BMJ Open Is Committed to Open Peer Review. As Part of This Commitment We Make the Peer Review History of Every Article We Publish Publicly Available

BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay- per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected]

http://bmjopen.bmj.com/ on September 30, 2021 by guest. Protected copyright. BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open

Vitamin D status in tuberculosis patients with diabetes, prediabetes, and normal blood glucose in China

ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2017-017557

Article Type: Research

Date Submitted by the Author: 04-May-2017

Complete List of Authors: Zhao, Xin; Beijing Hospital, Laboratory Medicine Yuan, Yanli; Jilin Provincial Academy of Tuberculosis Control and Prevention, Director Office Lin, Yan; International Union Against Tuberculosis and Lung Disease, TB and HIV-China Zhang, Tiejuan; Jilin Provincial Academy of Tuberculosis Control and Prevention, Prevention and Treatment Ma, Jianjun; Jilin Provincial Academy of Tuberculosis Control and Prevention, Prevention and Treatment Kang, Wanli; Beijing Chest Hospital, Capital Medical University, Epidemiology; Beijing Tuberculosis and Thoracic Tumor Research Institute, Epidemiology

Bai, Yunlong; Jilin Provincial Academy of Tuberculosis Control and http://bmjopen.bmj.com/ Prevention, Prevention and Treatment Wang, Yunlong; Meihekou City Tuberculosis Institute, Prevention and Treatment Shao, Hongshan; Dongfeng County Tuberculosis Institute, Treatment Dlodlo, Riitta; International Union Against Tuberculosis and Lung Diseases, TB and HIV Harries, Anthony; International Union against Tuberculosis and Lung Disease, TB and HIV; London School of Hygiene and Tropical Medicine

Primary Subject on September 30, 2021 by guest. Protected copyright. Infectious diseases Heading:

Secondary Subject Heading: Global health

Keywords: Vitamin D, Tuberculosis < INFECTIOUS DISEASES, Diabetes, China

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 1 of 29 BMJ Open

1 1 2 3 Vitamin D status in tuberculosis patients with diabetes, pre-diabetes, 4 and normal blood glucose in China 5 6 7 Investigators and Institutions: 8 9 Xin Zhao1 10 11 2 12 Yanli Yuan

13 3 14 Yan Lin 15 For peer review only 16 Tiejuan Zhang2 17 18 Jianjun Ma 2 19 20 Wanli Kang4, 5 21

22 2 23 Yunlong Bai 24 6 25 Yunlong Wang 26 27 Hongshan Shao7 28 29 Riitta A. Dlodlo8 30 31 Anthony D. Harries9, 10

32 http://bmjopen.bmj.com/ 33 34 35 36 1 Beijing Hospital, Beijing, China 37 38 2Jilin Provincial Academy of Tuberculosis Control and Prevention, Changchun, China 39 40 3 International Union Against Tuberculosis and Lung Disease, Beijing, China on September 30, 2021 by guest. Protected copyright. 41

42 4 43 Beijing Chest Hospital, Capital Medical University, Beijing, China 44 5 45 Beijing Tuberculosis and Thoracic Tumor Research Institute, Beijing, China 46 47 6 Meihekou City Tuberculosis Institute, Meihekou, China 48 49 7 Dongfeng County Tuberculosis Institute, Dongfeng, China 50 51 8 52 International Union Against Tuberculosis and Lung Diseases, Bulawayo, Zimbabwe

53 9 54 International Union Against Tuberculosis and Lung Diseases, Paris, France 55 56 10 London School of Hygiene and Tropical Medicine, London, UK 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 2 of 29

2 1 2 3 Address for correspondence: 4 5 Professor Yan Lin, 6 7 61151/152, No. 1 Xindonglu, Chaoyang District, Beijing 100600, China. 8 9 10 Telephone: +86 10 8532 1053; Email: [email protected] or 11 12 [email protected] 13 14 15 For peer review only 16 Details of Paper 17 18 19 Word Count of Abstract: 300 20 21 Word Count of the Full Paper: 2,614 22 23 References: 31 24 25 Tables: 4 26 27 Figures: 1 28 29 30 Short title: Vitamin D status in TB, preDM TB and DMTB patients in China 31

32 Key words: Vitamin D; Tuberculosis; Diabetes, China http://bmjopen.bmj.com/ 33 34 35 36 37 38 39 40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 3 of 29 BMJ Open

3 1 2 3 ABSTRACT 4 5 OBJECTIVE: The association between tuberculosis (TB), diabetes (DM) and vitamin 6 7 D status is poorly characterised. We therefore: i) determined vitamin D status in TB 8 9 10 patients in relation to whether they had normal fasting blood glucose (FBG), preDM 11 12 or DM and ii) assessed whether baseline characteristics in TB patients with normal 13 14 FBG, preDM and DM status were associated with vitamin D deficiency. 15 For peer review only 16 METHOD: In TB patients consecutively attending six clinics or hospitals in China, 17 18 19 we measured 25 hydroxycholecalciferol [25(OH)D3] at the time of registration using 20 21 electrochemiluminescence in a COBASE 601 Roche analyser by chemiluminescence 22 23 immunoassay. Data analysis was performed using the chi square test and multivariate 24 25 logistic regression. 26 27 RESULTS: There were 306 eligible TB patients, including 96 with smear positive 28 29 30 pulmonary TB, 187 with smear negative pulmonary TB and 23 with extra pulmonary 31

32 TB (EPTB). Of these, 95 (31%) had normal blood glucose, 83 (27%) had preDM and http://bmjopen.bmj.com/ 33 34 128 (42%) had DM. Median serum vitamin D levels were 16.1 ng/ml in TB patients 35 36 with normal FBG, 12.6 ng/ml in TB patients with preDM and 12.1 ng/ml in TB 37 38 39 patients with DM (P< 0.001). In TB patients with normal FBG, the odds of vitamin D 40 on September 30, 2021 by guest. Protected copyright. 41 deficiency (25(OH)D3 <20 ng/ml) were significantly higher in patients with smear 42 43 negative TB (P=0.003), cigarette smokers (P=0.001) and the cold season (P<0.001). 44 45 Severe vitamin D deficiency (25(OH)D3 < 10 ng/ml) was strongly associated with 46 47 living in a rural area for TB patients with preDM (P=0.043) and with being aged 60 48 49 50 years and over for DMTB patients (P=0.031). 51 52 CONCLUSION: Vitamin D levels are lower in TB patients with preDM and DM and 53 54 are also affected by certain baseline characteristics including older age and rural 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 4 of 29

4 1 2 3 residence. TB programmes need to pay more attention to vitamin D status in their 4 5 patients, especially if there is coexisting preDM or DM. 6 7 8 9 10 STRENGTHS AND LIMITATIONS OF THIS STUDY 11 12 • In contrast to many previous studies on this subject in China, the current study 13 14 enrolled a large number of patients with tuberculosis who were consecutively 15 For peer review only 16 registered in the routine programme setting in urban and rural settings 17 18 19 • Rural settings were further divided into nonpoverty and poverty areas to 20 21 assess the influence of this factor on 25(OH)D3 levels 22 23 • Blood samples for measurement of 25(OH)D3 were taken without collecting 24 25 detailed information about the patient’s history, and there is therefore no data 26 27 28 on duration of symptoms or signs which might have impacted on vitamin D 29 30 levels. 31

32 • The presence or absence of comorbidities that affect vitamin D status http://bmjopen.bmj.com/ 33 34 depended on medical documentation or patient selfreports and these may 35 36 37 have been missed 38 39 • 25(OH)D3 levels were only measured at one time before antiTB treatment 40 on September 30, 2021 by guest. Protected copyright. 41 and therefore there is no information about whether these changed during the 42 43 six months course of antiTB treatment 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 5 of 29 BMJ Open

5 1 2 3 INTRODUCTION 4 5 Despite the 47% decline in tuberculosis (TB) mortality since 1990, the disease 6 7 remains a major global public health threat. In 2015, 10.4 million people developed 8 9 10 active TB and 1.4 million people (without HIV infection) were estimated to have died 11 1 12 from the disease. Successful treatment for TB not only depends on a standardized, 13 14 supervised 6month course of antituberculosis drugs, but also on nutritional well 15 For peer review only 16 being and support.1, 2 17 18 19 Vitamin D is a secosteroid synthesized in the skin by the action of sunlight and 20 21 also ingested in the diet. It has pleiotropic effects on many organs, and plays a key 22 23 role in human innate and adaptive immunity.3, 4 In the context of TB, Vitamin D 24 25 assists mononuclear phagocytes to suppress the intracellular growth of 26 27 Mycobacterium tuberculosis (M. tb) after initial infection.5, 6 Vitamin D does this by 28 29 30 undergoing metabolic activation mainly in the liver to 25hydroxycholecalciferol [25 31

32 (OH)D3] and then in the kidney to 1,25 dihydroxycholecalciferol [1,25(OH)2D3], the http://bmjopen.bmj.com/ 33 34 latter being an immunologically active hormone that stimulates antimycobacterial 35 36 activity in vitro.3, 4 37 38 39 Previous reports have shown that people with vitamin D deficiency have a 40 on September 30, 2021 by guest. Protected copyright. 41 significantly higher risk of developing active TB compared with those who have 42 43 normal vitamin D levels.3, 7, 8 This has been particularly observed in migrants coming 44 45 to the United Kingdom from Asia.9 TB patients have also generally been reported to 46 47 have lower serum levels of vitamin D compared with normal controls.10, 11 These 48 49 50 studies have suggested that vitamin D supplementation during antiTB treatment in 51 52 areas endemic for vitamin D deficiency, particularly in poor communities, might 53 54 improve the response to antiTB treatment. Conversely, however, other studies have 55 56 shown no improved treatment response in terms of sputum culture conversion and 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 6 of 29

6 1 2 3 final treatment outcomes between those receiving vitamin D supplementation and 4 5 those without. 9, 12, 13 These conflicting reports beg the question as to whether vitamin 6 7 D deficiency is equally distributed amongst patients with different types and 8 9 10 categories of TB or amongst TB patients with other coexisting comorbidities, such 11 12 as diabetes mellitus (DM). 13 14 Chaudhary et al analyzed vitamin D status, using a radioimmunoassay method, 15 For peer review only 16 and reported no difference in serum vitamin D levels between TB patients and TB 17 18 19 patients with DM, although the prevalence of severe vitamin D deficiency was highest 20 14 21 in the group of DMTB patients. Conversely, a small scale study in urban China, 22 23 using liquid chromatographytandem mass spectrometry, reported significant 24 25 differences in vitamin D levels among TB, TB with preDM and DMTB patients.10 26 27 There have been no reports of studies looking at associations between vitamin D, DM 28 29 30 and TB in routine programmatic settings. We therefore: i) determined vitamin D 31

32 status in TB patients in relation to whether they had normal fasting blood glucose http://bmjopen.bmj.com/ 33 34 (FBG), preDM or DM and ii) assessed whether baseline characteristics in TB patients 35 36 with normal FBG, preDM and DM status were associated with vitamin D deficiency. 37 38 39 40 on September 30, 2021 by guest. Protected copyright. 41 METHODS 42 43 Design: 44 45 This was a crosssectional study carried out at six TB clinics and hospitals 46 47 within the routine health services in Jilin province, China. Total sample size was 48 49 50 calculated using a simplified 2sided significance level of 0.05 and a power of 0.8. We 51 52 assumed, based on previous studies, that at least 1015% of TB patients in China 53 54 would have serum vitamin D deficiency.15,16 Therefore, the minimum required sample 55 56 size was 6293. However, taking into account poor collection of serum samples in the 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 7 of 29 BMJ Open

