PEER REVIEW HISTORY ARTICLE DETAILS TITLE (PROVISIONAL)

Transcription

1 PEER REVIEW HISTORY BMJ Open publishes all reviews undertaken for accepted manuscripts. Reviewers are asked to complete a checklist review form ( and are provided with free text boxes to elaborate on their assessment. These free text comments are reproduced below. TITLE (PROVISIONAL) AUTHORS REVIEWER ARTICLE DETAILS Cohort profile: The Health and Prevention Enhancement (H- PEACE), a retrospective, population-based cohort study conducted at the Seoul National University Hospital Gangnam Center, Korea Lee, Changhyun; Choe, Eun Kyung; Choi, Ji Min; Hwang, Yunji; Lee, Young; Park, Boram; Chung, Su Jin; Kwak, Min-Sun; Lee, Jong-Eun; Kim, Joo Sung; Park, Sue K.; Cho, Sang-Heon VERSION 1 REVIEW Alfredo Lozada Austral University, Argentina 03-Oct-2017 This manuscript is a Cohort Profile that was collected in a retrospective manner. It was designed to investigate the association of abnormal pre disease diagnostic test results and biomarkers with the later development of NCD. These included malignancies, cardiac and metabolic diseases. Participants were included when they received a health check up between 2003 and In this period, individuals were recruited with an average age of 45.5 years. Only 32,6 completed the 2nd follow up assessment. Maybe more detail should be given on the eligibility and in diagnostic criteria. It should be mentioned why participants did not have repeated variables measured It is noteworthy that there was a male predominance in this young cohort of highly educated participants. Men had a very high percentage of current smokers and it was very low in women. This percentages were also different for gender in drinking. There were also marked gender differences in sistolic and diastolic blood pressure as well as in blood glucose. In this young and lean cohort it was remarkable in the table of abnormal findings to observe that 6% had CKD, 66% abdominal obesity and 29% fatty liver. It is also surprising to see a 54 % H. Pylori infection, a 36% atrophic gastritis and 37% of colonic adenomas. In table 4 showing baseline characteristics and comorbidities, the authors list past hypertension and past diabetes and maybe should be changed to just hypertension and diabetes. In the first and second paragraphs of page 30 the authors make recommendations on the the screening for GI Malignancies. Maybee 1

2 REVIEWER REVIEWER the authors should elaborate more in their interpretation, generalizability and conclusions. Shing Fung Lee Department of Clinical Oncology, Tuen Mun Hospital, Hong Kong 19-Nov-2017 The authors have presented the study rationale and methodology well, and they listed out the measurements with detail. I agreed that the major strength is that it is a large and organised database with lots of useful data and valuable genetic information. The questions and concerns are as follows: The cohort contained a ranges of age groups, from <35 to 65+, but sizes of older age groups, e.g , and 65+ were few as noted in Table 4. In studying non-communicable diseases e.g. cancers, which is mainly a disease of the elderly, do you think the age of patient cohort is an issue? Self-reported questionnaires were used to obtain a number of data items, there could be recall bias and response bias. What were the ethnic group distribution of the cohort? Did the cohort recruited only include Korean? For blood pressure, recently the American Heart Association has recommended changing the starting point for high blood pressure to be systolic blood pressure (BP) of 130 mmhg and a diastolic BP above 80 mmhg. In your study, in page 18 line 21, it seems you defined hypertension as above 140/90 mmhg. Will the author consider changing the definition of hypertension in the study? This point also applies to a lot of measurements in this study because the volunteer from 2003 to 2014 were included, and as the study continues, it is expected that definitions of some medical conditions will get updated during your study. As for the cancer incidence in Table 6, do you have the data of different cancer histology subtypes? e.g. Lung: adenocarcinoma, squamous cell carcinoma, small cell carcinoma, etc. Sudesh Raj Sharma Massey University New Zealand 23-Nov-2017 Revision Summary: - I acknowledge the efforts of the authors to prepare such an excellent manuscript. It was clear and well-structured. I did not find any major issues with the article and recommend that the article is suitable for publication. - Due to my limitations relating to clinical knowledge, I am not sufficiently qualified to assess the clinical test mentioned in the manuscript. Some minor revision - The link in line 12 (Page 10) is not working - The correct way to measure Waist Circumference is to 2

