Biomarkers in the pathogenesis and diagnosis of Crohn’s perianal fistulas: a systematic review

Abstract

Background and aims

Perianal fistulas in Crohn’s disease (PFCD) continue to pose a therapeutic challenge owing to low rates of healing and remission. Identifying biomarkers involved in their pathogenesis and diagnosis is essential for personalising care and improving treatment outcomes. This systematic review aims to highlight PFCD biomarkers and provide insights for clinical practice.

Methods

This systematic review, registered at PROSPERO (CRD42024575480) and conducted in accordance with PRISMA guidelines, includes English-language studies on PFCD biomarkers published through May 2025. Searches were performed in PubMed and Scopus. Study selection and risk-of-bias assessment were conducted independently by two reviewers, according to prespecified search and eligibility criteria.

Results

Of 5242 records initially identified, 3700 were screened, and 33 studies were included (24 addressing pathogenesis and 9 diagnostic biomarkers). Variants in OCTN1/2, NOD2, IRGM, TYK2, and VDR genes influence fistula risk and severity in Crohn’s disease, particularly polymorphisms in IRGM, NOD2, and genes within the JAK–STAT signalling pathway. Altered microbiota and the presence of antibiotic resistance genes support a bacterial role in pathogenesis. Key pro-inflammatory cytokines (IL-1β, TNF-α, IL-6) and cellular populations (CD4⁺ T-cells, macrophages) drive inflammation, while epithelial-mesenchymal transition and extracellular matrix remodelling contribute to fistula persistence. Recent evidence suggests that PFCD arises from complex immune-stromal cell interactions, involving cellular senescence, epigenetic alterations, and population heterogeneity. Diagnostic advances include metabolic profiling and urinary Raman spectroscopy, alongside established biomarkers such as granulomas, faecal calprotectin, and NGAL.

Conclusions

This review highlights several promising biomarkers involved in the pathogenesis and diagnosis of PFCD; however, heterogeneity across studies limits their immediate translation into clinical guidelines. Larger, well-designed studies are needed to validate these biomarkers and support personalised treatment strategies.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s10151-026-03376-8.

Introduction

Perianal fistulas (PF) are benign anorectal conditions that can significantly impair patients’ personal, social, and professional lives. Prompt and accurate diagnosis, together with appropriate treatment, are crucial for disease control and for improving patients’ quality of life [1]. However, PF recurrence rates following surgery are highly variable, ranging from 3 to 57%, and are influenced by factors such as anatomical classification, aetiology, and the therapeutic approach used [2].

While approximately 90% of PF cases are of cryptoglandular origin, other less common aetiologies include tuberculosis, trauma, hidradenitis suppurativa, human immunodeficiency virus (HIV) infection, sexually transmitted diseases, radiotherapy, and malignant diseases. Crohn’s disease (CD) represents a significant, though less frequent, cause. Recent studies indicate that PF and abscesses have a mean prevalence of 8–23 cases per 100,000 people, while perianal lesions associated with inflammatory bowel disease (IBD) affect approximately 4 cases per 100,000 individuals [3].

The pathogenesis of PF remains incompletely understood, and it is believed to involve distinct mechanisms in cryptoglandular fistulas (CGF) compared with perianal fistulas in Crohn’s disease (PFCD). The formation of these fistulas is a multifactorial process, potentially driven by anal gland and intestinal inflammation, epithelial-mesenchymal transition (EMT), and gut microbiota dysbiosis [4].

Cryptoglandular fistulas likely originate from infection of the perianal glands [5], which can progress to abscess and subsequent fistulisation. Pro-inflammatory cytokines drive tissue damage [6], while the microbiota appears to play a key role in sustaining the inflammatory process [7].

In PFCD, the secretion of pro-inflammatory cytokines such as tumour necrosis factor alpha (TNF-α) and transforming growth factor beta (TGF-β), combined with epithelial defects, can drive the EMT [8]. The microbiota may also contribute to both inflammation and EMT. Additionally, the activation of matrix metalloproteinases (MMPs) leads to inflammation, tissue damage, and, ultimately, fistula formation [9]. These complex interactions contribute to the low healing rate, high recurrence rate, and more aggressive behaviour of PFCD, which is often associated with a poorer overall prognosis in patients with CD [10].

The management of PFCD typically involves a combination of biological therapies and surgical procedures [1, 11]. This combined approach aims to control inflammation, preserve anal continence, and prevent disease recurrence. In contrast, the treatment of CGF is exclusively surgical or based on other local techniques, with no indication for systemic pharmacological therapy [12].

Distinguishing between these two types of PF is not always straightforward, especially when PF is the initial and sole manifestation of CD, which occurs in up to 10% of patients [13]. Therefore, accurate differential diagnosis is extremely important, as the therapeutic strategies differ markedly.

The primary objective of this review is to identify and synthesise the main biomarkers involved in the pathogenesis of PFCD. As a secondary objective, it aims to explore and identify biomarkers that may serve as diagnostic tools to distinguish CGF from PFCD, providing practical, patient-centred guidance for clinical decision-making.

Methods

The systematic review was registered in the International Prospective Register of Systematic Reviews (PROSPERO, protocol ID: CRD42024575480), and the findings were reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [14]. Deviations from the registered protocol were as follows: replacement of an author, modification of the review period, exclusion of the Embase database, and removal of age restrictions for participants.

Search strategy

A systematic review was conducted of full-text articles published in English between January 2000 and May 2025. The PubMed and Scopus databases were searched. The search strategy was developed using a combination of Medical Subject Headings (MeSH) and free-text terms related to Crohn’s disease and perianal fistula. The full search strategies are available in the Supplementary Material. Reference lists of potentially relevant articles and previous literature reviews were also screened to identify additional studies.

Study selection

Two reviewers (AS and CR) independently conducted database searches and removed duplicates using EndNote 21.

Titles, abstracts, and full texts of the articles were independently screened by the same two reviewers according to predefined eligibility criteria. In case of disagreement, a third senior reviewer (RM) independently assessed the full texts to reach a consensus.

Eligibility criteria

Retrospective or prospective studies, case series, case-control studies, and cohort studies with a minimum sample size of 10 patients were included. Studies involving patients with PFCD diagnosed by clinical criteria, magnetic resonance imaging (MRI), or endoscopic ultrasound (EUS), and/or histology were considered. Studies assessing any biomarker associated with the pathogenesis or diagnosis of PF, as well as those comparing biomarker levels between patients with PFCD and healthy controls, between patients with PFCD and patients with CD without PF, or between patients with PFCD and patients with CGF, were also included.

