**5.1 Propafenone, procainamide, digoxin and calcium channel blockers**

Given the availability of just a few trials with inconsistent results, propafenone is not currently recommended as first-line for POAF prophylaxis (Bradley et al., 2005; Dunning et al., 2006; Fuster et al., 2006; Eagle et al., 2004). Its use may be limited by its proarrhythmic effects in patients with structural heart disease. Current available evidence also does not support the use of procainamide for POAF prophylaxis. Although based on limited evidence, preoperative "digitalization" was historically used to prevent POAF. Currently, digoxin does not have an indication for POAF prophylaxis but can be used for rate control once atrial fibrillation occurs (Bradley et al., 2005). Only the non-dihydropyridine calcium channel blockers (non-DHP-CCB) diltiazem and verapamil, have evidence supporting their effectiveness for POAF prophylaxis from a meta-analysis evaluating twelve small studies encompassing 719 patients (Wijeysundera et al., 2003). However, two other meta-analyses found a non-significant reduction (Andrews et al., 1991) and even an increase in the risk of POAF (Woodend et al., 1998) with the CCBs. Because of this and the risk of atrioventricular block and low-output syndrome, especially in combination with beta-blockers, the guidelines recommend against routine use of CCBs for POAF prophylaxis and that the non-DHP-CCBs, diltiazem or verapamil, be reserved for rate control only once POAF has occurred (Bradley et al., 2005; Eagle et al., 2004).

### **5.2 Thiazolidinediones**

Thiazolidinediones (TZDs) may affect POAF through pleiotropic anti-inflammatory activity against macrophage activation and pro-inflammatory cytokines (Consoli & Devangelio, 2005; Ricote et al., 1998). One study evaluated a nested cohort study of diabetic patients from the AFIST I, II, and III trials (Giri et al., 2001; White et al., 2003; White et al., 2007a) assessed whether the use of TZDs affected the incidence of POAF in diabetic patients who were also receiving beta- blockers and amiodarone (Anglade et al., 2007). In addition to substantial pre- and postoperative beta-blocker use, 43.8% of control patients and 35% of

Strategies for the Prevention of Postoperative Atrial Fibrillation in Cardiac Surgery 223

In a meta-analysis of 10 clinical trials it was demonstrated that atrial pacing at the right atrium, left atrium or Bachman's bundle produced a decrease in atrial fibrillation (Fan et al., 2003). These 10 studies are limited by multiple pacing protocols, including using complex algorithms, fixed pacing and flexible algorithms. Eight of these studies demonstrated that bi-atrial pacing reduced the odds of POAF by 54% (OR=0.46; 95% CI 0.3-0.71). There was a significant lack of use of beta-adrenergic blocking drugs used in the post-operative phase in the meta-analysis at 56%. In a small group of patients (n=80) who underwent valvular surgery it was found that bi-atrial synchronous pacing for 72 hours decreased atrial fibrillation from 45% in the control group to 20% in the paced group (p=0.02) (Debrunner et al., 2004). It is noted that only 30% of this small group were exposed to pre-operative beta-

Pacing of the atria is not without risk. In a randomized trial of 100 patients it was found that atrial fibrillation occurred in 27.5% of the paced patients and 28.6% of the control group (Chung, 2003). There was an increase in atrial ectopy (10 fold increase) in the group of patients whom developed atrial fibrillation (Chung, 2003). It was hypothesized that inconsistent pacing in the atria, under sensing and intermittent loss of capture were factors in the increase in ectopy (Chung, 2003). A sub-analysis of patients paced at a lower rate (80 bpm) and use of an algorithm that maintained the atrial rate 50 ms above the intrinsic rate,