3 measure midway between the top of the hip bone and below the rib cage. Do you think your measurement technique might have affected your results in anyway? - The reviewer also provided a marked copy with additional comments. Please contact the publisher for full details. VERSION 1 AUTHOR RESPONSE We deeply thank to the reviewer for the valuable comments and suggestions. We answered the questions and suggestions of reviewer, which we address below. Reviewer: 1 1. Participants were included when they received a health checks up between 2003 and In this period, individuals were recruited with an average age of 45.5 years. Only 32,6 completed the 2nd follow up assessment. Maybe more detail should be given on the eligibility and in diagnostic criteria. It should be mentioned why participants did not have repeated variables measured [Response] As readers may feel confused in our writing, we again described our follow-up system in detail [Line ] and revised the Figure 1. In this section, we d like to explain our two follow-up systems including passive and active follow-up system of outcome variables. When we describe our active follow-up system, we d like to show the Korea health check-up system first and then we d like to describe our active follow-up system including repeated measurement of risk factors and active repeated assessment of outcome variables. The detailed information was as below [Line in the text]. In our country, the health checkup system are composed of two systems in which the NHIC is partially paid for participants basic health examination fee once every two years or in which the health examination participants pays all expenses for their health examination fees. The latter program includes more precise health examination testing that individuals want, and all costs must be paid by individuals. We are under the latter system and the health examination participants have to pay for all the health checkups under their voluntary visit. The participants taken health examination at our center may do repeated measurement at different centers without revisit to our center under the partial coverage of the NHIC. Although the health check-up services conducted in our center includes a variety of tests that the subject wants, including the basic health check-up program conducted by the NHIC, we have to make a lot of effort in order for the subject to return to our center because the former health check-up system covered by the NHIC is free-paid for the basic health check-up program. We are using a number of ways to encourage our return to our center. To encourage participants, we provide reminder calls every year and send health care information about people's next health check-up date via phone and letter. So the eligibility for follow-up assessment depends on the participants voluntary re-visit need and self-payment availability. Under this active follow-up system, we did a repeated measurement of risk factors and an assessment of outcome variables for 46,484 individuals (50.9%) of total 91,336 cohort members participated in the baseline health check-up once within 4 years from the first visit and 74,304 individuals (83.5%) were actively followed up once within 10 years from the first visit. Moreover, 50,049 participants (54.8%) completed the two or three repeated measurements (Fig 1). The median follow-up was 4.04 years (interquartile range [IQR] ). [Response] We also described this at the STRENTHS AND LIMITATIONS section [Line and in the text]. The detailed information was as below. [Line ] First, our study participants are composed of individuals who voluntarily visited our center and their data collection for the repeated measurements rely on participants' self-paid. Our system of enrolling cohort members includes a potential selection bias. 3

4 [Line ] Third, our active follow-up rates in assessing outcome variables and repeatedly measuring risk factors is not high, leading to selection bias. 2. It is noteworthy that there was a male predominance in this young cohort of highly educated participants. Men had a very high percentage of current smokers and it was very low in women. This percentages were also different for gender in drinking. There were also marked gender differences in systolic and diastolic blood pressure as well as in blood glucose. In this young and lean cohort it was remarkable in the table of abnormal findings to observe that 6% had CKD, 66% abdominal obesity and 29% fatty liver. It is also surprising to see a 54 % H. Pylori infection, a 36% atrophic gastritis and 37% of colonic adenomas. [Response] Thank you for the reviewer s comment. The findings the reviewer pointed out may be the representative characteristics of our cohort, which is composed of the Korea ethnicity only. -Consistent to our results, the population prevalence of nonalcoholic fatty liver disease (NAFLD) in Korea is more than 25%, like many Western countries.