Table 3.
Multivariate linear regression of LV circumferential strain and ejection fraction in relation to cardiovascular risk factors
| Ecc (%) |
Ejection fraction (%) |
|||
|---|---|---|---|---|
| Covariate | β ± SE | P value | β ± SE | P value |
| Age (per 10 years of age) | 0.30 ± 0.13 | 0.0236 | 0.80 ± 0.35 | 0.0226 |
| Sex (female vs. male) | −0.44 ± 0.20 | 0.0318 | 1.77 ± 0.54 | 0.0012 |
| Cohort (primary vs. secondary intervention) | −0.15 ± 0.17 | 0.3926 | 0.34 ± 0.46 | 0.4592 |
| Smoking (yes vs. no) | 0.64 ± 0.27 | 0.0188 | −0.16 ± 0.72 | 0.8184 |
| BMI (kg/m2) | 0.04 ± 0.02 | 0.0367 | 0.05 ± 0.05 | 0.3832 |
| Mean diastolic BP (mmHg) | 0.07 ± 0.02 | 0.0002 | 0.02 ± 0.05 | 0.7521 |
| Mean HDL cholesterol (mg/dL) | −0.02 ± 0.008 | 0.0236 | 0.02 ± 0.02 | 0.2383 |
| Mean HbA1c during the DCCT/EDIC study | 0.28 ± 0.10 | 0.0043 | −0.20 ± 0.26 | 0.4343 |
| Myocardial scar | ||||
| Nonischemic vs. none | 1.13 ± 0.58 | 0.0518 | −8.93 ± 1.54 | <0.0001 |
| Ischemic vs. none | −0.04 ± 0.58 | 0.9506 | −3.67 ± 1.53 | 0.0164 |
Only the most significant lipid variable (i.e., HDL cholesterol) was included in the final model. The model was also adjusted for CMR machine type. Positive regression coefficients indicate a decrease in function because Ecc has a negative value. Boldface values indicate significance or clear trend at P ≤ 0.05.