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. TITLE (PROVISIONAL) AUTHORS REVIEWER REVIEW RETURNED GENERAL COMMENTS ARTICLE DETAILS Association of perceived tinnitus with duration of hormone replacement therapy in Korean postmenopausal women: a crosssectional study Lee, Seong-Su; Han, Kyungdo; Joo, Young-Hoon VERSION 1 - REVIEW Takehiro Michikawa Centre for Health and Environmental Risk Research, National Institute for Environmental Studies, Japan 04-Sep-2016 This paper is presented to examine the associations between hormone replacement therapy and tinnitus in postmenopausal women (19-64 years). This study has some strengths, such as nationally representative sample and information on possible confounding factors. I have several comments that should be considered by the authors. Comments Methods 1. Study participants 1) page 5, line 31 there were 25,534 KNHANES participants with tinnitus aged 19 years or older I think that there were 25,534 KNHANES participants aged 19 years or older 2) page 5, line 35. Why did the authors exclude older women aged 65 years or older? 2. Tinnitus survey Did the authors use their original question regarding tinnitus? Other risk factors 3. Page 7, lines 46-51 I think that hypertension and diabetes should include the use of antihypertensive and diabetes agents. 4. Page 8, line 35 Several factors in Table 1 was not include in the section of other risk factors. Statistical analysis 5. This reviewer suggests that the authors reconsider the confounding factors in terms of the association between HRT and tinnitus. The authors should explain the reason why they include
each factor in multivariate model. 1) In Table1, socioeconomic status including educational level and income was associated with tinnitus. However, the authors did not include these factors in the model. Why? 2) In the previous report from the KNHNES, tinnitus was associated with gender, smoking, stress, sleep, hearing loss, hyperlipidemia, osteoarthritis, rheumatoid arthritis, asthma, depression, and thyroid disease history (Kim HJ et al. PLoS One. 2015;10(5):e0127578). Please, show the results after adjusting for sleep, hyperlipidemia, osteoarthritis, rheumatoid arthritis, asthma, depression, and thyroid disease history. 3) They stated HRT might be associated with hearing loss (page 4, line 31). Hearing loss is a risk factor of tinnitus. Therefore, I think that the association between HRT and tinnitus is partly explained by the association between HRT and hearing loss. Therefore, they should explain the meaning of the adjustment for hearing loss. 4) In Table 2, oophorectomy was weakly associated with the severity of tinnitus. This result is reasonable, because women who had oophorectomy often got HRT. However, is oophorectomy a confounding factor of the association between HRT and tinnitus? 6. The description of multivariate model was not consistent with th footnote of Table 3. Results 7. Please, include the descriptive information on HRT and other reproductive factors. 8. In Table 2, the authors presented the crude association between HRT and the severity of tinnitus. In Figure 1, the crude association between duration of HRT and tinnitus is shown. However, they only showed the multi-adjusted ORs of the association between HRT (yes or no) and tinnitus (yes, no) in Table 3. This reviewer would like to know the results of the association between HRT and the severity of tinnitus, and those of the association between duration of HRT and tinnitus in multivariate model. 9. In Table 2, crude association of breast feeding (p = 0.03) and oophorectomy (p = 0.05) with the severity of tinnitus was suggested. Did this association persist even after adjusting for several confounding factors? The authors may clarify this association in discussion. Additional comments Title 10. To help readers understand this article, this reviewer recommends that the authors include study design (cross-sectional study) in Title. Abstract 11. Please, include information on age of target participants in Abstract. Tables
REVIEWER REVIEW RETURNED GENERAL COMMENTS 12. 1) In Table 1 (footnote), information regarding men can delete. 2) In Table 2, describe the unit of the values, please. breast feeding (year?), pregnancy (number?) 3) Title of Table 3 is strange. Bamini Gopinath University of Sydney, Australia 06-Sep-2016 This cross-sectional study examined the association between HRT and prevalence of tinnitus in Korea. They found that longer duration of HRT was associated with development of tinnitus in Korean postmenopausal women. 1) Please specify age range on participants in the Abstract 2) In the Introduction (line 40), it is probably more correct to state that it was self-reported tinnitus rather than otolaryngologistdiagnosed tinnitus 3) Line 31, pg 6 it is stated that there were 25,534 KNHANES participants with tinnitus that is a high number of people with tinnitus do the authors mean to say that this number had tinnitus data? This seems more likely. 