ARRHYTHMIAS and conduction abnormalities in heart transplant recipients can be signs of acute and chronic rejection, as well as of coronary artery disease.1,2Here, we report a 13-month-old, 7.5-kg heart transplant recipient with recent onset of tachypnea, decreased peripheral perfusion, and nonsustained ventricular tachycardia who developed cardiac arrest following administration of neostigmine and glycopyrrolate after undergoing an uneventful endomyocardial biopsy during general endotracheal anesthesia to exclude rejection.
The patient was born with anomalous origin of the left coronary artery from the pulmonary artery and underwent a modified Takeuchi procedure (intrapulmonary artery baffle with autologous pericardial patch) at age 2 months. She subsequently developed ischemic cardiomyopathy secondary to occlusion of the left coronary artery repair, necessitating orthotopic heart transplantation at age 1 yr. The transplanted heart exhibited normal left ventricular systolic function, mild left ventricular hypertrophy, and mild restrictive physiology as evaluated by serial echocardiograms. Cardiac biopsies were negative for rejection; however, myocellular edema was present initially, consistent with postpreservation ischemia–reperfusion injury of the allograft.
A week before the events that are the subject of this report and while appearing well, the patient underwent uneventful placement of a right subclavian Broviac catheter during general endotracheal anesthesia. This anesthetic consisted of 50% N2O, 50% O2, fentanyl (3 μg/kg), ketamine (3 mg/kg), cisatracurium (0.5 mg/kg), and isoflurane (up to 0.5% end-tidal concentrations). She was in sinus rhythm, and her heart rate varied between 110 and 140 beats/min. Residual neuromuscular blockade was reversed with neostigmine (0.08 mg/kg) and glycopyrrolate (0.016 mg/kg) infused into a rapidly running peripheral intravenous catheter over 1 min; during this interval, the heart rate increased from 110 to 160 beats/min. The trachea was extubated without any complications.
Approximately 1 week later, she developed nonsustained ventricular tachycardia. An endomyocardial biopsy was therefore scheduled to assess for rejection. An echocardiogram performed the day before the biopsy showed qualitatively good biventricular function. Serum electrolytes, calcium, and magnesium concentrations were all within their respective normal ranges. However, before the procedure, the patient was noted to be tachypneic, with evidence of decreased peripheral perfusion. Her vital signs were as follows: heart rate, 135 beats/min; respiratory rate, 38 breaths/min; and blood pressure, 90/60 mmHg. After placing standard noninvasive monitors, anesthesia was induced with etomidate (0.2 mg/kg), fentanyl (2 μg/kg), and cisatracurium (0.2 mg/kg) and maintained with isoflurane (up to 0.5% end-tidal concentration). The average blood pressure was 90/50 mmHg. She was in sinus rhythm with a heart rate between 100 and 120 beats/min. Transvenous endomyocardial biopsies were obtained without complications. The capillary wedge pressure was 15 mmHg. She was weaned from anesthesia, and neuromuscular blockade was reversed with neostigmine (0.07 mg/kg) and glycopyrrolate (0.014 mg/kg) infused into a rapidly running peripheral intravenous catheter over 1 min. These were mixed together and administered in the same syringe. Subsequent event debriefing and case reviews yielded no evidence of drug administration errors, including incorrect doses or incorrect drugs. Her heart rhythm progressed from sinus bradycardia to asystole within 2–3 min, accompanied by circulatory collapse, which was unresponsive to cardiopulmonary resuscitation (which included external chest compressions, repeated epinephrine boluses, calcium gluconate, and attempts at direct current cardioversion) and transvenous ventricular pacing. Ice was placed around the head to induce hypothermia, and fentanyl (25 μg/kg), pancuronium (0.2 mg/kg), and midazolam (0.2 mg/kg) were administered. Cardiopulmonary resuscitation was continued while the patient was placed on extracorporeal membrane oxygenation in the cardiac catheterization laboratory.3Time from onset of cardiac arrest to full flow extracorporeal membrane oxygenation was 30 min; a percutaneous 6-French long sheath was positioned across the atrial septum to decompress the left ventricle. After achieving full extracorporeal membrane oxygenation flow and with improved coronary perfusion, ventricular function recovered. The patient was transferred to the cardiac intensive care unit with a blood pressure of 100/70 mmHg and a heart rate of 45–65 beats/min.
