Author + information
- Received April 29, 2016
- Revision received September 26, 2016
- Accepted September 30, 2016
- Published online April 17, 2017.
- Michala E.F. Pedersen, MD∗ (, )
- Milena Leo, MD,
- Manish Kalla, MD,
- Aniel Malhotra, MD,
- Matthew Stone, MD,
- Kelvin Wong, MD,
- Kim Rajappan, MD,
- Matthew Ginks, MD,
- Timothy R. Betts, MD and
- Yaver Bashir, MD
- Oxford University Hospitals NHS Foundation Trust, Department of Cardiology, John Radcliffe Hospital, Oxford, United Kingdom
- ↵∗Address for correspondence:
Dr. Michala E.F. Pedersen, Department of Cardiology, John Radcliffe Hospital, Headley Way, Oxford OX3 9DU, United Kingdom.
Objectives This study reports on the safety of early removal of pericardial drains after cardiac tamponade complicating atrial fibrillation catheter ablation (AFCA) procedures, the need for repeat pericardiocentesis, major adverse outcomes, as well as length of stay, and the need for opiate analgesia.
Background Tamponade from AFCA is traditionally managed by pericardiocentesis with delayed removal of the drain (typically 12 to 24 h later) in case of re-bleeding. A drain in situ often causes severe pain but ongoing blood loss is rare. Our institution adopted the practice of early removal of drains before leaving the laboratory if bleeding has stopped.
Methods The authors performed a retrospective descriptive analysis of 43 cases of tamponade complicating AFCA from 2006 to 2015, comparing patients in whom the drain was removed early (group early removal [ER]; n = 25) versus traditional delayed removal (group delayed removal [DR]; n = 18).
Results The groups were similar with respect to clinical/demographic characteristics, proportions of first-time versus re-do and pulmonary vein isolation versus pulmonary vein isolation + additional ablation. There were no deaths. No ER patients required drain re-insertion before discharge. The length of stay was shorter in the ER group (3 days; range 1 to 9 days) than in the DR group (4 days; range 2 to 60 days). The requirement for opiate analgesia was less in the ER group (8%) than in the DR group (72%).
Conclusions Early removal of pericardial drains after tamponade complicating AFCA procedures appears to be safe and effective, with re-insertion not required in this cohort. The traditional practice of leaving drains in situ for 12 to 24 h may result in more patient discomfort and longer hospitalization.
Atrial fibrillation catheter ablation (AFCA) is one of the more complex interventional procedures in cardiology and carries a risk of complications. Cardiac tamponade is the most frequently encountered procedure-related complication occurring in 1.2% to 1.3% of patients in 2 worldwide surveys (1,2) and 1.3% in a European survey (3). It is also the main cause of procedure-related mortality, responsible for 25% of fatalities (4,5).
The majority of episodes of cardiac tamponade complicating AFCA can be managed successfully by immediate percutaneous drainage and reversal of systemic heparinization with protamine (6–8). There are no recommendations as to the length of time to leave the drain in situ. The only guidance on management of cardiac tamponade complicating ACFA we are aware of is in the 2012 Heart Rhythm Society (HRS)/European Heart Rhythm Society (EHRA)/European Cardiac Arrhythmia Society (ECAS) Expert Consensus Statement on Catheter and Surgical Ablation of Atrial Fibrillation, which states, “Once the pericardial space has been drained, the patient needs to be monitored for ongoing bleeding with the drainage catheter in place. Rarely, if there has been a tear, percutaneous drainage may be inadequate and surgical drainage and repair is needed” (6).
Traditionally the drain is left in the pericardial space for approximately 12 to 24 h in case of re-bleeding (7,8). Because this has the potential to cause severe chest discomfort from pericardial irritation, and ongoing blood loss is rare, our institution has gradually adopted the practice of early removal of drains before leaving the catheter lab if the bleeding has stopped. The aim of this study was to compare clinical outcomes after cardiac tamponade among patients who had early removal of the pericardial drain versus the more traditional delayed removal.
