Emerging Microbes & Infections

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/temi20

Tuberculosis preventive treatment among

individuals with inactive tuberculosis suggested
by untreated radiographic abnormalities: a
community-based randomized controlled trial

Haoran Zhang, Henan Xin, Ying Du, Xuefang Cao, Shouguo Pan, Jianmin Liu,
Ling Guan, Fei Shen, Zisen Liu, Bin Zhang, Dakuan Wang, Boxuan Feng, Jiang
Du, Xueling Guan, Yijun He, Yongpeng He, Zhanjiang Zhang, Jiaoxia Yan, Qi
Jin, Lei Gao & for the LATENTTB TRIAL2-NSTM study team

To cite this article: Haoran Zhang, Henan Xin, Ying Du, Xuefang Cao, Shouguo Pan, Jianmin
Liu, Ling Guan, Fei Shen, Zisen Liu, Bin Zhang, Dakuan Wang, Boxuan Feng, Jiang Du, Xueling
Guan, Yijun He, Yongpeng He, Zhanjiang Zhang, Jiaoxia Yan, Qi Jin, Lei Gao & for the LATENTTB
TRIAL2-NSTM study team (2023) Tuberculosis preventive treatment among individuals with
inactive tuberculosis suggested by untreated radiographic abnormalities: a community-
based randomized controlled trial, Emerging Microbes & Infections, 12:1, e2169195, DOI:
10.1080/22221751.2023.2169195

To link to this article: https://doi.org/10.1080/22221751.2023.2169195

© 2023 The Author(s). Published by Informa View supplementary material

UK Limited, trading as Taylor & Francis
Group.

Published online: 27 Jan 2023. Submit your article to this journal

Article views: 2192 View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=temi20
Emerging Microbes & Infections
2023, VOL. 12, e2169195 (10 pages)
https://doi.org/10.1080/22221751.2023.2169195

RESEARCH ARTICLE

Tuberculosis preventive treatment among individuals with inactive

tuberculosis suggested by untreated radiographic abnormalities: a
community-based randomized controlled trial
Haoran Zhanga*, Henan Xina*, Ying Dua*, Xuefang Caoa*, Shouguo Panb*, Jianmin Liuc*, Ling Guanc,
Fei Shenc, Zisen Liub, Bin Zhangb, Dakuan Wangb, Boxuan Fenga, Jiang Dua, Xueling Guanc, Yijun Hea,
Yongpeng Hea, Zhanjiang Zhangb, Jiaoxia Yanb, Qi Jina, Lei Gaoa and for the LATENTTB TRIAL2-NSTM study
teama†
a
NHC Key Laboratory of Systems Biology of Pathogens, Institute of Pathogen Biology, and Center for Tuberculosis Research, Chinese
Academy of Medical Sciences and Peking Union Medical College, Beijing, People’s Republic of China; bCenter for Diseases Control and
Prevention of Zhongmu County, Zhengzhou, People’s Republic of China; cThe Sixth People’s Hospital of Zhengzhou, Zhengzhou, People’s
Republic of China

ABSTRACT
Epidemiological and interventional studies have been rarely conducted among those with positive interferon-γ release
assay (IGRA) results and radiologically inactive tuberculosis (TB) lesions on chest radiograph. This study aimed to estimate
the effectiveness and safety of a six-week twice-weekly regimen (rifapentine plus isoniazid) among this key population in
rural China. First, chest digital radiography was conducted to screen individuals with inactive TB lesions. Then, the
identified participants were further evaluated and eligible participants with IGRA-positive results were included in
subsequent randomized controlled trial (RCT). Of 44,500 recruited residents, 2,988 presented with radiographically
inactive TB among 43,670 with complete results of chest radiography and questionnaire, and 28.61% (855/2,988)
tested IGRA positive. Subsequently, 677 eligible participants were included in this RCT (345 in the preventive
treatment group and 332 in the untreated control group). The treatment completion rate was 80.00% (276/345), and
11.88% (41/345) participants reported side-effects including two cases of hepatotoxicity (0.58%, 2/345). In the
intention-to-treat analysis, the cumulative incidence rate of microbiologically confirmed active TB during a two-year
follow-up was 1.16 (95% confidence interval [CI]: 0.03–2.29) in the preventive treatment group and 1.51 (95% CI:
0.20–2.82) in the control group (p = .485). Subgroup analyses showed that the protective rates were 55.42% (95% CI:
10.33–93.07%) and 80.17% (95% CI: 25.36–97.96%) for participants with fibrosis and for those aged ≥60 years,
respectively. The expected treatment effect was not observed for the six-week regimen in this study. Future studies
with sufficient sample size are needed to verify our findings.

ARTICLE HISTORY Received 22 August 2022; Revised 18 December 2022; Accepted 11 January 2023

KEYWORDS Latent tuberculosis infection; preventive treatment; radiographically inactive tuberculosis; fibrosis; randomized controlled trial

Introduction for preventive treatment, which could incur a great

epidemiological value, needs to be firstly addressed
According to the estimation by the World Health
when developing local guidelines for LTBI
Organization (WHO), about a quarter of the world’s
management.
population was infected with Mycobacterium tubercu-
The 2020 updated WHO guidelines rec-
losis (MTB), and approximately 5%–10% MTB infec- ommended that individuals with a history of inactive
tions might progress to active tuberculosis (TB) in TB or fibrotic lesions should not be a contraindica-
their lifetimes [1,2]. Promoting the testing and treat- tion for preventive treatment, due to the increased
ment of latent tuberculosis infection (LTBI) among risk of active disease progression [3]. Secondary pre-
high-risk populations is a critical tool for the END ventive treatment has been practically recommended
TB. Obviously, population-based LTBI testing and for individuals with abnormal chest radiographic
treatment is infeasible, and given the vast pool of findings suggestive of inactive TB (untreated or
infections, strategies centred on high-risk groups will inadequate treated) in America and Canada due to
provide greater cost effectiveness. Identifying targets the 6–19 times increased risk of developing active