The most recent 2010 European AF guidelines recommend that biatrial pacing should be considered for prophylaxis (Class 2B recommendation based on Level A evidence) (ESC, 2010). Earlier publication in 2006 by EACTS for the guidelines for POAF after cardiothoracic surgery in 2006 (Grade A recommendation based on Level 1B studies) and in 2005 by the American College of Chest Physicians (ACCP) (Strength: B, Evidence: good, Net Benefit: small/weak) both similarly recommend biatrial pacing for prophylaxis (Dunning et al., 2006; Maisel & Epstein, 2005). (Table 1) Specifically, the 2005 ACCP guideline specifically recommends not using unilateral pacing of the right or left atrium. (Strength: I, Evidence: fair, Net Benefit: small/weak) (Maisel & Epstein, 2005). Furthermore, the 2006 EACTS guidelines recommend that temporary pacing should be used in high risk patients receiving beta-blockers and amiodarone for prophylaxis as protection from complications of bradycardia (Grade A recommendation based of Level 1B studies). The CCS guideline also recommends considering atrial pacing with or without a ventricular lead in patients with symptomatic bradycardia (Class 2A recommendation based on Level A evidence) and those patients who are not on a beta-blocker before surgery (Class 2A recommendation based on Level B evidence) (Mitchell et al., 2005a; Kerr & Roy, 2004). Last, the CCS guidelines strongly recommend placing temporary ventricular epicardial pacing electrode wires at the time of surgery to allow for backup pacing as necessary (Class 1 recommendation based on

The pathophysiology of posterior pericardiotomy is based upon adequate drainage of the pericardial space thereby reducing pericardial effusion (Biancari, 2010). Only the earlier European guidelines do include posterior pericardiotomy as a non-pharmacologic option for the prevention of POAF (Grade B recommendation based on Level 1B studies) (Dunning et al., 2006). A recent meta-analysis evaluating 763 patients found that patients who had a posterior pericardiotomy significantly reduced POAF (10.8% versus 28.1%, p=0.003; OR. 0.33, 95% CI 0.16–0.69) and early (6.9% versus 46.2% p<.0001) or late (0% versus 11.3%,

adrenergic blockade, and post-operative use was not collected.

demonstrated no difference in atrial fibrillation rates (Chung, 2003).

Level C evidence) (Mitchell et al., 2005a; Kerr & Roy, 2004).

**6.2 Posterior pericardiotomy** 

TZD patients received amiodarone. Despite this, the study was unable to show a significant reduction in POAF. This may have been due to a lack of power due to small sample size, dilution of effect from concomitant beta-blocker and/or amiodarone use, or increased fluid retention associated with TZD use. In this same analysis, statins did demonstrate a significant reduction in POAF (28% versus 37%, p<0.05). This suggests that the most likely reason TZDs were of no benefit is due to their risk of fluid accumulation thereby attenuating any anti-inflammatory effect (Lertsburapa K, 2008). At this time, TZDs can not be recommended as an option for POAF prophylaxis, either alone or in combination with betablockers.

### **5.3 Triiodothyronine**

The rationale behind the use of triiodothyronine (T3) for POAF prophylaxis lies in the observation that CPB results in a euthyroid sick or low T3 state (Klemperer et al., 1996). The mechanism by which T3 may prevent POAF is unknown (Reichert & Verzino, 2001). Interestingly, it has been shown that POAF is more common in patients with subclinical hypothyroidism when compared to those with normal thyroid function, after adjustments for other variables (Park et al., 2009). One demonstrated that intravenous administration of T3 starting at the time of cross clamp removal significantly decreases the incidence of POAF when compared to placebo (24% versus 46%, p=0.009) (Klemperer et al., 1996). All patients had a left ventricular ejection fraction of less than 40%. While T3 administration was associated with significantly higher postoperative cardiac indices and lower systemic vascular resistance, there was no significant difference in LOS (Klemperer et al., 1995). The authors previously reported data from this same study but included those patients with a history of preoperative atrial fibrillation (Klemperer et al., 1995). In this earlier study, there were no significant differences in the incidence of SVT between the two treatment groups. The authors do not report postoperative beta-blocker use but suggest that because the study population was more ill (ejection fraction <40%), beta-blockade may not be as effective and add-on therapy would be warranted. None of the guidelines currently recommend the use of T3 due to low quality of evidence (Bradley et al., 2005). Until more data becomes available supporting its for POAF prophylaxis, it should not be routinely utilized.
