The study included 331 subjects (101 normal weight, 114 obese, 116 T2DM) with adequate echocardiographic images to determine LA function. Demographic data are shown in table 1. There were significant differences between the groups pertaining to gender and race. The obese and T2DM groups had a higher percentage of females (p = 0.003) and a higher proportion of African American patients (p = 0.01). There was no significant difference in the BMI between the obese and T2DM groups and no significant difference in age among the three groups. Normal weight patients had significantly lower resting heart rates compared to obese and T2DM patients (p < 0.0001). Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were higher in obese and T2DM compared to normal subjects (p < 0.0001). There were no significant differences among the groups in LV shortening fraction or ejection fraction. Indexed LV mass was higher in the obese and T2DM groups compared to the normal group (p < 0.0001).
There were significant differences between the groups in several markers of LV diastolic function. Obese and T2DM patients had significantly higher average E/e’ (p < 0.0001) and lower septal (p < 0.0001) and lateral (p < 0.0001) e’ compared to the normal group. Additionally, there were significant differences in mitral inflow E/A ratios (p < 0.0001) with T2DM patients having the lowest ratio (Table 2). The direction of all these parameters indicates diastolic function abnormalities in obese and diabetic subjects.
We also observed differences between the obese and T2DM groups in the prevalence of abnormal diastolic function (Table 3). Of note, 65% of T2DM subjects were in the 75th percentile or higher for average E/e’ values based on the normal control group compared to 45% of the obese group (p < 0.0001). Similar significant differences were also seen in the lateral and septal e’ values (Table 3).
LA Function Differences Between The Groups:
Obese and T2DM had significantly lower reservoir (p < 0.0001), conduit (p = 0.007) and booster strain (p = 0.0002) compared to normal subjects (Table 2).
Strain rate analysis showed a significantly lower reservoir strain rate in obese and T2DM subjects (p = 0.01). The conduit strain rate was also significantly worse (less negative) in obese and T2DM subjects (p = 0.001). There were no significant differences among the groups in booster strain rate.
There were no differences among the groups in left atrial dimension indexed to body surface area (p = 0.2). However, indexed left atrial volume to body surface area was significantly lower in the obese and T2DM groups compared to the normal group (p < 0.0001) due to the larger body surface area in obese and T2DM groups. There was no significant difference among the groups in atrial emptying fraction (p = 0.3).
Univariate associations of left atrial strain with other echocardiographic measures:
Correlation analysis showed that higher LV mass was associated with lower reservoir and conduit strain (p < 0.0001), reservoir strain rate (p = 0.003), and worse conduit strain rate (p = 0.002). A higher indexed LA volume was associated with lower conduit strain and worse reservoir strain rate (Supplemental data).
Reservoir and conduit strain were positively associated with lateral and septal e’, and mitral E/A ratio. Conduit strain rate was negatively associated with lateral and septal e’, and mitral E/A ratio, while it was positively associated with E/e’ ratio. This indicates that a higher conduit strain rate (less negative, worse) correlated to a lower e’ and thus a higher E/e’ ratio (Supplemental data).
Predictors Of Abnormal Diastolic Function:
Multivariable analysis showed that higher conduit strain rate (less negative) was a significant predictor of being in the worst 25th percentile of mitral E/A (odds ratio (OR), 2.18 ; confidence interval (CI) 1.57–3.03 per unit increase in strain rate, p < 0.0001). Additionally, a lower reservoir strain rate was a significant predictor of being in the worst 25th percentile for septal e’ (OR, 2.15 ; CI 1.24–3.72 per unit decrease in strain, p = 0.006). Importantly, indexed LA volume and LA dimension were not predictors of diastolic dysfunction (Table 4).