Growth Patterns and Factors Associated with Short Stature in Saudi Children and Adolescents with Type 1 Diabetes Mellitus: A Retrospective Cross-Sectional Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In general

The experimental design is appropriate for assessing variables related to growth in diabetic children and adolescents.

The results presented support the hypothesis stated in the introduction.

The study employs a statistical design suitable for evaluating factors associated with short stature or growth retardation.

Because they named it that way, “Vitamin D status” yes, it is actually a serum concentration of vitamin D.

Comments on the Quality of English Language

none

Author Response

Reviewer # 1

In general

• The experimental design is appropriate for assessing variables related to growth in diabetic children and adolescents.

Thank you for your positive assessment. We appreciate the reviewer’s recognition that this study is appropriate for evaluating growth-related variables in diabetic children and adolescents. Our study was carefully structured to account for key clinical and developmental factors, ensuring that the design aligns with the research objectives. This confirmation supports the validity of our methodological approach.

• The results presented support the hypothesis stated in the introduction.

Thank you for this encouraging comment. We are pleased that the reviewer finds the results to be consistent with and supportive of the hypothesis outlined in the introduction. Our analysis was designed to directly address the proposed hypothesis, and we have ensured that the findings are clearly presented and appropriately interpreted in relation to it. This alignment reinforces the overall coherence of the study.

• The study employs a statistical design suitable for evaluating factors associated with short stature or growth retardation.

Thank you for this thoughtful comment. We appreciate the reviewer’s acknowledgment that statistical design is appropriate for evaluating factors associated with short stature and growth retardation. Particular care was taken in selecting analytical methods (propensity score matching and regression analysis) that account for potential confounders and allow for robust assessment of associations. This feedback supports the rigor and suitability of our analytical approach.

• Because they named it that way, “Vitamin D status” yes, it is actually a serum concentration of vitamin D.

Thank you for this clarification. We agree that the term “Vitamin D status” in our manuscript refers specifically to serum vitamin D concentration. To improve precision and avoid ambiguity, we will revise the text to explicitly state “serum vitamin D concentration” where appropriate, while retaining “Vitamin D status” only in contexts where it is clearly defined.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Although the study addresses an interesting topic with practical clinical applicability, the study design does not allow for a correct analysis of the risk factors for short stature in patients with type 1 diabetes.
Thus, although a control group is included, not all clinical-biological data are presented or compared, the analysis that would allow for the evaluation of the differences between patients with type 1 diabetes and controls, in terms of height and weight status and other factors independent of type 1 diabetes. No data are presented regarding lifestyle, BMI, HbA1c, vitamin D status in the control group, in order to evaluate their impact on short stature.

In the Material and method section, the cutoff level of the physical activity level to differentiate between yes and no is not mentioned.
Why was the value of 0.2 for p from the univariate analysis chosen as defining for the inclusion of parameters in the multivariate analysis (lines 179-180)?
In the Results section (lines 205-213) are mentioned criteria for defining different parameters that have already been included in the Material and method section.
In Table 2 for the parameters height, weight and BMI, although IQR is mentioned, the table does not mention IQR as an interval, probably the values ​​represent mean and standard deviation.
In patients with T1DM, the insulin dose/Kg should be included considering that they are more frequently overweight/obese compared to the control group, and the insulin dose can be associated with the stature status. TBR and TAR should also be included to see if frequent/severe hypoglycemia influences the short stature status.
The distribution of height, weight, and BMI across all age groups and in the control group should be included in the analysis and compared with that identified in patients with type 1 diabetes.
The prevalence, general and by age group, of short stature in the control group should be mentioned.
In Table 4, the percentages for the distribution of normal and short stature respectively by parameter categories are not reported correctly. Thus, for example, for the duration of diabetes:
- 6.67%, 80% and 13.33% of those with short stature had newly diagnosed diabetes, intermediate duration and long duration respectively
- 30.62%, 63.63% and 5.74% of those with normal stature had newly diagnosed diabetes, intermediate duration and long duration respectively.
The Discussion section addresses many theoretical data that are not included in the current study and thus, the results of the current study cannot be reported to these data.
Reference 14 is identical to reference 13.
Considering the previous comments, the study design is not suitable for a proper analysis of risk factors for short stature in patients with type 1 diabetes, the main limitations of the study being the incomplete analysis of potential risk factors known to influence height status, the incomplete comparison between the 2 groups included, the cross-sectional and retrospective design that does not allow the identification of a causal relationship, and the small sample size for the category of subjects with short stature.

Author Response

Reviewer # 1

In general

• The experimental design is appropriate for assessing variables related to growth in diabetic children and adolescents.

Thank you for your positive assessment. We appreciate the reviewer’s recognition that this study is appropriate for evaluating growth-related variables in diabetic children and adolescents. Our study was carefully structured to account for key clinical and developmental factors, ensuring that the design aligns with the research objectives. This confirmation supports the validity of our methodological approach.

• The results presented support the hypothesis stated in the introduction.

Thank you for this encouraging comment. We are pleased that the reviewer finds the results to be consistent with and supportive of the hypothesis outlined in the introduction. Our analysis was designed to directly address the proposed hypothesis, and we have ensured that the findings are clearly presented and appropriately interpreted in relation to it. This alignment reinforces the overall coherence of the study.

• The study employs a statistical design suitable for evaluating factors associated with short stature or growth retardation.

Thank you for this thoughtful comment. We appreciate the reviewer’s acknowledgment that statistical design is appropriate for evaluating factors associated with short stature and growth retardation. Particular care was taken in selecting analytical methods (propensity score matching and regression analysis) that account for potential confounders and allow for robust assessment of associations. This feedback supports the rigor and suitability of our analytical approach.

• Because they named it that way, “Vitamin D status” yes, it is actually a serum concentration of vitamin D.

Thank you for this clarification. We agree that the term “Vitamin D status” in our manuscript refers specifically to serum vitamin D concentration. To improve precision and avoid ambiguity, we will revise the text to explicitly state “serum vitamin D concentration” where appropriate, while retaining “Vitamin D status” only in contexts where it is clearly defined.

 

Reviewer # 2

Dear Reviewer, we sincerely thank you for your thorough, insightful, and constructive critique of our manuscript. Your comments have been instrumental in helping us identify the key methodological limitations of our study and have guided us toward a more honest, accurate, and appropriately cautious presentation of our findings. Especially, the table you asked for that compares different age groups We have carefully addressed each of your concerns point by point, and we believe the revised manuscript is substantially improved in terms of scientific rigor, transparency, and clinical interpretability. Below, we provide our detailed responses to each of your comments.

Comment 1: Although the study addresses an interesting topic with practical clinical applicability, the study design does not allow for a correct analysis of the risk factors for short stature in patients with type 1 diabetes.

We thank the reviewer for this important methodological critique. We acknowledge that the retrospective cross-sectional design of our study has inherent limitations when interpreting associations as causal risk factors. We fully agree that establishing temporality prerequisites for valid risk factor analysis is not possible with this design. Specifically, we cannot definitively determine whether longer diabetes duration preceded the development of short stature in all cases, nor can we exclude reverse causality or the influence of unmeasured confounding variables. To address this concern transparently, we have revised the manuscript in the following ways: Revised language throughout: We have replaced the term “risk factors” with more appropriate terminology, such as “factors independently associated with” or “correlates of” short stature. For example, the abstract and results sections now consistently refer to “independent predictors” only in the statistical sense (i.e., variables that remain significant in multivariable models), while explicitly stating that these should not be interpreted as causal risk factors.

