| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Eur J Cardiothorac Surg 2006;30:852-872
© 2006 Elsevier Science NL
a James Cook University Hospital, Middlesbrough, UK
b Guy's Hospital, London, UK
c Leiden University Medical Center, The Netherlands
d Papworth Hospital, Cambridge, UK
Received 12 April 2006; received in revised form 28 July 2006; accepted 4 September 2006.
* Corresponding author. Address: Papworth Hospital, Cambridge CB3 8RE, UK. Tel.: +44 1480 364299; fax: +44 1480 364744. (Email: sam.nashef{at}papworth.nhs.uk).
 |
1. Executive summary |
|---|
 Top 1. Executive summary 2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
|
2. Introduction |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
The aim of this guideline is to summarise the evidence in support of a variety of prophylactic and therapeutic regimens for the management of AF and to provide a comprehensive strategy to minimise the impact of this complication for patients undergoing cardiothoracic surgery.
|
3. Scope of the guideline |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
|
4. Methodology of the guideline |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
Guidelines often fall short of expectations due to a failure to consult those clinicians who are most likely to use them. For this guideline, every literature review has already been published in full in the ICVTS. Every topic was published online and clinicians were able to post comments on the topic over a 2-month period. These comments were then published together with the full paper in the ICVTS and are now available to all readers in full text online at www.icvts.org.
|
5. Levels of evidence and grading of recommendations |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
Systematic literature review was up to the end of 2005. Once recommendations are made, they are graded according to the quality of papers used to come to our conclusion:
|
6. Definitions of AF |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
These statements are in agreement with the American Heart Association (AHA) guidelines on AF published in JACC 2001 [7].
|
7. Prophylaxis for AF after cardiac surgery |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
Included were five meta-analyses, two systematic reviews and a cohort study of 630,000 patients from the STS database all providing strong evidence in support of ß-blocker prophylaxis.
All the identified meta-analyses concluded that ß-blockers significantly reduced the incidence of AF [9–13]. The largest meta-analysis was by Crystal et al. [11], published in 2002 in Circulation. They reported that across 27 randomised controlled trials (RCTs) with 3,840 patients, controls had an incidence of AF of 33% but patients receiving ß-blockers had an incidence of 19%, corresponding to a number needed to treat (NNT) of only seven. Of note, they have recently updated their findings as a Cochrane review [14]. Ferguson et al. [15] performed a large retrospective analysis of the STS surgical database containing 629,877 patients to look at the mortality and morbidity associated with peri-operative ß-blocker use. After propensity analysis, they found a decrease in mortality from 3.4% to 2.8% for patients receiving peri-operative ß-blockers.
The American Heart Association [16] strongly recommends routine pre-operative or early post-operative ß-blocker therapy as the standard of care for coronary artery bypass grafting (CABG). Thus, there is very strong evidence that ß-blockers reduce the incidence of AF. In addition, there is evidence that ß-blocker prophylaxis reduces length of stay, costs, mortality and morbidity. Evidence for amiodarone prohylaxis is similar (see Section 7.3) but this is based on fewer RCTs and many patients having heart surgery are already taking ß-blockers with an established safety record.
| Recommendation ß-Blockers should routinely be used as first choice for the prophylaxis of AF in all patients undergoing cardiac surgery, unless otherwise contraindicated.
If the patient is on ß-blockers, these should be continued up to the morning of surgery and restarted on the first post-operative day. If the patient is not on ß-blockers, one should be started pre-operatively or post-operatively unless otherwise contraindicated.
|
7.2 Sotalol versus standard ß-blockers for prophylaxis against AF
Evidence was sought for whether sotalol might be superior to standard ß-blockers for prophylaxis against AF during cardiac surgery. This search is fully documented in the ICVTS (Patel and Dunning [17]) together with a summary of all identified papers. We identified 51 papers using the presented search strategy. From these, 8 papers represented the best evidence on this topic.