CONTACT Lei Gao [email protected] NHC Key Laboratory of Systems Biology of Pathogens, Institute of Pathogen Biology, and Center for
Tuberculosis Research, Chinese Academy of Medical Sciences and Peking Union Medical College, No. 9 Dong Dan San Tiao, Beijing 100730, People’s
Republic of China
*These authors contributed equally to this study.
†
Team members of the LATENTTB TRIAL1-NSTM study were listed in the online data supplement.
Supplemental data for this article can be accessed online at https://doi.org/10.1080/22221751.2023.2169195.
© 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrest-
ricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
2 H. Zhang et al.

disease as compared to those with normal chest radi- Methods

ography [4,5]. Our previously prospective study con-
Trial design
sistently found that individuals with inactive TB
suggested by untreated chest radiographic abnormal- The study, which conducted in rural communities in
ities showed an increased risk of developing active Zhongmu County with an average TB incidence of
TB, this subgroup contributed about 30% of TB 57 per 100,000 for three consecutive years (2015–
cases occurred in our study population in rural 2017), consisted of two phases: a community-level
China [6,7]. However, little was known about the screening study based on chest digital radiography
burden of this specific population and their TB (DR) investigations (two-round screening survey, see
infection status in China. Under this circumstance, Figure 1) and an open-labelled two-arm RCT includ-
we firstly conducted a population-based large-scale ing a preventive treatment group and control group
screening study to identify eligible participants without intervention. The control group was settled
among rural residents to establish baseline results without preventive treatment because there was cur-
for further interventions. rently no recommendation for this population in the
In addition, our previous studies have proved guidelines of WHO or China [3,10]. The regimen
the efficacy of a six-week regimen (rifapentine was six weeks of twice-weekly rifapentine (at a maxi-
plus isoniazid) as 69% in two years [8] and 61% mum dose of 600 mg; at a dose calculated based on
in five years [9] among Chinese rural residents. body weight, with incremental adjustment for partici-
Therefore, the primary aim of this subsequent ran- pants who weigh ≤50 kg, for example, if a participant
domized controlled trial (RCT) was to evaluate the weighs 38–50 kg, the dose of drug he or she should
safety and two-year protective effect of this short- take is 450 mg; and at a dose of 600 mg for those
course regimen for TB preventive treatment who weigh >50 kg) plus isoniazid (at a maximum
among individuals with inactive TB suggestive dose of 600 mg; carried and rounded off at a dose of
untreated by radiographic abnormalities. This 15 mg/kg for those who weigh ≤50 kg, for example,
study was expected to promote the exploration of if a participant weighs 43 kg, the dose of drug he or
suitable TB preventive treatment for this specific she should take is calculated as 600-[15 mg/kg*(50–
population in China and areas with high burden 43)] = 495, so 500 mg will be taken; and at a dose of
of inactive TB. 600 mg for those who weigh >50 kg).

Figure 1. Flowchart of the participant enrolment. ALT, alanine aminotransferase; AST, aspartate aminotransferase; DR, digital radi-
ography; HCV, hepatitis C virus; QFT, QuantiFERON-TB Gold In-Tube; RCT, randomized controlled trial; TB, tuberculosis; ULN, upper
limit of normal; WBC, white blood cell. Of the 44,500 recruited rural residents, 43,670 participants aged 18–75 years had complete
results of chest radiography and questionnaire, 40,682 were excluded because of normal chest radiographic findings (n = 40,451),
suspect TB (n = 119) and other pulmonary diseases (n = 112). The rest 2,988 participants were identified with radiographically
inactive TB lesions. Among them, 855 individuals were identified to be QFT-positive and 830 signed the informed consent
form for the RCT. Finally, 677 eligible participants were included in the RCT and were randomly classified into two groups:
345 in the preventive treatment group and 332 in the control group without treatment.
 Emerging Microbes & Infections 3

This trial was registered at the Chinese Clinical The sample size was calculated with the aim of
Trial Registry (ChiCTR-1800018224). The protocol reducing TB incidence in the intervention group by
was approved by the ethics committees of the Institute 70% in two years with a risk of active TB development
of Pathogen Biology, Chinese Academy of Medical of 2.6% in intervention group [9], and 229 participants
Sciences, Beijing, China (IPB-2018-03-12). Written per group would provide a power of 80% based on a
informed consent was obtained from each study par- two-sided alpha level of 0.05. Assuming 20% treat-
ticipant. Proswell Medical Company (Beijing, China) ment discontinuation and 15% loss to a two-year fol-
independently monitored the study during its low-up, the final estimated sample size was 674
implementation. participants.

