At our institution, thoracic epidural anesthesia (TEA) is being used with increasing frequency to provide surgical anesthesia for breast surgery. Until recently, we have not seen any morbidity associated with this technique. Although profound bradycardia and asystole after spinal anesthesia have been described, [2–10]to our knowledge, there has been no clinical report describing such cardiovascular event after TEA. Therefore, we report a case of postoperative bradycardia and asystole complicating TEA performed on a healthy woman for surgical removal of bilateral breast tissue expanders.
A 35-yr-old, 60-kg, 165 cm woman, American Society of Anesthesiologists (ASA) physical status 1, underwent removal of bilateral breast tissue expanders and replacement with saline implants. Before this admission, she had undergone bilateral modified radical mastectomy and insertion of tissue expanders during TEA. Her perioperative course was uneventful. She requested TEA again for this procedure. On the morning of admission to the day surgery unit, her physical examination was normal, but she was anxious. After insertion of an intravenous catheter, she received 100 micro gram fentanyl and 2 mg midazolam intravenously. With the patient in the sitting position, an epidural catheter was inserted at the T5–6 level with loss of resistance technique. Licodaine 2% with 1:200000 epinephrine was given in incremental doses of 2 ml to a total of 12 ml. A surgical level of sensory anesthesia was obtained front C8 to T10. During the surgery, she was comfortable and hemodynamically stable. Intraoperatively, she was sedated with 150 micro gram fentanyl and 2 mg midazolam intravenously. To maintain surgical anesthesia, a total of 7 ml of 2% lidocaine with 1:200000 epinephrine was given in incremental doses throughout the 1-h operation. The epidural catheter was removed at the end of surgery without any complications. On arrival at the postanesthesia care unit, she was awake, relaxed, and comfortable. Vital signs were: blood pressure 102/55 mmHg, heart rate 65 beats/min in normal sinus rhythm, respiratory rate 12 breaths/min, and hemoglobin oxygen saturation of 96% on room air. The sensory level extended from T2 to T10. The patient was placed in a semi-sitting position. Ten minutes later, she complained of nausea; vital signs were still stable. Ondansetron (4 mg in 100 ml normal saline) was given intravenously over 15 min. Approximately 15 min later, she again complained of nausea. The postanesthesia care unit nurse noticed that heart rate had suddenly decreased from the 60s to the 40s. The cardiac rhythm was sinus bradycardia in the 40s for approximately 1 min, during which time blood pressure was normal. Then, she abruptly became unresponsive and pulseless, with 45 s of asystole on the electrocardiograph. Cardiopulmonary resuscitation was promptly started, and 1 mg atropine was administered intravenously. She responded immediately to treatment, with return of sinus rhythm of 75 beats/min, a stable blood pressure of 110/60 mmHg, and a respiratory rate of 12 breaths/min. Repeat physical examination was unremarkable; the patient was calm and hemodynamically stable. Neurologic examination revealed residual sensory blockade from T2 to T10 level 1 h after the last dose of local anesthetic had been administered via the epidural catheter. She was admitted overnight for close observation. No dysrhythmia or myocardial ischemia had been documented before or after the cardiac arrest. However, postoperative 24-h Holter monitoring revealed sinus rhythm with a heart rate from 40 to 120 beats/min. She was discharged less than 24 h after surgery, without any further complications.
Severe cardiovascular events after neuraxial blockade are well documented in the literature. In 1988, the ASA Closed Claim Study reported 14 cases of sudden cardiac arrest during spinal anesthesia. All cases involved healthy patients undergoing minor surgical procedures. The highest level of block averaged T4. Despite apparent appropriate resuscitation, 6 of the 14 patients died, and only 1 survivor had complete neurologic recovery. Caplan et al. postulated that respiratory changes due to oversedation may have been a triggering factor in approximately one half of the cases. However, several cases of severe bradycardia and cardiac arrest during spinal anesthesia without hypoxemia or oversedation have been reported. Several cases of sudden cardiac arrest during lumbar epidural anesthesia also have been observed. [11,12]Carpenter et al. prospectively studied the incidence of various complications during spinal anesthesia. They concluded that baseline heart rates less than 60 beats/min, beta-adrenergic-blockade, and ASA physical status 1 were the strongest predictors of bradycardia. But, unlike hypotension, peak block height had the weakest correlation to bradycardia.
