Rebound Pulmonary Hypertension and Cardiogenic Shock after Withdrawal of Inhaled Prostacyclin

A 74-yr-old man (height, 140 cm; weight, 52 kg) presented with congestive cardiac failure. The patient had undergone previous coronary artery bypass grafting in 1979 and 1999. His transthoracic echocardiogram revealed an ejection fraction of 35%, severe aortic stenosis (aortic valve area of 0.9 cm²), severe mitral regurgitation, significant PHTN (estimated pulmonary artery systolic pressure 69 mmHg), moderate tricuspid regurgitation, and normal right ventricular function. Coronary catheterization revealed occluded saphenous venous grafts but a patent left internal mammary graft to the left anterior descending artery. There was no significant coronary stenosis amenable to surgical intervention.

The patient presented to the operating room for mitral and aortic valve replacement. Hemodynamic monitoring included arterial, central venous, and pulmonary artery catheterization. Aprotinin was administered without incident. The induction of anesthesia was with titrated intravenous midazolam and fentanyl. Tracheal intubation was uneventful. Anesthetic maintenance was with titrated midazolam, fentanyl, and isoflurane 0.5% in oxygen. Neuromuscular blockade was achieved with pancuronium. The hemodynamics before cardiopulmonary bypass were as follows: central venous pressure, 12–14 mmHg; pulmonary artery systolic/diastolic pressures, 44–55 mmHg/25–30 mmHg; systemic blood pressure, 110–140 mmHg/60–80 mmHg; heart rate, 80–100 beats/min; cardiac index, 1.8–2.0 l · min⁻¹· m⁻². Intraoperative transesophageal echocardiography was performed (Omniplane II probe, Sonos 5500; Philips Medical Systems, Amsterdam, The Netherlands). The preoperative findings were confirmed. The patient underwent uncomplicated bioprosthetic aortic and mitral valve replacement on cardiopulmonary bypass (myocardial ischemic time, 140 min; cardiopulmonary bypass time, 207 min).

Separation from cardiopulmonary bypass was uneventful. The hemodynamics were as follows: central venous pressure, 14–18 mmHg; pulmonary artery systolic/diastolic pressures, 50–55 mmHg/30–35 mmHg; systemic blood pressure, 100–120 mmHg/50–60 mmHg; heart rate, 90–100 beats/min; cardiac index, 1.8–2.0 l · min⁻¹· m⁻². Intraoperative transesophageal echocardiography confirmed normal prosthetic valvular function, no change in the tricuspid regurgitation, moderate right ventricular dilation, and moderate biventricular dysfunction. Biventricular function was supported with infusions of epinephrine at 2 μg · kg⁻¹· min⁻¹and milrinone at 0.375 μg · kg⁻¹· min⁻¹. PHTN was managed with hyperventilation with 100% oxygen, titrated to arterial blood gas analysis. Despite aprotinin administration, there was significant coagulopathy after protamine reversal. This was corrected with transfusion of platelets and fresh frozen plasma. The patient was transferred to the intensive care unit while sedated on mechanical ventilation.

In the intensive care unit, the patient’s admission hemodynamics were as follows: central venous pressure, 14 mmHg; pulmonary artery systolic/diastolic pressures, 58 mmHg/30 mmHg; systemic blood pressure, 127 mmHg/66 mmHg; heart rate, 90–100 beats/min; cardiac index, 2.1 l · min⁻¹· m⁻². His PHTN increased despite sedation and hyperventilation with 100% oxygen (fig. 1) The right ventricular dysfunction was also progressive, as evidenced by a persistent increase in central venous pressure to 24–28 mmHg; cardiac index progressively decreased to below 2.0 l · min⁻¹· m⁻².

The milrinone infusion was increased to 0.5 μg · kg⁻¹· min⁻¹, and iPGI²was commenced at 50 ng · kg⁻¹· min⁻¹. Sedation was maintained with an infusion of fentanyl at 200–300 μg/h. These interventions decreased the PHTN by greater than 15%, expressed as systolic pulmonary arterial pressure/systemic systolic pressure. The PHTN was stabilized for 12 h in this fashion. Right ventricular function improved, as evidenced by a decrease in central venous pressure to 16–18 mmHg and a cardiac index of 2.2–2.6 l · min⁻¹· m⁻².