Study participants and sample size calculation Community-level screening and intervention
participant recruitment
For the enrolment screening, eligible participants
were required to (1) be aged 18–75 years; (2) be Between 18 September 2017 and 29 August 2018, a
local household registration of rural residents or con- community-level baseline screening was initially con-
tinuous residence at the study site for at least six ducted to identify participants with radiographically
months [11] and have no migration plan in the inactive TB using chest DR examination. To improve
next three years; (3) be self-committed to complete the accuracy of the radiographic diagnosis of inactive
the entire study period; and (4) be willing to sign TB, two-round chest DR screening combined with
the informed consent form. Exclusion criteria were MTB infection testing was conducted. In the first
(1) residents with current active TB or suspected round, participants who were initially identified with
TB (suspected TB was defined by chest radiographic radiographically inactive TB lesions by one radiologist
abnormalities that were suggestive of a probable would be introduced for QFT testing. Those with
active TB but without microbiologic evidence accord- QFT-positive results were then invited for the
ing to the National Health Industry Standard on second-round screening survey, in an interval of
Diagnosis for Pulmonary Tuberculosis [WS288- three to six months (between 30 August 2018 and 15
2017]) [12]; (2) residents with previous anti-TB treat- October 2018) to identify eligible participants for the
ment history confirmed by checking information RCT. Specifically, chest DR was repeated for QFT-
recorded in the national Tuberculosis Information positive participants with radiographically inactive
Management System (NTIMS); or (3) residents who TB lesions to exclude unstable lesions with pro-
were pregnant or expectant women. gression. In this round of screening, a panel of radiol-
The RCT inclusion criteria were (1) residents with ogists, consisting of one local radiologist and one
radiographical lesions suggestive of inactive TB, which radiologist from provincial infectious diseases hospi-
were adapted from the National Health Industry Stan- tal, read each image together and provided a common
dard on Classification of Tuberculosis (WS196-2017) report after discussion. If the participants were
[13], including fibrotic lesions, calcification, nodule, confirmed to have radiological findings suggestive of
and pleural incrassation; and (2) QuantiFERON-TB suspected TB, microbiological testing including spu-
Gold In-Tube (QFT, a commercial product of Inter- tum smear, culture and GeneXpert MTB/RIF assay
feron-γ release assay [IGRA]; Qiagen, Germantown, (Cepheid, Sunnyvale, CA, USA) would be done to
MD, USA) positive results (a cut-off value of exclude currently active TB. TB diagnosis was made
≥0.35 IU/mL recommended by the manufacturer). according to the National Health Industry Standard
The participants would be excluded if they met one on Diagnosis for Pulmonary Tuberculosis (WS288-
of the following exclusion criteria: (1) contraindica- 2017) [12]. The standard report form was shown in
tion for rifapentine or isoniazid, (2) history of LTBI Supplementary Table S1. Additionally, each partici-
treatment with rifapentine persistently for >14 days pant completed a standardized questionnaire to obtain
or with isoniazid intermittently for >30 days in the sociodemographic information, as well as blood rou-
past two years [14]; (3) serum alanine aminotransfer- tine examination and biochemistry tests. After com-
ase (ALT) and/or aspartate aminotransferase (AST) >2 prehensively considering the test results, eligible
times the upper limit of normal (ULN), and/or participants for the RCT were determined.
accompanied by liver impairment symptoms and
signs; (4) white blood cell (WBC) count <2.0 × 109
Randomization
L−1; (5) positivity of antibody to hepatitis C virus
(HCV) and/or positive for antibody to human immu- Eligible participants selected for the RCT were ran-
nodeficiency virus (HIV); (6) with renal insufficiency domly classified into the preventive treatment group
or degeneration; (7) addiction to drinking or actively or control group by sex using computer-generated
using street drugs; (8) cancer; and (9) mental disorders random numbers. All eligible participants were cen-
or disability. trally allocated a sequential study identifier by the
4 H. Zhang et al.

RCT statistician and then randomly assigned to one of defined as clinically diagnosed TB. Cases registered in
the two groups. Participants, study team, and statisti- NTIMS during the follow-up period were also
cal analysts were not blinded to treatment assignment. included in the present study. In consideration of
However, the panel members for evaluation and diag- the suboptimal accuracy of clinical diagnosis in this
nosis of TB were fully unaware of the trial-group special population, microbiologically confirmed TB
assignments. was regarded as the primary study endpoint in order
to more precisely evaluate the efficacy of the regimen.
Furthermore, to rule out initial misclassification of co-
Intervention implementation
prevalent cases, we excluded individuals diagnosed
At each village clinic visit, to prevent control contami- with TB during the first quarterly follow-up period
nation, the study drugs were dispensed and taken after from the analysis.
meals under direct observation by trained village Treatment completion, permanent discontinu-
doctors. Participants had no access to prescription ation, or death from any cause were also analysed.
drugs by themselves. During the pre-specified six- Treatment completion was defined as completion of
week phase, participants in the preventive treatment ≥11 of 12 doses and was comprehensively assessed
group were allowed to take the medicine within by pill counts and drug registration. AEs and study
three days in case of interruptions. However, for the drug-related side-effects were graded by researchers
last dose, considering the need for more physical according to the Common Terminology Criteria for
examination and information collection, participants Adverse Events (CTC-AE, version 4.0). The attribu-
were allowed for seven days’ delay in order to improve tion of drug-related side-effects was co-determined
the adherence [14]. The occurrence of adverse events by two external physicians using a scale ranging
(AEs), including gastrointestinal reactions, neurologi- from unrelated to definitely related (Supplementary
cal symptoms, hypersensitivity or allergy, influenza- Table S2).
like symptoms and other systemic drug reactions, Temporary or permanent discontinuation was
were recorded by the researchers on each visit before judged by physicians and principal investigators
dispensing medicine every time. Participants were based on clinical evaluation. For permanently discon-
asked to return to the clinic if they felt unwell at any tinued participants, follow-up lasted until the end of
time throughout the treatment course. To closely two-year follow-up.
monitor side-effects, scheduled clinical examinations
and evaluations, including blood routine examination
Statistical methods
and blood biochemistry tests (test indicators included
AST, ALT, total bilirubin, creatinine, and urea nitro- Analyses were performed using SAS (SAS Institute,
gen, etc.), were conducted once every two weeks version 9.4). Between-group comparisons for categ-
during the treatment period. Assessment of long- orical variables were estimated by Chi-squared test
term side-effects occurred at three-month visit after with a two-sided significance level of .05. The inten-
the last medication. tion-to-treat (ITT) analysis included all participants
enrolled in the two groups. The per-protocol analysis
included participants who completed the study regi-
Follow-up survey and endpoints
men. The cumulative incidence and incidence rate
All participants including untreated controls were fol- per 100 person-year with 95% confidence interval
lowed up for 24 months. At quarter follow-up, partici- (CI) was calculated at 24 months. The differences in
pants were investigated of suspected symptoms the incidence of active TB between the preventive
through door-to-door or telephone survey. At annual treatment and untreated control groups were assessed
follow-up, participants were invited for investigation using a one-sided significance level of .05. Due to lost
on suspected symptoms and chest DR examination. to follow-up issues, Kaplan–Meier time-event analy-
Participants with evidence of suspected TB were trans- sis was used. Person-year for each participant was
ferred to the local Center for Diseases Control and calculated based on the quarterly and annual fol-
Prevention for final diagnosis. Three sputum samples low-up records for 24 months after treatment. If
were collected from each participant for sputum the participants did not participate in the two-year
smear, culture and GeneXpert MTB/RIF assay. On survey, the duration for which they remained in the
the basis of National Health Industry Standard on study was calculated based on the records of the
Diagnosis for Pulmonary Tuberculosis (WS288- last completed quarterly follow-up. If the participants
2017) [12], individuals with positive results for culture had died of other causes, the person-years were calcu-
and/or GeneXpert MTB/RIF assay, were diagnosed lated based on the records of time of death. If necess-
with microbiologically confirmed TB cases. Response ary, Cox proportional hazards regression models and
to empiric TB treatment among participants suspected an unconditional multiple logistic regression analysis
as having TB but negative to microbiological tests was would be performed to identify respective potential
 Emerging Microbes & Infections 5