7 1 2 3 field of 15% and another 20% of blood samples losing their quality during long 4 5 distance transportation, we needed to recruit at least 129 TB patients. Previous 6 7 experience with this type of work in China has indicated the need to add 5% onto the 8 9 10 number of recruited patients due to missing key information variables and another 11 12 10% to include different types of TB. Our sample size was therefore at least 148 TB 13 14 patients. 15 For peer review only 16 17 18 19 Setting and Sites: 20 21 Previous reports suggested that the prevalence of vitamin D deficiency varies 22 23 depending on the economic situation of study subjects’ residential areas.6 Therefore, 24 25 we selected two TB hospitals in urban areas (Liaoyuan City Chest Hospital and 26 27 Tonghua City TB Hospital), three TB clinics in nonpoverty rural areas (Dongfeng 28 29 30 County TB Institue, Meihekou City TB Institute and Tonghua County TB Institute) 31

32 and one TB clinic in a poverty rural area (Daan County TB Institute), a State poverty http://bmjopen.bmj.com/ 33 34 county streamlined by China State Council with an average annual farmers’ income of 35 36 less than USD$189. The selection of the clinics and hospitals was based on broad 37 38 39 economic development levels, a sufficient number of TB patients stratified by 40 on September 30, 2021 by guest. Protected copyright. 41 different types of the disease, the availability of blood sample storage facilities and 42 43 the willingness of the staff to participate in the study without incentives. 44 45 46 47 Patient recruitment: 48 49 50 Patients who were consecutively diagnosed and registered with TB from 51 52 November 2015 to July 2016 were included in the study. The patients included were 53 54 ≥18 years and newly registered with any type and category of TB according to the 55 56 national and international guidelines.17, 18 Exclusion criteria included: a) pregnant or 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 8 of 29

8 1 2 3 lactating women, b) being positive for human immunodeficiency virus (HIV), c) 4 5 having Aspartate aminotransferase (AST) or Alanine aminotransferase (ALT) ≥3 6 7 times the upper limit of normal level as hepatic dysfunction may alter vitamin D 8 9 10 metabolism, d) receiving vitamin D or Vitamin D analogues for any reason, e) 11 12 receiving corticosteroid treatment for any reason, and f) having cancer or receiving 13 14 anticancer therapy. 15 For peer review only 16 17 18 19 Blood sample collection: 20 21 After the patients were registered, 5 ml venous blood sample was taken 22 23 immediately following an overnight fast of at least 10 hours. The samples were 24 25 centrifuged within 2 hours to separate the serum and stored at 20。C. All serum 26 27 28 samples were transported to the local Centres for Disease Control or blood bank 29 30 within 7 days and stored at 70。C until analysis was undertaken in two single batches. 31

32 http://bmjopen.bmj.com/ 33 34 35 Vitamin D measurement and definition: 36 37 The major circulating metabolite of vitamin D is 25hydroxycholecalciferol 38 39 [25(OH)D3], which was measured in the Laboratory Medicine Department of Beijing 40 on September 30, 2021 by guest. Protected copyright. 41 Hospital using the electrochemiluminescence (ECLIA) method that determines serum 42 43 44 25(OH)D3 levels in a COBASE 601 Roche analyser using a chemiluminescence 45 46 immunoassay (CLIA). Reagents were supplied from Roche with a normal 47 48 measurement range of 370 ng/ml. Holick’s definitions of vitamin D status were used: 49 50 25(OH)D3 ≥ 30 ng/ml = normal; 25(OH)D3 between 2029.9 ng/ml = insufficient 51 52 vitamin D; 25(OH)D between 1019.9 ng/ml = vitamin D deficiency; 25(OH)D 53 3 3 54 4 55 between 0 9.9 ng/ml = severe vitamin D deficiency. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 9 of 29 BMJ Open

9 1 2 3 Data collection and recording: 4 5 A standardized questionnaire was developed for data collection focusing on 6 7 demographic characteristics, types and categories of TB, symptoms, complications 8 9 10 and cigarette smoking status. Patients were diagnosed with DM either as a result of 11 12 this being already known (DM diagnosed at any time by a registered medical 13 14 institution and documented in the clinic notes) or as a result of FBG ≥ 7.0 mmol/L 15 For peer review only 16 being identified at the time of TB registration or at the time of TB diagnosis from 17 18 19 another health facility. PreDM was identified as a result of FBG 6.16.9mmol/L at 20 19 21 the time of TB registration based upon WHO definition. The questionnaire was 22 23 compiled by TB clinic staff and reviewed by the provincial and Union staff during 24 25 monitoring visits and before data analysis. 26 27 28 29 30 Data analysis and statistics: 31

32 Individual patient data were entered to an EXCEL file by the principal investigator. http://bmjopen.bmj.com/ 33 34 Nonparametric tests were used to compare median levels of vitamin D [25(OH)D3] 35 36 between groups. Categorical comparisons of various 25(OH)D3 levels between 37 38 39 patients with DM, preDM and normal FBG were carried out using the chi square test. 40 on September 30, 2021 by guest. Protected copyright. 41 Relationships between 25(OH)D3 and associated exposure variables were evaluated 42 43 with odds ratios (OR) and 95% confidence intervals using logistic regression. Levels 44 45 of significance were set at 5%. We selected variables with unadjusted ORs for which 46 47 the P value was <0.05 and included these in a multivariate logistic regression model. 48 49 50 51 52 Ethics approval: 53 54 The research proposal was reviewed and approved by the provincial 55 56 authorities in the implementing sites. The Ethics Advisory Group, International Union 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 10 of 29

10 1 2 3 Against Tuberculosis and Lung Disease, Paris, France, formally approved this study 4 5 (EAG number: 102/15). 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39 40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 11 of 29 BMJ Open

11 1 2 3 RESULTS 4 5 There were 317 TB patients consecutively registered in this study. Of these, 11 6 7 were excluded because of insufficient volumes of serum or absence of key 8 9 10 information variables. Of 306 eligible TB patients, 96 had smearpositive pulmonary 11 12 TB, 187 had smearnegative pulmonary TB and 23 had extra pulmonary TB (EPTB). 13 14 There were 215 males and 91 females, aged from 1892 years with a mean of 51.5 15 For peer review only 16 years. Of 306 TB patients, 95 had normal FBG, 83 had preDM and 128 had DM. The 17 18 19 group of TB patients with normal FBG was younger than the DMTB group 20 21 (P<0.001), but there were no other statistically significant differences between the 22 23 groups. 24 25 Serum Vitamin D [25(OH)D3] levels in TB patients with normal FBG were 26 27 higher than those found in preDM TB patients and DMTB patients (Table 1). The 28 29 30 proportions of patients with vitamin D deficiency and severe vitamin D deficiency 31

32 were 61.1% for those with TB and normal FBG, 74.7% for those with preDM and http://bmjopen.bmj.com/ 33 34 83.6% for those with DM (P<0.001) The median serum vitamin D levels and their 35 36 distributions in general were similar between Daan county (rural poverty county) and 37 38 39 the other sites (P>0.05), although there was a higher proportion (P=0.032) of patients 40 on September 30, 2021 by guest. Protected copyright. 41 with normal vitamin D levels in the nonpoverty areas (Figure 1). 42 43 Baseline characteristics of TB patients in relation to being vitamin D deficient 44 45 (25(OH)D3 <20ng/ml) are shown in Tables 24. The odds of a TB patient with 46 47 normal FBG having vitamin D deficiency were significantly higher in smear negative 48 49 50 TB patients, those who smoked over 10 cigarettes per day and patients diagnosed in 51 52 the cold season (From November to April) rather than in the warm season (From May 53 54 to October) on both univariate and multivariate analysis (Table 2). The odds ratio was 55 56 increased in EPTB patients although the numbers were small. Baseline characteristics 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 12 of 29

12 1 2 3 were no different with respect to vitamin D deficiency for TB patients with preDM 4 5 (Table 3) or TB patients with DM (Table 4). 6 7 In terms of severe vitamin D deficiency (serum 25(OH)D <10ng/ml), the 8 3 9 10 only significant differences were i) in TB patients with preDM among whom a higher 11 12 proportion living in rural areas were severely deficient compared with urban areas 13 14 (40.7% versus 16.7%, OR 3.429, 95% CI 1.04011.299, P=0.043) and ii) in TB 15 For peer review only 16 patients with DM among whom a higher proportion aged 60 years and above were 17 18 19 severely deficient compared with adults aged 4059 (39.6% versus 21.0%, OR 2.474, 20 21 95% CI 1.0865.623, P=0.031). 22 23 24 25 26 27 DISCUSSION 28 29 30 Our study presents findings of vitamin D status among different types of TB 31

32 patients and in relation to their DM status. http://bmjopen.bmj.com/ 33 34 First, the median level of vitamin D in TB patients with normal FBG at 35 36 16.1ng/ml was lower than that found in the general population (26.3ng/ml),20 and this 37 38 6, 10, 15 39 finding is similar to that reported from other studies. Once antiTB treatment is 40 on September 30, 2021 by guest. Protected copyright. 41 initiated, serum vitamin D levels may be further depressed through the action of drugs 42 43 such as rifampicin which increase liver metabolism,21 and TB clinic staff need to be 44 45 made aware of this fact. 46 47 Second, serum vitamin D levels in TB patients with preDM and with DM 48 49 50 were significantly lower than in TB patients with normal FBG, with a similar trend 51 10 52 having also been found in a small scale study in urban settings of China. This was 53 54 mirrored by the higher proportions of preDM and DM patients with vitamin D 55 56 deficiency and severe deficiency compared with patients who had normal glucose 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 13 of 29 BMJ Open

13 1 2 3 levels. However, these are not consistent observations with a report in India finding 4 5 no significant differences in mean vitamin D levels between groups with TB, DM or a 6 7 combination of both.14 Both TB and DM can be associated with suboptimal 8 9 10 nutritional intake, but in this observational crosssectional study we are unable to 11 12 attribute any causal association as to whether low vitamin D status increases the risk 13 14 of TB or DM or whether DM or TB increases the risk of vitamin D deficiency. 15 For peer review only 16 Third, of the 6 sites participating in this study, one was located in an area 17 18 19 characterised by State Poverty County streamlined by China State Council. Here, 20 21 there was a significantly lower proportion of TB patients with normal vitamin D status, 22 23 although the general distribution of vitamin D status from normal to severe deficiency 24 25 was no different compared with the other counties. Based on this finding, whether 26 27 there is a need for interventions focused on improving nutritional intake and fortifying 28 29 30 food with vitamin D supplements should be discussed. 31

32 Fourth, our study highlighted certain characteristics associated with vitamin D http://bmjopen.bmj.com/ 33 34 deficiency in TB patients according to their DM status. In TB patients with a normal 35 36 FBG, cigarette smoking was an important risk factor for vitamin D deficiency. This 37 38 39 association requires further investigation with possible biochemical reasons including 40 on September 30, 2021 by guest. Protected copyright. 41 links between smoking and micropeptic ulcers which in turn may lead to impaired 42 43 nutritional absorption.22, 23 The finding that patients with smearpositive pulmonary 44 45 TB were at lower risk of vitamin D deficiency compared with those who had smear 46 47 negative disease appears to be counterintuitive as smearpositive TB tends to be 48 49 50 associated with increased disease severity which in turn might be expected to increase 51 52 the risk of vitamin D deficiency. TB patients registered in the cold season had a 53 54 higher risk of vitamin D deficiency compared with those registered in the warm 55 56 season, a finding similar to that observed in Malawi12 and possibly due to less 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 14 of 29

