Glycemic Variability and Disordered Eating Among Adolescents and Young Adults with Type 1 Diabetes: The Role of Disinhibited Eating

Abstract

Background and Aims:

Disordered eating behaviors (DEB) are common among individuals with type 1 diabetes (T1D). Glycemic variability, potentially harmful in T1D, may reveal distinct characteristics between those with higher versus lower variability, particularly concerning DEB. Our aim was to evaluate the prevalence of DEB and associated risk factors among adolescents and young adults with T1D and to investigate unique factors associated with DEB across different levels of glycemic variability.

Methods:

An observational, cross-sectional study was conducted with 147 individuals with T1D, aged 13–21 years. Data were collected from medical charts, personal technological devices for assessing glycemic variability, and self-reported questionnaires, including assessments of DEB.

Results:

DEB were found in 62 (42.1%) individuals, and 41.5% achieved the glycemic variability (% coefficient of variation) target ≤36%. Among individuals with low glycemic variability, DEB were positively associated with diabetes distress (odds ratio [OR]: 1.14 [95% confidence interval or CI: 1.05–1.22], P < 0.001), longer diabetes duration (OR: 1.34 [95% CI: 1.05–1.70], P = 0.016) and lower socioeconomic-status (OR: 0.53 [95% CI: 0.31–0.90], P = 0.019). Among those with high glycemic variability, body mass index Z score (OR: 3.82 [95% CI: 1.48–9.85], P = 0.005), HbA1c (OR: 4.12 [95% CI: 1.33–12.80], P = 0.014), disinhibited eating (OR: 1.57 [95% CI: 1.14–2.15], P = 0.005), and tendency to lower socioeconomic status (OR: 0.75 [95% CI: 0.56–1.01], P = 0.065).

Discussion:

DEB are prevalent among adolescents and young adults with T1D and are associated with various risk factors. Factors associated with DEB vary across different levels of glycemic variability. Both low and high glycemic variability are associated with specific risk factors for DEB. One notable risk factor is diabetes-specific disinhibited eating among individuals with high glycemic variability, in contrast to those with low glycemic variability. Given these different risk factors, it may be prudent to adjust intervention programs to reduce DEB among T1D adolescents according to their glycemic variability levels.

Introduction

Disordered eating behaviors (DEB) are a range of eating pathologies that fail to meet a formal eating disorder diagnosis, which include a spectrum of abnormal behaviors such as binge eating, restricting food intake, rigid dietary rules, and purging to reduce body weight.¹ Among adolescents with type 1 diabetes (T1D), one of the most common chronic diseases among children and adolescents,² a higher prevalence of DEB has been observed in comparison with the general population.³ DEB in adolescents with T1D include not only DEB that often occur in the general population but also binge eating and purging followed by insulin restriction or omission, a dangerous form of purging calories via intentional hyperglycemia and glycosuria.^1,4

Previous longitudinal studies of adolescents and young adults with T1D have shown that DEB persist and increase in severity over time, and DEB and eating disorders were found to be both common and persistent, with new cases emerging well into adulthood.^5,6 Moreover, the combination of T1D and DEB is associated with poor glycemic control, increased risk of both short- and long-term T1D complications, and increased mortality.^4,7 DEB are common among individuals with T1D from various reasons. The standard nutritional treatment of T1D requires a significant focus on food and includes dietary rules and limitations related to diabetes treatment. To some extent this may lead to dietary restraint, attempting to restrain eating in an extreme or rigid manner, which may elevate the risk of disinhibited eating (eating more than intended in an uncontrolled manner) and lead to insulin omission.^8
–10 Moreover, intensive insulin therapy is associated with increased body weight, which may promote weight concerns and body dissatisfaction.^9,11,12 These factors, combined with disease-related distress and burden related to living with T1D, may contribute to DEB.⁹

In recent years, there is increased awareness on the impact of blood glucose level fluctuations (glycemic variability) on clinical outcomes of persons with diabetes.¹³ Glycemic variability is a physiological phenomenon that is important in the context of T1D. It encompasses fluctuations in blood glucose levels, including hypoglycemic episodes, postprandial levels, and variations across different days. In T1D, glycemic variability plays a crucial role, as alternating hyperglycemia, normoglycemia, and hypoglycemia are linked to insulin dosage.¹⁴ Previous studies have shown that higher glycemic variability could be harmful and associated with complications related to diabetes.¹³ Achieving the glycemic target without hypoglycemia and with limited glucose variability is a desirable objective for people with T1D.^15,16

Glycemic variability has been associated with several variables. In previous studies, it was found that adolescents who take more responsibility for T1D treatment and have higher treatment adherence have lower glycemic variability.^17,18 On the contrary, a longitudinal study among adolescents and young adults with T1D found that disease-related distress and anxiety were associated with higher glycemic variability.¹⁹ In addition, glycemic variability has been associated with eating behaviors, in particular, increased hunger, disinhibited eating, and an increase in food cravings.^10,20 Therefore, the characteristics of adolescents with higher glycemic variability may differ from those with lower glycemic variability, particularly when considering DEB. Factors associated with DEB may vary across different levels of glycemic variability, with both low and high glycemic variability potentially presenting distinct risk factors for DEB. Evaluating DEB risk factors relative to glycemic variability, whether low or high, may contribute to the knowledge regarding unique risk factors according to glycemic variability levels.

