Introduction

Over the last 80+ years, congenital cardiac anesthesia has developed as a subspecialty of pediatric and/or cardiac anesthesia, depending on one’s perspective. It is impossible to describe the evolution of congenital cardiac anesthesia without simultaneously referring to developments in the surgical and interventional management of congenital heart disease (CHD) due to the great interdependency of these fields. As pediatric anesthesia developed, surgical treatments and percutaneous interventions for children with CHD began to be devised, starting with the simple surgical ligation of a patent ductus arteriosus (PDA). Over the years, the continual development and refinement of surgical and interventional techniques has led to the introduction of staged palliations, increasingly sophisticated repairs of complex intracardiac lesions requiring cardiopulmonary bypass (CPB) and circulatory arrest, and most recently, of complex biventricular repairs. These advances were accompanied by concurrent changes in anesthetic management that overcame the technical challenges and mitigated the perioperative morbidity associated with these novel techniques.

This chapter will be organized primarily around the theme of how anesthesiologists met these new challenges using the anesthetic armamentarium that was available to them at the time. The secondary theme running through this history is the gradual shift from a purely intraoperative focus to one with a view of the entirety of the perioperative period. Not surprisingly, the drastic decreases in surgical mortality have led to an increased interest in the reduction of perioperative morbidity. The final theme is the progressive and ongoing expansion in the age range of patients routinely presenting for congenital cardiac surgery.

The first years: 1938–1954

The first successful operation for CHD was performed in August 1938 when Robert E. Gross ligated the PDA of a 9‐year‐old girl. The operation and the postoperative course were smooth, but because of the interest in the case, the child was kept in the hospital until the 13th postoperative day. In the report of the case, Gross mentions that the operation was done under cyclopropane anesthesia, and continues: “The chest was closed, the lung being re‐expanded with positive pressure anesthesia just prior to placing the last stitch in the intercostal muscles.” A nurse using a “tight‐fitting” mask delivered the anesthetic. There was no intubation and, of course, no postoperative ventilation. The paper does not mention any particular pulmonary complications, so it cannot have been much different from the ordinary postoperative course of the day [1].

In 1952, Dr. Gross published a review of 525 PDA ligations where many, if not all, of the anesthetics were administered by the same nurse anesthetist, under surgical direction [2]. Here he states: “Formerly, we employed cyclopropane anesthesia for these cases, but since about half of the fatalities seemed to have been attributable to cardiac arrest or irregularities under this anesthetic, we have now completely abandoned cyclopropane and employ ether and oxygen as a routine.” It is probably correct that cyclopropane under these circumstances with insufficient airway control was more likely than ether to cause cardiac arrhythmias. An intralaryngeal airway was used, which also served “to facilitate suction removal of any secretions from the lower airway” (and, we may add, the stomach). Dr. Gross claims that the use of this airway reduced the incidence of postoperative pulmonary complications. Without a modern, rigorous review of this series, it is hard to know what other particular anesthetic challenges were confronted by the anesthetists, but we may assume that intraoperative desaturation from the collapsed left lung, postoperative pulmonary complications, and occasional major blood loss from an uncontrolled, ruptured ductus arteriosus were high on the list.

The next operation to be introduced, the systemic to pulmonary artery (PA) shunt, was billed as “corrective” for the child with cyanotic CHD. The procedure was proposed by Helen Taussig as an “artificial ductus arteriosus” and was first performed by Albert Blalock and the surgical technician Vivien Thomas at the Johns Hopkins Hospital in 1944. In a very detailed paper, Drs. Blalock and Taussig described the first three patients to undergo the Blalock–Taussig shunt operation, recently recognized as Blalock‐Thomas‐Taussig shunt. Dr. Harmel anesthetized the first and third patients, using ether and oxygen in an open drop method for the former and cyclopropane through an endotracheal tube for the latter. The second patient was given cyclopropane through an endotracheal tube by Dr. Lamont. Whether the first patient was intubated is unclear, but it is noted that in all three cases, positive pressure ventilation was used to reinflate the lung [3]. Interestingly, in this early kinder and gentler time, the surgical and pediatric authors reporting the Blalock–Taussig operation acknowledged by name the pediatricians and house officers who took such good care of the children postoperatively, but still did not acknowledge the contribution of the anesthesiologists, Lamont and Harmel. Although intubation of infants was described by Gillespie as early as 1939, it is difficult to say precisely when intubations became routine [4].

