Intraoperative transesophageal echocardiography in cardiovascular surgery. Consensus document from the Spanish Society of Anesthesia and Critical Care (SEDAR) and the Spanish Society of Endovascular and Cardiovascular Surgery (SECCE)

Transesophageal echocardiography is a semi-invasive technique that allows an evaluation of cardiac morphology and function in real time and it is a quality standard in cardiovascular surgery. It has become a fundamental tool for both monitoring and diagnosis in the intraoperative period that allows decide the correct surgical planning and pharmacological management. The goal of this document is to answer the questions of when and how the perioperative TEE should be performed in cardiovascular surgery, what are their applications in the intraoperative, who should perform it and how the information should be transmitted. The authors made a systematic review of international guidelines, review articles and clinical trials to answer by consensus to these questions.

The importance of intraoperative transesophageal echocardiography in the surgical decision in cardiac surgery

OBJECTIVE

To determine the importance of intraoperative transesophageal echocardiography (IOTEE) in the surgical decision in patients undergoing cardiac surgery.

PATIENTS AND METHOD

Prospective observational study of patients undergoing cardiac surgery from January 2009 to May 2012, which was monitored with IOTEE by the anesthesiologist in charge. The data collected were: 1) type of surgery; 2) preoperative echocardiographic diagnosis (baseline ECHO); 3) echocardiographic diagnosis before entering cardiopulmonary bypass (CPB) (pre-CPB IOTEE); 4) any differences between the baseline ECHO and the pre-CPB IOTEE (new pre-CPB finding) and whether these differences modified the planned surgery, and 5) echocardiographic diagnosis after disconnection of CPB (unexpected post-CPB finding) and whether these post-CPB echocardiographic findings led to reinstating it. The software program SPSS(®) was used for data analysis.

RESULTS

The total number of patients studied was 1,273. Monitoring with IOTEE showed "new pre-CPB" findings in 98 patients (7.7%), and 43.8% of these led to a change in the scheduled surgery. Of these findings, the most frequent were abnormalities of the mitral valve that had not been diagnosed, and which led to a replacement or repair that had not been scheduled. The incidence of "unexpected post-CPB findings" was 6.2% (79 patients), and 46.8% of those required reinstating the CPB and modifying the surgery performed. The failed valve repairs and dysfunctional valve prostheses were the main causes that led to re-entry into CPB. In the remaining 42 patients, with "unexpected post-CPB findings", there were no changes in the surgical procedure as the echocardiographic findings were not considered to be significant enough to re-establish CPB and revise or change the surgical procedure.

CONCLUSION

Intraoperative monitoring with IOTEE by the anesthesiologist during surgery provides important information before and after the CPB that resulted in modifying surgical management.

[Perioperative management of atrial fibrillation]

Atrial fibrillation is a frequent complication in the perioperative period. When it appears there is an increased risk of perioperative morbidity due to stroke, thromboembolism, cardiac arrest, myocardial infarction, anticoagulation haemorrhage, and hospital readmissions. The current article focuses on the recommendations for the management of perioperative atrial fibrillation based on the latest Clinical Practice Guidelines on atrial fibrillation by the European Society of Cardiology and the Spanish Society of Cardiology. This article pays special attention to the preoperative management, as well as to the acute perioperative episode. For this reason, the latest recommendations for the control of cardiac frequency, antiarrhythmic treatment and anticoagulation are included.

Tricuspid valve surgery

INTRODUCTION

The tricuspid valve has been taken as a non-critical valve in terms of acute or late mortality in a number of conditions. Tricuspid functional regurgitation is a cause of late operations with an increased morbidity. A number of techniques have been described and used in clinical practice in the past forty years and include simple suture techniques and the use of support for annuloplasty with the use of different types of prosthetic rings. The experience accumulated over the years indicates that tricuspid annuloplasty is mandatory to improve late results, which are superior, in general, to replacement of the valve.

METHODS

The role of echocardiography in defining surgical planning, intraoperative results and follow-up is reviewed as echocardiography is a fundamental tool in cardiac surgery. Surgery for isolated lesions of the tricuspid valve has not received much attention and herein we report the results of the follow-up of a limited series of patients undergoing isolated tricuspid surgery.

