Prevalence of intestinal spirochaetosis in serrated polyposis syndrome compared with other colorectal polyp conditions: a case–control study in Japan

Abstract

Objective

Serrated polyposis syndrome (SPS) is the most common form of colorectal polyposis syndrome. Most SPS cases are sporadic, and the environmental factors underlying its pathogenesis remain poorly defined. Intestinal spirochaetosis (IS), an infection of the colonic epithelium by Brachyspira species, is frequently asymptomatic and its clinical relevance is uncertain. Here, we examined the potential association between IS and SPS.

Methods

In this retrospective case–control study at two centres in Japan, IS prevalence was assessed in 64 patients with SPS and 382 non-SPS controls who underwent colonoscopic polypectomy. Controls were classified as having only conventional adenomas (Ctrl-CA) or serrated polyps not meeting the diagnostic criteria of SPS (Ctrl-SP).

Results

Histologically confirmed IS was detected in 12/64 (18.8%) SPS cases and 6/382 (1.6%) controls (OR 14.5, 95% CI 5.2 to 40.2; p<0.001). IS prevalence increased stepwise across Ctrl-CA, Ctrl-SP and SPS groups (1/274 (0.4%), 5/108 (4.6%) and 12/64 (18.8%), respectively; OR 6.5, 95% CI 3.1 to 13.8; p for trend <0.001). Among the 12 IS-positive SPS cases, all had a history of cigarette smoking and altered bowel habits were common. Furthermore, IS detection rates did not differ based on anatomical locations of polyps, but considerable intraindividual variability was observed across serial colonoscopies, suggesting possible episodes of clearance and reinfection. Real-time PCR and sequencing analyses demonstrated that all IS-positive cases were attributable to Brachyspira aalborgi, a species commonly found in developed countries.

Conclusions

This study reveals a novel association between IS and SPS with distinct clinical characteristics in Japan. Considering the retrospective design and small sample size from a single geographic region, further investigation is required to clarify the significance and generalisability of this association.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Serrated polyposis syndrome (SPS) is thought to arise from a complex interplay of genetic and environmental factors. Intestinal spirochaetosis (IS) represents colonisation of the large bowel mucosa by Brachyspira species and is often detected incidentally in asymptomatic individuals.

WHAT THIS STUDY ADDS

• We observed a strong association between SPS and IS in Japanese cohorts, with a stepwise relationship between IS and serrated polyp burden. SPS patients with IS display a recognisable clinical profile, including cigarette smoking and altered bowel habits, and show fluctuating IS status over time on serial colonoscopies.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings raise the possibility that IS may act as a candidate microbial cofactor within the serrated pathway of colorectal tumourigenesis. Further studies, including clinical longitudinal IS assessment during SPS surveillance in different geographic regions and basic molecular research, are needed to evaluate the significance of this association.

Introduction

Intestinal spirochaetosis (IS) is an infection of the colonic epithelium by Brachyspira species and is microscopically characterised by a ‘false brush border’ resulting from spirochete colonisation on the luminal surface of colonocytes.¹ Although IS has a global distribution, its prevalence varies widely from 0% to 64%, depending on the region and/or population examined. Two Brachyspira species, Brachyspira aalborgi and Brachyspira pilosicoli, are known to infect humans. Most affected individuals, particularly those infected with B. aalborgi in developed countries, remain asymptomatic; its clinical relevance is largely unknown and B. aalborgi tends to be regarded as a commensal organism. However, Swedish cohort studies recently revealed an association between IS and irritable bowel syndrome (IBS), with evidence of mucosal inflammation.^{2 3}

Approximately 30% of colorectal cancers (CRC) arise from serrated polyps (SPs) through the ‘serrated pathway’ characterised by distinct genetic, transcriptional and cellular profiles compared with conventional adenomas (CAs).^{4 5} Serrated polyposis syndrome (SPS) is defined by the number, size and location of SPs, including hyperplastic polyps, sessile serrated lesions, traditional serrated adenomas and unclassified serrated adenomas.⁶ SPS is the most common form of colorectal polyposis syndrome, with a prevalence of 0.4%–0.8%. Patients with SPS have an increased risk of developing CRC; therefore, intensive endoscopic surveillance is recommended.^{7 8} While pathogenic germline variants in the E3 ubiquitin ligase RNF43 account for 2% of SPS cases,^{6 9} most SPS cases are considered sporadic.¹⁰ Although some associated factors, such as cigarette smoking and high body mass index, have been reported,^{7 11 12} their significance in the pathogenesis of SPS remains unknown. In recent years, the role of microbiota in colon tumourigenesis, including SPs, has attracted increasing attention.^{13 14} In this study, we aimed to investigate a potential association between IS and SPS in Japan.