4) A comparison of characteristics between participants and nonparticipants is useful as this could indicate whether any selection bias occurred, please include this data for the reader. 5) What about the presence of other chronic conditions e.g. cardiovascular disease, visual impairment, cognitive impairment? These could be potential confounders. 6) Also what about other vestibular conditions such as vertigo, these are likely to influence observed associations. If these and other data were not available there could be residual confounding which needs to be mentioned in the Discussion. 7) Not clear why they used a cut point of >40 db HL for defining hearing impairment, generally >25 db HL defines any level of hearing impairment whereas the cut point used in this study defines severe hearing loss. Could the authors clarify their use of this cut point? 8) Table 3 heading needs to clearly state that it is presenting data on associations between HRT and tinnitus prevalence, currently the heading seems incomplete. 9) Education level and income were significantly different between people with and without tinnitus (Table 1), why were these factors not adjusted for in the end? 10) Model 5 in Table 3 seemed to have adjusted for depression and anxiety, however, these factors are not mentioned in the Methods, please clarify? VERSION 1 AUTHOR RESPONSE Reviewer: 1 Reviewer Name: Takehiro Michikawa Institution and Country: Centre for Health and Environmental Risk Research, National Institute for Environmental Studies, Japan Please state any competing interests or state None declared : Non declared Please leave your comments for the authors below This paper is presented to examine the associations between hormone replacement therapy and
tinnitus in postmenopausal women (19-64 years). This study has some strengths, such as nationally representative sample and information on possible confounding factors. I have several comments that should be considered by the authors. Comments Methods 1. Study participants 1) page 5, line 31 there were 25,534 KNHANES participants with tinnitus aged 19 years or older I think that there were 25,534 KNHANES participants aged 19 years or older Response: According to the reviewer s comment, we changed this sentence. 2) page 5, line 35. Why did the authors exclude older women aged 65 years or older? Response: In Korea only a small proportion of menopausal women aged 65 years or older currently take hormone replacement therapy. So we excluded the older women. 2. Tinnitus survey Did the authors use their original question regarding tinnitus? Response: I really appreciate your comments. In this study, there was no objective assessment of tinnitus. We do not know the relative severity or grade of the tinnitus in the absence of objective testing and more detailed questions. An assessment of the true prevalence of tinnitus in the general population must be distinguished from the prevalence of tinnitus in clinical populations or those seeking care. Furthermore, distinctions need to be made between a reported self-perception of a tinnitus versus more rigorous forms of tinnitus assessment such as formal tinnitus test batteries (e.g., quantitative and qualitative tinnitus assessments). Although from one standpoint, one might consider the prevalence of tinnitus as measured by objective tinnitus testing to be a more scientific estimation of the prevalence of tinnitus, it is likely that patient-centered perceptions of tinnitus are more likely to be the drivers of healthcare seeking behaviors for tinnitus. In other words, there is significant importance attached to determining the population-based self-reported prevalence of tinnitus. Other risk factors 3. Page 7, lines 46-51 I think that hypertension and diabetes should include the use of anti-hypertensive and diabetes agents. Response: We have a mistake. According to the reviewer s comment, we changed this sentence. Abdominal obesity was defined as a WC > 80 cm (all-female cohort), hypertension was defined as systolic blood pressure 130mmHg and/or diastolic blood pressure 85 mm Hg or being on an antihypertensive drug treatment for patients with a history of hypertension. Diabetes mellitus was defined as a fasting plasma glucose level 126 mg/dl or being on medication use for elevated glucose. 4. Page 8, line 35 Several factors in Table 1 was not include in the section of other risk factors. Response: According to the reviewer s comment, we have added these sentences. Those who answered that they had been diagnosed by a medical professional with stroke, myocardial infarction, and angina were defined as having a Cardiovascular disease. Mental health surveys based on the same questions as those used in the KNHANES were provided to all participants. Participants reported their level of stress as none, mild, moderate, or severe. Depression was screened using the Korean version of the World Health Organization (WHO) Composite International Diagnostic Interview-Short Form (CIDI-SF), which was validated as a cost-effective screening instrument that could be easily integrated into health surveys.18 The WHO CIDI-SF
