Author ’ s response to reviews Title : Latent Tuberculosis Infection and Associated Risk Indicators in Pastoral Communities in Southern Ethiopia : A Community Based Cross-sectional Study

Comment-1: Pls include the sampling method in the methods (the introduction can be shorter) Response: We shortened the Introduction and the sampling method was included in abstract part (see abstract part) Comnet-2: Line 38if there was no significant difference pls provide a p value (Did you do a chi square test for this statistic)? I would prefer to see p values in the results Response: P-values were calculated for all variables, now presented for all results in the text and included in Table-3. Regarding chi-squared test (χ2), yes we did Pearson's χ2 for all categorical variables. Background: Comment-3: Page 3, Lines 19-29pls update with statistics from the latest global TB report Response: Now we used WHO, 2017 as reference. New TB case is the same i.e 10.4 million but death rate is reduced to 1.7 million Comment-4: Page 3, Line 41 – pls reference this statement: Response: Now reference is added Comment-5: Page 3, Line 49please start the sentence with “The World Health Organization (WHO)” Also, is that statement by WHO really about active TB disease and not latent TB infection? If so pls clarify and or remove this statement. Response: We clarified The End TB strategy statement and stated that systematic testing and treatment of LTBI in at-risk populations is a critical component in the elimination of TB. Comment-6: Page 4, line 4are there any papers that can support this claim? I wonder if there has been research on this previously? Page 4, lines 6-9I think it may be best to say that for the most part the NTP relies on passive case finding – is that what you mean? I would also be clear when you are talking about active TB and latent TB infection. We may not expect the NTP to be monitoring latent TB infection. I think in your background we need a better sense of why latent Tb infection is important in this community, i.e. what is the rationale for your study? Response: For whole page-4 questions see modified rationale part. Methods: Comment-7: Pls include reference to whether the interview was face to face (I assume so) Page 6, line 59pls correct typo present, not present Page 7, line 6pls correct typo bivariate, not bivariate There is no statement about ethics and there needs to be Response: All questions from methods part were corrected accordingly (see page 5 paragraph 4, line 4; page-6 paragraph 2 lines 13 and 18; and page 6-7) Results Comment-8: Line 28-30 – there is no need to state that blood samples were collected etc. I think there is no need to include every socio-demographic characteristic – you can refer to the table Response: Corrected accordingly (See page 7, result part first paragraph) Comment-9: I think if you present IGRA results that relate to quality tests it is best to describe this in your methods Response: Thank you, now we addressed it in data analysis part in the methods (see page-6 paragraph 2, lines 11-13) Comment-10: Page 8, line 39the numbers for the CI come before the abbreviation CI whereas in other parts of the paper it is the other way around – pls correct Response: Corrected accordingly Comment-11: Page 9, first linepls write the results in the present tense, i.e. the results are presented. Response: Corrected accordingly (see page 9 line 1) Comment-12: You also don’t seem to present any crude odds ratios in your results sectionwas there a reason for this? Also, did you consider putting only selected results into your multivariable regression model? Response: We presented all crude and adjusted odds ratios in Table-3. To address your comment we presented crude odds ratios in text part. Regarding selected results in multivariable regression we included all except Residence and current TB treatment status. In residence case 93.4% were from rural and only 6.6% (33) were come from Urban. In current TB treatment status cases only 18 (3.6%) were used treatment. This is to avoid bias that originates from disproportionate number. We computed multivariable regression in the presence and absence of them but no difference. Discussion Comment-13: I think it would be good to say whether results were fed back to individuals and whether they were offered preventive therapy. It might also be good to discuss your findings in the context of TB control in Ethiopia (i.e. is this in the national Tb strategic plan, etc). Response: In Ethiopia preventive therapy is not common medical practice except for HIVpositive subjects. The goal of this study’s research is to determine the prevalence of LTBI in the pastoralist population in a remote corner of the country and to compare the data with those of other regions in Ethiopia and with other countries. The QuantiFERON tests we used here are recommended by the CDC and WHO to assess LTBI prevalence. Publishing this data, Ethiopian public health authorities will become aware of the LTBI prevalence and make decisions as to whether DOTS preventative treatment is