14 1 2 3 sunlight and more time spent indoors during winter months.24 Reports from Viet Nam 4 5 and Afghanistan have found gender differences in vitamin D status to be important for 6 7 TB patients.25, 26 but this was not observed in our study, and this may be due to a 8 9 10 relatively higher general nutritional status for both men and women in the Chinese 11 27 12 population. 13 14 For TB patients with preDM and TB patients with DM, we found that rural 15 For peer review only 16 residence and older age of 60 years and above to be risk factors for severe vitamin D 17 18 19 deficiency. Despite rapid socioeconomic development in all areas of China over the 20 21 last few decades, living standards in rural areas still lag far behind those in urban 22 23 areas in terms of average annual income and per capita purchases for food.28 A recent 24 25 study in China highlighted that aging, a family history of DM, being overweight or 26 27 underweight were associated with preDM and TB, the former being in itself a risk 28 29 29 30 factor for the development of DM. Both DM and TB and old age are associated 31 30

32 with impaired nutritional intake and this might result in low levels of 25(OH)D3. A http://bmjopen.bmj.com/ 33 34 previous study revealed that elderly TB patients in China had a long delays in 35 36 accessing health services and this might also contribute to poor vitamin D status.31 37 38 39 The strengths of our study were the large number of TB patients enrolled and 40 on September 30, 2021 by guest. Protected copyright. 41 the consecutive registration process. There were, however, some limitations. We took 42 43 blood samples for measurement of 25(OH)D3 immediately after the diagnosis of TB 44 45 without collecting the patient’s history. We therefore do not know the duration of 46 47 symptoms or signs of TB, a factor which might have impacted on vitamin D levels. 48 49 50 The presence or absence of comorbidities depended on medical documentation or 51 52 patient selfreports, so we may have overlooked some conditions impacting on 53 54 vitamin D status because we did not actively screen for them. In addition, we only 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 15 of 29 BMJ Open

15 1 2 3 measured 25(OH)D3 levels at one time before antiTB treatment and therefore do not 4 5 know whether these levels changed during the six months course of antiTB treatment. 6 7 8 9 10 11 12 CONCLUSION 13 14 Serum vitamin D levels were lower in TB patients with preDM and DM 15 For peer review only 16 compared with those whose FBG is normal. Vitamin D levels are also affected by 17 18 19 certain baseline characteristics, which in preDM and DM include older age and rural 20 21 residence. TB Programmes need to pay more attention to vitamin D status in their 22 23 patients, especially if there is coexisting preDM or DM. 24 25 26 27

28 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39 40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47

48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 16 of 29

16 1 2 3 ACKNOWLEDGEMENTS 4 5 We thank all the staff at the six TB clinics and hospitals for their support in the blood 6 7 sample collection, storage and managing patients. 8 9 10 11 12 CONFLICTS OF INTEREST 13 14 No conflicts declared. 15 For peer review only 16 17 18 19 CONTRIBUTORS 20 21 Xin Zhao (XZ), Yan Lin (YL) and YY (Yanli Yuan) proposed the research idea. 22 23 Anthony D Harries (ADH), Riitta A Dlodlo (RAD) and Yan Lin (YL) contributed to 24 25 the study design. Xin Zhao (XZ) worked on laboratory analysis and quality assurance. 26 27 Wanli Kang (WK) and Yan Lin (YL) conducted data entry and statistics analysis. 28 29 30 Tiejuan Zhang (TZ), Jianjun Ma (JM), Yunlong Bai (YB), Yunlong Wang (YW) and 31

32 Hongshan Shao (HS) contributed to patient management, recording, reporting and http://bmjopen.bmj.com/ 33 34 blood sample transportation. Yan Lin (YL), Anthony D Harries (ADH) and Riitta A 35 36 Dlodlo (RAD) contributed to the writing of the first draft. All authors contributed to 37 38 39 the subsequent revisions of the paper. All authors read and approved the final paper 40 on September 30, 2021 by guest. Protected copyright. 41 for submission. Yan Lin, Yanli Yuan and Xin Zhao share the primary authorship. 42 43 44 45 FUNDING 46 47 There was no specific funding line for this study. The authors bore their own costs for 48 49 50 conducting this study. 51 52 53 54 DATA SHARING STATEMENT 55 56 The data are part of an ongoing study and are not available at this time. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 17 of 29 BMJ Open

17 1 2 3 REFERENCES 4 5 1. World Health Organization. WHO Report 2016. Global Tuberculosis Control 6 7 2015. Available at www.who.int/tb/publication/global_report/en. 8 9 10 2. Frieden T. Toman’s Tuberculosis. Second Edition. 2004. World Health 11 12 Organization, Geneva. 13 14 3. Zittermann A. Vitamin D in preventive medicine: are we ignoring the 15 For peer review only 16 evidence? Bri J Nutr 2003: 89 (5): 55272. Doi: 10.1079/BJN2003837. 17 18 19 4. Holick MF, Chen TC. Vitamin D deficiency: a worldwide problem with health 20 21 consequences. The American J of Clinical Nutr 2008; 87 (4): 1080 s1086s. 22 23 5. Chesdachai S, Zughaier SM, Li H et al. The effects of firstline anti 24 25 tuberculosis drugs on the actions of vitamin D in human macrophages. J of 26 27 Clinical & Translational Endocrinology 2016; 6: 2329. Doi: 28 29 30 org/10.1016/j.jcte.2016.08.005 31

32 6. Sutaria N, Liu CT, Chen TC. Vitamin D status, receptor gene polymorphisms, http://bmjopen.bmj.com/ 33 34 and supplementation on tuberculosis: A systematic review of casecontrol 35 36 studies and randomized controlled trials. J of Clinical & Translational 37 38 39 Endocrinology 2014; 1: 151160. Doi: org/10.1016/j.jcte.2014.08.001 40 on September 30, 2021 by guest. Protected copyright. 41 7. Zeng J, Wu G, Yang W et al. A serum vitamin D level <25nmol/L pose high 42 43 tuberculosis risk: a metaanalysis. PLOS ONE 2015. Doi: 44 45 10.1371/journal.pone.0126014. 46 47 8. ArnedoPena A, JuanCerdan JV, RomeuGarcia A et al. Vitamin D status and 48 49 50 incidence of tuberculosis among contacts of pulmonary tuberculosis patients. Int 51 52 J Tuberc Lung Dis 2014; 19 (1): 6569. http://dx.doi.org/10.5588/ijtld.14.0348. 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 18 of 29

18 1 2 3 9. Davies PO. Vitamin D and tuberculosis: more effective in prevention than 4 5 treatment? Int J Tuberc Lung Dis 2015; 19 (8): 876. 6 7 http://dx.doi.org/10.5588/ijtld.15.0506. 8 9 10 10. Zhan Y, Jiang L. Status of vitamin D, antimicrobial peptide cathelicidin and T 11 12 helperassociated cytokines in patients with diabetes mellitus and pulmonary 13 14 tuberculosis. Exper and Thera medi 2015; 9: 1116. Doi: 10.3892/etm.2014.2042. 15 For peer review only 16 11. Wang Q, Ma A, Christian Bygbjerg Ib et al. Rationale and design of a 17 18 19 randomized controlled trial of the effect of retinol and vitamin D 20 21 supplementation on treatment in active pulmonary tuberculosis patients with 22 23 diabetes. BMC Infectious Dis 2013; 13: 10421. 24 25 http://www.biomedcentral.com/14712334/13/104. 26 27 12. Sloan DJ, Mwandumba HC, Kamdolozi M et al. Vitamin D deficiency in 28 29 30 Malawian adults with pulmonary tuberculosis: risk factors and treatment 31

32 outcomes. Int J Tuberc Lung Dis 2015; 19 (8): 90411. http://bmjopen.bmj.com/ 33 34 http://dx.doi.org/10.5588/ijtld.15.0071. 35 36 13. Xia J, Shi L Zhao L et al. Impact of vitamin D supplementation on the outcome 37 38 39 of tuberculosis treatment: a systematic review and metaanalysis of randomized 40 on September 30, 2021 by guest. Protected copyright. 41 control trials. Chinese medical J 2014; 127 (17): 312734. Doi: 42 43 10.3760/cma.j.issn.03666999.20140702. 44 45 14. Chaudhary S, Thukral A, Tiwari S et al. Vitamin D status of patients with type 2 46 47 diabetes and sputum positive pulmonary tuberculosis. Indian J Endocriol Matab 48 49 50 2013; 17 (Suppl 3): S6703. Doi: 10.4103/22308210.123564. 51 52 15. Zhu H, Cheng Q, Gan J et al. Vitamin D status in healthy man and woman in 53 54 Shanghai. Chinese J Oste & Bon Min 2010; 3 (3): 157163. Doi: 55 56 10.3969/j.issn.16742591.2010.03.002. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 19 of 29 BMJ Open

19 1 2 3 16. Wang S, Shen G, Jiang S et al. Nutrient status of vitamin D among Chinese 4 5 children. Nutrients 2017; 9 (4): doi: 10.3390/nu9040319. 6 7 17. Ministry of Health and China CDC. Guideline of national TB control program. 8 9 10 Beijing: The Peking Union Medical College Publishing House; 2008. 11 12 18. World Health Organization. Treatment of Tuberculosis Guidelines. Fourth 13 14 Edition, 2009. WHO, Geneva, Switzerland. WHO/HTM/TB/2009.420. 15 For peer review only 16 19. World Health Organization and International Diabetes Federation. Definition 17 18 19 and Diagnosis of Diabetes Mellitus and Intermediate Hyperglycaemia. At 20 21 www.who.int/diabetes/publications/diagnosis_diabetes2006/en/. Geneva, 22 23 Switzerland: WHO, 2006. 24 25 20. Wei QS, Chen ZQ, Tan X et al. Relation of age, sex and bone mineral density 26 27 to serum 25hydroxyvitamin D levels in Chinese women and men. Orthp Surg 28 29 30 2015; 7 (4): 3439. Doi: 10.1111/os.12206. 31

32 21. Bikle DD. What is new in vitamin D: 20062007. Current Opinion in http://bmjopen.bmj.com/ 33 34 Rheumatology 2007; 19: 3838. 35 36 22. Crofton SJ, Simpson D. Tobacco: a global threat. Macmillan Education. 2002. 37 38 39 Available at www.macmillanafrica.com 40 on September 30, 2021 by guest. Protected copyright. 41 23. Eriksen M, Mackay J, Schluger N et al. The Tobacco Atlas (Fifth Edition). 42 43 American Cancer Society and World Lung Foundation. 2015. Available at www. 44 45 TobaccoAtlas.org. 46 47 24. Sedrani SH, Elidrissy AW, El Arabi KM. Sunlight and vitamin D status in 48 49 50 normal Saudi subjects. Ame J Clin Nutr 1983; 38: 12932. 51 52 25. Sarin P, Duffy J, Mughal E et al. Vitamin D and tuberculosis: review and 53 54 association in three rural provinces of Afghanistan. Int J Tuberc Lung Dis 55 56 2016; 20 (3): 3838. http://dx.doi.org/10.5588/ijtld.15.0303. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 20 of 29