[1-4] (Age- and gender- specific prevalence of NAFLD in our healthcare center was previously reported.3) Although, subjects in this cohort are relatively young and lean, lean NAFLD (or nonobese NAFLD) is more common in Asians including Koreans compared to those in Western countries. Asians generally have a higher percentage of visceral fat and body fat compared with people of other races of the same age, sex, and even BMI.[5-7] As, visceral obesity is major risk factor for lean NAFLD, this might have affected higher prevalence of NAFLD in this cohort composed of relatively lean subjects.[5] In addition, the term fatty liver in this paper includes not only NAFLD but also alcoholic fatty liver. The proportion of heavy alcohol drinking in this cohort is as high as 37.5% in men, and this might have affected on the high prevalence of fatty liver in this cohort. -In a nationwide multicenter study in Korea performed by our center for the prevalence of H.pylori infection, the seropositive rate of H.pylori was 59.6%.[8] This rate is quite consistent with our study result and consequently explain the relatively high percent of atrophic gastritis. -In our study the colonic adenoma were 37%. In our center, there is report with a title of Prevalence and risk of colorectal adenoma in asymptomatic Koreans aged years undergoing screening colonoscopy.[9] In that study the prevalence of adenoma was 22.2% in years age group and 32.8% in the years age group. This finding is quite consistent with our study. In the paper, we concluded that the prevalence of adenoma in subjects aged years was higher than in previous studies. Male sex and current smoking habits along were the risk factors. -Reference [1] Fan JG, Kim SU, Wong VW. New trends on obesity and NAFLD in Asia. J Hepatol 2017;67: [2] Bae JC, Cho YK, Lee WY, et al. Impact of nonalcoholic fatty liver disease on insulin resistance in relation to HbA1c levels in nondiabetic subjects. Am J Gastroenterol 2010;105: [3] Choi SY, Kim D, Kim HJ, et al. The relation between non-alcoholic fatty liver disease and the risk of coronary heart disease in Koreans. Am J Gastroenterol 2009;104: [4] Chang Y, Jung HS, Yun KE, Cho J, Cho YK, Ryu S. Cohort study of non-alcoholic fatty liver disease, NAFLD fibrosis score, and the risk of incident diabetes in a Korean population. Am J Gastroenterol 2013;108: [5] Kim D, Kim WR. Nonobese Fatty Liver Disease. Clin Gastroenterol Hepatol 2017;15: [6] Alberti KG, Zimmet P, Shaw J, Group IDFETFC. The metabolic syndrome--a new worldwide definition. Lancet 2005;366: [7] Nazare JA, Smith JD, Borel AL, et al. Ethnic influences on the relations between abdominal subcutaneous and visceral adiposity, liver fat, and cardiometabolic risk profile: the International Study of Prediction of Intra-Abdominal Adiposity and Its Relationship With Cardiometabolic Risk/Intra- Abdominal Adiposity. Am J Clin Nutr 2012;96: [8] Lim SH, Kwon JW, Kim N et al. Prevalence and risk factors of Helicobacter pylori infection in Korea: nationwide multicenter study over 13 years. BMC Gastroenterol. 2013;24;13:104. doi: / X

5 [9] Chung SJ1, Kim YS, Yang SY et al. Prevalence and risk of colorectal adenoma in asymptomatic Koreans aged years undergoing screening colonoscopy. J Gastroenterol Hepatol. 2010; 25(3): In table 4 showing baseline characteristics and comorbidities, the authors list past hypertension and past diabetes and maybe should be changed to just hypertension and diabetes. [Response] Thank your precise comment. We ve changed the term to Hypertension and Diabetes [24th, 25th raw in the table 4]. 4. In the first and second paragraphs of page 30 the authors make recommendations on the screening for GI Malignancies. Maybe the authors should elaborate more in their interpretation, generalizability and conclusions. [Response] As the reviewer recommended, we added some more detail interpretation for our recommendation based on the several guidelines and study results in the manuscript [Line ] as bellows; [Line ] For subjects with 1-2 adenomas less than 10 mm in colonoscopy for colon cancer screening, they were classified as a low-risk group having low 5-year colon adenoma incidence rates and were recommended to take further colonoscopic screening test after 5 year. For subjects with advanced adenoma, multiple adenoma 3, or larger adenoma sized 10mm in baseline colonoscopy for colon cancer screening, they were