In the current study, we aimed to assess the prevalence of DEB and the risk factors associated with DEB among adolescents and young adults with T1D and to explore specific risk factors associated with DEB according to glycemic variability, whether low or high.

Materials and Methods

Participants and study design

This was an observational, cross-sectional study including patients with T1D treated at the Pediatric Diabetes Clinic at Schneider Children’s Medical Center of Israel. The study was carried out during October 2022–February 2024.

The inclusion criteria were patients with T1D aged 13–21 years, with diabetes duration of more than 6 months and for whom we had sufficient blood glucose data (use of continuous glucose monitoring [CGM], with CGM usage ≥70%). Excluded from the study were adolescents with known eating disorders, medical conditions related to nutritional status (e.g., various syndromes and the use of medications that may impact weight or diabetes management), psychiatric disorders, or inability to complete the study questionnaires. The assessment was done using electronic questionnaires sent to the participants or by filling out the questionnaires during the clinic visit. Participants completed the questionnaires independently.

Data collection

Medical chart

Participant characteristics that were obtained from the medical chart included sex, age, socioeconomic status, diabetes duration, height, weight, HbA1c, insulin regimen (multiple daily injections or insulin pump), and use of CGM.

Capillary HbA1c was measured by the DCA 2000 analyzer (Bayer Diagnostics Inc.).

Body mass index (BMI) was calculated as weight (kg)/height² (m²) and standardized to age and sex according to the standards of the Centers for Disease Control and Prevention.²¹

Personal technological devices

Blood glucose data during the preceding 2 weeks were obtained from CGM of patients’ devices. Blood glucose data included mean glucose and glucose standard deviation (SD). In addition, percent of CGM blood glucose values by clinical targets,¹⁵ time in target range (70–180 mg/dL), time above target range (>180 mg/dL and >250 mg/dL), and time below target range (<70 and <54 mg/dL) were expressed as percentages of all data readings.

The coefficient of variation was suggested to be the main parameter representing glycemic variability.²² Coefficient of variation was calculated as SD divided by the mean glucose. Glucose values were obtained from 2 weeks of CGM, with CGM usage ≥70%.

Mean daily insulin dose (units/day) during the preceding 2 weeks was obtained from the insulin pump of patients’ technological devices.

Socioeconomic position (SEP) was analyzed based on the Israel Central Bureau of Statistics’ characterization and classification of statistical areas within municipalities and local councils by socioeconomic level of the population in 2015.²³ The SEP clusters are scored between 1 and 10, with 1 representing the lowest and 10 the highest rating.

The Rabin Medical Center Ethics Committee approved the study. Participants or their legal guardians provided informed consent approving their participation in the study.

Measures

Diabetes Eating Problem Survey-Revised

DEB were screened by using the Diabetes Eating Problem Survey-Revised (DEPS-R).²⁴ The DEPS-R is a 16-item diabetes-specific self-reporting screening tool for disordered eating. The DEPS-R questionnaire includes different features associated with DEB, such as a drive for thinness, as indicated by “I would rather be thin than have good control of my diabetes,” eating pathologies, such as “I skip meals and/or snacks”, and diabetes management. The DEPS-R has good internal consistency (Cronbach’s α = 0.86) and is validated for use in the pediatric population. The questionnaire is scored on a 6-point Likert scale from 0 to 5 (never to always), with higher scores indicating more DEB. A score ≥20 indicates an increased risk for DEB.²⁴ Acceptable internal consistency was observed in the current sample of adolescents and young adults (α = 0.82).

Problem Areas in Diabetes Survey–Pediatric Version

Burden related to living with T1D was assessed using the Problem Areas in Diabetes Survey–Pediatric Version (PAID-Peds).²⁵ The PAID-Peds is a 20-item diabetes-specific self-reporting screening tool that assesses burden associated with T1D. The PAID-Peds has good internal consistency (Cronbach’s α = 0.94) and is validated for use in the pediatric population. The questionnaire is scored on a 5-point Likert scale from 0 to 4 (agree to disagree), with higher scores indicating greater burden related to living with diabetes. Acceptable internal consistency was observed in the current sample of adolescents and young adults (α = 0.90).

Disinhibited eating

Disinhibited eating (i.e., the experience of eating when blood glucose is thought to be dropping or low) was assessed by using two items of behavioral indicators of disinhibited eating, which included relinquishing control over the type and amount of food when blood glucose is perceived to be low. In addition, one item of the emotional indicators was used to examine guilt, shame, or regret for disinhibited eating when blood glucose is perceived to be low. The disinhibited eating items were developed for a cross-sectional larger study,¹⁰ which included 276 adolescents and adults with T1D, and aimed to examine if individuals with T1D are less restrained in their eating when they think their blood glucose is low and whether this contributes to insulin omission. The disinhibited eating items are a diabetes-specific self-reporting tool to assess disinhibited eating when blood glucose is thought to be low. The items are scored on 6-point Likert scale from 0 to 5 (never to always), with higher scores indicating greater disinhibited eating when blood glucose is perceived to be low. Acceptable internal consistency was observed in the current sample of adolescents and young adults (α = 0.64).

Statistical analysis

IBM SPSS Statistics, version 27.0 (IBM Corp. Armonk, NY) was used for analysis.

