Abstract
Background
Subclinical tuberculosis has been increasingly recognized as a separate state in the spectrum of the disease. However, evidence on the transmissibility of subclinical tuberculosis is still inconclusive.
Methods
We re-analyzed the data from the 2007 combined tuberculosis prevalence and tuberculin surveys in Vietnam. Poisson regression with robust standard errors was conducted to assess the effect of clinical presentation of individuals with tuberculosis in the household on tuberculin skin test (TST) positivity among children aged 6–14 years who participated in the tuberculin survey, adjusting for child's age, smear status of the index patient, and other covariates.
Results
In the multivariate analysis, we found significantly increased risks for TST positivity in children living with patients with clinical, smear-positive tuberculosis, compared with those living with individuals without tuberculosis (adjusted risk ratio [aRR]: 3.04; 95% confidence interval [CI]: 2.00–4.63) and with those living with patients with subclinical tuberculosis, adjusting for index smear status (aRR: 2.26; 95% CI: 1.03–4.96). Among children aged 6–10 years, those living with patients with clinical, smear-positive tuberculosis and those living with patients with subclinical, smear-positive tuberculosis had similarly increased risks of TST positivity compared with those living with individuals without tuberculosis (aRRs [95% CI] of 3.56 [1.91–6.62] and 3.11 [1.44–6.72], respectively).
Conclusions
Our findings support the hypothesis that smear-positive subclinical tuberculosis contributes to Mycobacterium tuberculosis transmission. To eliminate tuberculosis in 2035, control strategies need to address subclinical presentations of the disease.
Keywords: tuberculosis, subclinical, transmission, epidemiology
Among children aged 6–10 years, those living with patients with clinical, smear-positive tuberculosis or with subclinical, smear-positive tuberculosis had similarly increased risks of tuberculin skin test positivity compared with those living in background households with no individuals with tuberculosis.
Tuberculosis (TB), caused by Mycobacterium tuberculosis (Mtb), is transmitted between humans through expelled respiratory droplets. In recent years, TB disease has affected approximately 10 million individuals, killing 1.2 to 1.5 million people each year globally [1]. Transmission of Mtb is mostly unrecorded: in several studies, only a small proportion of TB transmission was traceable from household members and known contacts [2, 3].
Tuberculosis has been dichotomized into latent TB infection, which is nonsymptomatic and noninfectious, and active disease, which is symptomatic and infectious. However, there is increasing recognition of a more nuanced spectrum of disease, which also includes incipient and subclinical TB [4]. Subclinical TB is commonly defined as a disease state caused by viable Mtb bacteria that do not cause typical symptoms, while other abnormalities can be detected using radiology or microbiological tests [4, 5]. Data from a recent systematic review suggest that active disease is not an inevitable and irreversible outcome for subclinical TB, with only 25% of patients with subclinical TB progressing to clinical TB [6]. Subclinical TB is likely missed when patients present to clinical services where the TB diagnostic pathway is started by symptom screening [7, 8]. The common understanding has long been that cough is necessary for Mtb transmission; however, recent bioaerosol and face-mask sampling studies challenge this paradigm [9, 10]. If indeed subclinical TB, defined as a lack of (prolonged) cough, could contribute to Mtb transmission [8, 11], this can frustrate the global efforts to meet the World Health Organization’s (WHO’s) EndTB Strategy target to eliminate the TB epidemic by 2035.
Vietnam, a high-TB-burden country, conducted 2 consecutive national TB prevalence surveys in 2007 and 2017. Nearly half of the patients with TB found in these surveys had no cough for 2 weeks or more and, in both surveys, the prevalence was more than 2-fold higher than the corresponding notification rate, suggesting many missed TB diagnoses and/or long durations of disease before diagnosis [12, 13]. To inform decisions around directing global attention toward averting Mtb transmission from patients with subclinical TB, it is essential to know how infectious they are compared with patients with clinical TB. The 2007 TB prevalence survey, combined with a tuberculin survey among children living in the same households, offered an excellent opportunity to answer this question.
