Postoperative atrial fibrillation does not impact on overall survival after esophagectomy in patients with thoracic esophageal cancer: results from a randomized, double-blind, placebo-controlled trial

Background: Administration of landiolol hydrochloride was found to be associated with reduced incidence of atrial fibrillation (AF) after esophagectomy for esophageal cancer in our previous randomized controlled trial (RCT). In addition, reduced incidence of AF was associated with reduction of other complications. Meanwhile, the effects of postoperative AF and other complications on long-term survival following esophagectomy are not well understood. Materials and Methods: Between March 2014 and January 2016, 100 patients with esophageal cancer were registered in an RCT trial and randomly allocated to receive either administration of landiolol or a placebo. We analyzed data from this RCT to better understand the effect of postoperative AF and severe associated complications on overall survival (OS) after esophagectomy for cancer. We also examined whether prophylactic administration of landiolol hydrochloride directly affects prolonged survival in patients with esophageal cancer. Results: The five-year rates of OS in the patients with and without AF were 60%, and 68.6%, respectively, there was no significant difference (P = 0.328). Five-year rates of OS of the patients with and without severe complications were 64.6%, and 67.5%, respectively (P = 0.995). The five-year rates of OS in the placebo and landiolol groups were 65.8% and 68%, respectively (P = 0.809). In multivariate analysis, high stage (stage III/IV) alone was an independent prognostic factor for esophageal cancer patients following esophagectomy. Conclusions: New-onset AF and the other severe complications were not associated with poorer long-term survival following esophagectomy. In addition, administration of landiolol hydrochloride after esophagectomy did not contribute to prolonging the OS.


INTRODUCTION
Esophagectomy is considered the optimum treatment against esophageal cancers. Despite improvements in surgical skills and the introduction of minimally invasive surgery, such as thoracoscopic esophagectomy, the overall rate of morbidity remains very high.
New-onset atrial fibrillation (AF) following esophagectomy occurs in 10-40% of patients with esophageal cancer [1][2][3][4]. Postoperative AF may directly increase the risk of thromboembolic events and is also known to indirectly increase rates of other complications, such as anastomotic leakage, pneumonia, and respiratory failure, through organ hypoperfusion [5][6][7]. The incidence of major postoperative complications in our previous study increased in patients that developed new-onset AF following subtotal esophagectomy [7]. Strategies targeting prevention of AF after esophagectomy may therefore be valuable.

Research Paper www.oncotarget.com
Landiolol hydrochloride is an ultra-short acting beta 1-selective blocker that can be administered intravenously. It has a half-life of only four minutes and it has a weaker negative inotropic effect than other intravenous beta 1-blockers [8][9][10]. We performed a randomized, double-blind, placebo-controlled comparative phase III trial (UMIN000010648) to examine if intravenous administration of landiolol hydrochloride is effective in the prevention of AF after esophagectomy. We also examined if reduced incidence of AF may reduce other postoperative complications. Administration of landiolol hydrochloride had association in our previous study with reduced incidence of AF after esophagectomy for esophageal cancer. This was associated with reduction in other postoperative complications [8].
The influences of new-onset AF after esophagectomy on long-term survival have not been widely studied. Two papers reported no significant impact on three-year survival [4,5]; while another two papers considered new-onset AF after esophagectomy to be an independent factor in poor prognosis [11,12]. The effect of postoperative AF on long-term survival following esophagectomy is therefore controversial.
Severe postoperative complications may make patients with esophageal cancer less likely to survive over the long term. Patients with esophageal cancer but without severe postoperative complications have been shown to have better long-term survival than patients with complications [13][14][15]. Elsewhere, no such relationship was shown between severe complications and long-term survival [16][17][18][19].
The current study uses our randomized controlled trial (RCT) data to investigate the effect of postoperative AF on long-term oncological outcomes following transthoracic esophagectomy for cancer. We also evaluate the influence of severe postoperative complications on overall survival (OS) and whether prophylactic administration of landiolol hydrochloride directly influences prolonged survival in patients with esophageal cancer. Table 1 shows patient characteristics. There were no significant differences between the two groups in age, gender, body mass index, percentage of patients receiving neoadjuvant chemotherapy, tumor location, pathological type of tumors, or Tumor-Node-Metastasis (TNM) stage. Table 2 shows surgical outcomes, with no differences between the two groups, which were well balanced.

Postoperative complications
New-onset AF was noted in 15 patients in the placebo group and in five patients in the landiolol group (Table 2). There were significant differences (P = 0.012). Overall incidence of postoperative complications of more than Clavien-Dindo Grade IIIa were significantly higher in the placebo group (32%) than in the landiolol group (8%) (P = 0.003).