20 1 2 3 26. Hopham LT, Nguyen ND, Nguyen TT et al. Association between vitamin D 4 5 insufficiency and tuberculosis in a Vietnamese population. BMC Infect Dis 6 7 2010; 10: 30610. 8 9 10 27. World Health Organization. WHO Report 2016. World Health Statistics 2016. 11 12 Available at www.who.int/gho/publications/world_health_statistics/en. 13 14 28. National Bureau of Statistics. China Statistical Yearbook 2014. China Statistics 15 For peer review only 16 Press, Beijing, 2014. 17 18 19 29. Wang Q, Ma A, Han X et at. Prevalence of type 2 diabetes among newly 20 21 detected pulmonary tuberculosis patients in China: A community based cohort 22 23 study. PLOS One 2013; 8(12): e82660. Doi: 10.1371/journal.pone.0082660. 24 25 30. Wang Q, Ma A, Han X et at. Is low serum 25hydroxyvitamin D a possible link 26 27 between pulmonary tuberculosis and type 2 diabetes? Asia Pac J Clin Nutr 2017; 28 29 30 26 (2): 2416. Doi: 10.6133/1pjcn.032016.02. 31

32 31. Lin Y, Enarson DA, Chiang CY et al. Patient delay in the diagnosis and http://bmjopen.bmj.com/ 33 34 treatment of tuberculosis in China: findings of case detection projects. Public 35 36 Health Action 2015; 5 (1): 6569. Doi: org/10.5588/pha.14.0066. 37 38 39 40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 21 of 29 BMJ Open

21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

P value P

0.001 0.001

128 = 12.1 0.001

69 (53.9) (53.9) 69 18 (14.1) (14.1) 18 0.627 N with DM with TB patients patients TB

=83 12.6 N

http://bmjopen.bmj.com/ TB patients patients TB with pre-DM pre-DM with BMJ Open

=95

16.1 23 (24.2) (9.6) 8 (2.3) 3 <0.001 19 (20.0) 28 (33.7) (29.7) 38 0.041 N normal FBG normal on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 9.9ng/ml) 9.9ng/ml) For peer review≤ only

Vitamin D statusD Vitamin patients with TB No (%) with normal level normal(≥30ng/ml) with No(%) No (%) with deficiency (10-19.9ng/ml) deficiency(10-19.9ng/ml) with No(%) 39 (41.1) 34 (41.0) FBG = fasting blood glucose; pre-DM = pre diabetes mellitus; DM = diabetes mellitus = DM diabetes prediabetes mellitus; glucose;= pre-DM fastingblood = FBG of25-(OH)D measurements by statusD determined Vitamin Median level (ng/ml)level Median (20-29.9ng/ml) insufficiency with No(%) 14 (14.7) deficiencysevere ( with No(%) 13 (15.7) Table 1: Vitamin D status in TB patients in China, stratified by those with normal FBG, pre-DM and DM FBG,and pre-DM withthosenormal by stratified TBChina, inin patients D statusVitamin 1: Table

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from 23 0.003 0.028 0.021 0.001 P value P Multivariate Multivariate OR adjusted CI) (95%

6.194(1.873-20.481) 25.681(1.411-467.424) 1.836(0.382-8.825) 5.619(1.293-24.424) 36.464(4.602-288.900) Univariate OR OR Univariate CI) (95% Reference 1.029(0.407-2.597) Reference 1.400(0.564-3.589) 1.246(0.411-3.889) Reference 1.330(0.526-3.359) Reference 2.533(1.045-6.141) 6.000(0.630-57.139) Reference 0.738(0.208-2.619) Reference 0.913(0.255-3.274) 1.598(0.559-4.570) 6.848(1.397-33.566) http://bmjopen.bmj.com/ BMJ Open (83.3) 5 (54.5) 6 (50.0) 6 on September 30, 2021 by guest. Protected copyright. No.(%) with No.(%)with D Vitamin deficiency N=58 42 (60.9) 16 (61.5) 18 (56.3) 27 (64.3) 13 (61.9) 14 (56.0) 44 (62.9) 15 (45.5) 38 (67.9) 52 (61.9) 23 (52.3) 14 (63.6) 15 (88.2) For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only 60 60 20

Male Female <40 40-59 ≥ Urban Rural Positive Negative EPTB New Retreatment 0 1-9 10-19 ≥

Characteristics Sex Age Residence Type/Sputum smear of TB Category Smoking cigarette/day Table 2: Characteristics of TB patients in China with normal blood glucose in relation to vitamin D deficiency (<20ng/ml) deficiency vitamin toD in relation glucose blood normal China in with patients ofTB Characteristics 2: Table

Page 23 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from Page 24 of 29 24 <0.001 14.402(4.070-50.965)

Reference 5.253(2.140-12.893) http://bmjopen.bmj.com/ BMJ Open on September 30, 2021 by guest. Protected copyright.

18 (40.9) 40 (78.4) For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only May-October November-April

intervals. confidenceCI= =ratio; ORodds tuberculosis; extra-pulmonary = EPTB tuberculosis; = TB

of TB Month registration

Pvalue adjusted OR(95% adjusted CI) Multivariate Multivariate

http://bmjopen.bmj.com/ BMJ Open Reference 2.571(0.758-8.720) 2.556(0.719-9.081) Reference 1.769(0.620-5.048) Reference 1.009(0.325-3.135) 0.375(0.076-1.858) Reference 3.589(0.754-17.802) Reference 3.200(0.769-13.135) 2.400(0.713-8.077) 3.600(0.385-33.637) Reference 0.665(0.236-1.876) Univariate OR (95% (95% CI) OR Univariate on September 30, 2021 by guest. Protected copyright. (55.6) 5 (85.7) 6 44(35.1) 18(30.8) 12(60.0) 27(79.4) 23(79.3) 16(66.7) 46(78.0) 20(76.9) 37(77.1) 45(70.3) 17(89.5) 20(62.5) 16(84.2) 20(80.0) No.(%) with No.(%)with D Vitamin deficiency N=62 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only 60 60 20 Male Female <40 40-59 ≥ Urban Rural Positive Negative EPTB New Retreatment 0 1-9 10-19 ≥

Sex Age Residence Type/Sputum smear of TB Category Smoking cigarette/day Table 3: Characteristics of TB patients in China with pre-DM in relation to vitamin D deficiency (<20ng/ml) (<20ng/ml) deficiency D vitamin relation to in pre-DM China in with patients ofTB Characteristics 3: Table Characteristics

Page 25 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from Page 26 of 29 26

http://bmjopen.bmj.com/ BMJ Open Reference Reference 1.110(0.407-3.024)

on September 30, 2021 by guest. Protected copyright.

25 (73.5) 25(73.5) 37(75.5) For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only May-October May-October November-April

intervals. confidence CI= tuberculosis; extra-pulmonary = EPTB tuberculosis; = TB

Month of TB Month registration

P value P adjusted OR (95% OR adjusted CI)

) ) http://bmjopen.bmj.com/ BMJ Open Reference 0.685(0.135-3.484) 1.424(0.253-8.031) Reference 1.033(0.401-2.660) Reference 0.954(0.335-2.715) 1.355(0.162-12.472) Reference 0.929(0.329-2.625) Reference 1.633(0.411-6.487) 0.884(0.289-2.706) 0.816(0.150-4.434) Reference 0.663(0.250-1.757 Univariate OR (95% CI) OR Univariate Multivariate on September 30, 2021 by guest. Protected copyright. (87.5) 7 (80.0) 8 76(85.4) 31(79.5) 11(84.6) 49(79.0) 47(88.7) 45(83.3) 62(83.8) 31(83.8) 69(83.1) 78(83.9) 29(82.9) 49(83.1) 24(88.9) 26(81.3) No.(%) with No.(%)with D Vitamin deficiency N=107 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only 60 60 20 Male Female <40 40-59 ≥ Urban Rural Positive Negative EPTB New Retreatment 0 1-9 10-19 ≥

Sex Age Residence Type/Sputum smear of TB Category Smoking cigarette/day Table 4: Characteristics of TB patients in China with DM in relation to vitamin D deficiency (<20ng/ml) (<20ng/ml) D vitamin to deficiency DMrelation in China in with patients ofTB Characteristics 4: Table Characteristics

Page 27 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from Page 28 of 29 28

http://bmjopen.bmj.com/ BMJ Open Reference Reference 1.509(0.569-3.999)

on September 30, 2021 by guest. Protected copyright.

31 (79.5) 31(79.5) 75(85.4) For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only May-October May-October November-April

intervals. confidence CI= tuberculosis; extra-pulmonary = EPTB tuberculosis; = TB

Month of TB Month registration

http://bmjopen.bmj.com/ BMJ Open on September 30, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only Figure 1: Vitamin D status of TB patients in the poverty rural area compared with other non-poverty areas inChina areas othernon-poverty with compared area rural TBinthe poverty patients of Dstatus Vitamin1: Figure

Page 29 of 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open

Vitamin D status in tuberculosis patients with diabetes, prediabetes, and normal blood glucose in China: a crosssectional study

For peer review only Journal: BMJ Open

Manuscript ID bmjopen-2017-017557.R1

Article Type: Research

Date Submitted by the Author: 07-Aug-2017

Epidemiology http://bmjopen.bmj.com/ Bai, Yunlong; Jilin Provincial Academy of Tuberculosis Control and Prevention, Prevention and Treatment Wang, Yunlong; Meihekou City Tuberculosis Institute, Prevention and Treatment Shao, Hongshan; Dongfeng County Tuberculosis Institute, Treatment Dlodlo, Riitta; International Union Against Tuberculosis and Lung Diseases, TB and HIV Harries, Anthony; International Union against Tuberculosis and Lung

Disease, TB and HIV; London School of Hygiene and Tropical Medicine on September 30, 2021 by guest. Protected copyright.

Primary Subject Infectious diseases Heading:

Secondary Subject Heading: Global health, Infectious diseases

Keywords: Vitamin D, Tuberculosis < INFECTIOUS DISEASES, Diabetes, China

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 1 of 30 BMJ Open

1 1 2 3 Vitamin D status in tuberculosis patients with diabetes, prediabetes, 4 and normal blood glucose in China: a cross-sectional study 5 6 7 Investigators and Institutions: 8 9 Xin Zhao1 10 11 2 12 Yanli Yuan

13 3, 4 14 Yan Lin 15 For peer review only 16 Tiejuan Zhang2 17 18 Jianjun Ma 2 19 20 Wanli Kang5, 6 21

22 2 23 Yunlong Bai 24 7 25 Yunlong Wang 26 27 Hongshan Shao8 28 29 Riitta A. Dlodlo3, 9 30 31 Anthony D. Harries3, 10

32 http://bmjopen.bmj.com/ 33 34 35 36 1 Beijing Hospital, Beijing, China 37 38 2Jilin Provincial Academy of Tuberculosis Control and Prevention, Changchun, China 39 40 3 International Union Against Tuberculosis and Lung Disease, Paris, France on September 30, 2021 by guest. Protected copyright. 41

42 4 43 International Union Against Tuberculosis and Lung Disease, Beijing, China 44 5 45 Beijing Chest Hospital, Capital Medical University, Beijing, China 46 47 6 Beijing Tuberculosis and Thoracic Tumor Research Institute, Beijing, China 48 49 7 Meihekou City Tuberculosis Institute, Meihekou, China 50 51 8 52 Dongfeng County Tuberculosis Institute, Dongfeng, China

53 9 54 International Union Against Tuberculosis and Lung Diseases, Bulawayo, Zimbabwe 55 56 10 London School of Hygiene and Tropical Medicine, London, UK 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 2 of 30

