Recurrent Var. Bacteremia and Review of the Literature on : The First Case Report

Background

Some fermented foods, such as natto and yogurt, are considered healthy, but some cases of Lactobacillus acidophilus bacteremia in patients with human immunodeficiency virus and in patients with Crohn’s disease who eat yogurt have been reported [1,2]. Bacillus subtilis (B. subtilis) var. natto is a Gram-positive rod-forming spore that belongs to the Bacillaceae family. The genome of B. subtilis var. natto was decoded in 2010 [3]. In general, the pathogenicity of B. subtilis is low, but it can be severe in an immunocompromised host [4], and when B. subtilis is detected in clinical specimens, it is often considered contamination [5]. However, several case reports of true infection, such as ocular infection, trauma wound infection, meningitis, and bacteremia, were reported [6]. We aimed to report the first case of recurrent B. subtilis var. natto bacteremia and portal vein suppurative thrombophlebitis even after the patient stopped eating natto. We also aimed to perform a literature review for B. subtilis intra-abdominal infection.

Case Report

A 41-year-old Japanese woman presenting with fever was admitted to St Luke’s International Hospital. One month before admission, she developed portal vein pyogenic thrombophlebitis. She was treated with amoxicillin/clavulanate, levofloxacin, and rivaroxaban. However, 1 week before admission, her antimicrobials were discontinued due to antibiotic-related diarrhea. She was admitted to St Luke’s International Hospital for recurrent portal vein pyogenic thrombophlebitis. Our patient had congenital liver fibrosis and splenomegaly due to portal hypertension status post-splenectomy and had a history of acute suppurative cholangitis caused by B. subtilis var. natto 3 years earlier. In the first episode of our patient, we incorrectly identified the first isolate of B. subtilis as contamination because the MALDI-TOF Biotyper could identify B. subtilis but not as B. subtilis var. natto because B. subtilis var. natto was not included in the reference database of MALDI-TOF Biotyper at the time of first isolation. She was on chronic suppression with oral cefaclor because of subsequent recurrent cholangitis according to susceptibility (Table 1). She had been frequently eating natto but had stopped since then. She remained afebrile while on prophylaxis. She discontinued cefaclor just before she developed portal vein pyogenic thrombophlebitis. She had no family history of genetic diseases. On admission, her consciousness was clear, and her vital signs were: temperature, 36.8°C; blood pressure, 120/68 mmHg; pulse rate, 60 beats/min; respiratory rate, 16 breaths/min; oxygen saturation, 98% (room air). Physical examination revealed hepatomegaly and tenderness in the right hypochondrium, but no other abnormal findings were observed. Laboratory data showed a white blood cell count of 11 000/μL (82% neutrophils), hemoglobin of 8.1 g/dL, platelet count of 41.1×103/μL, total bilirubin of 0.8 mg/dL, aspartate aminotransferase of 117 U/L, alanine aminotransferase of 89 U/L, alkaline phosphatase of 146 U/L, lactate dehydrogenase of 273 U/L, γ-glutamyl transpeptidase of 44 U/L, and C-reactive protein of 2.6 mg/dL. Contrast-enhanced computed tomography revealed a thrombus from the portal vein to the superior mesenteric vein, dilation of the common bile duct, mesenteric panniculitis, and numerous collateral blood vessels (Figure 1). The 2 sets of blood cultures on admission were positive, and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF Biotyper; Bruker Daltonics, Billerica, MA, USA) identified B. subtilis as the pathogen (Figure 2). We conducted a genetic analysis of B. subtilis and confirmed the presence of B. subtilis var. natto, which had been previously detected 3 years before. The final diagnosis was B. subtilis var. natto bacteremia with portal vein suppurative thrombophlebitis. According to the susceptibility referenced by Staphylococcus [23] (Table 1), treatment was commenced using meropenem 1 g every 8 h due to the patient’s history of piperacillin/tazobactam allergy.

The fever resolved, and the clinical findings improved. Thus, the patient’s antibiotic was switched back to cefaclor 500 mg 3 times daily, and she was discharged on day 11. Nine months after discharge, her clinical status was stable, and the portal vein thrombus disappeared while she was on prophylactic cefaclor (Figure 1).

The specific strain of B. subtilis detected in the blood culture was identified with the following methods. As B. subtilis var. natto has mutations in the BioF and BioW genes in enzymes that produce biotin [7], we performed a genetic analysis of B. subtilis obtained from blood cultures 3 years prior and those from the second episode. Our isolates had mutations in the BioF and BioW genes and were homologous to the sequence of B. subtilis var. natto, but were not homologous to B. subtilis subsp. subtilis. Therefore, we identified all the B. subtilis isolates in the blood cultures as B. subtilis var. natto (Figure 3).

Discussion

We encountered a case of recurrent B. subtilis var. natto bacteremia and portal vein pyogenic thrombophlebitis even after discontinuation of eating natto. We genetically confirmed that B. subtilis in this case was identical to the B. subtilis var. natto detected in the first episode. The spores of Bacillus species are found in soil, dust, water, and air. In an environment with a wide prevalence of Bacillus species, Bacillus spores can enter the gastrointestinal tract of animals, including humans, by ingestion. The Bacillus spores can then transit unimpeded through the stomach [8]. B. subtilis var. natto is a probiotic isolated from natto. In the process of natto fermentation, B. subtilis (natto) or B. subtilis var. natto (BEST195) in Bacillus species are used because B. subtilis var. natto strains produce viscous substances. Natto, a typical Japanese food, is made of soybeans fermented with B. subtilis var. natto (BEST195). The string-like texture of natto is produced by BEST195, but not by other strains, such as B. subtilis 168.

