Introduction

Postoperative atrial fibrillation (POAF) occurs in 13–40% of patients after open heart surgery and is associated with prolonged hospitalization, stroke, increased economic costs, and mortality [1,2,3,4]. Oral β-blockers are recommended to treat POAF after cardiac and thoracic surgery [5]. The 2017 European guidelines recommend perioperative oral β-blocker therapy as a Class IIa indication to prevent POAF [6]. In contrast, Japanese guidelines recommend statins to prevent POAF after cardiac surgery [7]. The effectiveness of antiarrhythmic drugs other than β-blockers to prevent POAF is not well established as class IIb recommendation in the Japanese guideline [7].

The transdermal patch comprising the selective β1-blocker bisoprolol (Bisono® Tape; Toa Eiyo, Tokyo, Japan) became commercially available in Japan in 2013. The first-pass effect is lower and drug bioavailability is higher with the transdermal administration of bisoprolol than that observed with oral administration of the drug [8]. Phase 2 clinical trials have shown that plasma concentrations of bisoprolol remained more steady with a lesser tendency to show spikes in concentration with the use of a transdermal bisoprolol patch (TBP) than with oral bisoprolol [9]. Godo et al. showed lower peak and higher trough concentrations of bisoprolol in users of the transdermal patch than that in users of the oral preparation [10]. Its efficacy in preventing POAF as well as hypotension and bradycardia is attributable to the stabilized blood concentration of bisoprolol available in the transdermal patch.

In this study, we compared the incidence of POAF between users of TBP and oral bisoprolol fumarate (OBF) in patients undergoing cardiac and/or thoracic aortic surgery.

Methods

Patients and methods

We performed a retrospective study based on our hospital database. Between April 2016 and February 2018, a total of 216 patients underwent cardiac and/or thoracic aortic surgery at Nerima Hikarigaoka Hospital. We excluded patients with a history of atrial fibrillation, those not using postoperative β-blockers, those who did not undergo cardiopulmonary bypass, those who underwent endovascular treatment, and those with perioperative extracorporeal membrane oxygenation use. Eventually, 108 patients were included in this study (Fig. 1). We routinely prescribed bisoprolol for the prevention of POAF postoperatively (if β-blocker use was not contraindicated). Between April 2016 and April 2017, patients were administered OBF, and after May 2017, patients were administered TBP instead of OBF. In this study period, the indication for surgery and perioperative patient care remained unchanged. We compared preoperative patient characteristics and in-hospital outcomes between 59 patients administered OBF and 49 patients administered TBP to evaluate differences in the effect of β-blockers. Patients’ preoperative characteristics and intraoperative variables in both groups are shown in Tables 1 and 2. POAF was defined as the new onset of postoperative atrial fibrillation lasting > 30 s recorded using a telemetry system throughout hospitalization. Average observation period identical to postoperative hospital stay was 23 ± 14 days (median 19 days). When POAF occurred, nurses reported to a doctor who finally confirmed the diagnosis of POAF. Blood pressure and heart rate were measured preoperatively, upon admission to the intensive care unit, and postoperatively. Measurements later than postoperative day (POD) 1 were performed at 6 a.m. Good pain control was achieved using oral or intravenous analgesia. Other anti-arrhythmic agents including landiolol were postoperatively used for the treatment of any arrhythmia, not before POAF occurred, in 17 patients (29%) in the OBF group and in 8 patients (16%) in the TBP group without significant intergroup difference (p = 0.170). Postoperative arrhythmia except for atrial fibrillation was basically treated with amiodarone. Thoracic epidural analgesia was not used. Follow-up data including survival and the incidence of atrial fibrillation were obtained via a patient chart review. The mean follow-up period was 16 ± 9 months. The Institutional Review Board of our hospital approved the use of patient data for this retrospective study.

Fig. 1
figure 1

Flow diagram shows the study population. ECMO extracorporeal membrane oxygenation

Full size image
Table 1 Preoperative patient characteristics per study group
Full size table
Table 2 Operative data
Full size table

Administration of β-blocker

The administration of β-blocker commenced on POD 1 and was continued throughout admission. The OBF dosage was 2.5 mg per day, and TBP dosage was 4 mg (equivalent to 2.5 mg of OBF) [11]. The dosage of β-blocker was adjusted based on age, cardiac function, renal function blood pressure, and heart rate in individual patient. As a result, a total of 34 patients (32%) had lower dosage of β-blocker.

