Table 2.
Arterial outcome variable | Ridge regression penalisation (λ) | SDCGM (st.β, 95% CI) | p value | MSGCGM (st.β, 95% CI) | p value |
---|---|---|---|---|---|
cf-PWV, SD (n = 643) | λ = 0.11 | 0.065 (−0.018, 0.167) | 0.160 | 0.059 (−0.043, 0.164) | 0.272 |
carDC, SD (n = 725) | λ = 0.12 | −0.003 (−0.097, 0.092) | 0.952 | 0.088 (−0.014, 0.184) | 0.102 |
cIMT, SD (n = 726) | λ = 0.12 | −0.007 (−0.123, 0.111) | 0.916 | 0.078 (−0.038, 0.207) | 0.198 |
ABI, SD (n = 816) | λ = 0.11 | −0.033 (−0.071, 0.002) | 0.060 | −0.008 (−0.032, 0.017) | 0.548 |
CWSmean, SD (n = 725) | λ = 0.12 | −0.059 (−0.169, 0.066) | 0.318 | 0.082 (−0.044, 0.204) | 0.180 |
CWSpuls, SD (n = 725) | λ = 0.12 | −0.045 (−0.145, 0.053) | 0.374 | 0.042 (−0.055, 0.138) | 0.410 |
Associations were adjusted for age, sex, educational level, BMI, smoking status, alcohol use, total-to-HDL-cholesterol levels, use of antihypertensive and lipid-modifying drugs, and the other CGM-assessed index. Further, cf-PWV was additionally adjusted for MAP and heart rate; carDC and CWSpuls were additionally adjusted for MAP; cIMT was additionally adjusted for office systolic BP; ABI was additionally adjusted for office systolic BP and heart rate; and CWSmean was additionally adjusted for brachial pulse pressure. All coefficients were estimated by use of ridge regression. Point estimates and 95% CIs were calculated by use of 1000 bootstrap estimates
Standardised regression coefficients (st.β) indicate the median difference (95% CI) associated with 1 SD higher SDCGM or MSGCGM
In the cf-PWV study population, 1 SD corresponds to 0.57 mmol/l for SDCGM, 1.3 mmol/l for MSGCGM, and 2.2 m/s for cf-PWV. In the carDC, cIMT, and CWS study populations, 1 SD corresponds to 0.57 mmol/l for SDCGM, 1.3 mmol/l for MSGCGM, 6.0 10−3/kPa for carDC, 152.7 μm for cIMT, 10.2 kPa for CWSmean, and 6.6 kPa for CWSpuls. In the ABI study population, 1 SD corresponds to 0.56 mmol/l for SDCGM, 1.3 mmol/l for MSGCGM, and 0.11 for ABI