Article Text


Modern management of perianal fistulas in Crohn’s disease: future directions
  1. Paulo Gustavo Kotze1,
  2. Bo Shen2,
  3. Amy Lightner3,
  4. Takayuki Yamamoto4,
  5. Antonino Spinelli5,
  6. Subrata Ghosh6,
  7. Remo Panaccione1
  1. 1 Division of Gastroenterology and Hepatology, Inflammatory Bowel Disease Unit, Cumming School of Medicine, University of Calgary, Calgary, Canada
  2. 2 Interventional IBD Unit, Cleveland Clinic Foundation, Cleveland, Ohio, USA
  3. 3 Division of Colorectal Surgery, Mayo Clinic, Rochester, Minnesota, USA
  4. 4 IBD Centre, Yokkaichi Hazu Medical Centre, Yokkaichi, Japan
  5. 5 Division of Colon and Rectal Surgery, Humanitas Clinical and Research Center, Rozzano, Italy
  6. 6 Institute of Translational Medicine, University of Birmingham, Birmingham, UK
  1. Correspondence to Dr Paulo Gustavo Kotze, Colorectal Surgery Unit, Cajuru University Hospital, Catholic University of Parana, Curitiba PR 80045-165, Brazil; pgkotze{at}


Perianal fistulae in patients with Crohn’s disease (CD) can be associated with significant morbidity resulting in negative impact on quality of life. The last two decades have seen significant advancements in the management of perianal fistulas in CD, which has evolved into a multidisciplinary approach that includes gastroenterologists, colorectal surgeons, endoscopists and radiologists. Despite the introduction of new medical therapies such as antitumour necrosis factor and novel models of care delivery, the best fistula healing rates reported with combined medical and surgical approaches are approximately 50%. More recently, newer biologics, cell-based therapies as well as novel endoscopic and surgical techniques have been introduced raising new hopes that outcomes can be improved upon. In this review, we describe the modern management and the most recent advances in the management of complex perianal fistulising CD, which will likely impact clinical practice. We will explore optimal use of both older and newer biological agents, as well as new data on cell-based therapies. In addition, new techniques in endoscopic and surgical approaches will be discussed.

  • crohn’s disease
  • perianal fistula
  • biologics
  • colorectal surgery
  • therapeutic endoscopy
View Full Text

Statistics from


Perianal fistulas comprise a common and disabling phenotype of Crohn’s disease (CD). In 1938, Penner and Crohn described three cases with perianal fistulae secondary to regional ileitis, noting the difficulty in managing what was then believed to be a complication of the disease.1 Perianal fistulising Crohn’s disease (PFCD) is associated with severe symptoms such as pain, embarrassing discharge and significant reduction in quality of life resulting from physical impairment and impairment of sexual and psychological function. In more severe cases, faecal incontinence can occur furthering morbidity. In refractory cases, many patients face the decision of having a stoma and/or requiring proctectomy.2

A population-based study from Olmsted County documented that the cumulative risk of a patient with CD to develop perianal fistulas was 22% at 10 years and 26% at 20 years after diagnosis.3 PFCD can be present in up to 43% of CD cases and in some cases (~5%) it can be the patient’s only manifestation of CD.4 In approximately 10% of patients, the presence of perianal fistulas can be the initial manifestation of CD.5 PFCD tends to be more prevalent in colonic disease and rectal involvement than in those with isolated ileal disease.6 Perianal fistulas in CD can be simple or complex (figure 1), and treatment algorithms vary according to the severity of the perianal involvement and this classification.7 Simple superficial fistulas, for instance, usually can be treated with fistulotomy, rarely require more complex medical (such as antitumour necrosis factor (anti-TNF)) or surgical therapies. Complex fistulas, however, by their nature require different strategies and often an individualised approach.

Figure 1

American Gastroenterological Association (AGA) classification of perianal fistulas in CD. Adapted from Sandborn et al. 7 Simple fistulas described as low fistulae, with no abscess or stenosis, with a single external opening, no associated rectovaginal fistula, with or without associated proctitis. Complex fistulas described as high fistulae, with single or multiple external openings, with or without associated abscesses, stenosis, rectovaginal fistula and/or proctitis. Courtesy of Amy Lightner (USA).

Even with the significant medical and surgical therapeutic advances in the management of CD over the last decades, the treatment of PFCD remains challenging. With the knowledge that active rectal inflammation is an important driver of the pathogenesis, and persistence of inflammation a barrier to fistula healing, there remains a need to optimise medical therapy to achieve mucosal healing in order to maximise the chance for successful outcomes.4 In addition, surgery is often needed to correct mechanical abnormalities in the perineum, mostly secondary to chronic inflammatory disease, and several surgical options have been studied and described.8 Thus, a combined medical-surgical management constitutes the best approach to achieve fistula healing.9 However, to date, the combination of anti-TNF agents (infliximab (IFX), adalimumab (ADA) or certolizumab pegol (CZP)) with examination under anaesthesia (EUA), seton placement and curettage of the fistula tracks leads to perineal fistula healing in approximately 50% of the cases.10–13

There is clearly room for improvement in the healing rates of complex perianal fistulas in CD, and several novel strategies, approaches and techniques are being studied. Optimisation of biological therapy, with different doses, local injections and the use of new biological agents provide flexibility to potentially achieve better results.14–16 Novel endoscopic techniques have been described representing a less invasive option to surgery.17 From the surgical standpoint, local injection of stem cells, novel techniques as the ligation of the inter sphincteric tract (LIFT), video-assisted anal fistula treatment (VAAFT) and the use of plugs are also under evaluation.18–21

The aim of this review is to demonstrate the evidence and discuss in detail the novel clinical, endoscopic and surgical approaches that constitute the modern multidisciplinary management of PFCD.

Conventional non-biologic medical therapy, biologics and perianal CD: the old and the new

The classification, diagnosis, clinical and radiological scoring systems and a detailed assessment of current strategies for the management of complex fistulising CD have been recently published in other reviews.2 22 23 In this section, we will focus on optimisation of medical therapy for PFCD and potential future strategies with biological agents in this phenotype of the disease.

Conventional therapy

There is no role for the use of aminosalycilates in the management of PFCD. Corticosteroids are also not recommended for this phenotype of CD, as they do not result in fistula closure and may in fact precipitate abscess formation. They may be used to treat associated luminal disease but this needs to be done with proper control of any perianal sepsis.2 22

Conventional agents that continue to be used include antibiotics and the thiopurines. The results of the main studies with antibiotics and azathioprine (AZA)/6-mercaptopurine (6-MP) in PFCD are summarised in table 1. Metronidazole and ciprofloxacin are common antibiotics used in the management of PFCD. However, the evidence surrounding their use is limited to small studies.24 Nevertheless, they often lead to symptomatic improvement with decrease in fistula drainage. A trial as stand-alone therapy is appropriate in simple fistula. In complex disease, their use as monotherapy is limited due to adverse events and intolerability associated with prolonged therapy.2 Many of the authors however will use these antibiotics in combination with other therapies in more complex disease to prevent the development of abscess formation. Ciprofloxacin has formally been studied in this manner and was shown to be effective in combination with ADA in a prospective trial, improving the rates of fistula closure.25 Although, ciprofloxacin and metronidazole are often used, other antibiotics with broad spectrum and good tissue penetration such as amoxicillin-clavulic acid are also used and may be associated with better tolerability in those without penicillin allergy.

