Transmissibility and Disease Progression of Asymptomatic Mycobacterium tuberculosis Infection, Lima, Peru

Disclaimer: Early release articles are not considered as final versions. Any changes will be reflected in the online version in the month the article is officially released.

Estimating the transmissibility of asymptomatic Mycobacterium tuberculosis infection can clarify its contribution to tuberculosis (TB) spread. We conducted a prospective cohort study in Lima, Peru, enrolling index TB patients and their household contacts (HHCs) and classifying patients by the presence of symptoms including cough, night sweats, weight loss, or fever. We followed HHCs with serial tuberculin skin testing and clinical evaluations. Among 4,296 child HHCs, adjusted estimates for baseline infection (prevalence ratio 0.62 [95% CI 0.37–1.03]), incident infection at 6 months (hazard ratio (aHR) 0.63 [95% CI 0.27–1.49]), and TB disease during 1 year of follow-up (aHR 0.74 [95% CI 0.35–1.56]) were all consistent with lower risk for infection and disease progression among HHCs of asymptomatic compared with symptomatic index patients. Although asymptomatic infections may be less transmissible than symptomatic infections, the high prevalence of asymptomatic patients in national surveys suggest that they may contribute substantially to transmission.

Tuberculosis (TB) remains a global health concern, with an estimated 10.8 million persons falling ill in 2024 ( 1 ). National TB prevalence surveys show that a large proportion of persons with bacteriologically confirmed Mycobacterium tuberculosis infection do not report symptoms during screening ( 2 ). This group is considered to have subclinical TB, which the World Health Organization (WHO) terms asymptomatic TB ( 3 ). Traditional TB control strategies focus on symptomatic persons seeking care at healthcare facilities or on screening algorithms that initiate testing among those who report symptoms ( 4 ). Asymptomatic patients and those with mild or nonspecific symptoms that are not recognized during symptom-based screening (i.e., minimally symptomatic TB) ( 5 ) are often missed and may contribute to transmission.

Estimating the transmissibility of asymptomatic TB may clarify its contribution to TB transmission. Although studies suggest that asymptomatic TB can be transmitted ( 6 , 7 ), it remains unclear how its infectiousness and potential for disease progression compares with symptomatic TB. Most evidence comes from national prevalence surveys that compared infection prevalence among household contacts (HHCs) of symptomatic and asymptomatic index patients but did not assess incident infection or incident disease ( 8 , 9 ).

Using data from a longitudinal cohort of TB index patients and their HHCs in Lima, Peru, we assessed the relative transmissibility of TB and risk for disease progression after exposure to TB patients who did not report symptoms when they tested positive for M. tuberculosis . We compared infection prevalence at enrollment among HHCs exposed to asymptomatic versus symptomatic index patients as a proxy for baseline transmission and evaluated incident infection at 6 and 12 months to assess ongoing transmission risk. We then compared 12-month disease incidence to estimate progression potential and examined transmission risk factors among asymptomatic index patients. Together, those analyses provide empirical estimates of the transmissibility and disease progression potential of asymptomatic TB and clarify its contribution to overall TB transmission.

We conducted a prospective cohort study among HHCs of index TB patients in Lima, Peru, during September 2009–September 2012. We recruited patients >15 years of age with newly diagnosed pulmonary TB from 106 district health centers in Lima. TB diagnosis required either microbiological evidence (positive sputum smear or culture) or a clinician’s judgment based on chest radiograph, clinical manifestation, or both. We collected demographic and clinical data for index patients, including age, sex, employment status, symptom status, HIV status, smoking, alcohol use, diabetes, socioeconomic status, and sputum smear results.

Within 2 weeks of index patient enrollment, we identified and enrolled HHCs and collected data on age, sex, HIV status, smoking, alcohol use, diabetes, bacille Calmette–Guérin (BCG) vaccination status, and body mass index (BMI). We assessed baseline M. tuberculosis infection using the tuberculin skin test (TST) among HHCs without a prior positive TST or TB disease. We considered a TST positive at > 10 mm induration in HIV-uninfected persons and > 5 mm in those with HIV. We retested HHCs with a prior negative TST at 6 and 12 months. We followed up with HHCs at 2, 6, and 12 months to assess TB symptoms and document interval diagnoses; we referred symptomatic participants for clinical evaluation and reviewed medical records to identify TB diagnoses during follow-up. Additional study design details have been published elsewhere ( 10 ).

Ethics committees at Harvard University and at the National Institute of Health in Peru approved the study. All study participants or their guardians provided written informed consent, and children <18 years old provided assent.

We stratified index patients by the presence of baseline symptoms included in the WHO 4-symptom screen (W4SS) ( 11 , 12 ): cough, night sweats, weight loss, and fever. We evaluated 4 outcomes: baseline TB infection among HHCs; incident infection over 6 months among those uninfected at baseline; incident infection over 12 months among those uninfected at baseline; and incident TB disease over 12 months among HHCs without coprevalent TB at baseline.

We classified HHCs as infected at baseline if they had TB disease or a positive TST at enrollment. To reduce misclassification of community-acquired infection as household transmission, we excluded HHCs with a prior history of TB disease or a positive TST from the analysis. We considered HHCs to have incident TB infection if they were uninfected at baseline and subsequently had a positive TST or experienced secondary TB disease. We considered HHCs to have incident TB disease if they received a diagnosis of TB disease at any time from 2 weeks after enrollment to the end of follow-up.

We included only index patients with microbiologically confirmed TB. To assess the association between index patient symptom status and baseline HHC infection, we used generalized estimating equations with modified Poisson regression, specifying an exchangeable correlation structure to account for clustering within households. We estimated prevalence ratios (PR) and 95% CIs. We evaluated associations with incident TB infection at 6 and 12 months and incident TB disease at 12 months using Cox frailty models and reported hazard ratio (HRs) and 95% CIs.

We first fitted univariable models, then multivariable models, adjusting for prespecified confounders and predictors of TB infection or disease. Covariates included index patient age group, HIV status, smoking, alcohol use, diabetes, and household socioeconomic status. Additional predictors of HHC infection included index patient sex and employment status and HHC sex, age group, HIV status, diabetes status, BCG vaccination status, smoking and alcohol use, and BMI category ( Appendix Table 1). A second multivariable model retained confounders and variables with p<0.10 in the initial multivariable analysis. Because sputum smear status might lie on the causal pathway between disease severity and transmission, we did not include it in regression models.

For analyses of TB infection, we restricted the primary sample to child HHCs 15 years of age because infection in that group is more consistent with recent household transmission. Sensitivity analysis included HHCs of all ages. Analyses of incident TB disease included HHCs of all ages, given the relative rarity of disease compared with infection. We conducted complete-case analyses, excluding 13% of observations ( Appendix Table 2), and assessed robustness using multiple imputation. Finally, we examined specific symptom patterns by classifying index patients into 4 groups: asymptomatic; cough only; no cough but > 1 of night sweats, weight loss, or fever; and cough plus > 1 of those symptoms. We evaluated associations between symptom patterns and baseline infection and 6-month incident infection among child HHCs.

We next explored factors associated with transmission or disease progression from asymptomatic index patients. For that exploratory analysis, we restricted the sample to HHCs of asymptomatic index cases. We estimated PRs for baseline TB infection and odds ratios (ORs) for incident TB infection at 6 months and incident TB disease at 12 months, to identify potential associated factors. Given the limited sample size in the subgroup, we limited analyses to univariable models examining associations between characteristics of asymptomatic index patients and each outcome among their HHCs.

We identified 3,109 microbiologically confirmed M. tuberculosis –infected index patients with known baseline symptom status ( Appendix Table 3), including 113 asymptomatic and 2,996 symptomatic patients. Asymptomatic patients were significantly more likely to be smear-negative (OR 3.15 [95% CI 2.14–4.65]; p<0.001).

We enrolled 12,230 HHCs ( Appendix Table 4), of whom 4,296 (35.1%) were 15 years of age ( Table 1 ). Most (80.9%) had a BCG scar; 62.4% were 30 years of age, 57.8% had normal BMI, 6.0% were current smokers, and 25.7% were current drinkers.

Among child HHCs, 23.0% exposed to symptomatic index patients and 15.0% exposed to asymptomatic index patients were TST-positive at baseline ( Table 2 ). Exposure to asymptomatic index patients was associated with lower baseline infection (crude PR 0.60 [95% CI 0.36–1.01]). Adjusted PR was 0.62 (95% CI 0.37–1.03) for model A, in which we adjusted for age, sex, HIV status, alcohol consumption status, BCG vaccination, and BMI category, and 0.61 (95% CI 0.36–1.03) in model B, in which we excluded sex and alcohol consumption from the adjusted variables. When all HHCs were included, the association was attenuated (adjusted PR 0.94 [95% CI 0.79–1.11]) ( Appendix Table 5). Multiple imputation yielded similar estimates, with lower prevalence among child HHCs (adjusted PR 0.58 [95% CI 0.35–0.95]) and comparable results among all HHCs (adjusted PR 0.88 [95% CI 0.74–1.04]) ( Appendix Table 6).

At 6 months, 14.5% of child HHCs exposed to symptomatic index patients and 10.3% exposed to asymptomatic index patients tested positive by TST ( Table 3 ). Exposure to asymptomatic index patients was associated with lower hazard of infection (crude HR 0.63 [95% CI 0.27–1.49], adjusted HR 0.62 [95% CI 0.26–1.51] in both multivariable models). Estimates were similar when all HHCs were included (adjusted HR 0.78 [95% CI 0.50–1.20]) ( Appendix Table 7).