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

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Eating disorders in people with diabetes is an interesting topic and certainly requires attention. 
In this article, people with diabetes were asked via social media to fill in a questionnaire. Responses were obtained from 47 people, mostly women. The authors themselves describe all their limitations.  It is difficult to draw conclusions from this study because the answers cannot be considered representative of people with type 1 diabetes. It would be better to have the questionnaire completed by a specific patient group in order to draw more correct conclusions. 

Suggestions: If the article is submitted to a journal dedicated to diabetes, the introduction section on diabetes can be much shorter because readers will be aware of the general aspects on diabetes.

 

Comments on the Quality of English Language

Small remark: as "data" comes from latin and as it is the plural form of the noun "datum", plural sounds better although it is in modern English used with both singular and plural verbs.

Author Response

Firstly, we sincerely appreciate your time and effort in reviewing this manuscript. Below, you will find our responses, Below, you will find our responses, which we have carefully reviewed and considered with the utmost attention, taking into account all suggestions and corrections.

Comments 1: The English could be improved to more clearly express the research.

Response 1: By carefully reading and re-reading the article, we believe that the quality of the English fits into the demanding context that a publishable article requires. However, we remain available to review and refine any specific phrase or word that you believe requires modification.

Comments 2: It is difficult to draw conclusions from this study because the answers cannot be considered representative of people with type 1 diabetes. It would be better to have the questionnaire completed by a specific patient group in order to draw more correct conclusions. 

Response 2: In this study, our objective was to recruit participants who had a confirmed diagnosis of T1DM and who constituted a representative sample of individuals living with T1DM in Portugal. This representativeness was ensured through the predefined criteria established for questionnaire responses. While we acknowledge that data collection from a more rigorously controlled and systematically selected sample could have enhanced the robustness of our findings, the constraints at the time led us to conclude that this approach was the most appropriate and possible for reaching a broader population for the study.

Comments 3: If the article is submitted to a journal dedicated to diabetes, the introduction section on diabetes can be much shorter because readers will be aware of the general aspects on diabetes.

Response 3: In this case, we aimed to ensure that the introduction was both comprehensive and concise, as its primary objective was to rigorously contextualize each of the pathologies and their relationship, according to the available literacy. Accordingly, we sought to provide a clear and accessible overview, allowing even readers with less experience or literacy in any of the discussed topics to gain a foundational understanding and appropriate contextualization of each subject. Consequently, this has resulted in a more extensive introduction.

Comments 4: Small remark: as "data" comes from latin and as it is the plural form of the noun "datum", plural sounds better although it is in modern English used with both singular and plural verbs.

Response 4: We carefully analyzed this suggestion, as it raised some doubts. Ultimately, we decided to retain the term 'data' in the singular form, as it appears to be the more commonly used convention in the majority of the literature we consulted, including the sources referenced in our article. Nonetheless, if in any specific instance the plural form of 'data' is deemed more appropriate, we sincerely appreciate the suggestion.

Thank you once again for your careful attention to our article.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors sought to determine the prevalence Portuguese young adults and adults with T1DM that were at risk for eating disorders (ED) based on an online questionnaire (N=47, 26-35y) and how the variables of gender, age group,method of insulin administration, carbohydrate counting and BMI influenced the risk of developing ED. Risk of eating disorders was determined with the EAT-26 psychometric test where a score >20 was classified as at risk for ED and <20 was classified as not at risk for ED. Based on this dichotomy, 29.8% of the sample was classified as at risk for ED and only BMI was associated with risk for an ED with a positive association between BMI and the EAT-26 score. The rationale for the study was unclear in the introduction and the significance of the results were unclear in the discussion.

Major.

