We detected 5 blood culture–positive C. butyricum bacteremia cases (0.08%) (Table 1) from a total of 6,576 persons who had blood cultures positive for any bacteria (7,484 total clinical strains, including bacteria other than C. butyricum). Bacteremia developed in all 5 patients during hospitalization; 3 patients were women and 2 were men. Four patients were immunocompromised: 2 had undergone transplantation, 1 was undergoing chemotherapy for esophageal and gastric cancers, and 1 was receiving multiple immunosuppressive treatments for dermatomyositis. Two of the 5 patients also had end-stage kidney disease and were on dialysis. The Charlson Comorbidity Index scores ranged from 1 to 6 points for each patient. Three patients underwent catheterization with either a central venous catheter or a peripherally inserted central catheter. Four patients were taking prescribed MIYA-BM at the time of bacteremia diagnosis, and 1 patient (no. 2) had been prescribed a different probiotic containing C. butyricum 1 month before the diagnosis of bacteremia. All 4 patients taking MIYA-BM were prescribed it >1 week prior to hospitalization, and MIYA-BM was discontinued following the diagnosis of bacteremia in all these patients. Despite a detailed review of the medical records, we were unable to identify the specific reason for prescribing probiotics in 2 patients. All 5 patients had fever and abdominal symptoms, such as diarrhea and pain. One patient (no. 3) with nonocclusive mesenteric ischemia died within 90 days.

A consistent pattern of antibiotic susceptibility was observed in all clinical strains (Table 2). Moreover, those results were consistent with those of previous reports on the antibiotic susceptibility of C. butyricum. C. butyricum has been reported to be susceptible to penicillin, ampicillin, cefmetazole, imipenem, meropenem, clavulanic acid/amoxicillin, tazobactam/piperacillin, clindamycin, moxifloxacin, and metronidazole but resistant to cefotaxime and ceftriaxone (11,29,30).

Whole-genome analysis of all 5 patient clinical strains revealed that they either exhibited complete homology or had a maximum divergence of only 19 mutations relative to CBM 588, which was extracted from the MIYA-BM tablets. This result indicates that all clinical strains had the same clone as the CBM 588 extracted from MIYA-BM (Table 3) (31–34). We performed genetic annotation of the detected mutations (Appendix Table). We performed phylogenetic analysis of C. butyricum by using the Type (Strain) Genome Server (35). All clinical isolates and probiotics strain were clustered on the same branches (Figure). Average nucleotide identity scores of clinical isolates against those of the probiotics strain were higher than against those of reference strains. This analysis further corroborated the genetic homology between all the clinical strains and the CBM 588 strain.

Our single-center, retrospective study determined that the prevalence of C. butyricum bacteremia was 0.08% among all cases with bacteria-positive whole-blood cultures and that all clinical strains were derived from the CBM 588 strain. Bacteremia developed in all patients during hospitalization. Out of 5 cases, 4 had received immunosuppressive treatment and 2 had intra-abdominal issues (1 case of esophageal and gastric cancer and 1 case of post–pancreas and kidney transplantation).

Ishikawa et al. reported a case series of 11 cases of C. butyricum bacteremia, including 3 self-experienced cases and 8 cases from a literature review (23). The study revealed that at least 8 cases developed bacteremia during their hospitalization for conditions unrelated to the bacteremia itself. Furthermore, most patients had intra-abdominal issues at the time of developing bacteremia. In 3 cases, C. butyricum bacteremia developed after intra-abdominal surgery. Among the 8 cases without intra-abdominal surgery, 6 cases occurred after various intra-abdominal conditions (2 cases of Crohn’s disease, 2 cases of gastrointestinal ulcers, 1 case of biliary tract infection, and 1 case of nonobstructive mesenteric ischemia). Our study results align with previous findings, emphasizing the need for vigilant monitoring of bacteremia development associated with probiotic use in patients with intra-abdominal issues or those undergoing immunosuppressive therapy during their hospitalization.

