Perioperative Challenges in Modified Blalock–Taussig Shunt in Neonates: A Single-center Experience

INTRODUCTION

The incidence of congenital heart disease (CHD) in Asia is about 9.3/1000 live births and nearly 25–30% of them constitute the cyanotic CHD (CCHD).^[1] Many such neonates have inadequate pulmonary blood flow and need palliative interventions early in life to alleviate severe cyanosis and prolong survival.^[2] These interventions include maintaining the patency of the ductus arteriosus by using a percutaneously inserted stent or by creating a surgical systemic to pulmonary artery (PA) shunt. A modified Blalock–Taussig shunt (MBTS) is the most common type of systemic to PA shunt, where the ipsilateral innominate or subclavian artery is connected to the PA using a prosthetic graft (polytetrafluoroethylene) to allow interim growth of pulmonary arteries till a suitable age is reached to undergo a definitive repair of the underlying cardiac condition.^[3] Despite considerable improvements in surgical and anesthesia management, the MBTS procedure in neonates is still associated with significant morbidity and mortality. It is also relevant to say that although MBTS is a widely performed palliative procedure, literature focusing exclusively on the perioperative anesthetic and monitoring challenges in neonates remains limited. This retrospective study aimed to evaluate perioperative anesthetic challenges and the significance of monitoring strategies in early detection and management of complications in neonates who underwent MBTS at our center.

MATERIALS AND METHODS

The medical records of all neonates who underwent MBTS in our institute from August 2016 to May 2018 were reviewed, and those who satisfied the inclusion criteria (managed by a single surgical and anesthesiologist team) were included in this retrospective study. This inclusion method was followed to maintain consistency in the outcome. Data from 14 such neonates were included in the final analysis. Those with a prior history of cardiac arrest, heart failure, and on pre-operative inotropes were excluded from the study protocol. The perioperative anesthetic management was done according to our institutional practice. No premedication was given. L-thyroxine at a dose of 10 μg/kg/day was continued for two neonates who were also suffering from congenital hypothyroidism. Prostaglandin E1 was continued in five neonates who were receiving it to keep the ductus arteriosus patent because of the duct-dependent circulation. In brief, neonates received ketamine or sevoflurane as an induction agent, depending on the presence or absence of an intravenous (IV) line, along with fentanyl 2–5 μg/kg and midazolam 0.1 μg/kg. Neuromuscular block was facilitated with rocuronium bromide for nasotracheal intubation in all except one neonate with cleft palate, in whom oral intubation was done. All had a femoral arterial line and a central venous catheter in the right internal jugular vein, in place. Anesthesia was maintained with intermittent doses of fentanyl, midazolam, and rocuronium bromide. In addition to standard monitoring (pulse oxymetry, invasive blood pressure, central venous pressure, nasopharyngeal temperature, and urine output) as per any other cardiac surgical patient, hematocrit (Hct), arterial oxygen saturation (SaO₂), fraction of inspired oxygen (FiO₂), partial pressure of O₂ (PaO₂) and CO₂ in arterial blood (PaCO₂), blood lactate, and blood sugar were also monitored. Bispectral index was monitored for depth of anesthesia, and to maintain an appropriate anesthetic dose, and near-infrared spectroscopy (NIRS) (INVOS 5100C, Somanetics, Troy, MI, USA) for continuous monitoring of cerebral perfusion was done in all. Prostaglandin E1 (0.05–0.1 μg/kg/min) infusion was continued until the completion of shunt anastomosis and establishment of shunt blood flow in neonates who were already receiving it.

