Does Combined Medical and Surgical Treatment Improve Perianal Fistula Outcomes in Patients With Crohn’s Disease? A Systematic Review and Meta-Analysis

Abstract Background The optimal treatment of perianal fistulizing Crohn’s disease [PFCD] is unknown. We performed a systematic review with meta-analysis to compare combined surgical intervention and anti-tumour necrosis factor [anti-TNF] therapy [combined therapy] vs either therapy alone. Methods MEDLINE, EMBASE, and Cochrane databases were searched systematically up to end December 2023. Surgical intervention was defined as an exam under anaesthesia ± setons. We calculated weighted risk ratios [RRs] with 95% confidence intervals [CIs] for our co-primary outcomes: fistula response and healing, defined clinically as a reduction in fistula drainage or number of draining fistulas and fistula closure respectively. Results Thirteen studies were analysed: 515 patients treated with combined therapy, 330 patients with surgical intervention, and 406 patients with anti-TNF therapy with follow-up between 10 weeks and 3 years. Fistula response [RR 1.10; 95% CI 0.93–1.30, p = 0.28] and healing [RR 1.06; 95% CI 0.86–1.31, p = 0.58] was not significantly different when comparing combined therapy with anti-TNF therapy alone. In contrast, combined therapy was associated with significantly higher rates of fistula response [RR 1.25; 95% CI 1.10–1.41, p < 0.001] and healing [RR 1.17; 95% CI 1.00–1.36, p = 0.05] compared with surgical intervention alone. Our results remained stable when limiting to studies that assessed outcomes within 1 year and studies where <10% of patients underwent fistula closure procedures. Conclusion Combined surgery and anti-TNF therapy was not associated with improved PFCD outcomes compared with anti-TNF therapy alone. Due to an inability to control for confounding and small study sizes, future, controlled trials are warranted to confirm these findings.


Introduction
Perianal fistulizing Crohn's disease [PFCD] is one of the most challenging phenotypes of Crohn's disease [CD] to manage.They occur in one in five patients with CD 1,2 and often follow a relapsing-remitting pattern consisting of frequent episodes of fistula drainage, perianal pain, and abscess formation. 3,4As a result, they have a far-reaching impact on the functional, psychosocial, and emotional health of patients living with CD [5][6][7][8][9] and on healthcare costs. 10espite two seminal clinical trials demonstrating the effectiveness of infliximab, an anti-tumour necrosis factor [anti-TNF] drug, for the induction and maintenance therapy of PFCD, 11,12 a substantial proportion of patients are unable to achieve sustained fistula remission. 12,13As a result, there is considerable interest in strategies to enhance the effectiveness of anti-TNF therapy. 14Current guidelines recommend a multidisciplinary approach involving gastroenterologists and surgeons in order to optimize outcomes.This entails an examination under anaesthesia [EUA] with seton[s] to facilitate tract drainage and to prevent abscesses from reforming followed by anti-TNF therapy.6][17] Furthermore, recent studies have raised doubts as to the beneficial impact of setons on long-term PFCD outcomes. 18,19xisting guidelines on the management of PFCD are based on observational data and expert opinion.A systematic review by Yassin et al. in 2014  20 demonstrated that combined surgical intervention with anti-TNF therapy resulted in higher rates of fistula healing than either therapy alone.However, their study did not separate surgical intervention from anti-TNF therapy in the cohort who received single modality therapy.Therefore, it is unclear if surgical intervention or anti-TNF therapy is the major driver of fistula healing in patients treated with combined modality therapy.][23][24] The aim of our current study is to compare the effectiveness of combined surgical intervention and anti-TNF therapy vs either modality alone on PFCD.

Study design
We performed a systematic review according to an a priori established protocol and guidance from the Cochrane handbook. 25Our study question was designed using the PICO construct: the study population involved patients with PFCD, the intervention was combined surgical and anti-TNF therapy, the comparison was each modality alone, and the outcome was fistula response and remission.Our study was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis [PRISMA] statement guidelines. 26

Study eligibility
We included randomized controlled trials, and observational studies [comparative cohort studies] that reported PFCD outcomes and compared combined surgical intervention and anti-TNF therapy with either modality alone.We excluded studies that contained patients with diverting stomas, ileal pouch anal anastomosis, and perianal fistulas related to other aetiologies [e.g.cryptoglandular fistulas].We also excluded studies that did not stratify outcomes by the type of single modality therapy [surgical intervention or anti-TNF therapy] and studies with insufficient data.We did not restrict studies based on patient age, publication year, publication language, or publication type [full manuscripts or conference proceedings].Studies were imported into COVIDENCE systemic review software [Veritas Health Innovation], a web-based collaboration software platform designed for study screening and eligibility in systematic reviews.Two reviewers [MF, YF and/ or HK] independently evaluated each of the references in duplicate to determine study eligibility.Studies were initially screened by title/abstract with use of pre-established screening questions and those that were identified as potential studies underwent full manuscript retrieval.Disagreements were resolved by consensus [MF and YF, when required by consultation with the senior investigator JM].Data were extracted in duplicate by two independent reviewers using a pre-designed data extraction template.