Strengthened limitations section: We have expanded the limitations paragraph to explicitly state:

“However, several limitations must be acknowledged. The retrospective cross-sectional design precludes causal inference and examination of growth trajectories over time. Our cross-sectional measurement of height at a single time precluded assessment of growth velocity, and we lacked data on parental height (to determine familial short stature), pre-diagnosis growth records, pubertal staging (Tanner), bone age, and bone mineral density all essential for comprehensive growth evaluation. We did not systematically screen for autoimmune comorbidities (celiac disease, hypothyroidism) or assess markers of malabsorption and nutritional deficiencies. Regarding glycemic control, a single HbA1c measurement reflects only 2-3 months and showed limited variability; we lacked average HbA1c over 12-24 months, as well as continues glucose monitoring-derived metrics including coefficient of variation (CV), time below range (TBR), and time above range (TAR). Lifestyle factors (diet, physical activity, adherence) were assessed using simple yes/no responses from medical records, introducing recall and social desirability bias. No quantitative dietary intake (energy, protein) was assessed, and socioeconomic status and parental education were not collected. All participants used insulin pumps, precluding comparison with multiple daily injections therapy. Insulin dose (units/kg/day) was not available. Finally, small subgroup sizes (e.g., long-standing diabetes, n=8; short stature, n=30) resulted in wide confidence intervals and limited statistical power. Accordingly, our findings should be viewed as hypothesis-generating rather than confirmatory.”

Revised conclusion: We have tempered the concluding statement to reflect the design limitation:

“This study demonstrates that 12% of children with T1DM were short. Diabetes duration is a significant independent predictor of short stature in children and adolescents with T1DM, with the prevalence of growth impairment increasing progressively in new-onset cases to those with long-standing disease. Notably, while poor glycemic control and vitamin D insufficiency were common in the cohort, neither emerged as statistically significant predictors of short stature in the multivariable analysis. Clinically, these results underscore the importance of longitudinal growth monitoring throughout the course of T1DM, particularly beyond the first five years following diagnosis, to enable early identification and intervention for at-risk children. Future prospective studies are warranted to elucidate the mechanistic pathways linking diabetes duration to growth failure.”

Added to strengths section:  This multicenter study demonstrates several important strengths, including the application of a standardized data collection form, the use of validated Saudi growth charts for anthropometric categorization, and a rigorous statistical framework that employed propensity score matching, univariate and multi-variable logistic regression with appropriate evaluations of model fit. The de-tailed inclusion of clinical, laboratory, and lifestyle parameters such as HbA1c, vitamin D levels, time in range, and self-reported adherence enabled a multi-dimensional assessment of factors linked to short stature. Finally, we also acknowledge that although the study design restricts causal interpretation, an important strength is the inclusion of an age- and sex-matched healthy control group. This comparison clearly shows that children with T1DM have markedly lower height percentiles, a finding that stands on its own and does not depend on causal inference.

We believe these revisions appropriately contextualize our findings and address the reviewer’s concern. We are grateful for the opportunity to improve the scientific rigor and clarity of our manuscript.
Comment 2 : Thus, although a control group is included, not all clinical-biological data are presented or compared, the analysis that would allow for the evaluation of the differences between patients with type 1 diabetes and controls, in terms of height and weight status and other factors independent of type 1 diabetes. No data are presented regarding lifestyle, BMI, HbA1c, vitamin D status in the control group, in order to evaluate their impact on short stature.

Response 2: As stated in the abstract and introduction, the primary aim of this study was: To assess growth patterns and identify factors independently associated with short stature among Saudi school-age children and adolescents with T1DM. The healthy control group was included solely to provide a population-based reference for height percentiles and prevalence of short stature—i.e., to answer the question: “Do children with T1DM have shorter stature than their healthy peers of the same age and sex?” The control group was not intended to:

• Serve as a comparative analytic sample for risk factor analysis (e.g., vitamin D, HbA1c, lifestyle factors).
• Enable assessment of whether factors associated with short stature in T1DM are also operative in healthy children, or
• Allow multivariable modeling including both groups to test for interaction by diabetes status.

Therefore, the absence of lifestyle, vitamin D, and detailed biochemical data in the control group is by design, not an omission, because those variables are not relevant to the control group’s sole function as a height reference.

Comment 3: In the Material and method section, the cutoff level of the physical activity level to differentiate between yes and no is not mentioned.

We thank the reviewer for this important observation. We acknowledge that our original manuscript did not specify a cutoff for physical activity. The honest reason is that no standardized cutoff was used in the original data collection. This was a retrospective cross-sectional study using data extracted from electronic medical records. Physical activity status was documented by treating physicians during routine clinical encounters as part of medical history. The documentation typically consisted of a subjective, non-standardized question such as:

• "Is the patient physically active?" (Yes/No)
• "Does the patient exercise regularly?" (Yes/No)

There was no consistent definition provided to physicians, no specified threshold (e.g., minutes per week, days per week), and no standardized questionnaire. The response was recorded based on clinical judgment and/or parent-reported impression at the time of the visit.

Comment 4: Why was the value of 0.2 for p from the univariate analysis chosen as defining for the inclusion of parameters in the multivariate analysis (lines 179-180)?

We thank the reviewer for this important methodological observation. As stated in our manuscript (footnote of Table 5), variables with a univariable p-value < 0.2 were automatically included in the multivariable model. These included: diabetes duration (p = 0.021, 0.010), age at onset after 10 years (p = 0.015), vitamin D insufficiency (p = 0.019), poor time in range (p = 0.093), balanced diet (p = 0.124), and age group 10–14 years (p = 0.136). For variables with p > 0.2, we made and a priori decision to retain those with established clinical significance in type 1 diabetes research, even if they did not meet the statistical threshold. Poor glycemic control (p = 0.313) falls into this category for the following reasons: Glycemic control is a cornerstone of T1DM management and has been repeatedly associated with growth outcomes, metabolic complications, and long-term prognosis in the literature [1].

• Omitting glycemic control entirely from the multivariable model could introduce confounding by indication—i.e., children with poorer control may differ systematically from those with better control in ways that also affect growth (e.g., adherence, nutritional status, disease severity).
• Preserving all glycemic categories (optimal, suboptimal, poor) allows for a complete assessment of the exposure–response relationship, even if the univariable association is not statistically significant due to limited power or small subgroup sizes.

We acknowledge that this two-step approach, including variables with p < 0.2 plus clinically important variables with p > 0.2 was not clearly articulated in our original manuscript. We have now revised the Statistical Analysis section to explicitly describe this hybrid variable selection strategy.