1. Results: table 1 shows that 4.3% of the sample were underweight and 63.8% of the sample were normal weight. Address in the discussion how do both percentages compare to literature for individuals in this age category with eating disorders.
2. Results lines 182-183. The data in Table 5 supporting the following statement is unclear. “Although the number ofparticipants using insulin pens is lower, the percentage at risk of developing ED is notoriously” Adding a statistic to support the claim would improve the manuscript since the number of people at risk versus not at risk is only different by 2 people and the difference in the percentages are more likely due to the small sample size.
3. Discussion lines 229-230. The results supporting the following statement are unclear “Although there was nostatistically significant association between the EAT-26 score and age group, this study may demonstrate agreater risk of adults with T1DM developing an ED”.
4. Methods: state whether the Eat-26 test has been validated for determining eating disorders and add the citations.
5. Introduction Lines 40-41: Clarify if the statistic (670 new cases of diabetes were diagnosed every year for every 100,000…) is specific for type I diabetes or diabetes in general.
6. Introduction liens 86-90: the introduction does not provide a strong rationale for the study. For example, there is 1 citation in this paragraph and it is unclear which population the paragraph refers to. Expanding this paragraph to include the statistics for Portugal would strengthen the rationale. What is the prevalence of eating disorders and separately what is the incidence of TIDM, what does the literature indicate for the incidence of ED and or DEB with T1DM in a comparable population to Portugal. Also, describe what is already known in the literature for the relationship between the following variables (gender, age group, method of insulin administration,carbohydrate counting and BMI) on the risk of developing
7. Abstract – state in the abstract that there was a positive association between BMI and the EAT-26 score.

Author Response

Firstly, we sincerely appreciate your time and effort in reviewing this manuscript. Below, you will find our responses, which we have carefully reviewed and considered with the utmost attention, taking into account all suggestions and corrections. Additionally, all changes and corrections are highlighted in red within this response for clarity.

Comments 1: Results: table 1 shows that 4.3% of the sample were underweight and 63.8% of the sample were normal weight. Address in the discussion how do both percentages compare to literature for individuals in this age category with eating disorders.

Response 1: In this case, after careful analysis, we would like to clarify that, given the topic of eating disorders, our objective is not to make BMI comparisons, as we do not consider this to be a key aspect to highlight within the broader context of our study. More importantly, the BMI data pertains exclusively to participants with T1DM and not to individuals with ED, as our study does not allow us to identify or diagnose eating disorders. These BMI values correspond to participants for whom the primary condition was T1DM, not ED. To draw such conclusions, it would likely be necessary to conduct a study within a controlled and specifically selected group, where one of the inclusion criteria would be a prior diagnosis of an ED.

Comments 2: Results lines 182-183. The data in Table 5 supporting the following statement is unclear. Although the number ofparticipants using insulin pens is lower, the percentage at risk of developing ED is notoriously” Adding a statistic to support the claim would improve the manuscript since the number of people at risk versus not at risk is only different by 2 people and the difference in the percentages are more likely due to the small sample size.

Response 2: In our interpretation, and following the objective we aimed to highlight in this statement, the results presented in the table show that 24 participants use an insulin pump, 22 use an insulin pen, and 1 utilizes both methods of insulin administration. Among the 22 individuals using an insulin pen (IP), approximately 45% are at risk of developing an ED, compared to 16.7% of the 24 individuals using an insulin pump (IIP). While we fully acknowledge that these results are not statistically significant—and we consistently emphasize this point—the observation we made aimed to highlight that, given the nearly equal number of participants using IIP and IP, the percentage of ED risk was considerably higher among IP users. We consider this an important observation, given the nearly equal number of participants in both groups and the higher ED risk indicated by the EAT-26 results for IP users. Despite this explanation, we have revised the paragraph to make it more specific and clear, avoiding potential misinterpretations by adding the relevant statistics for each variable. 

"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 ED is notoriously higher (45,5 % compared to 16,7 % at risk of developing ED using IIP)."

Comments 3:  Discussion lines 229-230. The results supporting the following statement are unclear “Although there was nostatistically significant association between the EAT-26 score and age group, this study may demonstrate agreater risk of adults with T1DM developing an ED”.