Methods

Study design and participants

For this two-centre case–control study, the SPS group comprised 64 patients diagnosed with SPS between 2016 and 2024 at the Kyoto Prefectural University of Medicine Hospital and the Kyoto Prefectural University of Medicine North Medical Center. To minimise investigator-driven selection of controls, the control group was selected by consecutive sampling and comprised all 382 non-SPS patients who underwent colonoscopic polypectomy between August 2024 and May 2025. In this retrospective study, we used colonoscopic polypectomy patients as controls because IS assessment required colonoscopy-derived formalin-fixed paraffin-embedded (FFPE) specimens with evaluable non-lesional marginal mucosa, allowing identical pathological criteria and specimen-based assessment in both SPS cases and controls. Patients who had a previous diagnosis or clinical suspicion of SPS, or who fulfilled SPS criteria on cumulative review of polypectomy specimens, were not included as controls.

For each patient in both cohorts, all colonoscopy-derived FFPE polypectomy specimens identified in the institutional pathology archives were retrieved and retrospectively reviewed. The reviewed specimens had been collected between 1998 and 2024 for the SPS group and between 2000 and 2025 for the control group, indicating that the underlying polypectomy material in both cohorts was obtained over overlapping time periods. Overall, 3082 polypectomy specimens from 446 patients were reviewed, including 1208 specimens from 64 SPS patients and 1874 specimens from 382 control patients, corresponding to means of 18.9 and 4.9 specimens per patient, respectively. Surveillance colonoscopies without polypectomy specimens were not included for the analyses. All archived colonoscopy-derived polypectomy specimens were available, and no specimen was excluded because of undiagnosable or inadequate material.

The control group was then subdivided into Ctrl-SP and Ctrl-CA according to the presence or absence of SPs not meeting the WHO diagnostic criteria for SPS. The diagnosis of SPS was established according to the WHO Clinical Diagnostic Criteria,⁶ requiring fulfilment of either or both of the following: criterion 1, ‘≥5 SPs proximal to the rectum, all being ≥5 mm in size, with ≥2 being ≥10 mm in size’; or criterion 2, ‘>20 SPs of any size distributed throughout the large bowel, with ≥5 being proximal to the rectum’.

Clinical data for both groups were obtained from medical records. Smoking was included a priori based on previous reports of its association with SPS. Altered bowel habits were recorded because previous studies have reported an association between IS and IBS or bowel symptoms.³ This variable was defined as physician-diagnosed IBS, constipation or diarrhoea in which symptoms attributable to organic gastrointestinal diseases were excluded, and bowel-regulating agents (eg, osmotic laxatives such as magnesium oxide, probiotic preparations such as Biofermin and herbal medicines such as Daikenchuto) were prescribed by physicians. This variable was not intended to represent Rome-defined IBS except in two cases of physician-diagnosed IBS. One SPS patient who had a history of inflammatory bowel disease with immunosuppressive therapy was excluded from this variable. No cases of inflammatory bowel disease were identified in the control cohort, and no cases of HIV-1 infection were present in either cohort.

Immunohistochemistry

Representative FFPE sections (4 µm thick) from all H&E-positive cases were subjected to immunohistochemistry using an anti-Treponema pallidum antibody (Bio-Rad, cat. no. 9147-1007; dilution 1:20) with a BenchMark Ultra automated stainer (Ventana Medical Systems, Tucson, Arizona, USA). Antigenic sites were visualised using 3,3’-Diaminobenzidine+liquid chromogen and sections were counterstained with haematoxylin. Two cases lacking a false brush border were included as negative controls to confirm the specificity of immunostaining.