2 1 2 3 Address for correspondence: 4 5 Professor Yan Lin, 6 7 61151/152, No. 1 Xindonglu, Chaoyang District, Beijing 100600, China. 8 9 10 Telephone: +86 10 8532 1053; Email: [email protected] or 11 12 [email protected] 13 14 15 For peer review only 16 Details of Paper 17 18 19 Word Count of Abstract: 280 20 21 Word Count of the Full Paper: 2,820 22 23 References: 34 24 25 Tables: 3 26 27 Figures: 1 28 29 30 Short title: Vitamin D status in TB, preDM TB and DMTB patients in China 31

Page 3 of 30 BMJ Open

3 1 2 3 ABSTRACT 4 5 OBJECTIVE: The association between tuberculosis (TB), diabetes (DM) and vitamin 6 7 D status is poorly characterised. We therefore: i) determined vitamin D status in TB 8 9 10 patients in relation to whether they had normal fasting blood glucose (FBG), preDM 11 12 or DM and ii) assessed whether baseline characteristics in TB patients, including their 13 14 DM status, were associated with vitamin D deficiency. 15 For peer review only 16 METHOD: In TB patients consecutively attending six clinics or hospitals in China, 17 18 19 we measured 25 hydroxycholecalciferol [25(OH)D3] at the time of registration using 20 21 electrochemiluminescence in a COBASE 601 Roche analyser by chemiluminescence 22 23 immunoassay. Data analysis was performed using the chi square test, odds ratios and 24 25 multivariate logistic regression. 26 27 RESULTS: There were 306 eligible TB patients, including 96 with smear positive 28 29 30 pulmonary TB, 187 with smear negative pulmonary TB and 23 with extra pulmonary 31

32 TB (EPTB). Of these, 95 (31%) had normal blood glucose, 83 (27%) had preDM and http://bmjopen.bmj.com/ 33 34 128 (42%) had DM. Median serum vitamin D levels were 16.1 ng/ml in TB patients 35 36 with normal FBG, 12.6 ng/ml in TB patients with preDM and 12.1 ng/ml in TB 37 38 39 patients with DM (P< 0.001). The study highlighted certain baseline characteristics 40 on September 30, 2021 by guest. Protected copyright. 41 associated with vitamin D deficiency (25(OH)D3 <20 ng/ml). After adjusting for 42 43 confounders, serum vitamin D deficiency was significantly more common in patients 44 45 being registered in the cold season (November to April) (P=0.006) and in those with 46 47 DM (P=0.003). 48 49 50 CONCLUSION: Vitamin D levels are lower in TB patients with preDM and DM and 51 52 are also affected by certain baseline characteristics that include being registered in the 53 54 cold season and having DM. TB programmes need to pay more attention to vitamin D 55 56 status in their patients, especially if there is coexisting preDM or DM. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 4 of 30

4 1 2 3 4 5 STRENGTHS AND LIMITATIONS OF THIS STUDY 6 7 8 • In contrast to many previous studies on this subject in China, the current study 9 10 enrolled a large number of patients with tuberculosis who were consecutively 11 12 registered in the routine programme setting in urban and rural settings 13 14 • Rural settings were further divided into nonpoverty and poverty areas to 15 For peer review only 16 assess the influence of this factor on 25(OH)D levels 17 3 18 19 • Blood samples for measurement of 25(OH)D3 were taken without collecting 20 21 detailed information about the patient’s history, and there is therefore no data 22 23 on duration of symptoms or signs which might have impacted on vitamin D 24 25 levels. 26 27 28 • The presence or absence of comorbidities that affect vitamin D status 29 30 depended on medical documentation or patient selfreports and these may 31

32 have been missed http://bmjopen.bmj.com/ 33 34 • 25(OH)D levels were only measured at one time before antiTB treatment 35 3 36 37 and therefore there is no information about whether these changed during the 38 39 six months course of antiTB treatment 40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 5 of 30 BMJ Open

5 1 2 3 INTRODUCTION 4 5 Despite the 47% decline in tuberculosis (TB) mortality since 1990, the disease 6 7 remains a major global public health threat. In 2015, 10.4 million people developed 8 9 10 active TB and 1.4 million people (without HIV infection) were estimated to have died 11 1 12 from the disease. Successful treatment for TB not only depends on a standardized, 13 14 supervised 6month course of antituberculosis drugs, but also on nutritional well 15 For peer review only 16 being and support.1, 2 17 18 19 Vitamin D is a secosteroid synthesized in the skin by the action of sunlight and 20 21 also ingested in the diet. It has pleiotropic effects on many organs, and plays a key 22 23 role in human innate and adaptive immunity.3, 4 In the context of TB, Vitamin D 24 25 assists mononuclear phagocytes to suppress the intracellular growth of 26 27 Mycobacterium tuberculosis after initial infection.5, 6 Vitamin D does this by 28 29 30 undergoing metabolic activation mainly in the liver to 25hydroxycholecalciferol [25 31

BMJ Open Page 6 of 30

6 1 2 3 final treatment outcomes between those receiving vitamin D supplementation and 4 5 those without. 9, 12, 13 These conflicting reports beg the question as to whether vitamin 6 7 D deficiency is equally distributed amongst patients with different types and 8 9 10 categories of TB or amongst TB patients with other coexisting comorbidities, such 11 12 as diabetes mellitus (DM). 13 14 Persons with DM have three times the risk of developing active TB compared 15 For peer review only 16 with normal people, for reasons that are not particularly well understood.14, 15 Vitamin 17 18 3, 16 19 D deficiency may play a role in this association. Proteomic analysis has suggested 20 21 that there are qualitative differences in the vitamin Dbinding protein precursor and 22 23 other proteins between TB patients with and without DM.17 Findings from an 24 25 intervention study have also shown that when monocyte derived macrophages from 26 27 DM patients with low vitamin D receptor expression were supplemented with vitamin 28 29 18 30 D, the cellular function of eliminating Mycobacterium tuberculosis improved. 31

32 Clinical studies have also been undertaken but have shown conflicting results. http://bmjopen.bmj.com/ 33 34 Chaudhary et al analyzed vitamin D status, using a radioimmunoassay method, and 35 36 reported no difference in serum vitamin D levels between TB patients and TB patients 37 38 39 with DM, although the prevalence of severe vitamin D deficiency was highest in the 40 19 on September 30, 2021 by guest. Protected copyright. 41 group of DMTB patients. Conversely, a small scale study in urban China, using 42 43 liquid chromatographytandem mass spectrometry, reported significant differences in 44 45 vitamin D levels among TB, TB with preDM and DMTB patients.10 There have 46 47 been no reports of studies looking at associations between vitamin D, DM and TB 48 49 50 within routine programmatic settings. We therefore: i) determined vitamin D status in 51 52 TB patients in relation to whether they had normal fasting blood glucose (FBG), pre 53 54 DM or DM and ii) assessed whether baseline characteristics in TB patients with 55 56 normal FBG, preDM and DM status were associated with vitamin D deficiency. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 7 of 30 BMJ Open

7 1 2 3 4 5 METHODS 6 7 Design: 8 9 10 This was a crosssectional study carried out at six TB clinics and hospitals 11 12 within the routine health services in Jilin province, China. Total sample size was 13 14 calculated using a simplified 2sided significance level of 0.05 and a power of 0.8. We 15 For peer review only 16 assumed, based on previous studies, that at least 1015% of TB patients in China 17 18 20, 21 19 would have serum vitamin D deficiency. Therefore, the minimum required 20 21 sample size was 6293. However, taking into account poor collection of serum 22 23 samples in the field of 15% and another 20% of blood samples losing their quality 24 25 during longdistance transportation, we needed to recruit at least 129 TB patients. 26 27 Previous experience with this type of work in China has indicated the need to add 5% 28 29 30 onto the number of recruited patients due to missing key information variables and 31

32 another 10% to include different types of TB. Our sample size was therefore at least http://bmjopen.bmj.com/ 33 34 148 TB patients. 35 36 37 38 39 Setting and Sites: 40 on September 30, 2021 by guest. Protected copyright. 41 Previous reports suggested that the prevalence of vitamin D deficiency varies 42 43 depending on the economic situation of study subjects’ residential areas.6 Therefore, 44 45 we selected two TB hospitals in urban areas (Liaoyuan City Chest Hospital and 46 47 Tonghua City TB Hospital), three TB clinics in nonpoverty rural areas (Dongfeng 48 49 50 County TB Institue, Meihekou City TB Institute and Tonghua County TB Institute) 51 52 and one TB clinic in a poverty rural area (Daan County TB Institute), a State poverty 53 54 county streamlined by China State Council with an average annual farmers’ income of 55 56 less than USD$189. The selection of the clinics and hospitals was based on broad 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 8 of 30

8 1 2 3 economic development levels, a sufficient number of TB patients stratified by 4 5 different types of the disease, the availability of blood sample storage facilities and 6 7 the willingness of the staff to participate in the study without incentives. 8 9 10 11 12 Patient recruitment: 13 14 Patients who were consecutively diagnosed and registered with TB from 15 For peer review only 16 November 2015 to July 2016 were included in the study. The diagnosis of TB was 17 18 19 made in line with recommendations from the WHO and the China National TB 20 22, 23 21 Program guidelines. Patients with suspected TB were investigated by sputum 22 23 smear microscopy. If sputum specimens were smear positive for acidfast bacilli, the 24 25 patient was diagnosed as smear positive pulmonary TB. If sputum smears were 26 27 negative, but the patient had TBrelated symptoms and chest radiography that was 28 29 30 compatible with active pulmonary TB, the patient was diagnosed as smear negative 31

32 pulmonary TB after a formal consultation by physicians. Extra pulmonary TB (EPTB) http://bmjopen.bmj.com/ 33 34 was diagnosed on clinical grounds, using additional circumstantial evidence such as 35 36 radiography, ultrasound and laboratory investigations according to the site of the 37 38 39 lesion: for a diagnosis of EPTB to be made there had to be at least one specimen with 40 on September 30, 2021 by guest. Protected copyright. 41 confirmed Mycobacterium tuberculosis on smear / culture or histological evidence of 42 43 TB or strong clinical and circumstantial evidence consistent with the presence of 44 45 active extra pulmonary disease. 46 47 The patients included were ≥18 years and newly registered with any type and 48 49 22, 23 50 category of TB according to the national and international guidelines. Exclusion 51 52 criteria included: a) pregnant or lactating women, b) being positive for human 53 54 immunodeficiency virus (HIV), c) having Aspartate aminotransferase (AST) or 55 56 Alanine aminotransferase (ALT) ≥3 times the upper limit of normal level as hepatic 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 9 of 30 BMJ Open

9 1 2 3 dysfunction may alter vitamin D metabolism, d) receiving vitamin D or Vitamin D 4 5 analogues for any reason, e) receiving corticosteroid treatment for any reason, and f) 6 7 having cancer or receiving anticancer therapy. 8 9 10 11 12 Blood sample collection: 13 14 After the patients were registered, 5 ml venous blood sample was taken 15 For peer review only 16 immediately following an overnight fast of at least 10 hours. The samples were 17

18 。 19 centrifuged within 2 hours to separate the serum and stored at 20 C. All serum 20 21 samples were transported to the local Centres for Disease Control or blood bank 22 23 within 7 days and stored at 70。C until analysis was undertaken in two single batches. 24 25 26 27 28 Vitamin D measurement and definition: 29 30 The major circulating metabolite of vitamin D is 25hydroxycholecalciferol 31

