Our study found that standard anti-TB therapy effectively treated clinically suspected tuberculous uveitis in 11 patients, despite the absence of detectable MTBC DNA in intraocular fluid via multiplex PCR, highlighting the therapeutic benefit of anti-TB treatment in patients with strong clinical indications of tuberculous uveitis. The paucibacillary nature of intraocular fluid significantly limits the sensitivity of PCR, a challenge well-documented in the literature [25, 26]. Previous studies have shown that the low number of pathogens and limited sample volumes often result in negative PCR results despite clinical evidence of infection [27, 28].

As uveitis is associated with infection, obtaining evidence of the pathogen in ocular tissue or intraocular fluid is crucial for diagnosing tuberculous uveitis. However, enhancing the sensitivity of MTBC detection in intraocular fluid remains a challenging issue. Histopathology has confirmed tuberculous uveitis as a paucibacillary infectious disease, characterized by a low number of intraocular pathogens and limited sample volumes, severely restricting the utility of culture and smear techniques [29]. PCR detection methods can overcome these limitations by exponentially amplifying nucleic acids, significantly improving detection sensitivity [30]. The introduction of PCR to diagnose tuberculous uveitis by scholars has notably enhanced detection rates. Kotake et al. [31] pioneered the use of PCR to detect the MPB64 target gene of MTB in the aqueous humor of patients with tuberculous retinal vasculitis, demonstrating its clinical utility in this context.

Indian scholars have reported that multiplex PCR detection can enhance the sensitivity of detecting MTBC in intraocular fluid. Sharma et al. [14] were among the first to utilize multiplex PCR targeting IS6110, MPB64, and protein b genes in patients clinically suspected of tuberculous uveitis, achieving a sensitivity of 77.8% and specificity of 100%. This sensitivity was higher than that of single-gene approaches such as IS6110 or MPB64 aloneand correlated with favorable treatment responses. Praveen et al. [32] conducted multi-target PCR on 114 samples suspected of ocular TB, detecting IS6110 and MPB64 sequences in 80 samples. Only 8 samples tested positive for IS6110 alone. Among the 71 patients who received anti-TB therapy, 91.5% showed positive treatment outcomes. Mohan et al. [33] similarly employed multiplex PCR to assess 13 patients suspected of tuberculous uveitis. Among those with positive MPB64 and protein b sequences, only one patient tested positive for the IS6110 sequence. Considering that 97.1% of MTBC isolates harbor the CYP141 gene, targeting this gene can achieve a sensitivity of 85.7% [15,16,17], significantly higher than the sensitivity of the protein b gene (22.2%) observed in previous studies. This current study aimed to detect MTBC in intraocular fluid of patients with clinically suspected tuberculous uveitis using multiplex PCR to enhance detection sensitivity. Three primer pairs targeting MPB64, CYP141, and IS6110 were designed based on previous reports from Indian scholars [14, 15, 32]. The specificity of these primers and the feasibility of experimental conditions were validated through repeated testing using DNA from internationally recognized standard TB strains and confirmed TB patients’ body fluids.

In this study, we utilized multiplex PCR to detect MTBC DNA in the intraocular fluid of 15 patients (18 eyes) with clinically suspected tuberculous uveitis, encompassing 12 cases of aqueous humor and 3 cases of vitreous body. However, no DNA amplification bands were observed in the lanes targeting the MPB64, CYP141, and IS6110 genes. Despite negative PCR results, based on clinical manifestations, multimodal imaging, and serological indicators, we maintained a high suspicion of tuberculous uveitis. Accordingly, all 15 patients received diagnostic anti-TB treatment, with those showing effective response subsequently receiving standard anti-TB therapy for at least 6 months. Ultimately, 11 patients (13 eyes) with suspected tuberculous uveitis responded effectively to treatment. In this study, despite effective treatment leading to a clinical diagnosis of tuberculous uveitis in 11 patients, modified multiplex PCR did not detect DNA fragments in their intraocular fluid. This contrasts with previous literature findings, prompting additional analyses. Firstly, we critically reviewed the multiplex PCR methodology employed, selecting IS6110, MPB64, and CYP141 as target genes, consistent with studies by scholars [14, 17, 20]. Secondly, we meticulously designed primers for these targets and rigorously verified their specificity, experimental protocols, and conditions using DNA from the TB standard strain H37Rv, as well as DNA samples from alveolar lavage fluid, cerebrospinal fluid, and pleural effusion of patients clinically diagnosed with pulmonary TB, tuberculous meningitis, and tuberculous pleurisy in our region. All intraocular fluid samples were promptly processed upon collection, utilizing the same QIAGEN DNA extraction kit from the study referenced by Indian scholars [14, 34], with extraction conducted strictly following manufacturer instructions and using consistent sample volumes (100 µL).