Table 1

Main studies with conventional therapy in perianal fistulising Crohn’s disease

Despite their historical uses there are no prospective trials of thiopurines such as AZA and 6-MP in PFCD. A meta-analysis that included studies where fistula closure was included as a secondary end point reported a clinical response rate of 54% with AZA as compared with 21% in placebo.26 These data need to be interpreted carefully as the definitions of response varied and it is difficult to ascertain what the true fistula closure rate is. In the era of alternative options, thiopurines should not be used as stand-alone therapy but continue to serve an important role in combination therapy with anti-TNF therapy. The use of calcineurin inhibitors and methotrexate in PFCD is limited to small uncontrolled series, and currently are not recommended as first-line therapies, mostly after the approval of anti-TNF agents.23

Anti-TNF therapy

Two prospective clinical trials have been published with anti-TNF agents in fistulising CD. Present et al 27 published the first clinical trial demonstrating the efficacy of IFX in healing abdominal and perianal fistulas associated with CD. The study included 94 patients, the majority (90%) having perianal fistulas. The primary end point was a 50% reduction in fistula drainage, which was observed in 68% of patients in the IFX 5 mg/kg dose and in 26% of the patients with placebo (P=0.002). Complete healing of the fistulas was a secondary end point, and it was achieved in 55% of patients on IFX 5 mg/kg versus 13% in the placebo group (P=0.001). Sands et al evaluated the efficacy of IFX in an induction and maintenance trial (A Crohn’s Disease Clinical Trial Evaluating Infliximab in a New Long-Term Treatment Regimen in Patients with Fistulising CD (ACCENT II)) prospective trial). In this randomised placebo controlled trial (n=304), patients received open-label induction with IFX 5 mg/kg at weeks 0, 2 and 6 and were then assessed for response.28 One hundred ninety-five patients (64%) achieved response at weeks 10 and 14 and were subsequently randomised to receive maintenance IFX 5 mg/kg every 8 weeks or placebo. The primary end point was the time to loss of response to IFX, which was longer in the IFX-treated group compared with placebo (>40 weeks vs 14 weeks, P<0.001). At week 54, complete healing of the fistulas was observed in 36% in the IFX group and in 19% in placebo group (P=0.009).

There are no other prospective randomised controlled data from other anti-TNF therapies exclusively in fistulising CD. However, subgroup analyses from the pivotal phase III programme as well as the results from real-world cohorts suggest efficacy. A subanalysis of the Crohn’s Trial of the Fully Human Antibody Adalimumab for Remission Maintenance study demonstrated the efficacy ADA in maintaining fistula healing in a significant proportion of patients up to 2 years after induction.29 A Spanish multicentre retrospective study also demonstrated the efficacy of ADA with remission rates in complex perianal fistulas (defined by absence of drainage after therapy) in CD of 41% at 6 months and 29% at 12 months.30 Another small prospective study with ADA suggested that its efficacy can be increased with the addition of ciprofloxacin in order to heal fistulas in CD.25 Combined therapy with ADA+ciprofloxacin was more effective than ADA+placebo in fistula closure after 12 weeks (65% vs 33%, respectively, P=0.009). However, this difference was not maintained through week 24 (P=0.22), suggesting that antibiotics might play a role in induction therapy in combination with an anti-TNF agent in perianal CD. A subanalysis from the Pegylated Antibody Fragment Evaluation in Crohn’s Disease: Safety and Efficacy 2 trial with certolizumab pegol (CZP) also demonstrated efficacy in healing perianal fistulas in CD.31 Complete healing was observed in 36% of the patients treated with CZP as compared with 17% of the placebo group (P=0.038).

In luminal CD, combination of therapy with AZA and IFX is superior to monotherapy with each drug alone, as demonstrated in the Study of Biologic and Immunomodulator Naive Patients in Crohn’s Disease (SONIC) trial.32 Although high-quality data do not exist in PFCD, this strategy would also seem reasonable given that PFCD is considered a severe form of the disease. The authors believe that anti-TNF therapy in combination with immunomodulators (particularly thiopurines) remains the cornerstone of medical therapy for patients with complex perianal fistula.

Combination of anti-TNF therapy and surgery

Retrospective studies have demonstrated that the healing rates of PFCD with anti-TNF is increased when combined with seton placement and/or fistula track curettage. Regueiro et al demonstrated that patients treated with IFX had improved rates of response when they underwent an examination under anaesthesia (EUA) and seton placement, as compared with the use of IFX alone.10 Patients with CD and perianal fistulas who had an EUA prior to IFX induction had a better response (100% vs 82.6%, P=0.014), lower recurrence rate (44% vs 79%, P=0.001) and delayed time to recurrence (13.5 vs 3.6 months, P=0.0001) when compared with patients receiving IFX in isolation. Similarly, Gaertner et al demonstrated that the combination of IFX and seton placement is better than seton placement alone in CD fistulas (62% vs 43%, P=0.007).12 A recent systematic review and meta-analysis demonstrated that the combination of anti-TNF and surgical intervention is superior to anti-TNF alone.33

A randomised controlled trial exploring the optimal combination of medical and surgical therapy is currently underway in PFCD. The multimodal treatment of perianal fistulas in Crohn’s disease: seton versus anti-TNF versus advancement plasty (PISA) trial is actively recruiting with the aim to compare what the optimal surgical and medical strategy may be.34 After seton placement and sepsis drainage, patients will be randomised into three groups: chronic setons+6-MP; anti-TNF+6 MP (seton withdrawal after 6 weeks) and mucosal advancement plasty. The need for fistula-related re-interventions will be analysed after 1 year. While awaiting the results of the PISA trial, the authors advocate EUA with seton placement in addition to anti-TNF therapy.

Optimised dosing of anti-TNF

In recent years, there has been great interest in serum trough levels of drugs and its relationship on efficacy outcomes. The majority of the published literature is with IFX and in general it is accepted (depending on the assay used as a reference) that a trough level of >3 μg/mL is associated with response in patients with luminal CD.35 However, a debate arises whether patients with PFCD would benefit from higher drug levels. This remains a controversial area. Data from the Present et al study27 did not demonstrate better outcomes in patients treated with higher doses of IFX. The primary outcome of 50% reduction in fistula drainage was 56% in the 10 mg/kg dosing arm compared with 68% in the 5 mg/kg arm. Similarly, the rate of complete closure of the fistulas was also numerically lower in the 10 mg/kg group compared with the 5 mg/kg IFX group (38% vs 55%). The effects of higher doses of IFX in fistula healing were not studied in the ACCENT II trial.23

However, a recent study from Israel evaluated the influence of serum levels of IFX in healing of perianal fistulas in CD.36 A total of 36 patients were analysed. After induction (week 14) the median levels of IFX were higher in patients who responded as compared with those with non-response (4.1 vs 0.14 μg/mL, P=0.01). On multivariate analysis, the authors found that IFX drug levels of 9.25 µg/mL at week 2 and 7.25 µg/mL at week 6 were the best predictors of fistula response by weeks 14 and 30. Another study from the USA also demonstrated similar results. Yarur et al 37 analysed 117 patients with PFCD who were treated with IFX for at least 24 weeks. They once again demonstrated that patients with fistula healing had higher serum IFX levels compared with patients with active drainage (15.8 vs 4.4 µg/mL, P=0.0001). In the same study, a quartile analysis was performed demonstrating that the rate of fistula closure was greatest in the patients in the highest two quartiles (7% in the 0–2.8 µg/mL; 25% in the 2.9–10 µg/mL; 42% in the 10.1–20.1 µg/mL and 48% in the 20.1–50 µg/mL). The authors concluded that for treating perianal fistulas, levels >10 µg/mL of IFX might be needed to improve the outcome of complete healing. A recent study from the Netherlands demonstrated similar results.38 Patients under active IFX treatment with closed fistulas had higher serum levels than those with draining fistulas (6.0 vs 2.3 µg/mL, respectively; P<0.001). For the first time, the relation of serum levels of ADA and fistula closure was studied, and a similar trend was observed. The median serum concentration in patients with ADA with closed fistulas was 7.4 µg/mL, as compared with 4.8 µg/mL in those with active draining (P=0.003).

