Clinical characteristics of massive air embolism complicating left atrial ablation of atrial fibrillation: lessons from five cases

Abstract

Aims

This study aimed to elucidate the clinical characteristics of massive air embolism occurring during atrial fibrillation (AF) ablation.

Methods and results

Of 2976 patients undergoing AF ablation, 5 patients complicated by serious air embolism were examined. Atrial fibrillation ablation was performed with the use of three long sheaths for circular mapping and ablation catheters under conscious sedation. Two patients had air spontaneously introduced through a haemostasis valve of the long sheaths, at the end of long apnoea caused by the sedation, even though the catheters were placed within the long sheaths. The remaining three patients, all of whom also exhibited long apnoea, had air entry at the circular mapping catheter exchanges. Air accumulated in the right and left ventricles, left atrial appendage, right coronary artery, and ascending aorta. Haemodynamic collapse and hypoxaemia occurred in all and two patients, respectively, and supportive treatment and the accumulated air were aspirated. ST elevation, haemodynamic collapse, and hypoxaemia persisted for 10–35 min; however, all patients recovered completely. After we changed the sedative to one with less respiratory depressive effects and the timing of the saline flush at the circular mapping catheter exchanges, we never experienced such serious complications any further.

Conclusion

Serious air embolism can occur in patients with long apnoea under sedation during AF ablation with the use of long sheaths. Supportive therapy and air aspiration were effective in resolving the complication. A sedative that causes less respiratory depression and the timing of the saline flush were important for preventing air embolism.

Introduction

An air embolism complicating catheter ablation is an uncommon but potentially catastrophic event that occurs as a consequence of the entry of air into the vasculature.^1–4 In particular, a cerebral air embolism can cause severe neurological complications that require aggressive resuscitative efforts or hyperbaric oxygen therapy.^1,2 Catheter ablation therapy of atrial fibrillation (AF) requires an access to the left atrium (LA), and most catheter manipulation and exchanges are done within the LA. Hence, air embolism occurring during these procedures may cause serious damage to the heart, lungs, or brain. Although careful attention should be paid to avoid this complication, no previous reports have yet systematically addressed this problem. The aim of this study was to elucidate the clinical characteristics of serious air embolism complicating AF ablation.

Methods

Study population

This study included a total of 2976 consecutive patients with drug-refractory AF (mean age 61 ± 10 years, 2372 males, 1833 paroxysmal AF) who underwent AF catheter ablation in our laboratory from June 2003 to April 2011. A total of 3870 AF ablation sessions performed in 2976 patients were analysed in this study. All patients gave their written informed consent before the catheter ablation.

All antiarrhythmic drugs were discontinued 7 days before the ablation session. All patients were effectively anticoagulated for >1 month, and transoesophageal echocardiography was performed to exclude any atrial thrombi before the ablation.

From June 2003 to July 2009, warfarin was discontinued 3–5 days before the AF ablation session, and heparin bridging was performed during the periprocedural period until the international ratio returned to 2–3 after the resumption of warfarin. After August 2009, warfarin was continued throughout the periprocedural period.

Catheter ablation of atrial fibrillation

The approach of the pulmonary vein isolation (PVI) performed in our institute has been previously described in detail.^5,6 Briefly, after a single transseptal puncture, three long 8 Fr sheaths (SL0, AF Division, St Jude Medical) were introduced into the LA. Two 5 Fr circular mapping catheters (Lasso, Biosense Webster) via the two long sheaths were placed in the superior and inferior pulmonary veins (PVs), respectively, and the left- and right-sided ipsilateral PVs were circumferentially and extensively ablated during an intravenous isoproterenol infusion (1 μg/min) under fluoroscopic and electrophysiologic guidance. Radiofrequency (RF) current deliveries were applied with an 8 mm tip non-irrigated ablation catheter (Japan Lifeline, Inc.) or 3.5 mm irrigated tip ablation catheter (Biosense Webster, Inc.) through another long sheath. Each of the three long sheaths was not managed with a continuous flushing of heparinized saline, and whether the tip of the long sheath was withdrawn to the right atrium or not was left to the discretion of each operator. The endpoint of the ablation was the elimination of all PV potentials.