We excluded studies involving patients with PF resulting from causes other than CD or cryptoglandular origin (including iatrogenic [postpartum or postsurgical], neoplastic, traumatic, HIV-related, or radiotherapy-induced fistulas), studies with a sample size of < 10 patients with fistulas, studies with unavailable full text, animal or in vitro studies, and studies focusing exclusively on clinical factors and treatment outcomes.

Quality assessment

The methodological quality of the included studies was assessed by consensus between AS and CR using the Newcastle-Ottawa Scale (NOS) [15], applied to case-control and cohort studies, as well as its adapted version for cross-sectional studies, while retaining the classic nine-star scoring to ensure consistency. Based on the total score, studies were classified as high (7–9), moderate (4–6), or low quality (0–3), reflecting the risk of bias. Translational and experimental studies involving human tissue and in vitro models were evaluated using the adapted Quality Assessment Tool for In Vitro Studies (QUIN), considering cell characterization, use of appropriate controls, replication of experiments, validation of results (RNA-seq, qPCR, Western blot, immunofluorescence, spheroid models), statistical analysis, and consistency of conclusions.

Results

A total of 5242 records were identified through database searches: 2527 from PubMed and 2715 from Scopus. Following the removal of 1542 duplicate records, 3700 studies were screened by title and abstract. Forty-six full-text articles were assessed for eligibility, of which 31 met the inclusion criteria and were included in the final systematic review [16–48]. Screening of reference lists from the included studies identified two additional eligible studies. The review process is summarized in the PRISMA flow diagram (Fig. 1).

Fig. 1.

PRISMA flow diagram showing the study selection process

The included studies were conducted across diverse geographic regions, with 19 studies from European countries, 7 from the US, 6 from Asia, and 1 from New Zealand. A comprehensive description of the characteristics of these studies is provided in Tables 1 and 2.

Table 1.

