BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

PEER REVIEW HISTORY

BMJ Open publishes all reviews undertaken for accepted manuscripts. Reviewers are asked to complete a checklist review form (http://bmjopen.bmj.com/site/about/resources/checklist.pdf) and are provided with free text boxes to elaborate on their assessment. These free text comments are reproduced below.

ARTICLE DETAILS

TITLE (PROVISIONAL) Age- and sex-specific reference intervals for non-fasting lipids and apolipoproteins in 7260 healthy Chinese children and adolescents measured with an Olympus AU5400 analyser: a cross-sectional study AUTHORS Liu, Junjie; Dai, Yanpeng; Yuan, Enwu ; Li, Yushan; Wang, Quanxian; Wang, Linkai; Su, Yanhua

VERSION 1 – REVIEW

REVIEWER Adagmar Andriolo Escola Paulista de Medicina REVIEW RETURNED 13-Mar-2019

GENERAL COMMENTS Very important topic. These data will be added to the other studies on reference intervals.

REVIEWER Martin Frank Strand Høyskolen Kristiania REVIEW RETURNED 25-Mar-2019

http://bmjopen.bmj.com/ GENERAL COMMENTS First, I would like to commend the authors on the scope and importance of the work they present here. They have gathered a unique dataset, with material/data from a large number of healthy pediatric participants, and the establishment of relevant reference intervals is very important both for diagnostics, but also to track population health over time.

At the same time there are several improvements the authors can on September 25, 2021 by guest. Protected copyright. do to more clearly present their results, and to get more out of the dataset they have gathered.

4. Are the methods described sufficiently to allow the study to be repeated? There are several elements that need to be addressed in the methods section in order to ensure repeatability. a) The methods lack information about how the participants was recruited b) The methods lack information about where (location/region) the participants were recruited from. This is also important in regards of the use of the reference intervals. c) *How was the non-parametric calculation of the reference intervals performed (software, short description of method)? d) The number of candidates that was excluded due to history with hypertension, DM, coronary heart disease etc. is lacking. e) Reference 11 (Harris Boyd) is not placed correctly in the methods text.

1

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from f) Does the laboratory undergo interlaboratory testing schemes to ensure commutability?

7. If statistics are used are they appropriate and described fully? Se above*

10. Are they presented clearly? I marked this as no, but the answer should be yes and no. The reference intervals are clearly defined and presented in a good form in the table that is provided. However, the results provided lack some information that would be valuable to the readers in interpreting the data. This study presents reference intervals based on 7260 participants, but there is very little information about the age and gender distribution of the study population. Figure 1, 2 and 3 repeats the info given in table 2 graphically, but as partitioning by age and gender is an important part of establishing reference intervals, the distribution across age should be visualized with age as a continuous parameter. Thus, the reader will be able to see the changes in lipid levels across age and gender, and this will also increase the ability of readers to compare lipid distribution across age/gender with similar studies in other regions/countries, as this is a common way to present this type of data. The authors could also add information about the prevalence of dyslipidemia based on the data they have gathered. This would be very valuable information as the prevalence of dyslipidemia among healthy children isn’t well known. a) Please add a graph or table that clearly shows/visualizes the number of participants of each age and gender. Examples: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5410076/ figure 1 and https://www.ncbi.nlm.nih.gov/pubmed/26948098 figure 2. b) I highly recommend re-making figure 1-3 as percentile curves (se same references as above) or adding percentile curves in addition to figure 1-3. http://bmjopen.bmj.com/ c) I suggest analyzing and presenting the prevalence of dyslipidemia from the study population. This will add to the aims of the study as they now are stated in the manuscript, but this is very important health information to extract when such a good dataset has been gathered. d) In table 1, the CV-percentages are listed. I suggest adding a sentence under results where you summarize CV-values (give

the range). on September 25, 2021 by guest. Protected copyright.

11. Are the discussion and conclusions justified by the results The discussion is a bit thin given the data presented. How does the reference intervals you have found compare to reference values in use now? How do they compare to published reference intervals in similar studies in other countries? The take home messages listed t the start of the manuscript could be addressed in greater detail in the discussion. Do you find evidence to support the first point listed under the take home messages when you compare your results to others?