A presumptive diagnosis of humoral rejection was made; thus steroids and intravenous gammaglobulin were administered and plasmapheresis was performed. In the cardiac intensive care unit, the patient's heart rate was 120 beats/min, and the rhythm was sinus. The arterial blood pressure was 150/84 mmHg. The patient was started on milrinone (1 μg · kg−1· min−1) and sodium nitroprusside (5 μg · kg−1· min−1). She was sedated with fentanyl and midazolam. The initial serum lactate level was 5.8 mm and decreased to 1.4 over the next 7 h. The biopsy showed grade 0 cellular rejection (i.e. , no cellular rejection; table 1) and changes consistent with humoral rejection, including coronary capillary endothelial swelling and disruption, focal areas of microvascular occlusion by thrombus and leukocytes, and interstitial edema. Serial cardiac ultrasounds demonstrated improvement in ventricular function. Milrinone and sodium nitroprusside were weaned off, and extracorporeal membrane oxygenation was discontinued 3.5 days afterward. No gross neurologic injury was documented, and the trachea was extubated 8 days after the arrest. The patient was discharged to the cardiac ward in satisfactory condition 2 weeks later.
The patient returned for cardiac biopsy 1 month after these events and appeared well. Cardiac ultrasound showed normal left ventricular systolic function. The biopsy was performed via the right femoral vein because of jugular venous obstruction. This procedure was performed without complications during intravenous sedation with ketamine (1 mg/kg) and morphine (0.1 mg/kg) and caudal anesthesia (1 ml/kg of 0.25% plain bupivacaine). The capillary wedge pressure was 8 mmHg. Histologic evidence of rejection was not present.
In this infant, ventricular tachycardia, clinical deterioration, and elevated cardiac filling pressures (roughly twice her previous and subsequent values) suggested the possibility of rejection; features of humoral rejection were present on endomyocardial biopsy. The development of asystole after the administration of neostigmine and glycopyrrolate in this patient was an unexpected event because the allograft had been in place for only 1 month, which likely excludes the possibility of parasympathetic reinnervation. It is difficult to discern whether this event was related to the ongoing rejection process, the use of neostigmine, or the combination.
Clinical signs and symptoms of rejection are variable, nonspecific (e.g. , tachycardia, malaise, fever), and frequently subtle. As demonstrated at autopsy, the conduction system, as well as the sinus and atrioventricular nodes, can be targets for rejection in heart transplant recipients. Sinus node dysfunction can be a sign of acute or chronic rejection in this population. Sinus arrest has occurred in heart transplant patients in the setting of rejection necessitating artificial pacemaker therapy.4Other rhythm abnormalities, such as atrial flutter and fibrillation, conduction block, and ventricular rhythm disturbances, can also occur.
It has been shown that neostigmine reduces the heart rate in adult heart transplant recipients without evidence of rejection and that the magnitude of reduction is less when compared with patients with native hearts. In heart transplant patients, slowing of the heart rate after neostigmine was more pronounced after 6 months post transplantation than in the first 6 months, indicating some degree of parasympathetic reinnervation in the donor heart.5However, when the heart transplant patients were given atropine, the heart rate increase in response to atropine was similar and slower than in the patients with native hearts, suggesting limited parasympathetic reinnervation of the transplanted heart. Sinus arrest and asystole have been noted after reversal of neuromuscular blockade with neostigmine and glycopyrrolate in adult heart transplant recipients, with no clinical evidence of rejection years after transplantation.6,7We postulate that our patient's conduction system was involved during the rejection episode. She had developed nonsustained ventricular tachycardia before the myocardial biopsy. The heart rhythm and function recovered after plasmapheresis, steroids, and immunosuppressive therapy.
Currently, at least two broad types of rejection episodes are recognized after transplantation: cellular and humoral. The former is characterized by several features that include primarily activated lymphocytic infiltration with a resultant local inflammatory process and the development of myocyte necrosis (table 1). Current immunosuppression regimens for the most part are targeted to the T-cell signaling pathways underlying cellular rejection. However, cellular rejection may account for a significantly lower number of rejection episodes accompanied by cardiac functional deterioration than previously thought.