includes questions such as In your lifetime, have you ever had 2 weeks or more when nearly every day you felt sad, blue, or depressed? and Have there ever been 2 weeks or longer when you lost interest in most things such as work or hobbies or things you usually like to do for fun? To assess depression, subjects answered yes or no to a question of whether they had experienced a depressed mood for 2 or more continuous weeks during the previous year. We have added this reference. 18. Gigantesco A, Morosini P. Development, reliability and factor analysis of a self administered questionnaire which originates from the World Health Organization's Composite International Diagnostic Interview Short Form (CIDI-SF) for assessing mental disorders. Clin Pract Epidemiol Ment Health 2008;4:8. Statistical analysis 5. This reviewer suggests that the authors reconsider the confounding factors in terms of the association between HRT and tinnitus. The authors should explain the reason why they include each factor in multivariate model. Response: According to the reviewer s comment, we have added these sentences. Multiple logistic regression analyses were performed to estimate the magnitude of the association of tinnitus with HRT, and five statistical models based on the characteristics of the variables were used. Model 1 included age. Socioeconomic and lifestyle-related characteristics including BMI, smoking status, alcohol intake and exercise habits were included based on the results from the univariate analysis (model 2). Model 3 adjusted for the variables in model 2 plus comorbidity such as diabetes mellitus, hypertension, education, income, stress level, breast feeding, bilateral oophorectomy and hysterectomy. Finally, model 4, 5 adjusted for the variables in model 3 plus hearing impairment, mental health, and noise exposure. 1) In Table1, socioeconomic status including educational level and income was associated with tinnitus. However, the authors did not include these factors in the model. Why? Response: According to the reviewer s comment, we have added educational level and income in the logistic regression models. 2) In the previous report from the KNHNES, tinnitus was associated with gender, smoking, stress, sleep, hearing loss, hyperlipidemia, osteoarthritis, rheumatoid arthritis, asthma, depression, and thyroid disease history (Kim HJ et al. PLoS One. 2015;10(5):e0127578). Please, show the results after adjusting for sleep, hyperlipidemia, osteoarthritis, rheumatoid arthritis, asthma, depression, and thyroid disease history. Response: According to the reviewer s comment, we have added asthma and sleep duration. 3) They stated HRT might be associated with hearing loss (page 4, line 31). Hearing loss is a risk factor of tinnitus. Therefore, I think that the association between HRT and tinnitus is partly explained by the association between HRT and hearing loss. Therefore, they should explain the meaning of the adjustment for hearing loss. Response: According to the reviewer s comment, we have added these sentences. The present study demonstrated that the prevalence of tinnitus among a group of HRT was significantly higher than among those without HRT after adjusting for sociodemographic factors and comorbidities including hearing impairment. There is a correlation that women with tinnitus have poorer hearing than those without tinnitus. This could indicate that these postmenopausal women without HRT, with little or no remaining endogenous estrogen production, have a higher rate of deterioration of hearing function after than before menopause, which is consistent with previous suggestions that estrogen might have a protective role in hearing function11,12. Subjects with HRT could have poorer hearing than those without HRT. However, when adjusting for hearing impairment and noise exposure in model 5, subjects with HRT still had a greater OR for tinnitus as compared to
the non-hrt group. 4) In Table 2, oophorectomy was weakly associated with the severity of tinnitus. This result is reasonable, because women who had oophorectomy often got HRT. However, is oophorectomy a confounding factor of the association between HRT and tinnitus? Response: Participants with oophorectomy demonstrated a higher proportion of annoying (22.5%) and not annoying (20.7) tinnitus (p=0.0531). The use of HRT in postmenopausal women was significantly associated with tinnitus after adjustment of the bilateral oophorectomy in the logistic regression models. 