Interventions
Combined modality therapy was defined as local surgical intervention by EUA with or without placement of seton[s] and anti-TNF therapy.Anti-TNF therapy was defined as induction and maintenance treatment with infliximab, adalimumab, certolizumab pegol, or golimumab.

Clinical outcomes and quality assessment
Our co-primary outcomes were fistula response and healing.The definitions for fistula response varied between studies and included: minimal fistula symptoms, reduction in the number of draining fistulas, reduction in fistula drainage, or reduction in fistula discomfort.Fistula healing was defined in most studies as the complete closure of fistula tracts clinically and when performed an absence of drainage by the finger compression test.One study included in their definition of fistula healing the lack of complaints by the patient. 27In one study, fistula response was the only reported outcome measure and was included in our overall analysis of fistula healing. 28Faecal diversion was assessed as a secondary outcome and was defined as diverting ileostomy/colostomy or proctectomy.

Study quality
Study quality was determined using the Good Research for Comparative Effectiveness [GRACE] Checklist, a consensusbased assessment tool for non-interventional studies of comparative effectiveness. 29The checklist contains 11 items: six items pertaining to the reporting of data and five items pertaining to the study methods [Supplementary Table 2b].The data reporting items evaluated the description of treatment exposure, primary outcome recording, objectivity, validity, and equivalency between comparison groups.The methods items evaluated population restrictions, concurrent comparators, confounding/ effect-modifying variables, 'immortal time bias', and meaningful analyses.Each item was graded as sufficient or insufficient.

Statistical analysis
Statistical analysis was performed using the Cochrane Review Manager [RevMan, v.5.4.1,The Cochrane Collaboration].Weighted risk ratios [RRs] with 95% confidence intervals [CIs]  for fistula outcomes were estimated by fixed effects models, the preferred method for meta-analysis when fewer than five studies are available and when there is minimal heterogeneity between studies. 30For analysis involving five or more studies or when significant heterogeneity was present defined as an inconsistency index [I 2 ] >50%, we used random effects models. 25We performed separate analysis comparing combined modality therapy vs either modality alone [anti-TNF therapy or surgical intervention].To ensure stability of our results, we performed multiple pre-planned sensitivity analysis where we limited our analysis to: [1] studies that used fistula healing as an outcome measure, [2] studies that reported fistula outcomes within 1 year of treatment, and [3] studies where <10% of patients were treated with surgical closure as part of the EUA procedure.

Interventions
There were a total of 1251 patients: 515 patients treated by combined surgical intervention and anti-TNF therapy, 330 patients treated by surgical intervention alone, and 406 patients treated by anti-TNF therapy alone.Among patients who underwent an EUA, 575 [69%] patients had seton[s] placed, 173 [21%] patients underwent a fistulotomy, and 92 [11%] patients underwent surgical closure [Table 2].There were minor differences noted in the types of surgical interventions between the combined modality therapy group and the surgical intervention alone group: setons (82%
Two studies compared fistula response between patients treated with combined modality therapy [n = 23 patients] vs anti-TNF therapy alone [n = 34 patients]. 28,33Fistula response occurred in 22 patients [96%] treated with combined modality therapy, compared with 29 patients [85%] treated with anti-TNF therapy alone.There was no significant difference in rates of fistula response in patients treated with combined modality therapy compared with anti-TNF therapy alone [RR 1.10; 95% CI 0.93-1.30,p = 0.28] [Supplementary Figure 1].Both of these studies had endpoints within 1 year and had <10% of patients treated with surgical closure, and thus further sensitivity analysis was not performed.
Five studies compared the rates of facal diversion between combined modality therapy [n = 83 patients] and anti-TNF therapy alone [n = 152 patients] 19,31,33,35,38 ; however, only two of these studies delineated which treatment modalities these patients received 19,28 [Table 4].Faecal diversion occurred in 11 patients [13%] treated with combined modality therapy and in 25 patients [16%] treated with anti-TNF therapy alone.There was no significant difference in the rates of faecal diversion in patients treated with combined modality therapy compared with anti-TNF therapy alone [RR 0.85; 95% CI 0.46-1.58,p = 0.61] [Supplementary Figure 2].

