Page 94 - Read Online

Quin et al. Vessel Plus 2022;6:41 https://dx.doi.org/10.20517/2574-1209.2021.119 Page 5 of 7

POAF had a significantly lower 10-year AF rate of 5.8%. Although the 10-year survival rate of 63% among
the POAF patients was also significantly lower than the 70% rate for non-POAF patients, this was unlikely
to have been directly related to POAF because it was not found to be independently associated with 10-year
risk-adjusted survival using multivariable regression analysis.

These findings appear consistent with the literature examining both the short and long-term incidence of
AF and survival among patients with POAF. The incidence of POAF after CABG, approximately 20%-30%,
does not appear to have changed significantly over time [7-10] . Patients who developed POAF seem to have
higher subsequent rates of AF, as was seen in the current investigation.

While the association of POAF with long-term AF appears consistent, the association with long-term
survival is less certain. Much of the literature seems to support an association of POAF with reduced long-
[8]
term survival. A retrospective study by Thorén and colleagues of 7145 CABG patients (1996 to 2012)
followed over a median follow-up of 9.8 years, found an AF rate for the POAF group of 16.2% vs. 5.7 % for
non-POAF patients and an association of POAF with overall mortality (HR 1.16; 95%CI: 1.06-1.27), cardiac-
specific mortality (HR 1.27; 95%CI: 1.10-1.47), and cerebrovascular mortality (HR 1.39; 95%CI: 1.04-1.86).
Similarly, meta-analyses show an association between POAF and mortality. Included is an investigation by
Eikelboom and colleagues showing a 10-year mortality rate of 29% among POAF patients (n = 9,389) vs.
[10]
[11]
23% among non-POAF patients (OR 1.5, 95%CI: 1.4-1.6). Another large-scale meta-analysis of 61 studies
(including 188,191 CABG patients) showed mortality rates of 18.6% vs. 13.3% comparing POAF patients vs.
no POAF patients, respectively.

However, in other studies, an association of POAF with late AF and reduced long-term survival was not
seen. Taha and colleagues’ Swedish registry study of CABG patients (2007 to 2015, n = 24,523) showed no
association of POAF with all-cause mortality; adjusted mortality rates were similar to those of non-POAF
patients at a median follow up of 4.5 years (adjusted hazard ratio 1.08; 0.98-1.18) . Similarly, our
[12]
investigation, which examined whether there was an independent association between POAF and survival,
did not find one. Rather, the association may have been confounded by other patient factors, such as
[13]
diabetes, HTN, and COPD, which are more prevalent among patients who develop POAF . In such cases,
preoperative identification of these patients and heightened attention to their postoperative management
might avert POAF in these cases. Strategies to do so include may include the use of beta-blockers, correcting
[14]
electrolyte imbalances, and judicious use of inotropes . As well, for patients who do not possess
traditionally known risk factors of POAF, research examining serologic markers may provide a useful
means of preoperatively identifying patients who are at higher risk for POAF and therefore may benefit
[15]
from closer attention during their postoperative recovery.

Limitations of this ROOBY 10-year follow-up investigation include the inherent challenges related to its
study’s retrospective, observational design. As the 10-year cause of death was not assessed, there was no
ability to differentiate noncardiac deaths from cardiac-related causes. Also, death could not be used as a
competing event for the comparison of POAF versus non-POAF evaluations of 10-year atrial fibrillation
rates. As many risk factors for survival were more prevalent in POAF patients, the 10-year survival
comparison between POAF versus non-POAF patients may have been biased. We did not have sufficient
granularity to determine the effect of diabetes or hypertension control. As well, the survival curves
presented are not risk-adjusted. Unfortunately, data on the use of anti-thrombotic agents, which may
impact survival during the 10-year follow-up period, were not collected. Finally, the ROOBY trial’s
population consisted primarily of male veterans with multiple co-morbidities; thus, these study results may
not be generalizable to women or non-Veteran populations. As regards the former, however, it is noted that