12. Are the study limitations discussed adequately? Some strengths/limitations are briefly listed after the abstract but is not discussed adequately in the discussion section. The elements listed after the abstract should be elaborated on more clearly in the discussion. Information about recruitment method, location, age

2

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

and gender distribution etc. is lacking from methods, and possible limitations to the study population are not addressed.

Corrections: Sentence 22, page 7: “lipids levels”. Normally when using two nouns in a row; only the lat is given the plural form: “lipid levels”.

Conclusion: I recommend a minor revision of the manuscript; with some more work on/additions to the text in the methods, results and discussion sections. I also recommend some work on additional figures in order to present the gathered data in a more useful form.

REVIEWER Tenna Ruest Haarmark Nielsen Department of Pediatrics, Copenhagen University Hospital Rigshospitalet, Denmark REVIEW RETURNED 17-Apr-2019

GENERAL COMMENTS This is a very nice study, with a large number of participants. However, I have som general and som specific comments listed below:

Title: Apparently – are they considered healthy or not? Consider eliminating that word, and instead list the number, which is very impressive for such a cohort.

P 2: Abstract – aim: Ethnic more than geographic factors influence lipids. P2 line 30: ”Outliers were removed” – this seems too detailed for an abstract. Should be kept in the main text. P 2 Results: Very short. Would be preferable, if the references were a bit described. Were there age differences? Were there http://bmjopen.bmj.com/ gender differences? P 2: Conclusions: This could be more to the point if taking into account which reference values have been used previously for Chinese children – how does this change with the new references? What does this study really bring to the world of new knowledge.

Introduction: Obesity is a considerable factor in terms of dyslipidemia in children on September 25, 2021 by guest. Protected copyright. and adolescents. This should be touched upon, also in a context that takes into account the prevalence and development of obesity in or at least in the world. Also, the level of lipids that seems to cause CVD may be different among ethnicities (I.e. In Chinese adults, variations, even within levels usually regarded as low concentrations of TC (3.8−4.7 mmol/L), are associated with increased mortality from coronary heart disease) (Chen Z, Peto R, Collins R, MacMahon S, Lu J, Li W. Serum cholesterol concentration and coronary heart disease in population with low cholesterol concentrations. BMJ. 1991;303(6797):276-282.)

Subjects and methods: How were participants enrolled? From where? What type of community? Can they be considered representative of the specific community/region/country? The cohort should be described in detail when the aim as stated is to generate reference values. Otherwise the reader can not decide

3

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from whether the population in the study can be compared to any other given population where the references may be used

How were exclusions made – based on journals – or based on interview? How were weight and height measured? (Types of equipment? With/without shoes= With/without clothes?) P 5 ln 6pp: How many participants were excluded based on diagnoses listed in this section? P 5 ln 17pp: Seems to be very low numbers for use of medication and especially for obesity? How were these individuals recruited? Have you considered whether the cohort is representative for the Chinese population in general? According this report from Unicef it seems that 5-10% of children/adolescents have obesity in China – depending on rural or urban geography (http://www.unicef.cn/en/uploadfile/2018/0423/2018042311440016 5.pdf) P 5 ln 20: What acute clinical symptoms were causes for exclusion? P 5 ln 35: Non-fasting values – although HDL and total cholesterol are little affected by whether samples are drawn in the fasting or non-fasting state, triglycerides are on average 20% higher in the non-fasting state, and presents with great variability. This will also affect LDL concentrations, if they are calculated by i.e. the Friedewald equation. How is the LDL concentration extracted from the Olympus AU5400? It seems the LDL is directly measured by clearance though. The fact that the values are non-fasting should be clearly stated in title and/or abstract. Furthermore, this should be discussed in the discussion.

P 6 ln 12 ref 12 does not refer to CLSI but to guidelines in CVD health and risk reduction. Should probably be ref 6? Please make sure other references are accurately referred to.

http://bmjopen.bmj.com/ The Harris-Boyd method seems to be only valid for Gaussian distributed data (https://doi.org/10.1373/clinchem.2003.027953). Have the authors tested whether their data fits this distribution? This should be stated in the methods section along with how, and if data was transformed to fit the model, then this should be stated.