[Variables with a p value less than 0.2 in univariate analysis were included in the multivariate logistic regression model to identify independent predictors of short stature. Second, variables already recognized as clinically important in T1DM research namely the glycemic control categories (optimal, suboptimal, poor) were retained in the model a priori, irrespective of their univariable p-values, to minimize confounding by indication and maintain clinical interpretability [29]]

Comment 5: In the Results section (lines 205-213) are mentioned criteria for defining different parameters are already included in the Material and method section.

The reviewer is correct. Definitions for short stature, glycemic control, diabetes duration, time in range, vitamin D status, and lifestyle factors were already provided in the Methods section. We have removed the redundant text from the Results section  and revised table footnotes to simply reference the Methods section. This eliminates unnecessary repetition and improves manuscript flow.

Comment 6: In Table 2 for the parameters height, weight and BMI, although IQR is mentioned, the table does not mention IQR as an interval, probably the values ​​represent mean and standard deviation.

The reviewer is correct. Upon re-examination of Table 2, we found an error: the values reported in parentheses were the Q3-Q1. We have corrected Table 2 to accurately report the median and interquartile range (Q1–Q3), as the data were non-normally distributed. The corrected table is shown below. All statistical comparisons (Mann-Whitney U tests) and conclusions remain unchanged.

Domain	Variable / Category	Normal Group (n=231)	T1DM Group (n=231)	Test / Statistic, Effect size, p-value
Height Category	Normal	149 (64.5%)	214 (92.6%)	χ²=78.45, V=0.412, p<0.001
	Tall	78 (33.8%)	3 (1.3%)
	Short stature	4 (1.7%)	14 (6.1%)
Height	Percentile, Median (IQR)	84.5 [42.5 – 97.1]	49.3 [20.9 – 77.2]	W=37364, r=0.35, p<0.001
	(cm), Median (IQR)	151.3 [130.6 – 166.6]	148.0 [133.0 – 156.5]	W=29937, r=0.11, p=0.023
Weight Status	Normal	209 (90.5%)	215 (93.1%)	χ²=10.82, V=0.153, p=0.001
	Underweight	0 (0%)	7 (3.0%)
	Overweight	18 (7.8%)	9 (3.9%)
Weight	Percentile, Median (IQR)	78.6 [50.8– 92.3]	58.2 [34.2– 79.5]	W=34347, r=0.25, p<0.001
	(kg), Median (IQR)	46.4 [31.3– 61.0]	44.0 [29.0– 56.5]	W=28796, r=0.07, p=0.140
BMI Classification	Normal	142 (61.5%)	171 (74.0%)	χ²=9.34, V=0.142, p=0.025
	Thinness	18 (7.8%)	9 (3.9%)
	Overweight	42 (18.2%)	33 (14.3%)
	Obese	29 (12.6%)	18 (7.8%)
BMI	Percentile, Median (IQR)	74.2 [33.0 – 88.8]	62.5 [33.1 – 83.2]	W=29036.5, r=0.08, p=0.101
	(kg/m²), Median (IQR)	20.5 [16.6 – 23.6)	19.0 [17.0 – 22.6]	W=28012, r=0.04, p=0.353

Comment 7: In patients with T1DM, the insulin dose/Kg should be included considering that they are more frequently overweight/obese compared to the control group, and the insulin dose can be associated with the stature status. TBR and TAR should also be included to see if frequent/severe hypoglycemia influences the short stature status.

We thank the reviewer for this suggestion. However, we wish to clarify our anthropometric findings, as they differ from the reviewer's assumption.

Clarification of Weight Status in Our Cohort

As shown in the table below (data from our study), weight percentiles were significantly lower in children with T1DM compared to healthy controls:

Group	Weight Percentile, Median (IQR)	p-value
Normal (control)	78.6 (50.8–92.3)	Reference
T1DM	58.2 (34.2–79.5)	< 0.001

Thus, children with T1DM in our cohort were not heavier than controls. In fact, they had significantly lower weight-for-age percentiles. The higher prevalence of overweight/obesity by BMI classification in the T1DM group (30.8% vs. 22.1% in controls, p = 0.025) reflects the fact that BMI is a ratio of weight to height squared  and since children with T1DM had shorter stature (height percentile 58.2 vs. 78.6 in controls, p < 0.001), a normal or even lower weight can result in a higher BMI classification. Given that our T1DM cohort had lower weight percentiles than controls, the concern that higher insulin doses due to overweight/obesity might confound the association with short stature is less relevant to our specific population. Nevertheless, we agree that insulin dose (units/kg/day) is an important clinical variable. However, these data were not consistently documented in the electronic medical records used for this retrospective study. We have added this to the limitations section.

[Second, while we assessed Time in Range (TIR) and found no significant association with short stature, we did not have access to Time Below Range (TBR) or Time Above Range (TAR). Frequent or severe hypoglycemia (reflected by TBR) could influence linear growth through counter-regulatory hormone responses (e.g., cortisol, growth hormone) or altered nutritional intake, and its absence from our analysis represents a data limitation. ]
Comment 8: The distribution of height, weight, and BMI across all age groups and in the control, group should be included in the analysis and compared with that identified in patients with type 1 diabetes.

We would like to thank you for this suggestion we created a new table describing different anthropometric measures across different age groups of normal versus children with T1DM

Height differences emerged progressively with age. In the youngest age group (5–9 years), there was no significant difference in height between T1DM and control children (p = 0.217), suggesting that growth impairment is not yet established at diagnosis or early in the disease course. However, by the 10–14 years age group, significant height differences appeared (p < 0.001, r = 0.32), and these differences became even more pronounced in the 15–18 years group (p < 0.001, r = 0.64). This pattern was further reflected in height percentiles, where T1DM children in the 15–18 years group had a median height percentile of only 45.9 compared to 97.1 in controls (r = 0.73, a large effect size).  Weight was significantly lower in T1DM children starting from the 5–9 years group (p = 0.012, r = 0.27) and continues into older age groups. Notably, weight percentiles in the T1DM group remain consistently around the 46th–59th percentile across all ages, while control children maintain weight percentiles between 70 and 81, indicating that T1DM children were not only shorter but also lighter than their healthy peers. This finding challenges the assumption that T1DM is associated with higher weight; rather, the higher prevalence of overweight/obesity by BMI classification in T1DM appears to be driven by shorter stature, not excess weight. BMI differences were more complex. In the 5–9 years group, T1DM children actually had slightly higher BMI than controls (p = 0.018, r = -0.25, with the negative effect size indicating lower values in controls). This reverses in older groups, where controls have higher BMI (p = 0.004, r = 0.26 in the 15–18 years group). This pattern suggests that the relationship between T1DM and BMI changes with age and disease duration, possibly reflecting pubertal changes, insulin therapy effects, or lifestyle factors. The effect sizes (r) show a clear gradient with age. For height and height percentile, effect sizes progress from small-to-medium in the youngest group (r = 0.13–0.15) to medium in the 10–14 years group (r = 0.32–0.40) and large in the 15–18 years group (r = 0.64–0.73). This dose-response relationship between age (and thus diabetes duration) and growth impairment is clinically significant and supports the study's central finding that disease duration, not glycemic control, is the key predictor of short stature.