After completing the PVI, the cavotricuspid isthmus (CTI) was also ablated to create bi-directional conduction block. If frequent atrial premature contractions were present, focal ablations were performed.

In addition to the PVI and CTI ablation, the patients with persistent AF underwent linear ablations (LA roof and inferior line and mitral isthmus line) dependent on the AF inducibility.

Sedation

All patients underwent catheter ablation under conscious sedation. For the first 420 consecutive patients, propofol was used, and the last 2556 patients were given dexmedetomidine. Propofol was administered intravenously, with a slow injection of 1.0 mg/kg, for the induction of sedation, and maintained, at 1.5–4.5 mg/kg/h during the procedure. Lower induction and maintenance doses were used for the elderly. The blood pressure, heart rate, respiratory condition, and arterial oxygen saturation measured by a pulse oximeter were carefully monitored.

Dexmedetomidine was started, at 1.5 μg/kg/h for the first 20 min, for the induction of sedation, and maintained, at 0.2–0.7 μg/kg/h during the procedure. The vital signs were similarly monitored as in the patient's given propofol.

Results

Clinical characteristics of serious air embolism during atrial fibrillation ablation

Serious air embolism occurred in five patients, whose clinical characteristics are shown in Table 1. A serious air embolism was defined as one that was accompanied by haemodynamic collapse persisting for >10 min. A total of 10 cases with transient ST elevation, usually for a few minutes, were excluded from this study, because we did not clarify the cause of the ST elevation as the subsequent coronary angiograms revealed normal coronary arteries. The frequency of this event did not change over the study period.

Two patients (Cases 3 and 4) had air spontaneously introduced through the haemostasis valves of the long sheaths after long apnoea episodes, just before a loud snore, induced by propofol, even though circular mapping and ablation catheters were placed within the long sheaths. A sucking noise of air entering was heard in both patients. The remaining three patients had air enter while exchanging the circular mapping catheters, and also exhibited long apnoea episodes induced by the propofol or dexmedetomidine. An accumulation of air was observed in the right and left ventricles (Cases 3 and 4), left atrial appendage (all patients), right coronary artery (RCA) (all patients), and ascending aorta (Cases 3 and 4).

ST elevation was observed in the inferior leads in all patients and right precordial leads in three. Coronary angiography revealed a slow flow or total occlusion of the RCA due to air bubbles in all patients, whereas no patients had an abnormal left coronary artery angiogram. Forceful injections of contrast medium or saline were made into the RCA, resulting in fragmentation of the large occlusive air bubbles. In one patient (Case 5) active withdrawal of the air was attempted by the use of a small infusion catheter inserted into the RCA. Despite the above treatment, the ST-segment exhibited alternate improvement and deterioration, because the accumulated air intermittently moved from the left atrial appendage to the RCA. Then, we attempted to remove the air from the left atrial appendage.

A haemodynamic collapse was observed in all patients, and then noradrenaline was administered to maintain the blood pressure, and pacing from the high right atrium or right ventricular apex was applied for bradycardia.

The oxygen saturation levels in two cases deteriorated, presumably due to a pulmonary artery air embolism, because they had air accumulation in the right ventricle. Although bag-valve-mask ventilation was temporally performed for their respiratory deterioration, no artificial ventilation was needed.

The haemodynamic collapse and hypoxaemia persisted for 10–35 min; however, ultimately, all patients recovered completely without any sequelae or cerebral air embolism. After their recovery, we re-started the ablation procedure, and successful PVI and CTI bi-directional block were achieved in all patients. Following the procedure, every patient was placed in the supine position for at least 12 h to prevent any new onset of cerebral air embolism due to any possible residual air in the heart. A follow-up echocardiogram revealed a normal wall motion, and no ischaemic changes were observed on the electrocardiogram. None of the patients exhibited any neurological signs over the clinical course.