A systematic search of electronic patient records and the morbidity and mortality audits was performed identifying patients who had had an atrial fibrillation (AF) procedure complicated by pericardial effusion or tamponade requiring drainage. The medical notes, including anesthetic charts, were reviewed for each case, and the relevant information extracted and added to information from the procedure database and the electronic patient’s records. All patients who had a drain inserted in the electrophysiology (EP) lab for a pericardial effusion at the time of an AF ablation since 2006 were included. We did not include pericardial effusions caused by other types of ablation, for example, ablation of ventricular tachycardia, and patients who required drain insertion at a later date, for example, due to a late post-cardiac injury syndrome. Finally, if the bleeding did not stop in the laboratory and immediate surgery was required, patients were excluded for analysis. All data collected for this paper are observational.
The cases were then divided according to whether the drain had been removed in the laboratory (early removal [ER] group) or whether the drain had been removed at a later stage (delayed removal [DR] group). After the pericardial effusion had been drained to dryness, the drain was clamped. During the subsequent observational waiting period of at least 30 min, repeated transthoracic echocardiograms (TTEs) were performed to monitor for re-accumulation. If there was re-accumulation, the effusion would again be drained to dryness, the drain clamped, and another period of monitoring started. Once the operator was satisfied that there was no re-accumulation, the patient would be deemed ready to leave the laboratory. At this point it was left to operator discretion whether the drain could be removed (ER) or left in for monitoring (DR). The practice of early removal was gradually established between 2008 and 2013, as confidence and experience grew. Thus, in our earlier cases the predominant strategy was to leave the drain in for up to 24 h and in later cases to take the drain out before leaving the EP laboratory.
The groups were assessed for similarity in baseline: 1) clinical and demographic characteristics; 2) proportion of first-time versus re-do ablation; and 3) proportion of pulmonary vein isolation (PVI) alone versus PVI plus additional treatments (e.g., linear lesions).
All procedures were performed under general anesthesia apart from 3 patients; 1 patient in the ER group and 2 patients in the DR group were performed under conscious sedation. Venous access was obtained via the femoral veins and a multipolar catheter placed in the coronary sinus (CS). In cases performed under general anesthesia, all transseptal punctures were carried out under transesophageal guidance. In cases carried out under sedation, 1 case was done with intracardiac echocardiography (ICE probe, Siemens, Munich, Germany) for transseptal puncture guidance, 1 was via a patent foramen ovale (PFO)/atrial septic defect (ASD) and the remaining case was done with fluoroscopic guidance only. The punctures were done using either the Brockenbrough (Medtronic, Inc., Minneapolis, Minnesota), Endrys (Cook Medical, Bloomington, Indiana) or the NRG RF (Baylis Medical, Montreal, Quebec, Canada) transseptal needle through a long introducer sheath, usually the Mullens (Cook Medical) or a sheath of the Fast-Cath Transseptal Guiding Introducer SL Series (St. Jude Medical, Austin, Texas). At the operator’s preference, either a second separate transseptal puncture was done or the ablation catheter was passed through a second sheath, usually an Agilis (Agilis NxT Steerable Introducer, St. Jude Medical), and fed across the same initial puncture.
Irrigated tip ablation catheters were used in all cases apart from 1 patient who had cryoballoon therapy (Arctic Front Advance, Medtronic, Inc.). Three cases were done with the Hansen robotic navigation system. Patients with paroxysmal AF received PVI only and patients with persistent AF received PVI and linear and/or complex fractionated atrial electrogram ablation. A circular catheter was used as a guide for pulmonary vein electrical isolation in all patients. The ablation strategy was always at the operator’s discretion.
