For mainly economic reasons, the recent trend in cardiac surgery has been to tracheally extubate patients during the immediate postoperative period [1,2]. However, early extubation may not be entirely benign and is not appropriate in certain patients [3,4]. For an early extubation to be safe and successful, at the time of extubation a patient must possess an appropriate sensorium, normothermia, hemodynamic stability, adequate pulmonary function, adequate urine output, and minimal chest tube output. Furthermore, effective pain management is vital to safe and successful early extubation. In fact, aggressive control of pain during the immediate postoperative period after cardiac surgery may decrease morbidity and mortality [5,6]. Intrathecal morphine produces intense and prolonged analgesia by stimulating opioid receptors in the substantia gelatinosa of the posterior spinal cord, and thus may be a useful adjunct in controlling postoperative pain and facilitating early extubation after cardiac surgery. This prospective, randomized, double-blind, placebo-controlled clinical study investigates the use of intrathecal morphine in patients undergoing coronary artery bypass grafting (CABG) surgery for its impact on early extubation and postoperative analgesic requirements.
After institutional review board approval and informed consent, patients scheduled for elective CABG and deemed early extubation candidates participated in the study. At our institution, we view all patients scheduled for elective CABG surgery as early extubation candidates, including those having reoperations and those with decreased left ventricular function (ejection fraction < 40%). Patients requiring preoperative intravenous inotropic drugs, intraaortic balloon support, or mechanical ventilation were excluded from participation in the study. Any patient exhibiting significant pulmonary, endocrine, metabolic, or neurologic pathology were also excluded.
Patients receiving preoperative intravenous (IV) or subcutaneous heparin or having a platelet count of less than 150,000/mm3 were not studied. If a patient had a history of abnormal bleeding or was receiving medications known to affect coagulation, documentation of a normal prothrombin time, a normal partial thromboplastin time, and a normal bleeding time were all required before participation in the study. If a patient had no history of abnormal bleeding and was not receiving medications known to affect coagulation, no preoperative laboratory tests of coagulation were required for participation in the study. By previous agreement with the cooperating cardiac surgeons, if a bloody tap occurred in the operating room, surgery would be delayed 24 h.
All patients received either oral diazepam (10 mg) or oral lorazepam (2 mg) immediately prior to transfer to the operating room and had their cardiac medications continued through the morning of surgery. Prior to arrival to the operating room, each patient was randomized to receive either intrathecal morphine (Group MS) or intrathecal normal saline (Group NS).
Prior to induction of general anesthesia, all patients were placed in the sitting position and the intrathecal space at the L3-4 or L2-3 level was entered (via midline approach) with a 22-gauge spinal needle. Once clear cerebrospinal fluid was obtained, either 10 micro g/kg of morphine (0.5 mg/mL solution diluted with normal saline to a total volume of 3.0 mL) or 3.0 mL of normal saline was injected into the intrathecal space and the needle then removed. All physicians and nursing staff caring for the patients intraoperatively and postoperatively were blinded to treatment group.
The intraoperative anesthetic technique was the same for all patients and consisted of IV fentanyl (20 micro g/kg), midazolam (10 mg, total), and vecuronium. All of the fentanyl was administered prior to sternotomy. Regarding midazolam, 6 mg was administered prior to sternotomy, 2 mg was administered during rewarming, and 2 mg was administered during sternal closure. If required, inhaled isoflurane and/or IV nitroglycerin were used for blood pressure control prior to initiation of cardiopulmonary bypass (CPB). Hypothermic CPB (to a lowest temperature of 26 degrees C) with a membrane oxygenator and crystalloid prime was used in all patients. Nonpulsatile flows were maintained between 2.4-2.8 L [center dot] min-1 [center dot] m-2 and, if needed, isoflurane was used by the perfusionist to maintain perfusion pressure in the range of 50-70 mm Hg. alpha-stat blood gas management was used in all patients. Separation from CPB was facilitated with IV inotropic and/or vasoactive drugs at the discretion of the anesthesiologist managing the case.