Response 3: In this case, we have revised the sentence to ensure it is clear that we are merely reporting our observations based on the results, providing a basis for discussion and comparison with the existing literature, despite the lack of a statistically significant association of our results.

“The results were divided into young adults (n = 38) and adults (n = 9), of which 28.9% of young adults were at risk of developing ED compared to 33.3% of adults. Although there was no statistically significant association between the EAT-26 score and age group, our results indicate that adult participants with T1DM exhibited a higher percentage of individuals at risk for developing an ED.”

Comments 4: Methods: state whether the Eat-26 test has been validated for determining eating disorders and add the citations.

Response 4: This statement had already been properly referenced; We have included a more concise sentence to avoid any ambiguity. It is also important to emphasize that the EAT-26, as stated by its authors, is not intended for diagnosing eating disorders but rather for identifying the risk of developing such conditions.

"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."

Comments 5: Introduction Lines 40-41: Clarify if the statistic (670 new cases of diabetes were diagnosed every year for every 100,000…) is specific for type I diabetes or diabetes in general.

Response 5: We have clarified this by adding a parenthesis. "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."

Comments 6: Introduction liens 86-90: the introduction does not provide a strong rationale for the study. For example, there is 1 citation in this paragraph and it is unclear which population the paragraph refers to. Expanding this paragraph to include the statistics for Portugal would strengthen the rationale. What is the prevalence of eating disorders and separately what is the incidence of TIDM, what does the literature indicate for the incidence of ED and or DEB with T1DM in a comparable population to Portugal. Also, describe what is already known in the literature for the relationship between the following variables (gender, age group, method of insulin administration,carbohydrate counting and BMI) on the risk of developing

Response 6: First of all, thank you for this thorough and insightful comment, which is highly valuable for improving the content of our introduction. 

In response to this comment, we have sought to improve the content of our introduction by incorporating the suggested revisions. Firstly, the reference for this information also does not specify the particular population from which this fact was derived, and for this reason, we do not have specific population data beyond the fact that it refers to patients with T1DM. Regarding Portuguese statistical data, unfortunately, Portugal does not have national statistics on patients with both T1DM and ED simultaneously. To address the prevalence of ED in the Portuguese population, we have included a reference to an older study, which, at present, remains the most robust available data on ED prevalence in Portugal. 
“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 ED in Portugal dates to 1999, in which 2,398 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.”

The incidence of diabetes in Portugal has already been mentioned in the article, based on data from the latest Annual Report of the Portuguese National Diabetes Observatory (2023) (Reference No. 3). However, this report does not distinguish between type 1 and type 2 diabetes. Additionally, we have incorporated a reference to an Italian study, which examines a Mediterranean population with conditions similar to those of the Portuguese population, providing specific data on ED prevalence among patients with T1DM, joining the already available data on our study. “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.”

Thus, we believe we have addressed the suggestions as effectively as possible within the constraints of the existing literature.

Comments 7: Abstract – state in the abstract that there was a positive association between BMI and the EAT-26 score.

Response 7: This point was already stated in our abstract initially.

Thank you once again for your careful attention to our article.

Reviewer 3 Report

Comments and Suggestions for Authors

This is an important topic to explore. A few comments follow but my main suggestion for improvement would be to also look at the EAT-26 score percentiles which may give a clearer picture of the level of risk. I would also suggest that the introduction and conclusions are a little wordy- it may hold the readers interests better to make these more concise.

Line 86-89- please include the prevalence of disturbed eating behaviours in the nondiabetic population.

Table 3. A trend to males being less at risk but not significant probably due to the low numbers

Table 5 line 184- there was a trend to a higher risk of eating disorders with basal bolus cf pump therapy

Line 198- add to interpret with caution due to low numbers (covered in discussion)

Line 216. The prevalence of female participants does not inform as to the risk of females developing an ED

Line 229. This study showed no evidence of more ED in young adults compared to adults

Author Response

Firstly, we sincerely appreciate your time and effort in reviewing this manuscript. Below, you will find our responses, Below, you will find our responses, which we have carefully reviewed and considered with the utmost attention, taking into account all suggestions and corrections. Additionally, all changes and corrections are highlighted in red within this response for clarity.