Main characteristics of the pathogenetic studies included in the analysis

Study (author, year, country)	Study type (period)	No. of participants	Aims	Biomarker	Outcomes
Genetic profile
Vermeire et al. [17] 2005 Belgium	Case-control (NR)	981 IBD (203 PFCD) 305 HC	To study mutation prevalence, assess phenotypic expression, and perform conditional analyses to examine evidence for gene-gene interactions	OCTN (OCTN1 C1672T, OCTN2-207) DLG5 (DLG5 G113A, DLG5_e26)	OCTN variants were specifically associated with perianal and penetrating CD
Palmieri et al. [18] 2006 Italy	Case-control (NR)	1510 total 444 CD (52 pCD) 455 UC 611 HC	To investigate these variants in CD and UC, their interaction with CARD15 gene, and their correlation with clinical subphenotypes including PFCD	OCTN1 (SLC22A4): missense polymorphism 1672C-T OCTN2 (SLC22A5): promoter polymorphism-207G-C CARD15: variants R702W, G908R, and L1007fsinsC	TC haplotype is more frequent in CD patients with perianal involvement and stricturing/fistulising disease behaviour Organic cation transporter gene cluster variants may confer susceptibility to both CD and UC, and to PFCD
Jakubowska et al. [19] 2008 Poland	Cross-sectional (NR)	230 total 98 CD (25 PFCD) 132 HC	Assess polymorphisms in CD and their association to intestinal location and EIMs, including perianal disease	CARD15/NOD2 G908R DLG5 R30Q OCTN1 F503L	CARD15/NOD2 G908R is associated with ileal disease but not with perianal disease
Henckaerts et al. [21] 2009 Germany	Cohort (1996–2009)	755 CD (48 PFCD) 367 HC	To examine the influence of recently discovered CD-associated susceptibility loci on changes in disease behaviour and to evaluate whether a genetic risk model for disease progression could be generated	NOD2: rs2066844, rs2066845, rs2066847 L23R: rs11209026 IRGM: rs13361189 ATG16L1: rs2241880 CDKAL1: rs6908425 IL12B: rs1363670 SLC22A4/5: Haplotype variants TNFSF15: rs4263839	The C allele of CDKAL1 rs6908425 significantly increases the risk of developing PF (OR 8.86), whereas the presence of at least one NOD2 mutation is associated with a lower risk (OR 0.56)
Latiano et al. [22] 2009 Italy	Case-control (2004–2005)	1754 total 823 CD (19 pCD) 353 UC 578 HC	To investigate whether these variants are associated with IBD in Italian patients with adult- and childhood-onset disease	IRGM: rs1000113, rs4958847	The IRGM rs4958847 polymorphism is associated with fistulising behaviour and PF in CD
Freire et al. [24] 2011 Portugal	Cohort (NR)	203 CD	To assess whether CARD15 mutations are associated with the risk of developing PFCD and whether these mutations can predict the response of PF to antibiotic treatment	CARD15/NOD2 mutations: R702W (SNP8), G908R (SNP12), and 3020insC (SNP13)	In CD, CARD15 mutations are not associated with to the risk or timing of PF development Patients with PF carrying CARD15 mutations showed poorer response to antibiotics (metronidazole ± ciprofloxacin)
Connelly et al. [25] 2012 USA	Cohort (1998–2011)	116 ileocolic CD (35 septic anal disease: abscesses/fistulas)	To identify SNPs associated with anal disease in general, septic anal disease specifically and the response to anti-TNF treatment	TAGAP: rs212388	SNP rs212388 showed a strong association with the presence and severity of anal disease in ileocolic CD Mutations in TAGAP may predict a milder form and course of pCD
Eglinton et al. [26] 2012 New Zealand	Case-control (2003–2005)	507 CD (137 pCD)	To identify the environmental, clinical, and genetic risk factors for pCD	NOD2/CARD15: R702W, G908R, 1007 fs NOD2: rs72796353 IL2_21 IL23R ATG16L1 NCF4: rs4821544 IGRM TNF-α IBD5 (OCTN1/2) NALP3 CARD8	Only NCF4 showed a significant association with pCD IBD5 (OCTN1/2); TNF-α, and IRGM were not significantly associated with perianal disease in this population NOD2 / CARD15 variants (R702W, G908R, 1007 fs) did not show a consistent association with pCD
Yang et al. [27] 2014 South Korea	Cohort (NR)	906 CD (390 pCD)	To investigate the association between TNFSF15 genotypes and the clinical course of CD	TNFSF15: rs6478108, rs4574921, rs3810936, rs6478109, and rs7848647	Non-risk allele homozygotes of TNFSF15 SNP rs6478108 were independent genetic predictors of stenosing and penetrating CD Rs4574921 was identified as an independent genetic predictor for the development of PFCD
Schnitzler et al. [29] 2015 Germany	Case-control (NR)	2256 total 1537 IBD (1073 CD; 464 UC) 719 HC	To evaluate the association of rs72796353 with IBD susceptibility and phenotype	NOD2 SNP: rs72796353 NOD2 mutations: rs2066844, rs2066845, and rs2066847	The NOD2 IVS4 + 10 A > C variant showed a significant association with PFCD in the absence of other NOD2 mutations, suggesting its role as a novel predictor of pCD
Can et al. [28] 2015 Turkey	Case-control (NR)	300 total 211 IBD (16 pCD) 89 HC	To investigate the involvement of TYK2 polymorphisms in patients with IBD	TYK2: rs280523, rs2304256, rs280519, and rs280496	The presence of rs2304256 (A) was associated with the presence of pCD (P = 0.03) H2/H4 diplotype was observed at higher rates in PFCD
Kaur et al. [31] 2016 USA	Case-control (NR)	1721 CD (368 pfCD)	To identify clinical, serological, and genetic factors associated with pCD	Genetic: SNPs in genes related to IBD (KIF3B, CRTC3, TRAF3IP2, JAZF1, NRIP1, MST1, FUT2, PTGER); autophagy (DAPK1); TNF-α pathways (NUCB2; DAPK1); IFN-γ pathways (DAPK1; NDFIP2) and extracellular matrix and scaffolding proteins (USH1C; NDFIP2; TMC07) and JAK-STAT pathway (STAT3) Serological: ASCA, OmpC, and pANCA	Perianal CD was associated with: higher ASCA IgA and OmpC antibody levels known IBD loci, including KIF3B, CRTC3, TRAF3IP2, JAZF1, NRIP1, MST1, FUT2, and PTGER genes involved in autophagy (DAPK1); TNF-α pathways (NUCB2; DAPK1); IFN-γ pathways (DAPK1; NDFIP2); and extracellular matrix and scaffolding proteins (USH1C; NDFIP2; TMC07) JAK/Stat pathway (STAT3)
Zhang et al. [35] 2019 China	Case-control (2012–2016)	750 total 490 CD (250 PFCD) 260 HC	To explore susceptibility genes for CD and pCD in the Han population from South China	Differences in polymorphisms at 90 genetic loci, including SNPs rs72553867 (IRGM gene of chr 5), rs4409764 (NKX2-3 gene of chr 10) and rs3731772 (AOX1 gene of chr 2), were observed between patients with PFCD, patients with CD, and the HC group	Three SNPs-rs72553867 (IRGM), rs4409764 (NKX2-3), and rs3731772 (AOX1) increase the risk of pCD
Cusato et al. [41] 2024 Italy	Case-control (2016–2022)	206 CD (29 PFCD)	To evaluate the association between SNPs in genes encoding enzymes, transporters, and receptors involved in the VD pathway and the occurrence of pCD	Genetic variants in VDR, CYP27B1, CYP24A1, and GC	The presence of the heterozygous genotype for Apal and BsmI significantly increases the risk of developing pCD
Gisbert-Ferrándiz et al. [42] 2024 Spain	Case-control (NR)	115 CD (41 PFCD)	To analyse the association of BsmI, ApaI, TaqI, and FokI SNPs in the VDR gene with the clinical characteristics of CD	VDR SNP: ApaI (rs7975232)	The aa genotype of the ApaI SNP in the VDR gene is associated with a lower risk of PF in CD and with reduced expression of fibrotic and inflammatory markers in intestinal fibroblasts
Microbiological profile
Tozer et al. [30] 2015 United Kingdom	Cross-sectional (NR)	51 total 20 PFCD 13 luminal CD 18 CGF	To characterize the microbiota of the fistulous tract and rectal mucosa in patients with CD and CGF	Studied bacteria: Bifidobacteria Clostridium coccoides-Eubacterium rectale cluster Bacteroides- Prevotella cluster Faecalibacterium prausnitzii Escherichia coli	Patients with CD without fistula and those with PFCD had fewer Bifidobacteria than those with CGF The tracts of PFCD do not appear to harbor specific bacterial populations Of all the fistula tracts, only one (from a CGF patient) exhibited bacteria that were associated with those of the rectal mucosa
Haac et al. [32] 2019 USA	Cross-sectional (NR)	18 PFCD	To characterize the gut microbiota in PFCD using matched stool and fistula samples	Bacteroides Streptococcus Bifidobacterium Achromobacter Corynebacterium	The microbiota of PF in patients with CD differs from that of stool
Breton et al. [38] 2022 USA	Cross-sectional (NR)	108 total 31 PFCD 36 control group CD 41 HC	To define the microbial signature associated with PFCD in children	Bacterial microbiota	The microbial community within CD-related anorectal fistulas is compositionally and functionally unique Resistance genes to ciprofloxacin and metronidazole were detected in all three types of samples
Immunological profile
Ruffolo et al. [20] 2008 Italy	Cross-sectional (NR)	41 total 17 pCD (14 PFCD) 7 intestinal CD 17 HC	To assess the cytokine profile in the rectal mucosa of patients with pCD and to examine its relationship with the systemic cytokine profile and inflammatory parameters and the need for surgery	Inflammatory cytokines: TNF-α, IL-1β, IL-6, IL-12, and TGF-β1	IL-1β IL-6, and TNF-α appear to play an important role in pCD Mucosal levels of IL-6 and IL-12 are predictors of recurrence and of need for surgery in patients with pCD
Haddow et al. [33] 2019 United Kingdom	Cross-sectional (2014–2015)	61 total 13 PFCD 48 CGF 6 HC	To systematically compare the clinical phenotypes, cytokine profiles, and phosphoprotein profiles of CGF and PFCD	Differences in 30 cytokines and 39 phosphoproteins in PFCD and CGF tissues, as well as in the rectal mucosa	At the internal opening of the fistula, PFCD had higher IL-12p70 levels and lower IL-1RA/IL-1β ratios compared to CGF patients HC rectal mucosa showed higher levels of six phosphoproteins compared to PFCD patients In the fistulous tracts, no differences in cytokines or phosphoproteins levels were observed
Bruckner et al. [37] 2021 Switzerland	Translational experimental in vitro (NR)	26 PFCD (15 non-inflamed; 11 inflamed) PBMCs: 9 PFCD; 2 CD; 5 HC	To investigate the role of T-lymphocytes (T cells) in PF pathogenesis	T cells: CD3⁺CD8⁻, CD3⁺CD8⁺, or CD45⁺CD3⁻	PBMCs from patients with inflamed PFCD contained higher numbers of CD3⁺CD8⁻ T lymphocytes and lower numbers of CD3⁺CD8⁺ T cells compared to those from non-inflamed PFCD patients and HC CD3⁺CD8⁻ and CD3⁺CD8⁺ T cells play a crucial role in inflamed PFCD through the secretion of TNF-α and IL-13
Rizzo et al. [39] 2023 Italy	Translational experimental in vitro (NR)	29 total 22 PFCD 7 rectal cancer patients	The role of ECM dysregulation in the development of PFCD, compared to normal rectal mucosa from patients with rectal cancer	Profiles of 8172 genes and of TNFAI6 (gene encoding TSG-6)	By mediating changes in ECM organization, TSG-6 triggers the EMT transcription factor SNAI1 through the activation of mechanosensitive proteins
Levantovsky et al. [44] 2024 USA	Cross-sectional (NR)	15 PFCD	To investigate the cellular and molecular mechanisms of PFCD and to identify key biomarkers and potential therapeutic targets	CHI3L1 OSM (Oncostatin M) Transcription factors JUN/AP-1 Genes associated with cellular senescence Myeloid–stromal crosstalk signal	Enrichment of myeloid cells and fibroblasts in fistula tracts Cellular crosstalk and senescence, confirmed by immunofluorescence CHI3L1 was highly expressed in stromal cells of fistula tissue Differences in CHI3L1 and OSM expression among monocytes from individuals of African and European ancestry Transcription factor marks associated with stress and inflammatory memory (including JUN/AP-1) and enrichment of genetic loci linked to CD
Metabonomic profile
Adegbola et al. [36] 2021 United Kingdom	Cross-sectional (NR)	50 total 20 PFCD 30 CGF	To identify potential biomarkers of disease, that may improve understanding of its pathogenesis	Differences in 41 metabolites between PFCD and CGF, including amino acids, nucleosides, acylcarnitines, and lipids	In PFCD: Increased metabolites: arginine, threonine, tyrosine, lysine, citrulline, proline, methionine, dimethylarginine, homoarginine, O-acetyl-l-carnitine, N-acetyl-lysine, and octanoyl-l-carnitine Decreased metabolites: adenosine, creatinine, hexanoyl-l-carnitine, monoacylglycerol 18:3, S-methyl-5-thioadenosine, decanoyl-l-carnitine, and palmitoyl-l-carnitine Elevated lipids: hexosylceramide (d18:1/16:0) and diglyceride (38:5)