Differences in the sedation between the patients with propofol and dexmedetomidine

Air embolism occurred in four patients in the propofol group, but in only one (Case 5) in the dexmedetomidine group. In the propofol group, most of the patients often suffered from chest pain during the RF delivery, especially at the lower sites of the left and right inferior PVs or posterior wall of the LA adjacent to the ostium of the left and right inferior PVs. Although we administered pentazocine to relieve the chest pain, some patients presented with restlessness presumably in response to their chest pain. We then had to increase the dose of propofol to the maximum level in order to keep the patients in an unconscious state to continue the procedure. Three of the four patients received the maximum dosage of the sedative when the air embolism occurred.

In the dexmedetomidine group, most of the patients, as with the patients given propofol, complained of chest pain at similar sites, but few patients presented with restlessness. Therefore, we only had to add pentazocine to relieve the chest pain, but did not have to increase the dosage of the dexmedetomidine. A low–medium maintenance dose was administered in most patients.

How to flush the long sheaths with heparinized saline

In order to minimize the chance of sheath-related air embolism or thromboembolism, we carefully flushed the sheaths immediately before inserting the catheters. However, it was possible that air was introduced into the sheath during the insertion of the circular mapping catheter, and so we changed the timing of the flushing to immediately after inserting the catheter before pushing it out into the LA, in June 2007. Since then, we have never experienced any similar serious air emboli.

Discussion

Major finding

Serious air embolism with haemodynamic collapse or hypoxaemia can occur with long apnoea episodes produced by sedation during left atrial catheter ablation of AF. In our series, air accumulated in the right and left ventricles, left atrial appendage, RCA, and ascending aorta. Supportive treatments for haemodynamic shock and respiratory deterioration, and aspiration of the accumulated air yielded a complete recovery without any sequelae. Changing the sedative to the one with less respiratory suppressive effects and the timing of flushing the sheaths contributed to preventing such serious air emboli.

Mechanisms of air embolism

Given that all patients had air embolism during long apnoea spells and snoring induced by the sedation, an increased negative intrathoracic pressure produced by the long apnoea episodes seemed to be associated with the occurrence of the air embolism. With the intrathoracic pressure much lower than the atmospheric pressure, air could be sucked into the vasculature or the sheath through the haemostasis valve. Cases 3 and 4 must have especially had a markedly increased negative intrathoracic pressure, because the air was sucked in even with the catheter placed within the long sheath. The remaining three patients had the air enter while the circular mapping catheter was being exchanged. While the circular mapping catheters were inserted into the sheath, the air seemed to enter the sheath.

A rapid removal of a catheter from a long sheath could create a vacuum, which could draw air in through the haemostasis valve. We have paid careful attention to slowly removing any catheter from the long sheaths during every ablation session. Therefore, the vacuum effect as a cause of an air embolism was less likely in our series.

The reason why air embolism specifically occurred in the RCA, not the left coronary artery seemed to be that the right coronary cusp was positioned at the superior aspect of the heart when the patients were in the supine position. The ST elevation in the right precordial leads was considered to reflect an occlusion of the conus branch of the RCA.

Cases 3 and 4 with sustained hypoxaemia may have had a pulmonary artery air embolism, because air accumulation was observed in the right ventricle. The position of the tip of the long sheath was at the discretion of each operator. Therefore, in our approach air embolism could occur on either the pulmonary or systematic circulation side depending on the location of the tip of the long sheath. If it was drawn into the inferior vena cava, the air could enter the pulmonary circulation.

Sedation for ablation procedures

Propofol is a short-acting intravenous sedative agent used for the induction and maintenance of general anaesthesia, or sedation in procedures such as endoscopy. Propofol is a respiratory depressant, frequently producing apnoea that may persist for >60 s, depending on factors such as the pre-medication, rate of administration, dose administered,⁷ and presence of hyperventilation or hyperoxia. Since it provided no analgesia, our patients often suffered from chest pain, which made us increase the dosage of the propofol. The higher dosages seemed to be associated with long apnoea spells and snoring, leading to the occurrence of the air embolism.

