Bacillus subtilis var. natto bacteremia: A case series and review

Abstract

Natto, a traditional Japanese fermented soybean food, contains Bacillus subtilis var. natto, which is generally regarded as safe but can rarely cause bloodstream infections in susceptible individuals. We report three cases of B. subtilis var. natto bacteremia confirmed by biotin operon analysis: a 63-year-old man receiving maintenance hemodialysis, a 35-year-old woman at 36 weeks of gestation, and a 62-year-old man whose bacteremia led to the diagnosis of colorectal cancer. All patients reported habitual natto consumption and had underlying conditions associated with impaired intestinal barrier function or altered immune status. Blood cultures yielded Bacillus subtilis, which was initially regarded as a contaminant. Targeted antimicrobial therapy resulted in resolution of bacteremia in all patients; the pregnant patient underwent emergency cesarean delivery, and the third patient subsequently underwent surgical resection for colorectal cancer. When considered together with previously published cases, our findings suggest that B. subtilis var. natto bacteremia primarily occurs in patients with impaired host defenses or intestinal barrier dysfunction. Biotin operon analysis identified characteristic genetic features of B. subtilis var. natto, consistent with a likely dietary origin associated with natto consumption. Because Bacillus species isolated from blood cultures are often dismissed as contaminants, B. subtilis var. natto bacteremia is likely underrecognized. Clinicians should consider this diagnosis in patients with relevant risk factors and obtain a detailed dietary history when Bacillus species are isolated from blood cultures.

Introduction

Bacillus subtilis var. natto is a subspecies of Bacillus subtilis and a Gram-positive, spore-forming, facultative anaerobic bacterium. Although it is genetically very similar to other B. subtilis strains, it is distinguished by its ability to produce large amounts of poly-γ-glutamic acid (γ-PGA) during fermentation. This property gives natto its characteristic viscosity and makes the strain particularly suitable for soybean fermentation; therefore, it serves as the starter culture for natto production [1]. Natto is a traditional fermented soybean food produced using B. subtilis var. natto as the starter culture, and the final product contains high concentrations of viable B. subtilis var. natto spores. Thus, consumption of natto is closely associated with the ingestion of viable B. subtilis var. natto spores. Fermented soy products have long been consumed in Asia for their nutritional and health benefits [2], and some epidemiological studies even suggest that natto intake may confer positive health effects, including a potential reduction in mortality [2]. Although B. subtilis is generally regarded as non-pathogenic, bacteremia caused by this organism has been reported, particularly in immunocompromised hosts [3], [4], [5], [6], [7], [8], [9].

B. subtilis var. natto strains are characterized by mutations in genes involved in the biotin biosynthetic pathway, such as bioF and bioW. Biotin operon analysis, which identifies these mutations, has been used as a molecular approach to support the identification of isolates as natto-associated strains of B. subtilis, thereby facilitating strain characterization [10]. Cases of bacteremia caused by B. subtilis var. natto have been increasingly recognized, and previous reports suggest that most clinical isolates identified as B. subtilis from blood cultures are actually B. subtilis var. natto when examined by molecular methods. [9].

In this case series, we present three episodes of bacteremia caused by B. subtilis var. natto and perform a review of previously published case reports to better define the clinical significance of this rare pathogen. Genetic analysis of the B. subtilis isolates from these three cases revealed a deletion in the bioF gene and a mutation in the bioW gene [11], confirming their homology with B. subtilis var. natto. The cases involved a hemodialysis patient, a woman at 36 weeks of pregnancy, and a patient in whom bacteremia led to the diagnosis of colorectal cancer. Although sporadic cases have been described, clinical characteristics, pathogenesis, risk factors, and management strategies associated with this organism remain poorly defined, and its mechanism of onset has not been systematically examined. By integrating our case series with past reports, we aim to clarify the pathogenesis, clinical features, and management considerations of B. subtilis var. natto bacteremia.

Case Report 1

A 63-year-old Japanese man with end-stage renal disease (ESRD) secondary to chronic glomerulonephritis and alcoholic cirrhosis had been undergoing maintenance hemodialysis for one year. During a dialysis session, he developed chills followed by a fever of 39.6°C. His medical history included hypertension and hypothyroidism. Notably, he had gained 6 kg compared to his dry weight (84.3 kg) in a single day, suggesting excessive dietary intake.