32 [25(OH)D ], which was measured in the Laboratory Medicine Department of Beijing http://bmjopen.bmj.com/ 33 3 34 35 Hospital using the electrochemiluminescence (ECLIA) method that determines serum 36 37 25(OH)D3 levels in a COBASE 601 Roche analyser using a chemiluminescence 38 39 immunoassay (CLIA). Reagents were supplied from Roche with a normal 40 on September 30, 2021 by guest. Protected copyright. 41 measurement range of 370 ng/ml. In brief, the assay uses a polyclonal antibody 42 43 44 directed against 25(OH)D3. Two incubations are required with the specified reagents, 45 46 following which the reaction mixture is aspirated into a measuring cell where 47 48 application of a voltage to the electrode then induces a chemiluminescent emission 49 50 which is measured by a photomultiplier. Results are determined via a calibration 51 52 curve.24 Holick’s definitions of vitamin D status were used: 25(OH)D ≥ 30 ng/ml = 53 3 54 55 normal; 25(OH)D3 between 2029.9 ng/ml = insufficient vitamin D; 25(OH)D3 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 10 of 30

10 1 2 3 between 1019.9 ng/ml = vitamin D deficiency; 25(OH)D3 between 0 9.9 ng/ml = 4 5 severe vitamin D deficiency.4 6 7 8 9 10 Data collection and recording: 11 12 A standardized questionnaire was developed for data collection focusing on 13 14 demographic characteristics, types and categories of TB, symptoms, complications 15 For peer review only 16 and cigarette smoking status. Patients were diagnosed with DM either as a result of 17 18 19 this being already known (DM diagnosed at any time by a registered medical 20 21 institution and documented in the clinic notes) or as a result of FBG ≥ 7.0 mmol/L 22 23 being identified at the time of TB registration or at the time of TB diagnosis from 24 25 another health facility. PreDM was identified as a result of FBG 6.16.9mmol/L at 26 27 the time of TB registration based upon WHO definition.25 The questionnaire was 28 29 30 compiled by TB clinic staff and reviewed by the provincial and Union staff during 31

32 monitoring visits and before data entry and analysis. http://bmjopen.bmj.com/ 33 34 35 36 Data analysis and statistics: 37 38 39 Individual patient data were entered to an EXCEL file by the principal 40 on September 30, 2021 by guest. Protected copyright. 41 investigator. Nonparametric tests (KruskalWallis H) were used to compare median 42 43 levels of vitamin D [25(OH)D3] between groups. Comparisons of characteristics of 44 45 TB patients between those with normal FBG, preDM and DM and categorical 46 47 comparisons of various 25(OH)D levels between patients with DM, preDM and 48 3 49 50 normal FBG were carried out using the chi square test. Relationships between 25 51 52 (OH)D3 and associated exposure variables were evaluated with odds ratios (OR) and 53 54 95% confidence intervals. We selected variables with unadjusted ORs for which the P 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 11 of 30 BMJ Open

11 1 2 3 value was <0.05 and included these in a multivariate logistic regression model using 4 5 SPSS 13.0 (SPSS Inc., Chicago, IL, USA). Levels of significance were set at 5%. 6 7 8 9 10 Ethics approval: 11 12 The research proposal was reviewed and approved by the provincial 13 14 authorities in the implementing sites. The Ethics Advisory Group, International Union 15 For peer review only 16 Against Tuberculosis and Lung Disease, Paris, France, formally approved this study 17 18 19 (EAG number: 102/15). A written informed consent was obtained from participants. 20 21 22 23 24 25 26 27 28 29 30 31

BMJ Open Page 12 of 30

12 1 2 3 RESULTS 4 5 There were 317 TB patients consecutively registered in this study. Of these, 11 6 7 were excluded because of insufficient volumes of serum or absence of key 8 9 10 information variables. Of 306 eligible TB patients, 96 had smearpositive pulmonary 11 12 TB, 187 had smearnegative pulmonary TB and 23 had EPTB. There were 215 males 13 14 and 91 females, aged from 1892 years with a mean of 51.5 years. Of 306 TB patients, 15 For peer review only 16 95 had normal FBG, 83 had preDM and 128 had DM. The group of TB patients with 17 18 19 normal FBG was younger (P<0.001), had a higher proportion with rural citizenship 20 21 (P=0.026), a lower proportion with retreatment TB (P=0.016) and a higher proportion 22 23 registered in the months of May to October (the warm season) compared with the 24 25 DMTB group: there no other statistically significant differences between the groups 26 27 (Table 1). 28 29 30 Serum Vitamin D [25(OH)D3] levels in TB patients with normal FBG were 31

32 higher than those found in preDMTB patients and DMTB patients (Table 2). The http://bmjopen.bmj.com/ 33 34 proportions of patients with vitamin D deficiency or severe vitamin D deficiency were 35 36 61.1% for those with TB and normal FBG, 74.7% for those with preDM and 83.6% 37 38 39 for those with DM (P<0.001) The median serum vitamin D levels and their 40 on September 30, 2021 by guest. Protected copyright. 41 distributions in general were similar between Daan county (the rural poverty county) 42 43 and the other sites (P>0.05), although there was a higher proportion of patients with 44 45 normal vitamin D levels residing in areas classified as nonpoverty (P=0.032) 46 47 (Figure 1). 48 49 50 Baseline characteristics of TB patients, including those with normal FBG, pre 51 52 DM and DM in relation to having vitamin D deficiency / severe deficiency are shown 53 54 in Table 3. After adjusting for confounders with logistic regression, the main 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 13 of 30 BMJ Open

13 1 2 3 characteristics associated with having a higher risk of vitamin D deficiency were the 4 5 month of TB registration (November to April – the cold months) and having DM. 6 7 8 9 10 DISCUSSION 11 12 Our study presents findings of vitamin D status among different types of TB 13 14 patients and in relation to their DM status. 15 For peer review only 16 First, the median level of vitamin D in TB patients with normal FBG at 17 18 26 19 16.1ng/ml was lower than that found in the general population (26.3ng/ml), and this 20 6, 10, 20 21 finding is similar to that reported from other studies. Once antiTB treatment is 22 23 initiated, serum vitamin D levels may be further depressed through the action of drugs 24 25 such as rifampicin which increase liver metabolism,27 and TB clinic staff need to be 26 27 made aware of this fact. 28 29 30 Second, serum vitamin D levels in TB patients with preDM and with DM 31

32 were significantly lower than in TB patients with normal FBG, with a similar trend http://bmjopen.bmj.com/ 33 34 having also been found in a small scale study in urban settings of China.10 This was 35 36 mirrored by the higher proportions of preDM and DM patients with vitamin D 37 38 39 deficiency and severe deficiency compared with TB patients who had normal glucose 40 on September 30, 2021 by guest. Protected copyright. 41 levels. However, these are not consistent observations with a report in India finding 42 43 no significant differences in mean vitamin D levels between groups with TB, DM or a 44 45 combination of both.19 Both TB and DM can be associated with suboptimal 46 47 nutritional intake, but in this observational crosssectional study we were unable to 48 49 50 attribute any causal association as to whether low vitamin D status increases the risk 51 52 of TB or DM or whether DM or TB increases the risk of vitamin D deficiency. 53 54 Third, of the 6 sites participating in this study, one was located in an area 55 56 characterised as a State Poverty County streamlined by China State Council. Here, 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 14 of 30

14 1 2 3 there was a significantly lower proportion of TB patients with normal vitamin D status, 4 5 although the general distribution of vitamin D status from normal to severe deficiency 6 7 was no different compared with the other sites. It is unclear then whether there is any 8 9 10 need for interventions focused on improving nutritional intake or fortifying food with 11 12 vitamin D supplements in this particular area and this needs further discussion. 13 14 Fourth, our study highlighted certain characteristics associated with vitamin D 15 For peer review only 16 deficiency in TB patients. TB patients registered in the colder months of the year 17 18 19 (November to April) had a higher risk of vitamin D deficiency compared with those 20 21 registered in the warmer months (May to October). This is similar to findings 22 23 observed in Malawi,12 and is possibly due to less sunlight and more time spent 24 25 indoors during winter months.28 The study sites were also situated in the northeast of 26 27 China with where the winter season is long and very cold winter season. The prices of 28 29 30 vegetable, fruit and other food are far more expensive than those in the summer time, 31

32 and this may lead to reduced nutritional intake, particularly for poor people. http://bmjopen.bmj.com/ 33 34 Reports from Viet Nam and Afghanistan have found gender differences in 35 36 vitamin D status to be important for TB patients.29, 30 but this was not observed in our 37 38 39 study, and this may be due to a relatively higher general nutritional status for both 40 31 on September 30, 2021 by guest. Protected copyright. 41 men and women in the Chinese population. 42 43 Our study also found that DM was an independent risk factor for vitamin D 44 45 deficiency after adjusting for other baseline characteristics. A recent study in China 46 47 found that older age, a family history of DM and being overweight or underweight 48 49 50 were associated with preDM and TB, with preDM itself being a risk factor for the 51 32 52 development of DM. Both DM and TB are associated with impaired nutritional 53 33 54 intake and this might result in low levels of 25(OH)D3. Among our study subjects, 55 56 the age of TB patients with DM was generally older than that of TB patients with 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 15 of 30 BMJ Open

15 1 2 3 normal FBG. A previous study in China showed that elderly TB patients experienced 4 5 long delays in accessing health services and this might also contribute to poor vitamin 6 7 D status.34 8 9 10 The strengths of our study were the large number of TB patients enrolled and 11 12 the consecutive registration process to prevent selection bias. There were, however, 13 14 some limitations. We took blood samples for measurement of 25(OH)D3 immediately 15 For peer review only 16 after the diagnosis of TB without collecting the patient’s history. We therefore do not 17 18 19 know the duration of symptoms or signs of TB, a factor which might have impacted 20 21 on serum vitamin D levels. The presence or absence of comorbidities depended on 22 23 medical documentation or patient selfreports, so we may have overlooked some 24 25 conditions impacting on vitamin D status because we did not actively screen for them. 26 27 In addition, we only measured 25(OH)D levels at one time before antiTB treatment 28 3 29 30 and therefore do not know whether these levels changed during the six months course 31

32 of antiTB treatment. http://bmjopen.bmj.com/ 33 34 35 36 CONCLUSION 37 38 39 Serum vitamin D levels were lower in TB patients with preDM and DM 40 on September 30, 2021 by guest. Protected copyright. 41 compared with those whose FBG is normal. Vitamin D levels are also affected by 42 43 certain baseline characteristics, with vitamin D deficiency being significantly higher 44 45 in those with DM and those registered in the colder months of the year. TB 46 47 Programmes need to pay more attention to vitamin D status in their patients, 48 49 50 particularly during the cold season and if there is coexisting DM or preDM. 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 16 of 30

16 1 2 3 ACKNOWLEDGEMENTS 4 5 We thank all the staff at the six TB clinics and hospitals for their support in the blood 6 7 sample collection, storage and managing patients. 8 9 10 11 12 CONFLICTS OF INTEREST 13 14 No conflicts declared. 15 For peer review only 16 17 18 19 CONTRIBUTORS 20 21 Xin Zhao (XZ), Yan Lin (YL) and Yanli Yuan (YY) proposed the research idea. 22 23 Anthony D Harries (ADH), Riitta A Dlodlo (RAD) and Yan Lin (YL) contributed to 24 25 the study design. Xin Zhao (XZ) worked on laboratory analysis and quality assurance. 26 27 Wanli Kang (WK) and Yan Lin (YL) conducted data entry and statistics analysis. 28 29 30 Tiejuan Zhang (TZ), Jianjun Ma (JM), Yunlong Bai (YB), Yunlong Wang (YW) and 31