Seven RCTs compared sotalol to a conventional ß-blocker. Study sizes ranged from 101 to 429 patients. All seven papers showed a greater reduction in the incidence of AF compared with conventional ß-blockers although, two of these studies did not reach statistical significance. The largest study was by Suttorp et al. [18,19], who performed a four-arm study comparing low or high doses of sotalol or propranolol in 429 patients. Sotalol 40 mg tds resulted in an incidence of 14% of AF compared with 19% incidence of low dose propranolol (p = ns). Auer et al. [20] studied 253 patients randomised to four regimes including sotalol or metoprolol. The incidence of AF was 32% with sotalol and 40% with metoprolol, although this was again non-significant. Sanjuan et al. [21] found a significant reduction from 22% to 10% comparing atenolol with sotalol in 253 patients. Janssen et al. [22] studied 130 patients randomised to sotalol, metoprolol or no therapy. Only 2.4% of patients receiving sotalol went into AF, compared with 15% in the metoprolol group and 36% of controls, which was a significant finding. No serious side effects were reported in any group. Parikka et al. [23] randomised 191 patients to either sotalol or metoprolol. AF occurred in 16% of patients receiving sotalol compared with 32% of patients receiving metoprolol (p < 0.01). Nystrom et al. [24] randomised 101 patients to high dose sotalol or (1/2) dose ß-blockers. The incidence of AF was 10% in the sotalol group compared with 29% in the ß-blocker group (p = 0.028). Abdulrahman [25] studied 191 patients randomised to sotalol or metoprolol. The incidence was 10% in the sotalol group and 22% in the metoprolol group.
Crystal et al. [11] briefly summarised these studies and found that the incidence of AF in the sotalol groups was 12% compared with 22% in the ß-blocker groups, which was a significant finding, and they showed that the number needed to treat with sotalol over standard ß-blockers was 10 to prevent an additional case of AF. In these studies, either 40 mg tds or 80 mg bd were low in side effects but doses higher than this caused a higher incidence of side effects.
| Recommendation Sotalol may be more effective than standard ß-blockers for the prevention of AF without causing an excess of side effects.
|
7.3 Amiodarone for prophylaxis against AF
Evidence was sought for whether amiodarone provided effective prophylaxis against AF during cardiac surgery. This search is fully documented in the ICVTS (Dunning et al. [26]) together with a summary of all identified papers. A total of 90 papers were identified using the presented search strategy. From these, 12 papers represented the best evidence on this topic.
Eleven RCTs and one meta-analysis were identified. Nine of the 11 RCTs showed a significant reduction in the incidence of AF, and the remaining two studies showed non-significant trends in the reduction of AF. Included in these papers were two very well conducted studies published in the NEJM and the Lancet. Daoud et al. [27] randomised 124 patients to either 7 days of pre-operative oral amiodarone or controls. The incidence of AF was 25% in the amiodarone group and 53% in the control group. Giri et al. [28] randomised 220 patients who were already receiving ß-blockers to either amiodarone or placebo. The incidence of AF was 23% in the amiodarone group and 38% in the control group (p = 0.01). Ten of these RCTs were summarised by Crystal et al. [11], who reported an incidence of AF of 22.5% in the amiodarone groups, and an incidence of 37% in control groups, giving a NNT of seven to prevent an additional case of AF.
Since our review, one additional study has been published in 2004 by Kerstein et al. [29]. They used a case-control design to study 51 CABG patients receiving intravenous amiodarone peri-operatively to 92 patients not receiving amiodarone during the same period. The incidence of AF was 6% in the amiodarone group compared with 26% in the controls. There was no significant difference in complications. Of note, most patients also received ß-blockers and this study is limited by its non-randomised design.
The incidence of complications was low in all these RCTs except the studies by Butler et al. [30], who found a significantly increased rate of bradycardias and pauses, and Hohnloser et al. [31] who had to stop therapy in 18% of patients receiving amiodarone. Also of note, the reported studies varied in their protocols with many of the studies continuing ß-blockers peri-operatively in their patients.
| Recommendations Amiodarone should be used for prophylaxis of AF in all patients undergoing cardiac surgery in whom ß-blocker therapy is not possible.
In high-risk patients receiving ß-blocker therapy for prophylaxis of AF, amiodarone may also be used as additional prophylaxis with an acceptably low incidence of complications. These patients should be protected from the complications of bradycardia with temporary pacing wires being placed intra-operatively.
|
7.4 Bi-atrial pacing for prophylaxis against AF
Evidence was sought for whether atrial pacing reduces the incidence of post-operative AF in patients undergoing cardiac surgery. This search is fully documented in the ICVTS (Ronald and Dunning [32]) together with a summary of all identified papers. Four hundred and fifty eight papers were identified using the presented search strategy. From these, 16 papers represented the best evidence on this topic.
We identified 13 RCTs recruiting between 21 and 230 patients. These varied markedly both in their placement of atrial pacing wires, pacing strategies and also in their definitions of AF. Eleven studies looked at biatrial pacing. Six found a significant benefit and five found no significant benefit. Eight studies used right atrial pacing. Two studies reported a significant benefit but six found no benefit. Ten of the 13 completed RCTs were identified in the meta-analysis by Crystal et al. [11] but we repeated the meta-analysis in order to update these data. We found that there was a significant benefit shown for biatrial pacing (OR 0.51, 95% CI 0.36–0.72) but combining the right atrial studies showed no significant benefit (OR 0.71, 95% CI 0.45–1.15).
