Unsuccessful Resuscitation Under Hypotensive Epidural Anesthesia During Elective Hip Arthroplasty

Hypotensive epidural anesthesia is a combination of extensive epidural anesthesia (upper sensory thoracic level at T4 or above), subsequent total blockade of the sympathetic nervous system, and consequent arterial hypotension (objective mean arterial blood pressure [MAP] approximately 55 mm Hg) [1,2]. Immediately after injecting the local anesthetic, a small dose (2-5 [micro sign]g/min) of epinephrine is given to maintain stroke volume, heart rate (HR), and cardiac output. The rate of epinephrine and fluid administration is adjusted to achieve the desired MAP [1]. Since the introduction of this method, a significant reduction in both blood loss [3] and (on retrospective analysis) mortality has been achieved in patients undergoing hip and knee arthroplasties [4]. Asystole and severe bradycardia during epidural anesthesia have been described [5]. This is a report of unsuccessful intraoperative resuscitation during hypotensive epidural anesthesia. Case Report A 79-yr-old man weighing 60 kg was admitted to our hospital for severe pain in the left hip (loosening of the socket). A socket replacement was planned. A preoperative visit the evening before the intervention established that the patient was in good general physical condition without significant previous illness. His cardiac history was unexceptional. A treadmill exercise test could not be performed because of the patient's limited hip mobility. The hemoglobin level was 12.8 g/dL; the other laboratory values were normal. His blood pressure (BP) was 160/90 mm Hg, and his HR was 68 bpm. The electrocardiogram (ECG) showed no pathologic findings. Premedication consisted of 20 mg of dipotassium clorazepate the previous evening and 10 mg on the day of the operation. On arrival in the operating room, arterial BP was 150/75 mm Hg and HR was 64 bpm. The ECG (lead II) showed a sinus rhythm. Two venous indwelling cannulas and one arterial radial cannula were inserted. After a back flow of blood from the peridural catheter at L1-2, puncture and introduction of the catheter at T12-L1 presented no problems. We obtained a sensory level at T3 after the administration of 15 mL of 2% lidocaine CO2 with epinephrine 1:200,000. Subsequently, 5 mL of 2% lidocaine was administered at an interval of 30-35 min to maintain a sensory level at T4 (tested by temperature discrimination). Immediately after injection of the local anesthetic, epinephrine 1-4 [micro sign]g/min was given continuously. The average MAP was 50-55 mm Hg. The patient received 4 L/min oxygen through a face mask. In the induction room, the patient's hemoglobin level was 11.6 g/dL (determined approximately 20 min after the first injection of the local anesthetic). Intermittent hypotension (MAP below 50 mm Hg) and bradycardia (HR below 60 bpm) were treated successfully by the administration of an epinephrine bolus (5 [micro sign]g), for a total of 10 boluses, over a period of 180 min (four administrations within 45 min before the intervention) (Figure 1) and an infusion of isoproterenol (1-3 [micro sign]g/min). Until the time of cardiac arrest (duration of operation 180 min), the estimated blood loss was 800 mL. Up to this time, the patient had received 1800 mL of lactated Ringer's solution, 600 mL of hydroxyethyl starch 6%, and a transfusion of 140 mL of his own blood (Table 1). The patient was sedated with thiopental (150 mg/h) but could be aroused easily. A few minutes before cardiovascular arrest occurred, it was possible to talk to the patient normally. Until this point, the course of the operation had been uneventful. The cup had already been cemented (no cardiovascular reaction) and the wound rinsed. Within a few seconds, the BP decreased to an unmeasurable level (Figure 1). At first, the ECG showed no irregularities. The patient could no longer be aroused, and the carotid pulse could not be felt. The ECG subsequently showed bradycardia with broadened ventricular complexes, followed immediately by an asystole. After a precordial blow with the fist, external heart massage was started. A second anesthetist immediately administered 1 mg of epinephrine IV (maximal latency 2 min after cardiovascular arrest), and the trachea was intubated. Despite continuous cardiac massage, repeated administration of 1 mg of epinephrine every 3 min, and administration of volume (400 mL of hydroxyethyl starch, 600 mL of blood), radial pressure could only be obtained after 10 min. However, the maximal MAP was 25 mm Hg and could not be significantly improved. The highest end-expiratory partial pressure of carbon dioxide was 10 mm Hg. Resuscitation was discontinued after 30 min. Blood gas analysis during resuscitation showed the following values (no blood retransfused at this juncture): pHa 7.41, PaCO2 23.2 mm Hg, base