Eating Disorders in Young Adults and Adults with Type 1 Diabetes Mellitus

Abstract

The literature describes an increased risk of eating disorders (EDs) in patients with Type 1 diabetes mellitus (T1DM) compared to the general population. This risk is mainly related to physical and psychosocial problems related to diabetes. EDs should be carefully assessed and treated in these patients since they are associated with poor glycemic control and significant repercussions of pathology. Background/Objectives: To study the presence of EDs in young Portuguese adults and adults with T1DM, and how gender; age group; method of insulin administration; carbohydrate counting; and body mass index [BMI] variables influence the risk of developing an ED. Methods: Data collection was carried out using an online questionnaire, which was disseminated through the media of several Portuguese diabetes associations. Results: The sample consisted of 47 participants, mostly female, with the age group between 26 and 35 years being most representative. A statistically significant association was found between the Eating Attitudes Test [EAT-26] scores and the BMI of the participants [p = 0.003]; other variables did not show statistically significant differences. Conclusions: To better understand the relationship between these two pathologies, further studies are needed, as well as the development of more screening instruments to assess the risk of EDs specific to T1DM, and preventive interventions and guidelines that can assist the various areas of health that support the population with T1DM.

1. Introduction

Type 1 diabetes mellitus (T1DM) is a pathology caused by autoimmune damage to the insulin-producing β-cells of the pancreatic islets, usually leading to an endogenous insulin deficit [1]. There is extensive knowledge about many aspects of this pathology, including its genetics, epidemiology, immunological phenotypes, and the difficulties that are associated with it [2]. Currently, it is a disease recognized as the result of a complex interaction between the genome, metabolism, immune system, and environmental factors that vary between each individual, as well as the interaction between pancreatic β-cells and the innate and acquired immune systems [1,2].

Globally, T1DM is increasing both in terms of incidence and prevalence, with annual increases in incidence between two and three per cent [2]. In 2021, the estimated prevalence of diabetes [Type 1 and 2] in the Portuguese population aged between 20 and 79 was 14.1%, i.e., around 1.1 million Portuguese in this age group have diabetes [3]. On average, in the last decade, 670 new cases of diabetes [Type 1 and 2] were diagnosed every year for every 100,000 residents in mainland Portugal [3]. Many environmental exposures are associated with T1DM, including diet in childhood and adulthood, vitamin D sufficiency, early exposure to viruses associated with pancreatic islet inflammation, and decreased diversity of the intestinal microbiota [2].

Optimal glycemic control in T1DM requires multiple-dose insulin regimens that mimic physiological insulin release, which can be administered via injection or with an insulin infusion pump (IIP).

The goals to be achieved in T1DM should be individualized based on many criteria including comorbidities, the patient’s psychosocial abilities, and the available care resources [2,3]. For best results, patients should be supported by multidisciplinary teams and include information on insulin adjustments, carbohydrate (CH) counting, and optimal use of available technology [2]. The main life-threatening complications are hypoglycemia and ketoacidosis, both extreme consquences of poor T1DM control. An additional complication worth mentioning is the effect on patients’ quality of life. Decreased quality of life in these patients is associated with a prediction of subsequent poor glycemic control [2].

Eating disorders have increased over the last 50 years and are related to changes in the food environment of populations [4]. EDs are characterized by persistent disturbances in eating or eating-related behavior that result in altered consumption or absorption of food, and significantly impair health or psychosocial functioning [1,5]. Disordered attitudes towards weight, body shape, and eating play a fundamental role in their origin and maintenance [4,6]. Current literature suggests that there is a divergence in the underlying etiology between anorexia nervosa and binge eating spectrum disorders. It is likely that interactions between genetic and environmental factors, at a crucial time in the developmental period, add to the complexity of shaping these disorders [4].

