Five hundred and thirty-five T1D patients using a CGM were followed in the Endocrinology Department of CHUSJ during the study period. Of these, we excluded 61 due to lack on available data and 232 due to having a percentage of active CGM time < 70%. The final number of individuals analyzed was 242 (Supplementary Fig. 1).

The baseline characteristics of the study population are shown in Table 1a, b. Among the cohort of 242 patients, 51.7% were men, the mean age (± standard deviation) was 36.6 ± 12.6 years and 56.7% had higher education. Concerning the T1D related parameters (Table 1a), the mean duration of this disease was 16.8 ± 10.3 years and 35.1% of the patients had an insulin pump. The mean baseline TIR was 56.7 ± 18.2% and the mean baseline CV was 38.5 ± 7.7%. The mean TDD was 52.0 ± 21.7 UI and the mean TDD/kg was 0.7 ± 0.3 UI/kg. Regarding the dyslipidemia related parameters (Table 1b), 69.8% of the patients were not treated with a statin and, of those who were, 5.5%, 45.2% and 49.3%, were treated, respectively, with a low, moderate, and high potency statin. None of these patients was under fibrates. The mean baseline HDL cholesterol level was 57.3 ± 14.5 mg/dL, and the mean triglyceride level was 84.1 ± 78.1 mg/dL.

When comparing baseline characteristics of participants with and without an insulin pump (supplemental Table 3), even though the first group had lower levels of HbA1C (p = 0.020), there were no statistically significant differences found in CGM parameters and lipid profile analysis. Moreover, when comparing baseline characteristics of participants with normal BMI and obesity (supplemental Table 4), the second group had lower levels of HDL cholesterol (p = 0.014) and higher levels of triglycerides (p = 0.001) as it would be expected.

In the cross-sectional evaluation (n = 242), higher TIR levels were associated with lower triglyceride levels (Table 2a). The mean triglyceride levels were 75.5 [57.0, 109.0] mg/dL in patients with TIR < 50%, 65.5 [51.0, 90.0] mg/dL in patients with TIR between 50 and 70% and 64.0 [51.0, 80.0] mg/dL in patients with TIR ≥ 70% (p = 0.006). TIR levels were not significantly associated with total cholesterol, HDL cholesterol, LDL cholesterol or non-HDL cholesterol levels.

Regarding the association of CV with lipid profile in the cross-sectional analysis, lower CV levels were associated with higher HDL cholesterol levels. As CV decreases, HDL levels are higher: 58.4 ± 14.6 mg/dL in patients with CV ≤ 36%, 58.8 ± 15.2 mg/dL in patients with CV between 36 and 45% and 51.8 ± 11.6 mg/dL in patients with CV > 45% (p = 0.035). Total cholesterol, LDL cholesterol, triglycerides and non-HDL cholesterol are not significantly associated with CV levels (Table 2b).

In Table 3, we present the unadjusted and adjusted results of the linear regressions of the association of HbA1c and CGM parameters with lipid profile. Higher TIR levels were associated with lower triglycerides levels in the unadjusted and adjusted analysis. Higher HbA1C, GMI, TAR and TA250 were also associated with higher triglycerides levels in both analyses. The only glycemic parameter associated with HDL cholesterol was CV (−0.30 mg/dL, 95%CI −0.54 to −0.06, per 1% increase in CV in the adjusted analysis, p = 0.015). Neither HbA1c nor CGM parameters were significantly associated with total cholesterol, LDL cholesterol and non-HDL cholesterol.

In the longitudinal evaluation, we included the 90 participants with two different moments of analysis (6 to 18 months apart) (characteristics of these participants are shown in supplementary Table 2A, B). Of these participants, 23 were treated with statins and 17 maintained the same statin treatment between periods. The time between evaluations was 10.5 ± 3.0 months. The mean variation of TIR was 0.1 ± 13.4%, of CV was −1.0 ± 6.8 % and of HbA1c was −0.1 ± 0.7 %. As shown in Table 4, there was a significant association between the variation of CV and the variation of HDL cholesterol levels, both in the unadjusted and adjusted models (−0.54 mg/dL, 95%CI (−0.95 to −0.13), per 1% increase in CV in the adjusted analysis, p = 0.010). The variation of TBR was significantly associated with the variation of HDL cholesterol in both models (−0.46 mg/dL, 95%CI (−0.91 to −0.00), per 1% increase in TBR in the adjusted analysis, p = 0.048). Similarly to the previous analysis, neither HbA1c nor CGM parameters were significantly associated with total cholesterol, LDL cholesterol and non-HDL cholesterol. Differently from the cross-sectional analysis, variation of TIR, HbA1C, GMI, TAR and TA250 were not significantly associated with the variation of triglycerides values.