METHODS
Survey Design and Procedure
In 2007, the Vietnam National TB Programme conducted a nationwide TB prevalence survey with the participation of 94 179 adults. Nested within this survey was a tuberculin survey to estimate the annual risk of TB infection in Vietnam. This tuberculin survey included 23 160 children aged 6–14 years living in the same households as the adults enrolled in the prevalence survey. All children were eligible for tuberculin skin testing (TST) by trained survey team members following the Mantoux method using RT23 2TU in Tween-80 (Statens Serum Institute, Copenhagen, Denmark). Excluded were those with skin rash or fever, those who were currently on TB treatment, and those whose parents did not consent to participate. Results were read 72 hours after administering the TST and recorded as the maximum transverse diameter of the induration in millimeters. The detailed survey procedures and study design for both surveys have been reported elsewhere [12, 14].
Definitions
Participants of this study were defined as eligible children in the age range of 6 to 14 years who received a TST and came back for result-reading after 72 hours. Index patients with TB were defined as eligible participants of the TB prevalence survey who were diagnosed with TB using either light microscopy on Ziehl-Neelsen–stained sputum smear samples or culture on Löwenstein-Jensen solid medium. Clinical index patients were defined as index patients with TB who reported cough for 2 weeks or more. Without reporting cough for 2 weeks or more, index patients with TB would be classified as index patients with subclinical TB. In our sensitivity analysis, we stratified subclinical TB into 2 separate categories, including subclinical-no-symptoms, who reported none of the recorded TB-suggestive symptoms (cough, hemoptysis, fever, night sweats, and weight loss), and subclinical-no-prolonged-cough, who did not report cough for 2 weeks or more but still reported at least 1 of the symptoms listed above. In the survey, TST-positive results were defined as typical skin indurations with a diameter of 10 mm or larger, and this threshold was used as the primary outcome in the previous publication [14]. However, with the high infant Bacillus Calmette–Guérin (BCG) vaccination coverage and high rates of nontuberculous mycobacterial infections in Vietnam, the cutoff point of 10 mm might include nonspecific reactions [15, 16]. Therefore, we chose a TST-positive cutoff point of 15 mm as the primary analysis outcome to increase specificity, and thus reduce bias in the analyses. The socioeconomic status (SES) of the household was measured as household expenditure level estimated from a set of household assets [17].
Data Analysis
Data analysis was performed using Stata 14SE (StataCorp, College Station, TX, USA). Our first analysis estimated the adjusted risk ratios (aRRs) of having TST-positive results among children living with clinical index patients in their household and among those living with subclinical index patients, compared with those living in households with no index patient (hereafter referred to as background households). In our second analysis, we directly compared TST-positive results between children with a subclinical index patient and children with a clinical index patient in the household. We used Poisson regression with robust standard errors to estimate the aRRs, as the number of children living with index patients in this study was relatively small, which could introduce bias in a logistic regression model [18, 19]. Since smear-positive TB is generally more infectious than smear-negative TB [20, 21], we took into account the smear status of the index patient. As smear status did not apply to background households, we combined in the first analysis the clinical and the smear status of index patients in a single variable with 5 values: background households with no index patient; subclinical, smear-negative; subclinical, smear-positive; clinical, smear-negative; and clinical, smear-positive. In the second analysis, we included the smear status of index patients as a covariate. We examined through bivariate analyses the following additional covariates for confounding the association between TST positivity and index patient status: region and area of living, the child's BCG vaccination status, household SES, ethnicity, and the TB treatment history of any adult household member. Covariates with a confounding effect, pragmatically defined as a difference between crude and aRRs of 10% or more, were included in a multivariate model after checking for collinearity and interaction. As the prevalence of TB infection measured by TST accumulates throughout the lifetime [22], the age of the children was included in all regressions. We assumed that, on average, the TST results in younger children compared with older children are more exclusively a reflection of recent exposures to Mtb. Therefore, we performed a subgroup analysis stratifying by younger (6–10 years) and older (11–14 years) age and verified its validity by fitting an interaction term between clinical/smear status of the index patient and age of the child. Finally, we performed 2 sets of sensitivity analyses: one in which we replaced the cutoff for TST positivity with 10 mm, the other in which we applied a stratified definition of subclinical TB into subclinical-no-symptoms and subclinical-no-prolonged-cough. Also, our analyses were repeated excluding all children living with index patients who did not have positive culture results. All analyses were conducted at the .05 significance level.