The baseline anticoagulant status was documented. In the earlier cases, Warfarin was held 3 days before the procedure and bridged with low molecular weight heparin and in later cases Warfarin was uninterrupted. Novel oral anticoagulants (NOAC) were routinely held 36 h pre-procedure. After transseptal puncture, all patients received intravenous heparin to achieve a target activated clotting time of 250 to 350 s. Each introducer sheath was continuously flushed with heparinized saline for the duration of the procedure.
Diagnosis and management of cardiac tamponade
Patients were closely monitored by the operator and by the anesthesiologist. If hypotension occurred, prompt echocardiography was performed to rule out or confirm a pericardial effusion. Cardiac tamponade was diagnosed when the pericardial effusion resulted in hypoperfusion and a drop in systolic pressure below 80 mm Hg. Pericardiocentesis was performed under fluoroscopic guidance by a subxiphoid approach. A pigtail angiographic catheter was introduced into the pericardial cavity to drain the blood. After the effusion was drained manually to dryness, the pigtail catheter was connected to a drain and ultimately removed when there was no evidence of re-accumulation of blood. If a long introducer sheath inadvertently had entered the pericardial space, this was used to drain the blood as far as possible. Simultaneously, protamine was administered to reverse heparinization. In patients on uninterrupted warfarin, if bleeding continued after protamine administration and initial drainage, warfarin was reversed with prothrombin complex concentrate (PCC) at the operator’s discretion.
As there were insufficient patients in this retrospective study to have confidence in inferential statistical tests, this was a descriptive study using standardized differences to assess group differences. To measure the effect size (d) between 2 groups, we used the approach of Yang and Dalton (9) who described methods to calculate the standardized difference for both continuous and categorical baseline variables.
Forty-five patients were identified having had catheter ablation for AF complicated by a pericardial effusion requiring a pericardial drain. Two patients were excluded as they required immediate surgery due to ongoing bleeding leaving 43 cases for analysis. Twenty-five of these had the drain removed early (ER group) and 18 patients had their drain removed at a later stage (DR group). As our practice of early removal of drains was gradually established between 2008 and 2013, there is some overlap in time between the 2 groups. The groups were similar with respect to baseline demographic and clinical characteristics (Table 1).
The 2 patients who required immediate surgery both exhibited ongoing re-accumulation of pericardial blood requiring continuous drainage. These ablations had been performed on uninterrupted warfarin (respective international normalized ratio [INR]: 2.0 and 2.3) and both received PCC. As bleeding was ongoing and as thrombus formed in the pericardial space, irreversible tamponade physiology developed requiring immediate surgery. The intra-operative findings were a 2- to 3-mm left atrial tear in 1 patient and a small hole in the roof in the other; the defects were repaired and clot evacuated.
Baseline anticoagulant status
More patients in the ER group underwent AFCA on uninterrupted warfarin. The average INR on the day of the procedure was 2.3 ± 0.51 in the ER group with all but 2 patients having uninterrupted warfarin. These 2 patients had NOAC withheld 36 h pre-procedure. In the DR group, 7 patients had stopped Warfarin pre-procedure and had an average INR of 1.2 ± 0.16. The remaining 11 patients in the DR group were on uninterrupted warfarin with an average INR of 2.2 ± 0.23. Details are listed in Table 2.
The 2 groups were similar in proportions of first time procedures versus re-do procedures (16 vs. 9 in the ER group and 10 vs. 8 in the DR group) or whether they were having a PVI-only procedure (11 in the ER group and 8 in the DR group) or a more complex ablation for persistent AF (14 in the ER group, 10 in the DR group), showing low standardized differences (Table 1).
There was only a small difference in the dose of heparin in the 2 groups. On average, the ER group received 10,705 ± 2,134 IU and the DR group 10,030 ± 4,014 IU (d = 0.21). The patients who had uninterrupted warfarin (n = 36) overall received less heparin compared to the patients who had had their anticoagulation stopped (n = 9): 13,316 ± 4,543 IU versus 9,258 ± 2,077 IU.