ASCA anti-Saccharomyces cerevisiae antibodies, CD Crohn’s disease, CGF cryptoglandular perianal fistula, Chr chromosome, EIMs, extraintestinal manifestations, ECM extracellular matrix, EMT epithelial-mesenchymal transition, HC healthy controls, IBD inflammatory bowel disease, IFN-γ interferon-gamma, IL interleukin, JAK/STAT Janus kinase/signal transducer and activator of transcription, NR not reported, OmpC outer membrane porin C, OR odds ratio, PBMCs peripheral blood mononuclear cells, PF perianal fistula, PFCD perianal fistulising Crohn’s disease, pCD perianal Crohn’s disease, SNP single-nucleotide polymorphism, TNF tumour necrosis factor, TNF-α tumour necrosis factor alpha, UC ulcerative colitis

Table 2.

Main characteristics of the diagnostic studies included in the analysis

Study (author, year, country)	Study type (period)	No. of participants	Aims	Biomarker	Outcomes
Histopathological findings
Ramzan et al. [16] 2002 USA	Cohort (1990–1994)	82 CD (21 with granuloma; 61 without granuloma) (19 pCD)	To determine whether the presence of granuloma presence is associated with severity of CD, including fistulising behaviour, perianal disease, disease relapses, and EIMs	Presence of granulomas in tissue biopsies examined microscopically in CD	There were no significant differences between the two groups regarding the presence of fistulising or perianal disease, disease severity, recurrence rate, or EIMs
Denoya et al. [23] 2011 USA	Cohort (2001–2007)	207 CD (43 PFCD)	To correlate the presence and distribution pattern of granulomas in resected specimens with clinical characteristics and outcomes in patients undergoing surgery for CD	Granuloma	Granulomas were associated with a higher frequency of EIMs and pCD, as well as lower BMI and greater prevalence in younger patients and females
Han et al. [43] 2024 China	Case-control (2015–2023)	4430 total (93 PFCD; 4335 with non-CD, non-TB diseases; 2 pulmonary TB)	To assess the histopathological features of granulomas in PFCD	Granuloma	Most PFCD cases show fewer than three granulomas (≤ 3) and no foreign body giant cells Ribbon-like granulomas are only observed in PFCD cases
Inflammatory profile
Stevens et al. [34] 2019 The Netherlands	Cross-sectional (2009–2018)	56 total 37 PFCD (19 with ulcers and 18 without) 19 CGF	To determine the diagnostic accuracy of FC to: a) differentiate between PFCD and CGF b) detect mucosal inflammation in pCD	Faecal calprotectin	FC levels were significantly higher in PFCD compared to CGF (µg/g) FC discriminates PFCD from CGF, even in the absence of intestinal ulcers In active pCD, an elevated FC does not accurately predict the presence of ulcers and should be interpreted with caution
Becker et al. [40] 2024 The Netherlands	Cross-sectional (2019–2022)	63 total 45 PFCD 18 CGF	To prospectively validate this and to analyse whether increased local fistula calprotectin levels are associated with fistula characteristics	Faecal calprotectin-local calprotectin (from fistula tract scrapings)	PFCD and CGF differ in FC levels Local fistula calprotectin production did not account for this difference, implying FC reflects the luminal condition A correlation exists between calprotectin levels in fistula scrapings and CD fistula severity
Perregaard et al. [48] 2025 Denmark	Cross-sectional (2018–2020)	63 total 31 PFCD 32 CGF	To determine the accuracy of FC in distinguishing between PF of CG and CD origin, as well as to compare characteristics in fistula	Faecal calprotectin	FC distinguishes PFCD from CGF, even in clinically managed CD patients with normal endoscopy A FC value ≥ 110 μg/g, combined with clinical suspicion of underlying IBD, should prompt further investigations
Ma et al. [47] 2025 China	Case-control (2021–2024)	193 total 96 PFCD 97 CGF	To evaluate NGAL’s potential as a non-invasive diagnostic biomarker for distinguishing between PFCD and CGF and to assess its correlation with disease severity in PFCD	NGAL Faecal calprotectin C-reactive protein Erythrocyte sedimentation rate	NGAL is identified as a promising biomarker for differentiating between PFCD and CGF NGAL levels show a positive correlation with PFCD severity
Immunological profile
Ouboter et al. [45] 2024 The Netherlands	Cross-sectional (2019–2023)	Cohort 1: 14 PFCD; 7 non-CD fistula Cohort 2: 10 PFCD; 11 non-CD fistula Cohort 3: 5 PFCD	To characterize the composition and spatial localization of disease-associated immune cells in both types of PF using high-dimensional analyses	HLA- DR⁺CD38⁺ effector Th1/17 cells	Increased activated Th1/17 cells (HLA-DR⁺CD38⁺) distinguish PFCD from CGF and represent a promising blood biomarker CD4⁺ HLA-DR⁺CD38⁺ T cells constitute a potential therapeutic target in CD-associated fistulas
Metabonomic profile
Zhu et al. [46] 2024 China	Case-control (July–September 2020)	110 total 55 PFCD 55 simple PF or non-PFCD	To develop a non-invasive detection approach to predict CD in patients with PF	Fifteen and eleven spectral peaks in the urine SERS pattern (phosphoprotein and metabolomics profiles)	The CD group showed increased Raman peaks at 574, 591, 1239, 1276, 1291, and 1499 cm⁻1, and decreased peaks at 480, 518, 641, 677, 698, 715, 732, 889, 1181, 1206, and 1566 cm⁻1