RESULTS

The correlation between echocardiographic measurements and surgical measurements was confirmed and was helpful at the time of the confirmation of repair (r=0.53). Forty-seven patients (18 repair, 29 replacement) underwent isolated surgery. Results of isolated tricuspid repair seemed to be superior when compared to those of tricuspid replacement. Survival was 20.7% for tricuspid valve replacement (N=18) and 50% for tricuspid valve repair (N=29) (p=0.04). Freedom from reoperation was 94.4±5.4% for repair and 67.3±12.1% for replacement (p= 0.0011).

CONCLUSIONS

The tricuspid valve continues to be a surgical challenge.

[Survey of perioperative hemostasis and transfusion management in cardiac surgery: how do anesthesiologists practice?]

OBJECTIVE

To determine practices related to control of perioperative hemostasis and transfusion in patients undergoing cardiac surgery in Spain, including the extent to which protocols are being used.

METHODS

A questionnaire was created to collect information from physicians in anesthesiology and postoperative recovery care between July 1 and September 20, 2007. The physicians were asked about practice in the 12 months prior to the survey.

RESULTS

Thirty-four hospitals responded. Seventy percent reported that they did not have protocols or guidelines for the control of hemostasis during cardiac surgery. Forty-four percent did not have information on the proportion of patients who received transfusions; 47% gave transfusions to 75% of patients. The standard preoperative tests were platelet counts, activated partial thromboplastin time, and prothrombin time. Acetylsalicylic acid and clopidogrel were suspended before surgery at 15 (44%) and 25 (73%) hospitals, respectively. In cases of resistance to heparin, additional doses of the drug were injected, in combination with plasma or antithrombin in 29% and 12% of the hospitals, respectively. In the intensive postoperative recovery care unit, only 1 hospital used thromboelastography. Only 1 other hospital used a platelet function analyzer.

CONCLUSIONS

Hemostasis, perioperative coagulation, and criteria for transfusion vary widely among the hospitals surveyed. Few guidelines are available and they are not often being followed. A high percentage of patients receive transfusions, although not all hospitals can cite a figure. New technology has not been widely applied.

[Serious mechanical complications associated with pulmonary artery catheters in cardiovascular and thoracic surgery]

OBJECTIVE

Although the use of pulmonary artery catheters (PACs) in managing critical patients is a subject of debate, they continue to be inserted in many cases and possible complications should be taken into account. Our objective was to review the serious or potentially serious complications associated with PACs in our hospital in the past 15 years.

PATIENTS AND METHODS

This was a retrospective study of seious mechanical complications of PAC use in patients who underwent vascular, cardiac, and thoracic surgery.

RESULTS

The study included the records 7540 patients; 9 cases of serious complications were detected. These complications included 5 cases of pulmonary artery rupture (3 of which resulted in death), 1 perforated internal mammary vein, 1 knotted catheter, 1 bent one, and 1 case of a PAC becoming trapped in the surgical suture.

CONCLUSIONS

The 0.12% incidence of complications is lower than rates found in the literature. Although these complications are rare, it is necessary to take precautions against their unexcepted appearance by carefully selecting the patients in whom PACs are placed and by paying special attention to the characteristic clinical and radiological signs of complications.

Transoesophageal echocardiography accurately detects cardiac output variation: a prospective comparison with thermodilution in cardiac surgery

BACKGROUND AND OBJECTIVE

Intraoperative Doppler ultrasound can be used to measure cardiac output by transoesophageal echocardiography. Recently, its reliability, when compared to the thermodilution technique, has been questioned. The purpose of this study was to compare intraoperative changes in cardiac output measured by echo-Doppler and by thermodilution in cardiac surgery. We also assessed the agreement between the techniques.

METHODS

Fifty cardiac surgical patients (38 male, 12 female, mean age of 63.4 +/- 14.3 yr) were prospectively included after approval by the Ethics Committee of the Institution. Cardiac output was assessed by thermodilution, with 10 mL saline at 12 degrees C, and simultaneously and blindly by echo-Doppler in deep transgastric view with pulsed wave Doppler at the level of the left ventricular outflow tract. Matched thermodilution cardiac output and echo-Doppler cardiac output measurements were taken three times at the end of expiration, both pre- and post-cardiopulmonary bypass.

RESULTS

Echo-Doppler measurements were obtained in 44 patients (88%). In three patients, Doppler recordings could not be obtained adequately, and three developed left ventricular outflow tract obstruction after bypass. Bland-Altman analysis revealed a bias of 0.015 L min(-1), with narrow limits of agreement (-1.21 to 1.22 L min(-1)) and 29.1% error. Echo-Doppler was accurate (92% sensitivity and 71% specificity, P = 0.008 by receiver operating characteristic curves) for detecting more than 10% of change in thermodilution cardiac output. There were no complications related to the study.