The authors believe pushing the dose of anti-TNF with a view to attain higher trough serum concentrations is beneficial. If this is related to improved mucosal healing instead of a direct action of the medication on fistula tracks, is still a matter of debate. More data from referral centres and prospective trials with different doses and different anti-TNF agents are warranted.

New biological agents and perianal fistulas in CD

As previously stated, specific prospective studies looking at the efficacy of other agents besides IFX in PFCD are lacking. Data from post hoc subgroup analyses with the newer biologic agents has been reported.


Vedolizumab (VDZ) is a humanised IgG1 monoclonal antibody that binds to the alpha4-beta7 integrin on lymhphocytes that home to the gut thus selectively modulating leucocyte trafficking to the bowel. Data from the pivotal Randomized, Placebo-Controlled, Blinded, Multicenter Study of the Induction and Maintenance of Clinical Response and Remission by Vedolizumab (MLN0002) in Patients with Moderate to Severe Crohn’s disease (GEMINI) 2 trial, a prospective randomised placebo-controlled study that analysed the efficacy of VDZ in CD included patients with draining perianal fistula.39 Although this represents a small proportion of patients (18 in the placebo, 17 in the VDZ every 8 weeks and 22 in the VDZ every 4 weeks groups), patients receiving VDZ every 8 weeks had higher rates of fistula closure than those in placebo or VDZ every 4 weeks. Observational data from the Efficacy of vedolizumab on extraintestinal manifestation in patients with inflammatory bowel disease (OBSERV-IBD) a French cohort, also analysed the efficacy of VDZ in perianal disease.40 From the 35 patients with active perianal disease at baseline (30 with fistulas and 5 with anal fissures), 15 had complete remission (42.9%) and 2 (5.7%) had partial remission at week 14. At week 54, 12 (34.3%) patients were still in complete remission. Currently, a specific study evaluating the efficacy of VDZ in perianal fistulas in CD is ongoing (A Randomized Double-Blind Phase 4 Study to Evaluate the Safety and Proportion of Subjects With Fistula Healing in 2 Dose Regimens of Entyvio (Vedolizumab IV) in the Treatment of Fistulizing Crohn’s Disease (ENTERPRISE) trial, NCT02630966).41 This prospective study will help to define the role of VDZ and other leucocyte trafficking inhibitors in the management of PFCD. At this moment, the authors would reserve VDZ as an option in those patients who have failed or become intolerant to anti-TNF therapy.


Ustekinumab (UST) is a monoclonal antibody that blocks the p40 subunit shared by both interleukins 12 and 23 (IL-12 and IL-23). This results in inhibition of the inflammatory cascade through multiple pathways and consequently reduces systemic inflammation.42 The efficacy of UST in moderate-to-severe CD was demonstrated in two pivotal induction trials, in anti-TNF refractory (A Study to Evaluate the Safety and Efficacy of Ustekinumab Induction Therapy in Patients with Moderately-to-Severely Active Crohn’s Disease (UNITI)) and anti-TNF-naïve patients (UNITI 2), as well as for maintenance of remission (IM-UNITI).43

The efficacy of UST in PFCD has recently been reported in abstract form.44 The data reported is a subgroup analysis of patients with active fistula at baseline in the phase IIb (Crohn’s Evaluation of Response to Ustekinumab Anti–Interleukin-12/23 for Induction (CERTIFI)) and phase III (UNITI) programmes. Although approximately 40% of the included patients in the UNITI 1, UNITI 2 and CERTIFI studies had a history of PFCD, a smaller proportion (10.8% and 15.5%) had active fistulas at baseline of those studies. In the CERTIFI maintenance phase, fistula response occurred in 47% (9/19) in the UST group as compared with 30% (6/20) in placebo at week 22. In patients from IM-UNITI study, fistula response was observed in 80% of patients in the combined UST group and 45.5% (5/11) in the placebo arm (P=0.64).

Several observational cohort studies have been reported as well. A small cohort from McGill University Health Centre was recently presented.45 Of a total of 62 patients treated with UST, 6 had active draining perianal fistulas. At 6 months of follow-up, 66% (4/6) had fistula response (reduction of 50% in the number of fistulas) and 33% (2/6) had fistula closure and complete healing. Wils et al published the Groupe d’Etude Thérapeutique des Affections Inflammatoires du Tube Digestif (GETAID) experience in 122 anti-TNF-exposed patients with CD treated with UST.46 Twelve patients were treated primarily for the indication of perianal fistula. A clinical benefit (defined as significant improvement in CD-related clinical symptoms and laboratory tests) was observed in 67% (8/12) of those patients. Currently, the authors consider UST a second-line therapy in patients with perianal fistula, similar to VDZ. More prospective data are needed with both agents to truly establish their position in the therapeutic algorithm. Table 2 summarises the main efficacy results of currently available biological agents in PFCD.

Table 2

Main studies with efficacy data from biological agents in PFCD

Table 3

Summary of the findings of the literature in local injections of anti-TNF agents in perianal fistulas in CD

Local injections of anti-TNF agents

The concept of local injection of therapeutic agents in areas of inflammation is not novel. In rheumatology, the local injections of corticosteroids into inflamed joints has been commonplace to improve patients’ symptoms, and recently, local injections of anti-TNF agents have been described to treat several types of arthritis.47

This approach has also been described in perianal fistulae. The use of local injections of IFX into perianal fistulas was first described by Poggioli et al. 48 The authors used variable aliquots of 15–21 mg per injection, in repeat sessions, to treat patients with PFCD. In the initial report, 10 of the 15 patients had complete healing after 3–12 sessions, with a mean follow-up of 18.2 (3–30) months. Only four patients had previous systemic IFX infusions, emphasising the potential benefit of the local action of IFX. Another series from Italy, with 11 patients, demonstrated similar results.49 Fistula response (reduction in 50% of drainage) was observed in 72.7% (8/11) and fistula closure occurred in 36.4% (4/11) of the cases. Only three patients (27.2%) did not respond to local therapy. A third Italian case series analysed the results of local IFX injections in PFCD over a longer period of time.15 Of the 12 patients reported, 8 had ongoing injections every 4–6 weeks and persistent closure was observed in 7 cases (87.5%, 95% CI 47.4 to 99.6) after 1 year, and in the remaining 5 cases there was improvement (62.5%, 95% CI 24.5 to 91.5) at the end of the median follow-up of 35 (19–43) months.

ADA local injections have also been described in PFCD. In a retrospective case series, once again Tonelli et al described the outcomes in 12 patients, in whom 20 mg of ADA diluted to 20 mL of saline was injected into fistula tracts every 2 weeks.50 The median number of injections per patient was 7 (4–16). With a mean follow-up of 17.5 (5–30) months, 75% (9/12) of the patients had absence of drainage and 25% (3/12) had improvement. All parameters of the perianal CD activity index improved when compared with baseline (P=0.002). Preliminary results of another case series demonstrated similar results, but with injections of 40 mg of ADA every other week.51 A summary of these published retrospective series is illustrated in table 3.