BMI body mass index, CD Crohn’s disease, CGF cryptoglandular perianal fistula, EIMs extraintestinal manifestations, FC faecal calprotectin, IBD inflammatory bowel disease, NGAL neutrophil gelatinase-associated lipocalin, pCD perianal Crohn’s disease, PFCD perianal fistulising Crohn’s disease, TB tuberculosis, Urine SERS pattern urine surface-enhanced Raman spectroscopy pattern

Of the 33 included studies, 25 were rated as high quality or low risk of bias, while 6 were rated as moderate quality, based on assessments performed using the NOS. The two translational and experimental studies involving human tissue and in vitro models were rated as high to moderate quality, according to the adapted version of the QUIN scale (Supplementary Table 1).

Biomarkers of pathogenesis

Genetic markers

Several genes have been implicated in the development and clinical course of pCD. Variants in the NOD2 (CARD15) gene were among the first identified as susceptibility factors for CD and have also been investigated in relation to perianal disease phenotypes. Evidence regarding their role in PFCD remains inconsistent. While some studies reported no significant association between NOD2 mutations and PFCD risk [19, 24], others suggested that specific polymorphisms may influence disease behaviour. For example, variants such as the G allele of rs1363670 (located near the IL12B gene) and the T allele of rs12704036 were associated with stricturing disease and early fistula development in males [21]. In addition, the rs72796353 polymorphism, although not associated with overall susceptibility to CD, was associated with PFCD (OR = 5.27; 95% CI 2.75–10.12 vs. NOD2 wild-type carriers; P < 0.001), particularly in patients lacking common NOD2 mutations [29].

Beyond NOD2, other loci may contribute to PFCD susceptibility. The rs6908425 variant in CDKAL1 gene and the absence of NOD2 mutations were associated with an increased risk of PFCD (OR = 8.86; 95% CI 1.13–69.78; P = 0.04 and OR = 0.56; 95% CI 0.38–0.83; P = 0.004, respectively), particularly in patients with colonic involvement and smoking habits [21]. Furthermore, NOD2 mutations were associated with poorer response to antibiotic therapy (metronidazole ± ciprofloxacin; P = 0.041), suggesting a potential genetic influence on treatment outcomes [24].

Variants in genes involved in innate immunity, particularly IRGM and NCF4, were associated with perianal disease. IRGM was identified as a genetic risk factor for the fistulising phenotype of CD, including PFCD [22, 35]. In the European population, the rs4958847 polymorphism at IRGM was associated with both CD and the development of PF (OR = 1.55; 95% CI 1.01–2.38; P = 0.045) [22], with similar findings in the Han Chinese population for rs72553867 [35]. Additional variants included rs4409764 at NKX2-3 gene and rs3731772 at AOX1 gene [35].

Multiple genetic loci were linked with more aggressive CD phenotypes. NCF4 variants were associated with pCD, alongside clinical risk factors such as younger age at diagnosis, complicated intestinal disease, and ileal disease location [26]. Genetic analysis demonstrated a significant association between NCF4 gene variants when comparing perianal Crohn's disease (pCD) patients with non-perianal CD patients and healthy controls; no significant associations were observed for IBD5, TNF-α, or IRGM genes [26].

Genetic variability within the JAK-STAT pathway was associated with PFCD (P < 0.001) [31], alongside distal colonic involvement, CD-related serology, and family history of IBD. TYK2 gene variants were also associated with CD and ulcerative colitis (UC) susceptibility, as well as disease phenotype, namely the rs2304256 (A) genotype was more frequent in perineal involvement (P = 0.03), whereas the rs2304256 (C) was less frequent. The H2/H4 diplotype (rs280519 AG/rs2304256 AA) was associated with increased risk of perianal involvement [28].

At the IBD5 locus, OCTN1 and OCTN2 variants were associated with pCD and penetrating CD but not with overall IBD susceptibility [17]. The SLC22A5 -207G-C and SLC22A4 1672C-T variants were also associated with increased risk of CD and UC (P = 0.007). The TC haplotype was associated with PFCD and stricturing-fistulising disease behaviour (P = 0.037) and with extensive disease and increased need for immunosuppression in UC [18].

Additional genetic associations included TNFSF15 rs6478108 (CC), identified as a predictor of CD complications, and rs4574921, which predicted PFCD development (HR = 2.39, 95% CI 1.20–4.73; P = 0.013), although treatment effects were not evaluated [27]. TAGAP rs212388 (G allele) was associated with reduced risk of perianal septic complications, such as abscesses and fistulas, and lower need for ostomy surgery [25], suggesting a potential protective effect against severe septic forms of PFCD.

VDR polymorphisms (ApaI Aa, BsmI Bb) were associated with development of pCD complications (P = 0.01 and P = 0.02, respectively) [41], whereas the Apal aa genotype was associated with reduced risk of fistulising behaviour [42].

Microbiological markers

Microbial dysbiosis was consistently reported in PFCD, particularly in rectal mucosa and fistula-associated samples [30, 32, 38]. PFCD was associated with a reduced bacterial load in the rectal mucosa, with decreased abundances of Bacteroides and Bifidobacterium compared with CGF [30]. Reduced Bifidobacterium abundance in rectal mucosa was also observed in patients with non-fistulising CD compared with CGF [30].

Fistula-associated microbiota differed from stool microbiota, with enrichment of Achromobacter and Corynebacterium within fistula tracts and a higher abundance of Bifidobacterium in stool samples [32]. Metagenomic studies in paediatric CD reported increased Proteobacteria, depletion of butyrate-producing bacteria, and the presence of antibiotic resistance genes in fistula-associated microbiota [38].

Overall, PFCD was associated with a reduction in commensal bacteria, enrichment of Proteobacteria and other opportunistic taxa, and altered microbial functional profiles [30, 32, 38].

Immunological markers

Elevated levels of IL-1β, IL-6, and TNF-α were reported in the rectal mucosa of patients with pCD and were associated with higher disease activity and surgical need. PFCD showed greater severity and anatomical complexity than idiopathic PF [33]. Tissue biopsies, cytokines, and phosphoprotein profiling revealed elevated IL-12p70 levels and a reduced IL-1RA/IL-1β ratio, indicating impaired anti-inflammatory regulation in PFCD [33].

