Background

The estimated incidence of perianal fistulas in Europe is 1.2–2.8 per 10,000 people [1,2,3]. Perianal fistulas appear in 30–50% of Crohn’s disease (CD) cases, and 80% of those fistulas are classified as complex [4, 5]. In this scenario, medical treatments are intended to promote long-term fistula healing, while preserving continence and avoiding diverting stomas [6]. However, these goals are often unmet with currently available therapies, particularly in relation to complex perianal fistulas, which are the most challenging to treat [7]. Treatment of complex perianal fistulas (CPF) in patients with CD is especially challenging for surgeons and gastroenterologists. Medical therapy is typically still recommended as a first line treatment, with surgery being reserved for sepsis control or laying open superficial tracks [8, 9].

During the last 30 years, many “classic” surgical techniques used to treat cryptoglandular complex perianal fistulas, such as core out, advancement flaps, or ligation of intersphincteric fistula tracks (LIFTs), have also been used to resolve CPF associated with CD [10]. However, during the last 10 years, a shift has occurred to a new and usually minimally invasive surgery (MIS), with support from biological approaches, such as stem cells, platelet rich plasma, or the use of fibrin or glue, to avoid touching the sphincter, hence preserving fecal continence.

The aim of this study was to conduct a systematic review and audit of the results of complex perianal fistula surgeries for patients with CD or cryptoglandular fistulas. We performed this review to revisit the concept of perianal complex fistula treatments and answer questions concerning the clinical evidence that has emerged regarding various surgical approaches to CD and cryptoglandular complex perianal fistulas, in terms of fistula healing, Health-Related Quality of Life (HRQoL), cost, and fecal continence.

Methods

Review design

The protocol (stored in PORIB) and reporting methodology for this systematic review was designed in accordance with the PRISMA-P guidelines [11]. The participants, interventions, comparisons, outcomes, and study design (PICO) strategy was followed to identify the populations (CPF in patients with CD or fistulas of cryptoglandular origin), intervention (surgery), comparisons (clinical trials or observational studies), and outcomes (clinical, economic, and quality of life) (Additional file 1).

Eligibility criteria

We included all the primary studies published in the medical literature related to the clinical outcomes, quality of life, or economic costs of complex perianal fistula surgeries for patients with CD or cryptoglandular-associated fistulas not related to CD. The eligible studies included randomized controlled trials (RCTs) and observational studies that referred to humans, were written in English, included adults 18+ years old, and were published in the 10-year period from 2/01/2010 to 2/29/2020. We excluded articles regarding a different population (Reason 1), type of fistula such as simple fistulas, internal fistulas, such as rectovaginal, anovaginal, rectourethral, or ileovaginal fistulas (Reason 2), non-surgical interventions, such as pharmacological or medical treatments (Reason 3), type of study such as reviews (Reason 4) and publications without clinical outcome, quality of life, or cost results (Reason 5).

Data sources and search strategy

In March 2020, a literature search strategy was designed using variations on the search terms “Crohn’s disease”, “Rectal fistula”, “Perianal”, “Fistulizing disease”, “Complex”, “Inflammatory Bowel Disease”, “Cryptoglandular”, “Surgical intervention”, and “Surgical procedures”. The following databases were searched: PubMed, EMBASE, the Database of Abstracts of Reviews of Effectiveness (DARE), and the Cochrane Central Register of Controlled Trials (CENTRAL). We also searched the 2018–2020 abstract books for the European Society of Coloproctology (ESCP), European Crohn’s and Colitis Organisation (ECCO), United European Gastroenterology (UEG) Week, American Gastroenterological Association (AGA), and the American Society of Colon & Rectal Surgeons. The search strategy is shown in Additional file 2.

Study selection and data extraction

Two reviewers independently reviewed each citation against our eligibility criteria in the following 2-stage process: (1) title and abstract and (2) full text. During the study selection and data extraction stages, disagreements between the reviewers were resolved through discussion or with a decision made by a third researcher.

Quality assessment

Each article was assigned one of the following quality of study scores, as designated by the Scottish Intercollegiate Guidelines Network (SIGN) [12], according to the level of evidence provided in the paper:

  • 1++: High quality meta-analyses, systematic reviews of RCTs, or RCTs with an extremely low risk of bias

  • 1+: Well-conducted meta-analyses, systematic reviews, or RCTs with a low risk of bias Meta-analyses, systematic reviews, or RCTs with a high risk of bias

  • 2++: High quality systematic reviews of case control or cohort studies.

  • High quality case control or cohort studies with a very low risk of confounding factors or bias and a high probability that the relationship is causal.

