Siblings’ Risk of Adenoid Hypertrophy: A Cohort Study in Children

Round 1

Reviewer 1 Report

The most important caveat to this manuscript are that they included almost no genetic, immunologic and environmental variables in their study. This affected especially their logistic regression analyses. The variables chosen by linear regression convey very closely related information - multicollinearity. Recurrent upper respiratory tract infections, rhinitis, snoring, adenoid mucus coverage, and non-A tympanogram are related information. Consequently, minor random variation of these variables can unpredictably influence how much of the association to the independent variable - adenoid hypertrophy is attributed to one variable vs another variable in the model. The introduction of environmental variables, such as breastfeeding, environmental tobacco smoking, would be an important added value. Immunological and microbiological variables will also be welcomed.

If the authors want to restrict their study to the variables that were included, which may have important degrees of multicollinearity, I would ask for the authors to consider a control group consisting of pairs of siblings with no symptoms suggestive of chronic adenoid hypertrophy. 

Author Response

We are very much thankful to the Reviewer for his valuable and constructive comments which we are sure would better the manuscript manifold. In the revised version, we have tried our best to answer the queries raised during the review process and incorporate all the changes and suggestions made by the reviewers.

The basic conclusion of the work is that the size of the adenoid in siblings reaching the same age is similar. This correlation justifies the goal of further searching for genetic, infectious or environmental factors influencing the development of AH. However, the work compared the size of the adenoid in siblings who were raised in the same household. Therefore, in most cases, the pair of children was raised with exposure to the same environmental factors. To demonstrate the influence of such factors on the size of the adenoid, siblings should be separated. Theoretically, such a study, although unrealistic, could be based on the basic conclusion of the current study.

The presence of exposure to cigarette smoke has been demonstrated by Evcimik (2015) as a risk factor for the development of AH [1]. However, at present, numerous preventive campaigns have been carried out in our country, and parents are fully aware of the risks for the children associated with exposure to tobacco smoke. Therefore, the rule is not to smoke in rooms where children reside, and those parents who violate this rule usually will not admit it. We have considered to include this factor to our previous performed studies but all parents did not admit to smoke cigarette in the room where the children are. Additionally, I ‘ve (Aleksander Zwierz) asked in the last week 20 parents of new admitted preschool children to the ENT outpatient clinic if someone in their family smoke in the room where reside children and the answer was “no”. That is why we did not analyze this factor and we hope that it won’t be important in the future.  Breastfeeding was indicated as a risk factor for snoring and OSAS [2,3]. Xu, who is the authors of this study concluded that further investigations are warranted to confirm the relationship between breastfeeding and OSAS and the mechanisms behind this relationship [2]. Presumably, the adenoid hypertrophy may play a role in this mechanism. But in the best of our knowledge there is no study which sowed relationship between breastfeeding and AH.  Demonstrating this relationship in large population research would certainly be interesting.  Only then would it be possible to assess the effect of the length of time of child feeding on the size of the adenoid in siblings. To meet expectations of the reviewer we included to the analysis of children residence in an urban or rural area. An environment variable for where you live (city vs village) has been added. This variable is included in Table 2. Our data show no differences in adenoid size between urban and rural patients.

Variables for logistic regression analyzes were not selected based on linear regression results but based on univariate analysis results.Relevant information can be found in the section Statistical analysis "Variables significantly related to adenoid size in a univariate analysis were included in the linear and logistic regression analyses." We know the problem raised by the reviewer - one of the most important conditions in the logistic regression model is the lack of multicollinearity of the independent variables. Multicollinearity of the independent variables leads to unreliable and unstable estimates. One of the assumptions of logistic regression is the lack of strong collinearity of the variables. The proof of the existence of multicollinearity of the independent variables in the logistic regression model is their correlation. In order to quickly, visually assess whether such collinearity exists for our regression analysis, the correlation matrix presented in the form of a table (Table Rev.) was determined using SPSS, with the search for such pairs of variables for which the absolute value of the correlation coefficient would be in the range from 0.8 to 1. The results presented in the Table Rev. indicate that the explanatory variables used for logistic regression do not present strong collinearity of the variables.

 

Table Rev.

