Accidental systemic administration of 1 litre of cardioplegia solution during paediatric cardiac surgery

Summary

Cardioplegia is used to induce cardiac arrest in order to facilitate cardiac surgery in patients supported by cardiopulmonary bypass. It is administered directly into the coronary vessels after the heart has been isolated from the systemic circulation. We describe the case of a 9‐year‐old boy who mistakenly received 1 l of high strength St Thomas’ Harefield cardioplegia solution delivered into the systemic circulation during cardiac surgery. Although the patient’s heart did not stop, the subsequent physiological derangements were severe. The presenting features were refractory hypotension and dilutional anaemia along with severe hyperkalaemia, hypermagnesaemia and hyperchloraemic metabolic acidosis. Local anaesthetic systemic toxicity from the procaine contained within the cardioplegia solution was also a concern. Treatment required vasopressor administration and an extended period of cardiopulmonary bypass while serum electrolyte concentrations were corrected by haemodiafiltration. The systemic administration of cardioplegia solution is a rare but important iatrogenic clinical emergency that anaesthetists working in cardiac centres should be aware of. This case demonstrates that full recovery is possible.

Introduction

Cardioplegia is necessary for cardiac surgery which requires a stationary heart. It is administered directly into the coronary vessels after the heart has been isolated from the systemic circulation with an aortic cross clamp. Our institution uses St Thomas’ Harefield formulation, the most commonly used cardioplegia solution for paediatric cardiac surgery in the UK [1]. It is mixed with cold blood at a ratio of 4:1 (four parts blood to one part cardioplegia) for induction of cardiac arrest, and 8:1 for maintenance. Typical doses of the combined blood‐cardioplegia solution are 30 ml.kg^‐1 for induction and 15–20 ml.kg^‐1 for maintenance [1]. In this report, we describe a case in which 1 l of undiluted high‐strength St Thomas’ Harefield cardioplegia solution, amounting to 45 ml.kg^‐1, was mistakenly administered into the systemic circulation of a 9‐year‐old boy. We are unable to find any similar cases, paediatric or adult, reported in the literature, or any published guidance for the detection or management of this clinical emergency.

Report

A 9‐year‐old boy (22.2 kg; body surface area 0.92 m²) was scheduled to undergo re‐sternotomy for removal of a HeartWare (Medtronic, Minnesota, USA) left ventricular assist device. The device had been implanted six years previously for decompensated heart failure arising from lymphocytic myocarditis. Although initially implanted as a bridge to transplantation, the device was decommissioned after five months in the context of sufficient left ventricular recovery. Echocardiography showed mildly impaired biventricular function with trivial mitral regurgitation. The patient had normal renal function with mild anaemia pre‐operatively, and had no pre‐existing arrhythmias.

After premedication with oral midazolam (0.5 mg.kg^‐1), inhalational induction of general anaesthesia was performed with sevoflurane in 100% oxygen. Intravenous (i.v.) access was then obtained, fentanyl (4 µg.kg^‐1) and rocuronium (1 mg.kg^‐1) were administered and the patient’s trachea was intubated. Anaesthesia was maintained with isoflurane in oxygen and air. Arterial and central venous cannulae were then inserted. Dissection of the chest was performed with redo sternotomy precautions. A blood prime was used for the cardiopulmonary bypass machine. Shortly after commencing cardiopulmonary bypass at 32°C, new and profound arterial hypotension developed with a mean arterial pressure of 30–40 mmHg. This hypotension persisted for 60 min and was poorly responsive to large and repeated doses of metaraminol (total 15.75 mg) and phenylephrine (total 1.75 mg), infusions of adrenaline (0.07 μg.kg^‐1.min^‐1) and vasopressin (1.5 milliunits.kg^‐1.min^‐1), a stress dose of hydrocortisone (2 mg.kg^‐1), and an increase in cardiopulmonary bypass flow rate. Sinus rhythm continued with an unchanged electrocardiograph pattern at a rate of 80–90 beats per min. Although processed electroencephalogram and bilateral cerebral oximetry readings did not significantly change, the patient’s pupils were noted to be fixed and dilated. The initial blood gas on cardiopulmonary bypass revealed extreme metabolic acidosis and hyperkalaemia (Table 1).

Table 1.

Arterial blood gas (ABG) results over the intra‐operative course.