Results:

P 6 ln 25: “healthy” – previously “apparently healthy” – this should on September 25, 2021 by guest. Protected copyright. be consistent – consider using “healthy” all around, as you have excluded those you believe to have illness affecting lipid levels. Otherwise consider simply stating “children and adolescents” in P 4 ln 56, where you have not yet considered whether they are healthy or not. Upper age group up to 13 years – you should consider puberty as a factor in this age group. Would be informative to have a general descriptive table, including a diagram indicating how many individuals at each age, not just in the different age groups.

Figures: Consider plotting the references over the continuous age span. This has been done by other groups i.e. the IDEFICS consortium in Europe (Int J Obes (Lond). 2014 Sep;38 Suppl 2:S67-75.). Consider plotting the 95% confidence intervals.

Discussion:

4

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

In general a discussion should be used to discuss the present study taking other previous findings into account. The current text in the discussion does not live up to this purpose, but rather repeats from the indtroduction and results. Considerable effort is needed in this section to discuss the findings from the present study including what this study adds in terms of new knowledge. Would be worthwhile to take tracking of CVD markers into account when stating the need for measuring lipids in childhood (lipoproteins are the CVD risk factors showing the strongest tracking properties (Twisk JWR, Kemper HCG, Mechelen W van, Post GB. Tracking of Risk Factors for Coronary Heart Disease over a 14-Year Period: A Comparison between Lifestyle and Biologic Risk Factors with Data from the Amsterdam Growth and Health Study. Am J Epidemiol. 1997;145(10):888-898)

P 7 ln 9 “Previous study” – should be “A previous study”, although one could also use the plural previous studies, as multiple studies have investigated this. Likewise further down in the text.

It is stated that lipid levels are linked to sex especially after puberty – but the authors have not described puberty in their cohort, even though at age 13 a lot of them should have at least entered puberty. Puberty should be taken into account, or the authors should at least discuss this as a limitation to the study. Have other risk factors for lipid levels been taken into account? I.e. diet, lifestyle, family disposition? How these factors may have influenced the results should be discussed. The factors are mentioned in the discussion from other studies, but are not related to the present study, which should be the main focus of the discussion. Further it is stated that differences in analytical methods compose a problem, but it seems there in this part of the discussion is not distinguished between different populations and different methods, which of course both at problems, but two different types of http://bmjopen.bmj.com/ problems.

Since this seems to be new references for a Chinese population, the discussion should also describe how these new references distinguish from previous used references i.e. other international references – how are the Chinese levels compared to i.e. other frequent references like the AHA references (Circulation.

2003;107(11):1562-1566.) and how are they compared to i.e. adult on September 25, 2021 by guest. Protected copyright. Chinese references/cut offs.

It is stated that references for local populations are relevant. How local? How is the references generated in the present study generalizable to the general Chinese population? Is the cohort representative ? And for which part of the population? If not, when can the present referneces be used?

The discussion lacks discussion of strengths and limitations.

It has already in previous studies been established, that lipids vary with age, gender and ethnicity. This study should discuss what new knowledge is brings – this could be how the present references would change clinical management, or how the present references are different from others. The conclusions would then be more to the point.

5

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

VERSION 1 – AUTHOR RESPONSE

Part B( Reviewer 1): 1. The Reviewer’s comment: Very important topic. These data will be added to the other studies on reference intervals. The authors’ answer: We admire the reviewer’s profound knowledge. And we really appreciate your good evaluation and valuable advice. Thank you very much!

Part C( Reviewer 2): 1. The Reviewer’s comment: First, I would like to commend the authors on the scope and importance of the work they present here. They have gathered a unique dataset, with material/data from a large number of healthy pediatric participants, and the establishment of relevant reference intervals is very important both for diagnostics, but also to track population health over time. At the same time there are several improvements the authors can do to more clearly present their results, and to get more out of the dataset they have gathered. 4. Are the methods described sufficiently to allow the study to be repeated? There are several elements that need to be addressed in the methods section in order to ensure repeatability. a) The methods lack information about how the participants was recruited The authors’ answer: We really appreciate your good evaluation and valuable advice. We also thank the reviewer for raising this important issue, which has been added in the revised manuscript. Please see line 13-17 on page 5 in the revised manuscript.

2. The Reviewer’s comment: b) The methods lack information about where (location/region) the participants were recruited from. This is also important in regards of the use of the reference intervals. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been added in the revised manuscript. Please see line 13-19 on page 5 in the revised manuscript.

