Assessment of tuberculosis contact investigation in Shanghai, China: An 8-year cohort study

Accepted Manuscript Assessment of tuberculosis contact investigation in Shanghai, China: An 8-year cohort study Qi Jiang, Liping Lu, Jie Wu, Chongguang Yang, Ravi Prakash, Tianyu Zuo, Qingyun Liu, Jianjun Hong, Xiaoqin Guo, Qian Gao PII:

S1472-9792(17)30298-6

DOI:

10.1016/j.tube.2017.10.001

Reference:

YTUBE 1630

To appear in:

Tuberculosis

Received Date: 5 July 2017 Revised Date:

27 September 2017

Accepted Date: 1 October 2017

Please cite this article as: Jiang Q, Lu L, Wu J, Yang C, Prakash R, Zuo T, Liu Q, Hong J, Guo X, Gao Q, Assessment of tuberculosis contact investigation in Shanghai, China: An 8-year cohort study, Tuberculosis (2017), doi: 10.1016/j.tube.2017.10.001. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

3

Qi Jiang1,2 #, [email protected]

4

Liping Lu3 #, [email protected]

5

Jie Wu4, [email protected]

6

Chongguang Yang1, 5, [email protected]

7

Ravi Prakash1, [email protected]

8

Tianyu Zuo1, [email protected]

9

Qingyun Liu1, [email protected] Jianjun Hong3, [email protected]

11

Xiaoqin Guo3 *, [email protected]

12

Qian Gao1,2 *, [email protected]

M AN U

10

RI PT

2

Assessment of Tuberculosis Contact Investigation in Shanghai, China: An 8-year Cohort Study

SC

1

13 Affiliations:

15

1 Key Laboratory of Medical Molecular Virology, School of Basic Medical Sciences and

16

Shanghai Public Health Clinical Center, Fudan University, Shanghai, China 200032

17

2 Shenzhen Center for Chronic Disease Control, Shenzhen, China 518000

18

3 Songjiang District Center for Disease Control and Prevention, Shanghai, China 201620

19

4 Shanghai Municipal Center for Disease Control and Prevention, Shanghai, China 200336

20

5 Department of Epidemiology of Microbial Diseases, Yale School of Public Health, New

21

Haven, Connecticut, USA 06510

24

EP

23

# Qi Jiang and Liping Lu contributed equally to this work.

AC C

22

TE D

14

25

*Correspondence:

26

Dr. Qian Gao, Key Laboratory of Medical Molecular Virology, School of Basic Medicine,

27

Fudan University, Room 1001, Zhidao B., Dongan R. No.131, Shanghai, China 200032 Tel-

28

+86 021 5423 7195 Fax- +86 021 5423 7195 E-mail- [email protected]; Xiaoqin Guo

29

E-mail- [email protected]

30 31

Word count: Abstract: 206 words

Text: 2868 words

32

1

ACCEPTED MANUSCRIPT ABSTRACT

34

Background

35

Tuberculosis (TB) contact investigation has been observed as a useful programmatic

36

tool in active case finding. We collected data of contact cases to evaluate the

37

effectiveness of TB contact investigation programme in Shanghai, China.

38

Methods

39

Since 2009, we screened and followed up the close contacts of bacteria-positive TB

40

cases in Songjiang, Shanghai and calculated the incidence of the disease in close

41

contacts and confirmed the transmission by genotyping and sequencing.

42

Results

43

A total of 4584 close contacts of 1765 contagious TB index cases were followed up

44

for an average of 4 years. About 62 contacts (333/100 000, 95% CI: 256-428)

45

developed TB excluding 6 co-prevalent cases. The contact cases consisted 1.50%

46

(39/2592) of all the bacteria-positive cases in population. Transmission links were

47

confirmed in 60% (9/15) familial contacts and 22% (2/9) in non-familial contacts.

48

Source cases come from more than close contacts and both index and contact cases

49

created other secondary cases.

50

Conclusions

51

Familial contacts are more likely to acquire TB from the index, indicating the priority

52

of family members in TB contact investigation in China. However, most non-familial

53

contacts were infected from sources in the community and contact cases attributed

54

little to TB burden. Thus, active case finding should be strengthened in general

55

population.