Humoral rejection is caused by a T-cell response mediated by alloantibodies that are mainly directed against human leukocyte antigen class I and II molecules. A major risk factor is likely to be increased antigen exposure and allosensitization from events such as previous surgeries, use of homograft material to repair congenital heart defects, multiple blood product exposures, and previous pregnancy8; all but the last of these factors were present in this patient. The result is an inflammatory process that is primarily mediated by alloantibodies and activated complement; it is characterized by capillary endothelial swelling and damage, intravascular coagulation and macrophage accumulation, interstitial edema and hemorrhage, pericapillary neutrophil infiltration, and finally focal ischemia. More advanced immunohistologic and immunofluorescence studies can reveal immunoglobulin (IgA, IgM, and/or IgG) and complement (C4d, C3d, or C1q) deposition on capillaries and CD68 staining of intracapillary macrophages. There is accumulating evidence that humoral rejection is a significant cause of “biopsy-negative” rejection episodes that are potentially significant causes of acute or subacute contractile dysfunction and graft failure, as well as overall mortality.8–10Simultaneous histologic evidence for both cellular and humoral rejection can be found in some patients.
Contrary to the adult heart transplant experience with longer-standing allografts developing bradycardia and asystole after neostigmine in the absence of rejection, our patient's young allograft developed asystole after administration of neostigmine and glycopyrrolate in the setting of probable humoral rejection. We felt comfortable reversing the neuromuscular blockade in this patient given the relatively brief interval since her transplant (i.e. , small likelihood of reinnervation) and because no bradycardia or asystole developed after the use of neostigmine and glycopyrrolate when she underwent anesthesia for the Broviac catheter placement a week earlier. We speculate that she developed the cardiac manifestations of humoral rejection over this period; clinically, she did manifest evidence of myocardial dysfunction such as tachypnea and decreased perfusion (although a contemporaneous echocardiogram showed “normal” systolic function). It is also tempting to speculate that the patient's ventricular tachycardia and enhanced sensitivity to cholinesterase inhibition were due to alloantibodies reactive against epitopes on cardiac conducting tissue; it is equally possible that these events were due to local inflammation and ischemia.
If correct, this sequence of events also highlights some of the perioperative and anesthetic management challenges posed by these patients. The clinical signs of rejection can be nonspecific and relatively insensitive. Echocardiograms are suboptimal for detecting rejection, particularly during the initial weeks and months after transplantation, in part because the technique is confounded by changes that occur as a consequence of “normal” recovery from ischemia–reperfusion injury (e.g. , increased myocardial mass and edema) and a relative insensitivity to diastolic dysfunction (at least using standard clinical examination techniques); furthermore, it has been our impression that echocardiographic evidence of systolic dysfunction due to rejection can lag behind the parenchymal rejection process. Other techniques to detect rejection, including use of expression microarrays and magnetic resonance methods, remain in various stages of development. As a result, endomyocardial biopsy remains the accepted standard in most centers for surveillance and detection of rejection. It is noteworthy that cellular rejection is typically quite heterogeneous, and thus biopsy (which typically obtains four to seven specimens from right ventricular endocardium only) is potentially a hit-or-miss proposition that can have a false-negative rate of between approximately 20% and 60%. It is also worth noting that a substantial number of pediatric transplant recipients require deep sedation or general anesthesia to tolerate the procedure (and hence obtain a diagnosis) successfully, and thus pediatric anesthesiologists are routinely confronted with providing this care without knowing the rejection or true functional status of the patient. Overall, this case points out the need to maintain a high index of suspicion for the presence of rejection—both cellular and humoral— and its manifestations in cardiac transplant recipients, regardless of their clinical findings. With specific reference to reversal of neuromuscular blockade, it raises the question of avoiding the use of muscle relaxation in instances where antagonism of the blockade will be necessary. Alternatively, using short-acting agents that permit adequate recovery of neuromuscular junction function without use of cholinesterase inhibiting agents or, perhaps in the near future, use of specific combinations of neuromuscular blocking-reversal agents that do not share this side effect profile (e.g. , rocuronium–sugammadex), should be considered.