risk factors related with TB incidence and drug- Completion rate of treatment
related side-effects.
As shown in Table 2, the rate of treatment completion
was 80.00% (276/345), with discontinuing treatment
due to refusal to continue (31.88%, 22/69), underlying
Results
disease management without hospitalization (27.54%,
Study participants 19/69), hospitalization due to an underlying disease
(15.94%, 11/69), drug side-effects (14.49%, 10/69),
As a pre-enrolment screening for the RCT, 44,500
unreachable (7.25%, 5/69), and other reasons (2.90%,
rural residents were recruited, 830 of whom were
2/69), respectively. Details on therapy completion
excluded (decline to undergo chest radiography
were presented in Supplementary Table S5.
screening (n = 506), without questionnaire results
(n = 135), with history of anti-TB treatment (n = Table 1. Characteristics of the study participants included in
133), and with duplicate results (n = 56)), and the intention-to-treat analysis.
43,670 were potentially eligible with completed Preventive Untreated
results of questionnaires survey and chest DR exam- treatment group control group (N P for χ 2
Characteristic (N = 345) = 332) test
ination. A total of 40,682 participants were further
Male 199 (57.68) 194 (58.43) .843
excluded from the second-round screening survey Age (years)
due to normal chest DR examination findings (n = Median (Q25–Q75) 63 (54–68) 64 (55–69) .309
<50 45 (13.04) 40 (12.05) .924
40,451), suspected TB (n = 119), and other pulmon- 50–59 90 (26.09) 85 (25.60)
ary disorders (n = 112). The rest 2,988 (6.84%) par- 60–69 140 (40.58) 133 (40.06)
70–75 70 (20.29) 74 (22.29)
ticipants were identified with radiographically Body mass index
inactive TB lesions and all of them were examined (BMI, kg m−2)
by QFT testing. Among them, 855 (28.61%, 855/ <18.5 14 (4.06) 10 (3.01) .855
18.5–<24.0 133 (38.55) 124 (37.35)
2,988) participants were identified to be QFT-posi- 24.0–<28.0 134 (50.38) 132 (39.76)
tive, 2121(70.99%, 2,121/2,988) were QFT-negative ≥28.0 64 (18.55) 66 (19.88)
Completed primary 133 (38.55) 116 (34.94) .330
and 12 (0.40%, 12/2,988) were indeterminate. Sub- school
sequently, the 855 QFT-positive participants with Married 309 (89.57) 301 (90.66) .633
Household per capita 153 (44.35) 157 (47.29) .443
radiographically inactive TB were invited for the income > RMB
second-round screening survey, and 830 of them 2500 yuan
Ever-smoker 114 (33.04) 103 (31.02) .574
signed the informed consent form for the RCT if Current alcohol 86 (24.93) 96 (28.92) .242
checked to be eligible. Among them, 153 were further drinker
excluded due to an ALT and/or AST result >2 ULN Close contact with a 9 (2.61) 7 (2.11) .669
patient with active
(n = 69), missing WBC results (n = 53), missing ALT TB
and/or AST results (n = 11), HCV-positive results History of type II 19 (.51) 19 (5.72) .903
diabetes*
(n = 9), WBC <2.0 × 109 L−1 (n = 2), and chest DR HBsAg positive 7 (2.03) 4 (1.20) .397§
results that were inconsistent with the first-round Self-reported history 0 (0) 1 (0.30) .490§
of silicosis
image readings (which were recognized as suspected Inactive lesions
TB in the second-round screening survey) (n = 9). identified by chest
radiograph#
None of the participants had ever received LTBI With fibrotic 129 (19.05) 114 (16.84) .459
treatment before, and none was HIV-positive. Finally, lesions
With nodules 25 (3.69) 23 (3.40) .872
677 participants were identified to be eligible for the With calcification 195 (28.80) 187 (27.62) .959
RCT and were randomly classified into two groups: With pleural 25 (3.69) 34 (9.94) .167
345 in the preventive treatment group and 332 in incrassation
Size of the largest
the control group without treatment (Figure 1). fibrosis lesion
Detailed characteristics of the participants were (mm)†
<20 98 (75.97) 82 (71.93) .473
shown in Table 1 and Supplementary Table S3. ≥20 31 (24.03) 32 (28.07)
Among the total screened participants, 39.32% Concomitant drugs
during treatment‡
(17,172/43,670) were males, and the median age was Yes 84 (24.35) – –
59 years. The prevalence of fibrotic lesions was No 261 (75.65) – –
2.81% (1,230/43,670) and accounted for 41.16% Data are presented as n (%). Abbreviation: Q25, quartile 25; Q75, quartile
75; TB, tuberculosis. ‡Data on taking concomitant drugs were monitored
(1,230/2,988) among any radiographically inactive and collected during clinic visits for study drug treatment; therefore,
TB lesions. For those with fibrotic lesions, QFT posi- related information on concomitant drug use in untreated control
group were not collected. #Every type of radiographically inactive
tivity was 25.45% (313/1,230). After randomization, lesions could include only a single lesion or one accompanied with
all the variables were evenly distributed between the other radiographically inactive lesions. Therefore, the sum of the sample
groups. The distribution of age, baseline IFN-γ level, sizes for inactive lesions identified by chest radiographic findings might
not be equal to the total of 677. *History of self-reported type II diabetes
and active TB by fibrosis size were shown in Sup- or fasting blood glucose ≥7 mmol L−1. †Classified by Q75 value. §Fisher’s
plementary Table S4. exact test.
6 H. Zhang et al.