The Kolmogorov–Smirnov test was applied to test the normality of continuous data. The data are expressed as number (percent) for categorical variables, means ± standard deviations (SD) for normally distributed variables, and medians and interquartile range (IQR) for skewed distributions. Variables were compared between groups using independent t-tests (for normally distributed variables) or Mann–Whitney test (for skewed variables) or chi-squared test (for categorical and ordinal variables).

A multiple logistic regression analysis stepwise forward likelihood ratio (LR) was performed to examine clinical characteristics associated with DEPS-R (as a dichotomized score [DEPS score <20 or ≥20]). The model included the following variables: potential confounders in the first block, gender, age, diabetes duration, and SEP were entered in the enter method, and in the second block, HbA1c, BMI Z score, time below the target range of 54–70 mg/dL, PAID-Peds score, and disinhibited eating score were entered in the stepwise forward LR method. The analysis was done for the whole group and stratified according to glycemic variability. Participants were divided into two groups using the median of glycemic variability as the cutoff (low or high glycemic variability). We categorized adolescents and young adults by the level of glycemic variability (low or high) according to their coefficient of variation: < coefficient of variation, median = low glycemic variability, and ≥ coefficient of variation, median = high glycemic variability.

Results

Participant characteristics

Of the 170 adolescents and young adults invited to answer the questionnaires, 147 agreed to participate (86.4% response rate). No significant differences in age, sex distribution, BMI Z score, or HbA1c levels were found between the 147 participants and 23 nonparticipants (data not shown).

Of the 147 adolescents and young adults (46.9% female) who completed the DEPS-R questionnaire, 62 (42.1%) scored above the cutoff on the DEPS-R (≥20), indicating a high risk for DEB, 32.7% achieved a target HbA1c level <7.0%, and 41.5% achieved a glycemic variability (% coefficient of variation) target ≤36%.¹⁶ The mean glycemic variability was 37.9 ± 6.2, and the median glycemic variability was 37.4 (IQR 33.9–41.4). Participants used the following CGMs: 48.9% Dexcom G6 (Dexcom Inc., San Diego, CA), 20.4% FreeStyle Libre 1 and 2 (Abbott Diabetes Care, Alameda, CA), and 30.6% Guardian 3 and 4 (Medtronic, Northridge, CA). The clinical characteristics of the entire study cohort, stratified by glycemic variability, are presented in Table 1. Participants with high glycemic variability, as compared with those with low glycemic variability, had decreased time in the target range of 70–180 mg/dL (56.5 [IQR 43.0–67.7] vs. 68.0 [IQR 52.5–77.0], respectively, P = 0.006); increased time above the target range of >250 mg/dL (13.0 [IQR 6.2–25.2] vs. 7.0 [IQR 3.0–11.7] %, respectively, P < 0.001); increased time below the target range of 54–70 mg/dL (3.0 [IQR 2.0–5.0] vs. 1.0 [IQR 1.0–3.0] %, respectively, P < 0.001); increased time below the target range of <54 mg/dL (1.0 [IQR 0.5–2.0] vs. 1.0 [IQR 0.0–1.0] %, respectively, P = 0.002); and a longer duration of diabetes (6.9 [IQR 4.8–10.4] vs. 3.7 [IQR 1.6–10.5] years, respectively, P = 0.001).

Table 1.

Clinical and Glycemic Characteristics According to Glycemic Variability in Adolescents and Young Adults

Variable	All N = 147	Low glycemic variability N = 73	High glycemic variability N = 74	P value
Age (years), mean (SD)	16.9 (2.2)	17.1 (2.3)	16.9 (2.2)	0.867
Female, n (%)	69 (46.9)	30 (41.1)	39 (52.7)	0.107
SEP median (range)	7 (4; 8)	7 (5; 8)	6.5 (3; 8)	0.229
BMI Z score, mean (SD)	0.46 (1.06)	0.35 (1.0)	0.56 (0.99)	0.222
Diabetes duration, median, years (IQR)	5.9 (2.7; 10.3)	3.7 (1.6; 10.5)	6.9 (4.8;10.4)	0.001
Diabetes duration, median, years (IQR)
CGMs n (%)
Dexcom G6	72 (48.9)	34 (46.5)	38 (51.3)
Guardian 3 and 4	45 (30.6)	25 (34.2)	20 (27.0)	0.360
FreeStyle Libre 1 and 2	30 (20.4)	14 (19.1)	16 (21.6)
CGM usage, median, % (IQR)	93.0 (86.0–93.0)	93.0 (86.0–94.7)	92.5 (82.7–93.0)	0.154
Insulin pump, n (%)	119 (80.9)	57 (78.1)	60 (81.1)	0.403
Daily insulin dose, median, units/kg/day (IQR)	0.86 (0.70; 1.03)	0.84 (0.67; 1.0)	0.87 (0.70; 1.10)	0.241
Daily insulin dose, median, units/kg/day (IQR)
HbA1c, median, % (IQR)	7.4 (6.8–7.9)	7.3 (6.7–7.9)	7.4 (6.9–7.9)	0.291
Glucose, median, mg/dL (IQR)	160.2 (145.0; 184.0)	156.0 (140.5; 176.5)	167.5 (151.2; 193.7)	0.012
Time in range (70–180 mg/dL), median, % (IQR)	60.0 (46.0;73.0)	68.0 (52.5; 77.0)	56.5 (43.0; 67.7)	0.006
Time in range (70–180 mg/dL), median, % (IQR)
Time below Range (70–54 mg/dL), median, % (IQR)	2.0 (1.0; 4.0)	1.0 (1.0; 3.0)	3.0 (2.0; 5.0)	<0.001
Time below Range (70–54 mg/dL), median, % (IQR)
Time below range (<54 mg/dL), median, % (IQR)	1.0 (0.0; 1.0)	1.0 (0.0; 1.0)	1.0 (0.5; 2.0)	0.002
Time below range (<54 mg/dL), median, % (IQR)
Time above range (180–250 mg/dL), median, % (IQR)	24.0 (18.0; 29.0)	23.5 (17.0; 30.7)	24.5 (18.3; 27.7)	0.707
Time above range (180–250 mg/dL), median, % (IQR)
Time above range	9.0 (4.0; 21.0)	7.0 (3.0; 11.7)	13.0 (6.2; 25.2)	<0.001
(>250 mg/dL), median, % (IQR)
DEPS-R score cutoff, n (%)	62 (42.1)	29 (39.7)	32 (43.2)	0.395
Disinhibited eating score, median (IQR)	8.0 (5.0; 10.0)	8.0 (5.0; 11.0)	7.0 (5.0; 9.75)	0.237
Disinhibited eating score, median (IQR)
PAID-Peds score, mean (SD)	39.9 (19.6)	40.3 (21.2)	39.3 (18.2)	0.747