Ethical Approval
The 2007 national TB prevalence survey and its sub-studies received scientific and ethical approval from the Institutional Review Board of the Vietnam National Lung Hospital.
RESULTS
In the tuberculin survey, there were 23 160 eligible children enumerated, of whom 22 585 (97.5%) were presented on the testing day and 21 487 (92.8%) had an available TST result. Among them, 21 298 lived in background households and, of these, 1528 (7.2%) and 3759 (17.7%) were TST-positive using the cutoff points of 15 mm and 10 mm, respectively. There were 189 children who lived with at least 1 index patient found in the TB prevalence survey. At the cutoffs of 15 mm and 10 mm, the numbers of children with positive TST results were 28 (14.8%) and 51 (27.0%), respectively (Figure 1).
Figure 1.
Overview of the study population. Clinical index patient: An index patient who reported cough for 2 weeks or more. Subclinical index patient: An index patient who did not report cough for 2 weeks or more. Abbreviation: TST, tuberculin skin test. aPercentage calculated as the number of children with positive TST result divided by the number of children living with index patients.
Clinical/Subclinical Tuberculosis in the Household Compared With Background Households
Among 189 children living with at least 1 index patient, 101 (53.4%) lived with clinical index patients and 88 (46.6%) lived with subclinical index patients. The numbers of children with a TST-positive result at the cutoff of 15 mm in these 2 groups were 20 (19.8%) and 8 (9.1%), respectively. Overall, the sex and age distributions of children in the 3 groups were similar. However, the proportions of children vaccinated with BCG, and of those living with smear-positive index patients, were lower among children living with subclinical index patients, while the proportion of households with an adult member with a history of previous TB treatment was lower among children living in background households (Figure 1; Table 1).
Table 1.
Characteristics of Children Included in the Analyses
| Characteristics of the Children | Living in Background Households | Living With Subclinical Index Patients | Living With Clinical Index Patients | |||
|---|---|---|---|---|---|---|
| n | % | n | % | n | % | |
| Total | 21 298 | … | 88 | … | 101 | … |
| Sex | ||||||
| Male | 10 963 | 51.5 | 43 | 48.9 | 52 | 51.5 |
| Female | 10 271 | 48.4 | 45 | 51.1 | 49 | 48.5 |
| Age groups | ||||||
| 6–10 years old | 10 539 | 49.5 | 46 | 52.3 | 52 | 51.5 |
| 11–14 years old | 10 759 | 50.5 | 42 | 47.7 | 49 | 48.5 |
| BCG-vaccinated | ||||||
| Yes | 17 602 | 82.7 | 65 | 73.9 | 90 | 89.1 |
| No | 3696 | 17.3 | 23 | 26.1 | 11 | 10.9 |
| TB treatment history of household members | ||||||
| No treatment history | 21 056 | 98.9 | 81 | 92.1 | 93 | 92.1 |
| History of ≤5 years | 242 | 1.1 | 7 | 7.9 | 8 | 7.9 |
| Smear status of index TB patient | Not applicable | |||||
| Smear-positive | … | … | 52 | 59.1 | 72 | 71.3 |
| Smear-negative | … | … | 36 | 40.9 | 29 | 28.7 |
Abbreviation: TB, tuberculosis.
Clinical index patient: An index patient who reported cough for 2 weeks or more. Subclinical index patient: An index patient who did not report cough for 2 weeks or more.