Drs. Harmel and Lamont reported in 1946 on their anesthetic experience of 100 operations for congenital malformations of the heart “in which there is pulmonary artery stenosis or atresia.” They reported 10 anesthetic‐related deaths in the series, so it is certain that they encountered formidable anesthetic problems during these surgical procedures [5]. In 1952, Damman and Muller reported a successful operation in which the main PA was reduced in size and a band was placed around the artery in a 6‐month‐old infant with a single ventricle (SV). They state that morphine and atropine were given preoperatively, but no further anesthetic agents are mentioned. At that time, infants were assumed to be oblivious to pain, so we can only speculate on what was used beyond oxygen and restraint [6].

Over the next 20 years, many palliative operations for CHD were developed and a number of papers describing the procedures and their anesthetic management appeared in the literature. In 1948, McQuiston described the anesthetic technique used at the Children’s Memorial Hospital in Chicago [7]. This is an excellent paper for its time, however, a number of the author’s conclusions are erroneous. The anesthetic technique for shunt operations (mostly Potts’ anastomosis) is discussed in some detail but is mostly of historical interest today. McQuiston explained that he had no experience of anesthetic management used in other centers, such as the pentothal–N₂O–curare used at Minnesota or the ether technique used at the Mayo Clinic. McQuiston used heavy premedication with morphine, pentobarbital and atropine, and/or scopolamine; this is emphasized because it was important “to render the child sleepy and not anxious.” The effect of sedation with regard to a decrease in cyanosis (resulting in making the child look pinker) is noted by the authors. They also noted that children with severe pulmonic stenosis or atresia do not decrease their cyanosis “because of very little blood flow,” and that these children have the highest mortality. McQuiston pointed out that body temperature control was an important factor in predicting mortality and advocated the use of moderate hypothermia (i.e., “refrigeration” with ice bags), because of a frequently seen syndrome of hyperthermia. McQuiston worked from the assumption that hyperthermia is a disease in itself, but did not explore the idea that the rise in central temperature might be a symptom of low cardiac output with peripheral vasoconstriction. Given what we now know about shunt physiology, it is interesting to speculate that this “disease” was caused by pulmonary hyperperfusion after the opening of what would now be considered as an excessively large shunt, stealing a large portion of systemic blood flow.

In 1950, Harris described the anesthetic technique used at Mount Zion Hospital in San Francisco [8]. He emphasized the use of heavy premedication with morphine, atropine, and scopolamine. The “basal anesthetic agent” was Avertin (tribromoethanol) administered rectally and supplemented with N₂O/O₂ and very low doses of curare. Intubation was facilitated by cyclopropane with FiO₂ titrated according to cyanosis throughout. Bucking or attempts at respiration were thought to be due to stimulation of the hilus of the lung and were treated with “cocainization” of the hilus [8].

In 1952, Dr. Robert M. Smith discussed the circulatory factors involved in the anesthetic management of patients with CHD [9]. The anesthetic agents recommended were mostly ether following premedication. He pointed out the necessity of understanding the pathophysiology of the lesion and also “the expected effect of the operation upon this unnatural physiology.” That is, he recognized that the operations were not curative. While most of the previous papers had been about tetralogy of Fallot (TOF), Dr. Smith also described the anesthetic challenges of surgery for coarctation of the aorta, introduced simultaneously in 1945 by Dr. Gross in the United States and Dr. Craaford in Sweden. He emphasized hypertension following clamping of the aorta and warned against excessive bleeding in children operated on at older ages using ganglionic blocking agents. This bleeding was far beyond what anesthesiologists now see in patients operated on at younger ages, before the development of substantial collateral arterial vessels [9].

The heart–lung machine: 1954–1970

From 1954 to 1970 the development of what was then called the “heart–lung machine” allowed for the surgical repair of complex intracardiac congenital heart defects. Early CPB technology was associated with significant morbidity and mortality, especially in smaller children weighing less than 10 kg. In Kirklin’s initial report of intracardiac surgery utilizing a mechanical pump–oxygenator system at the Mayo Clinic, the only reference to anesthetic management was a brief remark that ether and oxygen were given [10]. In Lillehei’s description of direct vision intracardiac surgery in humans using a simple, disposable artificial oxygenator, there was no mention of anesthetic management [11]. What strikes a “modern” cardiac anesthesiologist in these two reports is the high mortality: 50% in Kirklin’s series and 14% in Lillehei’s series. All of these patients were children with CHD ranging in age from 1 month to 11 years.