We performed a search using the keywords “Bacillus subtilis” and “bacterial infection” (details are in Appendix A) in the electronic databases PubMed, Embase, and Ichushi (Figure 4). We finally identified 17 papers describing 30 cases, including our case (Table 2). A review of the literature on the reported cases of B. subtilis with intra-abdominal infection revealed that these articles were published from 1988 to 2022 (Figure 5). We found 4 cases of B. subtilis var. natto infection, 3 in 2022[6,7,9] and our case in 2020. Regarding the countries of origin, 13 of 18 published articles, including ours, were from Japan; the median age of the patients was 72 years (range, 15–96 years), 14 patients (47%) were female, and 3 patients (10%) died. There was no recurrent case of B. subtilis var. natto except ours.

In this review, B. subtilis var. natto was identified by analyzing the draft whole-genome of each B. subtilis strain using a next-generation sequencer [6,7] and detecting nattokinase specifically produced by B. subtilis var. natto [10] or identifying clinical specimens with Miyagino-based natto (originally isolated BEST195) [3] using single nucleotide variant (SNV) calling.

Our way of identifying the B. subtilis var. natto in our case was the same as that in the case reported by Tanaka et al [7] with the support of the same laboratory.

There have been reports of bacteremia related to B. subtilis in Japan [11], but genetic analysis of B. subtilis var. natto has not been previously performed. In our literature review, we could not verify whether it originated from B. subtilis subsp. subtilis or B. subtilis var. natto. One reason for the small number of reports is that the genome of B. subtilis var. natto was decoded in 2010, and currently only a few medical institutions can examine it [3]. In the first episode of our case, we incorrectly identified the first isolate of B. subtilis as contamination because the MALDI-TOF Biotyper could identify B. subtilis but not as B. subtilis var. natto because B. subtilis var. natto was not included in the reference database of MALDI-TOF Biotyper at the time of first isolation. In this review, we considered B. subtilis var. natto as a true infection in Japan; further cases of B. subtilis var. natto infection will be accumulated in the future.

Congenital liver fibrosis is characterized by bile duct dysplasia due to defective ductal plate formation during the embryonic period of bile duct formation and fibrosis in the region of the hepatic portal vein. The main symptoms of congenital liver fibrosis are splenomegaly due to portal hypertension, rupture of esophageal varices, and recurrent cholangitis [12].

Bacterial translocation is a phenomenon in which bacteria penetrate the intestinal epithelium and invade other organs by hematologic or lymphatic spread. Bacteria that invade the blood vessels of the intestinal tract reach the liver through the portal vein, whereas those that invade lymphatically reach the mesenteric lymph nodes [13,14]. The risks of bacterial trans-location include intestinal disease, neutropenia, immunosuppression, low nutrition, prolonged transvenous nutritional management, high levels of stress, and portal hypertension [15,16]. Therefore, the mechanism of bacteremia caused by B. subtilis var. natto in our case may be retrograde cholangitis due to liver fibrosis or bacterial translocation due to portal hypertension.

Furthermore, portal vein pyogenic thrombophlebitis is a rare but serious complication of intra-abdominal infection and has been previously reported as a complication of appendicitis [17], but in recent years, reports of its secondary occurrence in diverticulitis, inflammatory bowel disease, and cholangitis have increased [18,19]. Cirrhosis and portal hypertension are also risks for portal vein thrombosis [20]. Therefore, our case had several risk factors for portal vein thrombosis.

In our case, the patient had completely stopped eating natto and was on long-term suppression with cefaclor, but she relapsed after discontinuing cefaclor. We could not determine the cause of the relapse, but we considered that B. subtilis var. natto can colonize the intestinal tract for a long time. Hence, we followed the patient up closely even after the treatment for B. subtilis var. natto infection was completed. Regarding antimicrobial susceptibility in Bacillus spp., vancomycin, aminoglycosides, and clindamycin are often used because they produce β-lactamase [21], but whether B. subtilis var. natto is susceptible is unknown other than in our case. In our case, we assessed the antimicrobial susceptibility, and the susceptibility of B. subtilis var. natto was pan-sensitive. Further studies on the susceptibility of B. subtilis var. natto are needed.

There is little evidence for the effectiveness of anticoagulation for portal vein pyogenic thrombosis. Some observational studies have shown that the use of anticoagulation led to rapid dissolution of the portal vein thrombus, with no risk of bleeding [22], but no randomized trials or other studies have been performed. In the present case, we used rivaroxaban because of the low likelihood of spontaneous dissolution with antimicrobials alone. The patient did not experience bleeding, and complete resolution of the thrombus in the portal vein was confirmed on follow-up.

Conclusions

We reported a case of recurrent B. subtilis var. natto infection even after discontinuing natto consumption; therefore, patients with B. subtilis var. natto infection should be closely followed up even after completing treatment.

Due to advancements in PCR identification techniques, case reports of infections caused by B. subtilis var. natto are increasing.