T-cell subsets (CD3 + CD8- and CD3 + CD8 +) were implicated in pathogenesis, associated with TNF-α overexpression and epithelial-mesenchymal transition (EMT), and involved IL-13 and IL-22 [37].

Extracellular matrix dysregulation was identified in PFCD, with increased TNFAIP6, altered ECM organisation, and upregulation of integrins, metalloproteinases, and collagen deposition [39].

Immune-stromal interactions were observed in PFCD, involving myeloid cells and fibroblasts, alongside senescence pathways and epigenetic changes (JUN/AP-1). Differential expression of CHI3L1 and OSM was reported between monocytes from individuals of African and European ancestry [44].

Metabolic markers

Distinct metabolite profiles were observed between PFCD and CGF [36]. Nineteen metabolites differed significantly between the two conditions. PFCD was associated with increased levels of arginine, threonine, tyrosine, lysine, citrulline, proline, methionine, dimethylarginine, homoarginine, O-acetyl-l-carnitine, N-acetyl-lysine, and octanoyl-l-carnitine. Lower levels of adenosine, creatinine, hexanoyl-l-carnitine, monoacylglycerol 18:3, S-methyl-5-thioadenosine, decanoyl-l-carnitine, and palmitoyl-l-carnitine were observed in PFCD than in CGF. Higher levels of hexosylceramide and diglyceride were also identified in PFCD [36].

Diagnostic biomarkers

Histopathological markers

The prognostic value of granulomas in CD remains uncertain. Ramzan et al. [16] reported no association between granuloma presence and disease severity, pCD, extraintestinal manifestations (EIMs), or recurrence, indicating limited prognostic utility and supporting their role primarily in differential diagnosis. In contrast, Denoya et al. [23] identified distinct clinical associations in patients with granulomas in surgical specimens, including higher prevalence in females (P = 0.017), younger age at presentation (P < 0.001), lower body mass index (P = 0.037), increased incidence of pCD (P = 0.012), and greater occurrence of EIMs (P = 0.026). More recently, Han et al. [43] described a distinct granuloma subtype in PFCD, characterised by a band-like architecture, smaller size, and absence of multinucleated giant cells, which may facilitate differentiation between CD-related and non-CD-related disease.

Inflammatory markers

Faecal calprotectin (FC) is a useful non-invasive biomarker for differentiating PFCD from non-Crohn’s fistulas. Its effectiveness has been demonstrated using a cut-off value of ≥ 150 μg/g, achieving a sensitivity of 81% and a specificity of 89% [34]. FC levels were higher in PFCD compared with CGF (354.3 vs. 47.3 μg/g, P = 0.003), and calprotectin measured directly in the fistulous tract was correlated with clinical and radiological severity, suggesting potential prognostic relevance [40]. In addition, FC was validated as a screening tool in patients with PF (OR = 12.5; 95% CI 3.77–41.4; P < 0.0001), with a cut-off of ≥ 110 μg/g associated with CD (sensitivity = 76% and specificity = 80%), even in patients without endoscopic activity or receiving treatment [48].

Serum neutrophil gelatinase-associated lipocalin (NGAL) demonstrated high diagnostic performance (AUC = 0.927; 95% CI, 0.890–0.964), exceeding that reported for FC. NGAL also correlated with clinical parameters and inflammatory markers, supporting its utility for both diagnosis and assessment of PFCD severity [47].

Immunological and metabolomic markers

A specific population of activated CD4⁺ T cells (HLA-DR⁺CD38⁺) with a Th1/Th17 profile was identified and showed a significant association with PFCD. These cells were detected in both fistula tissue and peripheral blood. Their interaction with CCR6⁺ B lymphocytes and macrophages suggested a distinct immunological mechanism, with potential diagnostic and therapeutic implications [45].

Urinary Raman spectroscopy combined with machine learning demonstrated discriminatory ability between PFCD and CGF, with sensitivity = 71%, specificity = 80%, and accuracy 76% (the leave-one-patient-out cross-validation), indicating systemic metabolic alterations. Significant differences in 11 spectral peaks were observed, supporting the potential of a urinary biochemical signature as a complementary diagnostic tool [46].

Discussion

In this systematic review of 33 articles, we provide a comprehensive synthesis of the current literature on biomarkers involved in the pathogenesis and diagnosis of PFCD. Overall, the studies reviewed consistently demonstrate that PFCD is a biologically heterogeneous and multidimensional condition, arising from the dynamic interaction among genetic predisposition, environmental exposures, dysregulated immune responses, metabolic alterations, and processes of tissue remodelling. This complexity underpins the marked variability in disease course, severity, and therapeutic response among patients.

Genetic studies further reinforce the concept that CD results from a complex interplay between hereditary susceptibility and environmental factors, rather than being driven by a single gene. Despite the growing number of polymorphisms associated with PFCD, most exhibit small effect sizes and limited predictive value when considered individually, which restricts their clinical utility as independent biomarkers. Nevertheless, several genetic variants have been consistently associated with specific clinical phenotypes, particularly perianal disease, supporting a potential role for genetic profiling in risk stratification.

In European cohorts, polymorphisms in the OCTN1/2 (SLC22A4/5) [17, 18], IRGM (rs4958847) [22], CDKAL1 (rs6908425) [21], and NOD2 (rs72796353) genes [29] have been reproducibly linked to perianal involvement. Similarly, in Asian populations, significant associations have been identified with the TNFSF15 rs4574921 (CC genotype) gene in Korean patients [27], as well as variants in IRGM (rs72553867), NKX2-3 (rs4409764), and AOX1 (rs3731772) in Chinese patients [35]. Additionally, variants in the NCF4 gene have been reported in cohorts from New Zealand [26], further supporting the contribution of pathways related to innate immune responses in the development of perianal disease.

Beyond associations with individual genes, broader signalling pathways appear to play a key role in disease aggressiveness. The JAK-STAT pathway, particularly mediated by variants in the TYK2 (rs2304256) gene [28], has been associated with more severe and fistulising forms of perianal disease [31], highlighting its relevance in the regulation of inflammation and immune responses. In contrast, variants in the TAGAP gene appear to confer a protective effect against severe septic perianal complications, suggesting a modulatory role in clinical expression and disease progression [25].

Consistently, the most robust associations involve variants in genes related to innate immunity and immune regulation, namely NOD2 (CARD15), IRGM, NCF4, and components of the JAK-STAT pathway, particularly TYK2, reinforcing the central role of immune dysfunction in the development of more severe perianal phenotypes. Associations with intermediate levels of evidence have also been reported, often population-dependent, including variants in OCTN1/2 (IBD5) and TNFSF15, which are associated with more aggressive phenotypes and fistulising complications, particularly in specific ethnic contexts.