On admission, his vital signs were blood pressure (BP) 110/70 mmHg, heart rate (HR) 100 bpm, respiratory rate (RR) 16 breaths/min, and oxygen saturation (SpO₂) 98% on room air. He was fatigued but alert and hemodynamically stable, with no signs of acute distress. Physical examination was unremarkable. His height was 177.0 cm and weight 84.3 kg (BMI 26.9 kg/m²). No other significant abnormalities were identified. Given the fever and chills in the absence of localized infection, bacteremia was suspected. Two sets of blood cultures were obtained, and empirical intravenous cefotaxime 1 g per day was initiated.

On hospital day 2, blood cultures grew Escherichia coli, Streptococcus parasanguinis, Streptococcus anginosus, Streptococcus salivarius, and B. subtilis. Although B. subtilis was initially isolated from a single culture set, subsequent cultures remained positive, indicating true bacteremia rather than contamination. B. subtilis was susceptible to all antimicrobial agents tested, including ampicillin (Table 1). Antimicrobial therapy was escalated to vancomycin plus cefotaxime and later adjusted to ampicillin plus cefotaxime based on susceptibility results.

Table 1.

Antimicrobial susceptibility testing results for Bacillus subtilis var. natto isolates from this case series.

Antimicrobial Agent	MIC interpretive criteria (µg/mL)			Case 1	Case 2	Case 3
	S	I	R
Penicillin	≤ 0.12	-	≥ 0.25	≤ 0.03	≤ 0.03	≤ 0.03
Ampicillin	≤ 0.25	-	≥ 0.5	≤ 0.12	≤ 0.12	≤ 0.12
Imipenem	≤ 4	8	≥ 16	≤ 0.12	≤ 0.12	≤ 0.12
Meropenem	≤ 4	8	≥ 16	≤ 0.25	≤ 0.25	≤ 0.25
Vancomycin	≤ 4	-	-	0.25	0.25	0.25
Erythromycin	≤ 0.5	1–4	≥ 8	≤ 0.12	≤ 0.12	≤ 0.12
Minocycline	≤ 4	8	≥ 16	≤ 0.12	≤ 0.12	≤ 0.12
Levofloxacin	≤ 2	4	≥ 8	0.12	0.12	0.12
Clindamycin	≤ 0.5	1–2	≥ 4	0.5	0.5	0.5
Trimethoprim-sulfamethoxazole	≤ 8	16	≥ 32	≤ 0.5	≤ 0.5	≤ 0.5

Footnote: Interpretive criteria were based on the CLSI breakpoints for Bacillus spp., because CLSI breakpoints are not available for Bacillus subtilis var. natto [17]. Abbreviations: I, intermediate; MIC, minimum inhibitory concentration; R, resistant; S, susceptible.

On the second day of antibiotic therapy, B. subtilis was again isolated—this time from two additional blood culture sets—suggesting persistent bacteremia. Blood cultures obtained on hospital day 5 were negative. An investigation of the infectious focus and entry route was performed, but none could be identified.

The patient completed a 14-day course of intravenous antimicrobial therapy with agents to which the organism was susceptible, administered from hospital day 2 to hospital day 18. He was discharged in stable condition on hospital day 19. This molecular finding strongly suggested a dietary origin of the bacteremia. A post-discharge dietary assessment revealed that the patient had been consuming one pack of natto daily. Genetic analysis later confirmed B. subtilis var. natto, strongly supporting a dietary origin of the bacteremia. He was advised to discontinue natto intake. At 6-month follow-up, no recurrence of bacteremia was observed.

Case Report 2

A 35-year-old primigravid woman at 36 weeks and 2 days of gestation presented with abdominal tightening, chills, and high-grade fever (38.8°C), which progressed to severe rigors. She reported epigastric discomfort but denied gastrointestinal or urinary symptoms. Her medical history was notable for gestational diabetes mellitus (GDM) and threatened preterm labor.

On presentation to the emergency department, she appeared cold and uncomfortable. Vital signs were: BP 132/78 mmHg, HR 127 bpm, RR 20 breaths/min, SpO₂ 98%, and body temperature 39.2°C. Her height was 147.8 cm, weight 45.2 kg, and BMI 20.7 kg/m². Abdominal examination was unremarkable. Speculum examination revealed white vaginal discharge with scant blood. Maternal sepsis was suspected, and an emergency cesarean section was performed the following day in response to her clinical status due to suspected maternal sepsis and fetal considerations.

Two of four aerobic blood culture bottles collected at admission grew B. subtilis. Repeat blood cultures obtained on hospital day 5 were negative. The isolate was susceptible to all antimicrobial agents tested, including ampicillin (Table 1). Urine culture was deemed contaminated, and the vaginal culture grew only normal flora.