After completion of CABG, patients were transferred to the intensive care unit (ICU). Postoperative care was standardized for all patients and extubation was accomplished at the earliest clinically appropriate time. Criteria for extubation in our ICU include an appropriate sensorium, normothermia, hemodynamic stability, adequate pulmonary function, adequate urine output, and minimal chest tube output. If a patient developed hypertension, tachycardia, and/or excessive movement at a time when extubation was not yet appropriate (for any reason), the ICU nurse was free to administer small amounts of IV midazolam. Patients whose tracheas were extubated within 12 h of ICU arrival received exclusively IV patient-controlled morphine analgesia with a dose of 2.0 mg, a lockout interval of 10 min, and a maximum 4-h limit of 30 mg. Patient-controlled analgesia was continued for 48 h. In patients who were not extubated within 24 h of ICU arrival, the reason for prolonged intubation (hemodynamic instability, oxygenation difficulties, etc.) was ascertained.
Postoperative complications and treatments were recorded daily for all patients until hospital discharge. After extubation, patients were directly questioned regarding occurrence of pruritus, nausea, and vomiting. Urinary retention was defined as the need for reinsertion of a urinary catheter because of inability to void after its removal. In general, the urinary catheter was removed the day after extubation. All patients had a creatine phosphokinase (CK) level assayed (via coupled enzymatic reactions based on Rosalki's modification of the Oliver procedure) and an electrocardiogram recorded at 4:00 AM on the first postoperative day. If the CK total was more than 200 IU/L, CK-MB levels were assayed (via monoclonal antibody technique) and the CK-MB Index calculated (CK-MB [divided by] CK-total x 100). Sensitivity of the CK-MB assay in our laboratory is 0.4 ng/mL. If the initial CK total was more than 200 IU/L, two additional CK total and CK-MB levels were measured 6 and 12 h later. Perioperative myocardial infarction was defined as a postoperative CK-MB Index > 3.0 and/or postoperative electrocardiographic evidence (new Q waves or ST segment elevation) of infarction.
Fisher's exact test was applied to categorical data. Student's t-test (two-tailed) was used to test the difference between means in the two groups. A P value of <0.05 was considered statistically significant and P values are reported only when significance was found. Results are expressed as the mean +/- 1 SD, unless otherwise indicated.
Of the 40 patients enrolled and participating in the study, 19 were randomized to Group MS and 21 to Group NS. When compared to Group NS patients, significantly more Group MS patients were receiving diuretics preoperatively (Table 1).
The range of dural puncture attempts in the 40 patients was 1-30 (median 2.0). Not a single bloody tap occurred in any patient. The mean time from dural puncture to systemic heparinization in the 40 patients was 136 +/- 45 min (range 70-280 min). Both groups received similar intraoperative anesthetic and surgical management (Table 2).
Four patients in Group MS could not be tracheally extubated within 24 h of ICU arrival (two secondary to hemodynamic instability, one secondary to oxygenation difficulties, and one secondary to excessive mediastinal bleeding). Five patients in Group NS could not be tracheally extubated within 24 h of ICU arrival (two secondary to hemodynamic instability, one secondary to oxygenation difficulties, one secondary to excessive mediastinal bleeding, and one secondary to cerebral infarction). Of the 31 patients tracheally extubated within 24 h of ICU arrival, the time from ICU arrival to extubation was significantly prolonged in Group MS patients when compared to Group NS patients (654 +/- 267 min, range 300-1080 min versus 456 +/- 153 min, range 220-720 min; P = 0.02). Of these 31 patients, midazolam administration from ICU arrival to extubation in Group MS patients (2.6 +/- 5.4 mg, range 0-21 mg) was similar to Group NS patients (2.5 +/- 3.4 mg, range 0-12 mg).
Twelve patients in Group MS and 16 patients in Group NS were extubated within 12 h of ICU arrival and received exclusively IV morphine via patient-controlled analgesia for 48 h (Table 3). Although postoperative IV morphine use was less in Group MS patients when compared to Group NS patients, the difference did not reach statistical significance. The three patients in Group MS who were tracheally extubated between 12 and 24 h of ICU arrival had extubation delayed because of prolonged ventilatory depression.