All eating disorders damage physical health considerably and disrupt social functioning [4]. Both the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) and the International Classification of Diseases (ICD-11) encompass six main eating disorders: anorexia nervosa, bulimia nervosa, and binge eating disorder, as well as avoidant-restrictive food intake disorder, pica, and rumination disorder [4]. Eating disorders can affect individuals of all ages, genders, and ethnicities. Adolescents and young adults are the groups particularly at risk. In adult populations, the clinical profile is dominated by binge eating spectrum disorders. Globally, the prevalence of eating disorders has increased by 25%, but only around 20% of affected individuals seek treatment [4]. Epidemiological data in Portugal are scarce and present certain methodological limitations, particularly the lack of nationally representative samples. However, the available data seems to indicate figures similar to those observed in other European countries. The most recent and comprehensive study on the prevalence of EDs in Portugal dates to 1999, in which 2398 female high school students from the Lisbon and Setúbal regions were assessed, estimating a prevalence of 0.37% for anorexia nervosa and 12.6% for partial syndromes [7].

EDs, or even disturbed eating behaviors (DEBs), are more prevalent in patients with T1DM than in the general healthy population. It has been estimated that the prevalence rate of DEB in patients with T1DM ranges from 8% to 55%, and in the non-diabetic population, it is approximately 32.5% [8]. The prevalence of EDs is around 21% in adults with T1DM [8]. In a study conducted in Italy with a sample of 211 insulin-treated diabetic patients with type 1 or type 2 diabetes, ED was observed in 13.3% of the participants [9]. In the presence of an eating disorder, the frequency of glucose monitoring decreases and there is worse glycemic control [8].

In T1DM, insulin therapy and control of CH intake can be triggers for an ED [8]. The omission of insulin with the purpose of weight loss is a fundamental characteristic of eating disorders in the context of T1DM [8]. An insulin deficit prevents the normal use of blood glucose as an energy source, this results in weight loss, regardless of energy intake, as the body is compelled to generate an alternative energy source through the breakdown of stored fat. Omitting insulin, or administering less insulin than necessary, is an unsafe method of preventing weight gain, which many refer to as diabulimia. It has been associated with poor metabolic control, ketoacidosis, prolonged hospitalizations and a high risk of morbidity and mortality [6].

In addition to insulin therapy, controlling T1DM involves intensive nutritional education and self-monitoring, including CH counting and food weighing. Living with T1DM can be a huge burden for young adults and can lead to complicated relationships with food, nutrition, and body shape [6,8]. Detailed meal planning with monitoring of food portions and CH intake are essential for managing glycemic changes but can lead to disturbed eating and body weight behaviors [8]. The management of T1DM and ED is complex and requires specific diagnostic tools designed for these patients. Perfectionism and frustration can result from anxiety about gaining complete control of the disease. This psychological burden can exacerbate maladaptive behaviors regarding therapeutic adherence, especially in adolescents and young adults [8].

2. Materials and Methods

2.1. Location, Duration, and Study Period

Data was collected using an online questionnaire on the Google Forms platform, which was disseminated through the media of various diabetes associations and through social networks. Data was collected and processed between March and May 2023.

2.2. Type of Study, Type and Technique of Sampling, and Sample Size

Firstly, the study is characterized as analytical, observational and, finally, cross-sectional, since there was no continuous evaluation of the population under study. Sampling was non-probabilistic, and the technique used was by convenience, as participants were selected according to specific criteria and at the convenience of the researcher.

2.3. Target Population, Study Inclusion and Exclusion Criteria

The study population consisted of patients with T1DM, young adults and adults aged between 18 and 60 of both sexes, living in Portugal. Those who did not have T1DM, did not live in Portugal, and were under 18 or over 60 were excluded from the study.

2.4. Instruments

The data collection instrument used was a questionnaire on the Google Forms platform, consisting of questions related to demographic data, data referring to T1DM, anthropometric data, and the EAT-26.

The EAT-26 is a standardized psychometric test, used as a validated instrument for assessing the risk of developing eating disorders by self-reporting symptoms and concerns characteristic of eating disorders [10].

The test consists of 26 questions divided into 3 scales. Each question includes 6 answer options, with responses ranging from 0 = “never” to 3 = “always”, except for question 26 where the score is inverted [8]. Individuals who score 20 points or more on the test are considered to be at risk of an ED or to have high concerns about weight or eating [10].

To conclude whether there is a statistically significant association between the variables, the sum of the scores of the EAT-26 questions was transformed into a variable, which was then divided into two classifications: “no risk” for a score below 20 and “at risk” for a score greater than or equal to 20.