Recurrence of tumors
Median follow-up periods in the placebo and landiolol groups were 60.5 months and 58 months, respectively. The percentage of patients receiving adjuvant chemotherapy was similar between the two groups. Fourteen patients in the placebo group and 13 patients in the landiolol group had recurrence of esophageal cancer (P = 0.822, Table 3). Table 3 shows the details of the recurrent patients. These 27 patients received chemotherapy, chemoradiotherapy or palliative surgery, and four patients survive at the time of writing. Table 4 shows the recurrence rates and mortality rates in the patients with stage I and II. When limited to the patients with stage I and II, the recurrence rates in the patients with and without AF were 15%, and 9%, respectively (P = 0.611). However, overall mortality rate was significantly higher in the patients with AF (39%) than in the patients without AF (14%) (P = 0.046). Figure 1 shows Kaplan-Meier curves of OS stratified by TNM stage. The five-year rates of OS in patients with stage I, II, and III / IV were 85.9%, 76.2%, and 46.5%, respectively (P = 0.001). Figure 2 shows Kaplan-Meier curves of OS stratified by the presence or absence of postoperative AF. The five-year OS rates in all patients with and without AF were 60%, and 68.6%, respectively, there were no significant differences in the OS rates between the two groups ( Figure 2A, P = 0.328). In stage I/II stratification, there were significant differences in the rates of OS between the two groups ( Figure 2B, P = 0.045), but in stage III / IV stratification there were no significant differences ( Figure 2C, P = 0.929). Figure 3 shows Kaplan-Meier curves of OS stratified by the presence/absence of postoperative complications (more than Clavien-Dindo grade IIIa).

Survival outcomes
Five-year rates of OS in the patients with and without postoperative complications were 64.6%, and 67.5%, respectively. There was no significant difference in the rates of OS between the two groups (P = 0.995). Figure 4 shows intention to treat Kaplan-Meier curves of OS between the placebo and the landiolol groups. Five-year OS rates in the placebo and landiolol groups were 65.8%, and 68%, respectively. There was no www.oncotarget.com

Univariate and multivariate Cox proportional hazard model analysis for OS
Univariate and multivariate OS analyses were calculated using Cox proportional hazard regression model. In univariate analysis, 3 of the 13 factors decreased OS in patients with esophageal cancer following esophagectomy ( Table 5, P < 0.10). The factors were diabetes, patients who underwent curative resection, and patients with TNM stage III/IV. Multivariate analysis revealed that among patients with esophageal cancers following esophagectomy, only the patients with TNM stage III/IV (P = 0.001; hazards ratio, 3.413) were independent associative factors for decreased OS (Table 5).

DISCUSSION
In our prospective RCT, administration of landiolol was associated with reduction in incidence of new-onset AF following esophagectomy for esophageal cancer. Furthermore, reduction in the incidence of AF related to landiolol administration was associated with reduction in the incidence of the other severe postoperative complications [8].
New-onset AF was unrelated, however, to poorer long-term survival after esophagectomy. When limited to stage I and II patients with esophageal cancer, the occurrence of AF after esophagectomy may have an association with poor survival. Therefore, the prophylactic administration of landiolol hydrochloride may prolong the overall survival in patients with stage I and II. Contrary to previous reports [13][14][15], overall severe complications were not associated with long-term survival of these patients. This study refuted the hypothesis that administration of landiolol hydrochloride leads to improvement of long-term oncological outcomes after esophagectomy for esophageal cancer. Only high stage of cancer (stage III/IV) was an independent prognostic factor for patients with esophageal cancer following esophagectomy.
New-onset AF after esophagectomy is due to the preexisting cardiac substrate conducive to the activation of AF. This reflects long-standing cardiac remodeling Abbreviations: AF, atrial fibrillation. and represents overall cardiovascular risk independent of acute onset of postoperative AF [12,20]. Postoperative AF may therefore be correlated with long-term survival following esophagectomy [11,12]. This hypothesis was not substantiated in the current study, however, which included patients in all stages. Patients with AF after esophagectomy may have increased risk of chronic AF in the future, and may have increased post-operative mortality, such as by thromboembolism, stroke or heart failure [12,20]. In the current study, however, none of the patients who had new-onset AF after esophagectomy developed chronic AF. To better understand the relationship between new-onset AF after esophagectomy and long-term survival, clarification by large-scale multicenter clinical trials is required.
The influence of severe postoperative complications on the long-term survival of patients with esophageal cancer remains controversial [13][14][15][16][17][18][19]. In our results, there were no differences in survival rates between patients with and without severe complications. Severe complications, such as anastomotic leakage, pulmonary complications and bilateral vocal code paralysis, may impact on shortterm surgical outcomes through a pathway of systemic inflammatory response. Long-term survival, however, may not be affected after the short-term surgical effect has disappeared [16]. There were no mortalities during this RCT, which was based in a high-volume center. In addition, the rate of severe complication (more than Clavien-Dindo grade IIIa) was low at 20%. All patients went through the same standardized multidisciplinary team process and were treated in a high-volume center. In general, patients with esophageal cancer who undergo esophagectomy in high-volume centers have been shown to have improved short-and long-term outcomes [15].
Invasion due to esophagectomy induces an increase in inflammatory cytokines. Perioperative inflammation is generally accepted to be related to tumor recurrence [8,21,22]. Landiolol hydrochloride has strong antiinflammatory effects, it therefore not only reduces postoperative complication rates, but may also help to reduce tumor recurrence after esophagectomy [8]. Indeed, beta-blockers in general have been reported to improve prognosis for various cancers [23][24][25][26][27]. The antitumor effect exerted by beta-blockers may be mediated by suppression of the sympathetic nervous system through blocking of catecholamines [22]. This RCT has already proven a significantly lower serum IL-6 level in the landiolol group than in the placebo group during the early postoperative period [8]. In addition, landiolol hydrochloride may inhibit attachment of circulating tumor cells to vascular endothelial cells [21,28]. In this study, however, the recurrence and survival rates were similar with and without administration of landiolol. In addition, there was no significant difference in the rates of relapse-free survival between the two groups (logrank P = 0.661, data not shown). In our study, 3 μg/kg/ min dose of landiolol was administered only during the 72 hr after esophagectomy. In previous papers, beta blockers improved prognosis for various cancers, but were administered over a long-term period [23][24][25][26][27]. To confirm whether prophylactic administration of landiolol directly influences prolonged survival in patients with esophageal cancer after esophagectomy, the dose and the duration of landiolol require examination.
This study had several limitations. First, due to the relatively small sample size (n = 100), findings from this study could not establish definitive evidence. It was also a single-center study. Multi-center large scale prospective RCT to evaluate whether intravenous administration of landiolol hydrochloride affects long-term survival after esophagectomy for esophageal cancer is therefore required. A second limitation is that the median follow-up period in this study was 60 months, so a longer follow-up period is needed. Thirdly, subsequent treatment after esophageal cancer recurrence was not defined, so differences in treatment after recurrence may have affected survival analysis.
In conclusion, new-onset AF and other severe complications were not associated with poorer long-term   survival after esophagectomy. In addition, administration of landiolol hydrochloride after esophagectomy did not contribute to prolonged OS of patients with esophageal cancer.