Of the two Group MS patients who experienced pruritus, one (who had a preoperative history of intractable pruritus) required treatment with diphenhydramine and one required treatment with diphenhydramine and naloxone (Table 4). Of the seven patients experiencing nausea, treatment with antiemetics was required in none of the Group MS patients and two Group NS patients. One Group MS patient who complained of nausea subsequently vomited. Both Group NS patients who were treated with antiemetics subsequently vomited. No patient had their postoperative course complicated by prolonged nausea and vomiting. Urinary retention occurred in one Group NS patient. In this male patient, the urinary catheter was removed on the first postoperative day, reinserted 8 h later, and then finally removed on the third postoperative day.
Postoperative atrial fibrillation occurred in six Group MS patients and nine Group NS patients, despite lack of preoperative atrial fibrillation in any of these patients. Pharmacologic therapy was required in all patients in both groups and cardioversion was required in one Group MS patient and four Group NS patients. Furthermore, one patient (Group MS) required atrioventricular node ablation on the seventh postoperative day. Postoperative ventricular tachycardia occurred in two patients in each group. Of the two Group MS patients, one required multiple defibrillations and subsequent treatment with amiodarone and one required no treatment. Of the two Group NS patients, one required treatment with lidocaine and subsequent insertion of an automatic implantable cardioverter defibrillator on the twelfth postoperative day and one required no treatment.
Five patients in Group MS and seven patients in Group NS experienced perioperative myocardial infarctions (defined as a postoperative CK-MB Index > 3.0). Of these 12 patients, six (two in Group MS and four in Group NS) also demonstrated postoperative electrocardiographic evidence of myocardial infarction (new Q waves or ST segment elevation). Three of the 12 patients who experienced perioperative myocardial injury sustained substantial cardiac morbidity. One patient (Group MS) required placement of an intraaortic arch balloon pump to facilitate separation from CPB and developed postoperative ventricular tachycardia requiring multiple defibrillations and subsequent treatment with amiodarone. One patient (Group NS) experienced intraoperative asystole after chest closure secondary to a clotted vein graft which required immediate mediastinal exploration, initiation of CPB, and placement of an intraaortic balloon pump to facilitate separation from CPB. The third patient (Group NS) developed postoperative atrial fibrillation requiring pharmacologic therapy and ventricular tachycardia requiring treatment with lidocaine and subsequent insertion of an automatic implantable cardioverter defibrillator on the twelfth postoperative day.
One patient in Group NS developed a perioperative cerebral infarction. This patient experienced intraoperative asystole after chest closure secondary to a clotted vein graft which was associated with approximately 20 min of profound hypotension. Postoperatively, no meaningful neurologic activity returned and the unresponsive patient was eventually discharged from the hospital to a chronic care facility 62 days postoperatively.
One patient in each group experienced intraoperative awareness. The Group MS patient, although unable to detail specific events, offered a vague (yet somewhat accurate) description of the intraoperative environment. The Group NS patient detailed specific intraoperative events, including verbal comments made by operating room personnel. In both patients, isoflurane was used prior to CPB, yet not used after separation from CPB.
All patients enrolled and participating in the study were eventually discharged from the hospital. The mean duration of postoperative hospital stay in Group MS (7.4 +/- 3.2 days, range 4-18 days) was no different than in Group NS (10.3 +/- 12.8 days, range 4-62 days). Seventeen patients in Group MS and 15 patients in Group NS were discharged from the hospital in 10 days or less.
The quality of postoperative analgesia and its relationship to outcome has recently received much attention [7-9]. Inadequate analgesia during the postoperative period may lead to many adverse hemodynamic (tachycardia, hypertension, vasoconstriction), respiratory (tachypnea, decreased tidal volume), metabolic (increased catabolism), immunologic (impaired immune response), and hemostatic (platelet activation) alterations [7-9]. Aggressive control of postoperative pain has been shown to decrease morbidity and mortality in high-risk patients after noncardiac surgery [10,11].