Table 2: Age-Stratified Comparison of Anthropometric Measures Between T1DM and Control Groups

Age Group	Parameter	T1DM Median (Q1–Q3)	Control Median (Q1–Q3)	p-value*	Effect Size (r)†
5–9 years	Height (cm)	119.0 (112.0–125.2)	122.0 (111.7–130.8)	0.217	0.13
	Weight (kg)	20.0 (18.0–24.0)	25.2 (18.8–31.6)	0.012	0.27
	BMI (kg/m²)	18.5 (16.0–23.0)	18.2 (13.4–21.0)	0.018	-0.25
	Height Percentile	55.5 (26.9–82.9)	52.1 (18.4–93.3)	0.946	-0.01
	Weight Percentile	46.1 (37.0–77.8)	70.6 (26.3–95.2)	0.185	0.14
10–14 years	Height (cm)	145.0 (138.0–152.0)	151.7 (144.6–158.7)	<0.001	0.32
	Weight (kg)	42.0 (32.4–50.0)	47.3 (38.7–54.7)	0.004	0.26
	BMI (kg/m²)	19.0 (17.1–22.2)	21.0 (17.0–23.8)	0.080	0.16
	Height Percentile	50.9 (19.2–76.7)	85.1 (37.0–97.0)	<0.001	0.40
	Weight Percentile	59.2 (32.4–82.4)	79.9 (57.3–91.8)	<0.001	0.32
15–18 years	Height (cm)	158.5 (151.0–167.2)	175.5 (164.7–184.8)	<0.001	0.64
	Weight (kg)	57.0 (48.0–65.0)	68.1 (53.9–79.0)	<0.001	0.37
	BMI (kg/m²)	19.0 (17.0–24.0)	22.4 (19.0–25.6)	0.004	0.26
	Height Percentile	45.9 (14.4–72.0)	97.1 (72.8–97.2)	<0.001	0.73
	Weight Percentile	59.2 (31.6–78.1)	80.8 (59.5–92.4)	<0.001	0.38

*† p-values from Mann-Whitney U test; Effect size r = rank-biserial correlation (interpretation: 0.1 = small, 0.3 = medium, 0.5 = large). Bold values indicate statistical significance (p < 0.05).*
The prevalence, general and by age group, of short stature in the control group should be mentioned.
Comment 9:  In Table 4, the percentages for the distribution of normal and short stature respectively by parameter categories are not reported correctly. Thus, for example, for the duration of diabetes:
- 6.67%, 80% and 13.33% of those with short stature had newly diagnosed diabetes, intermediate duration and long duration respectively. 30.62%, 63.63% and 5.74% of those with normal stature had newly diagnosed diabetes, intermediate duration and long duration respectively.

We thank the reviewer for this careful observation. The reviewer is correct that the percentages in the original Table 4 were not clearly presented. Specifically, the table should display row percentages (the proportion of normal vs. short stature within each category of a given parameter) rather than column percentages or raw frequencies alone.

We have revised Table 4 to present:

1. Frequencies (n) within each category
2. Row percentages (percentage of normal/short stature within that parameter category)
3. Appropriate statistical test results (chi-square or Fisher's exact)

The corrected table is shown below.

Among the 244 participants, 30 (12.4%) had short stature. Age group was not significantly associated with stature status (p = 0.126), although the highest proportion of short stature was observed in the 10–14 years group (17.5%) compared to the 5–9 years (8.2%) and 15–18 years (8.7%) groups. Age at onset of T1DM showed a statistically significant association with stature status (p = 0.036), with the highest proportion of short stature observed among children diagnosed between 5 and 10 years of age (17.5%), compared to those diagnosed before 5 years (12.5%) and after 10 years (6.5%). Diabetes duration was also significantly associated with stature status (p = 0.004), with the proportion of short stature increasing progressively from new-onset diabetes (3.0%) to intermediate duration (14.9%) and long-standing diabetes (25.0%). No significant association was observed between glycemic control and stature status (p = 0.534), although short stature was more frequent in the suboptimal (17.0%) and poor control groups (13.9%) compared to the optimal group (6.9%). HbA1c levels were similar between groups (p = 0.173). Vitamin D status showed a significant association with stature status (p = 0.042), with a higher proportion of short stature among participants with vitamin D insufficiency (19.3%), compared to deficiency (11.0%) and sufficiency (6.8%). However, median vitamin D levels were comparable between groups (p = 0.447). Balanced diet (p = 0.144), time in range (p = 0.19), treatment adherence (p = 0.358), gender (p = 0.705), and family history of diabetes (p = 1.0) were not significantly associated with stature status. Table 4

 

Table 4: Comparison of sociodemographic and clinical characteristics by stature status (normal vs short stature) (N = 244)

Variable	Category	Normal Stature (n = 214)	Short Stature   (n=30)	Test statistics, Effect size, p-value
Age Group (years)	5–9	45 (91.8%)	4 (8.2%)	χ²=4.44, V=0.135, p=0.126
	10–14	85 (82.5%)	18 (17.5%)
	15–18	84 (91.3%)	8 (8.7%)
Gender	Male	126 (88.7%)	16 (11.3%)	χ²=0.14, φ=0.024, p=0.705
	Female	88 (86.3%)	14 (13.7%)
Diabetes Duration	New onset	65 (97.0%)	2 (3.0%)	χ²=8.8, V=0.19, p = 0.004
	Intermediate	137 (85.1%)	24 (14.9%)
	Long-standing	12 (75.0%)	4 (25.0%)
Glycemic Control	Optimal	27 (93.1%)	2 (6.9%)	χ²=1.58, V=0.085, p = 0.534
	Suboptimal	39 (83.0%)	8 (17.0%)
	Poor	124 (86.1%)	20 (13.9%)
Age at Onset of T1DM	Before 5 years	14 (87.5%)	2 (12.5%)	χ²=6.4, V=0.162, p = 0.036
	5–10 years	99 (82.5%)	21 (17.5%)
	After 10 years	101 (93.5%)	7 (6.5%)
HbA1c (%)	Median [Q1-Q3]	9.6 [7.5-11.5]	9.6 [8.4-12.4]	W=2716.5, r=0.09, p=0.173
Serum vitamin D concentration	Deficient	65 (89.0%)	8 (11.0%)	χ²=6.32, V=0.161, p=0.042
	Insufficient	67 (80.7%)	16 (19.3%)
	Sufficient	82 (93.2%)	6 (6.8%)
	(ng/mL) Median [Q1-Q3]	26 [19-37]	25.5 [19.2-28]	W=3485.5, r=0.05, p=0.447
Balanced Diet	Yes	183 (86.3%)	29 (13.7%)	χ²=1.98, φ=0.09, p=0.144
	No	31 (96.9%)	1 (3.1%)
Time in Range (TIR)	Poor	121 (84.6%)	22 (15.4%)	χ²=3.32, V=0.117, p=0.19
	Fair	36 (90.0%)	4 (10.0%)
	Good	57 (93.4%)	4 (6.6%)
Treatment Adherence	Yes	176 (88.9%)	22 (11.1%)	χ²=0.85, φ=0.059, p=0.358
	No	38 (82.6%)	8 (17.4%)
Family History of Diabetes	Yes	46 (86.8%)	7 (13.2%)	χ²=0, φ=0, p=1
	No	168 (88.0%)	23 (12.0%)

 

The Discussion section addresses many theoretical data that are not included in the current study and thus, the results of the current study cannot be reported to these data.