Table 2. Completion of the treatment regimens and (74/345) and 2.37% (10/345), respectively. Among 41
attributions for treatment discontinuation. (11.88%) participants with drug-related side-effects,
Variables n (%) the proportions of gastrointestinal reactions (6.38%),
Number of completed doses neurological symptoms (2.90%), hypersensitivity or
11–12 doses (completed)* 276 (80.00)
<11 doses (uncompleted) 69 (20.00) allergy (1.74%), influenza-like symptoms (1.45%),
Attribution for treatment discontinuation and hepatotoxicity (0.58%) were observed. Detailed
Unreachable 5 (7.25)
Adverse drug effect 10 (14.49) drug-related side-effects for each dose were shown in
Refusal to continue 22 (31.88) Supplementary Table S9. Only one treated case
Hospitalization due to underlying disease† 11 (15.94)
Management required for underlying disease but 19 (27.54) (0.29%) was in hospitalization due to the side-effects.
without hospitalization† Two participants (0.58%) were identified with elevated
Other reasons including right arm trauma and traffic 2 (2.90)
accident
ALT/AST higher than normal level, which resolved to
*Defined as completed ≥90% doses (i.e. ≥11 doses) of drug therapy. the normal level after the administration of the medi-
†
Underlying disease management included management of hyperten- cations had ceased. None of the participants had
sion and high fasting blood glucose.
>3×ULN or >5×ULN, and no renal dysfunction was
identified.
Protective effect of the treatment regimen
In total, 15 TB cases (nine were microbiologically Risk factors
confirmed and six were clinically diagnosed) were As shown in Supplementary Table S10, univariate
identified during follow-up, eight in the preventive analysis showed that no considered variables were
treatment group and seven in the untreated controls statistically significantly associated with the incidence
(Table 3 and Supplementary Table S6). Supplementary of active TB.
Figure S1 shows the Kaplan–Meier curve of the time to Detailed information on concomitant medication
active TB according to the groups. In the ITT analysis use during preventive treatment was showed in Sup-
(Table 3), the cumulative incidence of total active TB plementary Table S11. The frequency of side-effects
was 2.32% (95% CI: 0.73–3.19%) in the preventive occurrence among treated participants with concomi-
treatment group and 2.11% (95% CI: 0.56–3.65%) in tant medications was significantly higher than that of
the untreated control group, respectively (p = .531). those without concomitant medications (27.38% vs.
The results did not remarkably change when including 6.90%, p < .001). Multiple logistic regression analysis
only microbiologically confirmed TB cases. The indicated that the risk of side-effects increased
cumulative incidence rate of microbiologically among the participants with concomitant medication
confirmed active TB during a two-year follow-up (adjusted odds ratio [OR]: 3.41, 95% CI: 2.58–4.68)
was 1.16 (95% CI: 0.03–2.29) in the preventive treat- (Supplementary Table S12).
ment group and 1.51 (95% CI: 0.20–2.82) in the con-
trol group (p = .485). In the per-protocol analysis
(Table 3), the cumulative incidence of microbiologi- Discussion
cally confirmed TB was 1.09% (95% CI: 0.22–3.14%) To the best of our knowledge, this is the first RCT on
in the preventive treatment group. No statistical sig- LTBI testing and treatment in populations with radio-
nificance was found when compared with the graphically inactive TB and without history of TB
untreated controls (p = .467) as well. Further subgroup treatment in rural China. This study showed a high
analyses showed that the respective protective rate was burden of radiographically inactive TB among the
55.42% (95% CI: 10.33–93.07%) (p = .454) and 80.17% study participants (6.84%, 2,988/43,670), and the
(95% CI: 25.36–97.96%) (p = .104) for the participants prevalence of LTBI was as high as 28.61% (855/
with fibrosis and for those aged older than 60 years, 2,988) among these participants. Unfortunately, the
although it did not reach statistical significance studied six-week regimen for LTBI treatment did not
(Table 3). There were no significant differences in show significant protective effect in the total study
the TB incidence by the type of radiographically inac- population. Subsequent exploratory subgroup ana-
tive TB lesions (Supplementary Table S7). lyses observed a positive protective effect among the
participants with fibrotic lesions (55.42%) and
among those aged ≥60 years (80.17%), but it did not
Adverse events
reach statistical significance.
For 12 deaths that occurred in the two groups during Our results suggested that there was a certain pro-
the study period (five in the preventive treatment portion of the general population in rural China who
group and seven in the untreated control group), might have been infected with MTB and left inactive
none of them were attributed to the LTBI treatment TB lesions after self-cured. Additionally, the preva-
(Supplementary Table S8). As shown in Table 4, the lence of LTBI was much higher among individuals
proportions of total and serious AEs were 21.45% with radiographically inactive TB (28.61%, 855/
Table 3. Incidence of active tuberculosis in the study groups.
Preventive treatment group Untreated control group
Cumulative incidence, Incidence rate per 100 Cumulative incidence, Incidence rate per 100 Difference in P-value for difference in Protective rate* % Protective rate† %
n/N (%, 95% CI) person-years (95% CI) n/N (%, 95% CI) person-years (95% CI) cumulative rate‡ cumulative rate¶ (95% CI) (95% CI)
ITT analysis
Microbiologically confirmed and clinically 8/345 8/650 7/332 7/620 0.21 .531 – –
diagnosed cases among total population (2.32, 0.73–3.91) (1.23, 0.62–2.41) (2.11, 0.56–3.65) (1.13, 0.55–2.31)
Microbiologically confirmed among total 4/345 4/650 5/332 5/620 −0.35 .485 23.18 23.46
population (1.16, 0.03––2.29) (0.62, 0.24–1.58) (1.51, 0.20–2.82) (0.81, 0.35–1.88) (1.83–83.04) (1.07–89.69)
Microbiologically confirmed cases among 1/129 1/243 2/114 2/218 −0.97 .454 55.42 55.43
individuals with fibrosis (0.78, 0.02–4.24) (0.41, 0.07–2.29) (1.75, 0.21–6.19) (0.92, 0.25–3.29) (10.33–93.07) (6.19–95.91)
Microbiologically confirmed cases among 0/31 0/59 1/32 2/62 −3.13 .508 100.00 100.00
individuals with fibrosis size ≥ 20 mm (0) (0) (3.13, 0.08–9.15) (3.23, 0.89–11.03) (0–55.10) (0–54.32)
Microbiologically confirmed cases among 1/210 1/397 5/207 5/388 −1.94 .104 80.17 80.62
individuals aged ≥ 60 years (0.48, 0.01–2.62) (0.25, 0.04–1.41) (2.42, 0.32–4.51) (1.29, 0.55–2.98) (25.36–97.96) (16.53–98.87)
PP analysis
Microbiologically confirmed and clinically 6/276 6/524 7/332 7/620 0.06 .586 – –
diagnosed cases among total population (2.17, 0.45–3.89) (1.15, 0.53–2.48) (2.11, 0.56–3.65) (1.13, 0.55–2.31)
Microbiologically confirmed among total 3/276 3/524 5/332 5/620 −0.42 .467 27.81 29.63
population (1.09, 0.22–3.14) (0.57, 0.19–1.67) (1.51, 0.20–2.82) (0.81, 0.35–1.88) (2.57–84.91) (1.68–91.23)
Microbiologically confirmed cases among 1/108 1/206 2/114 2/218 −0.82 .520 46.86 46.74
individuals with fibrosis (0.93, 0.02–5.05) (0.49, 0.03–3.10) (1.75, 0.21–6.19) (0.92, 0.25–3.29) (7.60–90.43) (4.48–94.26)
Microbiologically confirmed cases among 0/27 0/52 1/32 2/62 −3.13 .542 100.00 100.00
individuals with fibrosis size ≥ 20 mm (0) (0) (3.13, 0.08–9.15) (3.23, 0.89–11.03) (0–55.10) (0–54.32)
Microbiologically confirmed cases among 1/165 1/316 5/207 5/388 −1.81 .170 74.79 75.19
individuals aged ≥ 60 years (0.61, 0.02–3.33) (0.32, 0.06–1.78) (2.42, 0.32–4.51) (1.29, 0.55–2.98) (22.33–96.84) (14.48–98.19)