Data processing and statistical analysis were carried out using IBM SPSS Statistics version 28. To assess the significance of associations between qualitative variables, the Chi-square test of independence and Fisher’s exact test were used, with a 95% confidence level for a random error of less than 5%.

3. Results

3.1. Characterization of the Sample

This study resulted in a total sample of 47 participants with T1DM, aged between 18 and 60, with the 26–35 age group being the most representative (31.9%), as shown in

Table 1. Characterization of study sample.

		n	%
Age group	Between 18 and 25 years old	14	29.8
	Between 26 and 35 years old	15	31.9
	Between 36 and 45 years old	13	27.7
	Between 46 and 60 years old	5	10.6
Gender	Female	37	78.7
	Male	10	21.3
BMI	Underweight (<18.5 kg/m2)	2	4.3
	Normal Weight (18.5–24.9 kg/m2)	30	63.8
	Overweight (25–29.9 kg/m2)	15	31.9
	Obesity (≥30 kg/m2)	0	0
Recognition of the concept of diabulimia	Yes	14	29.8
	No	33	70.2
Practice of insulin omission	Yes	3	6.4
	No	44	93.6

. The majority of participants were female (78.7%).

Most of the participants (63.8%) had a normal BMI (18.5–24.9 kg/m²) and the average BMI was 23.4 kg/m², with a minimum and maximum value of 15.7 kg/m² and 28.9 kg/m², respectively.

Regarding recognition of the concept of diabulimia, around 70% of the sample did not recognize this term, and a total of 3 participants (6.4%) reported having already practiced intentional insulin omission.

3.2. EAT-26 Score

The classification of the questionnaire was divided into “not at risk” and “at risk”. Participants who scored less than 20 were not at risk of developing an ED and did not have significant diet and weight concerns. A total of 20 or more points classified participants with a potential risk of an ED who demonstrated high concerns about diet and weight.

From the sample obtained (

Table 2. Characterization of EAT-26 score.

		n	%
Classification	No risk	33	70.2
	At risk	14	29.8

), around 29.8% of the participants obtained a score greater than or equal to 20, i.e., were at risk, and 70.2% a score less than 20. The minimum score obtained was 1 and the maximum was 56. The average score was 15.3 with an associated standard deviation of 12.5.

3.3. Relationship Between Gender and EAT-26 Score

Analyzing the data in

Table 3. Gender characterization related to EAT-26 score.

		EAT-26 Score
Gender		No Risk	At Risk	Significance
Feminine	n	24	13	0.123
	%	64.9	35.1
Masculine	n	9	1
	%	90	10

, there is no statistically significant association (p = 0.123) between gender and the total EAT-26 score.

3.4. Relationship Between Age Group and EAT-26 Score

The results relating to the association between age group and EAT-26 score (

Table 4. Characterization of age group related to EAT-26 score.

		EAT-26 Score
Age Group		No Risk	At Risk	Significance
Young Adults 1	n	27	11	0.796
	%	71.1	28.9
Adults 2	n	6	3
	%	66.7	33.3

¹ Between 18 and 25 years old; ² Between 26 and 60 years old.

) show that there is no statistical significance (p = 0.796) between the two variables.

3.5. Relationship Between Insulin Administration Method and EAT-26 Score

Analyzing the relationship between the insulin administration method and the score obtained on the EAT-26 (

Table 5. Characterization of insulin administration method related to EAT-26 score.

		EAT-26 Score
Insulin Administration Method		No Risk	At Risk	Significance
Insulin Infusion Pump (IIP)	n	20	4	0.083
	%	83.3	16.7
Insulin Pen (IP)	n	12	10
	%	54.5	45.5
Both	n	1	0
	%	100	0

), there was no statistically significant association between the variables (p = 0.083). Although the number of participants using insulin pens is lower (n = 22 compared to n = 24 participants using IIP), the percentage at risk of developing an ED is higher (45.5% compared to 16.7% at risk of developing an ED using IIP). Given the non-significance of the results, this can only be acknowledged as an observational finding.

3.6. Relationship Between CH Counting Method and EAT-26 Score

Considering the results obtained (

Table 6. Characterization of CH counting method related to EAT-26 score.