Despite promising results, these previously described studies have several limitations that must be considered. There is a lack of standardised technique of injection in these case series. Where should injections be made; at internal opening, within the fistula track and external opening? Or should they be performed only in the external opening and wounds? Moreover, the overall tolerability and patient satisfaction is not well described bringing into question the applicability of the technique. Therefore, several issues need to be resolved if this is to become a more common practice including the preferred dose of each agent, the ideal frequency of injections, the combined use of systemic and local therapy and finally the possibility of a patient receiving a systemic biological agent from one class and a different drug administered locally from another class.

All the preliminary uncontrolled data suggest there can be a potential therapeutic effect with local injections of anti-TNF agents, as also described in a systematic review.52 Prospective studies could guide physicians in the future to select the proper patients for this type of treatment. To date, there are no data with other non-anti-TNF biologics.

Future strategies: cell-based therapy for PFCD

Recently, a host of mesenchymal stem cell (MSC)-based therapeutic approaches for the treatment of perianal disease have been introduced with the results from a number of phase I, II and III clinical trials being reported and showing promising results.

MSCs are non-haematopoietic multipotent cells, which can be isolated from the connective tissue of most organs, most commonly from adipose tissue and bone marrow.53 The exact mechanism by which MSCs heal perianal fistulas remains unknown, but there is extensive literature articulating the immunomodulatory and immunosuppressive properties of these cells. The downregulation of the immune system comes through the inhibition of dendritic cells from monocytes, the suppression of naïve and memory CD4+ cells, prevention of T cell activation by downregulating costimulatory molecules on mature dendritic cells and promotion of regulatory T cells.53–56 Interestingly, in vitro MSCs induce dendritic cells into a regulatory-like profile57 and shift the T cell population to a more tolerogenic pattern.58 Another unique property of MSCs that allows them to be valuable players in the tissue repair process is their ability to migrate to sites of tissue injury and inflammation.59 At the site of injury, MSCs are then able to differentiate into various cell types60 and engraft into various tissues where they promote tissue repair.61–64 These characteristics highlight the reasons why MSCs are an attractive candidate for the treatment of perianal fistulas in CD with cell-based therapy.

Autologous adipose-derived MSCs

The first successful report with MSCs for the treatment of perianal CD was described by the group of Garcia-Olmo, from Madrid, Spain. A female aged 33 years with a complex branching rectovaginal fistula was successfully treated clinically by the direct injection of 9×106 autologous adipose-derived MSCs. This prompted the group to perform a phase I clinical trial to study the safety of MSCs for the treatment of PFCD.65 Four patients with complex fistulising CD (perianal, rectovaginal, enterocutaneous) were treated with a direct injection of 3×106 autologous adipose-derived MSCs into each fistula tract. Healing was observed in three of four (75%) of patients at 8 weeks, and there were no adverse events at the 12 and 22 months follow-up visits. Due to the success of this phase I trial with regard to safety and demonstrable efficacy, the group then performed a phase II trial to determine the efficacy of direct injection of autologous adipose-derived MSCs for the treatment of perianal CD-related fistula.66 Forty-nine patients with complex PFCD were randomly assigned to treatment with fibrin glue alone versus fibrin glue with 2×106 adipose tissue-derived MSCs. Patients in the MSC group received a second dose of 4×106 MSC if fistula healing was not seen at 8 weeks. Fistula healing at 1 year, as defined by clinical cessation of drainage to external palpation, occurred in 71% of the MSC group as compared with 14% of the control group.

A later phase III RCT by the same group randomised 200 patients from 19 centres to receive 20 million MSCs (64 patients), 20 million adipose-derived MSCs plus fibrin glue (60 patients) or fibrin glue alone (59 patients) following cleansing of the tract and closure of the internal opening.67 Fistula healing was defined as reepithelization of the external opening and absence of collection >2 cm by MRI. A second dose of 40 million MSCs were injected into those study group patients who did not demonstrate healing at 12 weeks. While there were no significant differences in the healing rates at 24 and 52 weeks, the treatment groups had higher rates of healing than the controls, with 1 year healing rates of 57% and 52% as compared with 37% (P=0.13). Interestingly, on an exclusive subanalysis of Garcia-Olmo’s site patients, healing rates at 24 weeks were 54% and 83% in the treatment cohorts as compared with 18% in controls. This may be due to the extensive experience of administering cells at this particular centre, which could have impacted success rates.

Outside of Spain, additional groups have reported the direct injection of autologous adipose-derived MSCs for the treatment of PFCD. A phase I trial from Korea included 10 patients with perianal CD fistulas who were treated with direct injection of autologous adipose tissue- derived MSCs at a dose of 1×107, 2×107, 4×107 cells/mL based on the size of the fistula (total of 3–40×107 cells).68 After 8 weeks, complete external epithelialization of the fistula tract was observed in 30% of patients, and these results were sustained after 8 months. The same group then performed a phase II trial to determine efficacy. Thirty-three patients were treated by direct injection of 3×107 or 6×107 (per 1 cm of fistula length) autologous-derived adipose MSCs with a primary end point of clinical healing as defined by the cessation of drainage to palpation at the external opening level. A second injection of 1.5 times the number of cells was administered if healing was not achieved at 8 weeks. At 8 weeks and 1 year, 82% and 88% had sustained closure of the fistula, respectively.69 Of the six patients who had an incomplete closure, five of them achieved a >50% closure and decreased drainage at 1 year.

Autologous bone marrow-derived MSCs

Only one trial to date has studied the use of autologous bone marrow-derived MSCs for the treatment of perianal CD. Ten patients, refractory to previous anti-TNF therapy and surgical interventions, received a direct injection with 1.5 to 3×107 MSCs every 4 weeks until an improvement was obtained or when autologous MSCs were no longer available (2–5 injections). Complete closure of fistula tracts was observed in 67% patients at 8 weeks, which was sustained to 1 year.70

Autologous adipose-derived MSCs on a scaffold

While direct injection of MSCs with or without glue is technically simple and without potential injury to the sphincter complex, there are no data to demonstrate where the cells migrate, the local concentration or how long the cells stay in the local environment. The Mayo Clinic group developed a novel protocol for the impregnation of a Gore Bio-A Fistula Plug (Gore plug; W.L. Gore and Associates, Newark, Delaware, USA) at a concentration of approximately 20 million per plug for the delivery of MSCs to the fistula tract. The tract is curetted and the plug trimmed to fit the diameter of the tract, and placed 6 weeks after the adipose tissue harvest/seton placement. A phase I trial of 12 patients has been completed, which demonstrated complete clinical and radiographic (MRI) healing in 10 out of 12 patients (83%) at 6 months.71 While this is the only trial to use a matrix, or scaffold, to deliver MSCs, the initial evidence shows potential for improved healing rates with the use of a matrix material for the theoretical prolonged local exposure of MSC.

Allogeneic adipose-derived MSCs

While autologous cells may have a theoretical benefit of lower rates of rejection of improve ‘tolerability’, allogeneic cells can be delivered without the additional intervention of adipose and bone marrow harvest and expansion. Allogeneic cells can be stocked and ready to treat several patients, and prevent unnecessary delays in treatment. There have now been phase I, II and III clinical trails investigating the efficacy of direct injection of allogeneic adipose-derived MSCs in PFCD.

The first study using allogeneic adipose-derived MSCs was a multicentre phase I/II clinical trial studying safety and efficacy. Complete healing was defined as both clinical and radiographic, with complete cessation of drainage from the external opening and healing on MRI as defined by an absence of any abscess >2 cm. In the 24 patients who received a direct injection of 2×106 MSCs, followed by a second administration of 4×106 MSCs if the fistula closure remained unsuccessful at week 12, fistula healing occurred in 38% at week 12% and 56% at week 24.72

Following this phase I/II, a multicentre phase III RCT was performed through 49 centres in Europe and Israel.18 A total of 212 patients were randomised to a single injection of 12×106 MSCs or saline, and healing was again defined as both clinical and radiographic. At 24 weeks, the MSC-treated patients had significantly increased rates of healing as compared with placebo (50% vs 34%, P=0.024). Moreover, the MSC-treated patients had a significantly shorter time to clinical remission (6.7 vs 14.6 weeks). Interestingly, there was a high rate of closure in the placebo group. This is likely related to the curettage of the tract and ligation of the internal opening.