SC

M AN U

TE D

EP

AC C

56

RI PT

33

57

KEYWORDS

58

Tuberculosis; Contact investigation; Transmission

2

ACCEPTED MANUSCRIPT 1. INTRODUCTION

60

World Health Organization (WHO) has recommended contact investigation as an

61

important part of tuberculosis (TB) control programme, considering its contribution in

62

early identification of active TB [1]. Contact investigation provides a better chance of

63

cure and reduction in further transmission of the disease [1], especially in low- and

64

middle-income countries [2]. Developed countries have routinely investigated

65

contacts to identify and treat persons with active TB or latent TB infection (LTBI) [3,

66

4]. Since 1976, after the publication of first guidelines on TB contact investigation,

67

American National Programmes were not expanded until publication of updated

68

guidelines in 2005, suggesting that investigation and follow-up should be initiated or

69

expanded after assigning priorities to contacts based on the characteristics of

70

individuals and the features of exposure [5]. However, contact investigations were

71

inconsistently carried out in high-burden resource limited settings on the basis of no

72

or poor standards [6,7]. In 2007, Chinese Ministry of Health, proposed a guideline on

73

investigation among close contacts of infectious TB cases. These include the role of

74

clinicians to educate new smear-positive TB cases about the importance of contact

75

investigation, to record basic information of close contacts, identify the persons who

76

are directly exposed to the index case including his/her family members, colleagues

77

and classmates. Any close contacts with suspected symptoms must be informed for

78

further examinations [8].

79

Despite the benefits, scale-up of contact investigation programme encountered

80

barriers in resource-limited areas and its cost-effectiveness remained controversial [9].

81

A systematic review summarized the work in low and middle-income countries

82

revealed a pooled yield of 4.5% (4.3-4.8%) among close contacts for active TB and

83

2.3% (2.1-2.5%) for bacteria confirmed TB [10]. In rural Malawi, it has been reported

84

that contact investigations contributed to less than 10% of new TB cases [11].

85

Molecular epidemiological tools have helped in establishing the transmission link

86

between the index and contact cases, if they share the same genotype. But,

87

unexpectedly, less than half of the incident contacts were confirmed infection by the

88

index cases [11,12].

89

In the present study, in order to assess the yield of TB contact tracing and its role in

90

active case finding of the disease, we collected the information of contact cases

91

investigated and additionally followed up them in Songjiang, Shanghai since 2009.

AC C

EP

TE D

M AN U

SC

RI PT

59

3

ACCEPTED MANUSCRIPT We confirmed the transmission from index to contact cases by whole genome

93

sequencing (WGS) and tried to identify the risks of transmission between index and

94

close contacts in a large metropolitan area in China.

AC C

EP

TE D

M AN U

SC

RI PT

92

4

ACCEPTED MANUSCRIPT 95

2. METHODOLOGY

96

2.1. Study populations

97

Songjiang District Center for Disease Control and Prevention (CDC) has established a

98

routine

99

bacteria-positive TB patients about their close contacts and continuously followed up

100

them for the period of 2009-2016. Suspected individuals with TB symptoms were

101

screened for chest X-ray (CXR) and sputum tests, based on the national guidelines [8].

102

Diagnosed cases were registered in TB management system of Songjiang CDC and a

103

code was assigned to them.

104

2.2. Contact Investigation and Contact Tracing

105

“Index case” was defined as a newly notified bacteria-positive (smear-positive or

106

culture-positive) TB patient. CDC collected the information (name, age, gender and

107

their relationship) of close contacts that had direct contact with the index cases

108

including their family members, colleagues and classmates [8]. Close contacts were

109

screened in the first month after the diagnosis of the index cases and were followed up

110

twice a year during 2009-2011 and once in 2012-2016. Index cases were enquired

111

about clinical manifestation of their close contacts by clinicians. Once TB symptoms,

112

like cough or fever appeared in any of the contacts, they were further examined for

113

chest X-ray and sputum tests for TB diagnosis and were assigned as “contact case”

114

and treated when diagnosed with active pulmonary TB. The routine screening process

115

did not include latent TB screening or prevention treatment among the contacts. It is

116

considered that immune response to M. tuberculosis generates on an average of 42

117

days after exposure [13] and contact cases diagnosed with TB less than or equal to 42

118

days after an index patient’s TB diagnosis were considered “co-prevalent” cases.