excess -7.2 mmol/L, PaO2 53.4 mm Hg, SpO2 86.9%, and hemoglobin 7.3 g/dL. The autopsy did not disclose any pathologic findings.Figure 1: Time curve of saturation, heart rate, and arterial blood pressure. SpO (2) = oxygen saturation of hemoglobin (measured by pulsoximeter), HF = heart rate, DRUCK 1 = arterial blood pressure (invasive measurement). *Bolus of 5 [micro sign]g of epinephrine.Table 1: Intraoperative Fluid ManagementDiscussion This patient's circulating blood volume was 4.2 L. The hemoglobin had decreased from 12.8 g/dL to 7.3 g/dL, or 57% of the initial value. The hemoglobin decrease from 12.8 g/dL to 11.6 g/dL, even before the start of the operation, was caused by hemodilution. Hahn [6] has shown that this arterial hypotension (systolic BP decrease of 25% or more) induced by the epidural anesthesia leads to a fluid transfer from the tissue spaces (Starling's hypothesis), consequently causing an increase of circulating blood volume (in our case, from 4.2 to 4.6 L). Assuming an isovolemic blood replacement, the estimated blood loss (with a hemoglobin of 7.3 g/dL) would be 2.1 L [7] (see Appendix). The infused fluids had an intravascular effect of 1.2 L. According to a formula described by Mortelmans et al. [7], however, this would have led to a hemoglobin value of 8.9 g/dL (under isovolemic conditions). Our patient's blood loss therefore was even greater, and greatly underestimated. This explains why no effective BP was obtained during resuscitation and why the end-expiratory carbon dioxide was low. Caplan et al. [8] reported predisposing factors that could be responsible for a negative outcome in the event of cardiac arrest after spinal anesthesia. They postulated that vasodilation is caused by hypoxia and hypercarbia, thus reducing venous return. Although our patient was under light sedation, the oxygen saturation was 98% until the cardiac arrest. Arterial blood gas analysis showed a PaCO2 of 34 mm Hg; thus, these components probably did not play a decisive role. The early administration of a powerful alpha-adrenergic agonist to increase venous return to the heart and improve cardiac output, as recommended by Caplan et al. [8], was performed in our case. In hypotensive epidural anesthesia, however, there is also a total blockade of the sympathetic nervous system, similar to that in spinal anesthesia. This makes resuscitation more difficult, particularly when cardiovascular arrest is caused by hypovolemia. In his editorial on Caplan's work, Keats [9] concluded, "The CPR [cardiopulmonary resuscitation] is ineffective in the absence of blood in the heart." The studies of Bonica et al. [10,11] on the circulatory effects of peridural block (sensory level at T5 or above) have shown that, under normovolemic conditions, cardiac output remains unchanged. On the other hand, in the presence of a hypovolemia, it leads to a marked decrease of the HR, cardiac output, and MAP. With epinephrine, on the other hand, cardiac output is maintained (because of an increased HR), but the MAP decreases by 23%. Stanton-Hicks et al. [12] have shown that epinephrine administered with a local anesthetic into the epidural space contributes to circulatory stability by preserving HR and stroke volume, even when the arterial BP decreases. The sudden bradycardia followed by an asystole may be the result of parasympathetic activity (vagal response) similar to the "faint response" caused by a reduced venous return [13,14]. It may also be explained pathophysiologically as a stimulation of the inhibitory cardiac receptors, leading to a reflex bradycardia and vasodilation-Bezold-Jarisch reflex [15]. The causative factor, as shown in animal experiments [16], is sudden, severe hypovolemia that leads to a massive paradoxic stimulation of these inhibitory receptors because of vigorous contraction of an almost empty ventricle. We conclude that, during a hypotensive epidural anesthesia, the intravascular volume situation is of significant importance-even a small degree of hypovolemia (recognizable, partly, by the decreased central venous pressure and the decreasing diuresis) must be avoided or quickly treated. The increasing need for a vasopressor (in our case, intermittent administration of a bolus of epinephrine) to maintain MAP must be considered a sign of hypovolemia and be treated with intravascular fluids until the desired effect is achieved. This is a strong indication for extended monitoring. Appendix (Equation 1) where y = end component concentration, yo = initial component concentration, e = base of natural logarithm, x = amount of blood exchanged, BV = blood volume of the patient. This formula expresses the exponential relationship between the percentage of the blood volume (BV) exchanged without transfusing a certain component and the percentage of the initial value of the component that remains after the exchange.