advisable in marginalized pastoralist populations with high rates of LTBI and active TB. Comment-14: Table 1, I suggest that there is a space between your number and the bracket for your percentages, like this 23 (5.1%) Response: Corrected accordingly Comment-15: What does “sick during the survey” mean? Can you describe what this means? Response: Corrected as sick Comment-16: Is the imprison variable about a history of imprisonment? Response: Yes we have variables which indicate type of imprisonment, duration of imprisonment, whether presence of individuals with cough during imprisonment. Comment-17: Figures 1 and 2 are not good quality – can these images be provided in a higher resolution? Response: We redesigned all of the images to width of a higher resolution (1200) and dpi of 300 PNG picture to improve the quality of images. BMC Public Health operates a policy of open peer review, which means that you will be able to see the names of the reviewers who provided the reports via the online peer review system. We encourage you to also view the reports there, via the action links on the left-hand side of the page, to see the names of the reviewers. 2. Responses to Reveiwer-1 Reviewer reports: Judith Bruchfeld (Reviewer 1): This study describes the prevalence of LTBI in a pastoral population in Southern Ethiopia. The study has merits and the subject is pertinent to the WHO plan of TB elimination but due to flaws in study design there are important limitations which preclude some of the conclusions drawn. Commnet-1: One important limitation is the definition of latent TB and how this definition is then used in the study design. The definition of latent TB is given in the background of the manuscript line 34 as " the presence of M. tuberculosis in the body without signs and symptoms, or radiographic or bacteriologic evidence for the presence of TB." This definition of latent TB is somewhat confusing and should be rephrased according to e g Erkens et al EurRespir J 2010; 36: 925-949 ''Latent infection with M. tuberculosis'' is usually defined as presumptive infection with M. tuberculosis complex, as evidenced by a ''positive'' tuberculin skin test reaction and/or a positive interferon-c release assay (IGRA), without any sign of clinically or radiologically manifest disease." Response: We clarified the definition of LTBI according to WHO, 2015 which is almost similar to the definition of Erkens et al (see first sentence of second paragraph in introduction part) Comment-2: Screening for latent TB was performed by IGRA and showing a rather high positivity rate of more than 50%. However positive individuals were screened only for TB symptoms but not further examined with chest radiography to exclude active disease. This is not on line with the definition of latent TB. Could the high IGRA positivity rate be due to concomitant undiagnosed active TB? This should be addressed both regarding the study design, results and discussion sections, i e the high IGRA positivity rate could be due to both latent and active TB. Could undiagnose active TB haven been unevenly distributed in the study population e g by gender and therefore affecting results by logistic regression? Response: It is obvious that the drawback of IGRA test is differentiating LTBI from active TB. As indicated in rationale part of the study Pai et al, described that despite limitations, it is believed that IGRAs could improve existing information about the global epidemiology of LTBI. The study conducted by Joshi et al., 2007 on Health Care Workers with LTBI indicated nearly two-thirds of HCWs with latent TB infection had abnormal radiographic findings, of which only 3.4% had features suggestive of active TB. In USA, Ronald and colloquies on the evaluation of active and LTBI with imaging concluded that “Universal chest radiography in a large preemployment TB screening program was of low yield in the detection of active TB or increased LTBI reactivation risk, and it provided no assistance in deciding which individuals to prioritize for LTBI treatment”. Combined together we believed that a high role of radiological examination in differentiating LTBI from active in individuals with high IFNlimited setting like South Omo where high TB incidence communities with difficult to reach populations IGRA test that conducted in quality controlled condition with clinical examination for active TB and epidemiological data providing fruitful information on the epidemiology of LTBI. Regarding unevenly distribution of undiagnose active TB in the study population, we used relatively best sampling technique for multicenter study design seeing that everyone in the target population having an equal chance of being chosen. For example if we take a variable that raised by the reviewer, in our study the amount of males was 50.2% which is almost equal to that of females. In addition, we were checked for the most common individual-level confounders such as gender, age, BMI, health status, any treatment during blood sample collection and vaccination with BCG before classifyi