119

However, contacts diagnosed after the first month of screening were considered

120

“incident” cases.

121

2.3. Strains genotyping and WGS

122

As a routine work, the sputum samples were collected from all the suspected

123

pulmonary TB patients at clinics for smear and MGIT culture test. Culture-positive

124

samples were sent to Shanghai CDC for drug susceptibility testing and variable

125

number tandem repeats (VNTR) genotyping. A set of loci optimized with high

126

discriminatory power was chosen which includes nine general loci and

127

three-hypervariable loci [14]. Other patients in Songjiang harboring identical

tracing,

since

2009.

Clinicians

investigated

RI PT

contact

SC

of

AC C

EP

TE D

M AN U

network

5

ACCEPTED MANUSCRIPT genotypes (differing less than one locus) with index or contact cases were defined as

129

“clustered cases” and sequenced to confirm the transmission links using WGS.

130

M. tuberculosis isolates from index cases, contacts and clustered cases were subjected

131

for genomic DNA extraction using CTAB method [15]. WGS was performed

132

commercially at Yikon Genomics Co. (Jiangsu, China). The low quality bases at the

133

20 reads or length of less than 20 were first trimmed using Sickle

134

(https://github.com/ucdavis-bioinformatics/sickle) and were then mapped to the

135

reference

136

(https://sourceforge.net/projects/bowtie-bio/files/bowtie2/2.2.9/).

137

(http://www.htslib.org/) was used for calling the single-nucleotide polymorphisms

138

(SNPs) and fixed mutations (frequency ≥ 75%) were identified using VarScan

139

(version 2.3.9, https://sourceforge.net/projects/varscan/).

140

Although, a population-based study in Shanghai [16] observed that 12 SNPs can be a

141

potential threshold to define recent transmission, which is also reported in other

142

settings [17]. Therefore, we used 12 SNPs or less to confirm the transmission between

143

strains isolated from index and contact cases, suggesting that the index case and

144

contact case are involved in the same transmission chain; otherwise, contact cases

145

were not infected by their index cases. The transmission chain was constructed on the

146

basis of SNPs of strains using an in-house Perl script according to the method

147

described by Walker TM [18], as the backwards SNP accumulation is unlikely in M.

148

tuberculosis, so we reconstructed the putative transmission directions based on the

149

accumulating SNPs.

150

2.4. Statistical analysis

151

Demographic, clinical and genetic data were analyzed using Stata/SE 13.1 (StataCorp,

152

USA). Student’s T tests or non-parametric tests were used for comparison of numeric

153

data and Pearson χ2 tests for the proportions of categorical variables. Survival analysis

154

and competing risk regression were used to determine the cumulative incidence with

155

95% confidence intervals (CI) and risk factors of TB with sub distribution hazard risk

156

(sHR). A P value of <0.05 was considered as statistically significant.

H37Rv

(GenBank:

AL123456)

with

Bowtie

SAM

2

tools

AC C

EP

TE D

M AN U

SC

genome

RI PT

128

157

6

ACCEPTED MANUSCRIPT 3. RESULTS

159

3.1. Baseline Characteristics of the Contact Cohort

160

From 2009 to 2015, a total of 3914 tuberculosis cases were reported in Songjiang

161

district, out of which 58.4% (2287/3914) were bacteria-positive (smear- or

162

culture-positive). Out of all the infectious cases (index cases), 2055 (89.9%,

163

2055/2287) provided 5291 contacts in the contact investigation. Approximately, 44

164

retreated patients repeating 92 contacts and 24 patients sharing the same contacts

165

were excluded from the study. Finally, 5121 contacts from 1987 index cases were

166

enrolled in the study. Most of the index cases (82.8%, 1645/1987) did not generate