Emerging Microbes & Infections

Abbreviation: CI, confidence interval; ITT, intention-to-treat analysis; PP, per-protocol analysis. *Protective rate was calculated by using cumulative incidence. †Protective rate was calculated by using incidence rate per 100 person-years. ‡The
difference was the rate in the preventive treatment group minus the rate in the untreated control group. ¶One-sided p-value for Chi-squared test or Fisher’s exact test.

7
8 H. Zhang et al.

Table 4. Characteristics of the participants with adverse poor response to the preventive treatment. In the con-
events and study drug side-effects. trol group, only 1.51% study participants developed
Variables n (%) microbiologically confirmed TB during two-year fol-
Participants 345 low-up, the unexpected lower incidence of active TB
Adverse events
Participants with any of the adverse events 74 (21.45) among the control group might support this specu-
Death 0 (0) lation. In addition, we could not rule out the possi-
Occurrence of serious adverse events
Yes 10 (2.37) bility that our included participants might not be
No 64 (18.55) recently infected, as it is commonly observed that
Frequency per person
1 event 53 (15.36) half of the lifetime risk of active TB was occurred in
>1 event 21 (6.07) the first two years after infection [3]. All in all, the
Attributed to drug
Yes 41 (11.88)
unexpected lower incidence of active TB also limited
No 33 (9.57) the power of this study to identify a significant protec-
Classified by event severity grade tive effect in the study population. Second, we could
Grade 1 56 (16.23)
Grade 2 15 (4.35) not exclude the possibility that the studied short-
Grade 3 2 (0.58) course regimen was not effective in the population
Grade 4 1 (0.29)
Side-effects with inactive TB. After all, this study was the first
Participants with any of the side-effects 41 (11.88) time to practice a short-course regimen in this popu-
Specific side-effect
Gastrointestinal reaction 22 (6.38) lation worldwide. However, focusing on the study par-
Hyposensitivity or allergy 6 (1.74) ticipants with fibrotic lesions, especially those with a
Influenza-like symptoms
Neurological symptoms
5 (1.45)
10 (2.90)
fibrotic size ≥20 mm and elderly (≥60 years), a
Hepatotoxicity 2 (0.58) much better protective rate was observed. Although
Other drug reactions 3 (0.87) not reaching statistical significance, it indicated that
Categories occurring per person
1 34 (9.86) the six-week short-course regimen might produce a
2 5 (1.45) more positive effect if we could define the target popu-
3 2 (0.58)
Hospitalization due to side-effects 1 (0.29) lation with radiographically inactive TB more pre-
cisely. Therefore, further studies are urgently needed
to improve the definition of inactive TB with compre-
2,988) than that among the general rural adults from hensive consideration of the type of radiographic
the same study site (19.0%, 801/4,223) [15]. Therefore, lesions, current MTB infection status, and history of
for the first time as we know, our results disclosed a MTB exposure and anti-TB treatment. Such improve-
high burden of radiographically inactive TB with ment would be crucial for evaluating and expanding
LTBI in rural China. It was essential to explore suitable LTBI treatment among individuals with radiographi-
preventive treatment tools for such a key population. cally inactive TB in the future.
TB preventive treatment studies have been conducted To improve the guarantee on safety, we adapted the
since the 1950s [16,17], and evidences [14,18–20] and frequency of active side-effects monitoring by revising
recommendations from the WHO [3,21,22] supported it from once monthly to twice monthly based on pre-
its effectiveness for several populations, including for vious research. Therefore, some adverse reactions
individuals with inactive TB [23–25]. However, the were handled in time at the initial stage. On the
main recommended option for preventive treatment other hand, the participants also received more care,
was monotherapy with daily doses of isoniazid for at which improved their compliance. During the inter-
least six months. In this study, we tried to use a vention implementation, only two participants
newly developed six-week regimen for this population showed moderate liver dysfunction, which was signifi-
to improve compliance and feasibility of the preven- cantly lower than the findings in our previous trial that
tive treatment. However, we did not observe a signifi- performed among middle-aged and elderly individ-
cantly positive protective effect for this short-course uals with LTBI (1.17%) [8]. Increasing the monitoring
regimen as observed in our previous studies [8,9]. frequency of adverse reactions will increase the inter-
The finding might be explained by two potential vention cost, therefore, further studies are needed to
reasons. First, the definition of inactive TB at enrol- balance the risks and benefits of preventive treatment
ment in our study was based on four typically radio- more systematically according to the characteristics of
graphical inactive TB lesions adopted by Chinese the intervention target population.
industrial standard [13]. However, the ability of such Consistent with the recommendation by guidelines
a definition based on radiographic abnormalities to from America and Canada, we excluded participants
specifically represent an inactive TB need to be further with anti-TB treatment history because there was
explored, as other lung infections may leave similar insufficient evidence to include them as preventive
lesions. Therefore, the potential misclassification treatment targets. However, some evidence still
may have led to the inclusion of study participants suggested that this population was at a higher risk
who were not the actual high-risk participants with of developing disease in high-burden settings
 Emerging Microbes & Infections 9

