**3. Discussion**

Immunotherapy with BCG is an effective adjuvant treatment for superficial urinary bladder cancer, recommended by the international associations of urologists [5]. The procedure is generally safe; however, a significant number of mostly local adverse reactions

causes a big proportion of dropouts from treatment—based on the study of Sylvester et al. only 25–29% of patients receiving BCG instillations completed the three-year treatment regimen [15]. Another study reported systemic side effects in around 30% of patients, who received BCG treatment, with malaise as the most frequent one (15.5%) [16]. Lung infection occurred in less than 1% of subjects [16]. In the most recent literature review, Liu et al. present much lower percentages of systemic complications of BCG treatment–between 0.4 and 5.7% for different entities, again with the lowest incidence of respiratory tract infection [6]. Therefore, in our institution, the tertiary pulmonary hospital, the patients with BCG treatment side effects present very rarely, only if the infection is located in the lungs. As highlighted in the literature, a high grade of suspicion must be maintained to establish a proper diagnosis, as the symptoms may be nonspecific and delayed in time from the last instillation, sometimes even for years [17], although the reported median interval between the procedure and the onset of symptoms was 8 days [7]. Our patient developed symptoms of acute infection a week after the last instillation, nevertheless at that time his complaints were not attributed by doctors to BCG immunotherapy.

The early recognition of BCG-related side effects is difficult. Unfortunately, there are no known risk factors, apart from those related to the procedure itself, such as injury of the urinary bladder during instillation, or acute urinary tract infection at the time of instillation [6,7,15]. Data on clinical case series did not find a relationship between the frequency of complications and patients' age, smoking status, or immune system disturbances, such as immunosuppressive treatment, chemotherapy, or splenectomy [6,7]. Our patient had a history of chronic pancreatitis resulting from incomplete pancreatectomy and post-traumatic partial splenectomy, without clear indices of immune system impairment. Additionally, he probably had liver cirrhosis with increased portal pressure and esophageal varices.

The radiologic recognition of BCG-related lung disease is also difficult. Lung involvement may present either as a result of the hematogenic spread of *M. bovis*, mimicking "miliary tuberculosis" [7,18–22] or—more frequently—as interstitial pneumonitis with bilateral ground glass opacities [6,13,23]. In children, BCG infection occurs rarely after vaccination and may present as pneumonic infiltrates and lymphadenopathy [24]. Ground glass opacities are usually assessed as hypersensitivity reactions rather than infection [25,26]. In our patient, radiologic changes in the lungs were mostly interstitial, micronodules were few, ground glass opacities and infiltrates dominated in CT scan, hilar lymph nodes enlargement was also present—the whole presentation at the first lung CT scan was interpreted as NSIP. Even at later stages, after lesion progression, the ground glass opacities predomination caused doubts among radiologists, if the BCG-osis is the most probable differential diagnosis. The time of symptom onset in the middle of the severe COVID-19 pandemic and quickly progressive respiratory failure, suggested the possibility of SARS-CoV-2 lung disease, as well. Nevertheless, repeated PCR tests for SARS-CoV-2 were negative.

The correct diagnosis was finally based on the identification of *M. bovis* BCG in sputum and bronchial fluid cultures.

The BCG vaccine was developed as an attenuated live vaccine, derived from virulent strains of *M. bovis* species. The BCG vaccine *M. bovis* strain cannot be differentiated from other members of the *Mycobacterium tuberculosis* complex (MTBC) solely on the basis of phenotypic tests. An example is the immunogenic protein MPT64, which is widely used as a diagnostic marker to differentiate MTBC from nontuberculous mycobacteria. However, bacillus Calmette-Guerin vaccine strains with deletion of the RD2 region do not secrete MPT64. Such culture isolates may be falsely identified as NTM [27,28].

With the introduction of molecular techniques, it has become also possible to differentiate virulent mycobacteria from non-virulent BCG strains. The development of multiplex polymerase chain reaction techniques has provided rapid, sensitive and specific differentiation of BCG vaccine strains [29]. PCR technology identified genomic regions RD1, RD2 and RD3 that were not found in vaccine strains. The loss of virulence was found to be due

to a specific regulatory mutation in the RD1 region. This genomic region is present in all virulent human and bovine strains but was not identified in all BCG vaccines [28].

In Poland, two BCG vaccines are used in the treatment of bladder cancer, BCG-Tice and BCG-Moreau. The use of two different strains for immunotherapy in the same population, one with the *mpb64* gene (BCG-Moreau), and the other without (BCG-Tice), further complicates the problem of correct identification. Furthermore, the use of a BCG strain negative for MPT64 will lead to its misidentification as NTM, in mycobacterial cultures [30].

Our patient was treated with BCG-Tice, therefore, the culture isolates were negative for MPT64 protein. This case highlights the importance of careful use of the MPT64 kit.

Positive cultures for *M. bovis* BCG are rarely reported in patients with pulmonary involvement. In the cohort of Perez-Jacoiste Asin et al. urine cultures were positive in three of 11 (27%) patients with two of them having a miliary pattern on chest X-rays [7]. Itai et al. reported positive sputum culture in a patient with solitary pulmonary nodule related to BCG-osis [31]. Positive blood cultures for BCG were reported by Osorio Aira et al. in a patient with a sudden onset of general symptoms immediately after instillation, who developed miliary tuberculosis [21].

Occasionally AFB are also found in lung tissue [32]. We did not take the lung sample for microscopic evaluation, due to increasing hypoxemia. This procedure may help to establish the diagnosis in the absence of MTB in the bronchial washings if the granulomas are found in lung specimens.

There are no established treatment guidelines for BCG infection. A three-drug regimen including RMP, INH and EMB is widely used [33].

In our patient, the additional problem was liver function impairment, probably due to liver cirrhosis. This influenced the treatment, as the anti-tuberculous drugs are hepatotoxic. After worsening liver function tests were noted, we decided to withdraw INH continuing RMP as the most active drug. BCG culture resistance to INH was found and supported our decision. Some strains of BCG have intrinsic resistance to INH and this drug may not be the best first choice in the empirical treatment of BCG infections [34]. In our patient, INH was replaced by levofloxacin, a fluoroquinolone that is active against mycobacteria. Fluoroquinolones, i.e., levofloxacin or moxifloxacin are used in the therapy of mycobacterial lung diseases, in case of anti-tuberculous drug intolerance, resulting in infection eradication [7,12,17,21,31].

The addition of systemic corticosteroids may be necessary for patients with severe respiratory insufficiency in the course of BCGosis [17,35]. Prednisone added to antituberculous therapy in this patient, resulted in the quick resolution of respiratory failure.

## **4. Conclusions**

Pulmonary BCGosis is a rare, but important complication of BCG treatment in patients with superficial bladder cancer. It must be taken into consideration in every patient with a history of such treatment, presenting with fever and pulmonary infiltrates, regardless of the time from the last instillation. To identify MPT64-negative clinical isolates of *M. bovis* BCG, it is necessary to use modern molecular methods of identification.

**Author Contributions:** K.L., writing—original draft preparation; A.L., data collection, patient's medical care; I.B., radiological scans preparation and description; M.K., microbiological tests performance; E.A.-K., writing—review and editing; W.T., supervision; M.S., writing—review and editing, supervision. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was approved by the Ethics Committee of the National Research Institute fo Tuberculosis and Lung Diseases (KB-64/2018, 14 December 2018) as a part of research on mycobacterial diseases in human. The patient gave written consent for publication.

**Informed Consent Statement:** Written informed consent has been obtained from the patient to publish this paper.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.
