Maintained consciousness during witnessed asystole after spinal anesthesia for Cesarean section

Introduction

Cardiac arrest after spinal anesthesia is rare, recently reported to occur in 1.5/10,000 anesthetics¹. Despite its low frequency, it was believed to be associated with a high mortality rate, reported as >40% mortality in the closed claims analysis by Caplan et al.² We present a unique case of cardiac arrest after spinal anesthesia for Cesarean section. We discuss relevant physiological mechanisms to explain why this patient never lost consciousness during the asystolic episode, which could mislead clinicians into delaying treatment.

Case report

A healthy Spanish speaking 24 year old G₂P₀ female (70 kg, 65", BMI 26.5) at term was scheduled for an elective Cesarean section for breech presentation. The patient’s preoperative heart rate (HR) was 62 beats per minute (bpm), blood pressure was 126/70 mmHg. After administration of 1000 ml Ringer’s lactate solution, spinal anesthesia was performed with 1.6 cc of 0.75% bupivacaine (12 mg) in a sitting position, resulting in a bilateral T4 sensory tested by pinprick. No sedatives or prophylactic vasopressors were administered. The patient was laid supine immediately after spinal anesthesia with a left lateral tilt to prevent aorto-caval compression. SpO₂ remained at 98%. Three min later her HR decreased to 30 bpm, followed by 15 sec of asystole, which was immediately noticed via pulse oximeter and EKG. 1 mg atropine was administered intravenously (IV). The anesthesiologist’s loud voice shouting for resuscitative help appeared to startle the patient who turned towards him. Now, 3 beats at 40 bpm were noted, followed again by asystole for 30 sec. 70 mcg of epinephrine IV was given, the anesthesiologist repeated the call for help, and removed the blanket and gown from her chest to begin chest compressions. The patient again appeared startled by this action based on facial expressions while turning towards the anesthesiologist, and a single spontaneous heart beat reappeared on the monitor. Another beat followed within 2 sec, rapidly accelerating to 160 bpm (sinus tachycardia). The patient was rapidly prepared, and the fetus was delivered within 5 min with Apgar scores of 8 and 9. Afterwards, the events were discussed with her with the help of an interpreter. She had been unaware of the cardiac arrest, and denied having lost consciousness.

Discussion

Spinal anesthesia is associated with an increased risk of hypotension and bradycardia³. ASA 1 status, baseline HR <60 bpm, a prolonged PR-interval, beta-blockade, age <50 years, a sensory blockade above T6 and male gender have been described as independent risk factors^2–5. The differential diagnosis for asystole in this setting includes hypoxia due to respiratory depression, high or total spinal, local anesthetic toxicity when an epidural dose is administered IV, myocardial ischemia, vasovagal reaction, and bradycardic reflexes after neuraxial anesthesia. Our patient was unsedated, and oxygenation maintained prior to the arrest. In fact, in contrast to earlier studies that suspected oversedation², it is difficult to invoke hypoxemia as the primary cause of cardiac arrests during this case or historical reports because most asystolic arrests occurred in the setting of 95–100% oxygen saturations⁶. Spinal anesthesia itself is associated with bradycardia in 9–13% of patients³. Most of these effects are related to sympathetic blockade during spinal anesthesia. Sympathetic blockade is often two to six levels higher than the sensory level, so that a patient with a T4 sensory block may have completely blocked cardioaccelerator fibers originating from T1 to T4⁷. However, in a prospective study of 952 patients, Carpenter et al.³ found that the traditional risk factor peak block height had the weakest correlation with the severity of bradycardia.

A more important effect of sympathetic inhibition is a significant decrease in venous return to the heart that enhances cardiac vagal tone⁸. The decrease in venous return could further be accentuated by aorto-caval compression despite the application of a left lateral tilt position, since no particular position has been shown to completely abolish caval compression⁹. Cardiac vagal tone may be enhanced by three different mechanisms⁴: 1) Decreased venous return results in decreased stretch of atrial pacemaker cells resulting in bradycardia; 2) Firing of low pressure baroreceptors in the right atrium and vena cava and 3) A Bezold-Jarisch reflex, in which left ventricular mechanoreceptors are stimulated and paradoxically cause bradycardia. Earlier studies suggested that vasodilation caused by sympathetic blockade may be best treated by preemptive fluid administration^7,8. However, more recent data suggest that crystalloids alone (without vasopressors) are inconsistent in preventing hypotension¹⁰.

