Skip to main content

Table 4 Multivariable regressions of AGE effect on CT calcification outcomes (N = 187)

From: Skin collagen advanced glycation endproducts (AGEs) and the long-term progression of sub-clinical cardiovascular disease in type 1 diabetes

CAC score (quantitative)*

Covariate effects from multiple models

X2

P value

Geometric mean ratio (95 % CI)^

Furosine:

 Unadjusted for A1c

5.7

0.017

5.6 (1.4, 22.7)

 Adjusted for

  DCCT mean A1c

1.3

0.26

3.8 (0.4, 37.3)

  EDIC mean A1c

4.4

0.035

5.0 (1.1, 22.2)

DCCT mean A1c:

 Unadjusted

5.0

0.025

1.9 (1.1, 3.3)

 Adjusted for

  Furosine

0.2

0.68

1.2 (0.5, 3.0)

EDIC mean A1c:

 Unadjusted

0.6

0.52

1.1 (0.7, 1.8)

CAC >0**

Covariate effects from multiple models

t value

P value

Odds ratio (95 % CI)#

DCCT mean A1c:

 Unadjusted

2.2

0.031

1.3 (1.0, 1.6)

EDIC mean A1c:

 Unadjusted

0.8

0.41

1.1 (0.9, 1.5)

CAC >100**

Covariate effects from multiple models

t value

P value

Odds ratio (95 % CI)#

DCCT mean A1c:

 Unadjusted

0.4

0.68

1.1 (0.7, 1.6)

EDIC mean A1c:

 Unadjusted

0

0.71

1.1 (0.6, 2.0)

  1. Italic values indicate statistically significant associations
  2. * Models for CAC mean score were based on a Tobit-censored regression which evaluated the effect of AGE or HbA1c on the outcome of log (CT score)-log (lowest detectable CT score) allowing for unmeasurable values below the lower limit of quantification. Both unadjusted and adjusted models were adjusted for scanning site, baseline age, diabetes duration, sex, and DCCT baseline retinopathy cohort (primary vs. secondary). Adjusted models further included both AGE and HbA1c as risk factors.
  3. ** Models for CAC >0 and CAC >100 were based on a Generalized Linear Mixed Model (GLIMMIX) logistic regression which evaluated the effect of AGE or HbA1c on the prevalence of CAC >0 or CAC >100. Both unadjusted and adjusted models were adjusted for baseline age, diabetes duration, sex, and DCCT baseline retinopathy cohort (primary vs. secondary) with scanning site as a random effect. Adjusted models further included both AGE and HbA1c as risk factors
  4. ^Geometric mean ratio of CAC scores was the ratio of predicted CAC scores for a one unit increase in HbA1c.
  5. #Odds ratio was the ratio of odds for CAC > 0 or CAC > 100 based on a one unit increase in HbA1c.