Empiric intravenous ampicillin was initiated on hospital day 5 and continued for 4 days until hospital day 8, followed by oral amoxicillin through hospital day 14. The patient responded well to treatment and recovered without complication.

Genetic analysis of B. subtilis was performed using the same method as in Case 1, and it was determined to be B. subtilis var. natto. Post-discharge dietary assessment revealed that the patient had consumed natto three times weekly prior to pregnancy and had increased her intake to one to two packs daily during pregnancy. At follow-up, she remained clinically well without recurrence of bacteremia during follow-up. Notably, she continues to consume two packs of natto daily, which her child—now thriving—also consumes regularly. Neither has experienced subsequent episodes of bacteremia.

Case Report 3

A 62-year-old man with a history of cerebral infarction and hypertension presented with a 5-day history of loss of appetite and a 4-day history of sudden-onset right lower quadrant pain (numerical rating scale 8/10). He consumed one serving of awamori (a traditional distilled alcoholic beverage) daily. Abdominal CT performed at a local clinic revealed an enlarged appendix suggestive of perforated appendicitis, and he was referred to our hospital.

On admission, vital signs were blood pressure 142/86 mmHg, heart rate 102 beats/min, respiratory rate 16/min, body temperature 36.8°C, and oxygen saturation 97% on room air. Physical examination revealed marked tenderness in the right lower quadrant, with positive McBurney’s, Lanz’s, and psoas signs. CT demonstrated multiple diverticula in the ascending colon and an enlarged appendix.

Five sets of blood cultures were obtained, and B. subtilis was isolated from six of ten bottles. Genetic analysis identified the strain as B. subtilis var. natto. The isolate was susceptible to all antimicrobial agents tested, including ampicillin (Table 1).

Empirical therapy with ampicillin/sulbactam (ABPC/SBT) was started for suspected appendicitis or diverticulitis, but the intra-abdominal abscess did not improve. Preoperative colonoscopy revealed a transverse colon tumor, which was completely resected endoscopically (carcinoma in adenoma, pT1, HM0, VM0, Ly0, V0, PN0; EMR specimen).

Due to poor response to antibiotics, right hemicolectomy with lymph node dissection was performed for suspected microperforated cecal diverticulitis. During laparotomy, localized peritonitis was observed, although it had been controlled with antibiotic therapy. Subsequent histopathological examination showed disruption of the cecal diverticular wall with inflammation, fibrosis, and abscess formation extending to the cecum and appendix. No residual tumor was found in the transverse colon. The postoperative course was uneventful, and the patient was discharged on hospital day 45 (postoperative day 18).

Discussion

B. subtilis var. natto bacteremia is rare, and its pathophysiological mechanisms remain poorly understood. Building on previously published reports and the three cases presented in this study, we explore its pathogenesis, characterize the clinical features, and discuss implications for management.

Bacteremia caused by B. subtilis var. natto, including the three cases we experienced, has been reported in a total of nine cases (Table 2). All reported patients consumed natto habitually, although details regarding quantity and duration were inconsistently recorded. Where documented, intake ranged from one to two packs per day. Notably, in four cases, blood cultures remained positive despite the initiation of antimicrobial therapy. The Charlson Comorbidity Index (CCI) was calculated for each case to objectively quantify the burden of underlying comorbidities, which may reflect host susceptibility to infection. The CCI scores ranged from 0 to 6, with a mean score of 2.4. All patients in the present series reported habitual consumption of natto. The gastrointestinal tract was therefore considered the common anatomical portal of entry for Bacillus subtilis var. natto in all cases via oral ingestion. Accordingly, bacteremia was presumed to result from translocation of ingested bacteria across the intestinal mucosa into the bloodstream. The cases were categorized according to the presumed mechanism of translocation rather than the anatomical portal of entry, as follows: translocation associated with gastrointestinal perforation (n = 3), bacterial translocation across an intact intestinal mucosa (n = 5), and unknown mechanism (n = 1). One previously reported case suggested anal fissure as a possible contributing local factor [5]; however, no gastrointestinal perforation, overt immunosuppression, or clear evidence of increased intestinal permeability was identified, and the source of infection was ultimately considered indeterminate. Antimicrobial treatment varied across cases. Empirical therapy most often involved broad-spectrum agents such as meropenem or vancomycin, which were subsequently switched to narrower-spectrum antibiotics, including cefaclor or ampicillin, based on susceptibility results. Reported treatment duration ranged from 10 to 39 days; in one instance, lifelong suppressive therapy was administered [7]. When this outlier was excluded, the mean duration of therapy was 22.4 days. All patients achieved complete recovery, and no fatalities were reported.