While there is a clear benefit in the use of biatrial pacing, several papers reported technical difficulties, with loss of sensing, diaphragmatic pacing and LV pacing, which led to a number of patients being withdrawn from their respective studies. Thus, if biatrial pacing is contemplated, much care must be used when placing the wires. In addition, many different pacing strategies were used, but most paced at a rate of 80–90 beats/min with higher rates if the native rhythm went above 80 beats/min. Finally, the number of days of pacing varied among the studies. As the incidence of AF generally peaked around day 2 in these studies, 3–5 days of pacing seem necessary for full protection from AF.
| Recommendations Biatrial pacing significantly reduces the incidence of AF in patients undergoing cardiac surgery. Care must be taken in the placement of the wires in order to minimise avoidable complications.
|
7.5 Magnesium for prophylaxis against AF
Evidence was sought for whether magnesium reduces the incidence of post-operative AF in patients undergoing cardiac surgery. This search is fully documented in the ICVTS (Rostron et al. [33]) together with a summary of all identified papers. One hundred and thirteen papers were identified using the presented search strategy. From these, 21 papers represented the best evidence on this topic. Two further meta-analyses were also identified after this search was conducted [34,35].
We identified 20 RCTs that addressed this issue and 3 meta-analyses. One meta-analysis summarised data on 2069 patients from all but six of the RCTs that we identified. The meta-analysis was performed by Shiga et al. [36] in 2004, summarising papers that contained magnesium alone as prophylaxis and compared it to placebo treatment. Twenty three percent of patients in the magnesium groups suffered a supraventricular arrhythmia compared with 31% in the placebo group (p = 0.002). This gives a NNT of 13 to prevent one episode of supraventricular tachyarrhythmia (SVT). In addition, they found a significant reduction in the incidence of ventricular arrhythmias with a NNT of 14 to prevent one episode of ventricular tachyarrhythmia (VT). Shiga et al. also summarised the complications reported in 648 patients. They found no episodes of bradycardia or hypotension. However, important differences were found between all these studies and no one prophylactic regime was found to be superior to another. Regimes ranged from a single dose of 5 mmol in the cardioplegia solution to 110 mmol over the course of 3 days.
The second meta-analysis published in 2005 [34] summarised 20 studies with 2490 patients. They found that the incidence of post-operative AF reduced from 28% to 18% (OR 0.54, 95% CI 0.38–0.75). They also summarised the effect on length of stay and mortality but found no significant difference. Again, they did not recommend one specific magnesium prophylactic regime.
The most recent meta-analysis summarised only eight RCTs that compared magnesium with placebo [35]. They also found a highly significant reduction in relative risk with the addition of magnesium (RR 0.64, 95% CI: 0.47–0.87).
There were six studies that investigated over 200 patients. Toraman et al. [37] in 2001 performed an RCT in 200 patients, giving them either 6 mmol of magnesium both pre-operatively and post-operatively or placebo. Only 2 (2%) of patients receiving magnesium went into AF compared with 21 (21%) in the control group. Unfortunately, patients receiving ß-blockers or digoxin were excluded.
Forlani et al. [38] performed an RCT in 2001, separating 207 patients into four groups. Patients received either sotalol 80 mg bd or magnesium 1.5 g orally for 6 days post-operatively or both or neither treatment. Only 1 of 52 patients who received both treatments went into AF compared with 19 of 50 control patients.
Hazelrigg et al. [39] randomised 105 patients to receive 80 mg/kg of magnesium pre-operatively, then 8 mg/(kg h) post-operatively for 48 h or placebo in 97 patients. Thirty-two treatment patients went into AF compared with 41 control patients, which was a non-significant trend towards benefit. However, the reduction in AF was significantly different between groups on day 1.
Yeatman et al. [40] performed the largest study on magnesium prophylaxis, with 400 patients randomised in a double blind fashion to receive 40 mmol of 2 mmol/ml magnesium sulphate in the cardioplegia solution or controls. They found that the incidence of AF was 22% in the magnesium group compared with 29% in controls, which was non-significant, although the findings were significant in a subset analysis of urgent patients. The authors acknowledged that their dose of magnesium only produced a concentration of 5 mmol/l of cardioplegia, when actually they should have used a higher dose to obtain a concentration nearer 15 mmol/l of cardioplegia.
Bert et al. [41] performed a multi-arm study in 387 patients randomised into six groups of prophylaxis including 2 g of magnesium pre-operatively, post-operatively and for 4 days after the operation. Unfortunately, addition of magnesium had no beneficial effect as compared with ß-blockers, digoxin or controls.