 

	rURTI*	Rhinitis	Snoring	Adenoid mucus coverage**	Tympanogram
rURTI*	1,000
Rhinitis	-,393	1,000
Snoring	-0,138	0,040	1,000
Adenoid mucus coverage	-0,188	-0,168	0,216	1,000
Tympanogram	-0,032	-0,021	0,195	0,10	1,000

* rURTI - recurrent upper respiratory tract infections.** Adenoid mucus coverage in MASNA scale

 

We have considered to introduce a control group but the study is based on an invasive diagnostic method of flexible endoscopy, performing the test in healthy children would be inappropriate. Children with suspected adenoid hypertrophy constitute the vast majority of preschool patients in the ENT outpatient office. The only group that could be considered as a control, would be the children diagnosed with persistent hoarseness who underwent laryngeal flexible endoscopy using an endoscope inserted through the nose. This is a group of patients which we analyze in other studies. However, it is much less numerous, and in addition we definitely do not see family predisposition in this group so we don’t perform the examinations in siblings. Hence, we were able to search only single pairs of siblings that met the criteria for inclusion in such a comparison, which excludes the possibility of proper statistical analysis.

References:

1.Evcimik MF, Dogru M, Cirik AA, Nepesov MI. Adenoid hypertrophy in children with allergic disease and influential factors. Int J Pediatr Otorhinolaryngol. 2015 May;79(5):694-7. doi: 10.1016/j.ijporl.2015.02.017. Epub 2015 Feb 25. PMID: 25758194.

2.Xu Z, Wu Y, Tai J, Feng G, Ge W, Zheng L, Zhou Z, Ni X. Risk factors of obstructive sleep apnea syndrome in children. J Otolaryngol Head Neck Surg. 2020 Mar 4;49(1):11

3.Chng SY, Goh DY, Wang XS, Tan TN, Ong NB. Snoring and atopic disease: a strong association. Pediatr Pulmonol. 2004;38:210–6.

Reviewer 2 Report

 

Summary: The article focuses on a cohort of children that were seen for adenoid symptoms at an ENT clinic over a period of 5 years. The aim was to investigate the relation/risk of developing adenoid hypertrophy in a child whose older sibling also had the same (genetic predisposition) and it was informed by contradictions in the previous studies on this subject. This is a very important paper and the results were informed by a wide scope of tests that explored clinical issues around adenoid hypertrophy. The major weakness is the logical flow of the paper which makes it a little difficult to read and appreciate the clinical application, but this can be easily addressed.

 

Background: The authors need to provide literature on adenoid size and age. This would provide the rationale of doing the study among the siblings at the same age.

What other factors influence adenoid size (especially in the environment) and therefore could potentially affect comparisons done at the same age but at obviously different times.

What causes adenoid hypertrophy? What is the association with genetics to warrant sibling studies? Needs to be described in the background.

What is the potential utility of the study- the rationale?

The last sentence highlights the purpose of the study. According to the background information, the problem seems to be that we do not know if an older sibling has adenoid hypertrophy, the subsequent siblings will also have them. So the authors set out to recruit sibling pairs to compare adenoid size and analyse for the predictors (demographic and clinical). If this is the correct interpretation, then there is need to clarify the objective(s) to indicate so.

Ethics: Please provide information on whether the parents/guardians consented to the study. The title indicates that it was a cohort study (indicating prospective enrolment into the study) and this requires consent from the parents/guardians. If the study was a review of clinic data, then we need to be clear that consent was waved.

 

Methods

What is preliminary school children? Not a widely used term and requires definition

Study design not mentioned. It looks like it was a retrospective review of clinic data or prospective enrolment. How were the 98 pairs of siblings selected? Randomly? How many pairs of siblings were there altogether? What about families that had more than 2 siblings attended to during that period? How did you select the pair out of the many siblings?

 

The study setting needs more description. It is a specialized clinic- which is good but do all the children presenting with adenoid symptoms undergo the tests described in the manuscript?

 

Line 125 introduces the issue of second-born child not previously mentioned. Need clarification if in this study, only 1^st and 2^nd born children were considered or any order, and if the siblings had to follow each other eg 1^st and 2^nd or 2^nd and 3^rd and not 1^st and 3^rd?

Analysis- specify the type of variables for which Chi square or Fisher’s exact test , and Pearson correlation test were used (categorical /continuous). Clarify the association (between what (explanatory) and what (outcome) variables were considered/of interest.

