Page 92 - Read Online

Quin et al. Vessel Plus 2022;6:41 https://dx.doi.org/10.20517/2574-1209.2021.119 Page 3 of 7

terminating episodes is less certain. Patients with a history of preoperative atrial fibrillation were excluded.
This current investigation examined the 10-year rates of AF and death for these two comparative groups. At
the time of a 10-year follow-up, ROOBY patients’ AF and vital status were based upon electronic medical
records data extracted centrally by the VA Perry Point Cooperative Studies Program (CSP) Coordinating
Center’s dedicated nurse coordinators. The VA Corporate Data Warehouse and Medicare databases
provided independent verification for chart abstractions performed; additionally, 10-year deaths were
verified using both VA and non-VA vital status registries. Discrepancies between databases and chart
abstractions were adjudicated by the ROOBY-FS Endpoints Committee (EC); the ROOBY EC included
cardiologists, cardiac surgeons, and the CSP nurse coordinator. Using this combined approach to determine
[6]
survival status, the accuracy rate has been previously reported as > 99% .

Statistical analysis
10-year atrial fibrillation status and survival were compared between surviving POAF versus non-POAF
patients. Dichotomous variables were compared using either chi-squared analysis or Fisher exact tests.
Continuous variables were compared using Student’s t-tests and Wilcoxon rank-sum tests. Kaplan Meier
survival were curves compared between POAF vs. non-POAF groups using log-rank tests to evaluate
statistical significance. Multivariable logistic regression analysis was performed to identify the patient risk
factors associated with 10-year mortality; model eligible variables were screened using P ≤ 0.10. In the
regression to predict 10-year death, the final regression model included age, smoking history, chronic
obstructive pulmonary disease (COPD), peripheral vascular disease (PVD), hypertension, and renal
dysfunction, which was defined as creatinine > 1.5 mg/dL. As this was a ROOBY sub-analysis, the threshold
for statistical significance was pre-established at P ≤ 0.01; however, all P-values and 95% confidence intervals
have been reported for independent interpretation.

Approval for this ROOBY-FS investigation was obtained through the Perry Point Cooperative Studies
Program Coordinating Center (CSPCC), Northport VA Medical Center IRB (Northport, NY), and the
Colorado Multiple IRB for the Rocky Mountain Regional VA Medical Center (Aurora, CO). ROOBY IRB
Protocol No 1657208 (continuing renewal approved 11/2/2021); ROOBY-FS IRB Protocol No. 1657220
(continuing renewal approved 12/14/2021).

RESULTS
Of the 2203 enrolled patients in ROOBY, 100 were excluded from this sub-analysis due to documented
preoperative atrial fibrillation (n = 93) or unknown atrial fibrillation status (n = 7).

Of the 2103 ROOBY patients remaining, 551 patients (26.2%) developed post-CABG new-onset POAF.

Baseline patient characteristics for POAF vs. non-POAF are listed in Table 1. In general, patients who
developed POAF were older and had more medical co-morbidities than non-POAF patients.

At a 10-year follow-up, 69 patients lacked sufficient detail in the electronic medical record to determine
their long-term atrial fibrillation status; these patients were excluded from the current analysis. Among 10-
year survivors, atrial fibrillation rates were higher for the POAF group (18.9%, n = 64/338) as compared to
the non-POAF group (5.8%, n = 61/1048); P < 0.001.

The unadjusted 10-year Kaplan-Meier survival rate for POAF patients was 63%, which was lower than the
70% 10-year survival rate of non-POAF patients [Figure 1]. Factors examined for their association with 10-
year survival included age (OR 1.07; 95%CI: 1.06-1.09), COPD (OR 0.65; 95%CI: 0.51-0.82), a preoperative