Kaplan et al. [42] performed a study in 200 patients, giving 3 g of magnesium intravenously pre-operatively and post-operatively and for 3 days. No significant difference was found overall, although in a sub-analysis of patients who had low pre-operative serum magnesium, a significant reduction in AF was demonstrated.
With regard to which regime should be employed, Yeatman et al. [40], who performed the largest study, recommend 15 mmol/l in the cardioplegia solution, although they used a dose smaller than this in their study. Toraman et al. [37] found the greatest beneficial effect in their large study of 200 patients. They used 6 mmol magnesium sulphate infusion in 100 ml 0.9% NaCl solution (at 25 ml/h) the day before surgery, just after cardiopulmonary bypass and once daily for 4 days after surgery. As this study demonstrates the largest benefit in a well-conducted study, perhaps this should be regarded as the optimal regime so far investigated.
| Recommendations Prophylaxis with magnesium is an effective strategy to minimise the incidence of AF for patients undergoing cardiac surgery. This may safely be given in addition to other strategies to reduce the incidence of AF.
|
7.6 Prophylaxis against AF after lung surgery
Evidence was sought for whether there is a role for prophylaxis against AF for patients undergoing lung surgery. This search is fully documented in the ICVTS (Shrivastava et al. [43]) together with a summary of all identified papers. A total of 457 papers were identified using the presented search strategy. From these, 14 papers represented the best evidence on this topic.
We identified 14 RCTs that studied the prophylactic effects of diltiazem, ß-blockers, digoxin, verapamil, flecainide, amiodarone, magnesium and epidural anaesthesia. The quality and sizes of these studies varied greatly, ranging from 330 patients to only 30 patients. The strongest paper was by Amar et al. [44] and evaluated prophylactic diltiazem. They demonstrated a significant reduction in post-operative AF in lung resection in an RCT of 330 patients with the incidence of 26% in control patients reduced to 14% in diltiazem patients. Furthermore, using a loading dose of 0.25 mg/kg of diltiazem administered over 30 min followed by a continuous infusion (0.1 mg/(kg h)) for 18–24 h, there were no significant increased adverse reactions to drug treatment. This well conducted study gives a NNT of eight to prevent one episode of AF.
A second large study was performed by Van Mieghem et al. [45] in 199 patients in an unblinded RCT. They evaluated prophylactic verapamil and found that the incidence of AF reduced from 15% to 8%. This was a non-significant finding and, importantly, 23% of patients experienced bradycardia or hypotension as a result of verapamil and had to be withdrawn from the study. Interestingly, Van Mieghem's study was originally a three-arm trial of verapamil, amiodarone and controls but was stopped early after three patients in the amiodarone group developed adult respiratory distress syndrome (ARDS), with two patients dying of this complication. The authors then performed a retrospective review of 552 lobectomies and pneumonectomies [46] and found 55 patients who received amiodarone, of whom 6 developed ARDS (11%), but of 497 patients who did not receive amiodarone, only 9 (1.8%) developed ARDS.
Flecainide was examined in two studies by Borgeat et al. [47,48], with 30 patients in each study. Both studies showed reductions in AF from 40% to less than 10%, but as there were only 15 patients in each treatment group, these findings were on the margins of significance.
ß-Blockers have also been studied. Jakobsen et al. [49] randomised 30 patients to metoprolol versus controls and showed a significant reduction in AF from 40% down to 7%. Bayliff et al. [50] randomised 49 patients to propranolol and 50 to controls. They only showed some trends towards effectiveness and the propranolol group suffered a 50% rate of hypotension.
Oka et al. [51] looked at morphine versus bupivacaine given by epidural in 50 patients. The incidence of AF reduced from 28% in the morphine group to 4.3% in the bupivacaine group and there was no excess of side effects in this group.
Finally Terzi et al. [52] in 1996 performed an unblinded RCT in 194 patients, allocating 93 patients to prophylactic magnesium. They demonstrated a reduction in the incidence of AF from 23% to 10% in the magnesium group, without side effects.
In summary, digoxin and verapamil do not reduce the incidence of AF. Amiodarone may cause ARDS, and cannot be recommended. Flecainide and ß-blockers have been inadequately studied to make safe recommendations. Single RCTs have demonstrated evidence for bupivacaine epidural, magnesium and diltiazem for the prophylaxis of AF in patients undergoing non-cardiac thoracic surgery.
| Recommendations There is currently inadequate evidence to recommend routine prophylaxis against AF for all patients undergoing lung surgery.
|
7.7 Does off-pump CABG reduce the incidence of post-operative AF?
Evidence was sought for whether patients undergoing off-pump CABG (OPCAB) are at lower risk from AF. This search is fully documented in the ICVTS (Raja et al. [53]) together with a summary of all identified papers. One hundred and seven papers were identified using the presented search strategy. From these, 18 papers represented the best evidence on this topic. We identified five meta-analyses, six RCTs not included in the meta-analyses and a large cohort study that provided direct evidence for this topic. Of note, several studies split their results across multiple papers.