The analysis section needs more clarity for the readers to understand the following:

What were the variables of interest- explanatory versus outcome. It seems the outcome variable was adenoid size, and this was compared between the siblings.

Which variables were analysed as continuous versus categorical ( for associations/regression)

Some variables like snoring, adenoid mucus coverage and tympanogram have scales- how were these analysed to check for association with adenoid size?

In line 129, analysis for prediction of adenoid hypertrophy in second-born children was done and one of the approaches was to consider ‘second-born’ factors. What are these factors?

The sample size was 49 sibling pairs. Was this sample size sufficient to provide valid conclusions, especially on associations?

Whereas the environment is a key factor in adenoid hypertrophy, the data is very silent about it. Is the assumption that the siblings were exposed to the same environmental conditions from birth to the time of measuring the adenoid size?

Results

The first sentence (line 138-141) belongs describes approach to analysis and should be shifted under that section. What was the rationale of this grouping? In what order was the assessment done? For example, was it the 1^st born children that were examined first in all pairs?

Line 145- permanent may be better expressed as persistent.

Table 2: The information in this table requires more clarity in a narrative. Generally, means are used for normally distributed data while medians are used for skewed data. The table provides means (SD) and medians (IQR) for all variables. What relationship were the authors looking for?

 

Tables 3, 4 and 5, and the related narratives are a bit confusing. It is difficult to logically follow and derive the message. I would propose that the authors present the key results- based on the primary objective (which also requires clarity), and any other interesting information presented thereafter with a clear description.

In the discussion, authors make some comparisons with tonsillectomy, but the study was purely about adenoids. I would propose that the comparisons stick to adenoids and preferably in the same age group.

There seems to be quite a bit of emphasis on second-born children. It is not yet clear to me what the significance of this as it was not described in the background/literature.

Author Response

Summary: The article focuses on a cohort of children that were seen for adenoid symptoms at an ENT clinic over a period of 5 years. The aim was to investigate the relation/risk of developing adenoid hypertrophy in a child whose older sibling also had the same (genetic predisposition) and it was informed by contradictions in the previous studies on this subject. This is a very important paper and the results were informed by a wide scope of tests that explored clinical issues around adenoid hypertrophy. The major weakness is the logical flow of the paper which makes it a little difficult to read and appreciate the clinical application, but this can be easily addressed.

 Answer: We are very much thankful to the Reviewer for his valuable and constructive comments which we are sure would better the manuscript manifold. In the revised version, we have tried our best to answer the queries raised during the review process and incorporate all the changes and suggestions made by the reviewers.

Background: The authors need to provide literature on adenoid size and age. This would provide the rationale of doing the study among the siblings at the same age.

Answer: We have added the paragraph concerned on studies of adenoid size change in children and have added the literature.

What other factors influence adenoid size (especially in the environment) and therefore could potentially affect comparisons done at the same age but at obviously different times.

Answer: We have added the paragraph about environmental factors (cigarette smoke, pollution and mole exposure, breastfeeding) and discussed its suitability for the analysis. Also, seasons may be influenced on mucus coverage of the adenoid and tympanometry but not on the adenoid size what confirmed our previous performed study. We have taken this into account in the methodology section.

What causes adenoid hypertrophy? What is the association with genetics to warrant sibling studies? \Needs to be described in the background.

 Answer: The knowledge concerned about relations between genetic factors and AH is still not satisfied we have described it in discussion.

What is the potential utility of the study- the rationale?

Answer: Finding a correlation between the siblings is to justify further studies of both environmental and genetic factors.

The last sentence highlights the purpose of the study. According to the background information, the problem seems to be that we do not know if an older sibling has adenoid hypertrophy, the subsequent siblings will also have them. So the authors set out to recruit sibling pairs to compare adenoid size and analyse for the predictors (demographic and clinical). If this is the correct interpretation, then there is need to clarify the objective(s) to indicate so.

Answer: The purpose of this study was to analyze adenoid size in siblings when they reach the same age. Then we analyzed other factors that may allowed to predict AH in a younger sibling based on medical history (including FNE) of an older one.