In the bivariate analysis, only the TB treatment history of any member in the household confounded the association between index patient status and TST positivity in the child (Supplementary Table 1). Multivariate analysis showed a significant aRR of 3.04 (95% confidence interval [CI]: 2.00–4.63) among children living with clinical, smear-positive index patients compared with children living in background households. No statistically significant difference was observed in the risk of TST positivity of children living in background households and that of children living in the remaining household types (Table 2).
Table 2.
Adjusted Risk Ratios of TST Positivity (Cutoff 15 mm) Among Children Living With Prevalent Index Patients Compared With Children Living in Background Households, by Multivariable Analysis
| Potential Confounder | Total Children | 6–10-Year-Old Children | ||||
|---|---|---|---|---|---|---|
| No. | RR (95% CI) | P | No. | RR (95% CI) | P | |
| Clinical status of the index TB patient | 21 487 | … | 10 637 | … | ||
| No index TB patient | 21 298 | Reference | 10 539 | Reference | ||
| Subclinical TB, smear (−) | 36 | 1.19 (.41–3.51) | .749 | 21 | .81 (.12–5.57) | .831 |
| Subclinical TB, smear (+) | 52 | 1.21 (.52–2.84) | .654 | 25 | 3.11 (1.44–6.76) | .004 |
| Clinical TB, smear (−) | 29 | 1.50 (.53–4.27) | .445 | 16 | Omitteda | |
| Clinical TB, smear (+) | 72 | 3.04 (2.00–4.63) | <.001 | 36 | 3.56 (1.91–6.62) | <.001 |
| TB treatment history of household members | 21 487 | … | 10 637 | … | ||
| No treatment historyb | 20 993 | Reference | 10 387 | Reference | ||
| History of ≤5 years | 494 | 1.67 (1.32–2.13) | <.001 | 250 | 1.95 (1.91–6.62) | <.001 |
| Age groups | 21 487 | … | … | … | ||
| 6–10 years old | 10 637 | Reference | … | … | ||
| 11–14 years old | 10 850 | 1.59 (1.44–1.75) | <.001 | … | … | |
Abbreviations: CI, confidence interval; RR, risk ratio; TB, tuberculosis; TST, tuberculin skin test.
Clinical index patient: An index patient who reported cough for 2 weeks or more. Subclinical index patient: An index patient who did not report cough for 2 weeks or more.
Sixteen children living with smear-negative clinical index patients were omitted from the analysis because they were all TST-negative.
No TB treatment of a household member within 5 years prior to the survey.
The prevalence of TST positivity among children in the background households increased with age, from 2.65% (95% CI: 2.03–3.45%) among 6-year-old children to 9.56% among 14-year-old children (95% CI: 8.41–10.60%) (Supplementary Figure 1). Including a term for the interaction between the age of the child (6–10 vs 11–14 years) and the clinical-smear combined status of index patients improved the regression model's prediction (likelihood ratio test, P = .007). Among children aged 6–10 years, the aRR of TST positivity was 3.11 (95% CI: 1.44–6.76) in children living with patients with subclinical, smear-positive TB and 3.56 (95% CI: 1.91–6.62) in children living with index patients with clinical, smear-positive TB compared with children living in background households (Table 2).
Subclinical Tuberculosis in the Household Compared With Clinical Tuberculosis in the Household
Table 3 shows the aRRs of TST positivity in children living with subclinical index patients, compared with those having clinical index patients in the household. The covariates with a confounding effect in the bivariate analysis were the smear status of index patients, treatment history of any member in the household, and household SES (Supplementary Table 2). In the multivariate analysis, children living with clinical index patients had a significantly higher risk of TST positivity (aRR: 2.26; 95% CI: 1.03–4.96) compared with children living with subclinical index patients. In the subgroup analysis among children 6–10 years of age, the risk of TST positivity living with subclinical index patients was the same (aRR: 1.01; 95% CI: .39–2.63) as that in children living with clinical index patients (Table 3).
Table 3.