167

more than three contacts. Apart from those who did not provide the information of

168

relationship, familial contacts consisted of the majority (80.5%, 4026/5002) of the

169

contact cohort. An average age for contact cases was found to be 36 years, comprised

170

of 3.4% (175/5121) children less than five years and 9.2% (472/5121) were older than

171

60 years. Almost, half of the contacts (50.4%, 2582/5121) were male. By the end of

172

2016, 528 contacts (10.3%, 528/5121) were missing and 9 died during follow up and

173

were excluded from analysis. The remaining 4584 contacts of 1765 index patients

174

completed the follow-up study for contact tracing (Figure 1).

175

3.2. Incidence of Contact Cohort and Hazard Risk

176

We observed that 68 contact cases developed active TB disease. Out of these, 4 were

177

migrants and transferred out, with no record of bacterial infection, whereas, others

178

consisted 1.44% (64/4435) of TB cases in Songjiang during 2009-2016. A total of

179

39 (60.9%, 39/64) cases were bacteria-positive, contributing for 1.50% (39/2592)

180

bacteria-confirmed cases in the population, while, rest of the 25 cases contributed for

181

1.36% (24/1843) bacteria-negative cases.

182

Among the contact cases, six (8.8%, 6/68) began to cough within 42 days of the

183

diagnosis of their index cases and were identified in primary screening, whereas, the

184

other 62 cases developed TB later and were diagnosed in the follow-up. This resulted

185

in the cumulative risk of active TB among close contacts to 1.48% (95%CI

186

1.15%~1.88%), with 0.13% (95%CI: 0.05% to 0.28%) of co-prevalent TB and 1.35%

187

(95%CI: 1.04% to 1.73%) incident cases. Following up for an average of 4 years

188

(6,779,476 person-days), the annual incidence of close contacts was found to be

189

333/100 000 person per year (95% CI: 256 to 428), roughly ten folds than general

190

population in Songjiang.

AC C

EP

TE D

M AN U

SC

RI PT

158

7

ACCEPTED MANUSCRIPT Survival Analysis revealed a decline in the risk of TB every year with the incidence of

192

active TB among close contacts being 0.70%, 0.26%, 0.17% and 0.11% in the first

193

four years and a consistent figure of 0.11% in the following years, excluding the six

194

co-prevalent cases. Familial contacts didn’t gain higher incidence than non-familial

195

contacts (HR=0.554, 95%CI 0.276-1.117, P=0.099). Competing risk regression

196

observed that younger index created more contact cases (sHR=0.985, 95%CI

197

0.971-0.999; P=0.031). But, the incidence of close contacts between age groups or

198

genders did not reveal any significant difference. Previous TB history, delay in

199

diagnosis and smear status of the index cases did not show any risk for active TB

200

among close contacts.

201

3.3. Transmission network construction based on WGS

202

To confirm and gain insight into the transmission events between index and contact

203

cases, clinical isolates from bacteria-positive index and contact cases were obtained

204

and sequenced. Among 39 bacteria-positive pairs, only 24 pairs (61.5%) were

205

subjected to WGS for confirmation of TB transmission. The other 15 pairs (four

206

smear-positive culture-negative cases, six cases that did not provide strains and five

207

samples not recovered) could not be sequenced. Further stratification of

208

bacteria-positive pairs enrolled and excluded from the study did not reveal any

209

significant difference in demographic characteristics. Family members of the index

210

cases were observed to contribute for majority of contact cases (69.2%, 27/39).

211

The genetic distance for the isolates from the index and the contact cases was

212

calculated, out of which, 11 pairs (45.8%, 11/24) of contacts were confirmed for their

213

epidemiological links. The number of SNPs between the pairs is shown in Figure 2:

214

11 had 0~6 SNPs, 1 had 19 SNPs (Pair 19) and 12 had over 100 SNPs. Among family

215

contacts, 40.0% (95%CI 16-68, 6/15) of them were not infected by their index cases,

216

while, non-family contacts were more likely to be transmitted from unknown

217

exposure (77.8%, 95%CI 40-97, 7/9). It was observed that contacts with matched

218

strains developed active TB on an average of 10 (0-38) months after the diagnosis of

219

their index cases, while strain-unmatched contacts developed active TB within a time

220

of 25 (1-79) months, with no significant difference (z=-1.393, P=0.164).