Ethics: Please provide information on whether the parents/guardians consented to the study. The title indicates that it was a cohort study (indicating prospective enrolment into the study) and this requires consent from the parents/guardians. If the study was a review of clinic data, then we need to be clear that consent was waved.

Answer: This is a retrospective study, but all parents agreed for endoscopic examination of their children.

 

Methods

What is preliminary school children? Not a widely used term and requires definition

Answer: We changed this term to preschool children. The children age is 3 to 7 years of old. Thank you for drawing attention to this issue, because in many countries the age of preschool children is different. We’ve added this information in the text.

Study design not mentioned. It looks like it was a retrospective review of clinic data or prospective enrolment. How were the 98 pairs of siblings selected? Randomly? How many pairs of siblings were there altogether? What about families that had more than 2 siblings attended to during that period? How did you select the pair out of the many siblings?

Answer: As we described earlier, we retrospectively select all 82 pair of siblings from whole group of examined children of 3-7 years age. Both siblings should be diagnosed with use of flexible endoscopic examination of the nasopharynx by the one doctor. Then we analyzed the if the children were diagnosed in the same age (the permissible age different should not exceed 12 months)- this left 49 pairs of siblings and we analysed this group.

The study setting needs more description. It is a specialized clinic- which is good but do all the children presenting with adenoid symptoms undergo the tests described in the manuscript?

 Answer: Yes, all the children presenting with adenoid symptoms undergo the tests described in the manuscript. We have described it in the Line  89 “All children performed an ENT physical examination, flexible fiberoptic rhinoscopy, and tympanometry”

Line 125 introduces the issue of second-born child not previously mentioned. Need clarification if in this study, only 1^st and 2^nd born children were considered or any order, and if the siblings had to follow each other eg 1^st and 2^nd or 2^nd and 3^rd and not 1^st and 3^rd?

Answer: If large families were analyzed, only one pair of siblings from a given family was analyzed. We guided by the criterion of the smallest age difference between the siblings at the time of examinations. We ‘ve added this information in the manuscript. Thank you.

Analysis- specify the type of variables for which Chi square or Fisher’s exact test , and Pearson correlation test were used (categorical /continuous). Clarify the association (between what (explanatory) and what (outcome) variables were considered/of interest.

Answer: To determine differences between variables, statistical significance was estimated using the Chi-square method or Fisher's exact test for categorical variables and the Stu-dent's t-test for quantitative variables. Associations between quantitative variables were analyzed using Pearson's correlation.

The aim of this study was to compare adenoid size in preschool-age siblings when they reach the same age. Therefore, our outcome variable was adenoid size. This study focused on assessing the significance of clinical or demographic factors in the prognosis of adenoid hypertrophy in siblings. Therefore, our explanatory variables were age (years), sex, rURTI, rhinitis, snoring, adenoid mucus coverage (MASNA scale), tympanogram and thermal season.

The analysis section needs more clarity for the readers to understand the following:

What were the variables of interest- explanatory versus outcome. It seems the outcome variable was adenoid size, and this was compared between the siblings.

Answer: Yes, adenoid size was the main variable. Then we analyze other factors and their relations with AH to find the most important one.

Which variables were analysed as continuous versus categorical (for associations/regression)

Answer: According to Table 1 and Table 2, the categorical variables were: sex, rURTI, rhinitis, Snoring, adenoid size in Bolesławska scale (three categories), adenoid mucus coverage in MASNA scale (4 categories) and tympanogram (three categories). Continuous variables include age (years) and adenoid size (A/C ratio, %).

To assess what affects the adenoid size expressed as a continuous variable (outcome variable) we used linear regression. For the purposes of linear regression explanatory variables that are originally categorical or ordinal variables have been converted to continuous variables, which is a common practice. We inform about the numerical values assigned to individual categories of these variables in the Statistical analysis section.

"In our linear regression analysis, such variables as recurrent upper respiratory tract infections, rhinitis, snoring, adenoid mucus coverage, and type of tympanogram were assigned appropriate values (for recurrent upper respiratory tract infections (rURTI), rhinitis, snoring: 0, symptom not present; 1, symptom present; for adenoid mucus coverage: from 0 to 4 according MASNA scale; for tympanogram type: 0–A, 1–C, 2–B)."

In the logistic regression analysis, explanatory variables were categorical variables, as described in the Statistical analysis section.