Dexmedetomidine is a sedative with analgesic effects commonly used for sedation of patients in intensive care units. It has no clinically important adverse effects on the respiration,⁸ and produces only mild cognitive impairment. Thus, it is possible for healthcare provider to easily communicate with the patient.⁹ Actually, the patients given that sedative in our study told us about their chest pain during the procedure, and we could stop the RF energy delivery, and administer pentazocine to relieve their chest pain. Despite the lesser respiratory depressive effects of the dexmedetomidine, Case 5 demonstrated a long apnoea episode and subsequent air embolism, which implied the important condition for an air embolism to occur, was the long apnoea spells rather than the kind of sedative used.

Management of air embolism

The management of the air embolism was generally supportive.³ In the case of transient coronary air embolism, which is often seen as transient ST elevation, only observation or increasing the blood pressure with manoeuvres such as atrial pacing or the administration of vasopressor drugs is effective in most patients. However, persistent coronary air embolism involving a haemodynamic collapse may require some coronary intervention. In our series, a forceful injection of contrast medium or saline into the coronary artery may have helped break up the large air bubbles, and the withdrawal of the air by using a small infusion catheter also contributed to the improvement in the coronary artery blood flow.

Patients complicated with cerebral air embolism, which did not occur in our series, could experience neurological impairment. In such cases, aggressive resuscitative efforts and hyperbaric oxygen treatment may be required.^1,2

Previous reports on air emboli

Coronary air embolism during catheter ablation have been reported to occur during catheter exchanges within the LA,^3,4 and they recovered completely without any sequelae as well. On the contrary, as was described above, cerebral air embolism resulting in neurological impairment even occurred during catheter ablation.^1,2 The determinants of whether air introduction leads to a cerebral air embolism or not remains uncertain. It might be related to the volume of the air introduced into the LA, or the location of the three arterial branches from the aortic arch.

Possible measures to prevent air embolism

Most cases with air embolism complicating AF ablation seemed to recover reversibly as in our five cases. Nonetheless, because the air introduction during this procedure potentially results in devastating complications, all possible measures must be taken to avoid the air entry during this procedure.

If a conscious sedation is performed, sedatives with a lesser respiratory depressive effect should be recommended to prevent long apnoea spells. Once the apnoea is recognized, the dosage of the sedative should be decreased or a nasal airway could be effective to reduce the apnoea.

The circular mapping catheters are somewhat difficult to insert into the long sheaths because of the shape of the catheter tip. Air entry into the long sheath is likely to occur during catheter exchanges. In order to avoid pushing the air into the LA, it is suggested to flush the sheath with heparinized saline just after inserting the circular catheter into the sheath before pushing it out into the LA. In actual practice, some air can occasionally be drawn back.

Furthermore, a difference in the diameter between an 8 Fr sheath and a 5 Fr circular mapping catheter could be one of the factors of the air spontaneously being introduced. Thereby, a larger shaft diameter of the circular mapping catheter may be preferable, which would avoid a mismatch of the diameter between the circular mapping catheter and the long sheath.

Needless to say, rapid removal of the catheter from the sheath creating a vacuum should not be done.

Continuous flushing with heparinized saline through a haemostatic valve is thought to be essential to prevent retrograde flow and clotting, but is less likely to help avoid air embolism as in our cases. Based on the assumed mechanism of air embolism in our cases, this method cannot stop the introduction of air.

Additionally, three sheaths within the LA could be associated with a higher incidence of coronary artery air embolism. The long sheath should be withdrawn to the right atrium or inferior vena cava to reduce the possibility of coronary artery air embolism or systemic thromboembolism when it is not actively being used during the procedure.

Limitations

As this was a retrospective study, there might be the possibility that air embolism were less observed in the later dexmedetomidine group because of a learning curve, such as fewer catheter exchanges.

An asymptomatic cerebral air embolism could not be completely ruled out because we performed neither head computed tomography nor magnetic resonance imaging. However, no patients exhibited any neurological impairment after the ablation session.

Conclusion

Massive air embolism with haemodynamic collapse and hypoxaemia occurred during the left atrial ablation of AF when long apnoea episodes were produced by deep or moderate sedation. Once air embolism occurred, the appropriate supportive treatment and aspiration of the air were effective in achieving a complete recovery. A sedative with a lesser respiratory depressive effect and flushing the sheath after inserting the circular mapping catheters contributed to a less frequent incidence of this complication.

Conflict of interest: none declared.

References