CONCLUSIONS

The agreement between cardiac output by echo-Doppler and by thermodilution is clinically acceptable and transoesophageal echocardiography is a reliable tool to assess significant cardiac output changes in a population of selected patients.

Cardiovascular Response and Anesthetic Recovery in Electroconvulsive Therapy with Propofol or Thiopental

Propofol provokes a slight hypotensive effect that could mitigate the cardiovascular response to electroconvulsive therapy (ECT). In this study we compared the effects of propofol and thiopental for ECT anesthesia in seven women (22-67 years of age). Anesthesia was induced with either thiopental or propofol, and with atropine and suxamethonium for each treatment. The first anesthesia was assigned to thiopental or propofol at random; the next anesthesia was induced with the other drug, and alternated thereafter. Systolic blood pressure, diastolic blood pressure (DBP), and heart rate (HR) were recorded before anesthesia, after anesthetic induction, and 1 and 5 min after ECT. ECT-induced increases in DBP and HR were less marked with propofol than with thiopental. Seizure durations were decreased with propofol compared with thiopental.

Plasma magnesium in patients submitted to cardiac surgery and its influence on perioperative morbidity

BACKGROUND

To determine the changes in magnesaemia in cardiac surgical patients submitted to cardiopulmonary bypass (CPB) and their influence on perioperative morbidity.

METHODS

SETTING

the cardiovascular surgery department of a university hospital.

PATIENTS

60 patients of both sexes, mean age 60+/-12 yrs, operated on consecutively for myocardial revascularization or valve replacement.

INTERVENTIONS

plasma Mg2+ levels were measured preoperatively, during CPB, postCPB and throughout the first 24 hrs after operation. Preoperative plasma Mg2+ levels of these patients were compared with those of 15 non-cardiac surgical patients and 11 healthy volunteers.

RESULTS

Mean values of Mg2+ similar in the three populations although in the group of cardiac patients the number of hypomagnesaemic patients was significantly higher (16 patients=26.6%). In these 16 patients, preoperative hypomagnesaemia had a statistically significant relationship with the preoperative treatment with beta-blockers and previous history of arrhythmias (p<0.05). A progressive statistically significant decrease of Mg2+ was observed throughout the surgery that remained low at 24 hours postoperatively (p<0.05). Normomagnesemic patients needed significantly more shocks and electrical energy to obtain heart defibrillation after CPB. The incidence of both postoperative arrhythmias and postoperative low cardiac index (<2.5 L.m2) was statistically significantly more frequent in hypomagnesaemic patients (p<0.05).

CONCLUSIONS

Preoperative hypomagnesaemia was more frequent in this small sample of cardiac surgical patients than in non-cardiac surgical patients and was related to preoperative treatment with b-blockers. Hypomagne-saemia caused by CPB persisted 24 hrs after operation and was associated with higher incidence of both postoperative arrhythmias and low cardiac index.

[Changes in oxygen saturation in the jugular bulb during cardiac surgery]

OBJECTIVE

Heart surgery with cardiopulmonary bypass (CPB) leads to changes in supply and consumption of cerebral oxygen (DO2 and VO2C). Monitoring jugular bulb oxygen saturation (SjO2) detects changes in the DO2C/VO2C ratio that occur in patients undergoing heart surgery. The objective of this study was to determine the evolution of SjO2, of the arteriovenous difference of cerebral oxygen and of cerebral oxygen extraction, as well as the possible relation between those variables and changes in mean arterial pressure, hemoglobin counts and temperature in patients undergoing heart surgery with CPB.

PATIENTS AND METHOD

A prospective study carried out in 31 patients who underwent coronary valve surgery. To monitor SjO2, each patient's internal jugular vein was cannulated with an oximetric catheter in a retrograde direction to monitor SjO2.

