The Norwegian Atherosclerosis and Childhood Diabetes (ACD) study is a prospective population-based cohort study with follow-ups every fifth year. The cohort consists of subjects with childhood-onset T1D and a healthy control group with similar age and sex distribution.

All examinations were conducted at Oslo University Hospital, Ullevål. Written informed consent was obtained from all participants, and parents of those younger than 18 years, at baseline and at every follow-up.

At baseline, the cohort was recruited among children and adolescents with T1D aged 8–18 years from the South-East Health Region of Norway and registered in the Norwegian Childhood Diabetes Registry (NCDR). The control subjects being mainly friends of participating T1D subjects, with similar age and socioeconomic background. The extended details of the recruitment process and examinations have been well described previously.22

In 2006–2008, a total of 314 children and adolescents with T1D and 120 control subjects aged 8–18 years were included in the study. At the 5-year follow-up, approximately 80% of the baseline cohort participated, together with 53 new control subjects and 15 new T1D subjects, expanding the whole cohort to 502 participants (329 with T1D and 173 controls) (figure 1).

Flow chart demonstrating when the study started, every follow-up and how many individuals were included/lost in the study at baseline and 5 and 10 years of follow-up. T1D, type 1 diabetes. Created in BioRender. Rogn, A. (2025) https://BioRender.com/f33q316

At the 10-year follow-up, conducted in 2017–2019, 59% of the total cohort participated (n=298). Reasons for drop-out were mainly: (a) not managing to reach the subject through available contact information, (b) the subject having moved far away from the study site or abroad due to education or work, (c) lack of time for the subject, and (d) exclusion criteria like other chronic disease, pregnancy, recently given birth (within 3 months), pancreatic transplantation or newly discovered diabetes in control subjects.

The following parameters were assessed at all three time points: serum lipid profile, urinary albumin-creatinine ratio (u-ACR) and HbA1c. Fasting venous blood samples were taken in the morning.

The clinical examination was performed after blood sampling and included height, weight, waist circumference and blood pressure (BP) measurements in a standardized manner.

Information about medical history, current medications and diabetes microvascular complications, as well as family history of diabetes and CVD, was collected by interview and questionnaire.

Data on retinopathy were obtained through an ophthalmologic substudy that was conducted in collaboration with the ophthalmic department at Oslo University Hospital during the 10-year follow-up of the ACD study.

Risk factors analyzed in the present study were based on our previous articles22 and novel publications.23 24

The NCDR is a national childhood diabetes quality registry which registers the incidence of T1D in subjects and collects clinical data from children and adolescents with diabetes annually until the age of 18, on written informed consent.25 At baseline, we compared our T1D participants to all children and adolescents, registered in the registry, and found them to be representative for the whole childhood diabetes population in Norway at that time.22 25 We also received supplementary annual data from the NCDR for all our diabetes study participants until leaving the registry at 18 years of age, covering the years between 2005 and 2012, to investigate annual exposure to known risk factors and covariates such as total cholesterol, high-density lipoprotein-cholesterol (HDL-c), low-density lipoprotein-cholesterol (LDL-c), body mass index (BMI), BP and HbA1c.

Serum glucose, creatinine, and lipid profile were measured using standard methods. HbA1c was determined for all participants by high-performance liquid chromatography (Variant; Bio-Rad, Richmond, California) at the central DCCT-standardized laboratory (the normal reference range 4.0–5.9% (20–41 mmol/mol), and the intra-assay coefficient of variation <3%) at the clinical chemistry department of Oslo University Hospital, Oslo, Norway.

At the 10-year follow-up, cIMT was measured with a standardized protocol and strict quality control in the same manner as previously.22 Two sonographers performed the examinations with Siemens Acuson Sequoia 512 (Siemens Acuson; Mountain View, California) ultrasound scanner equipped with a linear array 14 MHz transducer. All scans were digitally stored on the internal hard drive for subsequent offline analysis. One experienced reader (Mario Gaarder), blinded to the subjects’ clinical details, performed all measurements using M’Ath software.22

Demographic and clinical data are presented as means with their SDs, and medians with the minimum or maximum value. Differences in continuous variables between subjects and controls were tested with the Student’s t-test for normally distributed data or the Mann-Whitney U test for skewed data. Correlation analyses between continuous variables were performed using Pearson’s correlation coefficient (r). Χ2 test for contingency tables was used to detect differences in categorical variables.

Linear mixed-effects models were fitted to cIMT as the outcome variable to account for the repeated measures by subject, missing values, and to adjust for covariates. Time, group (T1D vs controls), covariates, and time-by-group interaction were fixed effects in all models. All models included random intercept and slope, and an unstructured covariance matrix was used. The models were adjusted for the following covariates: age, BMI, HbA1c, total cholesterol, HDL-c, u-ACR and systolic blood pressure (SBP). Covariates that correlated >0.7 were not included in the model to avoid multicollinearity. Based on the linear mixed-effects model, we estimated mean values with 95% CIs for the time points for each group over a 10-year period. We also estimated within-group changes from baseline to 10 years of follow-up, and the between-group changes. To investigate whether the development over time between T1D subjects and controls differed among men and women, we repeated the analysis described above with stratification by sex.

To investigate the association between glycemic exposure and cIMT over time among persons with T1D, linear mixed-effects models were used. Time, glycemic exposure, and time-by-glycemic exposure interaction were fixed effects in all models. The models included random intercepts. The analysis was repeated with stratification by sex.

Annual glycemic exposure over time for T1D participants was calculated by trapezoidal integration of the area under the curve (AUC) from the time of the diagnosis and including our examinations. We derived the total area under the curve of the HbA1c curve (tAUCHbA1c) and the incremental area under the curve of the HbA1c curve (iAUCHbA1c>norm) >6.9% (52 mmol/mol), which considers the time spent above the normal range. Finally, we repeated the analysis described above for tAUC for total cholesterol, LDL-c, HDL-c, SBP and BMI from the time of diagnosis until the age of 18 years.

A significant level of 5% was used. Statistical analyses were performed using the IBM SPSS statistics V.28.0 (IBM SPSS: IBM) and STATA V.17.0 (StataCorp, College Station, Texas).

The datasets generated during and/or analyzed during the current study are not publicly available due to ethical rules of the NCDR that have supported us with additional data, and for reasons of patient privacy. However, some of the data can be available from the corresponding author on reasonable request.