DNA isolation, PCR analysis and DNA sequencing

DNA was extracted from the same FFPE sections used for immunohistochemistry using the NucleoSpin DNA FFPE XS kit (Takara Bio, Shiga, Japan) according to the manufacturer’s instructions. Real-time PCR was performed using the SYBR Green I reagent (Takara Bio, Shiga, Japan) on an ABI7000 (Applied Biosystems, Foster City, California, USA). Primers targeting the 16S rRNA region of the Brachyspira genus were adopted from a study by Westerman et al¹⁵: forward, 5′-TGGACTAATACCGCATATACTCTT −3′; reverse, 5′-ATGAGCCTGCGGCCTA −3′. This ‘species discrimination’ primer set enables amplification of three Brachyspira species (B. pilosicoli, B. aalborgi and B. ‘hominis’), with species differentiation inferred from melting-curve profiles. A deviation in melting temperature >0.5°C from the median value was considered significant. PCR products with significant calls were further confirmed by Sanger sequencing at Macrogen (Seoul, Republic of Korea). The resulting sequences were compared with reference sequences in GenBank: B. aalborgi (Z22781.1), B. pilosicoli (JF430710.1) and B. ‘hominis’ (AF228809.1).

Statistical analysis

All statistical analyses were performed using IBM SPSS Statistics, V.31.0 (IBM). Categorical variables were compared using Pearson’s χ² test or Fisher’s exact test, as appropriate. Associations with binary outcomes were evaluated using multivariable logistic regression to estimate ORs with 95% CIs, while associations with continuous outcomes were analysed using multiple linear regression. For analyses involving ordered categories (eg, Ctrl-CA, Ctrl-SP, SPS), linear trends were assessed by entering the category as an ordinal variable in logistic regression models and reporting p for trend. Colonoscopy-level and polyp-level associations were analysed using population-averaged logistic generalised estimating equation (GEE) models with patient ID as the clustering variable, an exchangeable working correlation structure and robust sandwich SEs. Firth penalised logistic regression and exploratory backward elimination were performed through the R Integration Plug-in using R V.4.0.5 and the logistf package V.1.24.1. All tests were two-sided, and a p<0.05 was considered statistically significant.

This study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline,¹⁶ and the completed checklist is provided in the online supplemental material.

Results

Histopathological detection of IS

This regional case–control study in Japan included 64 patients histopathologically diagnosed with SPS. As controls, we included 382 consecutive patients who underwent colonoscopic polypectomy (Ctrl-total), which was further subdivided into 274 patients with only CAs (Ctrl-CA) and 108 patients with SPs not meeting the WHO diagnostic criteria of SPS (Ctrl-SP).⁶

H&E-stained sections from all resected polyps, collected from multiple rounds of total colonoscopy (1998–2024 for the SPS group; 2000–2025 for the control group), were retrospectively reviewed for the presence of a ‘false brush border’. Cases were classified as IS-positive when two board-certified pathologists (TI and MH), who were blinded to clinical information including SPS/control status and clinical symptoms, independently identified a false brush border on H&E-stained sections in at least one specimen per patient. For each IS-positive patient, at least one representative section was further examined by immunohistochemistry using an anti-Treponema pallidum antibody to confirm the diagnosis. When H&E-based assessments were discordant, immunohistochemistry was performed for IS adjudication, and the immunostained sections were independently reassessed by the same two pathologists. After immunohistochemical adjudication, final interobserver concordance for IS classification was 100%.

As previously reported,¹⁷ the ‘false brush border’ was readily identified in non-lesional marginal mucosa, but was frequently absent from the luminal surface of CAs. This feature was also noted in SPs, although to a lesser extent (figure 1A). Immunostaining further demonstrated a clear distinction between lesional and non-lesional mucosa; the former showed complete absence or only scattered immunoreactivity, whereas the latter displayed a continuous immunopositive band along the luminal surface. Accordingly, IS detection largely relied on non-lesional mucosa, and specimens lacking such mucosa were considered unsuitable for IS assessment. However, all archived colonoscopy-derived polypectomy specimens reviewed in this study contained evaluable non-lesional marginal mucosa; therefore, no specimens were excluded for this reason. Among patients with IS, the frequency of IS-identified polyps did not significantly differ between CAs and SPs (OR 1.6, 95% CI 0.99 to 2.7; χ² test, p=0.051; figure 1B). These findings indicate that Brachyspira colonisation of the luminal surface diminishes in the lesional area, but the type of polyp (CA vs SP) does not significantly affect its detection.