Changes in heart rate are influenced by the balance between sympathetic and parasympathetic tone. Normally, the baroreceptor-mediated response to a decrease in systolic blood pressure is an increase in heart rate. During neuraxial blockade, a moderate decrease in mean arterial pressure is not usually associated with a significant increase in heart rate. It has been postulated that preganglionic sympathetic blockade induced by epidural anesthesia is responsible for this relative bradycardia, a phenomenon explained by the blockade of cardiac accelerator sympathetic fibers (T1-T4). In a study on the effects of TEA (level C8-L1) in patients during general anesthesia, Goertz et al. reported markedly reduced cardioacceleration in response to a decrease in blood pressure while preserving the cardiac slowing in response to increased blood pressure. They suggested that an increased heart rate response to decreased blood pressure was sympathetic in nature and that cardiac slowing secondary to increased blood pressure was mainly of parasympathetic mechanism. However, in these studies, the authors could not exclude the effects of general anesthesia as well as systemically absorbed local anesthetics on baroreceptor reflexes. In awake volunteers during TEA, Takeshima and Dohi reported no change in baroreflex sensitivity evaluated with the depressor test and a slowing in heart rate response to the pressor test. They hypothesized that sympathetic control of heart rate could function as an inhibitor of the vagus rather than as an active cardiac accelerator. Nevertheless, cardiac sympathectomy during TEA leaves parasympathetic tone unopposed. In our patient, severe nausea and administration of vagotonic drugs such as fentanyl may have further increased vagal tone, leading to acute bradycardia and asystole. In addition, the sinus dysrhythmia demonstrated by 24-h Holter monitoring after the surgery may indicate preexisting autonomic or conductive imbalance.
Venous pooling in the pelvis and lower extremities from sympathectomy during lumbar epidural anesthesia is a well established phenomenon. Recently, Hopf et al. showed that, with TEA, sympathetic blockade does exceed sensory block in the caudal direction. Assuming temperature changes in the foot reflect diminished sympathetic outflow, these investigators found that, despite sensory levels confined to upper or mid thoracic regions (T5-T9), there was an increase in pedal skin temperature (albeit submaximal compared with lumbar epidural), which indicated partial sympathetic blockade. Therefore, a decrease in venous return may have occurred in our patient, who was in the semi-sitting position, enhancing vagal tone and contributing to the onset of severe bradycardia. Ecoffey et al. demonstrated that head-up tilt in the presence of peripheral sympathectomy induces a paradoxical bradycardia because of enhanced vagal tone, as demonstrated by an increase in plasma vasopressin concentration and no change in plasma catecholamines concentration. Jacobsen et al. reported that a decrease in left ventricular diameters is associated with an increase in plasma pancreatic polypeptide (a maker of parasympathetic function) during lumbar epidural anesthesia. They hypothesized that during central volume depletion associated with epidural anesthesia, an increase in vagal activity, rather than blockade of cardiac sympathetic activity, is responsible for the bradycardia. This suggests that a decrease in venous return can be an important triggering factor in cardiac arrest during epidural anesthesia. We advise that the semi-sitting position be used with caution in patients with TEA.
Some patients may be predisposed to sudden onset of severe bradycardia or asystole. One of the explanations could be a Bezold-Jarisch-like reflex. The Bezold-Jarisch reflex originates in inhibitory mechanoreceptors and chemoreceptors concentrated in the inferoposterior wall of the left ventricle. Stimulation of these receptors by stretch or chemical substances increases parasympathetic activity and inhibits sympathetic activity, producing bradycardia, vasodilation, and hypotension. A decrease in ventricular volume would normally decrease receptor activity. However, a rapid decrease in ventricular volume may trigger a paradoxic increase in receptor activity, supposedly because of vigorous ventricular contraction around an almost empty chamber. Such a paradoxic Bezold-Jarisch reflex has been described in various clinical settings, including vasovagal syncope. Although our patient had not reported a vasovagal episode in the past, her high degree of anxiety in the presence of sympathetic blockade may have offset the ability of the cardiovascular system to respond to the decrease in venous return.
Finally, the patient received ondansetron intravenously before the cardiac event. Bradycardia has been reported after administration of ondansetron but without significant difference compared with a placebo. To date, asystole has not been related to this compound. Interestingly, in animal studies, all 5-hydroxytryptamine receptor antagonist (5HT3) inhibitors inhibit the 5-HT3 agonist-mediated Bezold-Jarisch reflex. Bradycardia was not seen with any of the antagonists alone. A direct cardiac effect of ondansetron cannot be excluded but seems very unlikely in this context. Most importantly, the etiology of her nausea, for which ondansetron was given, was not secondary to cerebral hypoperfusion because her vital signs, mental status, and oxygenation were normal. If she had been hypotensive, the treatment of choice would have been intravenous fluid administration and vasopressor therapy with ephedrine or phenylephrine.
In conclusion, we report severe bradycardia and asystole after segmental TEA in a healthy patient after breast surgery. In the presence of inhibition of cardiac sympathetic fibers, enhanced vagal tone, and decrease in central venous return, sudden onset of cardiac arrest can occur in some patients. Because of our experience, we recommend that bradycardia be treated early with an anticholinergic agent or sympathomimetic and that venous return be maintained at all times.