A high or total spinal anesthesia was unlikely, since our patient responded appropriately to pinprick sensations at or above the T4 level. Nevertheless, a T4 level block could completely block cardioaccelerator fibers originating from T1 to T4 resulting in severe bradycardia. Caplan et al. in their closed claims analysis noted the median time from local anesthetic administration to cardiac arrest to be 36±18 minutes²; however, most of these cases were non-obstetric. The onset of asystole in our case was 3 min from initiation of spinal anesthesia which could have resulted from a sympathetic block accentuated by at least partial aorto-caval compression by the gravid uterus. Supine hypotensive syndrome of pregnancy has been well described in the literature and can cause a marked bradycardia with a reduction in cardiac output and severe hypotension in a few subjects¹¹.

While women are 11 to 50 times less likely to develop bradycardia during neuraxial anesthesia than men¹², parturients typically have a higher HR and are therefore even less likely to develop bradycardia or cardiac arrest during neuraxial anesthesia⁵. Carpenter et al. reported a low baseline heart rate (<60 bpm), ASA 1 status (when compared to ASA 3 and 4) and beta-blockade as the strongest predictors of bradycardia³. With her baseline HR of 62 bpm our patient may have been at higher risk.

Case reports exist in which patients were amnestic for episodes of unconsciousness, as documented with carotid sinus syndrome¹³. Despite the language barrier, our patient exhibited signs of consciousness during asystole via her startle reactions and turning her head to unexpected shouting and light tactile stimulation. This demonstrates that a patient could remain conscious for up to 30 sec despite being asystolic in isolated cases. This delay in loss of consciousness could substantially delay recognition of a life threatening event, especially if a patient is temporarily disconnected from monitoring. It therefore underscores the importance of continuous monitoring during spinal anesthesia with pulse oximetry and EKG at all times.

There are several theoretical mechanisms that may be postulated to explain the maintenance of cerebral oxygenation in this scenario. First, supine position and volume loading may have contributed to venous pooling within the cerebral vasculature. Studies have also demonstrated that stroke volume and CVP increase during vagally induced bradycardia¹⁴, so that a few heart beats during the asystolic episode may have produced a disproportionally large amount of blood flow in this setting. Whether aortic compression by a gravid uterus enhances cerebral blood flow at the expense of lower extremity blood flow remains unknown. Through their experiments on healthy controls Ide et al. showed that an impaired ability to increase cardiac output during exercise with a large muscle mass appears to limit blood flow distribution not only to active muscle¹⁵ but also to such a vital organ as the brain, and that this flow restriction was by way of the sympathetic nervous system¹⁶. During central blood volume depletion, the increase in sympathetic nerve activity shifts the cerebral autoregulation curve to the right¹⁷ and the vasoconstrictor sympathetic nerve activity over-rides vasodilatation¹⁸. Ide et al. also showed that sympathetic blockade at the level of the neck eliminated the limitation to the increase in cerebral blood flow following cardio-selective β-1 adrenergic blockade¹⁶. Although a theoretical hypothesis, a T4 level blockade in our patient may have blunted the effect of reduced cardiac output on cerebral vasoconstriction, and may have helped maintain cerebral perfusion prior to the asystolic events.

An auditory startle stimulus produces a transient tachycardia 2 to 5 sec after the stimulus, consistent with our findings; the reflex is mediated by the vagus nerve and sympathetic ganglia¹⁹. The initial startle response was followed by 3 spontaneous heart beats and probably occurred before the atropine had time to circulate through the bloodstream. The second startle reaction perhaps permitted enough cardiac activity to circulate the resuscitative drugs without CPR. A deep inspiration accompanying the startle response may have further reduced vagal tone by a mechanism similar to that seen with respiratory sinus arrhythmia (RSA)²⁰. RSA is a cardiorespiratory phenomenon characterized by HR or R-R interval (RRI) fluctuations in phase with inhalation and exhalation. Typically, HR accelerates during inspiration and slows during expiration, and is prominent during high resting vagal tone; increased sympathetic stimulation attenuates it. Rate and depth of respiration also influence the magnitude of RSA with slow deep breathing producing maximal and rapid shallow breathing greatly attenuating RSA. Unfortunately, no EKG strips are available to confirm the above assumptions as the anesthesiologist was busy resuscitating the patient.

Conclusion

We present a unique case of sudden cardiac arrest after spinal anesthesia in an apparently healthy patient. Vagal predominance and sympathetic blockade along with activation of baroreceptor reflexes may contribute to bradycardia during spinal anesthesia. Cerebral perfusion may be maintained by venous pooling and a partial sympathetic blockade blunting the effect of reduced cardiac output on cerebral vasoconstriction. Transient tachycardia could be caused by a startle response and the effect of slow, deep respirations on RSA magnitude, allowing sufficient spontaneous cardiac activity to circulate the resuscitative drugs. Consciousness may be maintained even after asystole has occurred, rendering insufficiently monitored patients at high risk.