Table 2.

Clinical characteristics of patients with blood culture–confirmed Bacillus subtilis var. natto bacteremia.

Age/Sex	Underlying diseases	CCI	Natto consumption	Recurrent blood culture positivity†	Presumed route of infection	Antimicrobial therapy	Treatment duration (days)	Reference
56/F	None	0	Daily	No	GP (SC)	PIPC/TAZ + TEIC→ABPC/SBT	39	3
65/M	Colorectal cancer, DM	3	Daily	No	GP (SC)	MEPM	10	4
53/F	Anal fissure	0	Daily	Positive (days 3,7)	Unknown‡	VCM→ABPC/SBT→AMPC/CVA	36	5
70/F	ESRD, MM, DM	6	Daily	Positive (day 3)	BT	VCM	14	6
41/F	Congenital LF post-splenectomy	3	Daily	No	BT	MEPM→CCL	Lifelong (CCL)	7
67/F	Erosive esophagitis SSc, PM	1	Daily	No	BT	VCM + MEPM→VCM	19	8
63/M	ESRD Cirrhosis	5	1 pack /day	Positive (days 2,3)	BT	CTX→CTX + VCM→CTX + ABPC	19	Present case 1
35/F	Gestational DM Threatened PL	0	2 packs /day	No	BT	ABPC→AMPC	14	Present case 2
60/M	Cerebral infarction, colorectal cancer	4	3–4 days/week	Positive (days 1,2)	GP	ABPC/SBT	28	Present case 3

Data are presented for individual patients with blood culture–confirmed Bacillus subtilis var. natto bacteremia. Variables include age, sex, underlying conditions, Charlson Comorbidity Index (CCI), natto consumption, recurrent blood culture positivity, presumed route of infection, antimicrobial therapy, treatment duration, and reference. ^†Day 1 was defined as the first day on which blood cultures were positive. ^‡Not classified as gastrointestinal perforation or bacterial translocation.

Abbreviations: ABPC, ampicillin; ABPC/SBT, ampicillin/sulbactam; AMPC, amoxicillin; AMPC/CVA, amoxicillin/clavulanic acid; BT, bacterial translocation; CCL, cefaclor; CTX, cefotaxime; DM, diabetes mellitus; ESRD, end-stage renal disease; GI, gastrointestinal; GP, gastrointestinal perforation; LF, liver fibrosis; MEPM, meropenem; MM, multiple myeloma; PC, presented case report; PL, preterm labor; PM, polymyositis; SC, sigmoid colon; SSc, systemic sclerosis; PIPC/TAZ, piperacillin/tazobactam; TEIC, teicoplanin; VCM, vancomycin.

Bacterial translocation in B. subtilis var. natto bacteremia is likely mediated through three major mechanisms: (a) intestinal bacterial overgrowth, (b) increased intestinal permeability, and (c) host immunosuppression [12].

• (a)Intestinal bacterial overgrowth: All cases (Cases 1–3 in the present report and previously reported cases [3], [4], [5], [6], [7], [8]) had a history of habitual natto consumption, suggesting that repeated dietary exposure may contribute to an increased intestinal bacterial burden. Antimicrobial-associated disruption of the intestinal microbiota was documented in only one previously reported case [6], in which piperacillin/tazobactam had been administered prior to the onset of bacteremia.
• (b)Increased intestinal permeability: Three cases were considered to involve increased intestinal permeability (Case 3 in the present report and two previously reported cases [3], [4]). In the previously reported cases associated with gastrointestinal perforation, all patients developed secondary peritonitis, including one following a barium study and another associated with colorectal cancer. These conditions likely facilitated bacterial translocation by disrupting intestinal barrier integrity caused by inflammation and sepsis [13], [14], [15].
• (c)Host immunosuppression: Immunocompromised conditions were present in four cases (Cases 1–2 in the present report and previously reported cases [6], [7], [8], including corticosteroid therapy, splenectomy, pregnancy, and immunomodulatory treatment for COVID−19 (methylprednisolone and tocilizumab). Under normal conditions, bacteria that traverse the intestinal mucosa are rapidly controlled and eliminated by innate immune mechanisms, including the intestinal immune system and macrophages. These defense mechanisms are maintained by the gut microbiota and play an important role in preserving intestinal barrier homeostasis [16]. In immunocompromised patients, impairment of these defense mechanisms may permit bacterial translocation.