Adjusted Risk Ratios of TST Positivity (Cutoff 15 mm) Among Children Living With Subclinical Index Patients Compared With Children Living With Clinical Index Patients, by Multivariable Analysis
| Potential Confounder | Total Children | 6–10-Year-Old Children | ||||
|---|---|---|---|---|---|---|
| No. | RR (95% CI) | P | No. | RR (95% CI) | P | |
| Clinical status of the index TB patient | 189 | … | 98 | … | ||
| Subclinical TB | 88 | Reference | 46 | Reference | ||
| Clinical TB | 101 | 2.26 (1.03–4.96) | .041 | 52 | 1.01 (.39–2.63) | .978 |
| Smear status of index TB patients | 189 | … | 98 | … | ||
| Smear (−) | 65 | Reference | 37 | Reference | ||
| Smear (+) | 124 | 1.79 (.76–4.20) | .181 | 61 | 8.04 (1.09–59.26) | .041 |
| Household expenditure | 189 | … | 98 | … | ||
| High | 66 | Reference | 40 | Reference | ||
| Medium | 68 | .49 (.21–1.17) | .107 | 39 | .66 (.22–2.04) | .474 |
| Low | 55 | .53 (.24–1.22) | .138 | 19 | .64 (.19–2.22) | .485 |
| Age groups | 189 | … | … | … | ||
| 6–10 years old | 98 | Reference | … | … | ||
| 11–14 years old | 91 | 1.13 (.57–2.27) | .724 | … | … | |
Abbreviations: CI, confidence interval; RR, risk ratio; TB, tuberculosis; TST, tuberculin skin test.
Clinical index patient: An index patient who reported cough for 2 weeks or more. Subclinical index patient: An index patient who did not report cough for 2 weeks or more.
Sensitivity Analyses
When taking 10 mm as the TST-positivity threshold, only children living with clinical, smear-positive index patients had a significantly increased TST-positivity prevalence (aRR: 2.24; 95% CI: 1.69–2.99) compared with those living in background households. In the subgroup analysis among children 6–10 years of age, the aRR of TST positivity was 1.81 (95% CI: 1.02–3.35) in children living with patients with subclinical, smear-positive TB and 2.84 (95% CI: 1.93–4.17) in children living with clinical, smear-positive index patients, compared with children living in background households (Supplementary Table 3).
In the primary model in which subclinical TB was stratified into no-symptoms and no-prolonged-cough, neither was associated with a significantly increased risk of TST positivity compared with children living in background households (Supplementary Table 4). When this analysis was restricted to children 6–10 years of age, statistically significant higher risks of TST positivity compared with children living in background households were observed in children living with subclinical-no-prolonged-cough, smear-positive index patients (aRR: 2.81; 95% CI: 1.06–7.47) and in children living with subclinical-no-symptoms, smear-positive index patients (aRR: 3.71; 95% CI: 1.07–12.86) (Supplementary Table 4).
There were 22 children living with index patients who had smear-positive, culture-negative or culture-unknown test results. The analysis results showed no substantial change after excluding these from all analyses (data not shown).
DISCUSSION
Compared with children living in households with no index patients with TB, we found a significantly (3.04 times) increased risk of having positive TST results for children living with clinical, smear-positive index patients but not for those living with subclinical index patients. The risk of a positive TST was 2.26 lower for children living with subclinical index patients than for children living with clinical index patients, adjusting for index smear status. Among children in the younger age group (6–10 years), however, the increased risks of TST positivity were similar for children living with clinical, smear-positive index patients and those living with subclinical, smear-positive index patients. These findings were irrespective of how subclinical TB was defined (no symptoms or no prolonged cough) and robust to the cutoff used for defining a positive TST.