  • 2+: Well-conducted case control or cohort studies with a low risk of confounding factors or bias and a moderate probability that the relationship is causal.

  • Case control or cohort studies with a high risk of confounding factors or bias and a significant risk that the relationship is not causal.

  • 3: Non-analytic studies, e.g. case reports, case series.

  • 4: Expert opinions.

We extracted the intervention, study design, number of patients, and main conclusions from each selected article and added the 2019 impact factor from the Journal Citation Reports (JCR) to provide more information for the qualitative analysis.

Results

We identified 577 citations from PubMed, EMBASE, DARE, and CENTRAL, and 32 records from other sources (Fig. 1). After removing duplicate papers (75 papers), the titles and abstracts of 502 citations were screened, from which 342 were excluded. Subsequently, the full text of 160 citations was retrieved and assessed for inclusion. Eighty studies were finally retained. No additional studies were identified through tracking citations or by manually reviewing the references of included studies.

Fig. 1
figure 1

Flow chart of the article selection process

Full size image

We split the results into two categories, i.e., CD-associated CPF and CPF of cryptoglandular origin. We also differentiated between MIS and classic surgical techniques. The clinical results are shown in Tables 1 and 2.

Table 1 Results of the Crohn’s disease analysis
Full size table
Table 2 Results of the complex perianal fistulas of cryptoglandular origin analysis
Full size table

We found that the concepts of healing and fecal incontinence were defined differently in the various studies, making quantitative aggregation difficult; therefore, these issues must be considered when interpreting our results. Nevertheless, we were still able to obtain relevant information from the review.

Clinical outcomes

Of the 18 articles with clinical results that referred to CD, 4 reported level 1 evidence, one reported level 2 evidence, and 10 reported level 3 evidence. A total of 15 articles referred to MIS techniques, and three referred to classical techniques. Studies referring to the most current techniques were more common than those referring to traditional techniques, the evidence levels presented for the current techniques were higher, and the results were more favorable. Two studies employing treatments using derivatives of adipose tissue reported healing rates exceeding 70% [15, 25], and a significantly greater proportion of patients stem cells-treated achieved combined remission versus controls (56.3% vs 38.6%, p = 0.010) in a high-level evidence study [14]. In a study investigating a treatment using mesenchymal cells, in which healing was defined as fistula absence or less than 2 cm discharge on magnetic resonance imaging, the authors reported a healing rate of 80% at week 12 [20]. In a pilot study with 10 patients undergoing fistulectomies with platelet-rich plasma, one (10%) patient experienced a recurrence, and two (20%) patients had persistent fistulas after treatment. In two studies examining classical techniques (primarily seton, LIFT, or lay open), the healing rates were approximately 50–60% [28, 29]. Graf et al. observed that 62 (52%) patients achieved healing (absence of fistula symptoms, skin healing, and no evidence of a fistula on clinical examination) by the end of the follow-up period, but only 14 of the patients had healed after a single procedure, while the remaining 48 healed after a median of 4.0 (2–20) additional procedures [28].

Of the 52 references referring to clinical results for patients with fistulas of cryptoglandular origin, three of the articles reported level 1 evidence, six reported level 2 evidence, and 43 reported level 3 evidence. A total of 28 articles referred to MIS (25 articles reported the use of plugs), and 24 articles referred to classic techniques (six articles reported the use of seton, and four studies incorporated flaps). The MIS studies reported healing rates between 50% and 90%; a healing rate of 70% was reported for a study using derivatives of adipose tissue [42], an 80% healing rate was achieved using laser technology [51], a 70% healing rate was achieved using platelets [31], and healing rates ​between 50% and 90% were achieved using various plugs [34, 38, 41]. Additionally, each of these studies presented results for various other aspects of complex perianal fistulas, such as pain, quality of life, or continence; continence was most frequently reported in these studies, and improved results were typical. Decreases in Wexner scores after the use of autologous platelet growth factors [36] and Nitinol Clips [38] were also reported. No fecal incontinence was reported after procedures performed with over-the-scope-clips [41] or stem cells derived from autologous adipose tissue [42]. Recurrence and retreatment occurred in 2/10 cases [44] and 20/25 cases [45] using two kinds of plugs.

For some studies that incorporated classic techniques, the healing percentages were similar to those observed in the MIS studies (70–80% for setons [60] or flaps [61] and somewhat lower percentages for other techniques, such as a fistulotomies or fistulectomies [23, 40]); however, the overall function in these patients was lower, as more cases of incontinence were reported, either because the incontinence was not corrected or it appeared de novo. In a study investigating fistulotomies, only 26.3% of the patients had a perfect continence state, with a Vaizey score equal to 0 [74]. The recurrence or retreatment rates in these studies varied from 5.9 to 50% [62, 66, 67].