	First ABG under general anaesthesia before cardiopulmonary bypass	ABG from cardiopulmonary bypass circuit prime	First ABG on cardiopulmonary bypass	Final intra‐operative ABG at chest closure
pH a	7.45	7.26	7.12	7.35
Partial pressure of carbon dioxide; kPa a	3.6	4.6	5.1	5.7
Partial pressure of oxygen; kPa a	51.1	32.1	26.5	8.4
Sodium; mmol.l‐1	133	143	132	143
Potassium; mmol.l‐1	4.3	10.5	12.9	5.0
Chloride; mmol.l‐1	103	107	133	114
Calcium; mmol.l‐1	1.23	1.00	1.42	1.48
Glucose; mmol.l‐1	5.4	7.3	5.7	7.6
Lactate; mmol.l‐1	1.3	4.5	1.0	3.1
Base excess; mmol.l‐1	‐3.8	‐10.8	‐16.2	‐2.0
Bicarbonate; mmol.l‐1	22.0	16.5	12.3	22.7
Haematocrit; %	34	20	22	23

^a pH‐stat.

After discussion with the perfusionist, it became apparent that the full 1‐l bag of high strength St Thomas’ Harefield cardioplegia solution (Terumo BCT, Larne, UK)—containing sodium 147 mmol; potassium 84 mmol; magnesium 80 mmol; calcium 2 mmol; chloride 400 mmol; and procaine 1364 mg—had accidentally entered the cardiopulmonary bypass reservoir and been delivered to the patient’s systemic circulation via the aortic cannula. While priming the cardioplegia heater‐cooler device with blood, the perfusionist’s attention had been diverted to a monitor failure on the venous‐assist setup; the cardioplegia pumps had been left running and the contents of the cardioplegia bag had gradually emptied into the bypass reservoir (Fig. 1). The delivery of reservoir blood containing cardioplegia to the patient’s systemic circulation occurred after commencing cardiopulmonary bypass but before the aortic cross clamp had been applied. Once the error was realised, efforts were made to treat the adverse effects of the components of the cardioplegia solution. The overdose of procaine was treated with intralipid; serum magnesium exceeded 4 mmol.l^‐1, the highest recordable value in our laboratory, and serum potassium peaked at 13.8 mmol.l^‐1. This was treated with furosemide, sodium bicarbonate and haemodiafiltration via the cardiopulmonary bypass pump, requiring more than two additional hours of cardiopulmonary bypass time.

Figure 1.

Schematic diagram of the cardioplegia delivery system. When the cardioplegia roller pumps are turned on and the three‐way tap is open between the recirculation line and the reservoir, cardioplegia solution can enter the reservoir.

Once stabilised, a further small dose of blood cardioplegia was given through the aortic root to facilitate completion of surgery. Twenty‐seven millilitres of cardioplegia (mixed with 232 ml of cold blood) is all that was required to arrest the patient's heart. No further cardioplegia was given.

The patient made a full recovery, unhindered by this intra‐operative misadventure.

Discussion

Cardioplegia solution is usually only administered through the cardiopulmonary bypass machine. For this reason, any error in its administration is most likely to occur in a patient already supported by cardiopulmonary bypass. This is inherently protective because, under these circumstances, perfusion is not dependent on cardiac output. However, the situation can still be grave.

In this case, severe and resistant hypotension was the initial presenting sign. Although treatment began immediately, because hypotension is a non‐specific sign, it took 19 min to recognise its cause. Significant dilutional anaemia also occurred (Table 1) as the intended blood priming volume was effectively doubled by the addition of the cardioplegia solution. Interestingly, there was no rise in lactate during cardiopulmonary bypass despite the low mean arterial pressure and considerable vasoactive support. The fact that our patient was an older, acyanotic child undergoing an elective procedure may have been protective in this respect.

Hyperkalaemia was the most marked presenting feature in this case. The patient received 3.8 mmol.kg^‐1 of potassium chloride, causing the serum potassium concentration to rise to 12.9 mmol.l^‐1 (Table 1). This was noted on the first arterial blood gas following initiation of cardiopulmonary bypass. It was surprising that the heart continued to beat, and did so without obvious electrocardiographic changes. It is possible that the gradual addition of the cardioplegia solution to the reservoir, and dilution with the patient’s circulating volume before reaching the coronary circulation may have mitigated its potency. Haemodiafiltration through the cardiopulmonary bypass circuit and diuresis with furosemide 1 mg.kg^‐1 i.v. were used to eliminate excess potassium. This necessitated a further 151 min of cardiopulmonary bypass after the cross clamp was released. We elected to separate from bypass once the serum potassium was less than 5.5 mmol.l^‐1.