Bold indicates statistical significance.

BMI Z score, Z score for body mass index; CGM, continuous glucose monitoring; DEPS-R, Diabetes Eating Problem Survey-Revised; HbA1c, hemoglobin A1c; IQR, interquartile range; PAID-Peds, Problem Areas in Diabetes Survey–Pediatric Version; SD, standard deviation; SEP, socioeconomic position.

The association of DEB with participant characteristics according to glycemic variability level

When comparing participant characteristics among participants with DEB and those without DEB, more girls have DEB than boys (52.1% vs. 32.5%) and girls constituted 58.1% of the group with DEB (P = 0.012). In addition, those with DEB had a higher BMI Z score (0.76 ± 0.94 vs. 0.24 ± 1.08 kg/m², P = 0.003), higher HbA1c (7.6% [6.8–8.3] vs. 7.3% [6.7–7.6], P = 0.009), higher disinhibited eating score (9.0 [7.0–12.0] vs. 6.0 [4.0–9.0], P < 0.001), and higher PAID-Peds score (50.83 ± 18.33 vs. 31.92 ± 16.55, P < 0.001). When comparing the two groups stratified according to glycemic variability level, for participants with low glycemic variability we found that those with DEB had a higher disinhibited eating score (10.0 [8.0–12.5] vs. 7.0 [3.0–9.0], P < 0.001) and higher PAID-Peds score (55.44 ± 18.62 vs. 30.21 ± 16.49, P < 0.001). In addition, more girls had DEB than boys (56.7% vs. 27.9%) and girls constituted 58.6% of the group with DEB (P = 0.013). For participants with high glycemic variability, when comparing the two groups, we found that those with DEB had a higher BMI Z score (0.82 ± 1.05 vs. 0.37 ± 0.90 kg/m², respectively, P = 0.049), higher HbA1c (7.6% [6.9–8.4] vs. 7.3% [6.8–7.6], P = 0.040), higher disinhibited eating score (8.0 [85.0–11.7] vs. 6.0 [5.0–8.0], P = 0.016), higher PAID-Peds score (47.14 ± 17.50 vs. 33.20 ± 16.53, P < 0.001), and decreased time below the target range of 54–70 mg/dL (2.0 [1.0–4.0] vs. 4.0 [2.0–6.0], P = 0.020).

Table 2 shows the results of the multiple logistic regression analysis stepwise forward LR of DEPS-R score cutoff of the entire study cohort and stratified by the level of glycemic variability. In the final model, DEPS-R score ≥20 was associated with a higher BMI Z score (odds ratio [OR]: 2.34 [95% confidence interval or CI: 1.32–4.16], P = 0.004), higher HbA1c (OR: 2.38 [95% CI: 1.20–4.73], P = 0.013), higher disinhibited eating score (OR: 1.25 [95% CI: 1.04–1.50], P = 0.014), higher PAID-Peds score (OR: 1.06 [95% CI: 1.02–1.09], P < 0.001), and a lower SEP (OR: 0.70 [95% CI: 0.55–0.89], P = 0.004). For participants with low glycemic variability, in the final model, DEPS-R score ≥20 was associated with a longer diabetes duration (OR: 1.34 [95% CI: 1.05–1.70], P = 0.016), higher PAID-Peds score (OR: 1.14 [95% CI: 1.05–1.22], P < 0.001), and lower SEP (OR: 0.53 [95% CI: 0.31–0.90], P = 0.019). For participants with high glycemic variability, in the final model, DEPS-R score ≥20 was associated with a higher BMI Z score (OR: 3.82 [95% CI: 1.48–9.85], P = 0.005), higher HbA1c (OR: 4.12 [95% CI: 1.33–12.80], P = 0.014), higher disinhibited eating score (OR: 1.57 [95% CI: 1.14–2.15], P = 0.005), and tendency to lower SEP (OR: 0.75 [95% CI: 0.56–1.01], P = 0.065).

