We retrospectively investigated the medical records of 91 T1DM pediatric patients who underwent 24-hour ABPM between January 2009 and February 2010 at Seoul National University Children's Hospital (Seoul, Korea). Patients were younger than 19 years of age at evaluation and had participated in a previous ABPM study conducted in our hospital [5]. Among them, 10 patients who were confirmed to have microalbuminuria or retinopathy at their ABPM evaluation were excluded. A total of 81 patients (33 males and 48 females) with T1DM who were regularly followed for 8 years at Seoul National University Children’s Hospital outpatient clinic were included.

Laboratory tests included glycated hemoglobin A (HbA1c), lipid profile (total cholesterol, triglycerides, low-density lipoprotein [LDL] cholesterol, high-density lipoprotein [HDL] cholesterol), 24-hour urine microalbumin excretion, albumin-to-creatinine ratio, free thyroxine, thyroid-stimulating hormone, and triiodothyronine. Data on T1DM duration, height, weight, and age at diagnosis were collected. Mean HbA1c level over the total T1DM duration was recorded from patient files. Blood samples were obtained by venipuncture the morning after an 8-to 12-hour fast. Levels of HbA1c, lipids, free thyroxine, thyroid-stimulating hormone, and triiodothyronine were measured. Some patients could not fast for 12 hours because of risk of hypoglycemia. Total cholesterol, LDL cholesterol, and HDL cholesterol levels were measured by enzymatic calorimetry. The albumin-to-creatinine ratio (μg/mg) or 24-hour urine microalbumin excretion (mg/24 hr) was measured in a random urine or 24-hour urine sample collected from each subject, respectively, using turbidimetric immunoassay. Height (cm) was measured using the Harpenden stadiometer (Holtain Ltd., Crymych, Wales, UK), and weight (kg) was measured using a digital scale (150 A; Cas Co. Ltd., Seoul, Korea). Body mass index (BMI) was calculated as weight (kg) divided by height squared (m²). The standard deviation score (SDS) for height, weight, and BMI was assigned based on the 2017 Korean National Growth Charts for patients over 2 years old [7].

Microvascular complications were evaluated over approximately 8 years of follow-up after ABPM assessment. Screening for DR presence or absence was performed by trained ophthalmologists using direct fundoscopy with mydriasis. Age at first pathological eye examination was defined as age at DR onset. Microalbuminuria was defined as excretion of 30–300 mg of albumin per 24 hours or a spot urine albumin-to-creatinine ratio of 30–300 μg/mg on 2 of 3 urine collections.

Clinic BP was defined as the average of 2 sphygmomanometer measurements conducted with the patient in a sitting position for at least 5 minutes. ABPM was performed using the Tonoport V ambulatory BP system (General Electric, Milwaukee, WI, USA), with a suitably sized cuff placed around each patient's nondominant arm. The BP system was programmed to measure BP every 20 minutes from 8:00 AM to 10:00 PM and every 30 minutes from 10:00 PM to 8:00 AM. Self-reported sleep-wake times have been used to classify ABPM data into daytime and nighttime periods. The BP indices were calculated from 24 hours, daytime, and nighttime measurements [5]. Systolic and diastolic BPs and mean arterial BP (MAP) were analyzed according to normal values adjusted for sex and height, and their SDS were calculated [8]. The nighttime dip in BP was calculated using the following formula: systolic or diastolic nighttime dip=(daytime systolic or diastolic mean BP – nighttime systolic or diastolic mean BP)/daytime systolic or diastolic mean BP [5]. Hypertension was defined as daytime, nighttime, or 24-hour mean systolic or diastolic BP higher than the 95th percentile of the normal pediatric ABPM value, which was adjusted for sex and height [8]. Normal BP was defined as daytime, nighttime, and 24-hour systolic and diastolic BP <95th percentile for age and sex. Nondipper status was defined as nighttime systolic BP or diastolic BP <10% of the diurnal mean value. Patients were categorized into 3 circadian BP groups: dipper, nondipper normotension, and nondipper hypertension.

All continuous variables were tested for normality and are presented as mean±standard deviation or median with interquartile range. Student t-test or the Mann–Whitney U-test was used to compare continuous variables, and the chi-square test or Fisher exact test was used to compare categorical variables between the 2 groups. The Cox proportional hazards model was used to identify predictors of microvascular complications 8 years after ABPM evaluation. Event-free survival curves for total microvascular complications, DR, and microalbuminuria were constructed using the Kaplan-Meier method. Multivariate models were adjusted for age at T1DM diagnosis and mean HbA1c level from diagnosis to last follow-up. P<0.05 was considered statistically significant. Statistical analyses were performed using IBM SPSS Statistics ver. 22.0 (IBM Co., Armonk, NY, USA).

