Table 2.
Minimally adjusted linear regression of circumferential strain (Ecc, %) in relation to cardiovascular risk factors
| Covariate | Estimate of β ± SE | P value |
|---|---|---|
| Basic model1 | ||
| Age (per 10 years of age) | 0.26 ± 0.13 | 0.0498 |
| Sex (female vs. male) | −0.86 ± 0.18 | <0.0001 |
| Cohort (primary vs. secondary intervention) | −0.12 ± 0.18 | 0.4906 |
| Minimally adjusted risk factor models2 (basic model + 1 covariate) | ||
| Group (intensive vs. conventional) | −0.18 ± 0.18 | 0.2997 |
| Smoking (yes vs. no) | 0.68 ± 0.27 | 0.0125 |
| BMI (kg/m2) | 0.08 ± 0.02 | <0.0001 |
| BMI ≥30 kg/m2 (yes vs. no) | 0.64 ± 0.19 | 0.0008 |
| Natural waist circumference (cm) | 0.03 ± 0.01 | <0.0001 |
| Mean systolic BP (mmHg) | 0.04 ± 0.01 | 0.0007 |
| Mean diastolic BP (mmHg) | 0.09 ± 0.02 | <0.0001 |
| Hypertension (yes vs. no) | 0.50 ± 0.18 | 0.0055 |
| Mean HDL cholesterol (mg/dL) | −0.02 ± 0.01 | 0.0014 |
| Mean LDL cholesterol (mg/dL) | 0.01 ± 0.004 | 0.0070 |
| Mean total cholesterol (mg/dL) | 0.01 ± 0.004 | 0.0326 |
| Mean triglycerides (mg/dL) | 0.01 ± 0.002 | <0.0001 |
| Hypercholesterolemia (yes vs. no) | 0.21 ± 0.19 | 0.2808 |
| HbA1c (%) | ||
| Before CMR | 0.28 ± 0.07 | 0.0002 |
| Mean during DCCT | 0.14 ± 0.07 | 0.0397 |
| Mean during the EDIC study | 0.37 ± 0.09 | <0.0001 |
| Mean during the DCCT/EDIC study | 0.38 ± 0.10 | <0.0001 |
| Metabolic syndrome (yes vs. no) | 0.81 ± 0.19 | <0.0001 |
| Sustained AER ≥30 μg/min or ESRD (yes vs. no) | 0.61 ± 0.22 | 0.0055 |
| Any PDR (yes vs. no) | 0.58 ± 0.24 | 0.0169 |
| AER ≥300 μg/min or ESRD (yes vs. no) | 0.69 ± 0.41 | 0.0915 |
| LV end-diastolic mass (g) | 0.004 ± 0.004 | 0.2993 |
| LV end-diastolic mass/BSA | 0.001 ± 0.008 | 0.8725 |
| Myocardial scar | ||
| Any scar (ischemic or nonischemic) vs. none | 0.73 ± 0.43 | 0.0904 |
| Nonischemic vs. none | 1.32 ± 0.60 | 0.0274 |
| Ischemic vs. none | 0.14 ± 0.59 | 0.8156 |
| Nonischemic vs. ischemic | 1.18 ± 0.83 | 0.1544 |
| Normalized myocardial T1 (ms) (n = 200) | −0.002 ± 0.003 | 0.5515 |
Positive regression coefficients indicate a decrease in function because Ecc has a negative value. Normalized myocardial T1: postcontrast CMR-derived T1 times were normalized by gadolinium dose, glomerular filtration rate, and infusion-acquisition delay times. Lower values are related to higher diffuse fibrosis. AER, albumin excretion rate; BSA, body surface area; ESRD, end-stage renal disease; PDR, proliferative diabetic retinopathy.
1Basic model was also adjusted for CMR machine type.
2Minimally adjusted model, separate model for each covariate, was also adjusted for age, sex, primary vs. secondary cohort, and CMR machine type.