We thank the reviewer for this critical and constructive feedback. The reviewer is correct that portions of our Discussion section included theoretical or speculative content that extend beyond the actual findings of our study. Specifically, we discussed mechanistic pathways, compared our results to studies with different designs or populations, and made recommendations that are not directly supported by our retrospective cross-sectional data.

Reference 14 is identical to reference 13.

The reviewer is correct. References 13 and 14 were identical. This was an inadvertent duplication during manuscript preparation. We have removed the duplicate reference (formerly 14) and renumbered the entire reference list. All in-text citations have been corrected accordingly. We apologize for this error.

Considering the previous comments, the study design is not suitable for a proper analysis of risk factors for short stature in patients with type 1 diabetes, the main limitations of the study being the incomplete analysis of potential risk factors known to influence height status, the incomplete comparison between the 2 groups included, the cross-sectional and retrospective design that does not allow the identification of a causal relationship, and the small sample size for the category of subjects with short stature.

We thank the reviewer for this fair and comprehensive summary of our study's limitations. We fully agree with each point raised. We have revised the manuscript to replace all causal and "risk factor" terminology with statements of association. We acknowledge the incomplete analysis of potential confounders (insulin dose, parental height, pubertal staging, socioeconomic status, TBR, TAR) and have added these to the Limitations section. Regarding the control group, we clarify that it was included solely as

a height reference; comparing risk factors between groups was never an objective. The small sample size of the short stature group (n = 30) resulted in wide confidence intervals and limited statistical power, which we now explicitly note. In light of these constraints, we have revised our conclusions to be hypothesis-generating rather than confirmatory, and we explicitly state that prospective longitudinal studies are needed to validate our findings. We are grateful for the reviewer's meticulous and constructive feedback, which has significantly improved the scientific integrity of our manuscript.

 

 

 

1. Kim, S., et al., Long-Term Growth Outcomes of Children With Type 1 Diabetes According to Glycemic Control and Use of Continuous Glucose Monitoring: A Retrospective Cohort Study. Pediatr Diabetes, 2026. 2026: p. 9111583.

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript Longer Diabetes Duration, Not Glycemic Control, Predicts  Short Stature in Saudi Children and Adolescents with Type 1 Diabetes Mellitus: A Retrospective Cross-Sectional Study represents a retrospective cross-sectional study involving 250 children and adolescents with T1D (ages 5 - 18) and 267 healthy controls. Height was assessed using Saudi centile charts, and short stature was defined as <3rd centile. Multivariate analysis revealed that longer duration of T1D (2 - 5 years and >5 years) is an independent predictor of short stature, whereas HbA1c and vitamin D were not significant after adjustments. In terms of the current state of knowledge, the study is intriguing and addresses a relevant topic in the field of pediatric diabetology. To meet the highest standards, the publication requires further clarification.

The manuscript provides clear definitions of the discussed issues. Short stature is defined as <3rd percentile and is based on local growth charts. This is particularly important in populations with different anthropometric patterns, a factor that is often overlooked.
The inclusion of a healthy control group shows that the distribution of height in T1D differs from the population without the disease. This allows for a more accurate presentation of the issue, and the inclusion of a control group contributes to a more accurate analysis and evaluation. The study includes both metabolic parameters (HbA1c) and selected lifestyle factors (diet, physical activity, adherence), as well as vitamin D. Reference to the TIR index is also important, based on the use of the most recent technologies and assessment of metabolic control.
One aspect that warrants further discussion is the fact that a child’s growth process is dynamic, and it is not always possible to demonstrate this during an examination without conducting appropriate tests. To support a diagnosis of short stature, the issue must be approached more broadly. This requires an assessment of growth velocity. An additional factor should be a retrospective assessment - whether the child is short due to a genetic factor (parents and short stature also prior to diagnosis) or whether the child stopped growing properly after the diagnosis of T1D. Furthermore, the lack of an assessment of the impact of puberty and the pubertal growth spurt are also factors influencing the evaluation of the growth process. The conclusion that “duration of T1D predicts short stature” may be true, but without a growth trajectory, it is prone to interpretive errors (e.g., uneven distribution of puberty stages across disease duration groups). Thus, the factor that should be considered is the factor related to the height of the mother and father. Based on this, it is possible to calculate mid-parental height (target height) and assess the deviation from genetic potential.

In studies of children’s growth, it is impossible to correctly interpret growth without assessing the stage of puberty (Tanner). Here, the group in Stage III (the period of rapid growth) may prove to be the most reliable group that can realistically reflect the growth process. Bone age assessment can also be a useful indicator. Most studies show that at the time of diagnosis of type 1 diabetes in pediatric patients, the bone age of most children typically corresponds to their chronological age. Additionally, with the duration of diabetes (and poor metabolic control), it is observed that children with T1DM may have lower bone mineral density and altered bone microarchitecture. Therefore, the assessment of such parameters is a key factor in the analysis of the growth process. The manuscript does not include any control for this factor in the data. This should be noted as a study limitation.

The discussion mentions factors influencing growth rate, but there is no information on how comorbidities were confirmed or ruled out in the study participants. The authors mention autoimmune conditions (thyroid disease, celiac disease, etc.) as potential causes of growth problems, but do not indicate whether or how they investigated them or how frequently they occurred in the analyzed group. Factors related to malabsorption and associated deficiencies, such as those seen in celiac disease have a significant impact on growth velocity.

The authors themselves note that HbA1c reflects the “past 2 - 3 months” and that any hight increase requires a long-term perspective. A factor that better supports a proper analysis is, for example, the average HbA1c level over the past 12 - 24 months. This provides a much broader picture of the situation. In modern diabetes monitoring (especially when using CGM systems), TIR and CV parameters are key to assessing glycemic control, complementing traditional glycated hemoglobin (HbA1c) measurements. It would be worthwhile to include the CV index in the assessment - if not in this study, then as a suggestion for future research. In children with type 1 diabetes, achieving a stable CV is often more difficult due to fluctuating appetite, physical activity, and growth hormone surges. Thus, evaluating values below and above target ranges is also important. Assessing these parameters significantly impacts a child’s development, as severe hypoglycemia and chronic hyperglycemia affect health and eating behaviors/activity, and indirectly influence growth,

The manuscript lacks information on the treatment methods used. A more detailed assessment of the type of therapy (insulin pump vs. multiple daily injections (MDI) is a factor influencing disease control. Additionally, insulin requirements and dosing may also be factors affecting the patient’s level of glycemic control.
The assessment of lifestyle and diet may be subject to the risk of measurement error. Responses based solely on yes or no answers can lead to bias and do not provide adequate sensitivity. These are very important factors influencing blood glucose levels and the degree of glycemic control. In further studies, it is also worth examining dietary factors and the adequate intake of energy and protein, which significantly influence growth processes. It is also worth considering the socioeconomic status and educational background of parents, which can indirectly influence dietary choices as well.