The clinical characteristics of patients with T1DM are presented in Table 1. Mean patient age was 15.6±2.4 years, mean T1DM duration was 7.5±4.0 years, mean BMI was -0.04±0.86 (SDS), and mean HbA1c level was 8.9%±1.9% at the time of ABPM evaluation. Among the 81 patients, 33 were male, and 25 were in puberty. No patient had both microalbuminuria and DR between January 2009 and February 2010.

All BP values (SDS) were elevated (Table 2), especially nighttime systolic BP (1.03±1.26), diastolic BP (1.33±1.04), and MAP (1.33±1.23). All BP profiles (mmHg) are summarized in Supplementary Table 1. Hypertension (daytime, nighttime, and 24-hour mean hypertension) was observed in 42 patients (51.9%): daytime hypertension in 18 patients (23.5%), nighttime hypertension in 35 patients (43.2%), and 24-hour hypertension in 20 patients (24.7%). Fifty-one patients (62.9%) had nondipper BP. Using the results of 24-hour ABPM, patients were categorized into 3 circadian BP groups: dipper (n=30, 37.0%), nondipper normotension (n=18, 22.2%), and nondipper hypertension (n=33, 40.7%). There were no significant differences among the 3 groups in age at diagnosis, T1DM duration, sex ratio, puberty status, or mean HbA1c level from T1DM diagnosis to ABPM evaluation.

Thirteen of 81 patients had their office BP readings corresponding with hypertension. When ABPM was measured, 5 patients had high BP only during daytime, and 8 patients had hypertension in the 24-hour mean measurement, daytime, and nighttime; i.e., there was no evidence of white-coat hypertension. Among 68 patients with normal office BP, 34 patients had any ABPM hypertension (24-hour mean, daytime, or nighttime hypertension); i.e., 34 patients had masked hypertension. Their distribution is expressed in a Venn diagram (Supplementary Fig. 1). Since office BP was measured during the day, we compared office BP and daytime ABPM using paired t-tests (Supplementary Table 2). Office BP was significantly lower than daytime ABPM, apparently reflecting the characteristics of the patient group that had masked hypertension (38 of 81, 46.9%).

Mean age at follow-up was 23.3±2.6 years, mean T1DM duration was 15.2±4.2 years, BMI was 22.4±2.9 kg/m², and mean HbA1c level was 8.6%±1.1% (Table 3). Over the 8-year follow-up after ABPM evaluation, microvascular complications occurred in 8 patients (DR alone in 5, microalbuminuria alone in 2, and both in 1; Fig. 1), 7 of whom had nondipper BP. Mean HbA1c level (from diagnosis to last follow-up) was significantly lower in the non-DR group than DR group (8.5%±1.0% vs. 10.1%±0.7%; P<0.001) (Table 3). Mean clinic systolic BP, nighttime diastolic BP, and MAP were significantly higher in the DR than non-DR group (P=0.049, P=0.033, and P=0.017, respectively; Table 2, Supplementary Table 1). Although not statistically significant, the other BP profiles (clinic diastolic BP and MAP; 24-hour systolic BP, diastolic BP, and MAP; daytime systolic BP, diastolic BP, and MAP; and nighttime systolic BP, diastolic BP, and MAP) were also higher in the DR group (Table 2). The proportion of patients with DR was significantly higher in the nondipper hypertension group (83.3%) compared with the dipper and nondipper normotension groups (0% and 16.7%, respectively; P=0.021). Univariate Cox regression analysis after adjustment for T1DM duration showed that mean HbA1c level (hazard ratio [HR], 4.84; 95% confidence interval [CI], 1.97–11.92; P<0.001), age at diagnosis (HR, 1.33; 95% CI, 0.98–1.81; P=0.07), and puberty status at diagnosis (HR, 8.478; 95% CI, 1.095–65.666; P=0.041) were significantly related to development of microvascular complications (Table 4); none of the BP profiles showed a significant relationship. In a multivariate analysis adjusted for age at diagnosis and mean HbA1c level, neither the BP profile nor presence of dipper BP was related to development of microvascular complications. Mean HbA1c level (from diagnosis to last follow-up) was significantly higher in the microalbuminuria than nonmicroalbuminuria group (10.1%±0.6% vs. 8.6%±1.1%; P=0.019) (Table 3). Unlike DR, BP profiles were not significantly related to microalbuminuria (Table 2, Supplementary Table 1). In a univariate Cox regression analysis adjusted for T1DM duration, mean HbA1c level (HR, 3.78; 95% CI, 1.19–12.03; P=0.024) and age at diagnosis (HR, 1.54; 95% CI, 0.93–2.55; P=0.096), but not puberty status at diagnosis or BP profiles, were significantly related to development of microvascular complications (Table 4). In a multivariate analysis adjusted for age at diagnosis and mean HbA1c level, BP profile and presence of dipper were also not related to development of microvascular complications. Thus, age at diagnosis, mean HbA1c level from T1DM diagnosis to complication development, and puberty status at diagnosis were significantly predictive of DR development. For microalbuminuria, only mean HbA1c level was significantly predictive.