Study population
Participants of this study were patients with diagnosed esophageal cancer at the Wakayama Medical University Hospital (WMUH). Detailed inclusion criteria are described in our previous paper [8]. This was a randomized, double-blind, placebo-controlled comparative phase III trial [8], conducted in accordance with a protocol reviewed and approved by the WMUH Ethical Committee on Human Research (approval number 1235). The study protocol was registered at the University Hospital Medical Information Network (UMIN000010648).

Randomization and blinding
Between March 2014 and January 2016, the 100 patients registered in the study were randomly allocated to receive administration of either landiolol or a placebo. After transthoracic esophagectomy with systematic lymphadenectomy, patients were randomized on postoperative day 1 in accordance with the study protocol. Infusion of landiolol hydrochloride (Ono Pharmaceutical, Osaka, Japan) was started at 3 μg/kg/min for 72 hours. Glucose solution was administered at 3 mL/hr during the same period. Patients, medical staff and investigators were all blinded to the treatments except one clinical research coordinator in the central registry and two selected pharmacists in WMUH who were aware of the treatments.

Study endpoints
The primary end point of this RCT was the incidence of AF after esophagectomy. Secondary endpoints were incidence of all postoperative complications and OS.
AF was defined as an absent P wave before the QRS complex, with irregular ventricular rhythm shown on 12-lead electrocardiogram [3,7,8]. Occurrence of postoperative AF was defined as persistence of this arrhythmia for five minutes or more. Diagnosis of AF was re-confirmed by cardiologists. Postoperative complications were analyzed according to Clavien-Dindo classification [29].

Follow-up and additional treatments
For stage I/II patients who underwent R0 resection, follow-up with semiannual endoscopy and abdominal computed tomography were performed. We used adjuvant chemotherapy for patients that were higher than stage III.
For patients who had undergone R1 resection, additional chemoradiotherapy was recommended, but we chose the same kind of follow-up for cases where this was considered to be impossible because of severe coexisting diseases. Subsequent treatment was not defined in the case of recurring esophageal cancer. SPSS 22.0 software program (SPSS Inc., Chicago, IL) was used for all statistical analyses. Quantitative results are expressed as medians and ranges. Statistical comparisons between the placebo group and the landiolol group were performed with χ 2 statistics and Fisher's exact test; P < 0.05 was considered significant. OS was defined as the time from the esophagectomy to the date of death from any cause. Survival curves were generated using the Kaplan-Meier method and compared using the logrank test; P < 0.05 was considered significant. Univariate and multivariate Cox proportional hazard models were used to evaluate factors that independently affected OS. Factors associated with univariate P < 0.10 were included in the multivariate analysis, while factors associated with multivariate P < 0.05 were defined as independent factors associated with OS.

ACKNOWLEDGMENTS
We acknowledge proofreading and editing by Benjamin Phillis at the Clinical Study Support Center, Wakayama Medical University.

Ethics statement
Informed consent was obtained from all patients in accordance with the guidelines of the Wakayama