In the discussion and conclusions, it is worth highlighting the factors that influence the rate of growth. A distinction should be made between factors related to the effects of chronic disease, puberty, and comorbid conditions (celiac disease, hypothyroidism). Additionally, caution should be taken when interpreting the statement “HbA1 were not significantly associated with stature status ” because the measurement is cross-sectional, and the distribution of HbA1c shows little variability (most patients have poor control).

Author Response

Reviewer # 3

We sincerely thank the reviewer for the comprehensive and highly insightful evaluation of our manuscript, as well as for the constructive suggestions that significantly strengthen the scientific rigor and clinical interpretation of our findings. We fully acknowledge and appreciate the reviewer’s emphasis on the importance of longitudinal growth assessment, pubertal staging, genetic potential (including mid-parental height), bone age evaluation, comorbid autoimmune conditions, and more robust glycemic and lifestyle measures. These points collectively highlight key methodological aspects that are essential for a complete understanding of growth patterns in children and adolescents with type 1 diabetes mellitus. We agree that growth is a dynamic process influenced by multiple interrelated factors, including pubertal development, genetic target height, cumulative metabolic control, and associated autoimmune or nutritional conditions. As correctly noted, the absence of longitudinal data, Tanner staging, parental height, and bone health parameters in our study limits our ability to fully characterize growth trajectories and differentiate between underlying etiologies of short stature. Importantly, we recognize that several of these variables—such as growth velocity, bone age, mid-parental height, and comprehensive CGM-derived glycemic variability indices—would substantially enhance the precision of future analyses. These factors were not available in our retrospective dataset, and we now explicitly acknowledge them as key limitations of the current study. We are grateful for the reviewer’s suggestion that these parameters be incorporated into future research. We fully agree that prospective, longitudinal studies including pubertal assessment, genetic target height estimation, bone age evaluation, and detailed metabolic profiling (including extended HbA1c trends and CGM variability metrics such as CV, TBR, and TAR) are essential to better delineate the determinants of impaired growth in T1DM. We have revised the manuscript accordingly to incorporate these limitations and to ensure that our conclusions are appropriately framed as associative rather than causal.

Comment 1: The manuscript Longer Diabetes Duration, Not Glycemic Control, Predicts Short Stature in Saudi Children and Adolescents with Type 1 Diabetes Mellitus: A Retrospective Cross-Sectional Study represents a retrospective cross-sectional study involving 250 children and adolescents with T1D (ages 5 - 18) and 267 healthy controls. Height was assessed using Saudi centile charts, and short stature was defined as <3rd centile. Multivariate analysis revealed that longer duration of T1D (2 - 5 years and >5 years) is an independent predictor of short stature, whereas HbA1c and vitamin D were not significant after adjustments. In terms of the current state of knowledge, the study is intriguing and addresses a relevant topic in the field of pediatric diabetology. To meet the highest standards, the publication requires further clarification.

• The manuscript provides clear definitions of the discussed issues. Short stature is defined as <3rd percentile and is based on local growth charts. This is particularly important in populations with different anthropometric patterns, a factor that is often overlooked. The inclusion of a healthy control group shows that the distribution of height in T1D differs from the population without the disease. This allows for a more accurate presentation of the issue, and the inclusion of a control group contributes to a more accurate analysis and evaluation. The study includes both metabolic parameters (HbA1c) and selected lifestyle factors (diet, physical activity, adherence), as well as vitamin D. Reference to the TIR index is also important, based on the use of the most recent technologies and assessment of metabolic control.

 

We sincerely thank Reviewer 3 for these positive and encouraging comments. We are gratified that the reviewer recognizes the methodological strengths of our manuscript.

We appreciate the reviewer's acknowledgment of the following key strengths:

• Use of local growth charts: The reviewer correctly notes that defining short stature as <3rd percentile based on validated Saudi growth charts is particularly important given that different populations have distinct anthropometric patterns. This is often overlooked in studies that apply international references to populations for which they were not developed.
• Inclusion of a healthy control group: We agree that the control group allows us to demonstrate that the height distribution of children with T1DM differs meaningfully from the general population, and we appreciate the reviewer's recognition that this contributes to a more accurate analysis and evaluation.
• Comprehensive assessment of relevant variables: We thank the reviewer for highlighting the inclusion of metabolic parameters (HbA1c), lifestyle factors (diet, physical activity, adherence), and vitamin D status, all of which are clinically relevant to growth outcomes in T1DM.
• Reference to Time in Range (TIR): The reviewer correctly notes that using TIR reflects the application of recent continuous glucose monitoring technologies and provides a more modern assessment of metabolic control beyond traditional HbA1c measurement.

We are encouraged that despite the methodological limitations of the retrospective cross-sectional design (which we have thoroughly addressed in our responses to the reviewer's other comments), Reviewer 3 recognizes these important strengths of our work. We believe that these strengths, combined with the transparent acknowledgment of limitations, make our manuscript a valuable contribution to the literature on growth outcomes in children with T1DM, particularly in the understudied Saudi population.

Comment 2: One aspect that warrants further discussion is the fact that a child’s growth process is dynamic, and it is not always possible to demonstrate this during an examination without conducting appropriate tests. To support a diagnosis of short stature, the issue must be approached more broadly. This requires an assessment of growth velocity. An additional factor should be a retrospective assessment whether the child is short due to a genetic factor (parents and short stature also prior to diagnosis) or whether the child stopped growing properly after the diagnosis of T1D. Furthermore, the lack of an assessment of the impact of puberty and the pubertal growth spurt are also factors influencing the evaluation of the growth process.

One aspect that warrants further discussion is that growth in children is a dynamic process that cannot be fully assessed through a single clinical examination without appropriate longitudinal evaluation. Accordingly, a diagnosis of short stature requires a more comprehensive approach, including assessment of growth velocity over time rather than relying solely on cross-sectional height measurements. In addition, retrospective evaluation is important to distinguish between different aetiologies of short stature. In particular, it is necessary to consider whether reduced height is attributable to genetic factors, such as familial short stature, or whether impaired growth developed after the diagnosis of type 1 diabetes mellitus (T1DM). While we have excluded children with known genetic disorders associated with short stature, the absence of parental height data and pre-diagnostic growth records limits our ability to differentiate between pre-existing and disease-related growth impairment. Furthermore, the assessment of pubertal status and the pubertal growth spurt is essential in the evaluation of growth patterns. Puberty is a key determinant of final adult height, and its timing and progression may be altered in chronic conditions such as T1DM. The lack of pubertal staging (e.g., Tanner staging) in our study represents an important limitation, as it restricts interpretation of height percentiles, particularly in adolescent subgroups. We acknowledge the reviewer’s important point that growth assessment should ideally incorporate longitudinal data, genetic background, and pubertal status. Our study, however, provides only a single time-point measurement of height and therefore offers a limited snapshot of growth status, without the ability to evaluate growth trajectory or velocity.

