Molecular Identification of Mycobacterium tuberculosis and Mycobacterium bovis in Vitreous Fluid Samples from a Patient with Presumed Isolated Ocular Tuberculosis Complicated by Antituberculous Drug Toxicity

1. Introduction

Mycobacterium tuberculosis remains one of the leading infectious causes of morbidity and mortality worldwide and continues to pose a major global health challenge [1]. Molecular diagnostic techniques have become an essential component of contemporary mycobacterial diagnostics, enabling the rapid detection of members of the Mycobacterium tuberculosis complex (MTBC) directly from clinical specimens and thereby facilitating timely clinical decision-making [1,2].

The MTBC comprises several closely related species, including M. tuberculosis, M. bovis, M. africanum, M. microti, and others. Comparative genomic studies have demonstrated that members of the MTBC exhibit greater than 99.9% nucleotide sequence identity and are distinguished mainly by specific genomic deletions, known as regions of difference (RDs), which enable species-level classification within the complex [3]. In particular, the absence of RD1, RD9, and RD12 regions has been widely utilized to discriminate M. bovis and Bacillus Calmette-Guérin (BCG) strains from M. tuberculosis [3,4,5].

An important clinical scenario is the treatment of non-muscle-invasive bladder cancer (NMIBC) with intravesical BCG, an attenuated strain of M. bovis. Since its introduction, BCG instillation has become the standard adjuvant therapy for high-risk NMIBC because of its proven efficacy in reducing recurrence and progression rates [6]. Although generally safe, BCG therapy may rarely result in local or systemic complications caused by persistence or dissemination of the attenuated mycobacteria [7,8].

The increasing implementation of highly sensitive molecular assays has introduced new diagnostic challenges in this setting. Because BCG belongs to the MTBC, nucleic acid amplification tests may detect BCG-derived genetic material and identify it as MTBC or specifically as M. bovis, depending on the analytical method employed. Consequently, positive molecular findings in specimens from patients with a history of intravesical BCG therapy may lead to diagnostic uncertainty and potentially mimic an infection caused by wild-type M. tuberculosis complex organisms [3]. Molecular methods targeting RD regions or other lineage-specific genomic markers may facilitate the distinction between M. tuberculosis, M. bovis, and BCG strains and therefore provide clinically relevant information for patient management [1,2,5]

Here, we present a case illustrating a potential diagnostic pitfall in molecular pathology, in which the detection of MTBC-associated DNA required careful integration of clinical history, pathological findings, and molecular characterization to distinguish BCG-related findings from classical tuberculosis infection.

2. Clinical Background

The patient whose clinical course is reported here after written informed consent is a 76-year-old Caucasian male with severe COPD (GOLD IV) who was and still requires long-term oxygen therapy and is receiving non-invasive ventilation at home. Other relevant pre-existing conditions documented include arterial hypertension, a history of renal and bladder carcinoma, and benign prostatic hyperplasia.

In March 2026, a bursectomy with surgical wound debridement was performed due to putrid trochanteric bursitis on the right side with a fistula tract. The microbiological specimens taken intraoperatively showed no evidence of bacterial infection at the time of discharge. Subsequently, a chronic fistula tract persisted. By the end of August, there were no signs of local or systemic infection; regular changes of antiseptic dressings were recommended.

3. Clinical Case

Ophthalmological History

In July 2026, a diagnostic vitrectomy with vitreous material collection was performed due to intermediate uveitis of the left eye. The relevant specimen was sent to the Institute of Pathology on 23 July 2026.

Cytological examination revealed cellular proliferation with a marked increase in lymphocytes, isolated neutrophils and only a few macrophages. No clear giant cells or other specific granulomatous changes were described. No atypical cells were detected. The findings were assessed as consistent with tuberculous uveitis, but not as morphologically specific.

In the initial molecular pathological examination using RealAccurate Quadruplex Mycobacteria PCR, DNA from the Mycobacterium tuberculosis complex was detected. An additional multiplex PCR test for a broad spectrum of viral, bacterial and fungal neurotropic pathogens proved negative.

Following the first vitrectomy, a peripheral retinal detachment without macular involvement occurred. On 31 July 2026, a repeat 23-gauge vitrectomy was performed, involving vitreous sampling, endolaser coagulation and silicone oil endotamponade. Two retinal foramina were treated. In addition, yellowish choroidal granulomas, clinically suspected to be tuberculous, were observed on the lower vascular arc and nasally to the optic disc. On discharge, the retina was circumferentially adherent beneath the silicone oil.