We reviewed the literature again and identified limited reports on PCR detection of MTBC in intraocular fluid from patients with tuberculous uveitis in China. Apart from a single case where MTBC DNA was detected by tissue PCR post-ocular evisceration [35], no positive results have been reported in relevant studies using PCR technology to detect MTBC in intraocular fluid [31,32,33,34]. Mao et al. in China studied 46 patients with retinal vasculitis and 7 with choroidal tuberculoma who received effective anti-TB treatment. Among these, 5 patients underwent PCR testing for MTBC with no positive results observed [36]. Similarly, Jolly et al. [37] reported that among 19 patients suspected of ocular TB in Hong Kong from 2014 to 2019, 8 patients had negative results in PCR detection of Mycobacterium TB in intraocular fluid. Additionally, other studies noted that despite using RT-PCR, no positive MTBC results were found in two uveitis patients diagnosed with pulmonary nodules [38]. These findings prompted us to analyze potential causes:

The clinical presentation of tuberculous uveitis exhibits variability across different races and regions, leading to diverse clinical phenotypes. These variations suggest differing pathogenic mechanisms for tuberculous uveitis. Currently, two well-recognized pathogenic pathways include direct invasion by MTBC and autoimmune reactions triggered by molecular mimicry [39]. Tabbara et al. [40] proposed that endogenous endophthalmitis, chorioretinitis, and nodular scleritis primarily result from MTBC invasion, whereas anterior uveitis and retinal vasculitis are predominantly immune-mediated. Cyclitis, choroiditis, and multifocal choroiditis may involve elements of both pathways. Rishi et al. [41] demonstrated a high PCR detection rate (up to 100%) for MTBC in cases of active endophthalmitis, contrasting with lower sensitivities observed in non-rapidly progressive, non-suppurative ocular inflammations. Chinese patients more frequently present with retinitis and multifocal choroiditis [37, 42], potentially contributing to the lower MTBC detection rates in intraocular fluid observed in China. In our study, 76.92% of eyes with effective anti-TB treatment exhibited retinal vasculitis, suggesting that immune reactions may be the predominant pathogenic mechanism underlying tuberculous uveitis among Chinese patients. Other factors contributing to the poor detection of Mtb DNA in ocular fluids include the endemicity of TB, anatomical location and load of Mtb, and the integrity of the blood-retinal barrier [11, 43]. Additionally, our study noted that only 46.2% of patients had retinal vasculitis, which may further explain the low detection rates of Mtb DNA in our samples.

Second, in the context of detecting MTBC DNA in Chinese intraocular fluid using PCR techniques, several key considerations warrant further investigation. Firstly, while previous studies, primarily conducted by Indian scholars [14, 15, 32], have identified MPB64, CYP141, and IS6110 as effective target genes with high detection rates, the applicability of these targets in the Chinese population requires validation. Exploration of additional target genes specific to the Chinese demographic could enhance diagnostic sensitivity. Secondly, MTBC possesses a robust cell wall comprising mycolic acid, peptidoglycan, lipoarabinoglycan, and arabinogalactan [44], which provide natural protection against external factors, including DNA extraction methods. Overcoming technical barriers associated with breaking down this cell wall is crucial for improving DNA extraction efficiency. Furthermore, MTBC is a facultative intracellular pathogen residing within host cells, where it may undergo phagocytosis and lysosomal degradation. Consequently, some MTBCs may release DNA into intraocular fluid following interactions with phagocytes and lysosomes. Despite attempts in our study to directly detect MTBC DNA in intraocular fluid without prior extraction, no amplification bands were detected. Future research could explore the specific mechanisms of MTBC interaction with immune cells, particularly retinal pigment epithelial (RPE) cells known for their phagocytic capabilities similar to macrophages [45]. Addressing these challenges and exploring new avenues in PCR detection methodologies could advance our understanding and diagnostic capabilities for tuberculous uveitis in the Chinese population.

Thirdly, latent infection with MTBC may contribute to alterations in the body’s immune status, potentially leading to autoimmune or autoinflammatory diseases, and it has been suggested that latent MTBC infection could trigger abnormal host responses to self-antigens, thereby promoting conditions such as autoimmune uveitis [46]. Yang et al. [47] have investigated the association between TB infection and various forms of non-infectious uveitis. Their large-scale data analysis indicated that a positive T-SPOT status, indicative of latent TB infection, is an independent risk factor for diseases like Behçet’s disease. This finding underscores the potential immunological consequences of latent MTBC infection. Additionally, Peng et al. [48] have discussed the association between TB and specific forms of uveitis, such as serpiginous choroiditis. They noted characteristic clinical and fundus imaging features in cases of serpiginous choroiditis with positive PPD (purified protein derivative) and T-SPOT tests, suggesting a link to TB infection.

In the context of diagnosing tuberculous uveitis, multimodal imaging tests play a crucial role alongside clinical presentation, especially when PCR detection of MTBC in intraocular fluid yields negative results. Various findings from multimodal imaging can provide significant clues suggestive of tuberculous uveitis. These include granulomatous anterior uveitis associated with extensive posterior synechiae, retinal vasculitis with or without choroiditis, and serpiginous choroiditis, among others [49]. In our study, despite negative results from multiplex PCR testing for MTBC in intraocular fluid, we diagnosed 15 patients with clinically suspected tuberculous uveitis based on their ocular clinical presentations and comprehensive multimodal imaging examinations. Importantly, we opted to continue anti-TB therapy for these patients, leading to successful treatment completion in 11 patients. This underscores the clinical decision-making process where the absence of positive PCR results does not definitively rule out tuberculous uveitis, especially in the presence of suggestive clinical and imaging findings.

This study had several limitations that should be clarified. First, PCR detection necessitates low nucleic acid concentration in samples, theoretically allowing for testing of aqueous humor, vitreous stock, and diluted vitreous lavage fluid samples [50, 51]. However, due to cost and logistical constraints, primarily aqueous humor samples were used. Future research could consider including additional vitreous humor samples if feasible. Second, the study’s sample size was limited. Future studies with larger samples could be conducted to better characterize the clinical features of patients who respond effectively to anti-TB therapy.