3. The Reviewer’s comment: c) *How was the non-parametric calculation of the reference intervals performed (software, short description of method)? http://bmjopen.bmj.com/ The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been added in the revised manuscript. Please see line 9-14 on page 7 in the revised manuscript.

4. The Reviewer’s comment: d) The number of candidates that was excluded due to history with hypertension, DM, coronary heart disease etc. is lacking.

The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for on September 25, 2021 by guest. Protected copyright. raising this important issue. In this study, 7605 children and adolescents (4125 boys, 3480 girls, aged 0-13 years) were randomly recruited from the Health Management Center of the Third Affiliated Hospital of University (also known as Maternal and Children Health Hospital). In order to obtain apparently healthy individuals, only those who visited our hospital for a routine health checkup were enrolled into this study. We carefully investigated the history of the individuals enrolled in the present study. Only individuals without a history of hypertension, diabetes mellitus, coronary heart disease, renal disease, inherited metabolic diseases were included in this study. Individuals were excluded from this study for the following reasons: taking medications, acute clinical symptoms, and obesity. After applying exclusion, 345 were excluded from this study. Finally, This study population consisted of 7260 healthy children and adolescents.

5. The Reviewer’s comment: e) Reference 11 (Harris Boyd) is not placed correctly in the methods text. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, the error has been changed in the revised manuscript.

6

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

6. The Reviewer’s comment: f) Does the laboratory undergo interlaboratory testing schemes to ensure commutability? The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue. We regularly participate in an external quality assessment (EQA) schemes organized by the Chinese National Center for Clinical Laboratories. And accuracy (measured by Bias%) was calculated from EQA and presented in Table 1.

7. The Reviewer’s comment: f) 7. If statistics are used are they appropriate and described fully? Se above* The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been added in the revised manuscript. Please see line 9-14 on page 7 in the revised manuscript.

8. The Reviewer’s comment: a) Please add a graph or table that clearly shows/visualizes the number of participants of each age and gender. Examples: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5410076/ figure 1 and https://www.ncbi.nlm.nih.gov/pubmed/26948098 figure 2. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been added in the revised manuscript. Please see Table 2 and Figure 2 in the revised manuscript.

9. The Reviewer’s comment: b) I highly recommend re-making figure 1-3 as percentile curves (se same references as above) or adding percentile curves in addition to figure 1-3. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see Figure 3 in the revised manuscript.

10. The Reviewer’s comment: c) I suggest analyzing and presenting the prevalence of dyslipidemia from the study http://bmjopen.bmj.com/ population. This will add to the aims of the study as they now are stated in the manuscript, but this is very important health information to extract when such a good dataset has been gathered. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue. The criteria for diagnosis of dylipidemia, however, vary among population. Therefore, each country is encouraged to have its own criteria. Even in the same area, the standard should be changed with the development of the economy. Current cut-offs for dyslipidemia in children on September 25, 2021 by guest. Protected copyright. and adolescents are based on data from The Editorial Board of Chinese Journal of Pediatrics. (Editorial Board of Chinese Journal of Pediatrics, Subspecialty Group of Child Health Care, The Society of Pediatrics, Chinese Medical Associaton, Subspecialty Group of Cardiovascular Disease, The Society of Pediatrics, Chinese Medical Association, Subspecialty Group of Atherosclerosis, The Society of Cardiovascular Disease, Chinese Medical Association. Experts consensus for prevention and treatment of dyslipidemia in children and adolescents. Zhonghua Er Ke Za Zhi 2009: 426-8.) And diagnostic criteria for dyslipidemia in children and adolescents are based on fasting lipid levels. However, the lipid profile levels in our study were measured in non-fasting state. And there is lack of diagnostic criteria for dyslipidemia in children aged 0-2 years old. Therefore, we do not analyze and present the prevalence of dyslipidemia from the study population.

11. The Reviewer’s comment: d) In table 1, the CV-percentages are listed. I suggest adding a sentence under results where you summarize CV-values (give the range).

7

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been added in the revised manuscript. Please see line 4 on page 7 in the revised manuscript.