Table 2.

Multiple Logistic Regression Analysis Stepwise Forward LR of DEPS-R Score Cutoff of the Entire Study Cohort and Stratified by Glycemic Variability

	All			Low glycemic variability			High glycemic variability
	N = 147			N = 73			N = 74
	OR	95% CI	P value	OR	95% CI	P value	OR	95% CI	P value
Diabetes duration	—	—	—	1.34	1.05–170	0.016	—	—	—
SEP	0.70	0.55–0.89	0.004	0.53	0.31–0.90	0.019	0.75	0.56–1.01	0.065
BMI Z score	2.34	1.32–4.16	0.004	—	—	—	3.82	1.48–9.85	0.005
HbA1c (%)	2.38	1.20–4.73	0.013	—	—	—	4.12	1.33–12.80	0.014
Disinhibited eating score	1.25	1.04–1.50	0.014	—	—	—	1.57	1.14–2.15	0.005
PAID-Peds⁴ score	1.06	1.02–1.09	<0.001	1.14	1.05–1.22	<0.001	—	—	—

Bold indicates statistical significance.

Logistic regression model: dependent variable: DEPS-R as a binominal factor (≥20 = 1, <20 = 0). Independent variables: BMI Z score, HbA1c, time below the target range of 54–70 mg/dL, disinhibited eating, PAID-Peds. The model was adjusted for sex, age, diabetes duration, and SEP.

CI, confidence interval; OR, odds ratio.

Discussion

To our knowledge, this is the first study focusing on DEB risk factors and glycemic variability among adolescents and young adults with T1D. We found distinctive risk factors associated with DEB across varying levels of glycemic variability; both low and high glycemic variability are associated with unique risk factors for DEB. The study presents a novel finding indicating that adolescents and young adults experiencing high glycemic variability exhibit distinct risk factors for DEB, one of them characterized by diabetes-specific disinhibited eating. About 40% of our cohort had a DEPS-R score above the cutoff, indicating a high rate of DEB in the studied population of adolescents and young adults with T1D. Less than half of the adolescents and young adults in our study achieved the glycemic variability target. There were differences between participants with high glycemic variability and those with low glycemic variability, mainly in terms of time in range, a key metric that indicates the amount of time spent at the proposed target glucose range (70–180 mg/dL), as opposed to time above and below range, metrics that indicate the amount of time spent above the proposed target glucose range (>180 mg/dL and >250 mg/dL) or below the proposed target glucose range (<70 and <54 mg/dL).²⁶

Elevated BMI Z score, HbA1c, diabetes distress, diabetes-specific disinhibited eating, and lower socioeconomic status were associated with DEB. In stratification according to glycemic variability, we found that DEB were associated with higher diabetes distress, longer diabetes duration, and lower socioeconomic status among those with low glycemic variability, while for those with high glycemic variability, DEB were associated with BMI Z score, HbA1c, diabetes-specific disinhibited eating, and lower socioeconomic status.

DEB in our study were high, similar to a previous study,²⁷ and even slightly higher than in other studies. DEB were associated with elevated BMI Z score and HbA1c levels, consistent with prior findings.^28
–30 It is possible that the increased BMI Z score in our cohort contributed to the high prevalence of DEB, as in the general population.³¹ Elevated weight poses a clinical concern in people with T1D. Longitudinal studies have demonstrated a significant rise in the percentage of overweight and obesity among people with T1D.³² Moreover, a substantial increase in incidence rates of eating disorders in the post-COVID-19 pandemic period in the adolescent population has been shown in previous studies.³³ Furthermore, methodological differences might underlie the discrepancy in findings. In contrast to some previous studies, the assessment in our study was mostly done with electronic questionnaires sent to the participants, which may have promoted open self-disclosure.

When considering sex as a potential risk factor, our analysis did not identify it as an individual risk factor. However, our findings align with previous studies indicating a higher prevalence of DEB among females than males.^27,29 This is in line with findings in the general population, where eating disorders are more prevalent among females than males,³⁴ underscoring the importance of the need for special attention to females with T1D.

Findings from previous studies as in our study suggest that DEB were associated with low socioeconomic status. This association has been observed in both the general population and among individuals with T1D.^29,35 These findings underscore the critical need for targeted attention to this high-risk population.

Contrary to previous studies, the current study analyzed CGM data to evaluate the glucose profiles of study participants. We found that high glycemic variability was correlated with more time below the desired glucose range (hypoglycemia) and more time above the desired glucose target (hyperglycemia).

The literature on glycemic variability and DEB among individuals with T1D are scarce; a randomized clinical trial of a family-based behavioral nutrition intervention among adolescents with T1D showed that DEB were not associated with any measures of glycemic variability.³⁶ On the contrary, previous studies have found an association between glycemic variability and risk factors for DEB.^10,19 These associations may differ according to glycemic variability level. Factors associated with DEB may differ across various levels of glycemic variability, with both low and high glycemic variability showing distinct risk factors for DEB. Hence, our study aimed to investigate distinctive risk factors for DEB relative to varying levels of glycemic variability.