Molecular Pathological Findings and Mycobacterial Species Differentiation

Following notification of Mycobacterium bovis detection from an external source at a university hospital, the original vitreous material was re-examined in our molecular pathology department. Two specimens were tested positive for Mycobacterium tuberculosis complex by qPCR using the RealAccurate^® Quadruplex Mycobacteria PCR Kit (Pathofinder, Maastricht, The Netherlands), followed by subtyping using the VisionArray MYCO PreCise Master Mix 2.0 and the VisionArray MYCO Chip 2.0 (Zytovision/Zytomics, Bermerhaven, Germany), strictly following the manufacturer recommendations and EN/ISO 17020 accredited processes.

This analysis detected nucleic acid from Mycobacterium tuberculosis. The original detection was thus confirmed, and the organism was typed as M. tuberculosis within the Mycobacterium tuberculosis complex.

In parallel, an isolation attempt from a swab and ocular punctation taken at our Clinics for Opthalmology and send to our microbiology service resulted in growth of Mycobacterium bovis in two parallel attempts by our external microbiology services, also under accredited conditions (ISO 15189). The phenotypical subtyping revealed Mycobacterium bovis, BCG, i.e., the strain used for the bladder cancer treatment that was applied to the patient. Thus, although no Mycobacterium tuberculosis was identified by culturing because it was either present as a minor pathogen or no growth-competent strain but solely residual DNA was present, the results strongly indicated a co-infection or superinfection with culturable M. bovis being the likely dominant strain in the observed inflammatory setting.

Investigation for Extraocular Manifestations

A chest CT scan dated 3 August 2026 showed no evidence of pulmonary tuberculosis. Instead the presence of marked, apically prominent bullous pulmonary emphysema compatible with COPD was observed. Several sputum samples tested negative for tuberculosis. A bronchoscopy was not performed due to the very severe COPD.

Similarly, an abdominal CT scan dated 4 August 2026 revealed no evidence of an abdominal tuberculous manifestation. The documented urine analyses were unremarkable. Conventional organ-specific diagnostics initially revealed no evidence of a primary focus in the lungs, abdomen or urogenital tract. Consequently, the case was initially classified by an ophthalmologist as suspected isolated ocular tuberculosis.

Possible Neurological Involvement

Craniocerebral MRI scans on 23 July and 6 August 2026 revealed contrast-enhancing cortical and subcortical supratentorial lesions. These were assessed as presumably inflammatory but non-specific. Cerebral tuberculosis was possible on the basis of imaging morphology, but not confirmed. No significant progression was observed during follow-up.

Clinically, there was pronounced generalised weakness, psychomotor retardation, prolonged response latency and, as reported intermittently by relatives, confusion. Neurologically, a positive Babinski reflex on the left side, muscle reflexes that could be elicited predominantly on the left, and bilateral dysmetria were described. There was no evidence of meningism, focal paresis, aphasia or apraxia.

The Quantiferon test was negative, as were the serological tests for Borrelia and syphilis. The soluble IL-2 receptor was elevated at 7,125 U/ml.

On 19 August 2026, a lumbar puncture was performed without complications. Four tubes of slightly yellowish cerebrospinal fluid were collected for cell count, Reiber test, oligoclonal bands, IL-2 receptor, general culture, as well as tuberculosis-specific microscopy, PCR and culture. In the available records, the cerebrospinal fluid is initially described as showing no pathological findings.

Thus, a clinically suspected neurotuberculosis should currently be regarded only as a morphological or clinical suspicion, not as a confirmed diagnosis.

Anti-Tuberculosis Treatment and Drug Response

Following the detection of the pathogen in the eye, treatment with rifampicin 600 mg daily and isoniazid 300 mg daily was initiated upon discharge from the ophthalmology department on 8 August 2026. Although a four-drug regimen would have been indicated, due to the patient history of former renal carcinoma and risk of renal failure, a two-phase treatment regimen over three months was initially envisaged. Liver function tests were normal immediately prior to discharge.

Unfortunately, after a treatment phase of only six days of treatment initiation, the patient developed a marked deterioration, presenting with anorexia, very low fluid intake, adynamia, psychomotor retardation, jaundice and dehydration. Relatives reported increasing confusion and episodes that at times appeared hallucinatory.

