Trajectories of HbA1c levels in children and youth with type 1 diabetes.

PURPOSE: To illustrate the distribution of Hemoglobin A1c (HbA1c) levels according to age and gender among children, adolescents and youth with type 1 diabetes (T1DM). METHODS: Consecutive HbA1c measurements of 349 patients, aged 2 to 30 years with T1DM were obtained from 1995 through 2010. Measurement from patients diagnosed with celiac disease (n = 20), eating disorders (n = 41) and hemoglobinopathy (n = 1) were excluded. The study sample comprised 4815 measurements of HbA1c from 287 patients. Regression percentiles of HbA1c were calculated as a function of age and gender by the quantile regression method using the SAS procedure QUANTREG. RESULTS: Crude percentiles of HbA1c as a function of age and gender, and the modeled curves produced using quantile regression showed good concordance. The curves show a decline in HbA1c levels from age 2 to 4 years at each percentile. Thereafter, there is a gradual increase during the prepubertal years with a peak at ages 12 to 14 years. HbA1c levels subsequently decline to the lowest values in the third decade. Curves of females and males followed closely, with females having HbA1c levels about 0.1% (1.1 mmol/mol) higher in the 25th 50th and 75th percentiles. CONCLUSION: We constructed age-specific distribution curves for HbA1c levels for patients with T1DM. These percentiles may be used to demonstrate the individual patient's measurements longitudinally compared with age-matched patients.

Detecting intentional insulin omission for weight loss in girls with type 1 diabetes mellitus.

OBJECTIVE: Intentional insulin omission is a unique inappropriate compensatory behavior that occurs in patients with type 1 diabetes mellitus, mostly in females, who omit or restrict their required insulin doses in order to lose weight. Diagnosis of this underlying disorder is difficult. We aimed to use clinical and laboratory criteria to create an algorithm to assist in the detection of intentional insulin omission. METHOD: The distribution of HbA1c levels from 287 (181 females) patients with type 1 diabetes were used as reference. Data from 26 patients with type 1 diabetes and intentional insulin omission were analysed. The Weka (Waikato Environment for Knowledge Analysis) machine learning software, decision tree classifier with 10-fold cross validation was used to developed prediction models. Model performance was assessed by cross-validation in a further 43 patients. RESULTS: Adolescents with intentional insulin omission were discriminated by: female sex, HbA1c>9.2%, more than 20% of HbA1c measurements above the 90th percentile, the mean of 3 highest delta HbA1c z-scores>1.28, current age and age at diagnosis. The models developed showed good discrimination (sensitivity and specificity 0.88 and 0.74, respectively). The external test dataset revealed good performance of the model with a sensitivity and specificity of 1.00 and 0.97, respectively. DISCUSSION: Using data mining methods we developed a clinical prediction model to determine an individual's probability of intentionally omitting insulin. This model provides a decision support system for the detection of intentional insulin omission for weight loss in adolescent females with type 1 diabetes mellitus.

The effectiveness of Internet-based blood glucose monitoring system on improving diabetes control in adolescents with type 1 diabetes.

OBJECTIVE: To determine whether the use of an Internet-based blood glucose monitoring system could improve glycemic control in adolescents with type 1 diabetes mellitus (T1DM). METHODS: In a randomized, controlled clinical trial, a total of 70 adolescent subjects with T1DM were recruited. Subjects randomized to the intervention group (n = 36) were instructed to submit their blood glucose levels weekly by Internet to the diabetes care team during a period of 6 months. Subjects randomized to the control group (n = 34) did not submit results but were under routine follow-up. RESULTS: At baseline, patients were 15.1 ± 2.6 years of age with mean HbA1c of 8.3 ± 1.3%. At the 6-month follow-up period, no by-group differences in change from baseline to end of treatment HbA1c levels were detected. In the intervention group, 12/36 did not submit blood glucose levels and were classified as non-compliant. In a secondary exploratory analysis in which non-compliant patients were omitted, HbA1c values in the compliant intervention group declined from 8.5 ± 1.7% at baseline to 8.2 ± 1.2% at 6 months, while in the control group HbA1c values increased from 8.2 ± 1.1 to 8.4 ± 1.1%, this difference did not reach statistical significance. CONCLUSIONS: An Internet-based blood glucose monitoring system was not associated with improved glycemic control in adolescents with T1DM. Identification of a sub-group of compliant subjects who may improve metabolic control by using this tool is needed.

Long-term insulin pump treatment in girls with type 1 diabetes and eating disorders--is it feasible?

AIM: This study evaluated the feasibility of insulin pump treatment in adolescent girls with type 1 diabetes mellitus (T1DM) and eating disorders (ED). METHODS: Sixty-three girls >10 years old were included in the study. Forty-eight were treated with pump (no-ED-pump group); 15 had ED, of whom eight were treated with pump (ED-pump group), and seven were treated with multiple daily injections (MDI) (ED-MDI group). RESULTS: Girls in the ED-pump group had higher hemoglobin A1c (HbA1c) compared to those in the no-ED-pump group both before and after pump initiation (P = 0.007). In the ED-pump group, levels 0-6 months after pump initiation were slightly but not significantly lower compared to baseline reference values (8.48% vs. 8.84%, respectively; P = 0.42). In the no-ED-pump group there was a significant decrease in HbA1c level 0-6 months after pump initiation (8.03% vs. 7.67%, respectively; P = 0.004). Thereafter HbA1c levels were not different from baseline reference on both groups. The rate of hypoglycemic episodes was similar in the ED pump and no-ED-pump groups (0.9 vs. 1.0 episodes per 100 patient-years, respectively). Prior to diagnosis of ED, mean HbA1c levels in the ED-pump group and the ED-MDI group were similar: 9.24 ± 1.25% versus 8.47 ± 1.30% (P = 0.51). In the time intervals after the diagnosis of ED, mean HbA1c level of all measurements in the eight girls in the ED-pump group was significantly lower compared with the seven patients in the ED-MDI group: 9.07 ± 1.33% versus 10.40 ± 2.01% (P = 0.04). CONCLUSIONS: Treatment with an insulin pump was feasible in girls with T1DM and ED and resulted in significantly lower HbA1c levels compared to the ED-MDI group.

The impact of baseline hemoglobin A1c levels prior to initiation of pump therapy on long-term metabolic control.

OBJECTIVE: This study was done to identify factors influencing long-term metabolic control in youth with type 1 diabetes mellitus (T1DM) treated with an insulin pump. RESEARCH DESIGN AND METHODS: Data were obtained from retrospective chart review of 113 patients (52 males) with T1DM treated with an insulin pump for up to 7 years. Their mean +/- SD age at diagnosis of T1DM was 9.7 +/- 5.1 years, and that at pump therapy initiation was 13.8 +/- 6.1 years. Linear trends and changes in hemoglobin A1c (HbA1c) levels following pump insertion were evaluated according to gender, metabolic control prior to initiation of pump therapy, time from diagnosis of diabetes until pump therapy, age at initiation, and the duration of pump treatment. RESULTS: Mean HbA1c levels of patients with good baseline metabolic control (HbA1c level 9%) control groups (8.1 +/- 0.9% and 8.2 +/- 1.1%, respectively; P < 0.001). However, with time a significant trend for increasing HbA1c level was demonstrated in the group with good metabolic control (P value for trend = 0.004). HbA1c levels of patients with poor baseline metabolic control decreased significantly immediately after pump initiation (9.4 +/- 1.6% vs. 8.0 +/- 1.2%, P = 0.0001) and thereafter remained stable (P value for trend = 0.54). In the multivariable analyses, baseline HbA1c level