Consistent with previous studies, in our cohort, diabetes distress was associated with DEB.³⁷ Although DEB were not associated with hypoglycemia or hyperglycemia, there may be burdens related to T1D management, including careful attention to food choices and regular blood glucose monitoring, which may contribute to DEB. Additionally, negative emotions related to diabetes may lead to diabetes management problems and increased vulnerability to DEB. Interestingly, when stratifying according to glycemic variability, higher diabetes distress was associated with DEB only in participants with low glycemic variability. It is possible that adolescents and young adults who take more responsibility for T1D treatment and have higher treatment adherence,¹⁸ yet experience increased distress related to diabetes management, may be more associated with DEB. This emphasizes the need to help adolescents and young adults, including those with low glycemic variability, to develop skills to cope with diabetes distress when managing diabetes, especially in this high-risk population, in order to reduce DEB.

Evidence from previous studies, as in our study, suggests that DEB were associated with HbA1c,^27,29 but substantial research gaps remain regarding the nature of the association at different levels of glycemic variability. When stratifying according to glycemic variability, higher HbA1c was associated with DEB only in participants with high glycemic variability. Interestingly, the association between DEB and HbA1c changed based on the level of glycemic variability, and this association is particularly significant in instances of high glycemic variability.

While DEB were not associated with hypoglycemia or hyperglycemia, diabetes-specific disinhibited eating in the context of insulin treatment and fluctuation in blood glucose was found to be a risk factor for DEB and in particular among adolescents and young adults with high glycemic variability. Physiologically, low blood glucose has been linked to hunger and eating more freely.³⁸ However, previous studies have found that eating while the glucose level was low was associated with increased cravings for food, particularly those containing high levels of carbohydrates.³⁹ Moreover, when individuals perceive their blood glucose to be low, they may exhibit less restraint in their eating behaviors, potentially violating self-imposed dietary restrictions. This could contribute to DEB, including insulin omission, ultimately leading to poorer glycemic control.¹⁰

To the best of our knowledge, this study demonstrates for the first time that adolescents and young adults have distinct risk factors associated with DEB that vary across different levels of glycemic variability. Particularly, adolescents and young adults with high glycemic variability show unique risk factors for DEB, including diabetes-specific disinhibited eating. These findings may have significant implications because most adolescents and young adults with T1D do not meet the glycemic variability target. Therefore, they are at an increased risk for diabetes-specific disinhibited eating, which may lead to DEB and elevated HbA1c levels. In addition to the desire to minimize glycemic variability, strategies are needed to deal with diabetes-specific disinhibited eating during glycemic excursions.

The strengths of this study included the use of a validated, standardized, diabetes-specific screening tool. An additional strength is the use of CGM data to evaluate the glucose profiles that enabled the identification of the association between clinical characteristics and DEB according to different levels of glycemic variability. Another strength is the use of coefficient of variation as a measure of glycemic variability, which, in contrast to other methods measuring glycemic variability, is not influenced by mean blood glucose. Nevertheless, our study has some limitations.

First, the current study was cross-sectional and therefore did not allow for a causal conclusion and was exposed to potential confounding, although adjustments were made for various confounders. Thus, randomized clinical trials are needed to determine potential associations between clinical characteristics and DEB across glycemic variability levels. Second, the analysis of glucose data was restricted to a 2-week period with at least 70% of data collected during that period. While this duration is consistent with international consensus,^22,40 longer-term data collection could provide broader insights into glucose metrics and their implications for interventions in diabetes management. Third was the inability to compare CGM data with capillary glucose measurements; future studies should aim to integrate these data sources more comprehensively to improve understanding of real-world glucose management. Fourth was the lack of data related to the pubertal stage of participants, which included a relatively broad age range, although adjustments for age were made. Finally, filling out the questionnaires was done using both electronic and manual methods during clinic visits; however, no significant difference was found in DEPS-R scores between the two methods.

Conclusions

In conclusion, DEB are common among adolescents and young adults with T1D and were associated with higher BMI Z score, HbA1c, and diabetes distress, as well as lower socioeconomic status. Distinct factors associated with DEB vary across different levels of glycemic variability. Both low and high glycemic variability may exhibit distinct risk factors for DEB. Notably, diabetes-specific disinhibited eating was found to be a risk factor for DEB among adolescents and young adults with high glycemic variability, in contrast to those with low glycemic variability. Our findings suggest that regular screening for DEB and risk factors for DEB are essential and may have implications on DEB and glycemic control. Given the distinct risk factors for DEB based on levels of glycemic variability, it may be prudent to adjust intervention programs accordingly.

Footnotes

Authors’ Contributions

T.P.-L. was involved in conceptualization, contributed to methodology and data collection, carried out the formal analysis and investigation, wrote and prepared the original draft, reviewed and edited the article, and participated in supervision. R.E.-B. was involved in conceptualization, contributed to methodology, wrote and prepared the original draft, reviewed and edited the article, and participated in supervision. M.P. was involved in conceptualization and reviewed and edited the article. S.S. was involved in conceptualization, contributed to methodology, reviewed and edited the article, and participated in supervision. M.G.-K. was involved in the data collection and reviewed and edited the article. M.Y.-G. was involved in the formal analysis and investigation and reviewed and edited the article. A.L. contributed to the methodology and data collection. T.P.-L. is the guarantor of this work and, as such, has full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors critically revised the article and approved the final version.