RESULTS

Baseline SjO2 (68 +/- 7.4%), obtained after anesthetic induction, was similar to SjO2 before (65 +/- 6%) and after (67 +/- 8.2%) CPB. However, SjO2 upon starting CPB (60 +/- 8.6%) and during rewarming (63 +/- 3%) were significantly lower than at baseline. SjO2 was significantly higher during hypothermic bypass (78 +/- 5%) than at baseline. SjO2 ranged from a low of 60 +/- 8% as CPB was initiated to a high of 78 +/- 5% during hypothermic CPB. Mean arterial pressure was significantly lower at the start of bypass (44 +/- 6 mmHg) than anesthetic induction (83.5 +/- 13.1 mmHg) and the decrease correlated with a significant decrease in SjO2. Changes in mean arterial pressure were unrelated to significant changes in SjO2 at other moments, however. Nor was there a significant relation between changes in temperature or hemoglobin and the evolution of SjO2. At least one episode of SjO2 desaturation (= 50%) occurred in 29% of the patients, with the lowest values being recorded at the start of CPB and during rewarming.

CONCLUSIONS

The greatest risk of cerebral oxygen imbalance between supply and demand occurs at the start of CPB and during rewarming, as shown by decreases in SjO2 levels below baseline at those times.

[Esmolol in anesthesiology: pharmacology and indications]

Esmolol is a beta-adrenergic receptor antagonist with a relatively specific affinity for beta 1 adrenergic receptors. Its mechanism of action is therefore largely cardioselective and only high doses block beta 2 adrenergic receptors. The pharmacologic features of the drug give it rapid onset of beta-blocking action (distribution half-life = 2 minutes) and a short duration of action due to rapid clearance (clearance half-life = 9 minutes). The rapid metabolism of esmolol allows its beta-blocking activity to be lowered rapidly by changing the rate of infusion and obtaining rapid reversibility of effect in the minutes following interruption of the infusion. The esmolol dose is therefore manageable and individual adjustments can be made in function of a patient's clinical status. Such properties mean that esmolol is indicated for short-term treatment of hypertension and tachycardia during the perioperative period and in clinical situations that require easy unblocking of beta receptors. Hypertension and bradycardia are the most frequent complications associated with the administration of esmolol, such that blood pressure, heart rate and electrocardiographic data must be monitored.

Platelet function during cardiac surgery and cardiopulmonary bypass with low-dose aprotinin

OBJECTIVE

To determine whether two low-dose regimens of aprotinin influence platelet function.

DESIGN

Prospective, randomized, single-blinded trial.

SETTING

University teaching hospital performing 600 cardiac operations per year.

PARTICIPANTS

Fifty-nine patients scheduled for cardiac surgery undergoing cardiopulmonary bypass (CPB) of expected duration of 60 minutes or more.

INTERVENTIONS

Patients were randomized into three groups. Group C (control) included 21 patients who did not receive aprotinin. In group A2, 17 patients received 14,286 kallikrein inhibitor units (KIU)/kg (2 mg/kg) of aprotinin before surgery, followed by a continuous infusion of 7,143 KIU/kg/h (1 mg/kg/h) until the end of surgery. In group A4, 19 patients received 28,572 KIU/kg (4 mg/kg) of aprotinin before surgery, followed by the same infusion.

MEASUREMENTS AND MAIN RESULTS

Postoperative bleeding and transfusion requirements were significantly less in group A4. Changes in platelet number and function were similar in the three groups. Platelet aggregation was assessed in four periods: before CPB (T1), post-CPB (T2), and 2 hours (T3) and 4 hours (T4) after CPB. Platelet aggregation induced by adenosine diphosphate, 1 and 2 micromol/L; ristocetin, 1 mg/mL; and arachadonic acid (AA), 1.4 mmol/L, decreased at T2 (p < 0.001) in all groups, and for the ristocetin and AA groups, remained at less than baseline values at T3 and T4. In five patients from each group, platelet receptors for glycoprotein IIb-IIIa (GPIIb-IIIa) and expression of platelet activation markers, guanosine monophosphate 140 (GMP-140) and lysosomal protein, were measured by flow cytometry before and after CPB. Modifications in the expression of GPIIb-IIIa were always modest and without statistical significance. Platelet activation markers, GMP-140 or lysosomal protein, nearly doubled from baseline to post-CPB only in the A4 group, whereas they remained stable in both other groups (statistically not significant).

CONCLUSION

The two regimens of aprotinin, both considered low dosage, did not exert a protective effect on platelet function. Neither dose produced changes in platelet GPIIb-IIIa or platelet activation markers. However, bleeding and transfusion needs were decreased.