[6,7,26]. Therefore, LTBI management for this popu- Acknowledgements

lation, especially those who have been exposed, We appreciate Dr Dehua Tu from Beijing Institute of
should not be neglected. In addition, China has a Tuberculosis control, Dr Xinhua Zhou and Dr Hongfei
high burden of HBV infection, therefore, participants Duan from Beijing Chest Hospital for their support in
with HBV were not excluded from this study to chest film interpretation. We thank Yifan Zhao from the
prompt individuals co-infected with MTB and HBV Second High School attached to Beijing Normal University
for his support in the field work. We also thank all the inves-
to benefit from preventive treatment in future [27].
tigators in the study site for their contributions.
However, the epidemic level of HCV was relatively
low in China [28] and HCV infection was associated
with anti-TB treatment-induced hepatotoxicity [29– Disclosure statement
31]. Therefore, given safety concerns regarding No potential conflict of interest was reported by the author(s).
potentially hepatotoxicity, participants with HCV
were excluded from this study. Regarding partici-
pants with HIV infection, the reason for their exclu- Funding
sion was because the prevalence was relatively low in This work was supported by the National Science and Tech-
China [32] and there was lacking evidence of the nology Major Project of China [2017ZX10201302-002],
safety and efficacy of the six-week regimen among CAMS Innovation Fund for Medical Sciences (CIFMS)
immunodeficient populations. Therefore, exploratory [2021-I2M-1-037], and the Fundamental Research Funds
and confirmatory research should be conducted for the Central Universities [3332021092]. They were not
involved in the trial design, patient recruitment, data collec-
among these specific populations with a well-con- tion, analysis, interpretation, or any aspect pertinent to the
ducted RCT design in the future. study.
Our study has several limitations. First, some TB
cases especially subclinical TB cases might have
been under-diagnosed because only participants Data availability statement
with suspected TB were transferred for further This study is registered at www.chictr.org.cn with identifier
microbiological tests. The unexpected lower inci- ChiCTR-1800018224. The corresponding author can pro-
dence of active disease limited us to detect expected vide, upon request, individual participant data that underlie
treatment effects and get a conclusive finding, con- the results reported in this article after applying necessary
measures to guarantee that no individual is identified or
sidering the current sample size. Although not reach- identifiable.
ing statistical significance, a positive protective effect
was observed among the subgroups with fibrotic
lesion and among those aged ≥60 years, further References
studies with sufficient outcome events from larger [1] Hartman-Adams H, Clark K, Juckett G. Update on
sample size are needed to verify our results as calcu- latent tuberculosis infection. Am Fam Physician.
lated in Supplementary Table S13. Second, partici- 2014;89(11):889–896.
pants in this trial were not blinded because an [2] Leung CC, Rieder HL, Lange C, et al. Treatment of
untreated control group was used. To reduce the latent infection with Mycobacterium tuberculosis:
update 2010. Eur Respir J. 2011;37(3):690–711.
risk of potential bias, the panel members for evalu- [3] World Health Organization. WHO consolidated
ation and diagnosis of the active TB were fully una- guidelines on tuberculosis. Geneva: World Health
ware of the trial-group assignments. In addition, AE Organization; 2020.
evaluation was mainly based on laboratory measure- [4] American Thoracic Society. Targeted tuberculin test-
ments and graded using a standardized, widely used ing and treatment of latent tuberculosis infection.
MMWR Recomm Rep. 2000;49(Rr-6):1–51.
CTC-AE. Third, as there is no gold standard for
[5] Public Health Agency of Canada. Canadian tuberculo-
LTBI testing, infection status and time of infection sis standards. 7th ed. Ottawa (ON): Public Health
acquisition were not known with certainty. There is Agency of Canada; 2014.
a need to further explore whether these factors [6] Xin H, Zhang H, Yang S, et al. 5-Year follow-up of
would influence treatment effectiveness. active tuberculosis development from latent infection
In conclusion, a high burden of radiographically in rural China. Clin Infect Dis. 2020;70(5):947–950.
[7] Gao L, Li X, Liu J, et al. Incidence of active tuberculo-
inactive TB without anti-TB treatment history was sis in individuals with latent tuberculosis infection in
identified in rural China. No significant efficacy of rural China: follow-up results of a population-based,
the six-week regimen was observed for TB preventive multicentre, prospective cohort study. Lancet Infect
treatment among this key population, which might Dis. 2017;17(10):1053–1061.
have been influenced by ages and the type of radiologi- [8] Gao L, Zhang H, Xin H, et al. Short-course regimens
of rifapentine plus isoniazid to treat latent tuberculosis
cal lesion. A more precise definition of radiographi-
infection in older Chinese patients: a randomised con-
cally inactive TB is needed to verify our preliminary trolled study. Eur Respir J. 2018;52(6):1801470.
findings and to explore precise context-specific inter- [9] Xin H, Cao X, Zhang H, et al. Protective efficacy of 6-
vention for this population in future. week regimen for latent tuberculosis infection
10 H. Zhang et al.