Allogeneic bone marrow-derived MSCs

Only one trial has been performed studying bone marrow-derived allogeneic MSCs.73 A phase II trial in the Netherlands enrolled 21 patients with CD to receive a single injection of 1×107 MSCs (n=5), 3×107 MSCs (n=5) or 9×107 MSCs (n=5), or placebo (n=6). At 12 weeks, fistula healing, as defined by cessation of drainage and no fluid collection >2 cm on MRI, was seen in 47% of MSC-treated patients and 33% of placebo patients, and this was sustained up to week 24. The group noted that the dose of 3×107 had the best treatment effects, and no particular dosage was associated with any increased observed adverse events.

Benefits of cell-based therapy in PFCD

The results from the aforementioned trials are encouraging for the treatment of PFCD, both in terms of safety and efficacy, and are summarised in detail in online supplementary table 1. No trial to date has reported any serious adverse events. The most common adverse events reported are perianal pain or abscess formation at the injection site, with equivalent rates in the treatment and control groups. The efficacy might be superior to that reported with biologic therapy alone, and equivalent to some series using the LIFT19 or rectal mucosal advancement flap74 techniques for complex fistulas. However, as compared with these surgical techniques, the direct injection of MSC has no risk of injury to the internal sphincter, and, therefore, no risk of faecal incontinence—a significant advantage over these surgical options which report rates of incontinence up to 10%.74 Figure 2 illustrates in detail the results of a patient with PFCD treated with MSC-based therapy.

Supplementary data

Figure 2
Figure 2

Perianal fistula in a patient with CD after local injection of mesenchymal stem cells. (A) Day 1; (B) week 4; (C) week 24 (complete healing). Courtesy of Amy Lightner (USA).

The disparity in healing rates seen among the aforementioned clinical trials up to 88% at 1 year is likely a reflection of the way ‘healing’ is defined. In the early clinical studies and some of the phase II trials, healing was defined clinically as cessation of drainage to external palpation. These trials reported higher rates of 1 year healing up to 88%. However, in the phase III studies, fistula healing was defined as reepithelization of the external opening and absence of collection >2 cm by MRI, adding radiographic healing to clinical assessment. Reported healing rates in these studies were lower, approximately 50% at 1 year. This is likely a reflection of how fistula healing was defined. Interestingly, in the MSC plug study by Mayo Clinic, healing was again defined both by clinical and radiographic healing, and healing rates were still 83% suggesting that the matrix may increase the exposure and delivery of MSCs to the target areas of inflammation, thereby improving fistula healing.

Future studies

As we go forward in the era of cell-based therapy for PFCD, it will be important to clearly establish how to best define fistula healing, determine the optimal dosing of MSCs and determine if repeated injections of MSCs are beneficial. Perhaps the strongest recommendations for the definition of fistula healing consist in accessing both clinical and radiographic improvement. This is now widely practiced by both recently completed and ongoing clinical trials (see online supplementary table 2). Unfortunately, the dose of MSCs administered among the aforementioned trials was quite variable. Thus, we cannot yet determine optimal dosing to maximise healing rates. Similarly, the issue of redosing and the optimal timing for repeated injections remain unanswered as trials have demonstrated great variability. These will be important areas to define as we optimise treatment protocols for cell-based therapy, a promising field which will probably change the management of PFCD in a near future. Given these results and cost considerations, the authors believe that cell-based therapies may be positioned ahead of anti-TNF in the future.

The role of endoscopy in the management of perianal fistulas in CD

Medical and surgical management constitute the main modalities for PFCD; however, the role of novel endoscopic techniques is being explored.75 Diagnostic and therapeutic endoscopy for perianal CD can be performed in both outpatient or inpatient endoscopy suites or in operating rooms as a complement to examination under anaesthesia. While the techniques of endoscopic therapy are still in their infancy, it remains a very interesting area.

Endoscopic evaluation

Under proper sedation or anaesthesia, endoscopic evaluation can provide valuable information in the assessment and documentation of inflammation and ulceration in the colon, rectum or ileal pouch, anal canal and perianal skin. Endoscopy is an accurate modality for the diagnosis and measurement of bowel stricture, including those in the anal canal, distal rectum and distal pouch. With the use of proper accessory tools, such as guide wires and spray catheters, an experienced endoscopist can identify the primary and secondary openings of fistulas and openings of anastomotic sinuses.

It is important to have a roadmap when treating PFCD and the use of pelvic MRI is considered the gold standard. However, endoscopic ultrasound (EUS)76 can provide onsite adjunct assessment of perianal fistulas, sinuses, abscesses and demonstrate local lymphadenopathy. The same endoscope may be used in vaginoscopy to assess rectovaginal or pouch-vaginal fistulas.

Endoscopic fistulotomy

Fistulas in the distal bowel or perianal area which are amenable for endoscopic complete fistulotomy are mostly extrasphincteric. Endoscopic partial fistulotomy can be performed in intersphincteric fistulas, by endoscopic incision and drainage (figure 3). These fistulas at the distal bowel are often associated with surgery with disrupted suture or staple lines, for example, pouch-pouch-fistula along the staple lines or fistulas from the tip of the ‘J’ to the anastomosis. The common endoscopy tool for fistulotomies is electroincision with needle knife or isolated-tip knife. The principle of endoscopic fistulotomies is to completely open up the fistula track and place endoclips along the incised edges to prevent reformation of the fistula.

Figure 3

Endoscopic fistulotomy. Needle knife fistulotomy and abscess drainage. (A) Abscess near the secondary external opening of perianal fistula with a seton. (B) Needle knife in action. Courtesy of Bo Shen (USA).

Endoscopic drainage

Perianal abscesses, which can be an incidental finding at the time of endoscopy, can be incised and drained with endoscopy electroincision under conscious sedation and topical anaesthesia (figures 3 and 4). This may provide a definitive therapy for cryptoglandular abscesses. However, endoscopic incision and drainage only renders a temporary measure for CD fistula-associated perianal abscesses and a more definitive therapy, such as seton placement or fistulotomy, is normally needed.77

Figure 4

Endoscopic therapy of perianal fistula and abscess. Left: abscess drainage with needle knife. Centre and right: endoscopic seton placement. Courtesy of Bo Shen (USA).

Pelvic and perianal abscesses are often assessed and treated with surgical or interventional radiology approaches. However, pelvic abscesses may not be readily accessible by surgery or interventional radiology or the patient’s general condition is not feasible for more invasive approaches. For those situations, if they are related to fistulas or suture line leaks, endoscopic therapy can be attempted, by intraluminal drainage with wire-guided pigtail stents, as a paliative step towards future definitive therapy.

Perianal fistula-associated abscesses with or without underlying IBD can also be treated with endoscopy-guided seton placement.75 The seton can be placed by an endoscopist as an outpatient.77 The procedure may be particularly feasible in those with a history of prior placement of setons and recurrent formation of fistulas and abscesses, as those patients have an established track. The seton is introduced along with a guidewire, under the direct endoscopic view (figure 4). A similar approach can be attempted for the distal rectal or distal ileal pouch fistulas.