Abdominal ultrasound revealed no biliary tract dilation, no focal liver lesions, no ascites and no evidence of a mechanical cause. Liver elastography was classified as F0–1. Renal failure, in the context of marked dehydration, was assessed as predominantly prerenal.

Rifampicin and isoniazid, as well as other potentially hepatotoxic and nephrotoxic medicines, were discontinued. Following fluid replacement and supportive treatment, there was a continuous improvement leading to near-normalisation of liver and kidney parameters. Consequently, the diagnosis of a severe drug-induced hepatopathy with a cholestatic-hepatitic pattern and concomitant prerenal acute renal failure was made.

Following further improvement, anti-infective therapy was resumed on 31 August 2026. The available records do not clearly indicate which regimen was used or whether there was renewed exposure to rifampicin or isoniazid.

Laboratory parameters observed at the hospital admission are shown in Table 1.

Concomitant Pulmonary Course

During the hospital stay, there was an exacerbation of COPD with a decrease in oxygen saturation and mild hypercapnia. The settings for non-invasive ventilation were adjusted on the patient’s own device. There were no indications of an acute pulmonary infection, so no antibiotic therapy was administered.

The patient’s diagnosis of COPD GOLD Stage IV is relevant to this case report because it significantly limited the diagnostic options – particularly bronchoscopy – as well as the choice and tolerability of therapeutic strategies.

Preliminary Overall Assessment

In summary, there is a molecularly confirmed infection of the left eye with a pathogen from the Mycobacterium tuberculosis complex. In the local molecular typing, M. tuberculosis was detected in the original vitreous material and the result was confirmed in a repeat test. At the same time, the clinical reports indicate that M. bovis was detected externally.

Morphologically, there was predominantly lymphocytic inflammation without definitive giant cells or specific granulomatous changes in vitreous fluid. Clinically, however, choroidal granulomas suggestive of tuberculosis were observed. Pulmonary, abdominal or urogenital tuberculosis was not detected in the available diagnostic investigations. Cerebral involvement remained a suspected diagnosis.

4. Discussion and Conclusion

The presented case is of general interest for several reasons. First, the case appeared to be a sole ophthalmologic issue. However, based on an established combination of classical cytology followed by a two-step molecular analysis combined with an isolation attempt a complex diagnosis emerged that correlates well the clinical observations.

Thereby, the presence of M. bovis BCG in the eye may not only have led to the ocular infection but also may have triggered a local immune response similar to its original intention during bladder instillation. This would explain the detection of M. tuberculosis DNA by qPCR and molecular subtyping, i.e., the diagnostical procedures were performed at the terminal phase of a co- or superinfection with M. tuberculosis and M. bovis BCG, while M. tuberculosis appeared to be either a minor strain or be already successfully defeated by the local immune reaction, while M. bovis was present in extension due to the prior bladder cancer treatment.

This latter treatment most likely was the source of the local infection, which is a timely reminder that strict hand hygiene is crucial especially for multimorbid patients as our patient described here. This is not only relevant for the inpatient, but also and maybe even more for the outpatient setting, which requires medical staff to repeatedly train patients and remind them on basic hygienic rules.

Moreover, the case is a reminder that proper anamnestic investigations are highly relevant. Some details, e.g., the prior BCG instillation, were reported after a first external detection of M. bovis on repeated requests. The subsequent detection of M. tuberculosis thereby was initially a random result, that under other circumstances and without the additional isolation attempt may have led to an insufficient therapy, especially as a history of a predisposing chronic lung disease was reported.

While the ophthalmic complications of BCG instillation are a known but very rare condition [9,10], the present case is novel as to the very best of our knowledge it is the first case of a co-detection also of M. tuberculosis, which may have been the less dominant pathogen in the presented setting maybe due to the local BCG induced immune reaction.

In addition, diagnostics were further complicated in consequence of a negative Quantiferon assay which was in contrast to the local detection of M. tuberculosis DNA, and also a possible cerebral involvement was discussed during the severe side effects of initialized antibiotic therapy. The latter thereby was limited due to GOLD stage IV COPD and severe early hepatic and renal toxicity whilst on rifampicin and isoniazid.

Taken together, the combination of in-depth diagnostics, consisting of several microbiological assays and molecular pathology investigations in concert with the in-depth clinical investigations led to the recovery of the patient who has been in convalescence after a few weeks now. As a consequence of this case, our internal test algorithms now include a request for information on BCG instillation in cancer patients and ophthalmological clinical courses, and mainly in combinations thereof.