Main finding
The study revealed that current smoking, Golimumab treatment, a history of TB, age > 40, Chloroquine treatment and gender significantly influence the occurrence of LTBI in rheumatic patients.
Rheumatic patients who smoke are at 1.50 times higher risk for LTBI than non-smokers. Smoking activates T-lymphocytes through inflammatory chemokines, promoting further lymphocyte recruitment to encapsulate MTB and form granulomas. Nicotine in cigarettes alters tumor necrosis factor-alpha (TNF-α) function, limiting its bactericidal effect on MTB, which increases the risk of LTBI. Additionally, nicotine affects T-lymphocyte activity, reducing their capacity to eliminate MTB [36]. Therefore, priority should be given to reducing nicotine intake among patients with rheumatic diseases, and screening smoking patients for LTBI should be emphasised. However, we also found that a history of smoking did not show a significant impact. It is speculated that the development of LTBI in rheumatic populations may be more strongly associated with recent exposure (current smoking), while the cumulative effects of past smoking may be diluted by long-term smoking cessation or immune system recovery. In addition, the definition of “history of smoking” in the included studies may be inconsistent, such as changes in smoking cessation duration or smoking intensity, leading to insufficient statistical ability to detect associations.
TNF-α is critical for granuloma formation and maintenance, and its inhibition may increase the risk of new TB infections or the reactivation of latent TB. However, TNF-α antagonists can improve rheumatic diseases by downregulating local and systemic pro-inflammatory cytokines, reducing lymphocyte activation and migration to the joint site, and other mechanisms. These benefits have led to the widespread use of TNF-α antagonists in clinical practice [37]. Common TNF-α antagonists include Infliximab, Etanercept, Adalimumab, Golimumab, and Certolizumab pegol [38]. Several studies have highlighted that TNF-α antagonist therapy significantly increases the risk of LTBI in rheumatic patients [38,39,40]. Our study shows that Golimumab increases the risk of LTBI in these patients, while Etanercept does not appear to influence LTBI occurrence. Structural and functional differences among TNF-α inhibitors account for this result. Compared to Etanercept, Golimumab has a stronger binding affinity for transmembrane TNF-α (tmTNF-α), which enhances its neutralizing effect on tmTNF signaling and increases the potential for fungal growth recovery in granulomas [41]. In addition, Golimumab can also increase the risk of LTBI infection by weakening the bactericidal ability of macrophages and inhibiting T cell immunity. Etanercept mainly binds soluble TNF-α and has a weak inhibitory effect on tmTNF-α, so its impact on granuloma stability is relatively small. Based on these findings, LTBI screening may be prioritized for patients receiving Golimumab. In theory, the mechanism of Adalimumab is similar to that of Golimumab. But our meta-analysis revealed that Adalimumab had no significant effect on the occurrence of LTBI in rheumatic patients. The trial showed that Golimumab and Infliximab had similar affinities for tmTNF-α, slightly higher than Adalimumab [42]. Such biological differences may account for the differences in the effects of Golimumab and Adalimumab on the induction of LTBI. In addition, this may be due to limitations in the included studies, with one being a single factor analysis result and the other being a multiple factor analysis result, which to some extent distorts the research findings. Although Al-Sohaim.et’s study [43] also showed that Adalimumab had no effect on the occurrence of LTBI in rheumatic patients, based on its mechanism of action and the shortcomings of existing research, we believe it is still necessary to screen rheumatic patients who have received Adalimumab treatment for LTBI. This meta-analysis indicates that not all TNF inhibitors increase the risk of LTBI. Given the limited existing literature, these findings should be confirmed through further research.
Wang Wen et al. [44] highlighted that individuals with a prior history of TB have a fivefold increased risk of reinfection compared to the general population. In our study, a history of TB was identified as a risk factor for LTBI in rheumatic patients. Rheumatic patients with a previous TB history should undergo LTBI screening, and the results should guide the decision on whether to initiate prophylactic treatment. Although several studies suggest that TB contact may increase the risk of LTBI in rheumatic patients, this trend was not observed in our study. Potential reasons include: (1) insufficient sample size, and (2) variability in TB contact reporting across studies, where self-reports lacked detailed information on the duration and frequency of exposure.
Previous research has shown that LTBI is more common in men within the general population [45]. As the primary workforce and economic providers, men tend to engage in more social and employment activities, thereby increasing exposure to MTB. Additionally, differences in hygiene habits between men and women may contribute to this disparity. A similar pattern was observed in our study, where male rheumatic patients had a higher risk of LTBI than female patients. Consequently, male patients should be prioritized in LTBI screening, facilitating early detection and prophylactic treatment to reduce TB risk.
