Obstructive hydrocephalus can arise from a number of causes and treatment involves creating an extracranial drainage path for cerebrospinal fluid, usually with a ventriculoperitoneal (VP) shunt. Placement of a VP shunt involves incising and retracting the scalp, making a burr hole, and advancing a catheter into a ventricle, often with the assistance of image guidance and an endoscope. An abdominal incision is performed with dissection through the anterior rectus sheath. At this point, a shunt passer is inserted subcutaneously from the scalp incision to the abdominal incision.
Although each step of this operation can result in significant pain and discomfort, the advancement of the shunt passer is the most stimulating aspect of the procedure, and for this reason, the operation usually is performed under general endotracheal anesthesia.1 However, general anesthesia can be undesirable in patients with certain coexisting diseases, including severe pulmonary arterial hypertension (PAH). Even with smooth intraoperative hemodynamics, patients with PAH are prone to postoperative decompensation. Depending on the surgical procedure and underlying disease cause and severity, noncardiac surgery in patients with PAH carries a morbidity risk between 40% and 50% and a mortality risk between 4% and 24%.2 In patients with PAH associated with an intracardiac shunt, that is, Eisenmenger syndrome, perioperative risks are even more extreme,3 but given the reluctance to take these patients for anything but transplantation surgery, there is currently no published data on their overall perioperative risk of morbidity and mortality.
We present the case of a 32-year-old man with severe PAH and Eisenmenger syndrome undergoing urgent VP shunt placement under monitored anesthetic care. The patient provided written and verbal consent for publication of this case report.
A 32-year-old man (55.1 kg, 170 cm) presented to the operating room (OR) for VP shunt placement for symptomatic obstructing hydrocephalus secondary to a 3.2-cm colloid cyst in his third ventricle. His symptoms included a 6-month history of headache, nausea, and vomiting. His medical history was significant for severe PAH secondary to residual congenital ventricular septal defect (VSD) after attempted surgical closure at 3 years of age (group 1 PAH; Table 1).4 He had bidirectional shunting via a small membranous VSD and a large complex muscular VSD. The last cardiac catheterization showed a pulmonary to systemic shunt ratio of 1.22 with a pulmonary artery pressure of 120/53, a pulmonary vascular resistance (PVR) of 13.8 Woods units, and a cardiac output of 4.15 L/min. Before admission, he was a stable New York Heart Association Class II on home oxygen, sildenafil, and ambrisentan. He had previously been on inhaled treprostinil but had stopped this 2 weeks before admission because of persistent headaches. He had a previous episode of atrial tachycardia not requiring ongoing pharmacologic management. Reflux symptoms were controlled with famotidine. Laboratory studies were remarkable for polycythemia (hemoglobin 17.9 g/dL, hematocrit 52%) and normal renal function. Transthoracic echocardiography 1 month before surgery showed a hypertrophic right ventricular (RV) with moderate dilation and systolic dysfunction, preserved left ventricular systolic function (ejection fraction, 55%–60%), and severe tricuspid regurgitation with an estimated RV systolic pressure of 118 mm Hg.
Before entering the OR, IV lines were meticulously deaired, and 18-g and 20-g peripheral IV catheters were placed in the left hand and right forearm, respectively. Standard American Society of Anesthesiologist monitors and a lower body-forced air warmer were applied along with high-flow nasal cannula oxygen at 8 L/min and capnography. A central line was not placed because of the risk of pneumothorax. A total of 2 mg IV midazolam was given, and dexmedetomidine was loaded with a 1-μg/kg dose over 10 minutes then continued with an infusion at 0.4 μg/kg/h. Vancomycin was administered for antimicrobial prophylaxis, given patient allergies to penicillin and levofloxacin. A 20-g arterial catheter was inserted in the left radial artery for invasive blood pressure (BP) monitoring. Baseline vital signs were normal, sinus rhythm was 69 beats per minute, BP 118/68 mm Hg, and respiratory rate was 14 with an Spo2 of 92%. In response to a decrease in BP to 92/58 mm Hg, an epinephrine infusion was started at 0.02 μg/kg/min and a vasopressin 0.5 units IV bolus was given with immediate improvement in hemodynamics. Inhaled nitric oxide was started at 20 parts per million (PPM), and 75 mg amiodarone was slowly infused to provide atrial dysrhythmia prophylaxis.