To facilitate tracheal extubation, a gradual withdrawal of the fentanyl infusion was begun. The iPGI²was then withdrawn in stages over 3 h while the patient was sedated with a fentanyl infusion at 50–75 μg/h. Right ventricular function was maintained with a cardiac index above 2.0 l · min⁻¹· m⁻²and a central venous pressure of 10–12 mmHg. The PHTN increased dramatically after cessation of iPGI². This increase persisted, and the PHTN increased to parasystemic levels, beyond the 70% systemic level at which iPGI²was commenced. This rebound PHTN occurred despite unchanged mechanical ventilation and the absence of agitation.

Four hours later, the sedation was terminated to prepare for tracheal extubation. Despite significant PHTN, the patient was calm, with adequate hemodynamics. However, during a coughing spell, the PHTN deteriorated and became suprasystemic. Acute right ventricular failure ensued, with low pulmonary arterial pressure and a doubling of the central venous pressure from 10 to 20 mmHg; the patient became profoundly hypotensive. This cardiogenic shock responded to bolus epinephrine, sedation, and reinstitution of iPGI². The hemodynamics rapidly recovered: central venous pressure, 13 mmHg; pulmonary artery systolic/diastolic pressures, 73 mmHg/24 mmHg; systemic blood pressure, 95 mmHg/50 mmHg; heart rate, 90–100 beats/min; cardiac index, 2.0–2.2 l · min⁻¹· m⁻². Transthoracic echocardiography at this time documented severe PHTN (estimated pulmonary artery systolic pressure, 78 mmHg), severe right ventricular dysfunction, and severe tricuspid regurgitation; these echocardiographic findings were all compatible with acute right heart failure associated with acute severe PHTN. The patient was ultimately successfully withdrawn from iPGI². However, he later succumbed to sepsis, complicated by multisystem organ failure.

This case report details management of acute-on-chronic PHTN after mitral and aortic valve replacement. iPGI²allows direct administration to the pulmonary arterial bed for selective pulmonary vasodilation; its short half-life of 2.7 min and rapid hydrolysis in blood prevent systemic arterial effects. 6The patient demonstrated an acute-on-chronic deterioration of PHTN that perhaps may stem in part from depletion of endogenous pulmonary endothelial nitric oxide due to cardiopulmonary bypass. 6,7The lack of normal pulsatile pulmonary blood flow decreases nitric oxide production by constitutive nitric oxide synthase. 4Coupled to this endogenous nitric oxide deficiency is a decrease in right ventricular reserve due in part to ischemia during aortic clamping and possible incomplete myocardial protection during cardiopulmonary bypass. Thus, the right ventricular failure is evident as afterload is increased while pump function is decreased.

This case also illustrates a tiered approach to management of PHTN. 8Initial management of right ventricular failure includes hypocarbia, hyperoxia, inotropic support, and further afterload reduction with milrinone. Selective pulmonary vasodilation is indicated when these initial measures are inadequate. The addition of iPGI²in the intensive care unit facilitated adequate control of the PHTN.

During withdrawal of selective pulmonary vasodilation, right ventricular afterload was acutely increased. The right ventricle was unable to cope with any further increases in afterload. The rebound PHTN after iPGI²withdrawal was immediately apparent and is readily explained by the short half-life of iPGI², as has been described for nitric oxide. This rebound PHTN predisposed the patient to cardiogenic shock, with further right ventricular challenge.

Although iPGI²offers significantly cheaper clinical equivalency compared with inhaled nitric oxide, it too has the clinical challenge of rebound PHTN on withdrawal in high-risk patients, as illustrated by this case report. Inhaled iloprost is a stable derivative of prostacyclin that can be administered intermittently because its effects last 1–2 h. 9Inhaled iloprost has been used successfully in Europe for PHTN after cardiac surgery and may represent a solution for the rebound PHTN that we have observed with prostacyclin. 10,11Because the effects of inhaled iloprost last up to 2 h, rebound PHTN may be less likely because acute loss of pulmonary vasodilation is unlikely. Rebound PHTN is a feature of inhaled pulmonary vasodilators with a half-life of minutes, previously described with nitric oxide and now also observed with iPGI². Inhaled iloprost may allow gradual controlled withdrawal of perioperative inhaled selective pulmonary vasodilation.

Inhaled iloprost is not currently available in the United States. Based on its favorable pharmacokinetics, it has great promise in the management of perioperative PHTN after cardiac surgery. It may represent another advance in perioperative inhaled pulmonary vasodilator therapy. Its place in perioperative management of PHTN and its integration with inhaled nitric oxide and iPGI²will be elucidated with further investigation and clinical experience.