[Epidural hematoma after removal of an epidural catheter]

Epidural hematoma is a rare but serious neurological complication of epidural anesthesia. We report the case of a 61-year-old man with squamous cell carcinoma of the lung who suffered an epidural hematoma after undergoing right double lobectomy. Before anesthetic induction an epidural catheter was inserted to the D5-D6 space for postoperative analgesia. Surgery was without noteworthy events and the patient was extubated in the operating room; 5,000 IU of low molecular weight heparin was injected subcutaneously every 24 hours and 5 mg of methadone was provided by epidural catheter every 8 hours. After removal of the catheter three days after surgery, lumbar back pain and hypoesthesia, and weakness in both legs appeared. Epidural hematoma was suspected and treatment with 30 mg.kg-1 of methylprednisolone i.v. was started. Nuclear magnetic resonance imaging of the lumbar spine confirmed the presence of a hematoma at D6-D8. Neurologic symptoms improved in the following hours and additional surgery was not required. The patient was released without neurological symptoms 10 days after lung surgery. We discuss the prevalence, etiology and treatment of epidural hematoma related to epidural anesthesia.

Epidural anesthesia in a patient with single ventricle

BACKGROUND AND OBJECTIVES

Single ventricle is a complex congenital cardiopathy characterized by a unit ventricular chamber. Changes of the cardiovascular system in response to epidural anesthesia in these patients have yet to be clearly elucidated.

METHODS

A 26-year-old man with single ventricle underwent an orchiopexy under lumbar epidural anesthesia with intravenous sedation.

RESULTS

Orchiopexy was successfully performed and the patient was discharged from hospital 4 days later.

CONCLUSION

This case report emphasizes the issues of importance to anesthesiologists in regard to this cardiopathy and describes a successful technique of caring for a patient with a complex problem in an emergency situation.

Maintained unilateral analgesia

We describe a case of a patient subjected to what proved to be an epidural puncture with catheter placement resulting in persistent unilateral analgesia. The epidurographic study by contrast medium injection through the catheter showed unilateral distribution of the contrast following the cranio-caudal axis in the anterior epidural space.

[Effects of premedication in children with congenital cardiopathy]

OBJECTIVES

To evaluate the effects of premedication in children with congenital heart disease.

MATERIAL AND METHODS

Twenty-five children scheduled for surgical repair of congenital heart defects were studied. Two groups were formed based on whether cardiopathy was noncyanotic (group CNC) or cyanotic (group CC). Patients were premedicated rectally with 4 mg/kg pentobarbital and, 15 minutes later, with 0.15 mg/kg of morphine chloride by subcutaneous perfusion. SpO2 was monitored, as was the degree of sedation and airway obstruction prior to premedication (T1), 15 minutes after administration of pentobarbital (T2) and 30 minutes after morphine (T3).

RESULTS

In the children with cyanotic cardiopathy, SpO2 increased over T1 (75.5 +/- 8.7%) at times T2 (76.2 +/- 7.7%) and T3 (78.1 +/- 8%), although the change was not statistically significant. In group CNC, average SpO2 did not change, although one case of clinically significant desaturation due to hypoventilation was observed at T3. Adequate sedation was attained in 36% of patients at T2 and in 80% at T3 (p < 0.002). There were no cases of airway obstruction.

CONCLUSION

Premedication with 4 mg/kg pentobarbital rectally does not provide adequate sedation. Addition of 0.15 mg/kg subcutaneous morphine chloride increased the effect considerably, providing stability in SpO2 and even improving it in group CC.

[Transposition of the great arteries. Its anatomical correction. Our initial experience]

Arterial switch correction of transposition of the great arteries has been performed without mortality on 8 children, 7 neonates and one 6 years old child, at our institution. One infant and the 6 year old child has previous pulmonary artery banding and modified Blalock-Taussig shunts. The coronary cuffs were anastomosed to the neo-aorta using the "trap-door" technique described by Mee and in seven the Lecompte manoeuvre was used. Postoperative management included treatment for persistent pulmonary hypertension and left ventricular dysfunction. All patients were discharged in satisfactory condition, and continue to do so at a maximum follow-up of ten months. These results encourage us to elect the arterial switch procedure as the primary treatment of transposition of the great arteries.