After the patient was prepared and draped, the dexmedetomidine infusion was increased to 0.6 μg/kg/min, and a 10-mg ketamine bolus was given in anticipation of pain from local anesthetic infiltration. A total of 30 mL of 1% lidocaine with 1:200,000 epinephrine was infiltrated along the surgical tract from the scalp to the abdomen. Before insertion of the shunt passer, an additional 2 mg midazolam and 10 mg ketamine were given. The shunt passer was advanced without objection from the patient and with stable hemodynamics. Two additional doses of 10 mg ketamine IV were given during catheter placement, which took approximately 10 minutes. With endoscopic guidance, the tip of the VP shunt was placed into the right frontal horn and the septum pellucidum was fenestrated in several locations. At this point, the dexmedetomidine infusion was reduced to 0.3 μg/kg/min in anticipation of the conclusion of surgery. Arterial blood gas analysis was unremarkable (pH 7.4, Paco2 38 mm Hg, Pao2 111 mm Hg, base deficit 1.3 mmol/L, HCO3 22.9 mmol/L), and electrolytes were supplemented with 750 mg calcium chloride IV and 5 mEq potassium chloride IV in response to findings of mild hypocalcemia (1.1 mmol/L) and hypokalemia (3.1 mmol/L). During incision closure, the inhaled nitric oxide was tapered off over 25 minutes before leaving the OR; 1000 mg Tylenol IV was administered. A total of 500 mL Normosol® (Hospira Inc., Lake Forest, IL) was administered throughout the intraoperative period. The epinephrine infusion was weaned off as sedation wore off with no change in hemodynamic parameters. The patient remained in normal sinus rhythm without premature atrial or ventricular beats with an Spo2 >92% throughout the procedure; 4 mg ondansetron IV was given for postoperative nausea and vomiting prophylaxis.
The patient was transferred to the recovery room on facemask oxygen and then sent to the floor with nasal oxygen in place. On postoperative day 2, he reported nonspecific pleuritic chest pain and required oxygen supplementation to maintain Spo2 >90%. Workup for cardiac ischemia and pulmonary embolism was negative. Expectant management of his presumed increased right-to-left shunt with diuresis and pulmonary exercises stabilized his symptoms, and he was discharged home on postoperative day 3 with New York Heart Association class III symptoms.
VP shunt insertion typically requires general anesthesia, primarily because of the pain associated with the subcutaneous tunneling phase and the need to avoid patient movement during critical aspects of the surgery.1 However, our patient’s severe underlying cardiopulmonary disease increased his perioperative risk of morbidity and mortality, to the point where the risk of general anesthesia was prohibitive, and efforts to coordinate the procedure under monitored anesthesia care were deemed in his best interest.
The reasons for avoiding general anesthesia in patients with PAH are multifactorial but broadly based on the need to preserve myocardial contractility, baseline RV mechanics, and euvolemia and to avoid increasing RV workload. Positive pressure ventilation increases intrathoracic pressures, decreasing venous return, increasing RV end-diastolic volume, and subsequently decreasing RV cardiac output and pulmonary blood flow.5 Positive pressure ventilation also produces nonuniform ventilation patterns and may have a negative effect on lung perfusion.6 Mild myocardial depressant effects are seen with many of the IV and inhaled anesthetics in modern use, which may not be tolerated by patients with pre-existing ventricular dysfunction. General anesthesia combined with neuromuscular blockade interferes with thermoregulation, potentially increasing PVR because of hypothermia. In addition, hypercarbia, respiratory or metabolic acidosis, and hypoxic pulmonary vasoconstriction increase PVR and increase RV workload. In the population with PAH, any of these factors places these patients at risk for acute RV decompensation with subsequent hypoxemia and RV failure. In addition to these risks, in patients with Eisenmenger syndrome, following any minor decrease in systemic BP, an increase in intracardiac right-to-left shunting will occur, causing systemic hypoxemia and hemodynamic instability leading to cardiovascular collapse and death.3,7 Patients with Eisenmenger syndrome are also at increased risk for systemic air embolism, thromboembolic complications, arrhythmias, and depression of ventricular function because of acute increases in afterload.7
Regardless of the type of anesthesia to be administered, optimization of preoperative functional capacity and cardiopulmonary medications is essential for any patient with PAH. Specific intraoperative goals for this patient included prevention of hypercarbia, hypoxia, hypothermia, acidosis, excessive sympathetic stimulation, myocardial depression, and hypervolemia. Signs or symptoms of worsening ventricular function, systemic malperfusion, and hypoxemia must be aggressively sought, identified immediately, and treated expeditiously to avoid cardiovascular collapse. In addition to standard monitors, invasive BP monitoring and intraoperative arterial blood gas analysis, intraoperative transesophageal echocardiography, or pulmonary artery catheter placement can be considered by providers skilled at interpreting these modalities. There are important limitations to both these monitoring modalities—most notably, in the case described, transesophageal echocardiography usually requires general anesthesia with a protected airway and multiple intraventricular communications deemed placement of a pulmonary artery catheter too risky.