classified as a high-risk group having higher incidence rates of advanced adenoma or higher recurrence rates of adenoma and were recommended to take colonoscopic surveillance at 3-year after initial polypectomy [12]. This strategy of colonoscopic surveillance has been reflected in the guidelines for colonoscopic surveillance [34]. [Line ] For gastric cancer screening, the effect of screening endoscopy remains controversial, but population-based screening has been undertaken in Korea and Japan. For participants with intestinal metaplasia in gastroendoscopy, we also classified them into high-risk group and recommended an annual endoscopic screening, based on study results that people with strong risk factors such as male and an older age can quickly find early-staged endoscopically-treatable gastric cancer by taking annual gastroendoscopic screening [11] ============================================================================= ====================== We deeply thank to the reviewer for the valuable comments and suggestions. We answered the questions and suggestions of reviewer, which we address below. Reviewer: 2 1. The cohort contained a ranges of age groups, from <35 to 65+, but sizes of older age groups, e.g , and 65+ were few as noted in Table 4. In studying non-communicable diseases e.g. cancers, which is mainly a disease of the elderly, do you think the age of patient cohort is an issue? [Response] We totally agree with your opinion. [Line ] It has only been about 10~15 years since the comprehensive health check-up program is actively performed in Korea and in the past decade, a large proportion of peoples participated in the check-up due the support from the affiliated company s welfare policy. This might result in the relatively young population enrollment in our study population. But since the health check-up program is getting more general, more elderly population is taking the health check-up. New enrollment is another future target study of our center. On the positive side, since the enrolled cohort population is relatively young, it is possible studying the preclinical disease stage and its final long term outcomes. This might be the characteristics of our cohort that we could design a lot of prediction model for the non-communicable disease by using a data from the preclinical stages. 5

6 2. Self-reported questionnaires were used to obtain a number of data items, there could be recall bias and response bias. [Response] It is a correct comment. We additionally described the limitation of self-reported questionnaires at the STRENGTHS AND LIMITATIONS section. The detailed information was as [Line ] A self-record questionnaires are used to obtain the information of risk factors and past disease histories before the next visit of study participants to our center and this procedure leads to recall bias and response bias. 3. What were the ethnic group distribution of the cohort? Did the cohort recruited only include Korean? [Response] This cohort included only Koreans. We changed the sentence in the manuscript as below [Line 107] We changed the sentence of The H-PEACE study collected data from 91,336 individuals to that of The H-PEACE study collected data from 91,336 Koreans. 4. For blood pressure, recently the American Heart Association has recommended changing the starting point for high blood pressure to be systolic blood pressure (BP) of 130 mmhg and a diastolic BP above 80 mmhg. In your study, in page 18 line 21, it seems you defined hypertension as above 140/90 mmhg. Will the author consider changing the definition of hypertension in the study? This point also applies to a lot of measurements in this study because the volunteer from 2003 to 2014 were included, and as the study continues, it is expected that definitions of some medical conditions will get updated during your study. [Response] We ve encountered the report of 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults opened at November 2017 recommending a new definition for hypertension. Our manuscript was written before the declaration was done. Since information of the enrolled population have been collected before November, 2017, we think it is appropriate to define hypertension according to the guideline when the enrollment was done. We have the raw data of blood pressure and the definition can be applied to the study population. But for those who had already been under anti-hypertensive medication at the point of enrollment in our cohort, had followed the previous guidelines for indication of treatment. This makes it difficult to re-classify these patients according to the new guidelines. Since your comments are a really good point, we will give a good attention to the changes of guidelines in various diseases and apply to our study population during the follow ups. And we will give a big attention to clearly declare the definition of disease, we are applying according to the period of terms. This is briefly described in the limitation section [Line in the text]. 