Figure 1. Histopathological detection of IS. (A) Representative low-magnification and high-magnification histological images of IS in CA and SP specimens. In H&E-stained sections (left), insets show higher magnification views of boxed regions, highlighting the false brush border. Double-headed arrows indicate lesion areas. The false brush border is evident in non-lesional marginal mucosa (solid-line insets), but absent in lesional mucosa (dotted-line insets) in both CA and SP. In immunostained sections using anti-Treponema pallidum antibody (right), immunopositive brown bands along the luminal surface appear continuous in the non-lesional mucosa but are almost absent or much less prominent within lesion areas in both CA and SP. Scale bars: 250 µm (low-magnification panels) and 50 µm (high-magnification panels). (B) Total number of IS-positive and IS-negative specimens (SPs and CAs) from 18 patients classified as IS-positive (ie, at least one IS-positive specimen identified). No significant difference was observed in the rate of IS detection between SPs and CAs. CA, conventional adenoma; IS, intestinal spirochaetosis; SP, serrated polyp.

IS prevalence in SPS cases and controls

Clinical characteristics, including age, sex, associated CRC during the observation period, and current or past histories of cigarette smoking, are shown in table 1. IS was detected in 12 of 64 (18.8%) patients with SPS and in 6 of 382 (1.6%) patients in the control group; IS prevalence in the SPS group was more than 10-fold higher than that in the control group (OR 14.5, 95% CI 5.2 to 40.2; p<0.001). When the control group was subdivided into patients with Ctrl-SP and Ctrl-CA, IS prevalence demonstrated a significant stepwise gradient (Ctrl-CA<Ctrl SP<SPS; 1/274 (0.4%), 5/108 (4.6%), and 12/64 (18.8%), respectively; OR 6.5, 95% CI 3.1 to 13.8; p for trend <0.001).

Table 1. Summary table comparing the SPS group with all controls (Ctrl-total), stratified into Ctrl-SP and Ctrl-CA.

	SPS n=64	Ctrl-total n=382			P value
			Ctrl-SP n=108	Ctrl-CA n=274		P for trend
Age, year, median	71.0 (41–86)	73.5 (26–94)	73.0 (46–94)	74.0 (26–93)
Male sex, % (n)	51.6 (33)	63.1 (241)	70.4 (76)	60.2 (165)
Colorectal cancer, % (n)	9.4 (6)	12.8 (49)	14.8 (16)	12.0 (33)
Smoking, % (n)	67.2 (43)	48.2 (184)	56.5 (61)	44.9 (123)
Intestinal spirochetes, % (n)	18.8 (12)	1.6 (6)	4.6 (5)	0.4 (1)	<0.001	<0.001
IS(+) colonoscopy/total, % (n)	9.9 (38/385)	0.7 (6/817)	1.7 (5/298)	0.2 (1/519)	<0.001	<0.001
IS(+) polyp/total, % (n)	14.6 (176/1208)	1.4 (27/1874)	3.2 (25/772)	0.2 (2/1102)	<0.001	<0.001
Patients with IS at first available colonoscopy, % (n)	14.1 (9)	1.6 (6)	4.6 (5)	0.4 (1)	<0.001	<0.001

CA, conventional adenoma; IS, intestinal spirochaetosis; SP, serrated polyp; SPS, serrated polyposis syndrome.

To address potential detection bias due to more intensive surveillance and larger tissue sampling in SPS cases, we examined IS detection using three complementary approaches: the proportion of IS-positive colonoscopies, the proportion of IS-positive polyps, and patient-level IS positivity restricted to the first available colonoscopy. Colonoscopy-level and polyp-level analyses used logistic GEE models clustered by patient ID. In all three analyses, IS detection remained more frequent in SPS cases than in controls: 38/385 vs 6/817 colonoscopies (9.9% vs 0.7%; GEE OR 12.5, 95% CI 4.5 to 34.8; p<0.001), 176/1208 vs 27/1874 polyps (14.6% vs 1.4%; GEE OR 8.4, 95% CI 2.9 to 23.9; p<0.001) and 9/64 vs 6/382 patients at the first available colonoscopy (14.1% vs 1.6%; OR 10.3, 95% CI 3.5 to 30.0; p<0.001). Consistent stepwise increases were also observed across Ctrl-CA, Ctrl-SP and SPS groups in all three analyses: at the colonoscopy level (GEE OR per one-category increase 5.9, 95% CI 2.9 to 12.1; p for trend <0.001), at the polyp level (GEE OR per one-category increase 4.7, 95% CI 2.5 to 9.0; p for trend <0.001) and in the first-available-colonoscopy analysis (OR per one-category increase 5.3, 95% CI 2.5 to 11.3; p for trend <0.001).