221

To interpret transmission chain in community, we sequenced isolates that had the

222

same VNTR genotypes (up to one loci difference) with any of the index or contact

223

cases. With at most 12 SNPs in genetic distance between clinical strains, we

AC C

EP

TE D

M AN U

SC

RI PT

191

8

ACCEPTED MANUSCRIPT confirmed other clustered cases among 27% (3/11) genotype-matched pairs and 12%

225

(3/26) genotype-unmatched pairs and added up to 54% (13/24) contact cases whose

226

source cases were detected (Figure 3). For example, the index of Pair 3 was a

227

multidrug-resistant TB (MDR-TB) patient in 2009 and transmitted to a secondary

228

case in community (with 1 SNP difference), while her father developed TB five years

229

later and was observed being transmitted the disease by another case in 2014 instead

230

of her (with 7 SNPs difference). Both the genotype-matched and unmatched pairs

231

revealed several clustered cases, indicating uncontrolled transmission, which

232

continuously created secondary cases in following years in the community.

AC C

EP

TE D

M AN U

SC

RI PT

224

9

ACCEPTED MANUSCRIPT 4. DISCUSSION

234

From 2009 to 2016, we screened and traced 4584 close contacts of 1765 contagious

235

TB index to identify new active TB cases. On an average, following up 2.6 close

236

contacts per index for an average of four years revealed that only 1.53% (68/4435)

237

contact cases developed TB with an annual incidence of 370/100,000 among the

238

contact population. However, less then half (45.8%) of index-contact pairs were

239

ensured of creating close contact transmission. About 60.0% of familial contacts were

240

transmitted by the index, whereas, most (77.8%) of the incident non-familial contacts

241

were infected by other sources in community.

242

China has faced number of difficulties in contact investigation, which resulted due to

243

non-availability of proper definition and guidance, as well as less awareness for TB

244

prevention programs in general public. WHO defined a household contact as a person

245

who shared living space and a close contact that shared other enclosed space; such as

246

gathering or workplace during 3 months before current treatment of the index case

247

and recommended priority of contact investigation among children less than five

248

years and HIV infected people [2]. However, according to the definition of close

249

contacts in China [8], contact investigations are largely dependent on local

250

understanding and practice [19]. In our contact cohort, we could identify only 2.6

251

contacts per index case, which is consistent with other studies in China [20-21], but

252

much lower than other countries where this has been reported to be 5-12 contacts per

253

index [22-26]. Most of the contacts in our study were family members, while

254

non-familial contacts were neglected and could hardly be investigated due to patients’

255

privacy or due to fear of discrimination. Nevertheless, among pairs of index and

256

contact cases with confirmed transmission by WGS, mostly were family members.

257

Non-familial contacts were more likely to be infected beyond known household

258

contacts. In a national survey across 13 provinces in China, the TB detection rate in

259

familial contacts was also higher than non-familial contacts [27]. Due to the higher

260

yield of TB cases in familial contacts and the difficulty in questioning or following up

261

non-familial contacts, contact investigation should focus and intensify active case

262

finding among family members of TB cases in China.

263

Among the notified contact cases, nearly half of them (46%, 11/24) have confirmed

264

transmission link from their index cases and shared the same genotype, which is

265

higher than similar studies from Malawi (38%, 62/163) [11] and Vietnam (17%, 2/12)

AC C

EP

TE D

M AN U

SC

RI PT

233

10

ACCEPTED MANUSCRIPT [12]. Only, half (54%, 13/24) of the source cases of the contacts could be identified,

267

including 11 familial pairs and 2 pairs from community, indicating the existence of

268

many unknown infectious sources in the population. In a population-based cluster

269

analysis, familial contacts attributed to 6.1% of recent transmission and 33.5%

270

clustered cases were transmitted in community [28]. Brooks et al. also stated that

271

community transmission accounts for a higher proportion of contact cases than does