5. As for the cancer incidence in Table 6, do you have the data of different cancer histology subtypes? e.g. Lung: adenocarcinoma, squamous cell carcinoma, small cell carcinoma, etc. [Response] In our cancer registry, the data of different cancer histology subtypes were not available. [Line ] There is no information of cancer histology subtypes for cohort participants ascertained as new cancer development. Individual cancer development in study participants was ascertained with data linkage to the nationwide cancer registry data. Korea cancer registry data includes ICD10 code, T-code (Topography) and M-code (Morphology in primary cancer sites). For example, in stomach cancer, 51.3% were adenocarcinomas, 22.1% were adenocarcinomas, and gastric cancer and histologic NOS were only 4.9% [42]. Future studies will attempt to merge the individual data associations of M-code and T-code. ============================================================================= ==== We deeply thank to the reviewer for the valuable comments and suggestions. We answered the questions and suggestions of reviewer, which we address below. 6

7 Reviewer: 3 1) The link in line 12 (Page 10) is not working [Response] We ve corrected our website, as [Line 58, 165, 472, 504] 2) The correct way to measure Waist Circumference is to measure midway between the top of the hip bone and below the rib cage. Do you think your measurement technique might have affected your results in anyway? [Response] We totally agree with your concern. [Line ] in our cohort, we used to measure waist circumference as the definition of the smallest area around the belly button below the rib cage and above the hip bone, although the ideal way is to measure the midway between the top of the hip bone and below the rib cage. The former is usually called the natural waist, and we used this former definition for practical reason to reduce intraindividual measurement bias in waist circumference. In health check-up at our center, a lot of participants (about health examinees) everyday visit to center for health examination and many health technicians and nurses take measurement of waist and hip circumferences. For each participants, there is a very little chance to take measurement of waist at the time of re-visit by same nurses met at the time of cohort enrollment. The intra-individual measurement bias in measuring waist circumference may be problematic at our center and thus we thought that it was necessary to use the most practical and easy waist measurement together with repeated nurse training. We did a small pilot study for measuring waist by the two methods for 10 men and 16 women. The ideal measuring method (midway measurement) has a limit in consuming time and effort due to difficulty in method itself. Despite of short time and smaller effort in measuring waist, the natural waist measuring method showed an excellent agreement (intraclass correlation coefficients = 99% in men and 93% in women) with ideal method and there were no shift between obesity groups classified by each method. ============================================================================= ==== ###We made some changes in the format of tables to make the legibility more comfortable for the reviewers. There are no changes in context. REVIEWER VERSION 2 REVIEW alfredo lozada Austral University or Universidad Austral, Argentina 08-Jan-2018 This manuscript is a Cohort Profile that was collected in a retrospective manner.this cohort was developed to investigate the clinical effectiveness of diagnostic tests, examinations and biomarkers, performed during health check ups, and the later development of Non CommunIcable Diseases (NCD). These NCD included malignancies, cardiac and metabolic diseases. Participants were included when they received a health check up between 2003 and 2014.During this period, individuals were recruited with an average age of 45.5 years. Only 50,9 % completed a 2nd follow up assessment. Maybe it should be added in line 118, under the Title: Repeated Measurements, a text that was present in the previous manuscript that stated: Annual health exams are mandadatory for all workers under the Industrial and Safety Law in Korea. Also on line 132 it is not well understood the phrase.. may do repeated measurements under partial coverage of the NIHC It is noteworthy that there was a male predominance in this young cohort of highly educated participants. Men had a very high 7