Although in the primary analysis we did not find a significantly increased risk of TST positivity in children living with subclinical, smear-positive index patients, an increased risk was observed among the younger children. These findings should be interpreted with caution, as the results came from a subgroup analysis and Mtb can be transmitted not only within households but also in the community; bias in the analysis may occur when community transmissions are misinterpreted as household transmissions [23]. However, as contact data from Vietnam showed that older children are more likely to have non-household contacts who may have infectious forms of TB [24], recent exposures among older children do not consistently reflect household transmission. Therefore, infection prevalence in the younger age group plausibly reflects recent household transmission rather than remote or non-household transmission, suggesting that the risks observed in this group provide a better estimation of transmission from household index patients with TB observed concurrently in the survey. This is supported by the significant age interaction in the primary model, indicating that the age-differential association between the status of the index patient and TST positivity of the household contacts was not a chance finding. In this context where both BCG vaccine coverage and rates of nontuberculous mycobacterial infection are high [15, 16], different cutoffs for the TST represent different compromises between the test's sensitivity and specificity [25]. When, instead of 15 mm, an induration of 10 mm was chosen as the TST-positivity threshold, results were similar to the primary findings, indicating that these were not dependent on the choice of cutoff.
The smear status of the index patients was included in all analyses as studies have shown that smear-positive index patients are much more infectious than their smear-negative counterparts [20, 26]. However, studies have suggested that the association between sputum smear status and transmission is inconsistent [10, 27]. While these studies focused on clinical patients, other studies provide strong evidence that the presence of respiratory symptoms is not necessary for Mtb transmission [9, 28, 29]. In addition, subclinical index patients may still have atypical or intermittent cough due to upper respiratory tract infections or allergies, which are not recognized when the patients are interviewed [11, 30, 31]. Patients with TB also expel aerosols when normal activities are involved, such as tidal breathing [9]. A study of face-mask sampling suggested that cough is not necessary for Mtb transmission, as tidal breathing for 1 hour produced enough Mtb for microbiological confirmation long before Mtb could be detected in sputum, and cough was not associated with the amount of Mtb expelled [10]. This evidence, combined with the findings in our study, suggests that classic TB symptoms, especially cough, are not required for TB transmission and subclinical TB is likely to contribute a substantial part of TB transmission on a population scale. This was suggested by a recent analysis of multiple datasets to which our data contributed [32]. While this analysis took a somewhat different approach in analyzing background transmission, the results were in line with ours.
Infection prevalence reflects not only the infectiousness of the index patients but also the period that these persons have been infectious [6]. Derived from a cross-sectional survey, our data cannot provide the means to estimate the duration of infectiousness for subclinical versus clinical index patients. The 88 subclinical index patients found in our survey are likely a mixture of those who had the disease for a long time without symptoms and those who had acquired TB recently, which could possibly progress to clinical TB over a short period of time [6]. Therefore, the interpretation of the findings in terms of infectiousness of subclinical versus clinical TB is complex, since the duration of infectiousness of subclinical index patients in our survey population may have either been shorter or longer than that of clinical index patients. It is essential for future studies to disentangle infectiousness as such from the duration of infectiousness, since this duration would be the parameter that is most readily influenced by TB-control measures, such as active case finding.
Our study has limitations. Our data were taken from a cross-sectional survey, with no genotyping performed to distinguish transmission sources; therefore, we could not be sure that the source of transmission was from the same household. We had no follow-up data on the TST-positive children, so we do not know how many children ultimately progressed to TB disease. TST misclassification may have occurred, as the TST results were read by several survey team members, causing interreader variation, and a strong digit preference was found in our data, in which TST results were rounded off to multiples of 5 mm (TST results are presented in Supplementary Table 5). Most of the subclinical index patients in our analysis had abnormal chest X-rays, which could make them more infectious than the average pool of patients with subclinical TB. Finally, our data came from a single country, potentially limiting the generalizability of our results to other settings, and the number of TST-positive children living in households with index patients was small, limiting the precision of our estimates while making many expected associations nonsignificant.
In conclusion, the findings of our study support the hypothesis that subclinical TB contributes to the transmission of Mtb. To eliminate TB in 2035, it is essential for TB-control strategies to shift attention to subclinical TB. Further studies are needed to confirm the infectiousness of subclinical TB and its contribution to the transmission of Mtb globally.
Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Supplementary Material
Contributor Information
Hai Viet Nguyen, Vietnam National Tuberculosis Program, Ha Noi, Vietnam; Department of Global Health and Amsterdam Institute for Global Health and Development, Amsterdam University Medical Centers, Amsterdam, The Netherlands.
Edine Tiemersma, KNCV Tuberculosis Foundation, The Hague, The Netherlands.
Nhung Viet Nguyen, Vietnam National Tuberculosis Program, Ha Noi, Vietnam.
Hoa Binh Nguyen, Vietnam National Tuberculosis Program, Ha Noi, Vietnam.
Frank Cobelens, Department of Global Health and Amsterdam Institute for Global Health and Development, Amsterdam University Medical Centers, Amsterdam, The Netherlands.
Notes
Acknowledgments. The authors thank the Vietnam National Tuberculosis Program Board of Directors and all tuberculosis staff involved in this study and the participants living in the survey sites who participated in this study.
References
- 1. World Health Organization . Global tuberculosis report 2021. Geneva, Switzerland: WHO Press, 2022. [Google Scholar]
- 2. Glynn JR, Guerra-Assunção JA, Houben RM, et al. Whole genome sequencing shows a low proportion of tuberculosis disease is attributable to known close contacts in rural Malawi. PLoS One 2015; 10:e0132840. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. McCreesh N, White RG. An explanation for the low proportion of tuberculosis that results from transmission between household and known social contacts. Sci Rep 2018; 8:5382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Drain PK, Bajema KL, Dowdy D, et al. Incipient and subclinical tuberculosis: a clinical review of early stages and progression of infection. Clin Microbiol Rev 2018; 31:e00021-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Migliori GB, Ong CWM, Petrone L, D'Ambrosio L, Centis R, Goletti D. The definition of tuberculosis infection based on the spectrum of tuberculosis disease. Breathe 2021; 17:210079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Richards AS, Sossen B, Emery JC, et al. The natural history of TB disease—a synthesis of data to quantify progression and regression across the spectrum. medRxiv 2022. [Preprint]. Available from: 10.1101/2021.09.13.21263499. [DOI] [Google Scholar]
- 7. World Health Organization . WHO guidelines approved by the guidelines review committee. Systematic screening for active tuberculosis: principles and recommendations. Geneva, Switzerland: World Health Organization, 2013. [PubMed] [Google Scholar]
- 8. Stuck L, van Haaster AC, Kapata-Chanda P, Klinkenberg E, Kapata N, Cobelens F. How “subclinical” is subclinical tuberculosis? An analysis of national prevalence survey data from Zambia. Clin Infect Dis 2022; 75:842–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Wurie FB, Lawn SD, Booth H, Sonnenberg P, Hayward AC. Bioaerosol production by patients with tuberculosis during normal tidal breathing: implications for transmission risk. Thorax 2016; 71:549–54. [DOI] [PubMed] [Google Scholar]
- 10. Williams CM, Abdulwhhab M, Birring SS, et al. Exhaled Mycobacterium tuberculosis output and detection of subclinical disease by face-mask sampling: prospective observational studies. Lancet Infect Dis 2020; 20:607–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Esmail H, Dodd PJ, Houben R. Tuberculosis transmission during the subclinical period: could unrelated cough play a part? Lancet Respir Med 2018; 6:244–6. [DOI] [PubMed] [Google Scholar]
- 12. Hoa NB, Sy DN, Nhung NV, Tiemersma EW, Borgdorff MW, Cobelens FGJ. National survey of tuberculosis prevalence in Viet Nam. Bull World Health Organ 2010; 88:273–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Nguyen HV, Tiemersma EW, Nguyen HB, et al. The second national tuberculosis prevalence survey in Vietnam. PLoS One 2020; 15:e0232142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Hoa NB, Cobelens FG, Sy DN, Nhung NV, Borgdorff MW, Tiemersma EW. First national tuberculin survey in Viet Nam: characteristics and association with tuberculosis prevalence. Int J Tuberc Lung Dis 2013; 17:738–44. [DOI] [PubMed] [Google Scholar]
