Left Ventricular Wall Thickness and Functional Recovery after Atrial Fibrillation Ablation in Patients with Systolic Dysfunction

Figures

• 

  Figure 1. MDCT-based analyses of LVWT. Contrast-enhanced MDCT images were reconstructed using commercially available software. After automated chamber segmentation, end-diastolic and end-systolic phases were determined using the left ventricular volume–time curve. Endocardial (red) and epicardial (green) contours were generated and manually corrected as needed, and regional LVWT was measured as the mean contour-to-contour distance between the endocardial and epicardial borders. Patient-level mean LVWT and the SD of LVWT were calculated from the 16 end-diastolic segmental mean LVWT values. Representative color-coded bull’s-eye maps with 1-mm increments are shown for tachycardia-induced cardiomyopathy (A) and previous myocardial infarction (B). LVWT = left ventricular wall thickness; MDCT = multidetector computed tomography; SD = standard deviation.

• 

  Figure 2. Segmental left ventricular wall thickness. Segmental distribution of end-diastolic LVWT across the 16-segment model. Segmental values are shown as cohort-level mean (SD). Summary values below the map indicate the median (interquartile range) of patient-level mean LVWT and SD of LVWT. LVWT = left ventricular wall thickness; SD = standard deviation.

• 

  Figure 3. Associations between baseline LVWT metrics and LVEF improvement. Scatter plots illustrate the relationships between baseline LVWT metrics and the ΔLVEF from baseline to 6 months after AF ablation: (A) mean LVWT and (B) SD of LVWT. The solid red lines indicate linear regression lines, and the dottedred lines indicate 95% CIs. Restricted cubic spline analyses show the associations of (C) mean LVWT and (D) SD of LVWT with ΔLVEF. No clear evidence of nonlinearity was observed for either metric, supporting the interpretation of LVWT metrics as continuous markers rather than threshold-based criteria. P overall indicates the overall association, and P nonlinearity indicates departure from linearity. CI = confidence interval; LVEF = left ventricular ejection fraction; ΔLVEF = change in left ventricular ejection fraction; LVWT = left ventricular wall thickness; SD = standard deviation.

• 

  Figure 4. ΔLVEF stratified by LVWT phenotypes. ΔLVEF from baseline to 6 months after atrial fibrillation ablation is shown across 4 baseline LVWT phenotypes defined by the median mean LVWT (7.0 mm) and median SD of LVWT (1.2 mm). Values are median (interquartile range). Differences across phenotypes were significant by the Kruskal–Wallis test (P < .001). In post hoc pairwise comparisons, the reduced-thickness/high-heterogeneity phenotype showed a lower ΔLVEF than the other phenotypes. ΔLVEF = change in left ventricular ejection fraction; LVWT = left ventricular wall thickness; SD = standard deviation.

• 

  Figure 5. Clinical outcomes. Kaplan–Meier curves for (A) the composite of heart failure hospitalization or cardiovascular death and (B) atrial fibrillation recurrence. Patients were stratified by 4 baseline LVWT phenotypes based on the median mean LVWT (7.0 mm; preserved thickness ≥7.0 mm vs reduced thickness <7.0 mm) and the median SD of LVWT (1.2 mm; high heterogeneity >1.2 mm vs low heterogeneity ≤1.2 mm). These categories were used for descriptive phenotyping and should not be interpreted as diagnostic cutoffs. HRs and 95% CIs were estimated using univariable Cox proportional hazards models, with the preserved-thickness/low-heterogeneity phenotype as the reference group. CI = confidence interval; HR = hazard ratio; LVWT = left ventricular wall thickness; SD = standard deviation.