treatment in rural China: 5-year follow-up of a ran- latent tuberculosis in adults. N Engl J Med. 2018;379
domised controlled trial. Eur Respir J. 2022;60 (5):440–453.
(1):210235. [21] World Health Organization. Guidelines on the man-
[10] National Health Commission of the People’s Republic agement of latent tuberculosis infection. Geneva:
of China. Technical specification for tuberculosis pre- World Health Organization; 2015.
vention and control in China (2020 edition). [cited [22] World Health Organization. Latent tuberculosis infec-
2022 Oct 16]. Available from: https://tb.chinacdc.cn/ tion-updated and consolidated guidelines for pro-
ggl/202004/P020200414515703939844.pdf. grammatic management. Geneva: World Health
[11] Technical Guidance Group of the Fifth National TB Organization; 2018.
Epidemiological Survey. The office of the fifth [23] Snider DE Jr, Farer LS. Preventive therapy with isonia-
National TB epidemiological survey. The fifth national zid for “inactive” tuberculosis. Chest. 1978;73(1):4–6.
tuberculosis epidemiological survey in 2010. Chin J [24] Katz J, Kunofsky S, Damijonaitis V, et al. Effect of iso-
Antitubercul. 2012;34(8):485–508. niazid upon the reactivation of inactive tuberculosis;
[12] National Health and Family Planning Commission of final report. Am Rev Respir Dis. 1965;91:345–350.
the People’s Republic of China. Diagnostic criteria for [25] Grzybowski S, Ashley MJ, Pinkus G.
tuberculosis (WS288-2017). Elec J Emerg Infect Dis. Chemoprophylaxis in inactive tuberculosis: long-
2018;3(1):65–67. term evaluation of a Canadian trial. Can Med Assoc
[13] National Health Commission of the People’s Republic J. 1976;114(7):607–611.
of China. Classification of tuberculosis (WS196-2017); [26] Cheng J, Sun YN, Zhang CY, et al. Incidence and risk
2017. factors of tuberculosis among the elderly population in
[14] Sterling TR, Villarino ME, Borisov AS, et al. Three China: a prospective cohort study. Infect Dis Poverty.
months of rifapentine and isoniazid for latent tubercu- 2020;9(1):13.
losis infection. N Engl J Med. 2011;365(23):2155– [27] Liu J, Liang W, Jing W, et al. Countdown to 2030:
2166. eliminating hepatitis B disease, China. Bull World
[15] Gao L, Lu W, Bai L, et al. Latent tuberculosis infection Health Organ. 2019;97(3):230–238.
in rural China: baseline results of a population-based, [28] Yue T, Zhang Q, Cai T, et al. Trends in the disease bur-
multicentre, prospective cohort study. Lancet Infect den of HBV and HCV infection in China from 1990–
Dis. 2015;15(3):310–319. 2019. Int J Infect Dis. 2022;122:476–485.
[16] Comstock GW. Isoniazid prophylaxis in an undeve- [29] Kim WS, Lee SS, Lee CM, et al. Hepatitis C and not
loped area. Am Rev Respir Dis. 1962;86:810–822. Hepatitis B virus is a risk factor for anti-tuberculosis
[17] Mount FW, Ferebee SH. The effect of isoniazid pro- drug induced liver injury. BMC Infect Dis. 2016;16:50.
phylaxis on tuberculosis morbidity among household [30] Lomtadze N, Kupreishvili L, Salakaia A, et al. Hepatitis
contacts of previously known cases of tuberculosis. C virus co-infection increases the risk of anti-tuberculo-
Am Rev Respir Dis. 1962;85:821–827. sis drug-induced hepatotoxicity among patients with
[18] Swindells S, Ramchandani R, Gupta A, et al. One pulmonary tuberculosis. PLoS One. 2013;8(12):e83892.
month of Rifapentine plus Isoniazid to prevent HIV- [31] Agha MA, El-Mahalawy II, Seleem HE-DM, et al.
related tuberculosis. N Engl J Med. 2019;380 Prevalence of hepatitis C virus in patients with tuber-
(11):1001–1011. culosis and its impact in the incidence of anti-tubercu-
[19] Villarino ME, Scott NA, Weis SE, et al. Treatment for losis drugs induced hepatotoxicity. Eur Respir J.
preventing tuberculosis in children and adolescents: a 2015;64:91–96.
randomized clinical trial of a 3-month, 12-dose regi- [32] National Center for AIDS & STD Control and
men of a combination of rifapentine and isoniazid. Prevention. The Chinese Center for Disease Control
JAMA Pediatr. 2015;169(3):247–255. and Prevention (CDC). Annals of information on
[20] Menzies D, Adjobimey M, Ruslami R, et al. Four comprehensive prevention and treatment for AIDS,
months of Rifampin or nine months of Isoniazid for STD and Hepatitis C. Beijing, China; 2020.