11. The Reviewer’s comment: Are the discussion and conclusions justified by the results The discussion is a bit thin given the data presented. How does the reference intervals you have found compare to reference values in use now? How do they compare to published reference intervals in similar studies in other countries? The take home messages listed t the start of the manuscript could be addressed in greater detail in the discussion. Do you find evidence to support the first point listed under the take home messages when you compare your results to others? The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see line 9-22 on page 9 in the revised manuscript.

12. The Reviewer’s comment: Are the study limitations discussed adequately? Some strengths/limitations are briefly listed after the abstract but is not discussed adequately in the discussion section. The elements listed after the abstract should be elaborated on more clearly in the discussion. Information about recruitment method, location, age and gender distribution etc. is lacking from methods, and possible limitations to the study population are not addressed. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see line 13-17 on page 5 and -8 on page 10 in the revised manuscript.

13. The Reviewer’s comment: Sentence 22, page 7: “lipids levels”. Normally when using two nouns in a row; only the lat is given the plural form: “lipid levels”. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, the error has been changed in the revised manuscript.

14. The Reviewer’s comment: I recommend a minor revision of the manuscript; with some more work http://bmjopen.bmj.com/ on/additions to the text in the methods, results and discussion sections. I also recommend some work on additional figures in order to present the gathered data in a more useful form. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Thank you very much!

Part D( Reviewer 3): on September 25, 2021 by guest. Protected copyright. 1. The Reviewer’s comment: This is a very nice study, with a large number of participants. However, I have som general and som specific comments listed below: The authors’ answer: We admire the reviewer’s profound knowledge. We really appreciate your good evaluation and valuable advice. Thank you very much!

2. The Reviewer’s comment: Title: Apparently – are they considered healthy or not? Consider eliminating that word, and instead list the number, which is very impressive for such a cohort. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue. They are considered healthy in our study. We have changed in the revised manuscript.

3. The Reviewer’s comment: P 2: Abstract – aim: Ethnic more than geographic factors influence lipids.

8

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please on page 2 in the revised manuscript.

4. The Reviewer’s comment: P2 line 30: ”Outliers were removed” – this seems too detailed for an abstract. Should be kept in the main text. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been removed in the revised manuscript.

5. The Reviewer’s comment: P 2 Results: Very short. Would be preferable, if the references were a bit described. Were there age differences? Were there gender differences? The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been removed in the revised manuscript.

6. The Reviewer’s comment: P 2: Conclusions: This could be more to the point if taking into account which reference values have been used previously for Chinese children – how does this change with the new references? What does this study really bring to the world of new knowledge. reference intervals used previously did not take account into age The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see line 18-20 on page 2 in the revised manuscript.

7. The Reviewer’s comment: Obesity is a considerable factor in terms of dyslipidemia in children and adolescents. This should be touched upon, also in a context that takes into account the prevalence and development of obesity in China or at least in the world. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see line 6-8 on page 4 in the manuscript.

8. The Reviewer’s comment: Also, the level of lipids that seems to cause CVD may be different http://bmjopen.bmj.com/ among ethnicities (I.e. In Chinese adults, variations, even within levels usually regarded as low concentrations of TC (3.8−4.7 mmol/L), are associated with increased mortality from coronary heart disease) (Chen Z, Peto R, Collins R, MacMahon S, Lu J, Li W. Serum cholesterol concentration and coronary heart disease in population with low cholesterol concentrations. BMJ. 1991;303(6797):276- 282.) The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see line 11-13 on September 25, 2021 by guest. Protected copyright. on page 4 in the manuscript.

9. The Reviewer’s comment: Subjects and methods: How were participants enrolled? From where? What type of community? Can they be considered representative of the specific community/region/country? The cohort should be described in detail when the aim as stated is to generate reference values. Otherwise the reader can not decide whether the population in the study can be compared to any other given population where the references may be used The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. This cross-sectional study was performed in Zhengzhou, Henan Province, China. From June 2016 to January 2019, 7605 children and adolescents (4125 boys, 3480 girls, aged 0-13 years) were randomly recruited from the Health Management Center of the Third Affiliated Hospital of (also known as Henan Maternal and Children Health Hospital). The study individuals were those who visited our hospital for a routine health checkup. Only individuals who have resided in Zhengzhou, Henan

9

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

Province for at least six months are enrolled into this study. The subjects’ residential places covered all the regions (including the Erqi , , Guancheng Hui District, District, , , and Zhengdong New District). Please see line 13-22 on page 5 in the manuscript.

