Vasopressin for the Treatment of Refractory Hypotension After Cardiopulmonary Bypass

Exogenous arginine vasopressin (VP) can be more effective than optimal doses of epinephrine in improving organ blood flow during cardiopulmonary resuscitation, presumably as a result of higher arterial pressure [1-4]. VP has also dramatically increased blood pressure in patients with septic shock and in left ventricular assist device (LVAD) recipients with vasodilatory shock, in whom endogenous VP levels were found to be significantly lower than normal [5,6]. We describe a patient after cardiopulmonary bypass (CPB) with intractable hypotension despite maximal inotropic and mechanical support, who responded dramatically to the administration of VP. Case Report A 59-yr-old female patient (weight 100 kg, body surface area 2.1 m2) presented with severe aortic stenosis for aortic valve replacement. Her medical history was significant for hypertension and insulin-dependent diabetes. Her preoperative medications included prinivil, insulin, and estrogen. Echocardiogram and catheterization data revealed nonobstructed coronary arteries, concentric left ventricular hypertrophy, and mild to moderate global hypokinesis. All cardiac chambers were mildly enlarged. Mitral annular calcification with trace mitral regurgitation, a calcified aortic valve with a valve area of 0.5 cm2, and trace tricuspid regurgitation were also noted. Initial blood pressure (BP) was 115/72 mm Hg, heart rate (HR) was 80 bpm, and pulmonary artery pressure (PAP) was 70/30 mm Hg. After sedation with midazolam 2 mg and fentanyl 50 [micro sign]g, anesthesia was induced with etomidate 10 mg and succinylcholine 125 mg. Maintenance drugs included isoflurane 0.6 vol% prebypass, fentanyl 30 [micro sign]g/kg, midazolam 15 mg, and pancuronium 15 mg. Hemodynamic data after induction were: BP 105/65 mm Hg, HR 70 bpm, PAP 60/37 mm Hg, central venous pressure (CVP) 17 mm Hg, cardiac output (CO) 4.5 L/min, cardiac index (CI) 2.1 L [center dot] min-1 [center dot] m-2, and systemic vascular resistance (SVR) 1500 dynes [center dot] s-1 [center dot] cm-5. CPB was initiated, and after aortic valve replacement with a St. Jude valve, the mitral valve was noted to be torn. This was repaired, then replaced with a St. Jude valve. Total duration of CPB was 6 h 22 min. Perfusion was nonpulsatile with a prime of plasmalyte A with bicarbonate. On CPB, rectal temperature was 22[degree sign]C and hemoglobin (Hb) ranged from 6.0% to 8.6%. Arterial pressures ranged from 50 to 75 mm Hg with peak flows of 5.25 L/min. Crystalloid cardioplegia was used. Before the termination of bypass the heart was observed to be full and in sinus rhythm, but with significantly depressed contractility noted on the echocardiogram. After the administration of milrinone 20 [micro sign]g/kg and digoxin 0.25 mg, infusions of norepinephrine (NE) and epinephrine (E) were begun. After an unsuccessful attempt at weaning from CPB, an intraaortic balloon pump (IABP) was inserted, the infusion of NE was increased to 0.15 [micro sign]g [center dot] kg-1 [center dot] min-1, and the infusion of E was increased to 0.2 [micro sign]g [center dot] kg-1 [center dot] min-1. The patient was then weaned from CPB. Hemodynamic data with IABP at 1:1 augmentation were: BP 95/60 mm Hg, HR 112 bpm (sinus), PAP 29/19 mm Hg, CVP 12 mm Hg, CO 5.0 L/min, CI, 2.4 L [center dot] min-1 [center dot] m-2, and SVR 800 dynes [center dot] s-1 [center dot] cm-5. Rectal temperature was 36.2[degree sign]C. Shortly after protamine administration, the systolic BP decreased to 70 mm Hg with HR 120 bpm, PAP 51/29 mm Hg, and CVP 15 mm Hg. Diphenhydramine and hydrocortisone were given to treat the apparent reaction, and the BP slowly improved over time with the IABP at 1:1 augmentation, incremental increases in the NE and E infusions, and the addition of phenylephrine (P) at 0.5 [micro sign]g [center dot] kg-1 [center dot] min-1. Lung compliance had worsened, however, and pulmonary edema was noted. A tidal volume of 350 mL and a respiratory rate of 16 breaths/min produced a peak pressure of 38 cm H2 O. An arterial blood gas (ABG) analysis at that time revealed pH 7.18, PaCO2 52 mm Hg, PaO2 60 mm Hg (fraction of inspired oxygen = 1.0, 5 cm H2 O positive end-expiratory pressure), HCO3 19.6 mEq/L, Hb 10.2%, K+ 4.1 mg/L, glucose 426 mg%. Despite gradual increases in the E infusion to 0.5 [micro sign]g [center dot] kg-1 [center dot] min-1 and of the NE infusion to 0.3 [micro sign]g [center dot] kg-1 [center dot] min-1, as well as incremental doses of bicarbonate, albuterol, insulin and increases in minute ventilation, we were unable to close the chest without a further abrupt decrease in BP. We decided