Pigtail stenting and seton placement can also be placed for the treatment of abscesses or fistulas under the guidance of imaging modalities, either on-site EUS or preprocedural CT or MRI.

Endoscopic injection therapy

A variant of agents have been administered into the tracks of perianal fistulas in patients with IBD or no IBD, which is typically performed by colorectal surgeons. These agents include fibrin glue,78 plugs,79 hyperosmolar glucose (50%), doxycycline80 and n-butyl-2-cyanoacrylate (Histoacryl).81 Those agents can be administered via endoscopic procedures, as this can be technically feasible (figure 4). However, there is lack of data regarding results with this endoscopic approach in PFCD.

Endoscopic clipping and suturing

Attempts have been made to close the primary opening of a distal bowel or perianal fistula. The principle of endoscopic clipping and suturing may have derived from surgical procedures such as mucosal, muscle flaps or sleeves. There is scarce published data on endoscopic clipping or suturing in PFCD or enterocutaneous fistulas. Through-the-scope clips and over-the-scope clips in the treatment of surgery-associated fistulas or anastomotic leaks have been used with some success, in patients with IBD and no IBD (figure 5).82–85

Figure 5
Figure 5

Endocopic clipping of enterocutaneous fistula. (A) The primary opening from mucosal side. (B) Deployment of endoclips to the orifice. Courtesy of Bo Shen (USA).

Endoscopic clipping in the treatment of PFCD has been challenging. In a small retrospective study of 10 patients with refractory anal fistulas, including 6 patients with perianal CD, the procedure was technically successful in all patients and permanent fistula closure was achieved in 7/10 (70 %) within a median time for healing of 72 days.86 Its long-term efficacy remains to be investigated.

Various endoscopic suture devices have been used to treat perforations, anastomotic leaks or fistulas in patients with no IBD.87 88 The role of endoscopic suturing in the treatment of perianal fistulas is not currently well defined.

Endoscopic closure of the primary openings of CD-related distal bowel or perianal fistulas has been difficult and has not become a part of routine clinical practice. The epithelialization of the fistula track, the presence of inflammation and fibrosis at and around the fistula opening and systemic underlying disease may be contributing factors to these difficulties. Over-the-scope clipping treatment of perianal fistulas can induce excessive discomfort and its use in vaginal fistulas, with a significant risk of iatrogenic trauma, is not recommended.

Endoscopic stricture therapy

Strictures at the anal canal or distal rectum are common findings in patients with PFCD, although the prevalence is not exactly known. While the presence of perianal fistulas in CD has been shown to hamper the healing of anorectal strictures, the impact of anorectal strictures on the healing of perianal fistulas and abscesses still remains to be better investigated.89

Anorectal strictures in CD can be classified into strictures at: (1) distal bowel; (2) anal transitional zone; (3) anus; (4) ileorectal, colorectal, ileoanal and ileal pouch-anal anastomosis. We speculate that the association between fibrostenotic and penetrating disease in CD at the distal bowel, anal and perianal areas is reciprocal. In other words, the treatment of a stricture may benefit the disease course of a fistula or an abscess, and vice versa. We recommend that symptomatic and asymptomatic strictures at the anorectal regions need to be treated. Endoscopic balloon dilation and endoscopic stricturotomy with needle knife or isolated-tip knife can be attempted.90 We feel that endoscopic stricturotomy with circumferential incision provides better outcomes than isolated balloon dilation, with better patient’s tolerance, a lower risk for perforation, sphincter damage or iatrogenic vaginal fistula and longer duration of patency.75 90 However, it can be associated with a higher risk for bleeding than balloon dilation.

There appears to be an expanding role for these endoscopic techniques on the management of PFCD. For the time being, these techniques are mostly used by endoscopists in expert centres, but they could easily be employed more broadly by endoscopists with advanced therapeutic capabilities. The exact role, the when and the why remain unanswered, but it is definitely an area that deserves further attention.

Modern surgical techniques for perianal fistulas in CD

Once active proctitis is treated, setons are placed and later withdrawn, and there is no more active inflammation in the fistula tracks, a definitive surgical repair for complex perianal fistulas in CD can be considered. There are several alternatives that can be used in this situation, and the proper choice depends on the anal anatomy, surgeon’s expertise and type of fistula.22 91 Classically, in simple superficial intersphincteric fistulas, an open fistulotomy can be used, which is associated with significant healing rates.91 However, in complex fistulas, fistulotomies can lead to faecal incontinence, and sphincter preserving techniques are formally indicated.

Chronic complex fistulas are often associated with severe anal stenosis, which can lead to the need of diverting stomas or even proctectomy. For this reason, timely fistula treatment aiming at anorectal preservation and avoidance of incontinence is key. Diverting stomas can relieve symptoms related to chronic anal sepsis and incontinence, but the rates of successful establishment of recontinuity are disappointing. Proctectomy can be still a consequence of severe PFCD in 8%–40% of patients.91

The advancement flap is probably the most used technique, with healing in approximately 50% of the patients.2 This technique can be demanding, and has better results in cryptoglandular fistulae as compared with fistulas related to CD. Recently, modern alternatives have been described to treat perianal fistulas. These techniques were originally described in fistula of cryptoglandular origin, and later studied in the management of PFCD. In this section, we will discuss five different modern surgical options for the management of PFCD: the LIFT procedure, the VAAFT technique, the fistula laser closure (FiLaC), fibrin glue injections and fistula anal plugs.

LIFT procedure

The LIFT procedure was initially described to treat cryptoglandular fistulae by the Thai, in 2007.92 93 It consists of a perianal procedure, that can be technically demanding, but has four basic steps: opening of the intersphincetric groove; dissection and isolation of the fistulous track; ligation of the track with interrupted sutures and closure of the perianal wound. The effectiveness of this procedure has been described in several case series that included patients with CD fistulas.94

A prospective study from the Los Angeles Cedars Sinai group (USA), described the outcomes of 15 consecutive patients with PFCD treated with the LIFT procedure over a period of 15 months.19 The primary outcome of the study was complete healing of the fistulas at 2 and 12 months. A preoperative seton placement aiming at inducing fibrosis of the track was performed in 87% of the cases. This strategy to place a seton for 4–6 weeks before the LIFT was done with the view to improve the healing results. Of the 15 patients, 9 (60%) were female and the median age was 34.8 years. Complete healing was observed in 60% (9/15) of the cases in a follow-up of 2 months. In the 12-month follow-up period, 67% (8/12) patients had complete healing of the LIFT, but 3 developed novel fistulas in other sites. At univariate analysis, the significant factors for success were lateral fistulas, as compared with midline fistulas (P=0.02) as well as longer fistula tracks (P=0.02).

The same group recently presented the results of a retrospective analysis of a prospective database of 23 patients with CD who underwent the LIFT procedure, over a 58-month period.95 All patients were operated on by the same surgeon. Fistula healing was observed in 48% (11/23) of patients and the median time for LIFT failure was 8 months. Patients with small bowel CD had higher rates of LIFT healing (P=0.04) and patients with colorectal CD had higher rates of LIFT failure (P=0.02).

To date, most of the experience with the LIFT procedure comes from patients with no CD, and there are limited studies published exclusively in PFCD. Several modifications of the procedure are described and different associations with advancement flaps or other techniques can be an option in CD cases. The best choice for the LIFT procedure in CD would be in patients without proctitis, with lateral long fistula tracks, previous setons and small bowel disease. Comparative prospective studies with other techniques are needed in patients with PFCD in order to better define the role of this procedure in the treatment algorithms.