Driven by economic reasons, the recent trend in cardiac surgery has been to extubate patients during the immediate postoperative period [1,2]. However, in patients undergoing cardiac surgery, perioperative myocardial ischemia (diagnosed by electrocardiography and/or transesophageal echocardiography) is most commonly observed during this time (25%-38% incidence) and is likely related to outcome [12,13]. For this and other reasons, early extubation may be associated with risk and is not appropriate in certain patients [3,4]. When early extubation is planned, aggressive control of postoperative pain is essential. Just as in high-risk patients after noncardiac surgery [10,11], aggressive control of pain during the immediate postoperative period in patients after cardiac surgery may also decrease morbidity and mortality [5,6]. Mangano, et al.  demonstrated, in adults undergoing CABG, that aggressive control of pain (continuous IV sufentanil infusion) during the immediate postoperative period (first 18 hours) decreased the incidence and severity of myocardial ischemia diagnosed by electrocardiography. Furthermore, Anand and Hickey  demonstrated in neonates undergoing cardiac surgery that aggressive control of pain (continuous IV fentanyl or sufentanil infusion) during the immediate postoperative period (first 24 hours) decreased morbidity and mortality.
Intrathecal morphine produces intense and prolonged analgesia by stimulating opioid receptors in the substantia gelatinosa of the posterior spinal cord. The technique is simple, reliable, and produces predictable pharmacodynamic effects . The administration of intrathecal morphine to patients prior to cardiac surgery, by producing intense analgesia in the immediate postoperative period, may therefore facilitate early extubation and beneficially affect outcome.
Administration of intrathecal morphine prior to cardiac surgery is not a new technique [15-26]. However, of 12 citations, three are retrospective reports [17-19], four are observational studies lacking placebo controls [20,23-25], one is a case report , and one is simply a description of the technique in the Correspondence section of the publishing journal . The only three prospective, randomized, blinded, placebo-controlled studies investigating the use of intrathecal morphine for cardiac surgery are contradictory [15,21,22]. Chaney et al.  and Vanstrum et al.  demonstrated that administration of intrathecal morphine (4.0 mg and 0.5 mg, respectively) immediately prior to induction of anesthesia decreased postoperative analgesic requirements when compared to intrathecal placebo. On the other hand, Casey et al.  found that administration of intrathecal morphine (20 micro g/kg) to patients immediately after tracheal intubation did not decrease postoperative analgesic requirements when compared to intrathecal placebo. These three studies involved overnight mechanical ventilation of patients and thus did not investigate the ability of intrathecal morphine to facilitate early extubation.
Nine of the 40 patients participating in this study could not be tracheally extubated within 24 hours of ICU arrival. The main reason for inability to achieve extubation during the immediate postoperative period was our inclusive definition of an early extubation candidate. The characteristics that define the optimal early extubation candidate after cardiac surgery are controversial [1-4]. At our institution, all patients scheduled for elective CABG surgery are viewed as early extubation candidates, including those having reoperations and those with decreased left ventricular function (ejection fraction < 40%). Two of the four patients participating in this study who were undergoing reoperation were not able to be tracheally extubated within 24 hours of ICU arrival (one secondary to oxygenation difficulties and one secondary to excessive mediastinal bleeding). All four patients who could not be tracheally extubated within 24 hours of ICU arrival secondary to hemodynamic instability had preexisting decreased left ventricular function.