Anesthetic challenges multiplied rapidly with the introduction of CPB and early attempts at intracardiac repairs. At that time, most anesthetics were performed by nurses under the supervision of the surgeon utilizing ether (first with open drop ether administration and later using various non‐rebreathing systems) or cyclopropane. In most early textbooks, cyclopropane was the recommended drug for pediatric anesthesia despite the fact that it was both explosive and difficult to use. CO₂ absorption was necessary with cyclopropane in order to avoid hypercarbia and acidosis, which might precipitate ventricular arrhythmias. The use of a Waters’ absorber, however, could be technically difficult, especially as tracheal intubation was considered to be dangerous to the child’s “small, delicate airway.”

In all of the early reports, it is noted or implied that the patients were awake (more or less) and extubated at the end of the operation. In the description of the postoperative course, respiratory complications were frequent, in the form of either respiratory insufficiency or airway obstruction. The former problem was likely related to the morbidity of early bypass technology on the lung, while the latter problem was probably because “the largest tube which would fit through the larynx” was often used and/or that the red rubber tube was not tissue tested.

Arthur S. Keats, who worked with Denton A Cooley at the Texas Heart Institute and Texas Children’s Hospital starting in 1955, had significant experience with congenital heart surgery and anesthesia and provided the most extensive description of the anesthetic techniques used in this era. He described anesthesia for congenital heart surgery without bypass in 150 patients, the most common operations being PDA ligation, Potts’ operation, atrial septectomy (Blalock–Hanlon operation), and pulmonary valvotomy [12]. Patients were administered a generous premedication of pentobarbital or chloral hydrate along with meperidine or atropine. Endotracheal intubation was utilized, and ventilation was assisted using an Ayres T‐piece, to‐and‐fro absorption system, or a circle system. Cyclopropane was used for induction, and a venous cutdown provided vascular access. A succinylcholine bolus and infusion were used to maintain muscle relaxation. Light ether anesthesia was used for maintenance until the start of chest closure, at which point 50% N₂O was used as needed. Electrocardiogram (ECG), ear oximeter, and intra‐arterial blood pressure (IABP) recordings were used for monitoring during this period, as were arterial blood gases and measurements of electrolytes and hemoglobin. The following year, he published his experiences with 200 patients undergoing surgery for CHD with CPB, almost all of whom were children [13]. Ventricular septal defect (VSD), atrial septal defect (ASD), TOF, and aortic stenosis were the most common indications for surgery. The anesthetic techniques were the same as described earlier, except that D‐tubocurare was given to maintain apnea during the bypass.

Perfusion rates of 40–50 mL/kg/min were used in infants and children, and lactic acidemia after bypass (average 4 mmol/L) was described. No anesthetic agent was added during the bypass procedure, and “patients tended to awaken during the period of bypass,” but apparently without recall or awareness. Arrhythmias noted ranged from frequent bradycardia with cyclopropane and succinylcholine to junctional or ventricular tachycardia, ventricular fibrillation (VF), heart block, and rapid atrial arrhythmias. Treatments included defibrillation, procainamide, digitalis, phenylephrine, ephedrine, isoproterenol, and atropine. Eleven out of 102 patients with VSD experienced atrioventricular block. Epicardial pacing was attempted in some of these patients but was never successful. Fresh citrated whole blood was used for small children throughout the case, and the transfusion of large amounts of blood was frequently necessary in small infants. The mortality rate was 13% in the first series (36% in the 42 patients less than 1‐year‐old) and 22.5% in the second series (47.5% in the 40 patients less than 1‐year‐old). Causes of death included low cardiac output after ventriculotomy, irreversible VF, coronary air emboli, postoperative atrioventricular block, hemorrhage, pulmonary hypertension, diffuse atelectasis, and aspiration of vomitus. No death was attributed to the anesthetic alone. Reading these reports provides an appreciation of the daunting task of providing anesthesia during these pioneering times. In 1957, in addition to ECG, IABP, and oximeter, Dr. Digby Leigh noted the importance of capnography in cardiac surgery. He described the effect of pulmonary blood flow on end‐tidal CO₂ (EtCO₂) and the decrease in EtCO₂ after partial clamping of the PA during the Blalock–Taussig shunt procedure. However, it was not until 1995 that Smolinsky et al. reported the importance of EtCO₂ during PA banding [14–16].