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Author Disclosure Statement

The authors declare that there is no conflict of interest regarding the publication of this article.

Funding Information

No funding was received for this article.

References

Pinhas-Hamiel

, Hamiel

, Levy-Shraga

. Eating disorders in adolescents with type 1 diabetes: Challenges in diagnosis and treatment. World J Diabetes, 2015; 6(3):517–526; doi: 10.4239/wjd.v6.i3.517

Mayer-Davis

, Lawrence

, Dabelea

, et al. SEARCH for Diabetes in Youth Study. Incidence trends of Type 1 and Type 2 diabetes among youths, 2002–2012. N Engl J Med, 2017; 376(15):1419–1429; doi: 10.1056/nejmoa1610187

Pursey

, Hay

, Bussey

, et al. Diabetes and disordered eating behaviours in a community-based sample of australian adolescents. J Eat Disord, 2020; 8(1):5; doi: 10.1186/s40337-020-0282-y

Goebel-Fabbri

, Fikkan

, Franko

, et al. Insulin restriction and associated morbidity and mortality in women with type 1 diabetes. Diabetes Care, 2008; 31(3):415–419; doi: 10.2337/dc07-2026

Baechle

, Hoyer

, Stahl-Pehe

, et al. Course of disordered eating behavior in young people with early-onset Type I diabetes: Prevalence, symptoms, and transition probabilities. J Adolesc Health, 2019; 65(5):681–689; doi: 10.1016/j.jadohealth.2019.05.016

Colton

, Olmsted

, Daneman

, et al. Eating disorders in girls and women with type 1 diabetes: A longitudinal study of prevalence, onset, remission, and recurrence. Diabetes Care, 2015; 38(7):1212–1217; doi: 10.2337/dc14-2646

Scheuing

, Bartus

, Berger

, et al. German BMBF Competence Network Diabetes Mellitus. Clinical characteristics and outcome of 467 patients with a clinically recognized eating disorder identified among 52,215 patients with type 1 diabetes: A multicenter german/austrian study. Diabetes Care, 2014; 37(6):1581–1589; doi: 10.2337/dc13-2156

Moskovich

, Dmitrieva

, Babyak

, et al. Real-time predictors and consequences of binge eating among adults with type 1 diabetes. J Eat Disord, 2019; 7(1):7; doi: 10.1186/s40337-019-0237-3

Peterson

, Fischer

, Young-Hyman

. Topical review: A comprehensive risk model for disordered eating in youth with type 1 diabetes. J Pediatr Psychol, 2015; 40(4):385–390; doi: 10.1093/jpepsy/jsu106

10.

Merwin

, Moskovich

, Dmitrieva

, et al. Disinhibited eating and weight-related insulin mismanagement among individuals with type 1 diabetes. Appet, 2014; 81:123–130; doi: 10.1016/j.appet.2014.05.028

11.

Propper‐Lewinsohn

, Gillon‐Keren

, Shalitin

, et al. Disordered eating behaviours in adolescents with type 1 diabetes can be influenced by their weight at diagnosis and rapid weight gain subsequently. Diabet Med, 2023; 40(11):e15166; doi: 10.1111/dme.15166

12.

Russell-Jones

, Khan

. Insulin-associated weight gain in diabetes -causes, effects and coping strategies. Diabetes Obes Metab, 2007; 9(6):799–812; doi: 10.1111/j.1463-1326.2006.00686.x

13.

Ceriello

, Monnier

, Owens

. Glycaemic variability in diabetes: Clinical and therapeutic implications. Lancet Diabetes Endocrinol, 2019; 7(3):221–230; doi: 10.1016/S2213-8587(18)30136-0

14.

Suh

, Kim

. Glycemic variability: How do we measure it and why is it important? Diabetes Metab J, 2015; 39(4):273–282; doi: 10.4093/dmj.2015.39.4.273

15.

Battelino

, Danne

, Bergenstal

, et al. Clinical targets for continuous glucose monitoring data interpretation: Recommendations from the international consensus on time in range. Diabetes Care, 2019; 42(8):1593–1603; doi: 10.2337/dci19-0028

16.

Elsayed

, Aleppo

, Aroda

, et al. 6. glycemic targets: Standards of Care in Diabetes—2023. Diabetes Care, 2023; 46(Suppl 1):S97–S110; doi: 10.2337/dc23-S006

17.

Marker

, Noser

, Clements

, et al. Shared responsibility for Type 1 diabetes care is associated with glycemic variability and risk of glycemic excursions in youth. J Pediatr Psychol, 2018; 43(1):61–71; doi: 10.1093/jpepsy/jsx081

18.

Martyn-Nemeth

, Birlingmair

, Idemudia

, et al. Hypoglycaemic treatment adherence and the association with psychological, self-management and glycaemic characteristics in adults with type 1 diabetes. Nurs Open, 2019; 6(3):871–877; doi: 10.1002/nop2.267

19.

Déniz-García

, Díaz-Artiles

, Saavedra

, et al. Impact of anxiety, depression and disease-related distress on long-term glycaemic variability among subjects with Type 1 diabetes mellitus. BMC Endocr Disord, 2022; 22(1):122; doi: 10.1186/s12902-022-01013-7

20.

Merwin

, Dmitrieva

, Honeycutt

, et al. Momentary predictors of insulin restriction among adults with type 1 diabetes and eating disorder symptomatology. Diabetes Care, 2015; 38(11):2025–2032; doi: 10.2337/dc15-0753

21.

