Skip to main content

Table 3 Univariate and multivariate logistic regression analyses of risk factors of any LA abnormality

From: Incidence of left atrial abnormalities under treatment with dabigatran, rivaroxaban, and vitamin K antagonists

  Separate univariate models Multivariate modela
Clinical variable OR (95 % CI) p value OR (95 % CI) p value
Dabigatran vs. VKA 0.32 (0.07–1.46) 0.141 0.40 (0.08–1.88) 0.245
Rivaroxaban vs. VKA 0.55 (0.19–1.62) 0.28 0.65 (0.22–1.98) 0.453
CHADS2: 2 vs. 0–1 4.40 (1.54–12.54) 0.006 4.07 (1.42–11.69) 0.009
CHADS2: ≥3 vs. 0–1 3.80 (0.89–16.16) 0.071 3.19 (0.73–13.99) 0.124
CHA2DS2-VASC: 2 vs. 0–1 1.99 (0.38–10.55) 0.419   
CHA2DS2-VASC: 3 vs. 0–1 2.03 (0.36–11.41) 0.423   
CHA2DS2-VASC: ≥4 vs. 0–1 6.30 (1.28–30.95) 0.023   
HAS–BLED: 2 vs. 0–1 2.57 (0.87–7.60) 0.089   
HAS–BLED: ≥3 vs. 0–1 3.11 (0.70–13.80) 0.135   
Non-banded LAA vs. banded LAA 0.64 (0.21–1.99) 0.443   
Unknown LAA vs. banded LAA 0.00 (0.00,∞) 0.99   
Beta blocker 1.64 (0.21–12.84) 0.636   
Calcium blocker 0.21 (0.03–1.61) 0.133   
Dronedarone 1.49 (0.32–6.89) 0.61   
Statin 1.24 (0.49–3.13) 0.657   
NSAR 0.77 (0.17–3.47) 0.735   
  1. CI confidence interval, LAA left atrial appendage, NSAR non-steroidal anti-rheumatic agents, OR odds ratio, VKA vitamin K antagonists
  2. aFinal model resulting from an all-subset variable selection based on Akaike’s Information Criterion; the medication group (dabigatran vs. VKA and rivaroxaban vs. VKA) was defined as fixed covariate, and the significant variables in column 1 were considered as possible covariates