8 REVIEWER percentage of current smokers and was very low in women. This percentages were also very different for gender in drinking. There were also marked gender differences in BMI, sistolic and diastolic blood pressure as well as in blood glucose and tryglicerides. Maybe the term Morbidity Rate should be included in the title of table 5. Also in this table, it is confusing on lines 18 and 19 Severe and Hypothyroidism, and in line 44 and 45, gender differences should be stated for osteopenia and osteoporosis. In this young and lean cohort it was remarkable to observe in abnormal findings, that 6% had CKD, 66% abdominal obesity and 29% fatty liver. It is also surprising to see a 54 % H. Pylori infection, a 36% atrophic gastritis and 37% of colonic adenomas. Maybee the authors should elaborate more in their interpretation of this results. In the paragraph of line 390 to 395 the authors make the statement that NAFLD and Metabolic Syndrome are independent Risk factors for silent brain infarction. Maybee the authors should elaborate more in this interpretation and their results. On line 404 a should be deleted. On line 432 it is not clear because adenocarcinomas u mentioned twice. On line 452 it should be stated people and not peoples. On line 458 it should be stated models. Shing Fung Lee Department of Clinical Oncology, Tuen Mun Hospital, Hong Kong 09-Jan-2018 This article is in general well revised. Queries are answered to my satisfaction. VERSION 2 AUTHOR RESPONSE We deeply thank the reviewer for the valuable comments and suggestions. We answered the questions and suggestions of the reviewer, which we addressed below. Reviewer: 1 1. Maybe it should be added in line 118, under the Title: Repeated Measurements, a text that was present in the previous manuscript that stated: Annual health exams are mandadatory for all workers under the Industrial and Safety Law in Korea. [Response] Thank you for the nice recommendation. We ve added this phrase at the recommended location. [Line ] Annual health exams are mandatory for all workers under the Industrial and Safety Law in Korea. 2. On line 132 it is not well understood the phrase.. may do repeated measurements under partial coverage of the NIHC [Response] We have made some changes to make the meaning clear as described below. [Line 126] National Health Insurance Corporation (NHIC) pay [Line ] The participants who have undergone health examinations at our center may receive health checkups afterwards at different centers that are under partial coverage of the NHIC without revisiting our center. 3. Maybe the term Morbidity Rate should be included in the title of table 5. [Response] We changed the title as the reviewer recommended. [Table 5.] Morbidity rates after active follow-up with specific measurements in the Health and Prevention Enhancement (H-PEACE) study 4. Also in this table, it is confusing on lines 18 and 19 Severe and Hypothyroidism, [Table 5] Sorry for our carelessness. We added a line to make this clearer. 8

9 5. line 44 and 45, gender differences should be stated for osteopenia and osteoporosis. [Table 5] We added the numbers for male and females, respectively. 6. In this young and lean cohort it was remarkable to observe in abnormal findings, that 6% had CKD, 66% abdominal obesity and 29% fatty liver. It is also surprising to see a 54 % H. Pylori infection, a 36% atrophic gastritis and 37% of colonic adenomas. Maybee the authors should elaborate more in their interpretation of this results. [Response] Thank you for this comment. The findings the reviewer pointed out may be the representative characteristics of our cohort, which is composed of only the Korean ethnicity. As the reviewer recommend, we elaborated in our manuscript as described below. [Line ] Consistent with our results, the population prevalence of NAFLD in Korea is greater than 25%, like that in many Western countries.[19, 44, 45,47] Although subjects in this cohort are relatively young and lean, lean NAFLD (or non-obese NAFLD) is more common in Asians (including Koreans) than in Western populations. Asians generally have a higher percentage of visceral fat and body fat than people of other races of the same age, sex, and even BMI.[48-50] As visceral obesity is a major risk factor for lean NAFLD, this might have affected the higher prevalence of NAFLD in this cohort, comprised of relatively lean subjects.[48] In addition, the term fatty liver in this manuscript includes not only NAFLD but also alcoholic fatty liver disease. The proportion of subjects who drink alcohol heavily in this cohort is as high as 37.5% in men, which might affect the high prevalence of fatty livers that we observe. [Line ] In a nationwide multicenter study in Korea performed by our center for the prevalence of H. pylori infection, the seropositive rate of H. pylori was 59.6%.