Reston et al. [54] performed a comprehensive and well-balanced meta-analysis in 2003 [55] of the short term and mid-term outcomes of OPCAB versus conventional CABG. Selecting 28 studies, they found that there was a highly significant reduction in AF in the OPCAB group (odds ratio of 0.69 in favour of OPCAB). There was, however, significant heterogeneity between these studies that they could not account for. However, if only the RCTs are included, the difference was increased rather than decreased. They did caution that most studies excluded patients such as non-elective surgery, re-operation, renal failure and impaired left ventricular function.
While Reston et al. also found significant benefits in terms of stroke, myocardial infarction (MI) and mortality, a meta-analysis by Van der Heijden et al. [56] in 2004 that assessed only RCTs disagreed with their meta-analysis, finding that there was no significant difference in the combined end-point of MI, death or stroke. Although this study did not look at AF, it is interesting to note that this meta-analysis also included the Octopus study [57], the SMART study [58] and an RCT from Hawaii [59]. This calls into question whether the meta-analysis by Reston et al. is already outdated.
Ascione and Angelini performed a pooled meta-analysis of the BHACAS 1 and 2 [60]. They showed that the incidence of AF reduced from 37% to 13%. This was a highly significant finding. There was no difference in mortality or cardiac events.
Of the recent randomised trials not included in the Reston meta-analysis, the SMART trial of 200 patients randomised to either OPCAB or conventional CABG found no significant difference in AF but, with an incidence of 16% in the OPCAB group and 22% in the CABG group, there was a trend towards reduced incidence in OPCAB. In contrast, the PRAGUE-4 trial [61] that randomised 400 patients to OPCAB or CABG found no difference at all in the incidence of AF. The OPCAB group had an incidence of 20% compared with an incidence of 24% in the CABG group.
The OCTOPUS trial [62] found no difference in AF between the two groups with a 20% incidence in the OPCAB group and a 21% incidence in the CABG group. Muneretto et al. [63] performed an RCT in 176 patients comparing total arterial OPCAB with total arterial CABG. They found that the incidence of AF was 22% in the OPCAB group and 35% in the CABG group, which showed a strong trend towards a lower incidence in the OPCAB group (p = 0.06).
Gerola et al. [64] performed an RCT in 2004 in Brazil in 160 patients and found a low incidence of AF in both groups. The finding of 9% in the OPCAB group and 5% in the CABG group was far lower than other studies, and calls into question the measurement of AF in their study, which was not described in the protocol.
Lee et al. [59] in 2003 performed a small RCT in 60 patients, and found an incidence of AF of 23% in OPCAB group and 39% in the conventional group, but this was not statistically significant.
In summary, three of the five meta-analyses directly assessed AF in OPCAB versus conventional CABG. They all found a significant reduction in AF with OPCAB. Six further RCTs were identified that were published after several of these meta-analyses. None of them identified a significant difference individually but their results can be summarised as follows: there was a 17.8% AF rate in the OPCAB group (114/642) and a 23% rate of AF in the CABG group (144/620). This corresponds to an odds ratio of 0.76 with a probability of 1.6% that the results are non-significant. These studies are, therefore, in broad agreement with the already performed meta-analyses that found significant differences. Our summary of the recent RCTs gives a NNT of 20 to avoid one incidence of AF.
| Recommendation Off-pump CABG reduces the incidence of AF compared with conventional (on-pump) CABG.
|
7.8 Operative interventions to reduce the incidence of AF
Several search strategies were employed to search for intra-operative interventions that might reduce the incidence of AF. However, only four papers were found that documented any intra-operative intervention to reduce the incidence of post-operative AF [65–68]. These include three papers investigating aortic fat pad removal, and one paper performing a posterior pericardiotomy in order to reduce the incidence of AF. Thus, this search was not published in the ICVTS.
Melo et al. [68] performed a ventral cardiac denervation in 207 patients undergoing low risk coronary arterial surgery. The fat pads that surround the superior vena cava, the aorta and the anterior and right lateral aspects of the main pulmonary artery were excised prior to cardiopulmonary bypass. This took on average 5 min (±2 min). A total of 219 patients were identified to act as non-randomised controls. Of the patients who had ventral cardiac denervation, 15 developed AF (7%, 95% CI 4–12%) compared with 56 in the control group (27%, 95% CI 18–35%), which was a significant finding (p < 0.01). This prospective cohort study was limited by a lack of telemetry post-operatively, a 20% incidence of patients undergoing off-pump surgery, 20% non-use of ß-blockers, and its non-randomised design (level 3b study).
