Lacticaseibacillus paracasei bacteremia and mediastinitis secondary to esophageal invasion by left hilar pleomorphic carcinoma: case report

Abstract

Lacticaseibacillus paracasei bacteremia associated with probiotic use is rare. We report a 71-year-old man with advanced lung cancer who developed L. paracasei bacteremia while ingesting a commercially available probiotic-fermented milk. Autopsy demonstrated direct tumor invasion and perforation of the mid-esophagus, providing a structural route for bacterial translocation from the gastrointestinal lumen into the bloodstream. Genetic analysis showed that the blood isolate was genetically identical to the probiotic strain contained in the ingested product. Bacteremia resolved with antibiotics, but the patient died of respiratory failure from progression of the malignancy. This case highlights that probiotic organisms, despite an established safety profile, can cause bacteremia when immunocompromise coexists with disruption of the gastrointestinal barrier.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-026-12971-x.

Introduction

Lactobacilli are widely found in nature, are commensal microbes in humans, and are commonly used as probiotics [1]. The genus Lacticaseibacillus is generally recognized as safe and extensively used as probiotics. It rarely causes systemic infections such as bacteremia, accounting for approximately 0.1% of all bacteremia cases [2]. The L. paracasei strain Shirota, in particular, has been shown in animal models to have negligible pathogenicity compared with clinical isolates [3]. Reports of bacteremia due to this strain are extremely rare, with only one previous case documented in the literature [4]. We report the second case of bacteremia caused by L. paracasei strain Shirota in a patient with advanced lung cancer that had invaded the esophagus. This case report was written in accordance with the CARE guidelines.

Case presentation

The patient was a 71-year-old man with a medical history of type 2 diabetes and hypertension, both under treatment. He had a 54 pack-year smoking history (30 cigarettes/day for 36 years). Forty-two days before admission, he presented to a local physician with recurrent postprandial vomiting. Chest CT revealed a posterior mediastinal mass, raising suspicion of a mediastinal tumor (Fig. 1). Upper gastrointestinal endoscopy demonstrated external compression of the esophageal wall. Twenty-four days before admission, endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) of an enlarged mediastinal lymph node established the diagnosis of lung cancer (giant cell carcinoma, BRAF V600E-positive, high PD-L1 tumor proportion score).

Fig. 1.

Chest CT scan with contrast taken 42 days before admission. A mass shadow was observed in the posterior mediastinum at the level below the tracheal carina, compressing the esophagus and surrounding the thoracic aorta and left main bronchus

He was subsequently referred to our department three days before admission. No anticancer therapy had been initiated at the previous hospital. Given the advanced disease, we planned outpatient staging with PET-CT, followed by admission to initiate treatment. PET-CT performed one day before admission revealed a large hypermetabolic mass in the left pulmonary hilum with extensive mediastinal invasion and necrosis. Widespread metastases were identified in the lymph nodes, pleura, peritoneum, muscles, and bones, consistent with clinical Stage IVB disease (T4N3M1c2) (Fig. 2).

Fig. 2.

PET-CT scan taken the day before admission. A large, hypermetabolic mass was found in the left pulmonary hilum with extensive mediastinal invasion and necrosis. Furthermore, widespread metastases were found in the lymph nodes, pleura, peritoneum, muscles, and bones, consistent with clinical stage of IVB (T4N3M1c2)

On admission day, he was brought to the emergency department due to worsening general malaise and dyspnea. On arrival, his vital signs were as follows: temperature 39.0 °C, heart rate 110 beats/min, and respiratory rate 28 breaths/min, with an SpO₂ of 94% on 5 L/min of oxygen via a face mask. Physical examination revealed crackles in his chest. Serum C-reactive protein was 33.88 mg/dL. Given the high fever and elevated inflammatory markers, severe infection was suspected. Chest CT showed that the posterior mediastinal mass, containing foam-like material resembling food debris, had extended to the right mediastinum, suggesting an esophagomediastinal fistula. The lesion also extended into the right lung (Fig. 3).

Fig. 3.

A chest CT scan taken the day before admission. It showed that the posterior mediastinal mass, which contained foam-like shadows resembling food residue, had expanded to the right side of the mediastinum, suggesting a communication between the tumor and the esophagus. The shadow also extended into the right lung

He was admitted with a diagnosis of sepsis secondary to mediastinitis and pneumonia, and meropenem was initiated. He was managed with bowel rest and started on central venous nutrition. Due to his unstable general condition, mediastinal drainage was not performed. Although his inflammatory markers and oxygenation initially improved, chest CT on Day 8 showed progression of right lung infiltrates (Fig. 4). Sputum Gram stain showed gram-positive cocci, therefore linezolid was added for suspected MRSA. Blood cultures obtained on admission grew L. paracasei in all four bottles. Although antimicrobial susceptibility testing was attempted, accurate interpretation was not possible due to the lack of established interpretive criteria (breakpoints) for this organism. Therefore, antibiotic selection was guided by previously published data regarding the intrinsic susceptibility profile of L. paracasei. Meropenem was changed to ampicillin/sulbactam on Day 8, but the regimen was switched to clindamycin on Day 9. His respiratory status and inflammatory markers improved over time. However, recurrent aspiration due to esophageal obstruction increased sputum volume and led to repeated airway obstruction starting on Day 14. He died on Day 21 of hospitalization from respiratory failure.