10. The Reviewer’s comment: How were exclusions made – based on journals – or based on interview? The authors’ answer: We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see on page 6 in the manuscript.

11. The Reviewer’s comment: How were weight and height measured? (Types of equipment? With/without shoes= With/without clothes?) The authors’ answer: We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see line 5-7 on page 6 in the manuscript.

12. The Reviewer’s comment: P 5 ln 6pp: How many participants were excluded based on diagnoses listed in this section? The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see line 11-13 on page 6 in the manuscript.

13. The Reviewer’s comment: P 5 ln 17pp: Seems to be very low numbers for use of medication and especially for obesity? How were these individuals recruited? Have you considered whether the cohort is representative for the Chinese population in general? According this report from Unicef it seems that 5-10% of children/adolescents have obesity in China – depending on rural or urban geography (http://www.unicef.cn/en/uploadfile/2018/0423/20180423114400165.pdf) The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue. We only recruited apparently healthy children and adolescents from the Health Management Center of the Henan Maternal and Children Health Hospital. And the individuals enrolled into study are those who visited our institution for routine health checkup. The individuals http://bmjopen.bmj.com/ who are obvious obesity, overweight or underweight are not included into this study. 14. The Reviewer’s comment: P 5 ln 17pp: Seems to be very low numbers for use of medication and especially for obesity? How were these individuals recruited? Have you considered whether the cohort is representative for the Chinese population in general? According this report from Unicef it seems that 5-10% of children/adolescents have obesity in China – depending on rural or urban geography (http://www.unicef.cn/en/uploadfile/2018/0423/20180423114400165.pdf)

The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for on September 25, 2021 by guest. Protected copyright. raising this important issue. We only recruited apparently healthy children and adolescents when recruiting students as subjects. And the individuals enrolled into study are those who visited our institution for routine health checkup. Those subjects were randomly recruited from the Health Management Center of the Henan Maternal and Children Health Hospital. The individuals who are obvious obesity, overweight or underweight are not included into this study.

15. The Reviewer’s comment: P 5 ln 20: What acute clinical symptoms were causes for exclusion? The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue. Acute clinical symptoms such as fever and sore throat can affect lipid levels. (Nassaji M , Ghorbani R. Plasma lipids levels in patients with acute bacterial infections. International Journal of Infectious Diseases, 2012, 15:S51-S51; Farshtchi D, Lewis V J. Effects of three bacterial infections on serum lipids of rabbits[J]. Journal of Bacteriology, 1968,95(95):1615- 1621.)

10

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

16. The Reviewer’s comment: P 5 ln 35: Non-fasting values – although HDL and total cholesterol are little affected by whether samples are drawn in the fasting or non-fasting state, triglycerides are on average 20% higher in the non-fasting state, and presents with great variability. This will also affect LDL concentrations, if they are calculated by i.e. the Friedewald equation. How is the LDL concentration extracted from the Olympus AU5400? It seems the LDL is directly measured by clearance though. The fact that the values are non-fasting should be clearly stated in title and/or abstract. Furthermore, this should be discussed in the discussion. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. LDL cholesterol was directly determined and not calculated using the Friedewald formula. Please see the new title and abstract in the revised manuscript. And see line 3-8 on page 9 in the discussion of the revised manuscript.

17. The Reviewer’s comment: P 6 ln 12 ref 12 does not refer to CLSI but to guidelines in CVD health and risk reduction. Should probably be ref 6? Please make sure other references are accurately referred to. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, the error has been changed in the revised manuscript. Please see line 20 on page 7 in the revised manuscript.

18. The Reviewer’s comment: The Harris-Boyd method seems to be only valid for Gaussian distributed data (https://doi.org/10.1373/clinchem.2003.027953). Have the authors tested whether their data fits this distribution? This should be stated in the methods section along with how, and if data was transformed to fit the model, then this should be stated. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue. The Shapiro-Wilk (S-W) test was used to evaluate the distribution of data. In the event of a non-Gaussian, the data are first transformed to an approximate Gaussian distribution using Box-Cox transformation method. After this transform, data fit the model. Please see line 16-18 on page 7. http://bmjopen.bmj.com/

19. The Reviewer’s comment: P 6 ln 25: “healthy” – previously “apparently healthy” – this should be consistent – consider using “healthy” all around, as you have excluded those you believe to have illness affecting lipid levels. Otherwise consider simply stating “children and adolescents” in P 4 ln 56, where you have not yet considered whether they are healthy or not. The authors’ answer: We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. on September 25, 2021 by guest. Protected copyright.