VAAFT procedure

The VAAFT procedure is a sphincter-preserving technique that was initially described in non-CD fistulas in 2011.96 It is performed in two phases, diagnostic and operative. With the patient adequately positioned under spinal anaesthesia, a fistuloscope (Karl Storz, Tuttlingen, Germany) is introduced in the external opening and the fistula track is inspected under direct vision. This diagnostic phase allows the surgeon to precisely determine the location of the internal opening and the possible existence of secondary tracks, which can be often lost by palpation. Once the track is fully explored and defined, the operative phase starts with destruction of the track being performed from the external towards the internal opening. Cauterisation with a specific electrode is performed every 1 cm circumferentially, followed by cleansing of necrotic tissue. Finally, closure of the internal opening can be performed with stapling devices or with simple sutures. The advantages of this technique are basically the preservation of the sphincter and the precise direct identification of the anatomy, with proper definition of secondary tracks that can lead to recurrence, if not adequately treated.

Initial reports in patients with no CD reported healing rates of 73.5% generally 2–3 months after surgery, demonstrating the potential of this technique.96 Schwander20 demonstrated the first results of the utilisation of the VAAFT technique in PFCD. In a case series of 13 patients with CD in whom the procedure was attempted, it could be completed in 11 cases (median age 34 years, 64% females). The closure of the internal opening in this case series was performed with the addition of an advancement flap in all cases, and 4/11 had faecal diversion. With the VAAFT technique, secondary tracks not previously detected with palpation were identified in 64% (7/11) of the cases. After a mean follow-up of 9 months, complete healing of the fistulas occurred in 82% (9/11) and no changes in the Cleveland Clinic incontinence score were detected (P>0.05).

The VAAFT procedure is a promising technique with the main advantage of precise identification of the fistulous tracks, mainly related to hidden secondary tracks that cannot be identified with simple inspection and palpation under anaesthesia. A possible disadvantage is the cost of the equipment and the need for possibly long learning curves. The role of this procedure in CD fistulas still needs to be determined in comparison with other techniques, and more data are definitely needed.

Fistula laser closure

The first description of the use of laser to treat perianal fistulas was published in 2006. Moy and Bodzin97 described the efficacy of carbon dioxide laser in 27 patients with CD, with improvement in the majority of patients. Later, in 2011, Wilhelm described a new surgical technique for the management of perianal fistulas that consisted in a radial emitting laser probe (FiLaC, Biolitec, AG, Germany).98 The basic principle of this technique is to destroy the epithelium with consequent obliteration of the fistula track, in a similar way as the VAAFT procedure, but without direct visualisation of the interior of the fistula track. Another difference is that in the VAAFT procedure an electric cautery is used to obliterate the fistula tracks, whereas in the FiLaC the destruction of the track is performed by laser energy. The initial report of this technique also used the closure of the internal opening with advancement flaps, and the healing rates in this first non-CD fistula series was 81.8% (9 out of 11 patients).

The first cases with CD-related fistulas treated with this technique were described in two patients from an initial series of 33 cases of overall perianal fistulas, by Giamundo et al. 99 Later, the same authors described their experience in 45 cases, but no further CD cases were added.100

Wilhelm et al recently published the experience with the FiLaC method in 13 patients with PFCD, as a subgroup of a total of 117 patients with fistulas.101 Complete primary healing was observed in 69.2% (9/13) of the patients with CD included in the study after the operation. At the end of the study period, secondary healing (after a second operation) was observed in 92.3% (12/13) patients with CD-associated fistulas. All cases had an associated advancement flap to cover the internal opening after the laser procedure. As similar healing rates were observed between cryptoglandular and CD fistulas, the authors concluded that CD was not a predictor of futility, despite the small sample of patients analysed.

The main advantages of this laser procedure could be related to a shorter learning curve, more rapid recovery and sphincter preservation. Disadvantages such as the cost of the equipment and the absence of direct visualisation of the tracks (blind procedure) must be emphasised, as hidden secondary tracks could impair healing rates. The experience with the laser therapy in patients with CD is extremely limited. Randomised trials comparing FiLaC with other sphincter preserving techniques are required.

Fibrin glue injections

Another sphincter-preserving alternative for treating perianal fistulas in CD is the local injection of fibrin glue. Heterologous fibrin glue is constituted by a combination of two components. The first component is based on fibrinogen, factor XIII, plasminogen and aprotinin. The second component is based purely on human thrombin. When the two components are simultaneously injected (figure 6), a fibrin clot that mechanically seals the fistula tract is created. This clot gradually goes through fibrinolysis that activate tissue healing processes that can permanently obliterate the fistula tract. Higher concentrations of aprotinin are commonly used, as they can increase fibrinolysis inhibition and lead to long-lasting effects in healing.102

Figure 6

Fibrin glue injection in a patient with perianal fistulising Crohn’s disease. After cleansing of the track, injection of the glue is performed under anaesthesia. Courtesy of Professor Yves Panis (France).

Initial series of non-CD-related perianal fistulas treated with fibrin glue injections reported significant healing rates, varying from 29% to 69%. Some CD perianal fistulas were included in some of these series, with healing rates varying from 31% to 80%, but the number of patients with CD remained small, and solid conclusions could not be drawn regarding the role of fibrin glue in PFCD with these data.2

There are only two studies that aimed to verify the efficacy of fibrin glue in CD-related perianal fistulas. In 2005, Vitton et al published the results of 14 consecutive patients (mean age 42 years) treated with the fibrin glue.102 All patients had the injection of the glue guided by ultrasonography, under general anaesthesia. After 3 months of follow-up, 71% (10/14) of the patients had no drainage, 7% (1/14) had reduced drainage and 21% (3/14) had no improvement. The ultrasonographic examination demonstrated that the tract disappeared in 14% (2/14) and was not permeable in 52% (8/14) of the cases. At the end of mean follow-up period of 23.4 months (12–26), 57% (8/14) had healed tracts.

The first prospective trial with fibrin glue and PFCD was published by the GETAID group, in 2010.103 Grimaud et al performed a multicentre, open-label randomised controlled trial in patients with CD with perianal fistulas. The patients were randomised 1:1 into two groups after seton withdrawal (fibrin glue: observation). The use of anti-TNF agents was not permitted during the study. All patients underwent preoperative pelvic MRI to map the fistula tracks. The primary outcome was to evaluate clinical remission (defined as absence of draining, perianal pain or abscesses) at week 8. Clinical remission was observed in 38% (13/34) in the fibrin glue group as compared with 16% (6/37) in the observation group (OR 3.2; 95% CI 1.1 to 9.8, P=0.04). The difference in remission rates was more notable in the subgroup of simple fistulas between the two groups (50% in fibrin glue vs 18% in observation, P=0.08) as compared with complex fistulas (25% vs 11%, P=0.26). In a long-term analysis, 83% (11/13) patients in clinical remission at week 8, maintained remission at week 16. No post-treatment imaging test (MRI) was performed after randomisation.

As seen, data exclusively in PFCD with fibrin glue are scarce, even with data from a small prospective trial. There is significant heterogeneity in the surgical management preinjection of the glue as observed in most case series. Seton withdrawal and fistula track curettage might be the best approach before the glue, as there might be no role to place a fibrin clot in a epithelialized chronic tract. Association of fibrin glue injection with other methods, such as advancement flaps and systemic use of anti-TNF agents would also be interesting to explore in the future.