These mechanisms likely coexist in many patients. Importantly, they are not mutually exclusive. In many cases, intestinal bacterial overgrowth, increased intestinal permeability, and host immunosuppression likely interacted synergistically to facilitate bacterial translocation and the subsequent development of bacteremia.

A retrospective study from a Japanese hospital reported that approximately 1% of positive blood cultures contained B. subtilis, almost all identified as B. subtilis var. natto, with a 30-day mortality of 16% and frequent need for urgent intervention [9]. Our case series, together with previous reports, suggests that these emergent procedures were largely driven by concomitant gastrointestinal perforation rather than by the bacteremia itself. In our report and the cases summarized in Table 2, all patients recovered without fatal outcomes, in contrast to the Japanese cohort [9]. Persistent bacteremia observed in that cohort may have reflected inadequate source control or severe underlying illness rather than the pathogenicity of B. subtilis var. natto itself. Overall, when appropriate source control and antimicrobial therapy are achieved, clinical outcomes appear favorable.

Antimicrobial susceptibility testing was interpreted according to the Clinical and Laboratory Standards Institute (CLSI) M45 guidelines for Bacillus spp. [17]. In the present series, all isolates demonstrated low MIC values and were susceptible to ampicillin and the other antimicrobial agents tested, suggesting that acquired antimicrobial resistance is uncommon in B. subtilis var. natto. Successful management appears to depend primarily on timely source control and appropriate management of the underlying predisposing condition.

Biotin operon analysis is currently available only in specialized laboratories, and it is therefore not always feasible to determine whether an isolated B. subtilis strain is of natto origin. Although definitive subspecies identification is primarily useful for confirming the association between natto consumption and bacteremia or for epidemiological investigations, it is not required for routine clinical management. Nevertheless, when Bacillus spp. are isolated from blood cultures, they should not be dismissed as contaminants without careful clinical evaluation, as they may indicate underlying gastrointestinal pathology or impaired intestinal barrier function.

This study has several limitations. Being a retrospective study, data were limited, and critical clinical information was sometimes missing. The total number of reported cases to date is small, introducing the possibility of information and selection bias, so the findings should be interpreted with caution. Identifying populations at higher risk of infection from natto consumption, or the conditions that increase this risk, remains difficult, requiring further accumulation and analysis of cases. The underlying pathophysiology is not well understood; bacterial translocation due to intestinal bacterial overgrowth is considered a possible mechanism, but details on natto intake, storage conditions, and immunosuppressive status are unclear, and the relative contribution of cellular and humoral immunity remains undefined. These uncertainties highlight the need for further clinical and mechanistic investigation.

Conclusion

Although natto is widely consumed and generally considered beneficial, B. subtilis var. natto bacteremia can occur in patients with compromised immune function or impaired intestinal barrier integrity. Habitual natto consumption may increase the risk of bacterial translocation in susceptible individuals. Clinicians should avoid dismissing Bacillus spp. isolated from blood cultures as contaminants without appropriate clinical and dietary evaluation.

Author contributions

All authors were involved in the clinical management of the case. The first draft of the manuscript was prepared by Hikaru Ishizawa and Masashi Narita. All authors reviewed and revised the manuscript and approved the final version.

CRediT authorship contribution statement

Takashi Oshima: Resources, Investigation. Satoshi Kutsuna: Supervision. Hikaru Ishizawa: Writing – original draft, Data curation. Masashi Narita: Writing – review & editing, Writing – original draft, Project administration, Conceptualization. Takahiro Murakami: Resources, Investigation. Yu Kuriyama: Resources, Investigation. Go Yamamoto: Writing – review & editing, Methodology, Investigation. Akiko Okura: Investigation. Hitoshi Miyasato: Supervision. Nana Yara: Resources, Investigation.

Ethics declaration

Written informed consent to take part in the study and to publish the article has been obtained from all participants or their legal representatives. The privacy rights of participants have been observed.

This study was performed in compliance with relevant laws, regulatory frameworks and guidelines where the research took place. This study was approved by the Okinawa Prefectural Miyako Hospital, Okinawa Prefectural Nanbu Medical Center and Children's Medical Center. (Approval No. 25M004, R7-142).

This research follows the CARE guidelines and the CARE checklist.

Funding

No external funding was received for this study.

Declaration of Competing Interest

The authors declare no conflicts of interest.