HRQoL outcomes

Four HRQoL studies in this analysis were performed for CD, but only one of them showed a relationship between the results and the surgical technique used. In a post-hoc analysis of the ADMIRE-CD clinical trial, Panes et al. [83] observed that patients who experienced clinical or combined remission had lower (Perianal Disease Activity Index) PDAI scores for pain and discharge than those who did not experience remission. The scores were fourfold higher for patients who experienced clinical or combined remission in combination with magnetic resonance imaging. The scores were fourfold higher for patients who experienced clinical or combined remission in combination with magnetic resonance imaging. Other HRQoL studies referred to abdominal surgery [84] or perianal disease [85, 86] in general, so they did not refer to surgery.

Three studies with HRQoL outcomes for patients with cryptoglandular CPF were included in this review. Jayne et al. [87] compared the efficacy of the Surgisis anal fistula plug with various other techniques in a prospective, multicenter, randomized, unblinded, parallel arm clinical trial. A total of 304 patients were included in their study, and the authors observed no differences in the clinical healing rates (55%, 64%, 75%, 53%, and 42% for the fistula plug, seton cut, fistulotomy, advancement flap, and LIFT procedure, respectively) at 12 months. The baseline fecal incontinence rates were lower for the groups with little improvement after treatment. The mean total costs were £2738 (± £1.151) for the fistula plug group and £2308 (± £1.228) for the surgeon’s preference group. The Quality Adjusted Life Years (QALYs) were higher for the fistula plug group (0.829 ± 0.174) than for the surgeon’s preference group (0.790 ± 0.212), which establishes that there is a 35–45% chance that the fistula plug is as profitable as the surgeon’s preference for an availability to pay range of £20,000–30,000/QALY.

In a prospective study of 34 patients undergoing surgical treatment, Jayarajah [88] reported overall preoperative and postoperative incontinence rates of 18% and 38%, respectively. The total mean Fecal Incontinence Quality of Life (FIQL) score was 16.0 (Standard Deviation, SD ± 0.4) preoperatively and 16.1 (SD ± 0.4) postoperatively. The authors also observed a considerable difference in the scale that measures “depression/self-perception” before and after the intervention (p = 0.012). In a retrospective cross-sectional study, Visscher et al. [89] found that by the end of the follow-up period (mean follow-up of 7.8 years) 39/141 patients (34%) who underwent unspecified surgical procedures after an initial perianal fistula surgery still experienced incontinence. Surgical fistulotomies, drainage of multiple abscesses, and high transsphincteric or suprasphincteric abscesses were associated with incontinence to a significant degree. Incontinence was worse for patients who had surgery for CPF (Wexner score, 4.7 ± 6.2) than for those who had surgery for simple fistulas (Wexner score, 1.2 ± 2.1) (p = 0.001). Surgery for CPF was also associated with worse quality of life outcomes, including lifestyle (p = 0.030), depression (p = 0.077), and shame (p < 0.001).

Cost outcomes

Of the articles included in this review, only the Jayne et al. article [87] associated a technique with its economic cost. In this study, a mean total cost was associated with the group who underwent treatment with fistula plugs (£2738 ± £1151) and the rest of the treatments (£2308 ± £1228). Additionally, a QALY gain of 0.829 ± 0.174 was calculated for the group with fistula plugs compared to 0.790 ± 0.212 for the surgeon’s preference group (0.790 ± 0.212). The probabilistic incremental cost-effectiveness results were £10,993 (± £478,666), with a 35–45% chance that the fistula plug is as profitable as the surgeon’s preference for an availability-to-pay range of £20,000–30,000/QALY. Three studies showed the cost of CD, but none showed the cost of any surgical techniques [90,91,92].