Our study revealed a high degree of genetic similarity between the strains of C. butyricum extracted from MIYA-BM tablets and clinical strains identified through genetic analysis, strongly supporting the definition of probiotic-related bacteremia in all our cases. Reports on probiotic-related bacteremia are scarce. Although systematic reviews of cases of bacteremia after probiotic use have been reported (36), to the best of our knowledge, no studies have evaluated the genetic similarities among these reports. Our study offers evidence supporting a direct causal relationship between probiotic prescription and bacteremia. Nonetheless, the patients we identified as nos. 1 and 2 present lingering challenges. We observed 19 differences in terms of SNPs between the strains found in the blood culture of patient 1 and the CBM 588 strain, which was relatively higher than that of the other patients. However, it is common to evaluate strain dissimilarity using fewer than 100 SNPs. Notably, rapidly growing bacteria, such as Helicobacter pylori, can accumulate ≈30 SNPs within 6 months of acute infection (37). In fact, some studies have established a genetic similarity cutoff of 80 for carbapenem-resistant Klebsiella pneumoniae (38), suggesting that the genetic dissimilarity observed in this case could be reasonably acceptable. We also considered the possibility that long-term oral administration of probiotics in the past could have led to genetic mutations in the CBM 588 strain within the bodies of patients we examined. Patient 2, who had been prescribed a different probiotic containing C. butyricum 1 month before the diagnosis of bacteremia, developed bacteremia caused by the CBM 588 strain. We considered 2 possibilities for this observation: the patient had previously taken MIYA-BM and it had colonized in the patient’s gastrointestinal tract, leading to an infection; or the C. butyricum present in the probiotics the patient was taking had genetic similarities to the CBM 588 strain.

Our findings also bring to light the potential adverse effects related to the inappropriate prescribing of probiotics. In all cases where MIYA-BM was prescribed, probiotics were administered for >1 week. However, after a comprehensive review of the detailed medical records, we were unable to identify the appropriate reasons for prescribing probiotics in half of the cases. Probiotics exhibit various therapeutic and preventive effects in different medical conditions, such as averting antibiotic-associated diarrhea (39) and CDI (40), preventing hepatic encephalopathy in patients with liver cirrhosis (41), and managing symptoms in patients with ulcerative colitis (42). Although probiotics may demonstrate effectiveness in such specialized clinical scenarios, those scenarios were not observed in the cases we studied, in which probiotics appeared to have been prescribed indiscriminately over an extended period.

One limitation of our study was that it was a single-center, retrospective investigation. Multicenter studies are needed to elucidate the prevalence of C. butyricum bacteremia and the genetic origin of the strains. Another limitation was that patient 1 showed improvement with ceftriaxone use, although C. butyricum is resistant to it. There is a possibility of contamination resulting from such factors as polymicrobial bacteremia and the absence of central venous catheterization. However, it cannot be ruled out that patients with concurrent sacral pressure ulcers are at risk of developing polymicrobial bacteremia, including C. butyricum bacteremia. Also, the duration of probiotic use for each case patient was based on information documented in their medical records, and the precise prescription durations were not always clear. However, the actual prescription periods must exceed the durations documented in the medical records, because the recorded periods represent at least the minimum assessable timeframe. Moreover, although specific reasons for probiotic prescription were not evident in the medical records, unique justifications may have existed. Nevertheless, it is crucial to note that none of the patients had a history of prior antibiotic use, CDI, irritable bowel syndrome, or liver cirrhosis. Hence, the need for prolonged administration exceeding 2 weeks for therapeutic purposes seems unlikely.

In conclusion, our study demonstrates that all clinical strains of C. butyricum identified in the positive blood cultures of the 5 cases we analyzed were derived from the strain found in probiotics. Although this type of bacteremia is rare, careful monitoring is essential when bacteremia is caused by probiotics. Clinicians must avoid long-term, inappropriate prescription of probiotics for hospitalized patients with multiple comorbidities, including immunosuppressive treatment and intraabdominal problems, to prevent bacteremia caused by probiotics.