The pharmacological strategy for this case was prioritized according to the need to maintain respiratory drive to avoid hypercarbia and hypoxia and to blunt any pulmonary hemodynamic response to noxious stimuli while providing sedation deep enough to avoid patient movement at critical times. Dexmedetomidine was chosen for its ability to produce sedation and analgesia while preserving spontaneous ventilation. Midazolam was used to deepen the level of sedation because of its easy reversibility with flumazenil. Opioids were avoided because of their respiratory depressant effects. Systemic hypotension was controlled with an epinephrine infusion and vasopressin boluses to provide both vasoconstriction and inotropic effects. Ketamine boluses were used sparingly and cautiously and only in anticipation of especially stimulating events during the surgery. Inhaled nitric oxide was selected as a pulmonary vasodilator given the more favorable antiplatelet profile compared with epoprostenol8 and the need to minimize the risk of surgical bleeding. Many other options are available for inotropic and vasoactive support, including norepinephrine, milrinone, and dobutamine, and selection of specific agents often comes down to intraoperative hemodynamics and individual practitioner expertise and preference. In this case, low-dose epinephrine infusion, vasopressin boluses, and inhaled nitric oxide resulted in optimal hemodynamic conditions with a stable systemic BP with an absence of hypoxia and metabolic acidosis, evidence of PVR, and RV function remaining at baseline.
Effective communication with our surgical colleagues was important in devising the optimal anesthetic and surgical strategies. Surgically, a VP shunt was considered the simplest and most effective procedure to alleviate the patient’s symptoms. Thought was given to other shunt locations including ventriculoatrial and ventriculopleural shunts, but potential complications, including arrhythmias and pneumothorax, and the presence of intracardiac shunts made these approaches prohibitively risky for this patient.
This case report is the first published description of placement of a VP shunt under monitored anesthesia care. A multidisciplinary approach with involvement of both cardiac and neuroanesthesia teams, and careful perioperative planning, resulted in a successful outcome in a patient with otherwise extreme anesthetic risk because of severe PAH and Eisenmenger syndrome.
1. Prabhakar H, Rath GP, Bithal PK, Chouhan RS. Intracranial pressure and haemodynamic changes during the tunnelling phase of ventriculoperitoneal shunt insertion. Eur J Anaesthesiol 2005;22:94750.
2. Blaise G, Langleben D, Hubert B. Pulmonary arterial hypertension: pathophysiology and anesthetic approach. Anesthesiology 2003;99:141532.
3. Ammash NM, Connolly HM, Abel MD, Warnes CA. Noncardiac surgery in Eisenmenger syndrome. J Am Coll Cardiol 1999;33:2227.
4. Simonneau G, Gatzoulis MA, Adatia I, Celermajer D, Denton C, Ghofrani A, Gomez Sanchez MA, Krishna Kumar R, Landzberg M, Machado RF, Olschewski H, Robbins IM, Souza R. Updated clinical classification of pulmonary hypertension. J Am Coll Cardiol 2013;62:D3441.
5. Matthews JC, McLaughlin V. Acute right ventricular failure in the setting of acute pulmonary embolism or chronic pulmonary hypertension: a detailed review of the pathophysiology, diagnosis, and management. Curr Cardiol Rev 2008;4:4959.
6. Soni N, Williams P. Positive pressure ventilation: what is the real cost? Br J Anaesth 2008;101:44657.
7. Oechslin E, Mebus S, Schulze-Neick I, Niwa K, Trindade PT, Eicken A, Hager A, Lang I, Hess J, Kaemmerer H. The adult patient with Eisenmenger syndrome: a medical update after Dana Point Part III: specific management and surgical aspects. Curr Cardiol Rev 2010;6:36372.
8. Thunberg CA, Morozowich ST, Ramakrishna H. Inhaled therapy for the management of perioperative pulmonary hypertension. Ann Card Anaesth 2015;18:394402.