The study of Lee et al. \cite{Lee_2024} estimated the risk of incident tuberculosis (TB) in those diagnosed with latent TB infections (TBI) following screening. Under the current program in the Republic of Korea, only those under the age of 65 with TBI are recommended to undergo TB prevetive therapy (TPT), which is covered by national health insurance. However, despite this provision, the incidence rate of active TB among those aged over 65 was five times higher than those under age 65, sparking vigorous debate on whether insurance coverage for TPT should be extended to all age groups. Lee et al. suggested that TPT might be less effective in reducing the risk of TB incidence among those aged over 60 (60+), with estimated adjusted hazard ratios (HRs) above one, albeit insignificant (Figure 1 in \cite{Lee_2024}). However, the interpretation of this finding requires caution. Considering the significant impact of comorbidities on increasing the risk of TB, individuals with comorbidities are more likely to be assigned to TPT and more likely to develop TB (Table 3 in \cite{Lee_2024}), suggesting that the presence of comorbidity serves as “confounding by indication” (Fig. 1) \cite{Kyriacou_2016}. While the study attempted to address this issue through multivariate adjustment, residual confounding \cite{Vansteelandt_2014} may persist, particularly when there is a notable imbalance in the comorbidity distribution between TPT recipients and nonrecipients. Specifically, the distribution of individuals with comorbidities within each age group, which is lacking in (1), may skew toward the TPT group, particularly among 60+, who are known to have a higher prevalence of comorbidities than the younger population. In this scenario, the multivariable adjustment, heavily relying on model extrapolation with limited empirical information on the TPT effect among those without comorbidities, might not fully mitigate confounding factors \cite{Vansteelandt_2014}.