The patient received 3.6 mmol.kg^‐1 magnesium chloride resulting in a serum magnesium level in excess of what our laboratory could measure (>4 mmol.l^‐1). Hypermagnesaemia‐associated vasodilatation likely contributed to the patient’s hypotension. Aside from haemodiafiltration and furosemide, no specific therapy was administered (calcium is already contained within the cardioplegia solution). By the time of postoperative admission to the paediatric intensive care unit, serum magnesium concentrations had fallen to 1.65 mmol.l^‐1 and no symptoms or signs of hypermagnesaemia were identified. However, these would have been difficult to detect as the patient remained sedated to facilitate the continuation of tracheal intubation until postoperative day one when the serum magnesium level had normalised.

The patient developed severe metabolic acidosis (anion gap −0.4 mmol.l^‐1) (Table 1) as a consequence of receiving 400 mmol (18 mmol.kg^‐1) of chloride. This caused an acute rise in serum chloride concentration to 133 mmol.l^‐1, a value higher than the serum sodium. A total of 160 ml (7.2 ml.kg^‐1) of 8.4% bicarbonate solution (in divided doses) was required to correct the pH. This caused an acute rise in serum sodium concentrations to a peak of 150 mmol.l^‐1, and an increase in serum osmolality to 310 mOsml.kg^‐1.

Procaine is an ester local anaesthetic. It has myocardial protective effects, helping to induce cardiac arrest and reduce reperfusion arrhythmias. This makes it a useful additive to cardioplegia but problematic in the context of toxicity. Procaine has a maximum recommended dose of 12 mg.kg^‐1 and a median lethal intravenous dose in mice of 46–80 mg.kg^‐1 [2, 3]. Our patient received a systemic dose of 61 mg.kg^‐1. Shortly thereafter he was noted to have fixed and dilated pupils, raising suspicion of central nervous system local anaesthetic toxicity. Seizure activity was a particular concern and would be difficult to detect under the influence of neuromuscular blocking drugs. An anti‐convulsant (midazolam 0.2 mg.kg^‐1) and a loading dose of 20% intralipid (1.5 ml.kg^‐1) were administered i.v. Further intralipid was advised against by the toxicology team on the basis of its potential to agglutinate in extracorporeal circuits and cause mechanical failure of oxygenators. This problem has been reported in patients supported by extracorporeal membrane oxygenation [4], but its risk in the more porous cardiopulmonary bypass oxygenators is less clear. Procaine ordinarily has a half‐life of 6 min and a clinical duration up to 60 min [3]. However, it is likely that its metabolism was prolonged by hypothermia in this case. Vasodilation is a feature of procaine cardiovascular toxicity and may have contributed to our patient’s vasoplegic state. Pupil reactivity recovered by the end of the case and there were no postoperative neurological concerns.

This incident did not affect the surgery itself. The patient’s heart began beating spontaneously once potassium levels had normalised and remained in sinus rhythm. Myocardial contractility was seemingly unaffected. Moderate bleeding occurred before chest closure. This may have been exacerbated by coagulopathy associated with prolonged cardiopulmonary bypass but is unlikely to have been directly related to the cardioplegia itself.

The human factors which culminated in this incident deserve careful consideration to prevent future similar occurrences. In the cognitively demanding and often busy environment of the cardiac operating theatre, safety is most likely to be enhanced when solutions focus on systems change rather than efforts to improve individual staff vigilance. We have implemented two systems changes in response to this case: Firstly, we no longer recirculate any cardioplegia solutions into the bypass reservoir. The recirculation line is required to prime the cardioplegia delivery system when preparing the circuit, however, the three‐way tap to the reservoir is now turned off before the patient is connected to the circuit. The tap is then left off to the reservoir for the duration of the case (Figure 1). The final priming of the blood cardioplegia is performed into a sterile container within the surgical field. Secondly, a calibrated in‐line blood parameter monitoring system such as the CDI® 550 (Terumo, Ann Arbor, Michigan, USA), could have enabled earlier detection of the dilutional anaemia and rising potassium levels in this case before the full litre of cardioplegia solution had been delivered. Although we have routinely used in‐line blood parameter monitoring at our institution for some time, we now calibrate it early with both the manufacturer’s gaseous calibration solutions and with a blood gas drawn as soon as possible after commencing cardiopulmonary bypass. We would further recommend that anaesthetists and perfusionists seek an early second opinion when the patient’s course deviates significantly from what it expected.

Systemic administration of cardioplegia solution due to human error is a rare but important clinical emergency that anaesthetists working in cardiac centres should be aware of. This case, involving high strength St Thomas’ Harefield cardioplegia solution, presented with refractory hypotension, unexpected anaemia, severe metabolic acidosis, hyperkalaemia and hypermagnesaemia, and most probably an element of local anaesthetic systemic toxicity. This condition is treatable with mechanical support and haemodiafiltration and a full recovery is possible.