		EAT-26 Score
CH Counting Method		No Risk	At Risk	Significance
Grams of CH	n	20	8	0.846
	%	71.4	28.5
CH equivalents	n	6	2
	%	75	25
Both	n	7	4
	%	63.6	36.4

), the analysis of the relationship between the CH counting method and the EAT-26 classification reveals that there is no statistically significant relationship between the variables (p = 0.846).

3.7. Relationship Between BMI and EAT-26 Score

Table 7. Characterization of the BMI value related to the EAT-26 score.

		EAT-26 Score
BMI		No Risk	At Risk	Significance
Underweight (18.5 kg/m2)	n	0	2	0.003
	%	0	100
Normal Weight (18.5–24.9 kg/m2)	n	24	6
	%	80	20
Overweight (25–29.9 kg/m2)	n	9	6
	%	60	40
Obesity (≥30 kg/m2)	n	0	0
	%	0	0

shows the results regarding the relationship between the participants’ BMI and the score obtained on the EAT-26. There was a statistically positive association (p = 0.03) between BMI and the EAT-26 score. In this case, 100% (n = 2) of the underweight participants, 20% (n = 6) of the normal-weight participants and 40% (n = 6) of the overweight participants were at risk of developing an ED. Given the small sample, these results should be interpreted with caution.

4. Discussion

In the present study, 29.8% of participants with T1DM were at risk of developing an ED. According to the literature, between 14% and 35% of patients with diabetes are screened positive for an ED using EAT-26 [11,12]. The results of this research fall within this range.

With regard to gender, in this study around 78.7% (n = 37) belonged to the female gender, which represents the greater participation, which may explain the fact that there was no statistically significant association between gender and the EAT-26 score, even though the female population had a percentage of participants at risk of developing ED of 35.1%. Males accounted for 21.3% (n = 10) of the 47 total participants, of whom only 1 (10%) had a score equivalent to the risk of developing an ED, i.e., greater than 20 points on the EAT-26.

The differences in the prevalence of ED between males and females is a topic that has been discussed internationally, including in patients with T1DM. The literature is heterogeneous on this subject, divided between studies that support a difference between genders and others that do not. From our data, we can only conclude that there was a trend suggesting that males may be at lower risk; however, the difference was not statistically significant, likely due to the small sample size. Numerous studies have concluded that there is a higher prevalence of EDs in females, even within the population with T1DM [12,13,14,15]. However, there are several studies that indicate no significant differences between genders and the risk of developing an ED in populations with T1DM. According to Albaladejo et al., the prevalence of the risk of an ED was not significantly different between men and women, as other studies have shown [16,17,18,19]. Thus, part of this subject may be controversial, with no single conclusion to be drawn as to which gender has a greater risk of developing ED in the T1DM population.

This study also aimed to understand whether age modified the risk of developing an ED in individuals with T1DM. Participants were categorized into young adults (n = 38) and adults (n = 9), with 28.9% of young adults and 33.3% of adults identified as being at risk for developing an ED based on EAT-26 scores. While no statistically significant association was observed between age group and risk status, the slightly higher proportion among adults constitutes a descriptive finding only and does not support any inferential conclusions. The literature shows that an ED can occur in various age groups from adolescence to young adulthood and middle age [20,21,22]. In addition, there is a large majority of studies on ED in adolescents with T1DM compared to studies carried out in young adults and adults. Even so, there is a tendency for the older the age, the lower the risk of developing an ED in the T1DM population, concluding that young age is associated with a higher risk of developing an ED, contrary to the results of this study [13,17,19].

Considering the insulin administration method, 24 participants (51.1%) used an insulin infusion pump compared to 22 insulin pen users. In this case, IP users had a substantially higher percentage of being at risk of developing an ED, 45.5%, compared to 16.7% of IIP users, and almost half of these IP users had a score of over 20 on the EAT-26.

Evidence on the possible relationship between the insulin administration method, IIP or IP, and the development of an ED is scarce. Recently, a systematic review and meta-analysis of 17 studies, most of which included adolescents, showed that the use of diabetes technology could be associated with a lower risk of ED [19,23]. It is possible that the greater flexibility of IIP could improve eating habits and quality of life and could be associated with a lower practice of insulin omission, also correlating with a lower risk of developing an ED in T1DM [23]. Policola et al. concluded in their research that there were no significant differences between IIP or IP users and the development of an ED; however, the authors report that IP users had a greater bulimic tendency than IIP users [19]. Other studies support this conclusion, and several authors report that the use of IIP is associated with a lower prevalence of EDs since it allows greater flexibility and better control in T1DM [14,16,17].