Adverse reactions to vancomycin prophylaxis in cardiac surgery

Several adverse effects of vancomycin have been reported. The aim of this study was to assess the incidence of adverse responses to antibiotic prophylaxis with vancomycin in cardiac surgical patients. Prospectively, 116 consecutive patients (106 adults and 10 children) undergoing cardiac surgical procedures in this institution from January to June 1990 were studied. After the anesthetic induction, vancomycin, 1 g in adults and 10 mg/kg in children, was intravenously administered over 30 minutes. The infusion rate was slowed if any adverse effect was observed. As a control group, 10 similar patients were evaluated during the same period of 30 minutes after anesthetic induction but prior to vancomycin administration and surgical stimulation. Thirty-one patients (26.72%) developed an adverse effect, mainly hypotension (29 patients, 25%), which was considered severe in 15 patients (12.93%). Seven patients (6.03%) developed a maculopapular erythema that was associated with hypotension (Red-Man's syndrome) in 5 patients and with bronchospasm in 1 patient. The incidence of adverse reactions in children (20%) was similar to the overall incidence. Only 1 patient in the control group (10%) developed hypotension during the period studied. The incidence of adverse reactions was not related to age, body weight, vancomycin dose administered per kilogram body weight, type of surgical procedure, or associated disease. Mean duration of the infusion was similar in patients with and without adverse responses (34.60 +/- 12.41 minutes and 37.38 +/- 14.55 minutes, respectively). It is concluded that perioperative prophylaxis with vancomycin in cardiac surgery produces a high and unpredictable risk of significant hypotension.

[Use of a new inotropic agent, enoximone, in heart surgery]

We report our experience with the use of enoximone in 5 patients with severely depressed preoperative myocardial function who underwent cardiac surgery. In patients 1 and 2, enoximone was administered as the inotropic of choice before cardiopulmonary bypass (CPB) and a substantial improvement of cardiac index was achieved; in these patients, enoximone administration after CPB permitted to overcome low cardiac output which persisted after high dose dobutamine in patient 1, and in patient 2 right ventricular contractility improved. In patient 3 the use of enoximone permitted the discontinuation of CPB, which had not been previously possible with the association of dopamine and dobutamine. In this patient, adrenaline perfusion improved the low cardiac output syndrome but resulted in poorly tolerated side effects. However, in patient 4 the administration of enoximone during pre-CPB did not improve in a reduction in filling pressures without an increase in the cardiac index because the patient was hypovolemic. In patient 5, the administration of enoximone permitted to interrupt the infusion of dobutamine and to reduce the dose of sodium nitroprusside, which had resulted in significant tachycardia, with increased myocardial contractility and a reduction of vascular peripheral and pulmonary resistances. We conclude that enoximone, single or in association with other inotropics, should be considered a drug of choice in patients undergoing cardiac surgery who develop a low cardiac output syndrome.

Amrinone in perioperative low cardiac output syndrome

Amrinone has been shown to have a beneficial effect on left ventricular function in low output syndrome (LOS), but its use after open-heart surgery has not been extensively revised. We studied 10 patients with LOS post-cardiopulmonary bypass (CPB), who failed to respond to conventional treatment (vasoactive drugs plus intraaortic balloon pump) and were treated with amrinone, 0.75 mg.kg-1 followed by a continuous infusion of 5 to 10 micrograms.kg-1-min-1. One patient failed to respond to the treatment and subsequently died, but in the other nine patients blood pressure and cardiac index increased, left filling pressure decreased and they were successfully weaned from the CBP and survived. These results suggest that amrinone, either alone or combined with other inotropic drugs and mechanical support, is a valuable drug in the management of LOS after CPB.

[Tremor and epidural anesthesia]

To evaluate the influence of temperature of the injected anesthetic solution on the development of tremor during epidural anesthesia, 66 patients divided in three homogeneous groups were evaluated: group I (n = 22; bupivacaine 4 degrees C), group II (n = 24; bupivacaine 20 degrees C), and group III (n = 24; bupivacaine 37 degrees C). The incidence of tremor was 20% (4 patients) in group I, 9% (2 patients) in group II and 12.5% (3 patients) in group III. No significant differences were found between the groups. The overall incidence was 13.6%. The epidural injection of 5 ml of saline at 37 degrees C achieved the attenuation and/or disappearance of tremor in three (3/4) group I patients (4 degrees C) and in one (1/3) group III patient (37 degrees C), whereas it was ineffective in one patient from group I and one from group III. In the two patients from group II (20 degrees C) and in one from group III (37 degrees C), tremor was self-limited. We conclude that the incidence of tremor during epidural anesthesia is not correlated with the temperature of anesthetic solutions, and that the epidural injection of saline at 37 degrees C may give some therapeutic benefit.