Comments 1: This is an important topic to explore. A few comments follow but my main suggestion for improvement would be to also look at the EAT-26 score percentiles which may give a clearer picture of the level of risk. I would also suggest that the introduction and conclusions are a little wordy- it may hold the readers interests better to make these more concise.
Response 1: First of all, we appreciate your recognition of the importance of our topic.

In this case, we aimed to ensure that the introduction was both comprehensive and concise, as its primary objective was to rigorously contextualize each of the pathologies and their relationship, according to the available literacy. Accordingly, we sought to provide a clear and accessible overview, allowing even readers with less experience or literacy in any of the discussed topics to gain a foundational understanding and appropriate contextualization of each subject. Consequently, this has resulted in a more extensive introduction.
 
Regarding the percentiles of the EAT-26, the questionnaire provides information on whether there is a risk of developing an eating disorder (ED) based on the total score obtained across all items. After reviewing this aspect, and considering that the EAT-26 is designed to determine the presence of risk, we acknowledge that risk is identified when the total score reaches or exceeds 20 points. In our study, the objective was solely to assess the presence or absence of risk. Therefore, in our view, reporting a higher score that may indicate a greater risk would not be relevant within the context of our research. We hope to have correctly understood this point and provided an adequate justification.

Comments 2: Line 86-89- please include the prevalence of disturbed eating behaviours in the nondiabetic population.
Response 2: We added the sentence regarding the prevalence in nondiabetic population.
"ED, or even disturbed eating behaviors (DEB), are more prevalent in patients with T1DM than in the general 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% [7]. "

Comments 3: Table 3. A trend to males being less at risk but not significant probably due to the low numbers
Response 3: In Table 3, as well as in all our results tables, we consistently state when there is no statistical significance. Therefore, we would like to clarify whether this suggestion refers to a phrase intended to replace these statements in the results section or rather a way to incorporate this point into the discussion. Our objective is always to make it clear when conclusions cannot be drawn due to the lack of statistical significance, which is likely a consequence of the small sample size.

Comments 4: Table 5 line 184- there was a trend to a higher risk of eating disorders with basal bolus cf pump therapy
Response 4: In this case, we also believe that this conclusion is adequately stated through the sentence "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 ED is notoriously higher (45,5 % compared to 16,7 % at risk of developing ED using IIP)", while always emphasizing that there is no statistically significant association between the variables presented.

Comments 5: Line 198- add to interpret with caution due to low numbers (covered in discussion)
Response 5: We added the sentence "Given the small sample, these results should be interpreted with caution."

Comments 6: Line 216. The prevalence of female participants does not inform as to the risk of females developing an ED
Response 6: We have revised this sentence to clarify that we are neither drawing general conclusions due to the lack of a statistically significant association, something we aimed to consistently highlight in our results, nor establishing a relationship between the number of female participants and a higher risk. Our objective is solely to acknowledge that our study indeed has a higher participation of females, which may influence the results.
"Given the higher prevalence of female participants in this study, our findings align with the established evidence of an increased risk of ED in females with T1DM. Accordingly, we observe a trend suggesting that males may be at lower risk; however, this difference is not statistically significant, due to the small sample size. "

Comments 7: Line 229. This study showed no evidence of more ED in young adults compared to adults
Response 7: In this case, we have also revised the sentence to ensure it is clear that we are merely reporting our observations based on the results, providing a basis for discussion and comparison with the existing literature, despite the lack of a statistically significant association.
"Although there was no statistically significant association between the EAT-26 score and age group, our results indicate that adult participants with T1DM exhibited a higher percentage of individuals at risk for developing an ED."

Thank you once again for your careful attention to our article.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for the revision. Well adapted. I still think the introduction is a bit too extensive.

Author Response

We would like to express our sincere gratitude once again for your thorough review of our manuscript.

Comment 1: Thank you for the revision. Well adapted. I still think the introduction is a bit too extensive.