The CH counting method also showed no statistically significant association with the EAT-26 score; the literature does not address this distinction in studies carried out on patients with T1DM and related to ED. However, it would be relevant to investigate in the future due to the differences in requirements, practice, and responsibility that the different methods entail, and the importance of CH counting as an appropriate method for glycemic control in T1DM [24,25,26].

Regarding BMI and the EAT-26 score, there was a statistically positive association (p = 0.033). This association has been extensively described in the literature [13,14,16,17,27]. Given the small number of participants in this study, the data should be interpreted with caution. According to the literature, we would expect to see a relationship between a higher BMI and an increased risk of developing an ED [13,14,17,28]. In this case, 40% (n = 15) of the overweight participants had an EAT-26 score higher than 20, categorizing them as being at risk of developing an ED. Watt et al. showed in their study that a score higher than 20 points was associated with a higher BMI and that the prevalence of behaviors associated with an ED was 37% in overweight and obese participants compared to 21% in normal-weight participants [13]. Being overweight or obese can be associated with a desire to lose weight, which often results in a negative body image and maladaptive behaviors such as dieting, insulin restriction, and behaviors associated with ED, with binge eating disorder being the most prevalent in the T1DM population [29]. Insulin restriction is also referred to in the literature as a serious behavior in patients with T1DM, and is shown to be associated with T1DM-related stress, the feeling of losing control of the disease, and the fear of weight gain [22]. In our study, only 6.4% (n = 3) of the participants revealed the practice of insulin omission.

It can also be observed that 100% (n = 2) of the underweight participants were at risk of developing an ED, which contrasts with the trend in the T1DM population according to most of the literature. However, it is important to note that this may be associated with other types of ED that are less prevalent in the T1DM population, which should also be a cause for concern and warrant clinical attention [13,14,29]. The percentage of participants with normal weight at risk of developing an ED was 20% (n = 6).

5. Study Limitations

Firstly, the study sample is small, and despite dissemination through channels directly related to the T1DM population, adherence was lower than expected.

Due to the fact that the data was collected using a self-administered questionnaire, there is a possibility that participants may have misunderstood, misinterpreted, or incorrectly answered the questions, even though they were validated.

Another issue related to the questionnaire is that anthropometric data were obtained through it, which raises concerns about their accuracy, as the measurements were not taken by the researcher.

Another limitation to highlight is the lack of adaptation of the EAT-26 for individuals with diabetes, despite the inclusion of a question regarding insulin restriction. Therefore, a limitation of the study is the use of a non-adapted assessment tool for individuals with T1DM, despite the tool being validated.

It should be noted that, given the cross-sectional nature of the study, the results should be interpreted with caution, since cross-sectional data cannot determine directionality or causality.

6. Conclusions

Given the requirements of T1DM management regarding meal planning, dietary education, and monitoring the amount of CH ingested, it is not surprising that the population with diabetes is susceptible to developing disturbed eating behaviors. This research only revealed positive results in the relationship between BMI and the risk of developing an ED in individuals with T1DM, however the literature shows that the specificities of T1DM and its relationship with different variables studied are the subject of ongoing research, aimed at better understanding the relationship between these two pathologies.

The coexistence of T1DM and ED is of critical concern due to its significant impact on the health and well-being of these populations. There must be recognition of the complexity and importance of experience in dealing with both pathologies, in order to avoid serious T1DM-related outcomes. Frequent assessment of the risk of developing an ED in individuals with T1DM appears to be crucial for early diagnosis and intervention, given the associated risk of increased morbidity and mortality.

More studies are needed to better understand the relationship between these two pathologies, so that multidisciplinary clinical action can be carried out in the most appropriate and beneficial way for the population that will live with chronic T1DM, as well as the development of more ED risk screening tools specific to T1DM and preventive interventions and guidelines, that can help the various areas of health that support the population with T1DM.