Multivariable logistic regression adjusted for sex and smoking status showed that IS was associated with SPS status (adjusted OR 14.0, 95% CI 4.9 to 40.4; p<0.001; online supplemental figure S1A). In separate models evaluating IS detection and SP burden, the total number of SPs was associated with IS detection (OR 1.12 per polyp, 95% CI 1.1 to 1.2; p<0.001; online supplemental figure S1B) and IS was associated with a higher number of SPs (B 11.2, 95% CI 8.6 to 13.8; p<0.001; online supplemental figure S1C). These regression analysis results support an association between IS and SP burden, but should be considered exploratory because of the small number of IS-positive patients.

Characteristics of IS-positive SPS cases

Clinical and pathological characteristics of patients with SPS are summarised in table 2. Because only 12 SPS cases were IS-positive, clinical characteristics associated with IS among SPS cases were first assessed using univariable analyses. Compared with patients with IS-negative SPS, those with IS-positive SPS were more frequently male than IS-negative SPS cases (11/12 (91.7%) vs 22/52 (42.3%); OR 15.0, 95% CI 1.8 to 124.9; p=0.012), had a higher SP location L/R ratio (median 1.2 vs 0.29; OR 1.9, 95% CI 1.1 to 3.3; p=0.019), were more frequently current or former smokers (12/12 (100%) vs 31/52 (59.6%); Fisher’s exact test, p=0.006) and more frequently had altered bowel habits (7/12 (58.3%) vs 6/52 (11.5%); OR 10.7, 95% CI 2.6 to 44.8; p=0.001). Regarding altered bowel habits, six patients with IS-positive SPS had constipation-predominant symptoms (including one physician-diagnosed IBS with constipation) and one had diarrhoea-predominant symptoms. Meanwhile, two patients with IS-negative SPS had constipation-predominant symptoms and four had diarrhoea-predominant symptoms (including one physician-diagnosed IBS with diarrhoea). The number of CAs and the frequency of CRC did not differ significantly between the groups.

Table 2. Clinical characteristics of IS-positive and IS-negative SPS cases.

	IS(+) SPS n=12	IS(−) SPS n=52	P value
Age at diagnosis, median	65.0 (50–79)	66.5 (41–86)	0.337
Male sex, % (n)	91.7 (11)	42.3 (22)	0.012
Diagnostic criteria, % (n)
Criterion 1	66.7 (8)	78.8 (41)
Criterion 2	16.7 (2)	0 (0)
Criterion 1 and 2	16.7 (2)	21.2 (11)
SP location L/R ratio, median	1.22 (0.25–6.4)	0.29 (0.0–5.7)	0.019
No. of CAs, median	4.5 (0–24)	3.5 (0–24)	0.554
Colorectal cancer, % (n)	16.7 (2)	7.7 (4)	0.348
Smoking, % (n)	100 (12)	59.6 (31)	0.006
Altered bowel habits, % (n)	58.3 (7)	11.5 (6)	0.001
Constipation-predominant, % (n)	50.0 (6)	3.8 (2)
Diarrhoea-predominant, % (n)	8.3 (1)	7.7 (4)

CA, conventional adenoma; IS, intestinal spirochaetosis; L/R ratio, left colon (descending, sigmoid colon, and rectum)/right colon (cecum, ascending, and transverse colon) ratio; SP, serrated polyp; SPS, serrated polyposis syndrome.

Because smoking showed complete separation, its association with IS positivity was evaluated using Firth penalised logistic regression. The estimated association varied across parsimonious adjustment models and was not statistically significant after adjustment for sex. In an exploratory backward-elimination analysis beginning with smoking, male sex, altered bowel habits and SP location L/R ratio, smoking (OR 19.6, 95% CI 2.0 to 2707.0; p=0.006) and altered bowel habits (OR 10.9, 95% CI 2.5 to 58.6; p=0.001; online supplemental figure S2) were retained in the final model. However, all estimates were imprecise and were interpreted as exploratory. The majority of patients in both IS-positive and IS-negative groups fulfilled WHO diagnostic criterion 1, having at least two large SPs (≥10 mm).⁶ Histologically, the predominant SP subtype was sessile serrated lesion, followed by goblet cell-rich hyperplastic polyp and microvesicular hyperplastic polyp, in both IS-positive and IS-negative SPS groups (figure 2A).