20. The Reviewer’s comment: Upper age group up to 13 years – you should consider puberty as a factor in this age group. Would be informative to have a general descriptive table, including a diagram indicating how many individuals at each age, not just in the different age groups. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue. There are two possible limitations to this study. one limitation is the lack of data on pubertal stage of the participants. And we discuss this limitation in the revised manuscript. We have added a general table and a diagram in the revised manuscript. Please see Table 2 and Fig. 2 in the revised manuscript.

21. The Reviewer’s comment: Figures: Consider plotting the references over the continuous age span. This has been done by other groups i.e. the IDEFICS consortium in Europe (Int J Obes (Lond). 2014 Sep;38 Suppl 2:S67-75.). Consider plotting the 95% confidence intervals.

11

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see the Fig. 3 in the revised manuscript.

22. The Reviewer’s comment: In general a discussion should be used to discuss the present study taking other previous findings into account. The current text in the discussion does not live up to this purpose, but rather repeats from the indtroduction and results. Considerable effort is needed in this section to discuss the findings from the present study including what this study adds in terms of new knowledge. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see line 9-22 on page 9 in the revised manuscript.

23. The Reviewer’s comment: Would be worthwhile to take tracking of CVD markers into account when stating the need for measuring lipids in childhood (lipoproteins are the CVD risk factors showing the strongest tracking properties (Twisk JWR, Kemper HCG, Mechelen W van, Post GB. Tracking of Risk Factors for Coronary Heart Disease over a 14-Year Period: A Comparison between Lifestyle and Biologic Risk Factors with Data from the Amsterdam Growth and Health Study. Am J Epidemiol. 1997;145(10):888-898) The authors’ answer: We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see line 21-22 on page 8 in the revised manuscript.

24. The Reviewer’s comment: P 7 ln 9 “Previous study” – should be “A previous study”, although one could also use the plural previous studies, as multiple studies have investigated this. Likewise further down in the text. The authors’ answer: We thank the reviewer for raising this important issue, which has been changed in the revised manuscript.

25. The Reviewer’s comment: It is stated that lipid levels are linked to sex especially after puberty – but the authors have not described puberty in their cohort, even though at age 13 a lot of them should http://bmjopen.bmj.com/ have at least entered puberty. Puberty should be taken into account, or the authors should at least discuss this as a limitation to the study. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue. There are two possible limitations to this study. one limitation is the lack of data on pubertal stage of the participants. And we discuss this limitation in the revised manuscript.

26. The Reviewer’s comment: Have other risk factors for lipid levels been taken into account? I.e. on September 25, 2021 by guest. Protected copyright. diet, lifestyle, family disposition? How these factors may have influenced the results should be discussed. The authors’ answer: We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see 16-18 on page 9 in the revised manuscript.

27. The Reviewer’s comment: The factors are mentioned in the discussion from other studies, but are not related to the present study, which should be the main focus of the discussion. The authors’ answer: We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see 9-22 on page 9 in the revised manuscript.

28. The Reviewer’s comment: Further it is stated that differences in analytical methods compose a problem, but it seems there in this part of the discussion is not distinguished between different populations and different methods, which of course both at problems, but two different types of problems.

12

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see 9-22 on page 9 in the revised manuscript.

29. The Reviewer’s comment: Since this seems to be new references for a Chinese population, the discussion should also describe how these new references distinguish from previous used references i.e. other international references – how are the Chinese levels compared to i.e. other frequent references like the AHA references (Circulation. 2003;107(11):1562-1566.) and how are they compared to i.e. adult Chinese references/cut offs. The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see 9-22 on page 9 in the revised manuscript.