Kuczmarski

, Ogden

, Grummer-Strawn

, et al. CDC growth charts: United States. Adv Data, 2000(314):1–27.

22.

Danne

, Nimri

, Battelino

, et al. International consensus on use of continuous glucose monitoring. Diabetes Care, 2017; 40(12):1631–1640; doi: 10.2337/dc17-1600

23.

Israel Central Bureau of Statistics (CBS); 2019. Characterization and classification of geographic units by the socio-economic level of the population. n.d. Https://Www.Cbs.Gov.Il/He/Publications/DocLib/2019/1765_socio_economic_2015.

24.

Markowitz

, Butler

, Volkening

, et al. Brief screening tool for disordered eating in diabetes: Internal consistency and external validity in a contemporary sample of pediatric patients with type 1 diabetes. Diabetes Care, 2010; 33(3):495–500; doi: 10.2337/dc09-1890

25.

Markowitz

, Volkening

, Butler

, et al. Youth-perceived burden of type 1 diabetes: Problem areas in diabetes survey-pediatric version (PAID-Peds). J Diabetes Sci Technol, 2015; 9(5):1080–1085; doi: 10.1177/1932296815583506

26.

Care

, Suppl

. Glycemic targets: Standards of Medical Care in Diabetes−2021. Diabetes Care, 2021; 44(January):S73–S84; doi: 10.2337/dc21-S006

27.

Araia

, Hendrieckx

, Skinner

, et al. Gender differences in disordered eating behaviors and body dissatisfaction among adolescents with type 1 diabetes: Results from diabetes MILES youth—Australia. Int J Eat Disord, 2017; 50(10):1183–1193; doi: 10.1002/eat.22746

28.

Troncone

, Affuso

, Cascella

, et al. Diabetes Study Group of Italian Society of Paediatric Endocrinology and Diabetology. Prevalence of disordered eating behaviors in adolescents with type 1 diabetes: Results of multicenter Italian nationwide study. Int J Eat Disord, 2022; 55(8):1108–1119; doi: 10.1002/eat.23764

29.

Cherubini

, Skrami

, Iannilli

, et al. Disordered eating behaviors in adolescents with type 1 diabetes: A cross-sectional population-based study in Italy. Int J Eat Disord, 2018; 51(8):890–898; doi: 10.1002/eat.22889

30.

Nilsson

, Madsen

JOB

, Jensen

, et al. High prevalence of disordered eating behavior in Danish children and adolescents with type 1 diabetes. Pediatr Diabetes, 2020; 21(6):1043–1049; doi: 10.1111/pedi.13043

31.

Duncan

, Ziobrowski

, Nicol

. The prevalence of past 12-month and lifetime DSM-IV eating disorders by BMI category in US men and women. Eur Eat Disord Rev, 2017; 25(3):165–171; doi: 10.1002/erv.2503

32.

Volsky

, Shalitin

, Fridman

, et al. Dyslipidemia and cardiovascular disease risk factors in patients with type 1 diabetes: A single-center experience. World J Diabetes, 2021; 12(1):56–68; doi: 10.4239/wjd.v12.i1.56

33.

Matsumoto

, Kadowaki

, Takanaga

, et al. Longitudinal impact of the COVID-19 pandemic on the development of mental disorders in preadolescents and adolescents. BMC Public Health, 2023; 23(1):1308; doi: 10.1186/s12889-023-16228-z

34.

Galmiche

, Déchelotte

, Lambert

, et al. Prevalence of eating disorders over the 2000-2018 Period: A systematic literature review. Am J Clin Nutr, 2019; 109(5):1402–1413; doi: 10.1093/ajcn/nqy342

35.

Burnette

, Burt

, Klump

. The ignored role of disadvantage in eating disorders. Trends Mol Med, 2024; 30(4):305–307; doi: 10.1016/j.molmed.2023.11.006

36.

Colman

MHE

, Quick

, Lipsky

, et al. Disordered eating behaviors are not increased by an intervention to improve diet quality but are associated with poorer glycemic control among youth with type 1 diabetes. Diabetes Care, 2018; 41(4):869–875; doi: 10.2337/dc17-0090

37.

Propper-Lewinsohn

, Elran-Barak

, Gillon-Keren

, et al. Disordered eating behaviors among adolescents and young adults with Type 1 diabetes treated with insulin pumps and hybrid closed-loop systems. Diabetes Technol Ther, 2024; 26(4):229–237; doi: 10.1089/dia.2023.0500

38.

Ciampolini

, Bianchi

. Training to estimate blood glucose and to form associations with initial hunger. Nutr Metab (Lond), 2006; 3:42; doi: 10.1186/1743-7075-3-42

39.

Strachan

MWJ

, Ewing

FME

, Frier

, et al. Food cravings during acute hypoglycaemia in adults with Type 1 diabetes. Physiol Behav, 2004; 80(5):675–682; doi: 10.1016/j.physbeh.2003.12.003

40.

Battelino

, Alexander

, Amiel

, et al. Continuous glucose monitoring and metrics for clinical trials: An international consensus statement. Lancet Diabetes Endocrinol, 2023; 11(1):42–57; doi: 10.1016/S2213-8587(22)00319-9