"To analyze the associations between a significant increase in adenoid size (adenoid hypertrophy) in the second-born child and clinical factors, such as recurrent up-per respiratory tract infections, rhinitis, snoring, adenoid mucus coverage, and tympanogram type as categorical variables, a logistic regression analysis was performed. Adenoid hypertrophy was defined as an A/C ratio of >65%, based on the Bolesławska scale, where 65% is the cut-off point between IIo and IIIo adenoid hypertrophy."

Some variables like snoring, adenoid mucus coverage and tympanogram have scales- how were these analyzed to check for association with adenoid size?

Answer: The answer to this question was contained in the previous question. Let me copy the relevant part of the previous answer.

To assess what affects the adenoid size expressed as a continuous variable (outcome variable) we used linear regression. For the purposes of linear regression explanatory variables that are originally categorical or ordinal variables have been converted to continuous variables, which is a common practice. We inform about the numerical values assigned to individual categories of these variables in the Statistical analysis section.

"In our linear regression analysis, such variables as recurrent upper respiratory tract infections, rhinitis, snoring, adenoid mucus coverage, and type of tympanogram were assigned appropriate values (for recurrent upper respiratory tract infections (rURTI), rhinitis, snoring: 0, symptom not present; 1, symptom present; for adenoid mucus coverage: from 0 to 4 according MASNA scale; for tympanogram type: 0–A, 1–C, 2–B)."

In line 129, analysis for prediction of adenoid hypertrophy in second-born children was done and one of the approaches was to consider ‘second-born’ factors. What are these factors?

Answer: It was the same factors like the factors considered for first-born children. It is useful to compare groups, to show its similarity and to make comparison for whole group of children. It is shown (in table1) there is only difference in recurrent upper respiratory tract rate between siblings.  One of the reasons why the second born sibling gets sick less frequently, it seems to be earlier immunity to infections with pathogens that caused infections in an older child.

The sample size was 49 sibling pairs. Was this sample size sufficient to provide valid conclusions, especially on associations?

Answer: Second-born' factors include the clinical features of second-born children, such as: recurrent upper respiratory tract infections, rhinitis, snoring, adenoid mucus coverage (MASNA scale) and tympanogram, which we report in the section that was used in the logistic regression analysis (Table 4, part one - Second-born child factors).

Small group sizes do not allow the discovery of all factors affecting a given outcome variable (in our study - chronic adenoid hypertrophy), however, they allow the discovery of variables that most strongly affect this variable. We are pleased that we were able to discover such variables in our study.

Whereas the environment is a key factor in adenoid hypertrophy, the data is very silent about it. Is the assumption that the siblings were exposed to the same environmental conditions from birth to the time of measuring the adenoid size?

Answer: Yes

 

Results

The first sentence (line 138-141) belongs describes approach to analysis and should be shifted under that section.

Answer: We ‘ve moved this sentence. Thank you.

What was the rationale of this grouping? In what order was the assessment done? For example, was it the 1^st born children that were examined first in all pairs?

Answer: Yes, the 1 born child was analyzed first and second born child as a second. We use the word “usually” because this group contained also twins. It might be also interesting to analyzed twins separately but it was too small sample. We still collect new pair of examined twins maybe it will be useful in the future.

Line 145- permanent may be better expressed as persistent.

Answer: We changed the word (also in tables). Thank you.

Table 2: The information in this table requires more clarity in a narrative. Generally, means are used for normally distributed data while medians are used for skewed data. The table provides means (SD) and medians (IQR) for all variables. What relationship were the authors looking for?

Answer: We fully agree with the statement that "Generally, means are used for normally distributed data while medians are used for skewed data"

We wanted to show the results in both forms for greater transparency of the data, however, we agree with the reviewer's intentions suggesting some redundancy of the data, therefore we decided to remove the median (IQR) for all variables, especially since we used the student's t test to determine the differences for quantitative variables.

Tables 3, 4 and 5, and the related narratives are a bit confusing. It is difficult to logically follow and derive the message. I would propose that the authors present the key results- based on the primary objective (which also requires clarity), and any other interesting information presented thereafter with a clear description.