Davis and Jacobs [67] hypothesised that removal of the aortic fat pad and its extension to the pulmonary artery may, in fact, increase the incidence of AF. They conducted a pilot study and subsequent cohort study in 320 patients, with the treatment group protected from the standard procedures of aortic fat pad disruption. They found that there was no difference between the groups in the incidence of AF and concluded that the aortic fat pad did not impact the incidence of AF (level 3b study).
Alex and Guvendik [66] carried out a cohort study where 70 consecutive patients undergoing CABG but also having ventral cardiac denervation were compared with 70 consecutive patients who did not have this additional procedure. Denervation took around 5 min to complete. The AF rate was 34% in control patients compared with 29% in those patients who had ventral cardiac denervation, which was a non-significant finding (p = 0.03).
Farsak et al. [65] studied the impact of a posterior pericardiotomy, with a 4 cm incision made posterior to the left phrenic nerve. An RCT was performed in 150 patients undergoing CABG. Seven of the 75 patients receiving a posterior pericardiotomy (9.3%) went into AF, compared with 24 of 75 controls (32%). In addition, the incidence of early and late pericardial effusion was significantly lower in the pericardiotomy group. Of note, patients receiving ß-blockers were excluded from this study (level 1b study).
| Recommendations Ventral cardiac denervation has not been convincingly shown to affect the incidence of post-operative AF in patients undergoing CABG.
|
|
8. Treatment of AF |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
8.1 Potassium
The role of potassium in the treatment of AF after cardiac surgery is not fully understood. Standard practice after cardiac surgery is to maintain serum potassium at the upper half of the normal range (4.5–5.5 mmol/l). The traditionally accepted view is that hypokalaemia predisposes towards potentially fatal ventricular dysrhythmias and that avoiding the risk of hypokalaemia will increase the margin of safety. Whether this helps avoid or treat AF has never been proven scientifically, and is unlikely to be proven in the future as the perceived risk of ventricular dysrhythmias will pose ethical obstacles to the conduction of any randomised trial in which cardiac surgical patients are allocated to a low potassium study arm. In view of this, it seems sensible to continue to recommend correction of absolute or relative hypokalaemia to the upper half of the normal range.
8.2 Pharmacological antiarrhythmic treatment
Evidence was sought for the optimal medical treatment for stable patients going into AF after cardiac surgery. This search is fully documented in the ICVTS (Dunning et al. [69]) together with a summary of all identified papers. We identified 364 papers using the presented search strategy. From these, 12 papers represented the best evidence on this topic. Of note, for this search, if two papers were found investigating the same drug, only the better paper was included. Therefore, for this guideline, the search was repeated and all papers were tabulated (Table 1
). Several studies were found that were cohort studies without a control group and these were excluded [70–74]. In order to provide a summary of the 19 studies identified in this area, and of the 12 drugs studied, the regimes, the number of studies, with the number of patients included in the studies and the findings and complications found are summarised in Table 2
.
|
|
8.2.2 Digoxin
Five studies assessed digoxin for the treatment of AF [75]. Again, no placebo groups were included in these studies but complications with digoxin were particularly low, with three papers reporting no complications at all. However, digoxin was shown to be less effective than procainamide, flecainide, diltiazem and possibly sotalol.
8.2.3 ß-Blockers
Four studies assessed ß-blockers for the treatment of AF [80–83]. The drugs studied were sotalol, esmolol and metoprolol. High doses of ß-blockade did cause a high incidence of hypotension and esmolol had to be withdrawn in nearly 50% of patients included in the study investigating this drug. Again, no placebo groups were included which would allow a more objective assessment of the effect of ß-blockers versus other treatments but the identified studies showed superiority over digoxin and diltiazem.
8.2.4 Diltiazem
Four studies investigated diltiazem [82–85]. Regimes were broadly similar using 0.25 mg/kg over 2 min and then a 5 mg/h infusion. The main complication was hypotension and this was around 20%. Diltiazem was shown to be better than placebo and digoxin. Esmolol may be superior to diltiazem but with more complications.
8.2.5 Propafenone
Five studies looked at propafenone [76,79,86–88]. Regimes used were either an iv bolus of 1–2 mg/kg over 10–15 min, followed by an infusion of 10 mg/kg over 24 h or an oral dose of 400 mg. Propafenone caused a high incidence of hypotension and one study reported a mean drop in blood pressure of 9 mmHg. Propafenone was found to be superior to procainamide and perhaps amiodarone, but not different from ibutilide. The remaining drugs are summarised in Table 2.