272

household transmission in high-burden settings [29]. In our study, both index and

273

contact cases were continually generating secondary cases in the community. Active

274

cases in close contacts could contribute only 1.44% of the total TB cases (1.50% of

275

bacteria-positive cases and 1.36% of bacteria-negative cases). Although, we could not

276

exclude the contribution of non-culture-positive infectious cases and those unknown

277

contacts of latent TB infection, still less than one third transmission links were

278

confirmed, if incidental cases are excluded in one month and would be even more less,

279

if co-prevalent cases were defined in six months [30]. Thus, contact investigation

280

could not find the majority of TB cases in this setting and attributed much less than

281

that we thought. A modeling study estimated both household and community

282

transmission for designing interventions [31], emphasized the impact of transmission

283

beyond household and the demand for active case finding in community, such as

284

population-based screening and routine referral in primary medical care.

285

The incubation period of tuberculosis varies from few weeks to few decades, but

286

contact investigation has been focused on examination at only one point of time

287

during diagnosis of index case in China [8]. However, we observed the majority of the

288

incident contact cases in the follow-up and the investigation in first month after the

289

diagnosis of the index case could detect only 7.8% (5/64) of them, suggested that

290

longitude incidence was continuously higher than the general population in the

291

following years. The genotype matched contact cases developed active TB up to 40

292

months after the diagnosis of the index. Therefore, we suggest a longer follow-up of

293

the close contacts, as recommended in Japan for at least two to three years [32].

294

Contact tracing has also been suggested as part of the TB prevention programme in

295

England in 2015 even with a low TB incidence setting [33]. Therefore, it is suggested

296

that family should undoubtedly take the most priority in contact investigation when

297

considering the higher possibility of household transmission.

298

Our study had several limitations. There were only 24 (60.0%, 24/40) culture-positive

AC C

EP

TE D

M AN U

SC

RI PT

266

11

ACCEPTED MANUSCRIPT pairs genotyped and WGS, which compressed the sample size. The passive follow-up

300

procedure could also be a limitation. The migratory behavior of the population

301

reduced the efficiency of contact naming and tracing.

302

In conclusion, contact tracing of family members is especially important. Even

303

though, the WGS has also shown that strains isolated from culture-positive

304

index-contact pairs may still differ 40% (95%CI 16-68). Thus, screening of

305

population or referral should be proposed in primary medical care to detect new

306

bacteria-positive cases in a TB-prevalent setting.

RI PT

299

307

Acknowledgements We thank clinicians of the community medical care in Songjiang who

309

conducted the massive follow-up of the contact cohort over years.

310

Funding: Key Project of Chinese National Programs for Fundamental Research and

311

Development: Shanghai Municipal Science and Technology Commission (2017ZX10105012);

312

Three-Year Act on Public Health System Construction in Songjiang District, Shanghai China

313

(2015-2017); and National Natural Science Foundation of China (81402727).

314

Disclaimer The views expressed in this paper do not necessarily reflect those of the funding

315

body.

316

Competing interests None declared.

317

Ethical approval: Shanghai CDC routinely collected the demographic data and clinical

318

isolates from TB patients. The institutional review boards of Shanghai CDC approved the

319

analysis with the anonymous dataset.

TE D

EP

320 321

323

Figure legends

AC C

322

M AN U

SC

308

324

Figure 1. Work Flow of contact tracing. N=the number of index cases (the number of

325

contacts).

326

Figure 2. SNPs between index-contact pairs, by sputum smear status of the index cases.

327

Figure 3. Transmission network interpreted by WGS.

12

ACCEPTED MANUSCRIPT 328

References

329

1

330 331

World Health Organization. Global tuberculosis report 2016. Geneva: World Health Organization, 2016.

2

World Health Organization. Recommendations for investigating contacts of persons with infectious tuberculosis in low- and middle-income countries.

333

Geneva: World Health Organization, 2012.