Some emerging genetic markers, such as NKX2-3 and AOX1, remain insufficiently validated and require confirmation in independent cohorts before any potential clinical application [35]. Similarly, the role of the vitamin D receptor (VDR) pathway remains controversial, with inconsistent findings that may reflect methodological differences as well as the influence of environmental and nutritional factors [41, 42].

Substantial inter-study heterogeneity—encompassing ethnic variation, phenotypic definitions, and methodological differences—limits direct comparability and likely underlies the inconsistencies across findings. Nevertheless, genotyping of specific SNPs appears promising for predicting more severe phenotypes, reinforcing the emerging role of precision medicine in risk stratification and the personalised management of CD.

The gut microbiota also plays a key role in the pathogenesis of PFCD. According to the cryptoglandular theory, chronic bacterial infection is considered the main factor underlying the development and persistence of fistulas [5]. However, some authors have questioned this hypothesis, arguing that bacterial presence in both CGF and PFCD is often limited and may not be decisive for disease maintenance, with the composition and function of the microbiota potentially being more relevant [30, 49].

In this context, recent studies suggest the presence of phenotype-specific microbial signatures, with reduced microbial diversity and shifts in key taxa associated with disease severity [50].

Findings such as the enrichment of Achromobacter and Corynebacterium in fistula tracts and increased faecal Bifidobacterium levels [32], as well as the higher abundance of Proteobacteria and antibiotic resistance genes in the fistula microbiome [38], support a role for niche-specific dysbiosis. However, the predominance of observational study designs and the lack of functional analyses limit causal interpretation.

From a mechanistic perspective, microbial components such as muramyl dipeptide (MDP) may contribute to fistula formation in CD by inducing EMT. This process leads to the activation of inflammatory mediators, namely TNF-α, TGF-β, SNAIL1, and IL-13 [51, 52], which in turn promote the upregulation of matrix metalloproteinases MMP-3 and MMP-9, favouring extracellular matrix degradation and the perpetuation of fistulous lesions [53].

Beyond the bacterial component, emerging evidence also suggests a role for the mycobiota. The recent identification of the fungus Debaryomyces hansenii in inflamed intestinal tissue represents an important advance and suggests that fungal dysbiosis may act not only as an inflammatory trigger but also as a factor contributing to chronicity [54].

Overall, studies indicate relatively consistent trends of intestinal dysbiosis in PFCD; however, variability in sampling methods, study populations, and sequencing approaches continues to hinder the reliable identification of microbial signatures.

In the immunological context, the pathogenesis of PFCD involves complex inflammatory and structural mechanisms that extend beyond the classical TNF-α pathway [37, 51, 55]. Previous studies have shown elevated levels of multiple pro-inflammatory cytokines, including IL-1β, IL-6, TNF-α, and IL-12p70 [20, 33, 56, 57], which are associated with increased disease activity and fistulising complexity. However, inconsistent correlations with clinical outcomes limit their current clinical applicability.

In CGF, several studies have demonstrated elevated pro-inflammatory cytokine profiles in fistula tissue. However, these findings should be interpreted with caution because of methodological heterogeneity, differences in study populations, and variability in sampling sites. Increased levels of IL-1β and IL-8 have been reported in chronic fistula tissue compared with healthy controls, with IL-1β predominantly localised in the distal regions of the tract and IL-8 more abundant proximally [56]. Similarly, elevated expression of IL-1β, IL-8, IL-12p40, and TNF-α has been identified in the distal fistula tract, although these increases did not consistently correlate with specific clinical outcomes [52]. Additionally, increased IL-1β and IL-8 expression has been observed in chronic CGF tissues, along with histological features suggestive of EMT, a process implicated in tissue remodelling and fistula persistence [58]. Although these findings highlight a complex inflammatory microenvironment, variability across studies limits their direct clinical translation.

Despite these advances, significant knowledge gaps remain, including the precise role of IL-22 and population-specific differences in cytokine expression and molecular pathways. Overall, the evidence supports a multifactorial inflammatory environment but remains insufficient to justify broad therapeutic expansion beyond the TNF-α pathway, emphasising the need for more robust mechanistic and clinical studies.

Metabolomics has emerged as an innovative tool for characterising PFCD. Alterations such as increased dimethylarginine and diacylglycerols [36, 59] alongside reduced decanoyl-l-carnitine, suggest that metabolic disturbances may promote both EMT and tissue remodelling. However, current evidence does not establish causality, and these alterations may, at least in part, reflect chronic inflammation. The observation of elevated diacylglycerol levels in fistula tissue supports previous systemic metabolic findings [60]; however, it does not clarify the directionality of these alterations and instead reinforces the hypothesis that altered metabolic pathways are involved in fistula pathophysiology.

In parallel, surface-enhanced Raman spectroscopy (SERS) has recently identified highly sensitive and specific urinary biomarkers [46], making urine a valuable non-invasive option for early diagnosis and monitoring of PFCD activity—particularly important for this frequently underdiagnosed phenotype. However, these findings remain at an exploratory stage and require validation in larger, independent cohorts.

More recently, studies in cell and molecular biology have shown that PFCD results from a complex crosstalk between immune and stromal cells, marked by senescence, epigenetic alterations, and population-specific differences [44]. These mechanisms indicate that fistula formation results not only from chronic inflammation but also from profound changes in tissue architecture and cellular behaviour, reflecting a broader process of tissue reprogramming. Although these findings expand current understanding, their predominantly descriptive nature limits causal inference in vivo, and the hierarchy of mechanisms underlying fistula formation remains unclear.

Although granulomas are relevant for histological diagnosis, available evidence indicates that their prognostic value is limited [16, 23, 43]. More recent studies suggest that specific granuloma features may help identify clinically more complex cases [43]; however, these findings remain isolated and lack external validation. Therefore, their prognostic value and therapeutic utility remain uncertain, and they should be interpreted with caution, at most as an adjunct within the overall clinical context.

Among inflammatory markers, faecal calprotectin (FC) remains a valuable tool for distinguishing PFCD from CGF and for assessing local inflammatory activity [34, 40, 48]. Serum NGAL has also emerged as a promising non-invasive marker of perianal involvement in CD patients [47]; however, the evidence remains preliminary.

Recent findings in immunology and metabolomics, including the identification of specific T-cell subpopulations and alterations in urinary profiles, suggest that PFCD involves distinct biological mechanisms with potential diagnostic and therapeutic value [46]. Nevertheless, these findings require validation in larger cohorts using standardised methodologies before clinical implementation.