Lysosomal autophagy is a key defense line for clearing intracellular pathogen [46]– [47]. Normal autophagy can degrade invading pathogens (such as MTB and Salmonella) through lysosomes, while when autophagy function is inhibited, pathogens are more likely to escape immune clearance and form latent infections [48]. Chloroquine is a classic lysosomal autophagy inhibitor that neutralizes the acidic environment of lysosomes (pH elevation), prevents the fusion of autophagosomes with lysosomes, or interferes with lysosomal enzyme activity, resulting in the inability to degrade autophagic substrate [49]– [50]. Although there is no direct evidence in previous studies to suggest a correlation between the use of chloroquine and latent infections associated with mycobacteria. But this study shows that chloroquine can increase the risk of LTBI in rheumatic patients. This may be due to immune dysfunction in rheumatic patients, and long-term use of chloroquine may cause latent. Thus, screening for LTBI in rheumatic patients on these medications warrants emphasis.
Previous studies have shown that the prevalence of latent tuberculosis infection in the population is associated with age [51]– [52], with infection rates increasing with age. This is likely due to the decline in immune function with age, which affects the body’s ability to clear MTB. In rheumatic populations, a similar phenomenon is observed, with the risk of LTBI being higher in the group aged over 40 years compared to those aged 40 years or younger. Therefore, it is important to prioritize LTBI screening for rheumatic patients studies aged over 40 years.
Furthermore, the effects of various factors, history of smoking, corticosteroids, duration of disease, RF factor, history of diabetes, TB contact and history of BCG vaccination were analyzed. The results indicated that none of these factors significantly influenced the development of LTBI in rheumatic patients. It is worth noting that the meta-analysis from three studies [18, 20, 30] indicated immunosuppressive therapy did not significantly influence the occurrence of LTBI in rheumatic patients. However, it is important to note that “immunosuppressants” encompass a broad spectrum of agents, including traditional chemical immunosuppressants, biologics, targeted synthetic drugs (e.g., JAK inhibitors), and glucocorticoids. These drugs differ substantially in their mechanisms of action and their associated risks of LTBI. Consequently, when analyzed collectively, their opposing effects may offset one another, potentially masking subgroup-specific risks.
This study found that the pooled LTBI rate in rheumatic patients was 22%, comparable to the 20.7% found by Salma Athimni et al. [53]. The discrepancies between these studies may be attributable to differences in disease types, treatment regimens, and diagnostic methods employed. Given the considerable heterogeneity in our results, additional subgroup analyses were conducted based on ethnicity and LTBI detection methods. Subgroup analyses based on ethnicity revealed a reduction in heterogeneity, with the heterogeneity in the mixed-race group falling to 0, and that in Asian populations (the yellow race) also decreasing, suggesting that ethnicity may contribute to the observed variability. Our study indicated that the white, mix race and yellow rheumatic patients had a higher incidence of LTBI compared to black patients (other race), highlighting the influence of ethnicity as an intrinsic factor in the occurrence of LTBI. This may be related to economic, social and genetic factors. For example, white people are more likely to receive TB screening and use biological agents than black people, which increases the risk of TB. Since the black subgroup was based on only one study, the effect estimate may be unstable. More studies involving black people are needed in the future to validate this result. Further subgroup analyses based on LTBI detection methods revealed a significant reduction in heterogeneity within the PPD-based subgroup, while no substantial change was observed in other subgroups. This suggests that the choice of detection modality may account for some of the heterogeneity in study results. Previous meta-analyses have shown that IGRAs are more specific for low-risk, BCG-vaccinated individuals and more sensitive in diagnosing MTB co-infection in HIV-infected individuals [54]. Discrepancies between TST and IGRA results are common in LTBI testing. Notably, IGRA accuracy can be improved by extending the latency period of the stimulus and measuring IL-2 levels. However, our study found no statistically significant differences in LTBI infection rates between subgroups based on the detection method. In the preliminary phase of the study, subgroup analyses were performed based on TNF-α antagonist treatment status. Unfortunately, due to the limited data on drug treatments provided in the included studies, this analysis could not be conducted.
The advantage of this study was that it provided strong evidence for whether rheumatic patients are included in the target population for TPT, and clarified some risk factors for the occurrence of LTBI in rheumatic patients, providing a basis for determining which rheumatic patients should be prioritized for LTBI screening. Several limitations were present in our study: (1) Cross-sectional studies included in the analysis exhibited high bias due to issues with control selection and between-group comparability, while cohort studies had biases related to the selection of non-exposed populations, between-group comparability, and loss to follow-up. These factors may have influenced the reliability of our findings. In addition, most of the included literature were cross-sectional studies, which have the shortcoming of recall bias and may affect the accuracy of the research results. (2) Despite the inclusion of several influential factors in the studies, the limited number of studies meeting the inclusion criteria constrained our ability to perform meta-analysis on some of these factors. And the included studies lacked factors such as socioeconomic status and comorbidities (other than diabetes), so further research is needed to address these gaps. (3) Some of the forest plots exhibited moderate to high statistical heterogeneity, which could be attributed to the notable variability in results across the studies.