6. The description of multivariate model was not consistent with the footnote of Table 3. Response: According to the reviewer s comment, we have changed the descriptions of multivariate model in the statistical analysis section. Results 7. Please, include the descriptive information on HRT and other reproductive factors. Response: According to the reviewer s comment, we have added this sentence. Endogenous hormone-related risk factors included age at menarche, age at menopause, total reproductive years, number of pregnancies, age at first birth, and duration of breastfeeding. Data regarding reproductive factors were collected by asking the participants to recall the age at first and last delivery and the age at menopause, gravidity, parity and lactation. Exogenous hormone-related risk factors included duration of oral contraceptive use and starting age and duration of HRT. 8. In Table 2, the authors presented the crude association between HRT and the severity of tinnitus. In Figure 1, the crude association between duration of HRT and tinnitus is shown. However, they only showed the multi-adjusted ORs of the association between HRT (yes or no) and tinnitus (yes, no) in Table 3. This reviewer would like to know the results of the association between HRT and the severity of tinnitus, and those of the association between duration of HRT and tinnitus in multivariate model. Response: According to the reviewer s comment, we have added the association between duration of HRT and tinnitus in Table 3. Regarding duration of HRT, there was a significant trend between a longer period of HRT and an increased risk of tinnitus. 9. In Table 2, crude association of breast feeding (p = 0.03) and oophorectomy (p = 0.05) with the severity of tinnitus was suggested. Did this association persist even after adjusting for several confounding factors? The authors may clarify this association in discussion Response: The use of HRT was significantly associated with tinnitus after adjustment of the breast feeding and bilateral oophorectomy in model 3 of regression analyses. Additional comments Title 10. To help readers understand this article, this reviewer recommends that the authors include study design (cross-sectional study) in Title. Response: According to the reviewer s comment, we changed our title. Abstract 11. Please, include information on age of target participants in Abstract. Response: According to the reviewer s comment, we added the information on age in the abstract. Tables
12. 1) In Table 1 (footnote), information regarding men can delete. Response: According to the reviewer s comment, we erased the information regarding men in footnote. 2) In Table 2, describe the unit of the values, please. breast feeding (year?), pregnancy (number?) Response: In this study we considered the history of breast feeding or pregnancy instead of the duration of breast feeding or the number of pregnancy. 3) Title of Table 3 is strange. Response: According to the reviewer s comment, we changed the title of Table 3. Reviewer: 2 Reviewer Name: Bamini Gopinath Institution and Country: University of Sydney, Australia Please state any competing interests or state None declared : None declared Please leave your comments for the authors below This cross-sectional study examined the association between HRT and prevalence of tinnitus in Korea. They found that longer duration of HRT was associated with development of tinnitus in Korean postmenopausal women. 1) Please specify age range on participants in the Abstract Response: According to the reviewer s comment, we added the information on age in the abstract. 2) In the Introduction (line 40), it is probably more correct to state that it was self-reported tinnitus rather than otolaryngologist-diagnosed tinnitus Response: According to the reviewer s comment, we changed the sentence. 3) Line 31, pg 6 it is stated that there were 25,534 KNHANES participants with tinnitus that is a high number of people with tinnitus do the authors mean to say that this number had tinnitus data? This seems more likely. Response : A total of 2,736 women were eligible for the study among 25,534 KNHANES participants aged 19 years or older. Among the 2,736 participants older than 19 years of age, 607 had experienced tinnitus in the previous 12 months (prevalence: 22.2%). 4) A comparison of characteristics between participants and non-participants is useful as this could indicate whether any selection bias occurred, please include this data for the reader. Response : KNHANES is an ongoing cross-sectional survey of the ci vilian noninstitutionalized population of South Korea over a time span of 3 years. Each year, 10000 to 12000 individuals in 4600 households are selected from a panel to represent the Korean population by using the multistage cluster and stratified random sampling method that is based on the Na tional Census Data. The selected households are asked to participate in the survey, and the participation rate in the past several cycles has ranged from 79% to 84%. The first to third cycles (1998, 2001, and 2004) of the KNHANES and the three years of the fourth cycle collected basic health and nutrition data by conducting interviews and health exami nations. According to the reviewer s comment, we have added these sentences. 5) What about the presence of other chronic conditions e.g. cardiovascular disease, visual impairment, cognitive impairment? These could be potential confounders. Response: According to the reviewer s comment, we added the several factors such as cardiovascular disease, asthma, sleep duration.