Tracheostomy after cardiac operations was not unusual, and in some centers, it was done “prophylactically” a week before the scheduled operation. These practices were certainly related to primitive (relative to the present) techniques and equipment used for both endotracheal intubation and CPB. Postoperative ventilatory support did not become routine until later when neonatologists and other intensive care specialists had proven that it could be done successfully. Successful management of prolonged respiratory support was first demonstrated in the great poliomyelitis epidemics in Europe and the United States in 1952–1954 [17].

Halothane was introduced into clinical practice in the mid‐1950s and rapidly became the most popular agent in pediatric anesthesia as it allowed for a smoother induction. Halothane was also widely used for pediatric cardiac anesthesia despite its depressive effect on the myocardium and its significant risk of arrhythmia. During this period, adult cardiac anesthesiologists, following the practice reported by Edward Lowenstein in 1970 [18], began to use opiate‐based intravenous anesthesia. Initially, morphine in doses up to 1 mg/kg was given with 100% oxygen and this technique became the anesthetic of choice for adult cardiac patients, but vasodilation and hypotension associated with its use slowed the incorporation of this technique in pediatric cardiac anesthesia until synthetic opiates became available.

Both prior to the development of CPB and during its early implementation, a number of modalities were utilized to improve outcomes in infants. Inflow occlusion (IO) was an elegant technique that required close cooperation between the entire operating room team, but most especially between the surgeon and the anesthesiologist. The technique was as follows: The chest was opened in the midline. After pericardiotomy, a side clamp was placed on the right atrial (RA) free wall and an incision made in the RA, or proximal on the PA, prior to placing the vascular clamps used to occlude caval return. Before application of the clamps, patients were hyperventilated with 100% O₂. During IO, the superior vena cava (SVC) and inferior vena cava (IVC) inflow were occluded, ventilation held, and the RA or PA clamp released. The heart was allowed to empty and the septum primum was excised or the pulmonic valve dilated. After excision of the septum or valvotomy, one caval clamp was released to de‐air the atrium. The RA side clamp or the PA clamp was then reapplied and the other caval clamp was released. The heart was resuscitated with bolus calcium gluconate (range 30–150 mg/kg) and bicarbonate (range 0.3–3 mEq/kg). Occasionally, inotropes were administered, most often dopamine. It was important to titrate the inotropes so as not to aggravate rebound hypertension caused by endogenous catecholamines. The duration of the IO was between 1 and 3 minutes – terrifying minutes for the anesthesiologist, but quickly over.

Another modality used to try to improve the survival after shunt operations, PA banding, and atrial septectomy was to operate in the hyperbaric chamber, thereby benefiting from the increased amount of physically dissolved oxygen [19] (Figure 1.1). It was a cumbersome affair operating in crowded, closed quarters. There was only room for two surgeons, two nurses, one anesthesiologist, and one baby, as the number of emergency oxygen units limited access. Retired navy divers ran the chamber and kept track of how many minutes the personnel had been in the hyperbaric chamber in the previous week. Help was not readily available because the chamber was buried in a sub‐basement and people had to be sluiced in through a side arm that could be pressurized. The chamber was pressurized to 2–3 atmospheres so it was unpleasantly hot while increasing the O₂ pressure and cold while decreasing the pressure. People with glasses were at a disadvantage. Providing anesthesia was also a challenge. Infants were anesthetized with ketamine and nitrous oxide. As the pressure in the chamber increased, the concentrations of N₂O had to be decreased to avoid hypotension and bradycardia. The hyperbaric chamber did not seem to improve survival and was abandoned around 1974.

It was also during this era that the first infant cardiac transplant was performed by Kantrowitz in 1967 [20]. The recipient was an 18‐day‐old, 2.6 kg patient with severe Ebstein’s anomaly, who had undergone a Potts’ shunt on day 3 of life. The donor was an anencephalic newborn. The anesthetic technique is not described, and the infant died of pulmonary dysfunction 7 hours postoperatively.