The surgical approach was done by right or left posterolateral thoracotomy or median sternotomy, depending on the site of the arch or difficulty in approach to PA. The size of Gore-Tex graft polytetrafluoroethylene was selected on the basis of body weight: 30 mm for under 2.5 kg and 3.5 mm for more than 2.5 kg. The graft was used as a conduit between the subclavian artery and a branch of PA. FiO₂ was kept high to maintain pulmonary blood flow and thereby SaO₂ more than 80% except for neonates with patent ductus arteriosus (PDA) in whom FiO₂ was kept low to prevent ductal closure (titration of FiO₂ was done to keep SaO₂ around 80–85%, never beyond 85%). Before clamping of the subclavian artery, heparin was given in a dose of 150 units/kg to ensure an activated clotting time (ACT) of >250 s. Post-procedure, heparin reversal was not done, and FiO₂ was titrated to keep SaO₂ between 75% and 80%. This was done to avoid a high FiO₂-related increase in pulmonary blood flow. Intraoperative fluid management included continuation of maintenance fluids and or addition of normal saline or Ringer’s lactate solution as required. Dopamine infusion was initiated coinciding with skin incision, at a dose of 10 μg/kg/min, to achieve the therapeutic blood level just before the shunt flow was established, and to maintain cardiac output and keep the shunt patent. Post-procedure Hct was maintained between 40% and 45% to avoid shunt thrombosis at higher Hct. FiO₂ and PaCO₂ values were titrated to avoid any increase or decrease in shunt blood flow due to these two variables. The shunt patency was assessed throughout the hospital stay by clinical presence of a shunt murmur, SaO₂, and PaO₂ values. In case of suspected shunt block, 2-D transthoracic echocardiography (TEE) was performed to visualize the presence or absence of shunt flow. The pre-operative and immediate post-operative values for arterial blood pressures, nasopharyngeal temperature, Hct, SaO₂, FiO₂, PaO₂, lactate, blood sugar, and NIRS readings were noted. Any alteration in these values from normal was documented and corrected. Postoperatively, all neonates were mechanically ventilated and received continuous infusion of fentanyl at a dose of 1–2 μ/kg/h for analgesia. Muscle relaxants were not used routinely. No patient required any blood transfusion. In addition to dopamine, a few neonates received dobutamine, adrenaline, noradrenaline, or milrinone, alone or in combination, to maintain the hemodynamic parameters, depending on the need. Heparin infusion was started at a dose of 10 units/kg/h, to maintain ACT within 1.5–2 times of control, within 4–6 h of surgery. Once feeding was started, oral or through Ryle’s tube, aspirin 5 mg/kg/day was given to all, and was continued until a definitive surgery was undertaken. The days of inotrope use, intraoperative and post-operative complications, duration of intensive care unit (ICU) stay, and hospital stay were also recorded for each case.

RESULTS

The demographics, pre-operative diagnosis, and clinical characteristics are described in Table 1. Mean age and body weights of the 14 neonates were 14.4 days (4–23 days) and 3.2 kg (2.8–3.5 kg). 11 were male babies. One (7.14%) neonate had a cleft palate, and two (14.3%) with congenital hypothyroidism were on daily L-thyroxine (10 μg/kg). The most common anomaly was Tetralogy of Fallot (TOF), in 9 patients (64.28%). Five neonates (35.7%) had ductal dependent anomalies. [Transposition of great vessels(TGA) (2cases,14.28%),Pulmonary atresia with intact ventricular septum (PA,IVS)(one case,7.14%), Single ventricle (two cases,14.28%. These neonates were kept on prostacycline E1 (0.05-0.1mic/kg/min)infusion to maintain patency of ductus. Three neonates (21.42%)were on mechanical ventilatory support [Table 1]. Eight neonates had metabolic acidosis.

Surgery was performed through right thoracotomy in 9 (64.28%), left thoracotomy in 2 (14.28%), and median sternotomy in 3 (21.42%) cases. The duration of surgery, mechanical ventilation, ICU stay, hospital stay, and days of inotrope use is represented in Table 2. Apart from 3 cases (21.42%), all needed more than 48 h of mechanical ventilation. There was significant improvement in postoperative PaO₂ (from 20 ± 9.8 mmHg to 45 ± 4.2 mmHg, P < 0.001) and SaO₂ (from 60 ± 6.5 to 80 ± 4.5, P < 0.0001) when compared to the pre-operative values [Table 3].

Although the difference in blood sugar, nasopharyngeal temperature, and lactate levels in the pre-operative and immediate post-operative period was statistically significant, it was not clinically relevant [Table 3]. All patients were on dopamine postoperatively. The median duration of inotrope infusion was 7.5 days (range 3–15 days). In addition, 5 (35.7%) neonates were on dobutamine for 5–12 days (median 8), 3 on adrenaline for 8–15 days (median 10), 4 on noradrenaline for 6–16 days (median 10.5), and 2 on milrinone for 9 days [Table 4].