6) Also what about other vestibular conditions such as vertigo, these are likely to influence observed associations. If these and other data were not available there could be residual confounding which needs to be mentioned in the Discussion. Response: Unfortunately we could not evaluate the vestibular function in this study. According to the reviewer s comment, we mentioned in the discussion. 7) Not clear why they used a cut point of >40 db HL for defining hearing impairment, generally >25 db HL defines any level of hearing impairment whereas the cut point used in this study defines severe hearing loss. Could the authors clarify their use of this cut point? Response : In this study the PTA of both ears at each frequency was obtained for frequency analysis. The degree of hearing loss was categorized as mild (26~40 db), moderate (41~55 db), moderately severe (56~70 db) and severe (71~90 db) based on the standard of the International Organization for Standardization (ISO). Disabling hearing loss refers to hearing loss greater than 40 db in the better hearing ear in adults and greater than 30 db in the better hearing ear in children from the definition of WHO. So we defined hearing loss at a threshold of a 40 decibel hearing level in the ear with better hearing. 8) Table 3 heading needs to clearly state that it is presenting data on associations between HRT and tinnitus prevalence, currently the heading seems incomplete. Response: According to the reviewer s comment, we changed the title of Table 3. 9) Education level and income were significantly different between people with and without tinnitus (Table 1), why were these factors not adjusted for in the end? Response: According to the reviewer s comment, we have added educational level and income in the Logistic regression models. 10) Model 5 in Table 3 seemed to have adjusted for depression and anxiety, however, these factors are not mentioned in the Methods, please clarify? Response: According to the reviewer s comment, we have added these sentences. Mental health surveys based on the same questions as those used in the KNHANES were provided to all participants. Participants reported their level of stress as none, mild, moderate, or severe. Depression was screened using the Korean version of the World Health Organization (WHO) Composite International Diagnostic Interview-Short Form (CIDI-SF), which was validated as a costeffective screening instrument that could be easily integrated into health surveys.14 The WHO CIDI- SF includes questions such as In your lifetime, have you ever had 2 weeks or more when nearly every day you felt sad, blue, or depressed? and Have there ever been 2 weeks or longer when you lost interest in most things such as work or hobbies or things you usually like to do for fun? To assess depression, subjects answered yes or no to a question of whether they had experienced a depressed mood for 2 or more continuous weeks during the previous year. I hope the revised manuscript will better meet the requirements of the BMJ open for publication. I thank you again for the constructive review by the reviewers.
REVIEWER REVIEW RETURNED GENERAL COMMENTS VERSION 2 REVIEW Takehiro Michikawa Centre for Health and Environmental Risk Research, Japan National Institute for Environmental Studies 03-Nov-2016 On the whole, the authors adequately responded to my comments. However, I think they can reconsider their responses to some questions. Number of previous comment 1. Study participants 2) page 5, line 35. Why did the authors exclude older women aged 65 years or older? Authors Response: In Korea only a small proportion of menopausal women aged 65 years or older currently take hormone replacement therapy. So we excluded the older women. I agree with your response. Please, include this information in Text. 2 Tinnitus survey Did the authors use their original question regarding tinnitus? Authors Response: I really appreciate your comments. In this study, there was no objective assessment of tinnitus. My comment is whether questionnaire items you used were the same items used in the past studies. If you used the same items, you can cite the relevant literature. 5 2) In the previous report from the KNHNES, tinnitus was associated with gender, smoking, stress, sleep, hearing loss, hyperlipidemia, osteoarthritis, rheumatoid arthritis, asthma, depression, and thyroid disease history (Kim HJ et al. PLoS One. 2015;10(5):e0127578). Please, show the results after adjusting for sleep, hyperlipidemia, osteoarthritis, rheumatoid arthritis, asthma, depression, and thyroid disease history. Authors Response: According to the reviewer s comment, we have added asthma and sleep duration. The authors partly responded to my comment. Remaining factors? hyperlipidemia, osteoarthritis, rheumatoid arthritis, depression, and thyroid disease history. 9 In Table 2, crude association of breast feeding (p = 0.03) and oophorectomy (p = 0.05) with the severity of tinnitus was suggested. Did this association persist even after adjusting for several confounding factors? The authors may clarify this association in discussion. Authors Response: The use of HRT was significantly associated with tinnitus after adjustment of the breast feeding and bilateral oophorectomy in model 3 of regression analyses.