Figure 2. Intraindividual variability in IS detection across serial colonoscopies. (A) Histological subtypes of SPs among patients with SPS. Pie charts and counts compare the distribution of subtypes between patients with IS-positive and IS-negative SPS. Subtypes are defined as follows: SSL, sessile serrated lesion; SSLD, SSL with dysplasia; MVHP, microvesicular hyperplastic polyp; GCHP, goblet cell–rich hyperplastic polyp; TSA, traditional serrated adenoma; USA, unclassified serrated adenoma. (B) IS detection according to polyp location. No statistically significant difference in the rate of IS detection was observed between the right-sided (61.4%) and the left-sided colon (54.7%). (C) Longitudinal heatmap illustrating the proportion of IS-positive polyps at each colonoscopy for patients with SPS (columns, n=12). Rows represent serial colonoscopies in chronological order (earliest at the top). The intensity of each cell corresponds to the percentage of IS-positive polyps per total polyps at each colonoscopy, ranging from white (0%) to black (100%). The numbers within each cell denote IS-positive polyps/total polyps evaluated at that colonoscopy. CA, conventional adenoma; IS, intestinal spirochaetosis; SP, serrated polyp; SPS, serrated polyposis syndrome.

Spatial and longitudinal assessment of IS in SPS

Given that the proportion of IS-identified polyps among all polyps in IS-positive SPS cases was approximately 57.5% (176/306), we investigated factors influencing IS detection. As shown above, the type of polyp (CA vs SP) did not significantly affect its detection. We then evaluated whether the anatomical location of polyps impacted IS detection (figure 2B). Detection rates were marginally higher in the right-sided colon (61.4%, 78/127) than in the left-sided colon (54.7%, 98/179), consistent with previous reports.¹⁸ However, this difference was not statistically significant (OR 1.3, 95% CI 0.83 to 2.1; χ² test, p=0.245) and was insufficient to account for the substantial number of IS-negative polyps. We further assessed IS detection rates across separate endoscopic examinations. Notably, while most polyps from various colorectal regions within a single endoscopic procedure were IS-positive, some patients exhibited exclusively IS-negative polyps when examined at a different time (figure 2C). No patient received targeted eradication therapy for IS. This pronounced temporal variability in IS detection may be due to incidental antimicrobial exposure that could not be comprehensively captured in the medical record, or may reflect the natural history of this infection, with spontaneous remission and reinfection episodes.

Exclusive identification of B. aalborgi in IS-positive cases

Species identification among IS-positive specimens was conducted using real-time PCR amplification of the 16S rRNA region specific to the Brachyspira genus, followed by melt-curve analysis. This analysis was shown to reliably discriminate B. aalborgi from B. pilosicoli.¹⁵ All patients with IS, including both SPS and control cases, exhibited a single melting temperature (Tm) peak, indicative of infection by a single Brachyspira species (figure 3). We did not observe any discrepancies between histological IS positivity and PCR results. All but two patients exhibited an identical Tm peak, which sequencing confirmed as B. aalborgi. Of the remaining two cases, one patient with SPS was infected with B. ‘hominis’, a sublineage of B. aalborgi,¹⁵ and one patient in the Ctrl-SP group harboured B. aalborgi with a single nucleotide substitution. B. pilosicoli was not identified from any patients. Thus, all detected IS cases were attributable to B. aalborgi.

Figure 3. Identification of Brachyspira species. Real-time PCR amplification of 16S rRNA followed by melting-curve analysis showing three distinct melting temperature peaks (1−3) from 18 patients with IS. Bold curves indicate representative peaks subjected to Sanger sequencing. Resulting sequences were aligned with reference sequences of Brachyspira aalborgi (Ref-A) and Brachyspira ‘hominis’ (Ref-H). IS, intestinal spirochaetosis.

Discussion

This study demonstrates that IS was substantially more prevalent in SPS compared with other colorectal polyp conditions. We acknowledge several limitations relating to the analysis, data availability, and histological assessment. First, the total number of IS-positive patients was small (n=18), meaning that multivariable regression estimates may be unstable, particularly in analyses with IS as the outcome or in analyses restricted to SPS cases; these estimates should therefore be interpreted as exploratory rather than confirmatory. Second, some potentially relevant variables, such as incidental antimicrobial exposures, alcohol consumption and body mass index, were not systematically available, which may have resulted in residual confounding. Third, because IS assessment in this study relied on a visible false brush border in non-lesional marginal mucosa and PCR analysis was not performed in histologically negative specimens, false-negative cases may have remained undetected, particularly in the presence of intermittent, low-level or spatially heterogeneous Brachyspira colonisation. Finally, as the current cases were recruited from a single geographic region, our findings may have been influenced by geographical or population-specific factors, as discussed further below.