Combined modality therapy vs surgical intervention alone
36,37 Fistula healing occurred in 139 patients [56%] treated with combined modality therapy, compared with 160 patients [49%] treated with surgical intervention alone.Fistula healing was significantly more likely in patients treated with combined modality therapy compared with surgical intervention alone [RR 1.17; 95% CI 1.00-1.36,p = 0.05] [Figure 3].Our results remained stable in sensitivity analyses when limiting to studies with endpoints within 1 year [RR 1.33; 95% CI 1.07-1.65,p = 0.01] 27,33,34,37 and when limiting to studies with less than 10% of patients treated with surgical closure [RR 1.42, 95% CI 1.09-1.84,p = 0.01]. 33,34,37hree studies compared fistula response between patients treated with combined modality therapy [n = 194 patients] vs surgical intervention alone [n = 274 patients]. 32,33,36Fistula response occurred in 149 patients [77%] treated with combined modality therapy, compared with 177 patients [65%] treated with surgical intervention alone.Fistula response was significantly more likely in patients treated with combined modality therapy compared with surgical intervention alone [RR 1.25; 95% CI 1.10-1.41,p = 0.0004] [Supplementary Figure 3].Further sensitivity analyses were not performed as only one of these three studies reported fistula response outcomes within 1 year 33 and only one of these three studies had <10% of patients treated with surgical closure. 334]36 Faecal diversion occurred in 51 patients [23%] treated with combined modality therapy and in 71 patients [23%] treated with surgical intervention alone [Table 5].There was no significant difference in the rates of faecal diversion in patients treated with combined modality therapy compared with surgical intervention alone [RR 0.97; 95% CI 0.71-1.33,p = 0.86] [Supplementary Figure 4].

Study quality
Study quality ranged from 5 to 8 out of a total of 11 points using the GRACE quality assessment tool [Supplementary Table 2a].The majority of studies lost points for use of nonvalidated primary outcomes measures, lack of reporting or control of confounders, potential for 'immortal time bias', and lack of subsequent analysis to test key assumptions.