Discussion

We conducted a comprehensive systematic literature review to audit the clinical outcomes resulting from surgical treatment of complex perianal fistulas. Various databases and sources of information were analyzed to determine the similarities between surgical treatments for CD-associated fistulas and cryptoglandular complex perianal fistulas. Studies by Graf [28], Gingold [29], and Galis-Rozen [30] showed that classic techniques, such as lay open, LIFT, fistulectomies, and flaps, are still used to treat CD, despite the potential risk of exposure to the anal sphincter. The communicated healing percentage for CD was approximately 50% usually after the repeated procedures [28, 30], and the percentage of patients with incontinence rose to 60% (9/15) by the end of the follow-up period [29]. These results were surpassed decisively by those for MIS, a difference of 17% in remission rates was reported in stem cells-treated patients in a clinical trial [14] and healing rates close to 80%, with a reduction in the PDAI and improvements to the HRQoL in some patient series’ [13]. In a study with autologous adipose-derived stromal vascular fraction a significant reduction in the severity of perianal disease was shown, with PDAI reduction from 7.3 to 3.4 in week 48 (p = 0.045). According to Tozer et al. [93] we hypothesize that the inflammatory origin present in the CD fistula prevents its complete healing with conventional techniques. Therefore, the use of cellular mechanisms with anti-inflammatory potential may have a favorable result in healing and maintaining sphincter function beyond the use of a single conventional technique. However, these conclusions must be ratified in subsequent studies, since the studies published up to the date of the review sometimes corresponded to series with few patients and a short follow-up period.

In CPF results on healing rate were similar between MIS and classic techniques. In MIS we found that in a study with OTSC device® (over-the-scope-clip) [34] there was no appearance of fecal incontinence, an improvement of Wexner score with the use of autologous platelet growth factors, from 3.0625 to 1.125 in a year, p = 0.0195 [35]; no deterioration of continence was observed with Nitinol Clip [38] as a result of the last 5 years. In classic techniques we found a study with a slight continence deficit in patients treated with Mucosal advancement flap [61], no improvement in continence was reported by Balciscueta et al. [62], and cases of fecal incontinence were detected as complications in 8 patients treated with seton [64].

We observed that during the 10-year period from 02/01/2010 to 02/29/2020, a shift occurred in the treatment of cryptoglandular CPF from classic surgical techniques to a MIS/biological approach; this shift has allowed better facilitation of the healing and preservation of sphincter function. Of the 18 papers that referred to CD, 13 reported on investigations regarding MIS, and four reported on classic techniques. We found that there was not a significant difference between the number of articles for each of the procedures, as there were 27 articles referring to MIS and 22 referring to classic techniques. Therefore, we suggest that a paradigm shift is beginning to occur, making MIS a first order treatment for CPF of cryptoglandular origin.

Notably, procedures like LIFT are recommended in some guidelines as a first option for patients with CD [94], although this is not supported by substantial evidence, as we found only one reference that advocated for this recommendation [29]. Therefore, we suggest that these recommendations should be reevaluated. Given what has been published, we believe that it is safer not to divide any part of the anal sphincter when treating perianal CD.

A critical point in the topic we are dealing with is the use of new pharmacological treatments such as anti-TNF for the management of patients with CD fistula. Although the objective of our study was to analyze the surgical techniques and pharmacological treatments that were excluded, we believe that it is appropriate to highlight this aspect. Treatment with infliximab indeed has good results in these patients, but it is also true that after 1 year of treatment the response rate can fall to 23% [95]. These results have recently been improved with the use of mesenchymal cells with annual response rates of 59% [96].

Nevertheless, these initial observations must be analyzed with consideration for the following limitations: First, our preliminary hypothesis was that the level of evidence from the selected studies could be low, and this would create a high risk of bias. The selected articles have confirmed this hypothesis, especially in the case of the articles about classic procedures. This made it difficult to carry out a more rigorous comparison of the results, such as meta-analysis or network meta-analysis. The different ways to present healing and continence results also made it difficult to aggregate the results.

We determined that there are no published studies that have specifically investigated the relationship between surgical techniques and quality of life. Therefore, we suggest that studies capable of determining the impact of the various surgical procedures from the patient’s perspective should be designed, as the HRQoL is only a secondary or tertiary variable in currently published studies. Only 6 of the 80 (7.5%) total references (Serrero, both in 2017 [15] and 2019 [13], Gingold [29], El-Said [59], Gottgens [74], and Herreros [48]) reported quality of life results. Consequently, we encourage the development of robust quality of life and cost-effectiveness studies, as both these variables are factored into our conclusions.

In conclusion, our review shows that patients with CD experience a higher rate of healing after MIS techniques than patients who undergo classic surgical techniques, and the healing rate for complex anal fistulas with cryptoglandular origins appears similar between classic and minimally invasive techniques. Additionally, the incontinence rate for patients that undergo minimally invasive surgical techniques is better than that of patients who undergo classic techniques. Therefore, we recommend moving to MIS-based techniques, in conjunction with new biological technologies like stem cells, plugs or Adipose-Derived Stromal Vascular Fraction use, because these techniques seem to be supported by recently published clinical evidence.