334

3

RI PT

332

Erkens CG, Kamphorst M, Abubakar I, et al. Tuberculosis contact

335

investigation in low prevalence countries: a European consensus. Eur Resp J

336

2010;36(4):925-49. doi: 10.1183/09031936.00201609. 4

Cavany SM, Sumner T, Vynnycky E, et al. An evaluation of tuberculosis

SC

337

contact investigations against national standards. Thorax 2017 doi:

339

10.1136/thoraxjnl-2016-209677.

340

5

341 342

M AN U

338

CDC. Guidelines for the Investigation of Contacts of Persons with Infectious Tuberculosis. Morbidity and Mortality Weekly Report 2005;54(RR-15).

6

Zachariah R, Spielmann MP, Harries AD, et al. Passive versus active

343

tuberculosis case finding and isoniazid preventive therapy among household

344

contacts in a rural. Int J Tuberc Lung Dis 2015;7:1033-39. 7

Pothukuchi M, Nagaraja SB, Kelamane S, et al. Tuberculosis contact

TE D

345 346

screening and isoniazid preventive therapy in a South Indian district:

347

operational issues for programmatic consideration. PloS one 2011;6:5-8.

348

8

Department of Disease Control M, Department of Medical Administration M, Prevention CCfDCa. Guidelines for Implementing the National Tuberculosis

350

Control Program in China (2008). Beijing: Beijing Union Medical College

351

Press 2008.

353 354 355 356

9

AC C

352

EP

349

Chang KC, Leung CC & Tam CM. Household contact investigation of tuberculosis in low-income and middle-income countries: public-health impact. Lancet Infect Dis 2009;9(1):3-4. doi: 10.1016/S1473-3099(08)70287-1.

10 Morrison J, Pai M, Hopewell PC.. Tuberculosis and latent tuberculosis

357

infection in close contacts of people with pulmonary tuberculosis in

358

low-income and middle-income countries: a systematic review and

359

meta-analysis. Lancet Infect Dis 2008;8(6):359-68. doi:

360

10.1016/S1473-3099(08)70071-9. 13

ACCEPTED MANUSCRIPT 361

11 Glynn JR, Guerra-Assuncao JA, Houben RM, et al. Whole Genome Sequencing

362

Shows a Low Proportion of Tuberculosis Disease Is Attributable to Known

363

Close Contacts in Rural Malawi. PloS one 2015;10(7):e0132840. doi:

364

10.1371/journal.pone.0132840.

365

12 Buu TN, van Soolingen D, Huyen MN, et al. Tuberculosis acquired outside of households, rural Vietnam. Emerg Infect Dis 2010;16(9):1466-8. doi:

367

10.3201/eid1609.100281.

368

RI PT

366

13 Urdahl KB, Shafiani S, Ernst JD. Initiation and regulation of T-cell responses in tuberculosis. Mucosal immunology. 2011;4(3):288-293. doi:

370

10.1038/mi.2011.10.

371

SC

369

14 Luo T, Yang C, Pang Y, et al. Development of a hierarchical variable-number tandem repeat typing scheme for Mycobacterium tuberculosis in China. PloS

373

one 2014;9(2):e89726. doi: 10.1371/journal.pone.0089726.

M AN U

372

374

15 Del Sal G, Manfioletti G & Schneider C. The CTAB-DNA precipitation method:

375

a common mini-scale preparation of template DNA from phagemids, phages

376

or plasmids suitable for sequencing. Biotechniques 1989;7:514-20. 16 Yang C, Luo T, Shen X, et al. Transmission of multidrug-resistant

378

Mycobacterium tuberculosis in Shanghai, China: a retrospective

379

observational study using whole-genome sequencing and epidemiological

380

investigation. Lancet Infect Dis 2017;17(3):275-84. doi:

381

10.1016/s1473-3099(16)30418-2

TE D

377

17 Walker TM, Lalor MK, Broda A, et al. Assessment of Mycobacterium

383

tuberculosis transmission in Oxfordshire, UK, 2007–12, with whole

384

pathogen genome sequences: an observational study. Lancet Respir Med,

386 387

AC C

385

EP

382

2014;2(4):285-292. doi:10.1016/S2213-2600(14)70027-X.

18 Walker TM, Ip CL, Harrell RH, et al. Whole-genome sequencing to delineate Mycobacterium tuberculosis outbreaks: a retrospective observational study.