Despite our control group comprising only polyp specimens, the observed prevalence of IS, 1.6% in the entire control group and 0.4% in the Ctrl-CA group, was comparable to previously reported rates in other regions of Japan and to national data from Sweden,^{2 17} suggesting that these figures approximate baseline prevalence levels in the general population of these countries. Our sensitivity analyses suggest that high IS prevalence in the SPS cases was not attributable to detection bias arising from increased surveillance or to differences in IS colonisation between SPs and CAs. The IS prevalence in the Ctrl-SP group was intermediate between those of the Ctrl-CA and SPS groups, and the number of SPs was strongly correlated with IS. Thus, IS detection exhibits a stepwise relationship with SP burden. Continued follow-up will be important to determine whether IS-positive cases in the Ctrl-SP group ultimately meet the diagnostic criteria for SPS in the future.

IS-positive SPS cases in our study were associated with altered bowel habits and cigarette smoking. An association between IS and IBS was recently demonstrated³ although altered bowel habits defined in this study did not represent Rome-defined IBS except for two cases. The link between IBS and colorectal neoplasia remains controversial¹⁹ and is particularly undetermined with respect to SPs. Cigarette smoking is a known risk factor for SPS.^{7 11} In our exploratory multivariable analysis, smoking was weakly associated with SPS. A recent European multicentre cohort reported that the patients with SPS who smoked more frequently exhibited diminutive SPs (those meeting current WHO criterion 2) and had a reduced risk of CRC.¹¹ Despite the limited sample size, all patients with IS-positive SPS in our cohort had a history of current or former smoking, but the majority satisfied WHO criterion 1, and there was no observed difference in associated CRC risk. Although a direct relationship between IS and smoking has not been previously established, smoking is known to alter the composition of the gut microbiota.^{20 21} In this respect, it is noteworthy that standard microbiota analyses may overlook the presence of Brachyspira species.²²

Previous studies examining the association between IS and colorectal polyps have reported inconsistent results, often owing to small sample sizes, lack of polyp subtype differentiation or single time-point analyses.¹⁸ In our study, we observed considerable within-patient fluctuations in IS detection across serial colonoscopies. This fluctuation may be due to incidental antimicrobial exposure or may reflect the natural history of this infection, with spontaneous remission and reinfection episodes. To date, the dynamics of Brachyspira species within the human gut remain largely unknown. Furthermore, negative results might have been caused by other factors such as re-localisation from the luminal surface rather than complete clearance of Brachyspira species.³ Importantly, relying on the absence of a false brush border at a single time point may underestimate Brachyspira exposure. Prospective or long-term retrospective studies are required to clarify the exact relationship between IS and SPs.

Geographical variability may influence this association, as prior reports linking IS and SPs have originated from Japan.^{18 23 24} In our study, all detected IS cases were attributable to B. aalborgi, consistent with previous reports indicating that B. aalborgi accounts for the majority of IS cases in developed countries.^{2 17} Nevertheless, recent investigations have revealed substantial genetic diversity within B. aalborgi,²² raising the possibility that—analogous to Helicobacter pylori²⁵—specific B. aalborgi strains with region-specific distribution may be associated with SPS. Further studies with broader geographic sampling and strain-level molecular characterisation will be needed to clarify whether this phenomenon is broadly applicable or specific to certain geographic regions.

In conclusion, our analysis revealed that B. aalborgi infection is prevalent among SPS patients in Japan, underscoring the importance of longitudinal IS monitoring in SPS surveillance cohorts. However, any causal relationship remains speculative, and the biological significance of this association requires further study, including the potential roles of a shared intestinal microenvironment that favours both SP development and Brachyspira colonisation, SPS-related mucosal changes that provide a niche suitable for Brachyspira species and Brachyspira-induced chronic inflammation that is involved in colonocyte metaplastic changes and SP development.^{2 3 26}

Data availability statement

Data are available on reasonable request.