The dosage of bisoprolol was down-titrated in 22 patients (37%) in the OBF group and 12 patients (25%) in the TBP group without significant difference (p = 0.212). The mean % dosages to desirable dosage (i.e., actual mean dosage/4 mg in the transdermal group and actual mean dosage/2.5 mg in the oral group) were 86.2 ± 23.4% in the transdermal group and 81.8 ± 39.6% in the oral group, respectively (p = 0.471). The rates of down-titration or discontinuation of bisoprolol during hospitalization were 51% (30 patients) in the OBF group and 39% (19 patients) in the TBP group without significant intergroup difference (p = 0.247). The β-blockers were administered at 7 a.m. If hypotension (systolic blood pressure < 90 mmHg) or bradycardia (heart rate < 60 beats per minute) occurred, the β-blocker was down-titrated or discontinued. Other β-blockers were not used concomitantly. Medications other than β-blockers that had been administered preoperatively were re-initiated once oral intake was possible.

Statistical analysis

The data in this study were expressed as mean ± standard deviation for continuous variables and as frequencies and percentages for categorical variables. Patient characteristics and echocardiographic data were compared between groups using the Fisher’s exact test for categorical variables and the unpaired Student’s t test for continuous variables. Prolonged vasopressor administration was defined as use of vasopressor later than postoperative day 1. High dose of diuretics was defined as postoperative use of oral furosemide of 40 mg or more daily.

Logistic regression analysis was performed to identify predictors of POAF. Variables included age, sex, body mass index, the New York Heart Association class 3/4, hypertension, diabetes, dyslipidemia, ischemic heart disease, stroke, chronic obstructive pulmonary disease, chronic kidney disease (serum creatinine level > 1.5 mg/dl), hemodialysis, left ventricular ejection fraction, left atrial diameter, brain natriuretic peptide level, preoperative medication used, type of surgical procedure performed, and use of TBP. Variables with p values < 0.25 using univariate analysis were subjected to multivariable logistic regression analysis. Overall survival after discharge was estimated using the Kaplan–Meier method, and comparisons between two groups were performed with log-rank test. All analyses were performed using the SPSS software version 25 (IBM Corp., Armonk, NY, USA). A p value < 0.05 was considered statistically significant.

Results

Patient characteristics and operative data in the groups administered transdermal and the oral bisoprolol

As shown in Table 1, the likelihood of hypertension in patients administered TBP was higher than that in patients administered OBF. No intergroup difference was observed in variables including age and renal dysfunction, which are reportedly associated with the incidence of POAF. Preoperative medication use did not differ between groups. Operative data are shown in Table 2. No intergroup differences were observed in type of surgical procedures, cardiopulmonary bypass, and aortic cross-clamp times.

Clinical outcomes

Early outcomes are presented in Table 3. Overall hospital mortality was 0% in the transdermal and 3.4% in the oral group, and this difference was statistically non-significant. Among the two deaths in the oral group, one patient had POAF. The incidence of POAF was 24.5% (12 patients) in the transdermal and 45.8% (27 patients) in the oral group. Among all cases, the rates of POAF was 35% (14/40) in isolated valve surgery, 33% (1/3) in isolated coronary artery bypass grafting (CABG), 35% (18/52) in isolated aortic surgery, 63% (5/8) in valve surgery and CABG, 0% (0/1) in valve and aortic surgery, and 25% (1/4) in others. Regarding the timing of onset of POAF, the median interval between surgery and the onset of POAF was 4 days, and this intergroup difference was statistically non-significant. Among patients with POAF, 69% (27 patients) developed POAF within 5 days postoperatively. In our study, the treatments for POAF included intravenous amiodarone, pilsicainide or landiolol. No intergroup difference was observed in stroke rates, re-exploration for bleeding, prolonged ventilator use, and intensive care unit stay. As shown in Table 3, no intergroup difference was noted in the rates of prolonged vasopressor administration, use of temporary pacing, high dose of diuretics and discontinuation of β-blocker. Regarding the effect of discontinuation/tapering of β-blocker on the incidence of POAF, the rates of POAF were similar between the normal dose and discontinuation/tapering dose groups (normal dose vs. discontinuation/tapering dose, 41% (24/59) vs. 31% (15/49), p = 0.318). In 16 patients who discontinued β-blocker, POAF occurred in 25% (2/8) in the transdermal group and in 63% (5/8) in the oral group without significant intergroup difference (p = 0.315).

Table 3 In-hospital outcomes
Full size table

Blood pressure and heart rate

Blood pressures and heart rates were compared preoperatively, upon admission to the intensive care unit, and on PODs 1, 2, 3, 5, and 7. No intergroup differences were observed in blood pressures and heart rates.

Predictors of postoperative atrial fibrillation

We assessed the independent risk factors for POAF using a multivariable logistic regression model. Multivariable analysis showed that the use of TBP was an independent predictor for reduced occurrence of POAF (odds ratio 0.21, 95% confidence interval 0.05–0.84, p = 0.027). Other predictors included coronary artery bypass grafting, and hypertension, as shown in Table 4.

Table 4 Predictors of postoperative atrial fibrillation
Full size table

Follow-up data

Overall survival was similar between the groups at average 16 months of follow-up (p = 0.485). One-year survival was 100% in the transdermal group and 94.6% in the oral group. Paroxysmal atrial fibrillation occurred in 1 patient (2.0%) of the transdermal group and in 2 patients (3.4%) of the oral group at follow-up (p = 1.000). Of them, one patient in the oral group underwent catheter ablation 3 months after surgery.