We added the following limitation However, several limitations must be acknowledged. The retrospective cross-sectional design precludes causal inference and examination of growth trajectories over time. Our cross-sectional measurement of height at a single time precluded assessment of growth velocity, and we lacked data on parental height (to determine familial short stature), pre-diagnosis growth records, pubertal staging (Tanner), bone age, and bone mineral density all essential for comprehensive growth evaluation. We did not systematically screen for autoimmune comorbidities (celiac disease, hypothyroidism) or assess markers of malabsorption and nutritional deficiencies. Regarding glycemic control, a single HbA1c measurement reflects only 2-3 months and showed limited variability; we lacked average HbA1c over 12-24 months, as well as continues glucose monitoring-derived metrics including coefficient of variation (CV), time below range (TBR), and time above range (TAR). Lifestyle factors (diet, physical activity, adherence) were assessed using simple yes/no responses from medical records, introducing recall and social desirability bias. No quantitative dietary intake (energy, protein) was assessed, and socioeconomic status and parental education were not collected. All participants used insulin pumps, precluding comparison with multiple daily injections therapy. Insulin dose (units/kg/day) was not available. Finally, small subgroup sizes (e.g., long-standing diabetes, n=8; short stature, n=30) resulted in wide confidence intervals and limited statistical power. Accordingly, our findings should be viewed as hypothesis-generating rather than confirmatory.

]

Comment 3: The conclusion that “duration of T1D predicts short stature” may be true, but without a growth trajectory, it is prone to interpretive errors (e.g., uneven distribution of puberty stages across disease duration groups). Thus, the factor that should be considered is the factor related to the height of the mother and father. Based on this, it is possible to calculate mid-parental height (target height) and assess the deviation from genetic potential.

First, the reviewer correctly notes that without growth trajectory data, our conclusion that "duration of T1D predicts short stature" is prone to interpretive errors. Specifically, uneven distribution of pubertal stages across diabetes duration groups could confound the observed association, as children at different pubertal stages have different growth velocities and height percentiles. We have addressed this by removing all causal and "predictor" language from our conclusions and replacing it with statements of association (e.g., "diabetes duration is associated with short stature").

Second, the reviewer correctly emphasizes the importance of parental height and mid-parental target height. Without these data, we cannot determine whether a child's short stature reflects the effect of T1DM or simply familial (genetic) short stature. This is a fundamental limitation of our study, as a child who is short due to genetic factors would be short regardless of diabetes duration. We have added this as a major limitation and have revised our conclusions according.

 

Comment 4: In studies of children’s growth, it is impossible to correctly interpret growth without assessing the stage of puberty (Tanner). Here, the group in Stage III (the period of rapid growth) may prove to be the most reliable group that can realistically reflect the growth process. Bone age assessment can also be a useful indicator. Most studies show that at the time of diagnosis of type 1 diabetes in pediatric patients, the bone age of most children typically corresponds to their chronological age. Additionally, with the duration of diabetes (and poor metabolic control), it is observed that children with T1DM may have lower bone mineral density and altered bone microarchitecture. Therefore, the assessment of such parameters is a key factor in the analysis of the growth process. The manuscript does not include any control for this factor in the data. This should be noted as a study limitation.

The reviewer correctly notes that it is impossible to correctly interpret growth without assessing pubertal stage. Children in different Tanner stages have vastly different growth velocities, and the same height percentile may have completely different implications for a prepubertal child versus a child in the midst of the pubertal growth spurt (Tanner Stage III). We agree that the lack of Tanner staging is a major limitation of our study. The reviewer correctly notes that bone age assessment is a useful indicator for evaluating growth potential. While most children have bone age corresponding to chronological age at T1DM diagnosis, prolonged disease duration and poor metabolic control may lead to delayed bone age and reduced final height. Without bone age data, we cannot determine whether the shorter stature observed in our T1DM cohort reflects delayed skeletal maturation (with potential for catch-up growth) or irreversible growth impairment. The reviewer raises an important point about bone health. Children with long-standing T1DM and poor metabolic control may have lower bone mineral density and altered bone microarchitecture, which could contribute to altered growth patterns. Our study did not assess any bone health parameters, and we acknowledge this as a limitation.

Comment 5: The discussion mentions factors influencing growth rate, but there is no information on how comorbidities were confirmed or ruled out in the study participants. The authors mention autoimmune conditions (thyroid disease, celiac disease, etc.) as potential causes of growth problems, but do not indicate whether or how they investigated them or how frequently they occurred in the analyzed group. Factors related to malabsorption and associated deficiencies, such as those seen in celiac disease have a significant impact on growth velocity.

We thank the reviewer for this important observation. The reviewer is correct that we mentioned autoimmune conditions (thyroid disease, celiac disease, etc.) as potential causes of growth problems but did not indicate how they were investigated or how frequently they occurred in our cohort. We acknowledge that we did not systematically screen for autoimmune thyroid disease, celiac disease, or markers of malabsorption and associated nutritional deficiencies (e.g., iron, vitamin D, calcium), all of which can independently affect linear growth in children with T1DM. Consequently, we cannot exclude the possibility that undiagnosed or untreated autoimmune comorbidities contributed to the short stature observed in some of our participants. We have revised the Discussion to explicitly state that these conditions were not assessed, added a sentence to the Limitations section noting this absence, and recommended that future studies include systematic screening for thyroid autoimmunity, celiac disease, and nutritional absorption markers. We apologize for this omission and thank the reviewer for highlighting it.

Comment 5: The authors themselves note that HbA1c reflects the “past 2 - 3 months” and that any hight increase requires a long-term perspective. A factor that better supports a proper analysis is, for example, the average HbA1c level over the past 12 - 24 months. This provides a much broader picture of the situation. In modern diabetes monitoring (especially when using CGM systems), TIR and CV parameters are key to assessing glycemic control, complementing traditional glycated hemoglobin (HbA1c) measurements. It would be worthwhile to include the CV index in the assessment - if not in this study, then as a suggestion for future research. In children with type 1 diabetes, achieving a stable CV is often more difficult due to fluctuating appetite, physical activity, and growth hormone surges. Thus, evaluating values below and above target ranges is also important. Assessing these parameters significantly impacts a child’s development, as severe hypoglycemia and chronic hyperglycemia affect health and eating behaviors/activity, and indirectly influence growth.

We thank the reviewer for this highly insightful and technically important comment. The reviewer correctly identifies several limitations of our glycemic control assessment and provides valuable suggestions for future research.

The reviewer is correct that a single HbA1c measurement reflects only the preceding 2-3 months, which is insufficient to capture cumulative glycemic exposure over the years during which growth impairment develops. We agree that average HbA1c over 12-24 months would provide a much broader and more accurate picture. The reviewer correctly notes that in modern diabetes monitoring, particularly with CGM systems, TIR and coefficient of variation (CV) are key parameters for assessing glycemic control, complementing HbA1c. CV reflects glycemic variability, which in children with T1DM is often difficult to control due to fluctuating appetite, physical activity, and growth hormone surges. High glycemic variability (even with acceptable HbA1c) may independently affect growth, health, eating behaviors, and physical activity. We added in the recommendation. The reviewer emphasizes that evaluating values below (hypoglycemia) and above (hyperglycemia) target ranges is important, as both severe hypoglycemia and chronic hyperglycemia affect health, eating behaviors, activity, and