Page 17 of 30 BMJ Open

17 1 2 3 REFERENCES 4 5 1. World Health Organization. WHO Report 2016. Global Tuberculosis Control 6 7 2015. Available at www.who.int/tb/publication/global_report/en. 8 9 10 2. Frieden T. Toman’s Tuberculosis. Second Edition. 2004. World Health 11 12 Organization, Geneva. 13 14 3. Zittermann A. Vitamin D in preventive medicine: are we ignoring the evidence? 15 For peer review only 16 Bri J Nutr 2003: 89 (5): 55272. Doi: 10.1079/BJN2003837. 17 18 19 4. Holick MF, Chen TC. Vitamin D deficiency: a worldwide problem with health 20 21 consequences. The American J of Clinical Nutr 2008; 87 (4): 1080 s1086s. 22 23 5. Chesdachai S, Zughaier SM, Li H et al. The effects of firstline anti 24 25 tuberculosis drugs on the actions of vitamin D in human macrophages. J of 26 27 Clinical & Translational Endocrinology 2016; 6: 2329. Doi: 28 29 30 org/10.1016/j.jcte.2016.08.005 31

32 6. Sutaria N, Liu CT, Chen TC. Vitamin D status, receptor gene polymorphisms, http://bmjopen.bmj.com/ 33 34 and supplementation on tuberculosis: A systematic review of casecontrol 35 36 studies and randomized controlled trials. J of Clinical & Translational 37 38 39 Endocrinology 2014; 1: 151160. Doi: org/10.1016/j.jcte.2014.08.001 40 on September 30, 2021 by guest. Protected copyright. 41 7. Zeng J, Wu G, Yang W et al. A serum vitamin D level <25nmol/L pose high 42 43 tuberculosis risk: a metaanalysis. PLOS ONE 2015. Doi: 44 45 10.1371/journal.pone.0126014. 46 47 8. ArnedoPena A, JuanCerdan JV, RomeuGarcia A et al. Vitamin D status and 48 49 50 incidence of tuberculosis among contacts of pulmonary tuberculosis patients. Int J 51 52 Tuberc Lung Dis 2014; 19 (1): 6569. http://dx.doi.org/10.5588/ijtld.14.0348. 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 18 of 30

18 1 2 3 9. Davies PO. Vitamin D and tuberculosis: more effective in prevention than 4 5 treatment? Int J Tuberc Lung Dis 2015; 19 (8): 876. 6 7 http://dx.doi.org/10.5588/ijtld.15.0506. 8 9 10 10. Zhan Y, Jiang L. Status of vitamin D, antimicrobial peptide cathelicidin and T 11 12 helperassociated cytokines in patients with diabetes mellitus and pulmonary 13 14 tuberculosis. Exper and Thera medi 2015; 9: 1116. Doi: 10.3892/etm.2014.2042. 15 For peer review only 16 11. Wang Q, Ma A, Christian Bygbjerg Ib et al. Rationale and design of a 17 18 19 randomized controlled trial of the effect of retinol and vitamin D supplementation 20 21 on treatment in active pulmonary tuberculosis patients with diabetes. BMC 22 23 Infectious Dis 2013; 13: 10421. http://www.biomedcentral.com/1471 24 25 2334/13/104. 26 27 12. Sloan DJ, Mwandumba HC, Kamdolozi M et al. Vitamin D deficiency in 28 29 30 Malawian adults with pulmonary tuberculosis: risk factors and treatment 31

32 outcomes. Int J Tuberc Lung Dis 2015; 19 (8): 90411. http://bmjopen.bmj.com/ 33 34 http://dx.doi.org/10.5588/ijtld.15.0071. 35 36 13. Xia J, Shi L Zhao L et al. Impact of vitamin D supplementation on the outcome of 37 38 39 tuberculosis treatment: a systematic review and metaanalysis of randomized 40 on September 30, 2021 by guest. Protected copyright. 41 control trials. Chinese medical J 2014; 127 (17): 312734. Doi: 42 43 10.3760/cma.j.issn.03666999.20140702. 44 45 14. Stevenson CR, Critchley JA, Forouhi NG, et al. Diabetes and the risk of 46 47 tuberculosis: a neglected threat to public health. Chronic Illness 2007; 3: 228 – 48 49 50 245. 51 52 15. Jeon CY, Murray MB. Diabetes mellitus increases the risk of active 53 54 tuberculosis: a systematic review of 13 observational studies. PLoS Medicine 55 56 2008; 5: e152. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 19 of 30 BMJ Open

19 1 2 3 16. Mathieu C, Gysemans C, Giulietti A, Bouillon R. Vitamin D and diabetes. 4 5 Diabetologia 2005; 48: 124757. 6 7 17. Zhang X, Ma A, Sun S et al. Proteomic analysis of plasma in adult active 8 9 10 pulmonary tuberculosis patients with diabetes mellitus. Clin Lab 2015; 61 (10): 11 12 148190. 13 14 18. LopezLopez N, GonzalezCuriel I, CastanedaDelgado J et al. Vitamin D 15 For peer review only 16 supplementation promotes macrophages antimycobacterial activity in type 2 17 18 19 diabetes mellitus patients with low vitamin D receptor expression. Microbes 20 21 Infect 2014; 16 (9): 75561. Doi: 10. 1016/j.micinf. 2014. 06. 010. 22 23 19. Chaudhary S, Thukral A, Tiwari S et al. Vitamin D status of patients with type 2 24 25 diabetes and sputum positive pulmonary tuberculosis. Indian J Endocriol Matab 26 27 2013; 17 (Suppl 3): S6703. Doi: 10.4103/22308210.123564. 28 29 30 20. Zhu H, Cheng Q, Gan J et al. Vitamin D status in healthy man and woman in 31

32 Shanghai. Chinese J Oste & Bon Min 2010; 3 (3): 157163. Doi: http://bmjopen.bmj.com/ 33 34 10.3969/j.issn.16742591.2010.03.002. 35 36 21. Wang S, Shen G, Jiang S et al. Nutrient status of vitamin D among Chinese 37 38 39 children. Nutrients 2017; 9 (4): doi: 10.3390/nu9040319. 40 on September 30, 2021 by guest. Protected copyright. 41 22. Ministry of Health and China CDC. Guideline of national TB control program. 42 43 Beijing: The Peking Union Medical College Publishing House; 2008. 44 45 23. World Health Organization. Treatment of Tuberculosis Guidelines. Fourth 46 47 Edition, 2009. WHO, Geneva, Switzerland. WHO/HTM/TB/2009.420. 48 49 50 24. Roche. Vitamin D3 (25OH)Cobas operational menu. V2 English, Roche, 51 52 200708, Switzerland. 53 54 25. World Health Organization and International Diabetes Federation. Definition 55 56 and Diagnosis of Diabetes Mellitus and Intermediate Hyperglycaemia. At 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 20 of 30

20 1 2 3 www.who.int/diabetes/publications/diagnosis_diabetes2006/en/. Geneva, 4 5 Switzerland: WHO, 2006. 6 7 26. Wei QS, Chen ZQ, Tan X et al. Relation of age, sex and bone mineral density 8 9 10 to serum 25hydroxyvitamin D levels in Chinese women and men. Orthp Surg 11 12 2015; 7 (4): 3439. Doi: 10.1111/os.12206. 13 14 27. Bikle DD. What is new in vitamin D: 20062007. Current Opinion in 15 For peer review only 16 Rheumatology 2007; 19: 3838. 17 18 19 28. Sedrani SH, Elidrissy AW, El Arabi KM. Sunlight and vitamin D status in normal 20 21 Saudi subjects. Ame J Clin Nutr 1983; 38: 12932. 22 23 29. Sarin P, Duffy J, Mughal E et al. Vitamin D and tuberculosis: review and 24 25 association in three rural provinces of Afghanistan. Int J Tuberc Lung Dis 26 27 2016; 20 (3): 3838. http://dx.doi.org/10.5588/ijtld.15.0303. 28 29 30 30. Hopham LT, Nguyen ND, Nguyen TT et al. Association between vitamin D 31

32 insufficiency and tuberculosis in a Vietnamese population. BMC Infect Dis http://bmjopen.bmj.com/ 33 34 2010; 10: 30610. 35 36 31. World Health Organization. WHO Report 2016. World Health Statistics 2016. 37 38 39 Available at www.who.int/gho/publications/world_health_statistics/en. 40 on September 30, 2021 by guest. Protected copyright. 41 32. Wang Q, Ma A, Han X et at. Prevalence of type 2 diabetes among newly detected 42 43 pulmonary tuberculosis patients in China: A community based cohort study. 44 45 PLOS One 2013; 8(12): e82660. Doi: 10.1371/journal.pone.0082660. 46 47 33. Wang Q, Ma A, Han X et at. Is low serum 25hydroxyvitamin D a possible link 48 49 50 between pulmonary tuberculosis and type 2 diabetes? Asia Pac J Clin Nutr 2017; 51 52 26 (2): 2416. Doi: 10.6133/1pjcn.032016.02. 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

Page 21 of 30 BMJ Open

21 1 2 3 34. Lin Y, Enarson DA, Chiang CY et al. Patient delay in the diagnosis and 4 5 treatment of tuberculosis in China: findings of case detection projects. Public 6 7 Health Action 2015; 5 (1): 6569. Doi: org/10.5588/pha.14.0066. 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

BMJ Open Page 22 of 30

22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

value 0.82 0.82 0.03 0.60 0.02 0.11 P <0.001 <0.001

χ 0.38 7.30 2.80 8.29 21.38 21.38 10.40

128) DM DM 8(6.3) 8(6.3) N= 10(7.8) ( 93(72.7) 93(72.7) 35(27.3) 59(46.1) 27(21.1) 32(25.0) 89(69.5) 89(69.5) 39(30.5) 13(10.2) 62(48.4) 53(41.4) 54(42.2) 74(57.8) 37(28.9) 83(64.8) http://bmjopen.bmj.com/ BMJ Open

83)

N= 7(8.4) 7(8.4) ( 9(10.9) Pre-DM Pre-DM 57(68.7) 57(68.7) 26(31.3) 20(24.1) 34(41.0) 29(34.9) 24(28.9) 59(71.1) 26(31.3) 48(57.8) 64(77.1) 64(77.1) 19(22.9) 32(38.6) 19(22.9) 25(30.1) No. (%) of the TB patients (%)ofNo. the on September 30, 2021 by guest. Protected copyright.

95)

N= 6(6.4) 6(6.4) ( 84(88.4) 84(88.4) 11(11.6) 44(46.3) 12(12.6) 22(23.2) 17(17.9) 69(72.6) 69(72.6) 26(27.4) 32(33.7) 42(44.2) 21(22.1) 25(26.3) 70(73.7) 33(34.7) 56(58.9) Normal FBG NormalFBG For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only

0 1-9 1-9 <40 <40 ≥60 ≥20 ≥20 New EPTB Male Male Rural 40-59 40-59 10-19 10-19 Urban Urban Female Female Positive Positive Negative Negative Retreatment Retreatment

Sex Age Age smear Smoking Smoking Residence Residence cigarette/day cigarette/day Type/Sputum Type/Sputum Category of TB TB of Category Table 1: Characteristics of the 306 recruited TB patients TB 306 patients recruited ofthe Characteristics 1: Table Characteristics

Page 23 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from Page 24 of 30 24

0.89 0.89 0.046

6.16 0.25

39(30.5) 39(30.5) 89(69.5) 25(19.5) 103(80.5) 103(80.5)

http://bmjopen.bmj.com/ BMJ Open

34(41.0) 34(41.0) 49(59.0) 14(16.9) 69(83.1) on September 30, 2021 by guest. Protected copyright.