Of the 31 patients tracheally extubated within 24 hours of ICU arrival, the mean extubation time was significantly prolonged in Group MS patients when compared to Group NS patients (10.9 hours versus 7.6 hours, respectively). Twelve of 15 Group MS patients and all 16 Group NS patients tracheally extubated during the immediate postoperative period were extubated within 12 hours of ICU arrival. The three patients in Group MS who were tracheally extubated between 12 and 24 hours of ICU arrival had their extubation delayed because of prolonged ventilatory depression. Intrathecal morphine initiates dosedependent ventilatory depression which is intensified by concomitant administration of IV analgesics and sedatives . Other factors known to increase the risk of ventilatory depression after intrathecal morphine are advanced age, significant coexisting disease, lack of opioid tolerance, general anesthesia, increased intrathoracic pressure, and patient position . The prolonged mean extubation time in Group MS patients was likely caused by the additive ventilatory depressant effects of intrathecal morphine and intravenous fentanyl used for the intraoperative anesthetic. In the primary author's clinical practice involving thoracotomies, the use of the same amount of intrathecal morphine as used in this study (10 micro g/kg) and smaller doses of IV fentanyl (2-4 micro g/kg) allows extubation of patients in the operating room at the end of surgery. Perhaps if smaller amounts of IV fentanyl had been used for the intraoperative anesthetic in this study, extubation times between the two groups would not have been different.
Although postoperative IV morphine use was less in Group MS patients when compared to Group NS patients, the difference was not statistically significant. This study originally intended to investigate 100 patients. However, for reasons stated below, we altered the protocol after interim analysis of data from the first 40 patients. Perhaps if all 100 patients would have been studied with the original protocol, the difference in postoperative IV morphine use between groups would have reached statistical significance. Larger amounts of intrathecal morphine than used in this study likely would have initiated more intense postoperative analgesia in Group MS patients . However, increasing the amount of intrathecal morphine would likely have also extended extubation times even further in this group. The postoperative analgesic benefit of intrathecal morphine may have been masked by the moderate amount of IV fentanyl used for the intraoperative baseline anesthetic. Perhaps if smaller amounts of IV fentanyl had been used for the intraoperative anesthetic in this study or a larger number of patients studied, the difference in postoperative IV morphine use between the two groups would have been significantly different.
Whether or not aggressive control of postoperative pain in patients after cardiac surgery affects outcome remains unsettled [5,6]. Morbidity and mortality may also be decreased after cardiac surgery by attenuating the stress response to CPB [5,6]. It is highly unlikely that the dose of intrathecal morphine used in this study (10 micro g/kg) significantly attenuated the stress response to CPB, for previous work done by our group revealed that much larger doses (4.0 mg) of intrathecal morphine were unable to significantly attenuate this response, as measured by blood norepinephrine and epinephrine levels . In this study, there was no difference in morbidity and mortality between the two groups, likely secondary to small sample size.
The four classic side effects of intrathecal morphine are pruritus, nausea and vomiting, urinary retention, and ventilatory depression . Despite the relatively large dose of intrathecal morphine used, none of the patients in Group MS (including the one who required treatment with naloxone) experienced severe pruritus. It is impossible to attribute postoperative nausea and vomiting observed in both groups to either intrathecal or IV morphine, since the incidence of nausea and vomiting is similar with either mode of opioid delivery . No patient in Group MS experienced urinary retention after removal of the urinary catheter or ventilatory depression after tracheal extubation.
Reports of intraoperative awareness during cardiac surgery are quite variable in terms of patient characteristics, medications used intraoperatively, temporal onset and sensations reported, and postoperative psychologic sequelae . The Group MS patient who experienced awareness was a 34-year-old, 81-kg female who described vague sensations of pain and hearing music during the operation. This patient was taking a hypnotic drug (flurazepam) and an antidepressant (sertraline) on a regular basis preoperatively. The Group NS patient who experienced awareness was a 48-year-old, 100-kg female who described distinct sensations of pulling on the chest and verbal comments made by operating room personnel during the operation. This patient was not taking any psychotropic medications on a regular basis preoperatively. In both patients, isoflurane was used prior to CPB but not after separation from CPB. In neither patient did postoperative psychologic sequelae develop. Perhaps if larger amounts of IV midazolam had been used for the intraoperative baseline anesthetic and/or isoflurane had been used after separation from CPB, intraoperative awareness in these two patients would not have occurred.