Answer: We wanted to present the results according to the order of analysis and application in subsequent stages of more advanced statistical techniques. The main and key results (association between adenoid size in siblings) was presented beneath the first paragraph which showed the patients group characteristic. This order seems right to us in the subsequent stages we presented other data from the study. The most important results were are also presented in the conclusions.

In the discussion, authors make some comparisons with tonsillectomy, but the study was purely about adenoids. I would propose that the comparisons stick to adenoids and preferably in the same age group.

Answer: Thank you, our study was concentrated on AH and siblings and of course we would like to make comparison with other studies concerned only on adenoid hypertrophy. Unfortunately, although the problem of AH is widespread, little is a study of this problem in the family aspect. Hence, we were forced to refer to studies that analyzed patients with adenoid hypertrophy including children with coexisting hypertrophy of the palatine tonsils.

There seems to be quite a bit of emphasis on second-born children. It is not yet clear to me what the significance of this as it was not described in the background/literature.

Answer: Experiences and observations related to the medical history and examination of an older child can be used to make it easier and faster to diagnose a younger child. We have added this statement in the conclusion.

 

 

Reviewer 3 Report

The authors are to be commended for studying a commonly held belief, that there are "good ENT patient families".  They evaluated siblings' likelihood of having adenoid hypertrophy by using flexible nasopharyngoscopy of siblings at the same age and found significant similarities.

This may mean that a second child does not need an extensive work-up but can be referred earlier or that a family should understand that the problem may arise in other children. 

 

Author Response

Thank you for your interest and acceptation of our work.

Round 2

Reviewer 1 Report

The authors's reply to the problem of multicollinearity was to elaborate a a pairwise correlation matrix. However, looking at correlations only among pairs of predictors is notoriously limiting. The authors should rely on variance inflation factor (VIF) to detect multicollinearity. Consequently, the results of Table 3 may suffer from serious multicollinearity despite the correlation matrix done by the authors.

Moreover, the relationship of snoring in the second-born child and an A/C ratio of >65% in the first-born child, one of their most important conclusions, was not very well investigated. They should have stratified the data and made separate 2 x 2 tables for the two categories of the predictor snoring in the second-born child, so to have two odd-ratios, and not only one (odds ratio of A/C ratio >65% in the first-born child yes/no, in the presence of snoring; odds ratio of A/C ratio >65% in the first-born child yes/no, in the absence of snoring).

Some other points:

(1) The authors improved significantly the introduction, including the change in size of the adenoids according to the age, tobacco smoke exposure, and breastfeeding. However, they selected references that found breastfeeding as a risk factor for OSAS, ignoring those that arrived to a completely different conclusion - breastfeeding is a protective factor for OSAS.

(2) Lines 68: urban planning - area of residence

(3) Lines 134-135: the authors should adopt a more standard interpretation of tympanometry results:

A: normal middle-ear status (good?)

B: there are two types of B, low ear canal volume, and high ear canal volume; why did they consider it the worst outcome? It is more adequate to avoid the use of the adjectives good, worse, and worst, and instead describe each type, for example, B - tympanic membrane with decreased mobility

C: significantly negative pressure in the middle ear

(5) Line 153: adenoid size volatility - what did the authors mean by that?

(6) Table 2: for variables that are not dichotomous, but the authors treated as dichotomous, do the P values (results of chi-square tests) apply to the blended variables, but not to the complete range of variable alternatives?; for example, MASNA scale, P value - MASNA 0 & 1 versus MASNA 2 & 3; snoring, were there two chi-square tests, although the P-values are identical - yes or no; no, occasionally, and persistent? The authors should make the results of the statistical tests clearer, perhaps showing the chi-square values.

 

 

 

 

Author Response

The authors's reply to the problem of multicollinearity was to elaborate a a pairwise correlation matrix. However, looking at correlations only among pairs of predictors is notoriously limiting. The authors should rely on variance inflation factor (VIF) to detect multicollinearity. Consequently, the results of Table 3 may suffer from serious multicollinearity despite the correlation matrix done by the authors.

Answer: As suggested by the reviewer, we present the results of collinearity based on VIF values (Table 1). Calculation programs such as SPSS allow you to automatically calculate the VIF and the second measure of predictor collinearity: the tolerance factor, which is the reciprocal of the VIF.