30. The Reviewer’s comment: It is stated that references for local populations are relevant. How local? How is the references generated in the present study generalizable to the general Chinese population? Is the cohort representative ? And for which part of the population? If not, when can the present referneces be used? The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue. This cross-sectional study was performed in Zhengzhou, Henan Province, China. From June 2016 to January 2019, 7605 children and adolescents (4125 boys, 3480 girls, aged 0-13 years) were randomly recruited from the Health Management Center of the Third Affiliated Hospital of Zhengzhou University (also known as Henan Maternal and Children Health Hospital). The study individuals were those who visited our hospital for a routine health checkup. Only individuals who have resided in Zhengzhou, Henan Province for at least six months are enrolled into this study. The subjects’ residential places covered all the regions (including the , Jinshui District, Guancheng Hui District, , Shangjie District, Huiji District, and Zhengdong New District). Therefore, the cohort is representative for Zhengzhou, Henan Province, China. 31. The Reviewer’s comment: The discussion lacks discussion of strengths and limitations. The authors’ answer: We thank the reviewer for raising this important issue, which has been changed in the revised manuscript. Please see 5-8 on page 10 in the revised manuscript. http://bmjopen.bmj.com/

32. The Reviewer’s comment: It has already in previous studies been established, that lipids vary with age, gender and ethnicity. This study should discuss what new knowledge is brings – this could be how the present references would change clinical management, or how the present references are different from others. The conclusions would then be more to the point.

The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for on September 25, 2021 by guest. Protected copyright. raising this important issue, which has been changed in the revised manuscript. Please see 9-12 on page 10 in the revised manuscript. Finally, we appreciate very much for your time in editing our manuscript and the reviewers for their valuable suggestions and comments. If you need any other information, please contact us without hesitation. Enwu Yuan Department of Clinical Laboratory The Third Affiliate Hospital of Zhengzhou University No.7 Front Kangfu Street Er’qi District Zhengzhou 450052 Henan China Tel: +86 13937112492.

13

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

Fax: +86 0371 66992000 E-mail: [email protected] Thanks very much for your attention to our paper. Sincerely yours, Enwu Yuan

VERSION 2 – REVIEW

REVIEWER Martin Frank Strand Høyskolen Kristiania, Norway REVIEW RETURNED 22-May-2019

GENERAL COMMENTS The authors have improved the manuscript substantially, and have addressed all the issues brought up in the initial review-round. They present unique and important results that are of use both for further research and for clinical practice. In my opinion the manuscript is now ready to be accepted, with one clarification and two proof changes: 1 - page 4 line 38: the sentence about economy is a bit unclear, and should be edited to more clearly define what the authors want to convey. 2 - page 7 line 25: add "for" in This method is reasonable ... sample sizes 3 - page 9 line 59: add "a" in reference intervals using ... local population.

VERSION 2 – AUTHOR RESPONSE http://bmjopen.bmj.com/

Part B( Reviewer 2):

1. The Reviewer’s comment: page 4 line 38: the sentence about economy is a bit unclear, and should be edited to more clearly define what the authors want to convey.

on September 25, 2021 by guest. Protected copyright.

The authors’ answer: We really appreciate your valuable advice. We also thank the reviewer for raising this important issue, which has been added in the revised manuscript. Please see line 14-15 on page 4 in the revised manuscript.

2. The Reviewer’s comment: page 7 line 25: add "for" in This method is reasonable ... sample sizes

The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been added in the revised manuscript. Please see line 10 on page 7 in the revised manuscript.

3. The Reviewer’s comment: page 9 line 59: add "a" in reference intervals using ... local population.

14

BMJ Open: first published as 10.1136/bmjopen-2019-030201 on 18 August 2019. Downloaded from

The authors’ answer: We admire the reviewer’s profound knowledge. We thank the reviewer for raising this important issue, which has been added in the revised manuscript. Please see on page 10 in the revised manuscript.

Finally, we appreciate very much for your time in editing our manuscript and the reviewers for their valuable suggestions and comments. If you need any other information, please contact us without hesitation.

Enwu Yuan

Department of Clinical Laboratory

The Third Affiliate Hospital of Zhengzhou University

No.7 Front Kangfu Street

Er’qi District

Zhengzhou

450052

Henan

China

Tel: +86 13937112492.

Fax: +86 0371 66992000 http://bmjopen.bmj.com/

E-mail: [email protected]

Thanks very much for your attention to our paper.

Sincerely yours,

Enwu Yuan on September 25, 2021 by guest. Protected copyright.

15