This study originally intended to investigate 100 patients. However, after analysis of data from the first 40 patients which revealed a significant prolongation of extubation time in Group MS patients when compared to Group NS patients, we decided to alter the protocol by revising the intraoperative anesthetic. Subsequent patients now receive substantially less IV fentanyl and slightly more IV midazolam intraoperatively. Hopefully, with these two alterations, postoperative extubation will not be delayed in patients receiving intrathecal morphine and intraoperative awareness will not occur.
The safety of an intrathecal injection immediately prior to systemic heparinization required for CPB deserves comment. Although extremely rare, subarachnoid and epidural hematomas have occurred in patients when a diagnostic or therapeutic lumbar puncture has been followed by systemic heparinization [29-31]. None of the previous citations investigating the use of intrathecal morphine for cardiac surgery report development of subarachnoid or epidural hematoma [15-26], yet the risk may be as high as 0.35% . However, certain precautions may decrease risk [29,32]. It has been recommended that the technique should not be used in patients who demonstrate a known preoperative coagulopathy from any cause, surgery should be delayed 24 hours if a bloody tap occurs, and the time from lumbar puncture to systemic heparinization should exceed 60 minutes [29,32]. A potential limitation of using intrathecal morphine for cardiac surgery is possible delay of surgery in the event of a bloody tap. We observed no clinical evidence of central neuraxial hematoma formation in any patient enrolled in our investigation. Furthermore, although data are not presented, Fitzpatrick and Moriarty  have reported use of the technique in more than 3400 patients and Vanstrum et al.  in approximately 1000 patients without occurrence of subarachnoid or epidural hematoma.
In conclusion, intrathecal morphine offers promise as a useful adjunct in controlling postoperative pain in patients after cardiac surgery. The technique has been applied to a large number of patients undergoing cardiac surgery [15-26] and appears to be safe, providing certain precautions are observed. However, the optimal dose of intrathecal morphine in this setting, along with the optimal intraoperative anesthetic that will provide significant analgesia yet not delay extubation in the immediate postoperative period, remains to be elucidated. Unless the technique is "fined-tuned" to achieve these two goals, the use of intrathecal morphine in patients undergoing cardiac surgery may delay early extubation of the trachea.
1. Cheng DCH. Early extubation after cardiac surgery decreases intensive care unit stay and cost. Pro: early extubation after cardiac surgery decreases intensive care unit stay and cost. J Cardiothorac Vasc Anesth 1995;9:460-4.
2. Higgins TL. Pro: early endotracheal extubation is preferable to late extubation in patients following coronary artery surgery. J Cardiothorac Vasc Anesth 1992;6:488-93.
3. Guenther CR. Early extubation after cardiac surgery decreases intensive care unit stay and cost. Con: early extubation after cardiac surgery does not decrease intensive care unit stay and cost. J Cardiothorac Vasc Anesth 1995;9:465-7.
4. Siliciano D. Con: early extubation is not preferable to late extubation in patients undergoing coronary artery surgery. J Cardiothorac Vasc Anesth 1992;6:494-8.
5. Mangano DT, Siliciano D, Hollenberg M, et al. Postoperative myocardial ischemia; therapeutic trials using intensive analgesia following surgery. Anesthesiology 1992;76:342-53.
6. Anand KJS, Hickey PR. Halothane-morphine compared with high-dose sufentanil for anesthesia and postoperative analgesia in neonatal cardiac surgery. N Engl J Med 1992;326:1-9.
7. Liu S, Carpenter RL, Neal JM. Epidural anesthesia and analgesia; their role in postoperative outcome. Anesthesiology 1995;82:1474-506.
8. Weissman C. The metabolic response to stress: an overview and update. Anesthesiology 1990;73:308-27.
9. Kehlet H. Surgical stress: the role of pain and analgesia. Br J Anaesth 1989;63:189-95.
10. Tuman KJ, McCarthy RJ, March RJ, et al. Effects of epidural anesthesia and analgesia on coagulation and outcome after major vascular surgery. Anesth Analg 1991;73:696-704.