 

The following interpretations were adopted for the collinearity coefficient:

• A VIF equal to 1 means that the predictors are not collinear.
• If the VIF is < 4 (criteria < 3 or < 5 can also be found in the literature), a lack of collinearity of predictors in the model can be assumed, while when the value of the coefficient is in the range of 4-10, the possibility of collinearity between
• For a VIF value above 10, strong collinearity of the predictors is assumed.In such a case, the analyzed model should be corrected, e.g. by removing one of the predictors with a high VIF or using ridge regression.

 

According to the above assumptions, the variables (predictors) used by us to create the regression model are not collinear- 1<VIF<2.

 

Table 1. Linear regression analysis for the prediction of adenoid size with Collinearity Statistics.

Coefficientsa
Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.	95.0% Confidence Interval for B		Collinearity Statistics
		B	Std. Error	Beta			Lower Bound	Upper Bound	Tolerance	VIF
	rURTI	9.044	3.641	.239	2.484	.015	1.812	16.275	.618	1.617
	Rhinitis	2.730	3.762	.067	.726	.470	-4.742	10.202	.674	1.484
	Snoring	7.468	1.775	.329	4.207	<.001	3.942	10.994	.935	1.070
	Tympanogram	6.254	1.847	.269	3.385	.001	2.585	9.923	.906	1.103
	Adenoid mucus coverage	3.985	1.671	.192	2.385	.019	.666	7.303	.880	1.136
a. Dependent Variable: Adenoid size

 

 

Moreover, the relationship of snoring in the second-born child and an A/C ratio of >65% in the first-born child, one of their most important conclusions, was not very well investigated. They should have stratified the data and made separate 2 x 2 tables for the two categories of the predictor snoring in the second-born child, so to have two odd-ratios, and not only one (odds ratio of A/C ratio >65% in the first-born child yes/no, in the presence of snoring; odds ratio of A/C ratio >65% in the first-born child yes/no, in the absence of snoring).

Answer: The analysis of the impact of two predictors (of snoring in the second-born child and an A/C ratio of >65% in the first-born child) on A/C ratio of >65% in the second-born child is the result of analyzes obtained from logistic regression, in which only these two predictors turned out to be significant. Therefore, to ensure some continuity with the antecedent results (logistic regression), we chose to show a crosstab for significant predictors functioning as a combined single variable. The stratification suggested by the reviewer does not correspond to this idea. Our approach is also supported by the fact that second-born children who do not snoring have an A/C ratio of <=65%, however second-born children who snoring have variable A/C ratio (>65%, <=65 %).

Some other points:

(1) The authors improved significantly the introduction, including the change in size of the adenoids according to the age, tobacco smoke exposure, and breastfeeding. However, they selected references that found breastfeeding as a risk factor for OSAS, ignoring those that arrived to a completely different conclusion - breastfeeding is a protective factor for OSAS.

Answer: We’ve add the different other authors conclusion statements concerned about OSAS and breast-feeding.

(2) Lines 68: urban planning - area of residence

Answer: Corrected, thank you.

(3) Lines 134-135: the authors should adopt a more standard interpretation of tympanometry results:

A: normal middle-ear status (good?)

B: there are two types of B, low ear canal volume, and high ear canal volume; why did they consider it the worst outcome? It is more adequate to avoid the use of the adjectives good, worse, and worst, and instead describe each type, for example, B - tympanic membrane with decreased mobility

C: significantly negative pressure in the middle ear

Answer: Corrected thank you.

(5) Line 153: adenoid size volatility - what did the authors mean by that?

Answer: The "adenoid size volatility" means the variability that we can observe for adenoid size. For a better understanding, we changed "adenoid size volatility" to "adenoid size variability"

(6) Table 2: for variables that are not dichotomous, but the authors treated as dichotomous, do the P values (results of chi-square tests) apply to the blended variables, but not to the complete range of variable alternatives? for example, MASNA scale, P value - MASNA 0 & 1 versus MASNA 2 & 3; snoring, were there two chi-square tests, although the P-values are identical - yes or no; no, occasionally, and persistent? The authors should make the results of the statistical tests clearer, perhaps showing the chi-square values.