indirectly influence growth. While we assessed TIR, we did not have access to TBR or TAR, as previously noted. We added the following paragraph to discussion. [ To begin with, HbA1c reflects average blood glucose over only the previous 2–3 months, whereas linear growth is a long-term, cumulative process shaped by metabolic conditions over years rather than weeks. [6] As diabetes duration emerged as the strongest predictor of short stature, it likely serves as a proxy for cumulative glycemic exposure (collinearity between the two variables), which a single cross-sectional HbA1c measurement cannot capture. Second, the GH/IGF-1 axis disturbances characteristic of T1DM including increased GH secretion and decreased IGF-1 levels may persist even in the setting of acceptable HbA1c values, as subcutaneous insulin therapy cannot fully replicate physiological portal insulin delivery to the liver [31]. IGF-1 levels correlate more strongly with diabetes duration and cumulative metabolic control than with isolated HbA1c measurements [38] . Third, the study population exhibited uniformly poor glycemic control overall (median HbA1c 9.6%, with 58.8% classified as poorly controlled), resulting in limited variability that may have reduced the ability to detect a significant association. Additionally, the cross-sectional design captures HbA1c at a single time point, which may not reflect the longitudinal glycemic trajectory that influences growth over time. Finally, as noted by Santi et al. [39], with contemporary diabetes management and intensive insulin regimens, many children with T1DM achieve normal growth regardless of moderate variations in HbA1c, suggesting that factors such as disease duration, timing of onset, and nutritional status may exert more direct effects on linear growth than short-term glycemic metrics alone.]

Comment 6: The manuscript lacks information on the treatment methods used. A more detailed assessment of the type of therapy (insulin pump vs. multiple daily injections (MDI) is a factor influencing disease control. Additionally, insulin requirements and dosing may also be factors affecting the patient’s level of glycemic control.

We thank the reviewer for this important observation. The reviewer is correct that our manuscript did not provide information on the treatment methods used in our cohort. Specifically, we did not specify whether participants were using insulin pumps or multiple daily injections (MDI), which is a factor that can influence glycemic control and potentially growth outcomes. Clarification: In our cohort, all participants were using insulin pumps. We have added this information to the Methods section.

Comment 7: The assessment of lifestyle and diet may be subject to the risk of measurement error. Responses based solely on yes or no answers can lead to bias and do not provide adequate sensitivity. These are very important factors influencing blood glucose levels and the degree of glycemic control. In further studies, it is also worth examining dietary factors and the adequate intake of energy and protein, which significantly influence growth processes. It is also worth considering the socioeconomic status and educational background of parents, which can indirectly influence dietary choices as well.

We thank the reviewer for this important methodological observation. The reviewer is correct that our assessment of lifestyle and dietary factors using simple yes/no responses is subject to measurement error, recall bias, and social desirability bias, and lacks the sensitivity to capture the complexity of these behaviors. We acknowledge that we did not assess quantitative dietary intake, including total energy and protein intake, both of which significantly influence growth processes in children. Furthermore, we did not collect data on socioeconomic status or parental educational background, factors that can

indirectly influence dietary choices, treatment adherence, glycemic control, and ultimately growth outcomes.

Comment 8: In the discussion and conclusions, it is worth highlighting the factors that influence the rate of growth. A distinction should be made between factors related to the effects of chronic disease, puberty, and comorbid conditions (celiac disease, hypothyroidism). Additionally, caution should be taken when interpreting the statement “HbA1 were not significantly associated with stature status ” because the measurement is cross-sectional, and the distribution of HbA1c shows little variability (most patients have poor control).

We thank the reviewer for these important suggestions. We have revised the Discussion to distinguish between factors influencing growth rate, specifically differentiating between the effects of chronic disease (metabolic disturbances, altered GH/IGF-1 axis), puberty (timing and tempo of the pubertal growth spurt), and comorbid conditions (celiac disease, autoimmune hypothyroidism), while noting that our study did not systematically screen for these comorbidities. Additionally, we have added caution when interpreting the finding that HbA1c was not significantly associated with short stature, acknowledging that the cross-sectional measurement reflects only the preceding 2-3 months, and the limited variability in HbA1c distribution in our cohort (median 9.6%, with 59.8% classified as poor control) reduces statistical power and may obscure a true association. These caveats have been incorporated into the Discussion and Conclusions sections.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

In general, the authors responded to most of the comments submitted.
Considering the cross-sectional design of the study, the lack of data on TBR, TAR, markers of glycemic variability and total daily insulin dose/Kg, and in order to avoid underestimating the involvement of glycemic control in stature status in patients with type 1 diabetes, I believe that the title of the manuscript is more appropriate as follows: Growth patterns and factors associated with short stature in Saudi children and adolescents with Type 1 Diabetes Mellitus: A Retrospective Cross Sectional Study.
Although they were clearly and extensively mentioned in the manuscript, the limitations of the study remain and influence the correct interpretation of the results and clinical applicability.

Author Response

Thank you very much for your careful review of the revised manuscript and for acknowledging the authors’ responses to the previous comments.

We sincerely appreciate your constructive suggestion regarding the manuscript title. We agree that, given the cross-sectional design and the absence of data on TBR, TAR, glycemic variability markers, and total daily insulin dose/kg, the revised title more accurately reflects the scope and limitations of the study. We will therefore adopt the proposed title:

“Growth patterns and factors associated with short stature in Saudi children and adolescents with Type 1 Diabetes Mellitus: A Retrospective Cross Sectional Study.”

We also acknowledge your important point regarding the study limitations. While they have been clearly stated and expanded in the manuscript, we agree that they remain essential for the correct interpretation of the findings and should be carefully considered when assessing clinical applicability. We will ensure that this is appropriately emphasized in the revised version to avoid any potential overinterpretation of the results.

Once again, we thank you for your valuable and insightful feedback, which has helped improve the clarity and scientific rigor of the manuscript.

Reviewer 3 Report

Comments and Suggestions for Authors

I would like to thank you for submitting the revised version of the manuscript. The current, revised version includes significant modifications that provide a more accurate description of the issues examined. It addresses key topics that were only briefly touched upon in the original version. Additionally, it is worth noting that the “Limitations” section has been expanded. This is particularly important for the purpose of conducting further research on the subject. While this is not necessarily a disadvantage of the current study, it highlights the essence of the areas under investigation and what may remain important in future research. 

It is important to refer to the vitamin D and glycated hemoglobin in greater detail, which has been added to the current version of the manuscript.  

The current version of the manuscript is valuable. However, it is necessary to review the document again to avoid repetition, ensure the use of appropriate academic terminology, and maintain consistency throughout the paper with regard to the sections discussed. 

Author Response

Thank you for your thoughtful review and for acknowledging the improvements made in the revised version of the manuscript.

We sincerely appreciate your positive evaluation regarding the expanded content, particularly the more comprehensive discussion of key issues and the strengthened “Limitations” section. We agree that these additions help to better contextualize the findings and highlight important directions for future research.

We also appreciate your recognition of the improved discussion on the relationship between vitamin D and glycated hemoglobin, which we agree is an important aspect of the manuscript.

In response to your remaining comments, we will carefully re-review the entire manuscript to:

• Remove any instances of repetition,
• Refine the use of academic terminology where needed,
• Ensure consistency across all sections, particularly in the presentation and discussion of key concepts.

We thank you again for your constructive feedback, which has been invaluable in improving the quality of the manuscript.