The procedure was completed in 10 of 18 cases in the DR group and in 13 of 25 cases in the ER group. Of the 10 patients with completed procedures in the DR group, 4 had this done with a drain or sheath in the pericardial space. Of the 13 patients with completed procedures in the ER group, 8 patients had their effusion noted at or after the completion of ablation and 5 patients had their procedures completed with a drain or sheath in the pericardial space.
Nine patients had their ablation procedure completed with a drain in situ (4 in the DR group and 5 in the ER group). Three of 5 patients in the ER group received heparin post-tamponade and protamine reversal at the end. The other 2 ER patients received protamine at the time of pericardial drainage but no further heparin was administered. Two of 4 patients in the DR group were reheparinized after pericardial drainage and the other 2 were not for the remainder of the procedure. These decisions were taken at individual physician discretion in the past. Following a clinical governance review, no ablation procedures are now completed post-tamponade without re-heparinization at our institution.
Management of pericardial effusion
Pericardial effusions were documented on echocardiography and heparin was reversed with protamine in all cases. Pericardial drains were positioned under x-ray guidance using the subxiphoid approach; there were no complications from the placement of the drain per se. In 3 cases a long transseptal sheath had inadvertently perforated into the pericardial space and this sheath was initially used to drain the effusion. In 1 case this was sufficient to drain the effusion to dryness and there was no re-accumulation after removing the long sheath. In 2 cases, hemopericardium re-accumulated after pulling back the long sheath and a subxiphoid drain was needed.
The blood volumes drained were similar in the 2 groups (839 ± 612 ml vs. 961 ± 841 ml in the ER and DR group, respectively) (d = 0.17). PCC was administered equally in the 2 groups (9 patients in the ER group, 6 patients in the DR group). Vitamin K was administered together with PCC in 4 patients in the ER group and in 2 patients in the DR group (Table 2).
The ER group had their pericardial drains removed in the lab before returning to the ward after a median of 1.5 h (range 0.5 to 3.5 h) waiting time confirming no re-accumulation. In the DR group, patients had their drain removed later, after a median of 17.0 h (range 5.0 to 23 h). They had the same duration of observation time for reaccumulation before leaving the EP laboratory (median 1.5 h; range 0.5 to 4.0 h). After drain removal, careful hemodynamic assessment and monitoring was continued on the ward. In no patient was there later re-accumulation of pericardial effusion as shown by no significant further drainage (>20 ml) as well as no effusion on repeat TTE before drain removal.
All patients were alert, well and hemodynamically stable, off any support, before leaving the laboratory.
Outcome after cardiac tamponade
No patient in either group required repeat pericardiocentesis. Two patients of 24 in the ER group needed opiate analgesia (8%), whereas 13 of 18 patients (72%) in the DR group needed opiates, reflected in a high effect size (d = −1.73). In the remaining 1 patient in the ER group, we have been unable to locate an entry clarifying whether opiate analgesia was administered or not. The length of stay appeared shorter in the ER group (median, 3 nights; range 1 to 9 nights) than in the DR group (median, 4 nights; range 2 to 60 days), the effect size being moderate at −0.48. Warfarin was restarted earlier in the ER group, on day 2 (median; range day 1 to day 7), and on day 3 in the DR group (median; range day 2 to day 14) (d = 0.75). Day 1 is considered to be the day of the procedure (Table 3).
Major adverse outcomes
One patient had a prolonged hospital stay due to sepsis and multi-organ failure but recovered and was discharged after 60 days in hospital.
Post-cardiac injury syndrome
Two patients in either group developed a post-cardiac injury syndrome (p = NS). The presentation of this was late pericarditic pain ± effusion and symptoms, but no clinical evidence of sepsis. One patient presented 2 weeks after the procedure with symptoms of pericarditis, a small pericardial effusion, and raised inflammatory markers. One patient presented after 3 weeks with pericardial and pleuritic pain, bilateral pleural effusions, and raised inflammatory markers. The remaining 2 patients both presented after 2 weeks with similar symptoms and additionally recurrence of a significant pericardial effusion, 1 that required drainage. All 4 patients were treated successfully with oral steroids.