44(46.3) 44(46.3) 51(53.7) 18(18.9) 77(81.1) For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only

No No October October Yes April April November- May-

registration registration Month of TB of Month Poverty area Poverty FBG = fasting blood glucose; Pre-DM = pre-diabetes mellitus; DM = diabetes mellitus; TB = tuberculosis; EPTB = extra-pulmonary tuberculosis extra-pulmonarytuberculosis = =EPTB tuberculosis; TB mellitus; DMdiabetes =mellitus; pre-diabetes =glucose; Pre-DM fastingblood = FBG year of the months cold the = April – ofyear; November the months = warm the October – May

value P

<0.01 <0.01

128 = 12.1 <0.01

69 (53.9) (53.9) 69 18 (14.1) (14.1) 18 0.63 N with DM with TB patients patients TB

=83 12.6 N

http://bmjopen.bmj.com/ TB patients patients TB with pre-DM pre-DM with BMJ Open

=95

Vitamin D statusD Vitamin patients with TB No (%) with normal level normal(≥30ng/ml) with No(%) No (%) with deficiency (10-19.9ng/ml) deficiency(10-19.9ng/ml) with No(%) 39 (41.1) 34 (41.0) FBG = fasting blood glucose; pre-DM = pre diabetes mellitus; DM = diabetes mellitus; TB = tuberculosis = TB mellitus; = DM diabetes prediabetes mellitus; glucose;= pre-DM fastingblood = FBG of25-(OH)D measurements by statusD determined Vitamin Median level (ng/ml)level Median (20-29.9ng/ml) insufficiency with No(%) 14 (14.7) deficiencysevere ( with No(%) 13 (15.7) Table 2: Vitamin D status in TB patients in China, stratified by those with normal FBG, pre-DM and DM FBG,and pre-DM withthosenormal by stratified TBChina, inin patients D statusVitamin 2: Table

Page 25 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from Page 26 of 30 26 0.438 0.230 P value value P

Multivariate Multivariate OR adjusted CI) (95%

Reference (0.67-2.54) 1.30 (0.75-3.34) 1.58

Univariate OR Univariate (95%CI) Reference (0.71-2.12)1.22 Reference (0.92-3.24) 1.72 (1.21-4.90) 2.43 Reference 1.11(0.65-1.91) Reference (0.88-2.64) 1.52 (0.46-3.59) 1.29 Reference (0.77-2.95) 1.51 Reference (0.86-3.72) 1.79 (0.79-2.77) 1.48 2.71 (0.98-7.49) http://bmjopen.bmj.com/ BMJ Open (73.9) 17 (80.0) 52 (68.1) 92 (79.3) 46 (75.9) 60 (85.3) 29 No.(%) with No.(%)with VitaminD deficiency (71.4)65 (75.3) 162 41(63.1) (74.6) 103 (80.6) 83 75(72.8) (74.9) 152 (68.8) 66 (77.0)144 (72.6) 175 on September 30, 2021 by guest. Protected copyright. 91 91 65 96 23 65 79 34 Total Total N=306 215 138 103 103 203 187 241 135 58 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only

60 60 20

<40 40-59 ≥ Urban Rural Positive Negative EPTB New Retreatment 0 1-9 10-19 ≥ Female Male Characteristics Category of TB TB of Category Sex Age Residence Type/Sputum smear Smoking cigarette/day Table 3: Characteristics of TB patients in China in relation to vitamin D deficiency and severe deficiency (<20ng/ml) (<20ng/ml) severeanddeficiency to D deficiency vitamin China relation in in patients ofTB Characteristics 3: Table

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from 27 0.006 0.090 0.003 Reference (1.25-3.69) 2.15 Reference 1.77(0.91-3.43) (1.40-5.15) 2.68

Reference (1.46-4.17) 2.47 Reference (0.99-3.59)1.88 (1.74-6.06) 3.25

http://bmjopen.bmj.com/ BMJ Open 74 (63.2) (63.2) 74 (81.0) 153 (61.1) 58 (74.7) 62 (83.6) 107

on September 30, 2021 by guest. Protected copyright. =7.910, P=0.341, indicating the good of fit of logistic regression models is good. is good. models regression logistic of of fit good the indicating P=0.341, =7.910, 95 95 83 2

117 189 128

χ For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only May-Oct Nov-April Normal Pre-DM DM

CI=ratio;intervals. confidence odds glucose;= OR blood fasting FBG= tuberculosis; extra-pulmonary = EPTB tuberculosis; = TB

TB of Month registration level FBG

P value is shown for the multivariate adjusted OR adjusted the isfor shown multivariate Pvalue year of the months cold the = April – ofyear; November the months = the warm October – May fit: of good show Hosmer –Leme Page 27 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from

BMJ Open Page 28 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

32 http://bmjopen.bmj.com/ 33 Figure 1: Vitamin D status of TB patients in the poverty rural area compared with other non-poverty areas in 34 China 35 36 67x53mm (300 x 300 DPI) 37 38 39 40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from 7, 8, 7, and 9 11-12 7, 8, 9 7, 10and 5 and 5 6 9 (cross-sectional9 study) Not relevant Not 7 7 7 and 7 8 1, 6 1, 7 and

6 6

10 10 Not relevant Not 1 1 3 3 Not relevant Not Not relevant Not Not relevant Not Reported Reported on# page

http://bmjopen.bmj.com/ BMJ Open —Give the—Give criteria, eligibility and sourcesthe methodsand of selection of participants on September 30, 2021 by guest. Protected copyright. —Give the—Give eligibility criteria, and sourcesthe methodsand case of ascertainment and control —If —If applicable, explain of matching how cases and controls addressed was —For matched studies,matching criteria give theand number of controlsper case —Give eligibilitythe criteria, and sourcesthe methods and of selectionof participants. Describe —If applicable, howexplain loss to addressedfollow-up was —For matched matching criteriastudies, give number of exposedand unexposed and Checklist for cohort, case-control, and cross-sectional studies (combined) (combined) studies cross-sectional and case-control, for cohort, Checklist For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Cohort study Cohort study Cohort study

) Describe statistical all methods, including thosecontrol confoundingused for to ) ) Indicate)the study’scommonly design a with inused term the the ortitle abstract ) ) Describe any methods used tosubgroups examine interactionsand ) ) inProvide the abstract informativean and balanced summaryof what was done and what was found ) Explain missinghow data were addressed a a a d c b b b For eachof interest,For variable sourcesgive of data details and of methods of assessment (measurement). Describe comparability comparability of assessment isthere methods more if thanone group

criteria, if criteria, if applicable collection methods of follow-up Case-control study selection. Give the rationalethe selection. Give forthe of caseschoice and controls Cross-sectional study ( ( ( ( ( Recommendation and why and Case-control Case-control study For peer review Case-control study only 3 3 objectives, State includingspecific pre-specifiedany hypotheses 1 1 ( 5 5 Describe setting,the locations, relevantand dates, including periods exposure,of recruitment, follow-up, and data 9 9 Describe efforts to any addresssourcespotential of bias 6 6 ( 7 7 define Clearly outcomes, exposures,all potential predictors, confounders, and effect modifiers. diagnostic Give 4 4 Present key of elements study design early in paperthe 10 10 Explain studythe how size was arrived at 12 12 ( STROBE 2007 (v4) checklist of items to be included in reports of observational studies in epidemiology* epidemiology* in studies observational of reports in be included to of items checklist (v4) 2007 STROBE Item Item # Statistical Statistical methods Quantitative variables 11 Explain quantitativehow were variables handled in analyses.the If applicable, describe which groupings chosen were Study size Bias Data sources/Data measurement 8* Variables Participants Setting Study design Methods Objectives Background/rationale 2 Explain scientific the background rationale investigation the for and being reported Introduction Title andTitle abstract Section/Topic

Page 29 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2017-017557 on 25 September 2017. Downloaded from Page 30 of 11 and 12 and 11 12 12 14 and 15 and 14 12, 13 1412, and Not relevant Not Not relevant Not 11 11 Table 1 Table 11 and 12 and 11 Not relevant Not Not relevant Not Not relevant Not Not relevant Not

12, 13 1512, and

Not relevant Not Not relevant Not

Not relevant Not 11 and 12 and 11 Not relevant Not

http://bmjopen.bmj.com/ BMJ Open Report numbers eventsof outcome summary measuresor If describe applicable, analytical methods account taking of sampling strategy — Report numbers in each category,exposure summaryor measures of exposure on September 30, 2021 by guest. Protected copyright. —Summarise —Summarise follow-up(eg, time average and total amount) —Report numbers eventsof outcome or summary measures over time sectional studysectional - For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Cohort study ) Report) category boundaries when continuous variables were categorized ) Give unadjusted estimates ifapplicable, and, confounder-adjusted estimatestheir and precision95% (eg, ) Describesensitivity any analyses ) If )relevant, If considertranslating estimates of relative risk forintorisk meaningful absolute periodtime a c b a e from similar studies,from otherand relevant evidence and magnitudeand of potential any bias Cross-sectional Cross-sectional study— Case-control Case-control study— Cohort study potential confounderspotential confirmed eligible,confirmed included in thecompleting study, follow-up,analysed and confidence interval). confidence Make clear which confounders were adjustedthey wereincluded andfor why Cross ( For peer review only presentthe which article is based ( (

(c) number Indicate (b) participants of missingwith eachfor data of interest variable Consider (c) diagram use flow of a reasons Give (b) non-participationfor stage each at 18 18 Summarise resultskey to study objectivesreference with 17 17 other Report analyses done—eg analyses of subgroups interactions, and sensitivity and analyses 20 20 Give cautiousinterpretationa of overall results considering objectives, limitations, multiplicity of resultsanalyses, 16 16 ( 21 21 Discuss the (external generalisability of validity) studythe results 22 22 sourcethe Give of funding rolethethe and of fundersfor presentthe if studyapplicable, originaltheand, for study on 19 19 Discuss of thelimitations taking intostudy, sourcesaccount biasof potential or imprecision. Discussboth direction 14* characteristics (a) Give participantsof study (eg demographic, clinical, social) and information on exposures and 15* 13* (a) numbers Report of individuals each at stage of study—eg numbers potentially eligible, examined eligibility,for An Explanation and Elaboration article discusses itemeach checklist and gives methodological backgroundand published examples of transparent reporting. The STROBE checklist is best used in thiswith conjunction (freely available article onsites the Medicine of PLoS Web http://www.plosmedicine.org/,at Annals of Internal Medicine at http://www.annals.org/, Epidemiologyhttp://www.epidem.com/). at and Information on InitiativeSTROBE the is www.strobe-statement.org.available at information*Give separately cases for and controls studiesin case-controlfor applicable, if and, exposedunexposed and groups in cohort and cross-sectional studies. Note: Funding Other information Generalisability Interpretation Limitations Key Key results Discussion Other analyses

Main resultsMain

Outcome Outcome data

Descriptive data

Participants Results

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60