Fig. 4.

Chest CT scan on day 8. The posterior mediastinal tumor had expanded, and the shadow in the right lower lobe had progressed to lung abscess. The right pleural effusion had also increased

To determine the source of bacteremia, the patient’s blood isolate was sent to the Yakult Central Institute. Species-specific and strain-specific PCR [4] and ELISA [5] confirmed that the isolate was genetically and immunologically identical to the L. paracasei (the previous taxonomic nomenclature was Lactobacillus casei) strain Shirota contained in commercially available probiotic-fermented milk (Fig. 5). On detailed questioning, he reported that he had voluntarily been regularly consuming this probiotic-fermented milk containing the Shirota strain to “boost immunity,” as solid food intake had been limited by esophageal stenosis.

Fig. 5.

Molecular and immunological identification of Lactobacillus paracasei isolated from blood culture. (A) PCR analysis was performed using L. paracasei species-specific primer sets (Lanes 2–4) and L. paracasei strain Shirota–specific primer sets (Lanes 6–8). PCR products were resolved on a 1.5% agarose gel. Lanes 1 and 5, 100 bp DNA size markers; Lanes 2 and 6, L. paracasei ATCC 25,302T; Lanes 3 and 7: L. paracasei strain Shirota (positive control); Lanes 4 and 8, patient blood isolate. Amplified products were detected in both species- and strain-specific PCR, confirming genetic identity with the Shirota strain. (B) ELISA using monoclonal antibody specific to the L. paracasei strain Shirota No positive reaction was observed with the type strain of L. paracasei ATCC 25,302^T (well 1), but positive reactions were observed with the positive control Shirota strain (well 2) and the patient blood isolate (well 3). These results confirm that the bloodstream isolate was immunologically identical to the strain contained in the commercially available probiotic-fermented milk

Pathological autopsy revealed pleomorphic carcinoma of the left pulmonary hilum (predominantly giant cell component), staged as pT4N3M1c2 (Figs. 6 and 7). Tumor invasion of the mediastinum and esophagus was evident, along with metastases to the right pleura, right diaphragm, thyroid, stomach, small and large intestines, mesentery, lesser omentum, and lymph nodes (supraclavicular, mediastinal, and hilar). Additional findings included mediastinitis, pulmonary abscess, bilateral aspiration and bacterial pneumonia, candidiasis, and pulmonary edema. The cause of death was determined to be respiratory failure due to the primary malignancy.

Fig. 6.

Gross pathological findings of the mediastinum after fixation. The tumor extended from the hilar lesion to encircle approximately three-quarters of the circumference of the descending thoracic aorta, with invasion into the mid-esophagus. At the cut surfaces, the esophageal wall was disrupted (dashed lines indicate the muscular layer; solid lines indicate the epithelium). Enlarged right supraclavicular lymph nodes were also observed. The site of esophageal rupture was in continuity with a lung abscess in the right lower lobe, extending into the mediastinum

Fig. 7.

Histopathological and immunohistochemical findings of the tumor. (A) Hematoxylin and eosin (H&E) staining shows diffuse proliferation of markedly atypical giant cells and spindle-shaped cells. (B) High-power view demonstrates pleomorphic tumor cells with large nuclei, prominent nucleoli, and multinucleation. Some areas contain intermingled spindle cells. (C–H) Immunohistochemistry: the tumor cells were diffusely positive for epithelial markers (AE1/AE3, CK5/6, MOC31) and also positive for the mesenchymal marker vimentin. A subset of spindle cells showed positivity for SMA. In contrast, markers for squamous cell carcinoma (p40), adenocarcinoma (TTF-1, napsin A), mesothelioma (WT1, D2-40, calretinin), and other lineage markers (S100, desmin, CDK4, MDM2, SMARCA4) were negative. These findings support the diagnosis of pulmonary pleomorphic carcinoma

Although giant cell carcinoma component predominated, spindle cell carcinoma elements were also present, supporting the diagnosis of pleomorphic carcinoma. The tumor originated in the left pulmonary hilum, extensively invaded the mediastinum, and ruptured the mid-esophagus wall. A lung abscess was present in the right lower lobe adjacent to the esophageal perforation site. Aspiration pneumonia and bronchopneumonia were observed throughout both lungs.