388

Lancet Infect Dis. 2013;13(2):137-146. doi:

389

10.1016/S1473-3099(12)70277-3.

390

19 Liu E, Cheng S, Wang X, et al. A systematic review of the investigation and

391

management of close contacts of tuberculosis in China. J Public Health

392

2010;32(4):461-6. doi: 10.1093/pubmed/fdq032.

393

20 Qin Y, Fang H, Li H, et al. [Analysis of incidence in clost contacts of 14

ACCEPTED MANUSCRIPT 394

smear-positive tuberculosis cases]. Chin J Antituber 2008;30:1.

395

21 Jia Z, Cheng S, Ma Y, et al. Tuberculosis burden in China: a high prevalence of

396

pulmonary tuberculosis in household contacts with and without symptoms.

397

BMC Infect Dis 2014;14:64.

398

22 Anger HA, Proops D, Harris TG, et al. Active case finding and prevention of tuberculosis among a cohort of contacts exposed to infectious tuberculosis

400

cases in New York City. Clin Infect Dis 2012;54(9):1287-95. doi:

401

10.1093/cid/cis029.

402

RI PT

399

23 Hirsch-Moverman Y, Cronin WA, Chen B, et al. HIV counseling and testing in tuberculosis contact investigations in the United States and Canada. Int J

404

Tuberc Lung Dis 2015;19(8):943-53. doi: 10.5588/ijtld.14.0642.

SC

403

24 Bayona J, Chavez-Pachas AM, Palacios E, et al. Contact investigations as a

406

means of detection and timely treatment of persons with infectious

407

multidrug-resistant tuberculosis. Int J Tuberc Lung Dis 2003;7(12 Suppl 3):9.

408

M AN U

405

25 Denholm J, Leslie D, Jenkin GA, et al. Long-term follow-up

409

of contacts exposed to multidrug-resistant tuberculosis in Victoria, Australia,

410

1995-2010. Int J Tuberc Lung Dis 2012;16(10):6. doi: 10.5588/ijtld.12.0092. 26 Izumi K, Ohkado A, Uchimura K, et al. Evaluation of tuberculosis contact

TE D

411 412

investigations in Japan. Int J Tuberc Lung Dis 2017;21(2):188-95. doi:

413

10.5588/ijtld.16.0508.

415

27 Zhang T, Chen W, Cheng S. [Progress in tuberculosis patient close contact investigation.] Disease Surveillance 2009;24(3):213-7.

EP

414

28 Yang C, Shen X, Peng Y, et al. Transmission of Mycobacterium tuberculosis in

417

China: a population-based molecular epidemiologic study. Clin Infect Dis

418

AC C

416

2015;61(2):219-27. doi: 10.1093/cid/civ255.

419

29 Brooks-Pollock E, Becerra MC, Goldstein E, et al. Epidemiologic inference

420

from the distribution of tuberculosis cases in households in Lima, Peru. J

421 422

Infect Dis 2011;203(11):1582-9. doi: 10.1093/infdis/jir162. 30 Sloot R, Schim van der Loeff MF, Kouw PM, et al. Risk of tuberculosis after

423

recent exposure. A 10-year follow-up study of contacts in Amsterdam. Am J

424

Respir Crit Care Med 2014;190(9):1044-52. doi:

425

10.1164/rccm.201406-1159OC.

426

31 McIntosh AI, Doros G, Jones-López EC, et al. Extensions to Bayesian 15

ACCEPTED MANUSCRIPT 427

generalized linear mixed effects models for household tuberculosis

428

transmission. Statistics in Medicine 2017 doi: 10.1002/sim.7303.

RI PT

SC

Med J 2015;350 doi: ARTN h810 10.1136/bmj.h810.

M AN U

432

33 Lipman M, White J. Collaborative tuberculosis strategy for England. Bmj-Brit

TE D

431

Japan.] Chin J Antituberc 2017;39:22-7.

EP

430

32 Mori T. [Problems with contact investigation in tuberculosis case-finding in

AC C

429

16

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT