Interpreting Nonsevere Tuberculosis Criteria in Children From Low-incidence Settings

To the Editors: We read with interest the article by Heimbach-Weber et al reporting a retrospective cohort study including 159 children younger than 16 years treated for tuberculosis (TB) between 2010 and 2023 at the Austrian National Reference Centre for Paediatric Tuberculosis, in which 104 patients (65.4%) initially met the World Health Organization (WHO) criteria for nonsevere disease, as defined according to the SHINE trial.1,2 Closely mirroring these findings, in our study based on data from the Spanish Pediatric TB Research Network (pTBred), which included 780 children with TB, 61.1% were classified as having nonsevere disease.3 Across both studies conducted in high-income, low-TB-incidence settings, children with nonsevere TB were more commonly identified through contact tracing (74% and 60% in the Vienna and pTBred studies, respectively), were more often asymptomatic at diagnosis (74% and 38%), and showed lower rates of microbiologic confirmation (36% and 32%) than children with severe TB, collectively indicating early diagnosis, very mild forms of disease, and a low bacillary burden in a large proportion of patients. Accurate classification of TB severity in children is clinically relevant, as recent WHO guidelines recommend a shortened 4-month treatment regimen for children with nonsevere, drug-susceptible TB, based on evidence from the shorter treatment for minimal tuberculosis in children (SHINE) trial.4 However, as the SHINE trial was conducted exclusively in low- and middle-income, high-TB burden settings, evidence supporting the safety and effectiveness of shortened treatment regimens for children with nonsevere TB in high-income, low-incidence countries such as Austria or Spain remains limited. In the Vienna study, the authors performed additional analyses incorporating chest computed tomography findings and bronchoscopy results, only performed in a subset of patients, and used a conservative assumption that all positive Xpert MTB/RIF (Cepheid, Sunnyvale, CA) results reflected a medium or high bacillary load; based on this, 28 of 104 children initially classified as having nonsevere TB were reclassified as having severe disease, reducing the proportion ultimately considered nonsevere from 65.4% to 47.8%. The Austrian investigators conclude that using the WHO criteria for nonsevere TB in isolation carries a substantial risk of underestimating disease severity, which may result in insufficient treatment and potentially lead to treatment failure. In the SHINE trial, however, neither chest computed tomography nor bronchoscopy was part of the study. Xpert MTB/RIF testing was systematically performed, and Mycobacterium tuberculosis was detected in 85 of 1202 children (7%); among cases with available semiquantitative results, all were classified as low (n = 11) or very low (n = 51), with no medium or high bacillary loads reported.2 Molecular test categories in WHO guidance are primarily intended to support diagnosis and treatment decision-making, rather than being validated markers of disease severity.4 Finally, all children enrolled in the SHINE trial were required to be symptomatic at diagnosis, in contrast to the European cohorts, in which a substantial proportion of children were asymptomatic. In summary, while we agree that the results of the SHINE trial cannot be directly extrapolated to all low-incidence, high-resource settings such as Austria or Spain, we believe that its results should likewise not be dismissed on the basis of diagnostic investigations that were not part of the SHINE study design. In pediatric TB, a careful and conservative approach to treatment decisions is always recommended; nevertheless, available data indicating earlier diagnosis and generally milder disease in contemporary European cohorts would seem to support, rather than challenge, the rationale for shorter treatment regimens. Importantly, systematic long-term follow-up of children treated with shortened regimens will be essential to confirm safety and effectiveness across settings.