Anal fistula plugs

Another interesting alternative that aims at improving fistula healing in PFCD is the anal fistula plugs (AFP). The basic principle of the action of the AFP is to occlude the fistula tracks and promote healing (figure 7). There are two types of plugs that are commonly used in perianal fistulas. The bioabsorbable plug (Surgisis, Cook Surgical, Bloomington, Indiana, USA) is a xenograft made of lyophilised porcine intestinal submucosa. Its advantages include resistance to infections, absence of foreign body reaction and allowance for the repopulation of cells and tissues from the patient in a period of approximately 90 days.104 The other available device is the synthetic plug (GORE Bio-A, W.L. Gore and Associates Flagstaff, Arizona, USA), that is composed by polyglycolic acid and trimethylene carbonate, two absorbable synthetic materials moulded in form of a cone that allows anatomical adaptation to the fistula tracks and fixation to the internal openings.105

Figure 7
Figure 7

Anal fistula plug insertion and fixation in a patient with perianal fistulising Crohn’s disease. Courtesy of Professor Yves Panis (France).

Two systematic reviews were published aiming to analyse the efficacy of the AFP in the management of perianal fistulas in CD. Both outlined the scarcity of studies in the literature, as most are retrospective case series or prospective cohorts with significant heterogeneity. The first review included 20 studies and reported both CD-related and non-CD fistulas.79 Only 42 patients with were analysed, from a total of 530 included patients. The authors concluded that the overall rate of success for PFCD was 54.8% (23/42), with 95% CI 0.39 to 0.70.

The second systematic review included 12 studies and a total of 84 patients with CD-related fistulas were analysed (1–20 per study), with a mean age of 45 (18–72) years.106 The mean follow-up from this review was 9 (3–24) months. The success rate of AFP in all studies combined, defined as complete closure of the fistula tract, was 58.3% (49/84), with 95% CI 47 to 69. The overall success rate of the bioabsorbable (Surgisis) plug was 60% (48/80), with 95% CI 48 to 71 while for the synthetic plug (GORE Bio-A) was 25% (1/4), with 95% CI 1 to 81. As seen, the majority of patients were treated with the bioabsorbable AFP. Recurrence was analysed across five studies and occurred in 13.6% (13/22) of the cases.

The GETAID group presented the results of the first prospective randomised controlled trial in CD-related fistulas to test the efficacy of the bioabsorbable AFP.21 The study included 106 patients (54 in the AFP group and 52 in the control arm). All patients previously had seton placement for PFCD and did not have local sepsis. After randomisation in a 1:1 ratio, patients in the AFP group had the insertion after seton withdrawal with fixation of the plug in the internal opening with vicryl 2.0 stitches. Patients in the control group simply had the setons removed and kept current medical therapy. The primary end point was clinical remission at week 12 (defined as the absence of any drainage by all fistula openings spontaneously or after gentle finger compression, or absence of perianal abscess). Fistula closure at week 12 was achieved in 31.5% (17/54) in the AFP group (95% CI 19.5 to 45.5) and in 23.1% (12/52) in the control group (95% CI 12.5 to 36.8), without statistical difference (relative risk (RR)=1.31; 95% CI 0.59 to 4.02; P=0.1931). There was no difference in terms of the complexity of the fistulas analysed. Clinical remission occurred in 33.3% of the complex fistula and 30.8% of the simple fistulas in the AFP group as compared with 15.4% and 25.6% in controls, respectively (RR(Relative Risk)=2.17 in complex fistulas and RR=1.20 in simple fistulas; P=0.45). In summary, the authors concluded that the AFP was not superior of seton withdrawal alone, and recommend individualised therapy for patients with PFCD, including utilisation of other techniques in the armamentarium.

The initial excitement over AFP for CD-related fistulas has dissipated after this important prospective randomised study. Initial healing rates in small case series achieved 80%, but systematic reviews demonstrated lower rates for approximately 55%–60% and the only prospective trial demonstrated healing rates of <40%. It is important to emphasise that there are several variables that can justify this heterogeneity of the results among most of the studies, such as presence of active proctitis, associated medical therapy and variation of fistulas complexity. Despite this, the AFP remains an alternative in the management of complex perianal fistulas in CD, and its use may be dictated by a surgeon’s or centre’s expertise and what alternatives are available.

Final messages and future directions

The treatment of perianal fistulising CD remains one of the most difficult challenges in the treatment of patients with IBD. Two decades ago, the introduction of biologic therapy via the first successful randomised controlled trial in fistulising CD ushered in a new area of hope for physicians and patients. Despite the introduction of new therapeutic options for luminal CD, dedicated studies to PFCD have been lacking. Yet, subgroup analyses and open-label cohorts suggest that there is a role for these agents in clinical practice. The most promising area is definitely in the realm of MSCs and the next several years should accurately define their position in the management of PFCD, where the authors predict they will play a valuable role as stand-alone or part of a combination therapy approach.

An algorithm (figure 8) is proposed that encompasses the modern management of PFCD. Although some of the elements may be restricted to tertiary centres with particular expertise, the expectation is that such an algorithm will be more broadly applicable as endoscopic therapeutic techniques are adopted. Future data with new biological agents are presented and MSC injections are approved by regulatory bodies. The positioning of each therapeutic option will certainly evolve as data emerges but for the time being it will be driven by patients’ characteristics, physicians’ preference and/or experience and availability in local practice.

Figure 8

Suggested algorithm for the future management of perianal fistulising Crohn’s disease. EUA, examination under anaesthesia; IMM, immunomodulators; TNF, tumour necrosis factor; LIFT, ligation of intersphincteric fistula tract; VAAFT, video-assisted anal fistula treatment; FiLac, fistula laser closure; MSC, mesenchymal stem cell; IL, interleukin.

Perhaps the most important point in the overall management is to have a good therapeutic relationship with experienced colorectal surgeons and therapeutic endoscopists, as there are several new techniques that can be employed to improve complete fistula healing in these patients. This concept of a collaborative multidisciplinary approach using the latest medical therapies combined with modern surgical and endoscopic techniques holds promise for a brighter future for patients.


View Abstract


  • Contributors PGK, BS, AL, TY, AS, SG and RP drafted the article. RP, PGK and SG reviewed the final version and all authors gave final approval for publication.

  • Disclaimer PGK has received speaking and consultancy honoraria from AbbVie, Janssen, Pfizer, Takeda and UCB. BS has received speaking and consultancy honoraria from AbbVie, Janssen and Takeda. AL and AS have served as consultants to Takeda. SG has served as Steering Committee member to Pfizer, AbbVie, Janssen, Aerpio, Receptos, Takeda, Bristol-Myers Squibb, Novo Nordisk, Shire, Ferring and has received research support from AbbVie and speaker fee from AbbVie, Takeda, Falk Pharma and Janssen. RP has received research/educational support from AbbVie, Abbott, Ferring, Janssen, Schering-Plough, Centocor, Millennium, Elan, Procter & Gamble and Bristol-Myers Squibb. He has served as a consultant for AbbVie, Abbott, Amgen, Aptalis, AstraZeneca, Baxter, Eisai, Ferring, Janssen, Merck, Schering-Plough, Shire, Centocor, Elan, GlaxoSmithKline, UCB, Pfizer, Bristol-Myers Squibb, Warner Chilcott, Takeda, Cubist, Celgene, Gilead Sciences and Takeda. RP has also participated on speaker’s bureaus for AbbVie, AstraZeneca, Janssen, Schering- Plough, Shire, Ferring, Centocor, Elan, Prometheus, Warner Chilcott and Takeda. He has attended Advisory Boards for AbbVie, Abbott, Amgen, Aptalis, AstraZeneca, Baxter, Eisai, Ferring, Genentech, Janssen, Merck, Schering-Plough, Shire, Centocor, Elan, GlaxoSmithKline, UCB, Pfizer, Bristol-Myers Squibb, Warner Chilcott, Takeda, Cubist, Celgene and Salix.

  • Competing interests None declared.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.