Discussion
In this systematic review and meta-analysis consisting of 13 studies with 1251 patients, there were no significant differences between combined surgical and anti-TNF therapy compared with anti-TNF therapy alone for PFCD response and healing.In contrast, combined surgical and anti-TNF therapy was more effective in achieving PFCD response and healing compared with surgical intervention alone.Our results remained consistent with multiple sensitivity analysis where we limited the analysis to studies where PFCD outcomes were assessed within 1 year and to studies where <10% of patients underwent surgical closure at the time of EUA.The majority of studies included in our analysis were retrospective, did not control for confounding, and had have low scores on our quality assessment, underscoring the importance of future, randomized controlled trials to determine the true benefit of combined modality therapy for PFCD in patients with CD.
A previous systematic review with meta-analysis consisting of 797 patients found that complete fistula remission occurred more frequently in patients treated with combined modality therapy [180/349 patients; 52%] when compared with either anti-TNF therapy or surgical intervention alone [191/448   patients; 43%].Similarly, in a retrospective multicentre study from Europe and Israel, medical therapy combined with surgical drainage resulted in higher rates of complete fistula healing [52% vs 42%; p = 0.04] and lower rates of repeat surgical intervention [25% vs 59%; p = 0.001] compared with either therapy alone. 22However, neither of these studies separated anti-TNF therapy from surgical intervention in the single modality cohort.Therefore, the results of our work add to these studies by suggesting that anti-TNF therapy may be the main driver of fistula healing in patients treated with combined modality therapy.
There are number of potential explanations for why combined therapy was not associated with better fistula outcomes compared with anti-TNF therapy alone.It is possible that patients who underwent combined modality therapy had more severe disease and as a result were less likely to achieve fistula response or healing.Alternatively, it is possible that anti-TNF therapy is the major driver of fistula healing, and that local surgical drainage along with seton placement does not provide additional benefit, particularly in the absence of drainable abscesses.It is also unclear what percentage of EUA involved fistula tract curettage, a technique aimed to improve PFCD outcomes by de-epithelializing the tracts. 39Finally, it is also possible that any beneficial effect from surgical intervention may have been lost from delayed removal of setons, since setons by design prevent fistula closure.In keeping with this hypothesis, Gaertner et al. recently demonstrated that delayed removal of setons was associated with lower rates of fistula closure. 32t is important to recognize that few patients in our metaanalysis underwent fistula closure procedures such as advancement flaps or ligation of intersphincteric fistula [LIFT] procedures.PISA-II, a landmark patient preference randomized controlled trial, recently demonstrated that anti-TNF therapy combined with surgical closure improved radiological healing of PFCD compared with anti-TNF therapy. 23hile this study was not included in our meta-analysis since all patients in both treatment arms underwent an EUA and seton insertion, it demonstrates that surgical closure may be a potential strategy performed at the time of EUA to improve PFCD outcomes.However, it is important to recognize that fistula closure can only be performed in selected patients with optimal outcomes occurring when rectal inflammation and anal strictures are absent. 4013]41 This was also supported by PISA-I, which demonstrated that EUA with placement of setons alone resulted in higher rates of surgical reintervention [10/15 patients] compared with combined surgical and anti-TNF therapy [6/15 patients]. 18This study was not included in our analysis since it did not report fistula response or remission as an outcome.There can be considerable variability in the types of interventions performed during an EUA.Fewer patients in the combined therapy group underwent fistulotomy compared to surgical intervention alone [9% vs 38%], suggesting that the combined therapy group may have had more complex disease.Therefore, our results may have underestimated the effectiveness of anti-TNF therapy.In contrast, we do not believe that differences in surgical closure procedures between the cohorts impacted our results, since there were similar rates of surgical closure procedures [9% vs 14%] and our results remained stable in a sensitivity analysis where we limited our analysis to studies where <10% of patients underwent surgical closure.This is the most comprehensive meta-analysis to date evaluating the impact of combined surgical intervention and anti-TNF therapy compared with either modality alone.The major strengths of our study were our stratified analysis by the type of single modality therapy and the consistency of our results in multiple sensitivity analyses.Our stratified analysis by type of single modality therapy allowed for assessment of the individual effect of surgical intervention and anti-TNF therapy separately in relation to combined modality therapy.
Our study contains a number of important limitations.First, only one study in our analysis was a randomized   controlled trial and, among the observational studies, few controlled for confounding such as fistula complexity, concurrent luminal disease, smoking, and anal stenosis.As a result, there may have been bias by indication.Second, the majority of studies in our analysis contained small sample sizes which may impact the precision of our point estimates and limit the generalizability of our findings.Third, standardized protocols for EUA were not used in any of the studies and none reported the proportion of procedures that performed curettage of fistula tracts.Although we performed a sensitivity analysis limiting to studies where <10% of patients underwent surgical closure to help mitigate some of this bias, it remains possible that additional differences in surgical technique existed between study cohorts.Finally, the majority of studies used clinical healing of fistula tracts as their outcome measure without radiological confirmation.However, to date, no widely accepted definition of radiological healing of fistula tracts exists.Notwithstanding these limitations, there are a number of conclusions that can be drawn from our study.First, surgical therapy should not be used as the sole therapy for complex PFCD.This is consistent with the results of PISA-I and in keeping with societal guidelines. 40,42,43Second, exams under anaesthesia ± setons may not be universally required prior to initiating anti-TNF therapy for PFCD.However, in situations where abscesses are present, we still believe that source control, with a dedicated exam under anaesthesia, remains a critical step prior to initiating anti-TNF therapy.
In summary, perianal fistula response and healing in PFCD occurred more frequently in patients treated with combined surgical intervention and anti-TNF therapy compared with surgical intervention alone but not with anti-TNF therapy alone.While these results suggest that combined modality therapy may not be universally required in all patients with PFCD, prospective controlled studies will be necessary to confirm these findings.

Figure 2 .
Figure 2. Fistula healing in combined modality therapy vs anti-TNF therapy alone.

Figure 3 .
Figure 3. Fistula healing in combined modality therapy vs surgical intervention alone.

Table 1 .
Our search strategy was guided by an experienced information specialist [KF] at the University of Ottawa Health Sciences Library in collaboration with our research team.MEDLINE, EMBASE, and Cochrane Central Register of Controlled Trials [CENTRAL] were searched from inception to December 2023 using the MeSH terms summarized in Supplementary To ensure completeness of our search we also reviewed the bibliographies of each included paper, and relevant review articles.Further, we manually searched published abstracts from major international conferences [Digestive Disease Week, European Crohn's and Colitis Organization, and United European Gastroenterology Week] up to 2023.

Table 1 .
Characteristics of included studies

Table 3 .
Fistula healing or fistula response outcomes of included studies

Table 4 .
Faecal diversion outcomes of combined modality therapy vs anti-TNF therapy alone

Table 5 .
Faecal diversion outcomes of combined modality therapy vs surgical intervention alone *Study did not specify whether the faecal diversion surgery was a diverting ostomy or proctocolectomy.