I understand the aim of this study is to examine the association between HRT and tinnitus. However, they demonstrated the results of the association between other reproductive factors and tinnitus. As a result, breast feeding (p = 0.03) and oophorectomy (p = 0.05) with the severity of tinnitus was suggested. Please, show the results of the association between breast feeding and oophorectomy after adjustment for several confounding factors. Or, I recommend that the authors delete the results regarding other reproductive factors, including oral contraceptive, lactation, pregnancy, and oophorectomy. VERSION 2 AUTHOR RESPONSE I thank the editors and reviewers of the BMJ open for taking their time to review my article. I have made some corrections and clarifications in the article after going over the reviewer s comments. The changes are summarized below. 1. Study participants 2) page 5, line 35. Why did the authors exclude older women aged 65 years or older? Authors Response: In Korea only a small proportion of menopausal women aged 65 years or older currently take hormone replacement therapy. So we excluded the older women. I agree with your response. Please, include this information in Text. Response: According to the reviewer s comment, we added that sentence. 2 Tinnitus survey Did the authors use their original question regarding tinnitus? Authors Response: I really appreciate your comments. In this study, there was no objective assessment of tinnitus. My comment is whether questionnaire items you used were the same items used in the past studies. If you used the same items, you can cite the relevant literature. Response: According to the reviewer s comment, we have added this reference. 18. Seo JH, Kang JM, Hwang SH, Han KD, Joo YH. Relationship between tinnitus and suicidal behaviour in Korean men and women: a cross-sectional study. Clin Otolaryngol. 2016;41:222-227. 5 2) In the previous report from the KNHNES, tinnitus was associated with gender, smoking, stress, sleep, hearing loss, hyperlipidemia, osteoarthritis, rheumatoid arthritis, asthma, depression, and thyroid disease history (Kim HJ et al. PLoS One. 2015;10(5):e0127578). Please, show the results after adjusting for sleep, hyperlipidemia, osteoarthritis, rheumatoid arthritis, asthma, depression, and thyroid disease history. Authors Response: According to the reviewer s comment, we have added asthma and sleep duration. The authors partly responded to my comment. Remaining factors? hyperlipidemia, osteoarthritis, rheumatoid arthritis, depression, and thyroid disease history. Authors Response: According to the reviewer s comment, we have added hyperlipidemia,
osteoarthritis, rheumatoid arthritis, depression, and thyroid disease history in Table 1. 9 In Table 2, crude association of breast feeding (p = 0.03) and oophorectomy (p = 0.05) with the severity of tinnitus was suggested. Did this association persist even after adjusting for several confounding factors? The authors may clarify this association in discussion. Authors Response: The use of HRT was significantly associated with tinnitus after adjustment of the breast feeding and bilateral oophorectomy in model 3 of regression analyses. I understand the aim of this study is to examine the association between HRT and tinnitus. However, they demonstrated the results of the association between other reproductive factors and tinnitus. As a result, breast feeding (p = 0.03) and oophorectomy (p = 0.05) with the severity of tinnitus was suggested. Please, show the results of the association between breast feeding and oophorectomy after adjustment for several confounding factors. Or, I recommend that the authors delete the results regarding other reproductive factors, including oral contraceptive, lactation, pregnancy, and oophorectomy. Authors Response: According to the reviewer s comment, we have erased Table 2. I hope the revised manuscript will better meet the requirements of the BMJ open for publication. I thank you again for the constructive review by the reviewers. BMJ Open: first published as 10.1136/bmjopen-2016-013736 on 10 July 2017. Downloaded from http://bmjopen.bmj.com/ on 15 March 2019 by guest. Protected by copyright.