There are several weaknesses with the available studies in this area. Very few use a placebo group and, therefore, it is impossible to compare results of the various drugs across studies against placebo as a baseline. In addition, comparison across the studies is further hampered by a wide variation in the definition of success for each drug, with time spans from 30 min to several days being used to define successful cardioversion. Thus, it is currently impossible definitively to recommend one drug over another for the treatment of AF. The complication profile is an important issue in the selection of a drug to use against AF, and in fact it may be more important to choose treatment in a particular hospital where there is widespread familiarity with the drug and its complications.
| Recommendation No drug has definitively been shown to be superior to any other. Amiodarone, digoxin, ß-blockers and diltiazem are among the most studied drugs and seem to be associated with relatively few complications.
|
|
9. Anticoagulation for patients with de novo AF after cardiac surgery |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
There are several questions that must be addressed for this topic:
Addressing the issue of reduction in stroke risk in patients with AF, Hart et al. [90] analysed results from 16 trials by meta-analysis. They demonstrated that anticoagulation with warfarin reduced the relative risk of stroke both in comparison to placebo and aspirin and that warfarin is, therefore, by far the best long-term treatment in patients with AF. The NNT to prevent one stroke per year was 37 in the primary prevention group and 12 in the secondary prevention group (patients with a history of stroke) when compared with placebo. These results were consistent for disabling and non-disabling strokes. It is interesting to note that, though the incidence of intracranial haemorrhage was twice that of placebo, the difference was not statistically significant. The mean international normalised ratio (INR) achieved was 2.0–2.6 in primary prevention trials and 2.9 in a single secondary prevention trial. In addition, since the above meta-analysis was performed, Hylek et al. [91] published a cohort study of 13,600 patients in the New England Journal of Medicine, showing that an INR above 2.0 significantly improved survival among patients with AF who suffer a stroke.
The next issue is whether AF after cardiac surgery significantly increases the risk of stroke. Villareal et al. [92] reported that in a cohort of 6500 patients undergoing CABG, patients who went into AF had a much higher incidence of stroke (5.2% vs 1.7%) and also an increased risk of short and long-term mortality. These patients were, however, at significantly higher risk in a large range of categories including age, heart failure, chronic lung disease and underlying coronary arterial disease and thus, some caution should be used when analysing these figures. However, adverse long-term mortality persisted after case-control matching.
Stamou et al. [93] performed a retrospective analysis of 19,500 patients who had undergone CABG, of whom 333 had suffered a stroke. Multivariate analysis showed that AF was an independent predictor of stroke, increasing the odds of stroke by 1.7. However, multiple other high risk factors also predicted stroke and thus, the stroke group was a much higher risk group than those who did not suffer a stroke. Almassi et al. [3] also showed a 5% stroke rate in patients with AF compared with 2.5% in the sinus rhythm group, performing a similar study to Stamou in a cohort of 3855 patients. Creswell et al. [2] found that the incidence of stroke was 3.3% if the patient was in AF compared with 1.4% in those with sinus rhythm in a cohort of 3983 patients.
We found no studies that demonstrate that anticoagulation (immediate or delayed) of patients, who go into AF after cardiac surgery, significantly reduces this increased risk of stroke. However, addressing the issue of the safety of immediate anticoagulation, Malouf et al. [94] in a cohort study on 144 cardiac surgical patients, performed an echocardiogram on all patients. They found a 16% incidence of tamponade requiring drainage in patients receiving early warfarin, with no such tamponade in controls. In addition, 32% of the anticoagulated patients had a large pericardial effusion on echocardiography compared with 4% in controls. As a caveat, these patients received warfarin not heparin, and a large proportion had a period of excessive anticoagulation at some stage. Despite this, their figures are a cause for concern.
The American College of Cardiology, American Heart association and the European Society of Cardiology have joint guidelines for the management of patients with AF. In general, they suggest managing post-CABG AF in a similar fashion to AF in non-surgical patients [7]. They recommend the use of anticoagulant treatment in high-risk patients with a target INR of 2.0–3.0. They further recommend that anticoagulation with heparin or an oral anticoagulant is appropriate when AF persists more than 48 h. However, they only quote two pre-1990 papers in support of this guideline.
Daoud [95] supported these recommendations in a review published in 2004, stating that warfarin anticoagulation should be started after 48 h. He further recommends that in higher risk patients, even if sinus rhythm returns, warfarin should be continued for 4 weeks as there is a delay in return of atrial contractility after a period of AF.