Discussion

The principal finding in this study was that a TBP was associated with lower rates of POAF in patients undergoing cardiac and/or thoracic aortic surgery. Regarding negative inotropic effects, postoperative blood pressure and heart rate were similar between the TBP and the OBF groups.

Reportedly, several risk factors including advanced age, hypertension, obesity, low left ventricular ejection fraction, renal dysfunction, and left atrial volume are associated with POAF after open heart surgery [2, 12,13,14]. Among various risk factors, activation of the sympathetic nervous system plays an important role in the pathophysiology of POAF. Sympathetic stimulation increases intracellular calcium, which promotes automaticity and triggered activity. Increased sympathetic activity prior to onset of AF has been observed [5]. β-blocker prevents POAF by blocking sympathetic tone. Yokota et al. showed that preoperative use of β-blocker was an independent predictor of POAF [15]. However, in our study, preoperative use of β-blocker was not associated with the incidence of POAF. Multivariable analysis in our study showed that non-use of the transdermal patch, coronary artery bypass grafting, and hypertension were independent predictors of POAF.

Reportedly, perioperative administration of oral β-blockers prevents POAF after open heart surgery [16,17,18]. The American College of Cardiology Foundation/American Heart Association guidelines recommend the use of oral β-blockers and amiodarone for the prevention of POAF after coronary artery bypass grafting [19]. However, regarding routine use of amiodarone for prevention of POAF, we had concerns about adverse effects of amiodarone. Hillis et al. showed that POAF typically occurs within 5 days postoperatively, with a peak incidence on POD 2 [19]. In our cohort, 69% of POAF cases occurred within 5 days postoperatively. No intergroup difference was observed in the timing of POAF.

Several studies have reported the effectiveness of intravenously administered β-blockers for the treatment and prevention of POAF. Sakamoto et al. showed that the conversion of POAF to sinus rhythm in patients administered landiolol (a short-acting intravenously administered β-blocker) was more common and quicker than that in patients administered diltiazem. Additionally, the risk of heart failure related to landiolol was lesser than that related to diltiazem [20]. Sezai et al. showed that 48-h administration of intravenous landiolol in patients undergoing coronary artery bypass grafting prevented POAF [21]. They also reported that landiolol prevented POAF in patients with left ventricular dysfunction undergoing cardiac surgery [22]. Their study showed that landiolol was associated with lower levels of brain natriuretic peptide and a shorter hospital stay, without requiring discontinuation of the drug for hypotension or bradycardia [22]. Perioperative intravenous administration of β-blockers has been shown to be effective and safe in the prevention of POAF. However, cost-effectiveness of intravenous β-blockers remains a concern. In addition, the superiority of intravenous β-blocker over oral β-blocker has not been clarified in terms of the prevention of POAF.

Several reports in the literature describe the efficacy and safety of a TBP in patients with hypertension and heart failure. Momomura et al. have reported that switching from oral bisoprolol to the transdermal patch was safe and effective in patients with chronic heart failure [23]. Only a few reports have described the effect of a transdermal patch on clinical outcomes after cardiac surgery; however, our data showed that the transdermal patch could be safely administered similar to oral bisoprolol in this group of patients.

The perioperative administration of TBP scores over oral bisoprolol because: (1) The plasma concentration–time profile of the transdermal patch is more stable than that of oral β-blockers and prevents rapid changes in blood pressure and heart rate. (2) A transdermal patch can be easily removed if hypotension or bradycardia occurs. Although the blood level of a transdermal patch can be sustained after removal, it may be possible to remove the patch before the blood concentration reaches the maximum blood level as the absorption of β-blocker in patients with the patch is slower compared with oral type [9, 24]. (3) Heart failure-induced edema of the intestines can block the gastrointestinal absorption of orally administered β-blockers. Some authors have reported that use of cardiopulmonary bypass was associated with gastrointestinal injury and edema formation [25, 26]. The percutaneous absorption of bisoprolol incorporated in the transdermal patch may overcome this drawback. (4) Transdermal bisoprolol is safely administered even in patients with impaired swallowing.

Study limitations

Limitations of our study are: (1) this was a single-center study involving a small sample size. (2) Patients included in this study underwent various types of surgeries, which could affect the incidence of POAF, although no intergroup difference was observed in the types of surgeries performed. (3) The retrospective study design serves as a drawback. Randomized control studies are required to confirm the superiority of the transdermal patch over oral bisoprolol.

Conclusions

The incidence of new-onset POAF in patients administered TBP was lower than that in patients administered oral bisoprolol after cardiac and/or thoracic aortic surgery. The TBP patch is an effective and safe β-blocker drug delivery system to prevent POAF.