[51] This rate is quite consistent with our study result and consequently explains the relatively high percentage of atrophic gastritis that we observed. [Line ] In our study, the colonic adenoma incidence rate was 37%. Our center performed a previous study on the prevalence and risks of colorectal adenoma in asymptomatic Koreans aged years undergoing screening colonoscopies.[52] In that study, the prevalence of adenoma was 22.2% in the year age group and 32.8% in the year age group. This finding is quite consistent with that in our cohort. In the paper, we concluded that the prevalence of adenoma in subjects aged years was higher than that in previous studies. Male sex and current smoking habits were among the risk factors. 7. In the paragraph of line 390 to 395 the authors make the statement that NAFLD and Metabolic Syndrome are independent Risk factors for silent brain infarction. Maybee the authors should elaborate more in this interpretation and their results. [Response] Thank you for the comment. As the reviewer recommend, we have elaborated on this in our manuscript as described below. [Line ] We reported that metabolic syndrome was an independent risk factor of silent brain infarction using brain MRI data (OR, 2.18; 95% CI, , p=0.001).[40] The prevalence of metabolic syndrome and its components were higher in subjects with silent brain infarction than in those without. This is the first study to demonstrate an association between metabolic syndrome and silent brain infarction. This finding might help identifying healthy people at increased risk of developing silent brain infarction. [Line ] We identified a relationship between NAFLD and the risks of coronary heart disease and arterial stiffness based on coronary CT and the cardio-ankle vascular index (CAVI), respectively.[16-19] To elucidate the relationship between NAFLD and the risk of coronary heart disease, we used the coronary artery calcification score as measured by coronary CT. This measurement showed that patients with NAFLD are at increased risk for coronary atherosclerosis (OR, 1.28, 95% CI, , p=0.023).[17] In the study on the association between NAFLD and arterial stiffness, we used CAVI, a new measurement of arterial stiffness. In an age-, sex-, and BMIadjusted model, NAFLD was associated with a 42% increase in the risk for arterial stiffness, and this increased according to the severity of NAFLD[16]. 8. On line 404 a should be deleted. 9

10 [434] Second, a self-record questionnaires are used to obtain the information of risk factors and past disease histories -recorded questionnaires are used to obtain information on risk factors and past disease histories before the next visit by the study participants to our center 9. On line 432 it is not clear because adenocarcinomas u mentioned twice. [Table 5] Sorry for our carelessness. We have edited the errors as shown below. [Line ] For example, among the stomach cancer types, 51.3% were adenocarcinomas, 22.1% were tubular adenocarcinomas, and only 4.9% were histologic NOS [42]. 10. On line 452 it should be stated people and not peoples. 11. On line 458 it should be stated models. ================================================== 1. The website mentioned that the reference was incorrect. We have changed this as shown below. [Reference 42] Korea Ministry of Health and Welfare, Korea National Cancer Center Annual Report of the Korea Central Cancer Registry. Korea National Cancer Center, [Available at 2. The reference number is updated due to additional references are added in response to the review s opinion. 3. We change the authorship of H-PEACE study investigators from the author list to the acknowledgements list. REVIEWER VERSION 3 REVIEW Alfredo Lozada Austral University, Argentina 04-Mar-2018 This manuscript is a Cohort Profile that was collected in a retrospective manner.this cohort was developed to investigate the clinical effectiveness of diagnostic tests, examinations and biomarkers, performed during health check ups, and the later development of Non CommunIcable Diseases (NCD). These NCD included malignancies, cardiac and metabolic diseases. Participants were included when they received a health check up between 2003 and 2014.During this period, individuals were recruited with an average age of 45.5 years. Only 50,9 % completed a 2nd follow up assessment. Maybee it should be changed in page 6 and line 137 to encourage their return to our. After the changes made I feel it is suitable for publication. 10