L. paracasei was not identified on histopathology. However, gram-positive and gram-negative cocci, as well as Candida, were observed in the lungs, consistent with polymicrobial aspiration pneumonia. Postmortem blood cultures identified Enterococcus faecium, suggesting that the infection control remained incomplete.

Discussion

This case represents an extremely rare instance in which L. paracasei strain Shirota, a probiotic with established safety, became the caused bacteremia in the setting of multiple predisposing factors. The primary pathophysiological mechanism was likely bacterial translocation facilitated by tumor-related perforation of the esophageal wall. Autopsy findings confirmed that the mediastinal tumor (a pleomorphic carcinoma) had invaded and perforated the esophageal wall. Furthermore, because of severe esophageal stenosis, the patient had voluntarily been regularly consuming liquid probiotics to “boost immunity.” This frequent exposure of the perforation site to a high load of viable organisms likely facilitated direct translocation of the organism into the bloodstream.

Invasive infections by Lactobacillus species are uncommon because of their low virulence. In a rabbit model of infective endocarditis, Asahara et al. demonstrated that the pathogenicity of L. paracasei strain Shirota was minimal, with rapid clearance from the bloodstream and from cardiac vegetations after inoculation [3]. These findings suggest that, in immunocompetent hosts, this strain is effectively eliminated by innate immune defenses, including macrophage-mediated bactericidal activity [3]. However, our case highlights that structural disruption of the gastrointestinal tract is a key risk factor. Additional contributing factors included the patient’s immunocompromised state due to advanced malignancy and malnutrition.

The Shirota strain has been reported to have low susceptibility to many empiric antimicrobials, with low susceptibility to most aminoglycosides, cephalosporins, quinolones, carbapenems, β-lactams, fosfomycin, sulfamethoxazole, and trimethoprim [6]. In contrast, it remains susceptible to tetracycline and macrolides [6]. Accurate identification of the causative organism is therefore essential to guide appropriate antimicrobial therapy. In this case, recognition of the patient’s probiotic intake and the known susceptibility profile of the Shirota strain supported switching therapy to clindamycin, which was associated with clearance of bacteremia. Unfortunately, the advanced progression of the underlying malignancy precluded recovery; however, improvement in inflammatory parameters and autopsy findings support that antimicrobial therapy controlled the Shirota strain bacteremia.

Although probiotics are generally considered safe, serious complications such as bacteremia and infective endocarditis have been documented [7]. This is the second reported case of bacteremia caused by an L. paracasei strain genetically identical to the Shirota strain. The first involved an 8-month-old girl who developed bacteremia after ingesting probiotics containing the Shirota strain following jejunojejunostomy for intestinal ganglion cell hypoplasia [5]. These cases highlight that, although exceedingly rare, even probiotics with proven benefits [7, 8] can cause bacteremia in severely immunocompromised patients or in those with structural disruption of gastrointestinal barriers due to malignancy, surgery, or other conditions.

Conclusion

We report a case of mediastinitis due to esophageal invasion by pleomorphic carcinoma, complicated by bacteremia caused by L. paracasei, genetically identical to the strain contained in commercially available probiotic-fermented milk. This case indicates that probiotics can cause bacteremia in immunocompromised patients with structural disruption of the gastrointestinal barrier.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Kushima and Yanagihara contributed to the conception and design of the study, data acquisition, and drafting of the manuscript. Koga and Hamasaki performed the pathological diagnosis, contributed to data interpretation, and critically revised the manuscript. Asahara and Yamamoto contributed to data acquisition and analysis, and critically reviewed the manuscript. Shundo, Hamada, and Fujita contributed to data analysis and interpretation, and critically reviewed the manuscript. All authors read and approved the final manuscript and agree to be accountable for all aspects of the work.

Funding

No funding was provided or sought.

Data availability

The datasets generated and analyzed in this report, including uncropped gel and ELISA images, are available from the corresponding author upon reasonable request (Contact: toyoshi.yana@fukuoka-u.ac.jp).

Declarations

Ethics approval and consent to participate

It was taken and written informed consent was obtained from the patient’s next of kin for the publication of clinical details and imaging findings included in this manuscript.

Consent for publication

Written informed consent was obtained from the patient’s next of kin for the publication of clinical details and imaging findings included in this manuscript.

Competing interests

Takashi Asahara and Shuta Yamamoto are employees of the Yakult Central Institute for Microbiological Research. However, all medical decisions, including diagnosis, treatment, and clinical interpretation, were made independently by the attending physicians (Kushima, Yanagihara, and Fujita) without any influence from Yakult. Pathological diagnoses were made independently by Kaori Koga and Makoto Hamasaki. Kushima, Yanagihara, Koga, Shundo, Hamada, Hamasaki, and Fujita have not received any research funding, financial support, or other benefits from Yakult or any other commercial entities related to this work.