Across domains, evidence base remains largely observational and heterogeneous, limiting causal inference and highlighting the need for prospective, standardised validation studies. Collectively, advances in genetics, microbiota, immunology, metabolomics, and cellular biology underscore the multidimensional nature of PFCD (Table 3). Biomarkers derived from these domains may improve diagnostic accuracy, refine risk stratification, and support personalised therapeutic decision-making.

Table 3.

Factors associated with risk and activity of PFCD

Domain	Key factors/impact	References
Genetics	Variants [OCTN1/2 (SLC22A4/5), NOD2, CDKAL1, TNFSF15, TYK2, IGRM, AOX1, NKX2-3, VDR] → increased risk of perianal phenotype; population differences (e.g., CHI3L1, OSM) modulate susceptibility	[17, 18, 21, 22, 26–29, 31, 35, 41]
Microbiota	Achromobacter, Corynebacterium, Proteobacteria, Debaryomyces → inflammation and fistula maintenance	[32, 38, 53]
Inflammation	IL-1β, IL-6, TNF-α, IL-12p70, IL-13, IL-22, EMT → high activity and complex fistulas; crosstalk between immune and stromal cells	[20, 33, 37, 39, 50, 54–56]
Cellular senescence/ epigenetics	Senescence pathways, epigenetic modifications, transcription factor footprints (JUN/AP-1) → memory of inflammation and persistent fibrotic/fistulising response	[44]
Tissue remodelling/ metabolomics	MMP-3/9, dimethylarginine, diacylglycerols → fistula persistence and progression; fibroblast activation and ECM remodelling	[36, 50–52, 58, 59]
Diagnostic biomarkers	Faecal calprotectin, serum NGAL, urinary SERS → non-invasive assessment of activity	[34, 40, 46–48]
Histology	Recent morphological features and presence of granulomas → differentiation of complex fistulas	[16, 23, 43]

This review employed a structured methodology to identify, select, and critically appraise relevant studies, thereby minimising investigator bias and ensuring an objective synthesis of the available evidence. However, several limitations should be acknowledged. There is considerable heterogeneity among the biomarkers evaluated across the included studies, which hinders comparability and reduces the consistency of conclusions regarding potential microbial or inflammatory signatures. In addition, the included studies were conducted in different populations, which may have introduced selection bias and limited the generalisability of the findings to broader clinical settings. Most studies were retrospective and did not adequately control for confounding factors, which may have influenced the observed associations between biomarkers and disease phenotypes. Consequently, the inability to perform a meta-analysis further limits the robustness and precision of the conclusions, thereby reducing their clinical applicability, particularly regarding the use of biomarkers for diagnosis, stratification, and prognosis.

Future directions

Progress toward clinically actionable biomarkers for PFCD will depend on well-designed, large-scale, multicentre prospective studies, ideally including robust validation cohorts to ensure reproducibility and generalisability. Standardisation of biomarker assessment methodologies will be essential to facilitate comparability across studies. In parallel, the integration of multi-omic approaches—including genomics, transcriptomics, proteomics, and metabolomics—may provide a more comprehensive understanding of the molecular mechanisms underlying PFCD. Longitudinal study designs will be particularly important to clarify temporal dynamics and disease evolution. Further work should also focus on the development of integrated biomarker panels combining genetic, microbial, immunological, and metabolic data, which may improve patient stratification and support more precise clinical decision-making. Ultimately, these approaches are expected to refine disease phenotyping and support the transition towards personalised, multimodal therapeutic strategies in PFCD.

Conclusions

PFCD is a biologically complex and heterogeneous condition driven by the interplay of genetic susceptibility, immune dysregulation, microbial imbalance, metabolic alterations, and tissue remodelling. This multifactorial pathophysiology underlies its variable clinical behaviour and therapeutic response. Although a wide range of biomarkers has been associated with disease phenotype and activity, the current evidence remains largely observational and heterogeneous, limiting their immediate clinical applicability.

Overall, emerging data support the potential value of integrated biomarker approaches for improving disease characterisation and stratification in PFCD, although their incorporation into routine clinical practice requires further validation.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (DOCX 16 KB) ^{(13.3KB, docx)}

Supplementary file2 (DOCX 97 KB) ^{(74.3KB, docx)}

Abbreviations

AOX1

Aldehyde oxidase 1

AUC

Area under the curve

CARD15

Caspase recruitment domain-containing protein 15

CDKAL1

Cyclin-dependent kinase 5 regulatory subunit associated protein 1-like 1

CGF

Cryptoglandular fistulas

CHI3L1

Chitinase-3-like protein 1

ECM

Extracellular matrix

EIMs

Extraintestinal manifestations

EMT

Epithelial-mesenchymal transition

FC

Faecal calprotectin

IRGM

Immunity-related GTPase family M

JAK-STAT

Janus kinases-signal transducers and activators of transcription

JUN/AP-1

Jun proto-oncogene/activator protein 1

MMPs

Matrix metalloproteinases

NGAL

Neutrophil gelatinase-associated lipocalin

NKX2-3

NK2 homeobox 3

NOD2

Nucleotide-binding oligomerization domain-containing protein 2

NOS

Newcastle–Ottawa scale

OCTN1/2

Organic cation transporter

OSM

Oncostatin M

pCD

Perianal Crohn’s disease

PF

Perianal fistula

PFCD

Perianal fistulas in Crohn’s disease

QUIN

Quality assessment tool for in vitro studies

SERS

Surface-enhanced Raman spectroscopy

SLC22A4/5

Solute carrier family 22 member 4/5

SNAIL1

Snail family transcriptional repressor 1

SNP

Single nucleotide polymorphism

TAGAP

T-cell activation Rho GTPase activating protein

Th1/17

T helper 1/17 cells

TNFAIP6

Tumour necrosis factor alpha-induced protein 6

TNFSF15

Tumour necrosis factor superfamily member 15

Author contributions

ACS: Conception and design, acquisition of data, analysis and interpretation of data and drafting the article; RM: analysis and interpretation of data and critical revision; LL: critical revision; CR: acquisition of data and analysis and interpretation of data; PL: critical revision; MDS: study supervision; Final approval of the article: All authors approved the final version of the manuscript.

Funding

Open access funding provided by FCT|FCCN (b-on). Open Access funding was provided by the Unidade Local de Saúde de Santo António.

Data availability

This systematic review did not generate new datasets. All data were extracted from published studies identified through systematic searches of PubMed and Scopus and are available within the original articles cited in the reference list.

Declarations

Ethical approval

This study is based on previously published studies and did not involve the collection of human data or any intervention involving human participants.

Conflict of interest

The authors declare no competing interests.

Publication statement

We confirm that this manuscript including related data, figures, and tables has not been previously published and is not under consideration elsewhere.