11. Yeager MP, Glass DD, Neff RK, Brinck-Johnsen T. Epidural anesthesia and analgesia in high-risk surgical patients. Anesthesiology 1987;66:729-36.
12. Smith RC, Leung JM, Mangano DT, SPI Research Group. Postoperative myocardial ischemia in patients undergoing coronary artery bypass graft surgery. Anesthesiology 1991;74:464-73.
13. Leung JM, O'Kelly B, Browner WS, et al. Prognostic importance of postbypass regional wall-motion abnormalities in patients undergoing coronary artery bypass graft surgery. Anesthesiology 1989;71:16-25.
14. Stoelting RK. Intrathecal morphine-an underused combination for postoperative pain management [editorial]. Anesth Analg 1989;68:707-9.
15. Chaney MA, Smith KR, Barclay JC, Slogoff S. Large-dose intrathecal morphine for coronary artery bypass grafting. Anesth Analg 1996;83:215-22.
16. Kowalewski R, MacAdams C, Froelich J, et al. Anesthesia supplemented with subarachnoid bupivacaine and morphine for coronary artery bypass surgery in a child with Kawasaki disease. J Cardiothorac Vasc Anesth 1996;10:243-6.
17. Taylor A, Healy M, McCarroll M, Moriarty DC. Intrathecal morphine: one year's experience in cardiac surgical patients. J Cardiothorac Vasc Anesth 1996;10:225-8.
18. Kowalewski RJ, MacAdams CL, Eagle CJ, et al. Anaesthesia for coronary artery bypass surgery supplemented with subarachnoid bupivacaine and morphine: a report of 18 cases. Can J Anaesth 1994;41:1189-95.
19. Swenson JD, Hullander RM, Wingler K, Leivers D. Early extubation after cardiac surgery using combined intrathecal sufentanil and morphine. J Cardiothorac Vasc Anesth 1994;8:509-14.
20. Fitzpatrick GJ, Moriarty DC. Intrathecal morphine in the management of pain following cardiac surgery; a comparison with morphine i.v. Br J Anaesth 1988;60:639-44.
21. Vanstrum GS, Bjornson KM, Ilko R. Postoperative effects of intrathecal morphine in coronary artery bypass surgery. Anesth Analg 1988;67:261-7.
22. Casey WF, Wynands JE, Ralley FE, et al. The role of intrathecal morphine in the anesthetic management of patients undergoing coronary artery bypass surgery. J Cardiothorac Anesth 1987;1:510-6.
23. Cheun JK. Intraspinal narcotic anesthesia in open heart surgery. J Korean Med Sci 1987;2:225-9.
24. Aun D, Thomas D, St. John-Jones L, et al. Intrathecal morphine in cardiac surgery. Eur J Anaesthesiol 1985;2:419-26.
25. Jones SEF, Beasley JM, Macfarlane WR, et al. Intrathecal morphine for postoperative pain relief in children. Br J Anaesth 1984;56:137-40.
26. Mathews ET, Abrams LD. Intrathecal morphine in open heart surgery [correspondence]. Lancet 1980;2:543.
27. Chaney MA. Side effects of intrathecal and epidural opioids. Can J Anaesth 1995;42:891-903.
28. Wong KC. Narcotics are not expected to produce unconsciousness and amnesia [editorial]. Anesth Analg 1983;62:625-6.
29. Owens EL, Kasten GW, Hessel EA. Spinal subarachnoid hematoma after lumbar puncture and heparinization: a case report, review of the literature, and discussion of anesthetic implications. Anesth Analg 1986;65:1201-7.
30. Ruff RL, Dougherty JH. Complications of lumbar puncture followed by anticoagulation. Stroke 1981;12:879-81.
31. Brem SS, Hafler DA, Van Uitert RL, et al. Spinal subarachnoid hematoma; a hazard of lumbar puncture resulting in reversible paraplegia. N Engl J Med 1981;303:1020-1.
32. Vandermeulen EP, Van Aken H, Vermylen J. Anticoagulants and spinal-epidural anesthesia. Anesth Analg 1994;79:1165-77.