The American Heart Association, supported by several authors, recommends warfarinisation while in AF, with an INR of 2–3 and anticoagulation within 48 h of the onset of AF after cardiac surgery. Chronic AF increases the risk of stroke and anticoagulation with warfarin provides the optimal protection from this risk, with a NNT of only 37 to prevent a stroke; this number drops to 12 if there is a history of stroke. It is also clear that AF after cardiac surgery doubles the risk of stroke, but no studies have yet demonstrated a drop in this risk with immediate anticoagulation. In addition, one study provides some evidence of increased risk of pericardial effusion and tamponade with early anticoagulation. A further compounding factor is the fact that in many patients, AF after cardiac surgery can be of relatively short duration because of either spontaneous or therapeutic cardioversion. These factors affect the choice of anticoagulant to be used in AF after cardiac surgery, not least because AF may have resolved by the time adequate anticoagulation with warfarin has been achieved.
| Recommendations After cardiac surgery, patients with AF should be anticoagulated while in AF and full anticoagulation should be started within 48 h of the onset of AF due to a doubling of their risk of stroke. This can be achieved with warfarin (INR 2–3), intravenous heparin or full-dose low molecular weight heparin.
|
|
10. Is antero-posterior position superior to antero-lateral position for placement of electrodes for external cardioversion of AF? |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
Five prospective RCTs were identified. None of these were specifically regarding patients who had cardiac surgery. Five studies were performed in a total of 628 patients. Two studies showed no difference between groups, one study showed a statistically significant benefit in favour of antero-lateral positioning and two studies showed a statistically significant benefit in favour of antero-posterior positioning. If all these studies are combined, there was an 82% conversion rate with antero-posterior positioning and a 77% conversion rate with antero-lateral positioning (p
= 0.09). Of note, all studies showed that if one position failed to convert the patient, success could still be achieved by alternating the paddle position.
| Recommendation There is no clinically significant difference between antero-lateral and antero-posterior pad positioning for cardioversion, although if cardioversion fails in one of these positions, the alternative position may still succeed.
|
|
11. Does magnesium offer additional benefit in patients having anti-arrhythmic treatment for AF after cardiac surgery? |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
Despite searching Medline, Embase, CINAHL, Cochrane and American Heart Association databases and guidelines, we found no studies that looked into the effect of using magnesium to treat patients going into AF after cardiac surgery. We extended the search to papers that might aid in a decision as to whether magnesium may potentially aid rate control or cardioversion in cardiac surgical patients.
Shiga et al. [36] performed a comprehensive meta-analysis in 2004, looking at the benefit of prophylactic magnesium in the prevention of AF post cardiac surgery. They identified 17 RCTs, comprising 2069 patients. In the pooled magnesium groups, the incidence of SVT was 23%, but in the control group it was 31% (p = 0.002). In addition, the incidence of ventricular tachycardia was also significantly lower, and the mean serum magnesium was significantly higher than in the control groups. Magnesium reduced the incidence of AF by 29% across the 17 trials performed.
In the general medical literature, we found seven papers that looked at either addition of magnesium or magnesium alone in the therapy of AF. Four of the seven papers demonstrated a significant benefit. Kalus et al. [98] considered the efficacy of magnesium as an adjunct to ibutilide in 321 medical patients in AF. This was a retrospective multicenter cohort study where the authors reviewed the case notes of patients in atrial flutter or fibrillation in whom cardioversion with ibutilide had been attempted. The rate of conversion was 67.2% versus 58.2% for patients in AF and 78.3% versus 64.4% for those in atrial flutter (ibutilide and magnesium versus ibutilide only) resulting in a 34% reduction in the need for elective DC cardioversion. The remaining studies were small randomised trials in general medical patients, with patient numbers ranging from 13 to 57 patients [99–104]. The literature on magnesium prophylaxis and non-cardiac-surgical literature on magnesium for AF suggests that magnesium may be of benefit, but there are currently no studies to support the use of magnesium therapy for AF after cardiac surgery.
| Recommendation There is currently no evidence in cardiothoracic patients going into AF that addition of magnesium is of benefit in facilitating the return to sinus rhythm or controlling the ventricular rate.
|
|
12. Ward management of patients developing AF |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
Recommendations
(Grade E recommendations based on expert consensus)
|
|
13. Protocol for the management of AF |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
|
|
Acknowledgments |
|---|
|
References |
|---|
|
Top1. Executive summary2. Introduction3. Scope of the...4. Methodology of the...5. Levels of evidence...6. Definitions of AF7. Prophylaxis for AF...8. Treatment of AF9. Anticoagulation for patients...10. Is antero-posterior position...11. Does magnesium offer...12. Ward management of...13. Protocol for the...References |
|---|