Response 1: As per your suggestion and other reviewers, we have revised and shortened some sections of the introduction, including reducing the discussion on T1DM symptoms and insulin treatment.

Reviewer 3 Report

Comments and Suggestions for Authors

Thankyou for your responses- please see comments below

Two of the reviewers have suggested that the introduction is too wordy. I appreciate your response but your readers already know the usual symptoms of diabetes and how to manage it with insulin- please shorten this.

Lines 92-97 are confusing with an unreferenced prevalence of ED of 21% in adults with type 1 diabetes followed by a referenced prevalence of 13.3%

Comment from reviewer 2 : comment 2- I cannot see where you have emphasised the point that the difference in ED between IP and IIP users is not statistically significant. Please remove the words notoriously and substantially and acknowlede the nonsignificance of the difference in words.

Comment 3 reviewers 2 &3 and response- Are you trying to say that 28.9% is significantly less than 33.3%? If not you cant say that your results indicate that adult participants with T1DM exhibited a higher percentage of individuals at risk for developing an ED.

Reviewer 3- You don't seem to understand the point. The higher prevalence of females participants may just indicate that females are more likely to participate in a study- you can only say from your data that there was a trend suggesting that males may be at lower risk however the difference was not statistically significant due to the small sample size.

 

Author Response

We sincerely appreciate your ongoing careful review and the detailed attention you have dedicated to our manuscript. Your constructive feedback has significantly contributed to enhancing the quality of our work.

Comment 1: Two of the reviewers have suggested that the introduction is too wordy. I appreciate your response but your readers already know the usual symptoms of diabetes and how to manage it with insulin- please shorten this.

Response 1: In accordance with your recommendations, we have made revisions to various sections of the introduction and reduced the content related to T1DM symptoms and insulin treatment and other points. We believe that the revised version is now more concise and direct, assuming a background knowledge of type 1 diabetes.

Comment 2: Lines 92-97 are confusing with an unreferenced prevalence of ED of 21% in adults with type 1 diabetes followed by a referenced prevalence of 13.3%

Response 2: The reference had been omitted due to the initial revisions, but it has now been reinstated. The prevalence of 13.3% refers to a different study aimed at demonstrating the prevalence in a population similar to that of Portugal, as no studies on the prevalence of ED in individuals with T1DM or T2DM in Portugal are available. This addition was suggested by a reviewer to better complete this paragraph.

Comment 3: Comment from reviewer 2 : comment 2- I cannot see where you have emphasised the point that the difference in ED between IP and IIP users is not statistically significant. Please remove the words notoriously and substantially and acknowlede the nonsignificance of the difference in words.

Response 3: Regarding the emphasis on the lack of statistical significance, this is addressed in lines 180-182: “Analyzing the relationship between the insulin administration method and the score obtained on the EAT-26 (Table 5), there was no statistically significant association between the variables (p = 0.083).” The word "notoriously" has been removed, and we did not find the use of the word "substantially." Additionally, we have included the phrase: "Given the non-significance of the results, this can only be acknowledged as an observational finding," to avoid any further misinterpretation.

"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 ED is higher (45,5 % compared to 16,7 % at risk of developing ED using IIP). Given the non-significance of the results, this can only be acknowledged as an observational finding."

Comment 4: Comment 3 reviewers 2 &3 and response- Are you trying to say that 28.9% is significantly less than 33.3%? If not you cant say that your results indicate that adult participants with T1DM exhibited a higher percentage of individuals at risk for developing an ED.

Response 4: In response to your correction, we modified the text as follows: Lines 229-233 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.

Comment 5: Reviewer 3- You don't seem to understand the point. The higher prevalence of females participants may just indicate that females are more likely to participate in a study- you can only say from your data that there was a trend suggesting that males may be at lower risk however the difference was not statistically significant due to the small sample size.

Response 5: We have removed the sentence regarding the prevalence in females and have rewritten it as follows: Lines 218-220 “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.”