Answer: In line with the title of Table 2, it shows the results for adenoid size among various subgroups of patients formed by demographic or clinical factors. According to the first variable, which is sex, we present the average (+/- SD) adenoid size in the group of girls and in the group of boys, and then we present the result of statistical analysis to assess whether the size of the adenoid differs between girls and boys, using the appropriate test for the quantitative variable - Student's t-test. When comparing more than two groups, ANOVA was used.

For a better understanding of the statistical tests used, we have corrected the statistical analysis section.

The second column in table 2 indicates the number of groups compared. For the variable Adenoid mucus coverage (MASNA scale), one basic analysis was performed, taking into account all categories of this variable. Additional category variants for this variable were not created due to the lack of their potential use in further analyses, although looking at the means for individual groups, comparing MASNA 0 & 1 versus MASNA 2 & 3 would also give a statistically significant result. For snoring, the P-values are not identical, however both P-values are below 0.001.

Reviewer 2 Report

Abstract

Methods- mention the study design and how data was collected.

Conclusion not clear. Is it that assessing adenoid size in the first born can help predict the adenoid size in the second born sibling? In which case, having an older sibling with III◦ AH is a good predictor of the same in the sibling that follows them?

 

Line 15-16 states that the study aimed at  assessing the significance of clinical factors in the prognosis of adenoid hypertrophy (AH) in siblings. However, line 79-80, it indicates that the aim was to analyse the adenoid size in siblings when they reach the same age and substantiate a correlation between AH and adenoid symptoms. These statements mean different things. It is better to be consistent in stating the aim of the study.

Line 85-87 indicates that the 82 pairs of siblings had visited the outpatient  ENT. What is not clear is whether they also presented with symptoms of AH. If one wanted to find out if having AH in an older siblings predicts the occurrence of the same in the younger one, then assessing the young sibling at the time when they visit the clinic for AH symptoms certainly indicates that the problem already exists, and there is nothing really to predict. The comparator in this case may not have been very appropriate creating selection bias and this needs to be stated as a limitation. If this was not the case, the authors need to clarify on the characteristics of the young siblings, the same way they indicated that the older siblings had been at the clinic because of AH symptoms.

Line 149- Student’s t-test is for continuous variables. Categorical variables are quantitative too.

Author Response

Methods- mention the study design and how data was collected.

Answer:  We’ve add the study design and exclusion criteria

Conclusion not clear. Is it that assessing adenoid size in the first born can help predict the adenoid size in the second born sibling? In which case, having an older sibling with III◦ AH is a good predictor of the same in the sibling that follows them?

Answer: Thank you, we’ve corrected the conclusions in the abstract. Now seems to be clearer.

 

Line 15-16 states that the study aimed at  assessing the significance of clinical factors in the prognosis of adenoid hypertrophy (AH) in siblings. However, line 79-80, it indicates that the aim was to analyse the adenoid size in siblings when they reach the same age and substantiate a correlation between AH and adenoid symptoms. These statements mean different things. It is better to be consistent in stating the aim of the study.

Answer: Corrected. Thank you

Line 85-87 indicates that the 82 pairs of siblings had visited the outpatient  ENT. What is not clear is whether they also presented with symptoms of AH. If one wanted to find out if having AH in an older siblings predicts the occurrence of the same in the younger one, then assessing the young sibling at the time when they visit the clinic for AH symptoms certainly indicates that the problem already exists, and there is nothing really to predict. The comparator in this case may not have been very appropriate creating selection bias and this needs to be stated as a limitation. If this was not the case, the authors need to clarify on the characteristics of the young siblings, the same way they indicated that the older siblings had been at the clinic because of AH symptoms.

Answer: Let me cite the text from manuscript: “We retrospectively analyzed a group of 1,247 preschool children (3–7 year of age) who visited a medical outpatient ear, nose, and throat (ENT) clinic with symptoms suggestive of chronic AH between 2016 and 2021.” All patients had the symptoms suggestive AH. Symptoms doesn’t mean the same as AH.

Line 149- Student’s t-test is for continuous variables. Categorical variables are quantitative too.

Answer:

In the "Statistical analysis" section, we have introduced information on the statistical tests used depending on the type of variables used (continuous variables or categorical variables).

The sentence: To determine differences between variables, statistical significance was estimated using the Chi-square method or Fisher’s exact test for categorical variables and the Student’s t-test for quantitative variables.