Eight neonates faced challenges in the perioperative period. Two (14.28%) had intraoperative hypercyanotic spell and hemodynamic instability during isolation of the pulmonary vessels. This was accompanied by a sudden drop in NIRS value from the baseline (35 and 42, respectively), lasting for <3 min. They were managed by transient withholding of surgery, and bolus of phenylephrine (0.02 mg/kg) and 5–6 mL/kg fluid, intravenously. One (7.14%) patient had loss of shunt murmur and hypoxemia on post-operative day 1, indicating shunt under perfusion and graft thrombosis. Re-exploration and thrombectomy were done, and the patient’s clinical profile improved [Table 5]. Another patient developed severe cardiorespiratory dysfunction, necessitating tracheostomy, on the 10^th post-operative day. He had four failed extubation attempts. He was on continuous ventilation, and additional adrenaline (1 μg/kg/min), noradrenaline (1 μg/kg/min), and milrinone (5 μg/kg/min) infusions. The duration of ICU (15 days) and hospital stays (21 days) was the longest in this patient. Two (14.28%) neonates had an increase in blood lactate, pulmonary plethora in chest X-ray, high SaO₂ (>85%), and low diastolic blood pressure. Shunt overflow was the cause. They were managed with reduction of FiO₂ to <0.4, fluid restriction, and higher positive end expiratory pressure (>8 cm H₂O).

There were two (14.28%) mortalities (on post-operative days 5 and 11), because of severe biventricular dysfunction. Both were on pre-operative mechanical ventilatory support. In the follow-up, the graft was patent in all 12 survivors at the time of discharge from the hospital.

DISCUSSION

MBTS is an effective palliative procedure in some CCHD to improve pulmonary blood flow and alleviate cyanosis till a definitive surgery is planned.^[3] However, despite the best possible perioperative management, neonates undergoing this procedure have a high incidence of morbidity and mortality because of pre-operative ventilation, surgical stress, the effect of long-term sedation and muscle relaxants on immature myocardium, as well as respiratory muscles.^[4,5] Perioperative management also becomes more complicated due to the presence of extracardiac problems and immature neonatal physiology to handle additional perioperative stress. Possible surgical injury to nearby structures further complicates the post-operative course, which did not happen to any of our cases.^[6] Most of the induction agents are well tolerated by these neonates. We preferred ketamine in those with an IV access and sevoflurane in others because of the cardio-stable properties of these two induction agents.^[3,7] Long-term mechanical ventilation is not unusual in neonates undergoing cardiac surgery. At our institute, it is a standard protocol to go for nasal intubation to prevent accidental extubation, which is more common in neonates with an orotracheal tube.

Although hypercynotic spells leading to unstable hemodynamics are more common in 2–6 months of age, it is a possibility in neonates too.^[3,8] This happens in neonates with significant right ventricular outflow tract (RVOT) obstruction. Precipitating factors such as pain, agitation, dehydration, or fever can lead to RVOT spasm, leading to a decrease in pulmonary blood flow and hypercyanotic spells. These episodes are characterized by tachypnea, hyperpnea, and worsening of cyanosis, which, if left untreated, may lead to cardiac arrest. We had all the preparation to combat such an incidence in our patients.

For surgical access for MBTS, posterolateral thoracotomy is the usual approach in neonates, and most of our cases underwent the same.^[3,4,6] Although sternotomy allows more central placement of shunt and thus more uniform growth of pulmonary arteries, the author’s “institute surgeons prefer posterolateral thoracotomy because of the concerns with resternotomy at a later date for definitive surgery. Out of the three neonates who underwent median sternotomy approach, two were on pre-operative mechanical ventilation and one had PA with an intact ventricular septum, but pulmonary vascular anatomy was unclear on TEE. The former two were too sick to tolerate prolonged lung compression as happens during lateral thoracotomy, and in the latter, the initial thought was for a central shunt placement.

We avoided hyperventilation and excessive inhalational agents all throughout because of their tendency to decrease pulmonary vascular resistance, increase systemic to PA flow ratio (QP:QS), and decrease diastolic pressure, with resultant concerns for myocardial ischemia and systemic perfusion.^[5,8]

We planned to put an arterial line in the femoral vessels because, at times, the subclavian artery is confused with PDA and clamped during anastomosis.^[9] The absence of a femoral arterial trace can diagnose accidental clamping immediately. None of our cases had post-clamp release (after the shunt is opened) hemodynamic instability and acidosis, which may occur due to systemic run-off to PA and ischemia reperfusion injury (following clamp removal).^[10] We avoided this situation with pre-emptive administration of IV fluids, sodium bicarbonate, and initiation of inotrope infusion.