to take the patient to the intensive care unit (ICU) with the chest open. At that time, BP was 95/60 mm Hg, HR was 105 bpm (sinus), PAP was 49/27 mm Hg, CVP was 16 mm Hg, CO was 3.5 L/min, CI was 1.7 L [center dot] min-1 [center dot] m-2, SVR was 870 dynes [center dot] s-1 [center dot] cm-5, rectal temperature was 36.6[degree sign]C, pH was 7.30, PaCO2 was 51 mm Hg, PaO2 was 58 mm Hg (fraction of inspired oxygen 1.0, 5 cm H2 O positive endexpiratory pressure) HCO3 was 24.9 mEq/L, Hb was 10.1 g/dL, K+ was 3.8 mmol/L, and glucose was 404 mg/dL. On ICU arrival, the patient experienced a further decrease in BP to 75/30 mm Hg, with HR 112 bpm, PAP 49/26 mm Hg, and CVP 17 mm Hg. This induced us to begin an infusion of VP (8 arginine vasopressin injection 20 U/mL). After the rapid infusion of 10 U of VP, the BP increased to 150/53 mm Hg, with HR 110 bpm PAP 51/29 mm Hg, and CVP 18 mm Hg. CO was measured at 4.8 L/min, CI was 2.3 L [center dot] min-1 [center dot] m-2, and SVR was 1600 dynes [center dot] s-1 [center dot] cm-5. The VP infusion was titrated to 14 U/h to maintain the BP >100/45 mm Hg, and the P infusion was discontinued. The NE infusion was rapidly reduced to 0.03 [micro sign]g [center dot] kg-1 [center dot] min-1, and the E infusion was decreased to 0.1 [micro sign]g [center dot] kg-1 [center dot] min-1. The VP infusion was continued at 14-20 U/h for 6 h, then reduced to 7 U/h until Postoperative Day (POD) 1. It was discontinued on POD 2 as the patient continued to improve hemodynamically. Urine output, which had been negligible after CPB, began at 20 mL/h on arrival in the ICU and was 100-200 mL/h by POD 1. On POD 3, the chest was closed and the IABP was removed. On POD 7, the patient was extubated, and she was discharged from the hospital on POD 32. She has continued to do well without negative sequelae. Discussion VP has long been recognized as a potent hormonal regulator of blood volume and BP through antidiuresis and its release either in response to osmotic changes or as a baroreflex-mediated response [7]. The reason for the dramatic response of some patients with hypotension to VP remains conjectural, however. Unlike animals, which show an attenuation of hypotension with VP infusions, there seems to be negligible pressor effects in healthy, hypotensive humans, despite significant vasoconstriction, particularly of the splanchnic vessels [8-10]. In certain pathophysiological states, however, the pressor affects of VP are greatly enhanced. Hypotension has been reversed, minimized, or prevented in patients with idiopathic orthostatic hypotension, patients on dialysis, diabetics, and those with septic shock [5,11-13]. Of particular interest is an observation that VP reversed hypotension in LVAD recipients with vasodilatory shock after CPB, most of whom had inappropriately low plasma VP levels [6]. These low levels of VP are in contrast to the high plasma VP levels described during and after CPB in many previous studies [14-19]. Enhanced pressor effects in these various types of hypotension may implicate autonomic insufficiency in these patients, which could result in impaired baroreceptor-mediated secretion of VP [5,10,13]. The presence of diabetes in our patient may therefore partially explain our result. Another possible explanation is that depletion of VP secretory stores in the neurohypophysis may occur during prolonged hypotension, and exogenous VP thus improves the response to other vasopressors [5]. Either mechanism is possible in cardiac patients, in whom CPB can stimulate a systemic inflammatory response that may include endotoxemia and a secondary vasodilation similar to that seen in septic shock [5,6,20,21]. Of additional interest is the observation that the CO in our patient increased from 3.5 to 4.8 L/min with VP, despite a SVR increase from 870 to 1600 dynes [center dot] s-1 [center dot] cm-5, with markers of preload remaining nearly constant. This is in contrast to the effect seen in septic shock, in which CO decreases with VP administration [5]. An increase in CO may have occurred in our patient secondary to improved coronary flow as a result of increased perfusion pressure, but increased inotropy is an additional possibility, as an increase in developed and resting tension has been demonstrated in guinea pig papillary muscle after the administration of VP [22]. There was no hemodynamic indication in our patient that VP caused significant coronary vasoconstriction, aggravation of collateral dependent myocardial ischemia, or other organ ischemia. We conclude that, in some cases of significant hypotension after CPB, despite maximal inotropic support, exogenous VP may provide a dramatic improvement in hemodynamic conditions.