excessive dosages did not appear to be the major cause. Renaud [9] showed that resting minute ventilation, tidal volume, and respiratory frequency were not affected by epidural fentanyl infusion. However, the ventilatory response to carbon dioxide and the minute ventilation at a carbon dioxide tension of 55 mm Hg were depressed in young orthopedic patients in his study. Epidural fentanyl has also been shown to cause greater respiratory depression than a similar intramuscular dose [lo]. Profound respiratory depression has also been reported after epidural bupivacaine hydrochloride and fentanyl administrations [ll, 121. Therefore, patients who receive epidural fentanyl should be observed closely.

1476 SAWCHUK ET AL Ann Thorac Surg 1993;5514724 The magnitude of the observed changes in carbon dioxide tension and ph in both the thoracic and lumbar epidural fentanyl groups was similar to the changes in arterial blood gases noted in previous studies of patients receiving epidural morphine and fentanyl after thoracotomy [13, 141. The changes in FEV, and FVC at 24 hours were also similar to those previously reported by Shulman and associates [l]. We did not find any significant differences in FEV, and FVC at 24, 48, and 72 hours between the two groups. Because of the relatively small number of patients studied and the large variations from patient to patient, we have only a 40% power for detecting a 50% difference in these variables. Further studies with larger numbers of patients are needed to determine the effects of epidural fentanyl analgesia on postoperative respiratory function. We decided to stop the study because of the significant difference in the incidence of somnolence and respiratory depression in the lumbar group. We further observed that epidural infusion rates can be gradually reduced without compromising the pain relief. Five of the thoracic epidural catheters were dislodged during the study versus none of the lumbar catheters. All the catheters were secured in a similar fashion, and we believe that the thoracic catheters were subjected to more movement because of proximity to the surgical site and rubbing against the dressing during postoperative mobilization. Coe and co-workers [15] did not observe any difference in dose requirements or incidence of side effects between lumbar and thoracic epidural fentanyl for postthoracotomy pain. These results are contrary to our observations and may be due to differences in surgical procedures. All of our patients had pulmonary procedures, whereas 22 of the 52 patients in the study of Coe and colleagues had esophagogastrectomy or hiatal hernia repair. The more diverse patient group may have masked potential differences in their study. In conclusion, thoracic epidural fentanyl infusion is superior to lumbar infusion for postthoracotomy analgesia because it uses a lower dose and is associated with a lower incidence of adverse side effects. References 1. Shulman M, Sandler AN, Bradley JW. Post-thoracotomy pain and pulmonary function following epidural and systemic morphine. Anesthesiology 1984;61:569-75. 2. Badner NH, Sandler AN, Koren G, et al. Lumbar epidural fentanyl infusions for post-thoracotomy patients: analgesic, respiratory and pharmacokinetic effects. J Cardiothorac Anesth 199O;4:54%51. 3. Melendez JA, Cirella VN, Dephin ES. Lumbar epidural fentanyl analgesia after thoracic surgery. J Cardiothorac Anesth 1989;3: 1 W. 4. Lomessy A, Magnin C, Vaiule JP, et al. Clinical advantages of fentanyl given epidurally for postoperative analgesia. Anesthesiology 1984;61:466-9. 5. Gough JD, Williams AB, Vaughan RS, et al. The control of post-thoracotomy pain. A comparative evaluation of thoracic epidural fentanyl infusions and cryoanalgesia. Anaesthesia 1988;43:780-3. 6. Chamberlain D, Bodily MN, Olssen GL, et al. Comparison of lumbar versus thoracic epidural fentanyl for post-thoracotomy analgesia using patient-controlled dosage. Reg Anaesth 1989;14(2S):26. 7. Chapman CR, Kyrjala ML. Measurement of pain. In: Bonica JJ, ed. The management of pain. Philadelphia: Lea & Febiger, 1990:580-94. 8. Gourlay GK, Murphy TM, Plummer JL, et al. Pharmacokinetics of fentanyl in the lumbar and cervical CSF following lumbar epidural and intravenous administration. Pain 1989; 38253-9. 9. Renaud B. Ventilatory effects of continuous epidural infusions of epidural fentanyl. Anesth Analg 1988;67971-5. 10. Negre I, Gueneron JP, Ecoffey C, et al. Ventilatory response to carbon dioxide after intramuscular and epidural fentanyl. Anesthesiology 1987;66:707-10. 11. Brockway MS, Noble DW, Sharwood-Smith GH, et al. Profound respiratory depression after extradural fentanyl. Br J Anaesth 1990;64:243-5. 12. Weightman WM. Respiratory arrest during epidural infusion of bupivacaine and fentanyl. Anesthesia 1991;19:2824. 13. Sjostrom S. Pharmacokinetics of epidural morphine and meperidine in humans. Anesthesiology 1987;67877-88. 14. Ahuga BR. Respiratory effects of epidural fentanyl. Anaesthesia 1985;40:949-55. 15. Coe A, Sarginson R, Smith MW, et al. Pain following thoracotomy. A randomized, double-blind comparison of lumbar versus thoracic epidural fentanyl. Anaesthesia 1991; 46:91&21.