- 15. Nguyen HT, Bui DD, Nguyen CV, et al. Tuberculosis epidemiology in six provinces of Vietnam after the introduction of the DOTS strategy. Int J Tuberc Lung Dis 2006; 10:963–9. [PubMed] [Google Scholar]
- 16. Tran BN, Quy HT, Qui NC, Nguyen LT, Dinh SN, Cobelens F. Decrease in risk of tuberculosis infection despite increase in tuberculosis among young adults in urban Vietnam. Int J Tuberc Lung Dis 2010; 14:289–95. [PubMed] [Google Scholar]
- 17. Nguyen HB, Tiemersma EW, Dinh SN, et al. Household expenditure and tuberculosis prevalence in VietNam: prediction by a set of household indicators. Int J Tuberc Lung Dis 2011; 15:32–7. [PubMed] [Google Scholar]
- 18. Nemes S, Jonasson JM, Genell A, Steineck G. Bias in odds ratios by logistic regression modelling and sample size. BMC Med Res Methodol 2009; 9:56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Knol MJ, Le Cessie S, Algra A, Vandenbroucke JP, Groenwold RH. Overestimation of risk ratios by odds ratios in trials and cohort studies: alternatives to logistic regression. CMAJ 2012; 184:895–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Asadi L, Croxen M, Heffernan C, et al. How much do smear-negative patients really contribute to tuberculosis transmissions? Re-examining an old question with new tools. EClinicalMedicine 2022; 43:101250. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Tostmann A, Kik SV, Kalisvaart NA, et al. Tuberculosis transmission by patients with smear-negative pulmonary tuberculosis in a large cohort in the Netherlands. Clin Infect Dis 2008; 47:1135–42. [DOI] [PubMed] [Google Scholar]
- 22. Ku C, Dodd PJ. Forecasting the impact of population ageing on tuberculosis incidence. PLoS One 2019; 14:e0222937. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Kendall EA. When infections don't reflect infectiousness: interpreting contact investigation data with care. Clin Infect Dis 2021; 73:e3456-e8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Horby P, Pham QT, Hens N, et al. Social contact patterns in Vietnam and implications for the control of infectious diseases. PLoS One 2011; 6:e16965. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Berkel GM, Cobelens FG, de Vries G, Draayer-Jansen IW, Borgdorff MW. Tuberculin skin test: estimation of positive and negative predictive values from routine data. Int J Tuberc Lung Dis 2005; 9:310–6. [PubMed] [Google Scholar]
- 26. Hernández-Garduño E, Cook V, Kunimoto D, Elwood RK, Black WA, FitzGerald JM. Transmission of tuberculosis from smear negative patients: a molecular epidemiology study. Thorax 2004; 59:286–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Jones-López EC, Acuña-Villaorduña C, Ssebidandi M, et al. Cough aerosols of Mycobacterium tuberculosis in the prediction of incident tuberculosis disease in household contacts. Clin Infect Dis 2016; 63:10–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Patterson B, Wood R. Is cough really necessary for TB transmission? Tuberculosis (Edinburgh, Scotland) 2019; 117:31–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Asadi S, Wexler AS, Cappa CD, Barreda S, Bouvier NM, Ristenpart WD. Aerosol emission and superemission during human speech increase with voice loudness. Sci Rep 2019; 9:2348. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Yousaf N, Monteiro W, Matos S, Birring SS, Pavord ID. Cough frequency in health and disease. Eur Respir J 2013; 41:241–3. [DOI] [PubMed] [Google Scholar]
- 31. Field SK, Escalante P, Fisher DA, Ireland B, Irwin RS. Cough due to TB and other chronic infections: cHEST guideline and expert panel report. Chest 2018; 153:467–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Emery JC, Dodd PJ, Banu S, et al. Estimating the contribution of subclinical tuberculosis disease to transmission—an individual patient data analysis from prevalence surveys. medRxiv 2022. [Preprint]. Available from: 10.1101/2022.06.09.22276188. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.