These findings suggest that early removal of pericardial drains after cardiac tamponade complicating catheter ablation for AF is safe and effective in patients where ongoing bleeding has stopped. The findings are based on retrospective data from a highly experienced center. The data does not include patients with uninterrupted NOACs.
In the ER series the drain was taken out after a mean of 1.5 h (range 0.5 to 3.5 h) monitoring for re-accumulation. In the DR group patients were monitored in the EP laboratory for re-accumulation for the same duration (mean, 1.5 h; range 0.5 to 4.0 h) and when no re-accumulation was confirmed, was sent to the ward for ongoing monitoring of a median 17 h (range 5 to 23 h). There was no further drainage in any patient and the drain was removed on the ward after a repeat TTE. No patient needed repeat pericardiocentesis during their hospital admission. Patients with the drain left in situ required more opioid analgesia for pain relief and had a longer post-operative length of stay.
Duration of pericardial drainage
We are not aware of any studies or guidelines on the length of monitoring for potential re-accumulation of pericardial bleeding aside from the 2012 HRS/EHRA/ECAS Expert Consensus Statement: “Once the pericardial space has been drained, the patient needs to be monitored for ongoing bleeding with the drainage catheter in place” (6). There is only limited published data on the duration of the drainage. Bunch et al. (8) undertook a retrospective observational study of 15 consecutive patients requiring emergency pericardiocentesis during catheter ablation procedures for AF or left atrial tachycardia over a period of 5 years. Of these, 2 cases had ongoing bleeding and required surgery. In the remaining 13 cases the patients were managed on the cardiac care unit and the drain removed after a mean of 1.6 ± 1.5 days (range 1 to 6 days). The authors do not elucidate on the presence of any further drainage during the monitoring period in the cardiac care unit.
A study on uninterrupted warfarin for AFCA by Latchamsetty et al. (7) report that cardiac tamponade is no more severe or difficult to manage in the presence of therapeutic anticoagulation with Warfarin. In this series of 40 patients the pericardial drain was removed after a median of 1 day (interquartile range [IQR]: 1 to 2 days) in the group with interrupted Warfarin (n = 23) and after a median of 1 day (IQR: 0 to 1 days) in the group with therapeutic INR (n = 17). The authors do not report the additional blood volumes drained during the monitoring period.
In a very recent paper on AFCA on uninterrupted factor Xa inhibitors, Gianni et al. (10) report 11 patients who required pericardiocentesis. Two patients required surgery, 1 because bleeding did not stop and 1 because coagulation of pericardial blood prevented effective drainage. The remaining 9 patients were successfully treated with pericardial drainage only, with the drain “removed within 24 h.” The authors describe that these patients remained hemodynamically stable and no recurrence of effusion was noted during the observation period.
Duration of stay and need for opioid analgesia
The length of stay (LOS) in our series appeared to be shorter in the ER group than in the DR group, and although the effect was moderate this may have both clinical and health service benefits. In 3 previous studies reporting LOS after cardiac tamponade, all using the conventional approach of leaving a pericardial drain in for a day, 1 study reported a similar LOS and 2 studies a longer LOS. Latchamsetty et al. (7) recorded a median length of stay of 4 days (IQR: 3 to 5 days) in the interrupted Warfarin group and 2 days (IQR: 2 to 3 days) among patients undergoing AFCA on uninterrupted Warfarin. Bunch et al. (8) report an average LOS of 5.3 days (range 3 to 12 days) in their cohort of 13 patients with pericardial effusion treated successfully with pericardiocentesis. In a study comparing the cryoballoon and radiofrequency ablation, 2 patients suffered tamponade treated by pericardiocentesis and required 4 adjunctive hospitalization days (11).