Some advanced monitoring remains the mainstay of a successful outcome, allowing for early detection of changes in parameters so that timely corrective measures can be taken. The authors in the present study measured regional cerebral oxygen saturation (rSO₂) by NIRS, and no significant change (fluctuation was within 20% of baseline) was noticed in the pre-operative and post-operative values. This latest technology helps the anesthesiologists to detect any overt neurological injury during a cardiac surgical procedure, as reported previously by Lamba et al. and Spaeder et al.^[11,12] Votava-Smith et al. demonstrated that even term neonates with CHD showed altered cerebral blood flow (CBF) and impaired cerebral autoregulation.^[13] Yamamoto et al. suggested that infants with CHD have immature cerebral autoregulation because changes in rSO₂ are predominantly dependent on changes in systemic blood pressure.^[14] Moreover, changes in QP:QS caused by MBTS may lead to changes in systemic blood flow and oxygenation, implying that this may affect cerebral circulation in these neonates.^[15] Post MBTS QP:QS increases, so systemic blood flow may decrease, with a decrease in CBF, even if minimal.

Our routine practice after MBTS is to regulate FiO₂ by keeping SpO₂ stable, and blood transfusion to keep Hct around 40–45%. We maintained systemic blood pressure using dopamine and added adrenaline and noradrenaline whenever required. These strategies might contribute to the stabilization of cerebral perfusion pressure and MBTS flow dynamics. We also emphasize that there was a possibility of an increase in CBF proportional to the rise in systemic blood pressure. Intraoperative blood transfusion can also inhibit CBF by causing an increase in Hct and blood viscosity.

The neonates in our series who were extubated within 48 h had TOF’s with a mild grade of pulmonary stenosis; an anatomy much simpler than the others, similar to the results of previous authors.^[4,5,8] Shunt thrombosis is a serious complication following MBTS procedure, which can lead to sudden cardiac arrest or even death if left untreated. Several factors were studied as causative agents, but findings were inconclusive, and it is mostly considered of unknown etiology.^[16] In spite of the fact that the size of the shunt was small, the incidence of shunt block was quite low in our series. This is because of the continuation of heparin infusion in the early post-operative period for 3–4 days, and aspirin thereafter. Mean pre-operative SaO₂ and PaO₂ were 60 ± 6.5% and 20 ± 9.8 mmHg, which increased to 80 ± 4.5% and 45 ± 4.2 mmHg, respectively, after shunt placement. These findings are similar to those of previous investigators.^[17] Graft patency rate, duration of mechanical ventilation, ICU stay, and hospital stay in our study are also comparable with the results of other authors.^[15-17]

Various predictors of mortality after MBTS described in literature include body weight (<2 kg), small shunt size (<4 mm), univentricular heart, pre-operative ventilation, sepsis, complex cardiac anomalies, and need for cardiopulmonary bypass.^[17-19] Mortality, in our study, was due to severe ventricular dysfunction, and both these neonates were on mechanical ventilation even before surgery, which denotes their critical pre-operative status. Rao et al. (2000) reported sepsis as the main cause of mortality in their patients.^[19] According to the Society of Thoracic Surgeons congenital heart surgery database, mortality after congenital heart surgery was significantly higher in pre-operative mechanically ventilated neonates compared to those who did not receive the same.^[20] Mechanical ventilation contributes to poor outcomes in general by its impact on heart–lung interaction, and alteration in systemic to pulmonary blood flows. The need for sedatives and muscle relaxants further increases the risk of pulmonary complications, leading to increased incidence of post-operative morbidity and mortality.

Maintenance of intravascular fluid status with normal saline, Glucose, or Ringer’s lactate solution is important, as is maintenance of Hct below 40–45% by blood transfusion if required, to prevent both shunt overflow and shunt underflow complications, along with regulation of FiO₂, and maintenance of SaO₂ and PaO₂.

CONCLUSION

Physiologically, neonates are not miniature children or infants. The MBTS in neonates though, has excellent results, possesses challenges for the team of anesthesiologists due to associated comorbidities, challenges due to intrinsic neonatal physiology apart from the lesion itself. Intense perioperative monitoring is the mainstay of a good outcome. We suggest the need for multicenter prospective research to make strategies to reduce further perioperative morbidity and mortality in this subset of patients.