We found that 72% of our patients in the DR group required opioid analgesia versus only 8% in the ER group. Bunch et al. (8) reported that 53% of their retrospective cohort of patients with a pericardial drain in situ for 1.5 ± 1.5 days (range 1 to 6 days) following cardiac tamponade had described pericardial pain; they believed it to be secondary to pericarditis due to irritation from the drain.
Suggested algorithm for pericardial effusion complicating AFCA
Two patients were excluded for analysis due to ongoing bleeding requiring continuous aspiration and both were found to have left atrial tears at surgery. In all other patients of our cohort it was possible to drain the effusion to dryness with no evidence of re-accumulation over a continuous observation period of at least 30 min (overall mean, 1.5 h after drain placement). No patients subsequently developed a later recurrence of pericardial effusion. Therefore, we suggest that in cases fulfilling these criteria, it is safe to remove the drain after a 30-min period without re-accumulation. Following drain removal, ongoing careful hemodynamic assessment and monitoring is required (Figure 1).
This was an observational retrospective descriptive study in a single center with multiple operators. A randomized prospective study would be more valuable to assess the benefits of early removal of pericardial drains after tamponade during catheter ablation for AF ablation. However, such a study would be protracted and difficult due to low incidence of cardiac tamponade complicating AF ablation.
Early removal of pericardial drains after cardiac tamponade complicating AFCA procedures appears to be safe and effective, with re-insertion not required in our experience. The traditional practice of leaving drains in situ for 12 to 24 h may result in more patient discomfort and longer hospitalization.
COMPETENCY IN MEDICAL KNOWLEDGE: Leaving in a drain in the pericardial space after cardiac tamponade complicating an AFCA procedure may at first appear to be the safest, and maybe easiest, strategy. However, the findings of this paper suggest that as long as bleeding has stopped taking out the pericardial drain is safe with no need for later re-insertion. Adopting such a practice of early removal is largely to the benefit of the patients (patient care) who will experience less discomfort and shorter hospital admissions. Also, we hope our findings have expanded the clinical skills of the clinicians to include the option of early removal of pericardial drains with the knowledge that this paper have shown it to be safe. Additionally, there may be benefits to health economics.
TRANSLATIONAL OUTLOOK: A randomized prospective study would be more valuable to assess the benefits of early removal of pericardial drains after tamponade during catheter ablation for AF ablation. However, such a study would be a challenge, protracted, and difficult due to low incidence of cardiac tamponade complicating AF ablation, even if it was carried out as a collaboration between multiple institutions. We believe the strategy of early removal is safe; our cohort of 43 patients where none needed drain re-insertion is a reasonable number, but more data on this subject should be collected and published.
Dr. Betts is supported by the Oxford Biomedical Research Centre, John Radcliffe Hospital, Oxford, United Kingdom. Dr. Leo has received a research fellowship from St. Jude Medical. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- atrial fibrillation catheter ablation
- delayed removal
- early removal
- length of stay
- prothrombin complex concentrate
- pulmonary vein isolation
- transthoracic echocardiogram
- Received April 29, 2016.
- Revision received September 26, 2016.
- Accepted September 30, 2016.
- 2017 American College of Cardiology Foundation
- Cappato R.,
- Calkins H.,
- Chen S.A.,
- et al.
- Cappato R.,
- Calkins H.,
- Chen S.A.,
- et al.
- Cappato R.,
- Calkins H.,
- Chen S.A.,
- et al.
- Belhassen B.
- Calkins H.,
- Kuck K.H.,
- Cappato R.,
- et al.
- ↵Yang D, Dalton JE. A unified approach to measuring the effect size between two groups using SAS. SAS global forum 2012, Statistics and Data Analysis, Paper 335. Available at: http://support.sas.com/resources/papers/proceedings12/335-2012.pdf. Accessed October 2016.
- Gianni C.,
- Di Biase L.,
- Mohanty S.,
- et al.