Introduction

Treatment of recurrent and complex anorectal fistulas is challenging. Surgical therapy for fistulas in patients with inflammatory bowel disease (IBD) is associated with high recurrence rates. Fistulas in patients with IBD result from chronic relapsing inflammation with subsequent intestinal fibrosis and stricture formation, and could be triggered by the epithelial–mesenchymal transition [1,2,3]. The etiology of the fistula reflects the existing comorbidities, which should also be addressed intraoperatively (e.g., perineal defects and scar tissue with compromised local vascularization as a result of previous colorectal procedures). Depending on the location of the fistula, different approaches can be used, including transabdominal, transperitoneal, or local techniques. A number of local surgical methods exist, such as excision, advancement flaps, and fibrin glue or biomesh application. Autologous muscle flaps—such as the Martius, sartorius, and gracilis flaps—are muscles flaps—such as the Martius, sartorius, and gracilis flaps—are further methods for fistula closure. However, no consensus regarding the preferred method of choice has been reached. Combining different procedures after an insufficient fistula closure may be useful to improve the efficacy of a single method. Larger recent studies [4,5,6,7,8,9,10] on GMT, including those of Ulrich et al. [6] and Wexner et al. [9], have demonstrated some excellent results.

Materials and methods

Study design and patients

A two-center retrospective study was performed from January 2000 to May 2018 in patients diagnosed with IBD. All patients underwent GMT at the surgery departments of the University Hospital Regensburg and the St. Josef Hospital Regensburg, Germany, between January 2000 and May 2018 to treat anovaginal, rectovaginal (RVF), rectourethral (RUF), and pouch–vaginal fistulas. All gracilis transpositions were performed by a single surgeon (AF). IBD patients who underwent GMT strictly due to fecal incontinence and not for fistula treatment were excluded. Follow-up procedures in cases of fistula recurrence were performed by other surgeons. Prior to GMT, all patients had undergone at least one failed attempt at fistula repair using various other procedures such as fistulotomy, fistulectomy, mucosal advancement flap, fistula plug, fibrin glue application.

Surgical technique

GMT was performed according to a standardized procedure [11]. We combined the GMT in all cases but one with a fecal diversion performing a loop ileostomy if no stoma existed. We prefer performing a GMT when a stoma is already established to reduce the local infection beforehand. Antibiotic prophylaxe with cefuroxime und metronidazole was given perioperatively. The fistula tract was identified by probing (Fig. 1) and excised completely. The defect on the rectal side was minimized with interrupted sutures. The gracilis muscle was identified and an incision measuring approximately 10–12 cm was made on the medial side of the thigh (Fig. 2). The distal muscle tendon was then identified by palpation near the pes anserinus and divided after a small incision medial to the knee joint. The muscle was mobilized and the small collateral blood vessels in the distal part of the muscle were ligated using ties for the bigger blood vessels and coagulation for the smaller ones. During mobilization, great care was taken to find a tension-free position of the neurovascular bundle on the lateral proximal side of the gracilis muscle. A subcutaneous tunnel between the ipsilateral proximal medial side of the thigh and the perineal region was created by blunt preparation. The muscle then was pulled through the tunnel and placed exactly underneath the suture line closing the rectal side of the fistula (Fig. 3), covering the rectum, placing well-vascularized tissue in the defect area (Fig. 4). Rotation of the muscle around its axis was carefully avoided. The distal tendon of the muscle was then fixed with Prolene sutures (Ethicon, Bridgewater, NJ and Cincinnati, OH, USA) to the contralateral pubic or ischial periosteum after transposition through an additional subcutaneous tunnel and a small incision on this side. Finally, the gracilis muscle was covered by a soft tissue flap (Fig. 5). If the medical history of the patient indicated fecal incontinence, the distal muscle tendon was transposed further to encircle the anal canal, and fixed at the ipsilateral pubic or ischial periosteum, thus creating a loop (gamma-loop; Fig. 6). This procedure potentially increases anal continence by contributing additional muscle tension. A secondary dynamization of the gracilis muscle remains an additional option for optimization of anal sphincter function. If no fecal diversion existed preoperatively, a protective stoma was added simultaneously. Only 1 patient opted against a protective stoma.

Fig. 1
figure 1

Complex rectovaginal and perineal fistula combined with a perineal defect

Full size image
Fig. 2
figure 2

Mobilization of the gracilis muscle

Full size image
Fig. 3
figure 3

Transposition of the gracilis muscle

Full size image
Fig. 4
figure 4

Fixation of the gracilis muscle on the contralateral pubic periost (tunneling technique)

Full size image
Fig. 5
figure 5

Vaginal flap to cover the gracilis muscle

Full size image
Fig. 6
figure 6

Types of gracilis muscle transposition (GMT) [10]. a Circular GMT, b transverse GMT

Full size image

Outcome assessment

The primary outcome measure was complete closure of the fistula by the first follow-up (approximately 3 months postoperatively) without additional follow-up operations. In patients with anovaginal, rectovaginal, and pouch–vaginal fistulas, the initial follow-up was performed 3 months postoperatively. A sensitive sign of the presence of a residual fistula is a feeling of air-trapping and vaginal discharge. All patients had clinical examination, including anoscopy, air insufflation, and fistula probing (in selected cases under general anesthesia). In case of uncertainty regarding a residual fistula, a water-soluble contrast enema, anorectal endosonography, or pelvic magnetic resonance imaging was performed. Stoma reversal was performed upon definitive closure of a fistula tract.

Secondary outcomes included fistula closure after follow-up procedures in cases of fistula persistence and stoma closure rate. Fistula-free time in cases of late recurrences was also documented.

Data analysis

Information regarding patients’ demographic characteristics, number of previous operations, types of fistula, fistula closure, complications, follow-up procedures, and stoma closure was collected retrospectively from medical records. Data were gathered in a Microsoft Excel Sheet Version 2010 (Microsoft Corporation, Redmond, WA, USA). Data are presented using summary statistics in the form of means and median values for quantitative data and percentages for qualitative data.

Ethical considerations

Due to the retrospective nature of the study, formal consent was not required. The study was approved by the local ethics committee of the authors’ study center.

Results

A total of 32 patients (2 male, 30 female; mean age 39 years, range 24–55 years) with IBD underwent gracilis transposition or re-transposition for a recurrent fistula at the authors’ institutions between January 2000 and March 2018 and attended at least the first follow-up visit at 3 months. Two patients with less than 3 months of follow-up were also included (one because of graft ischemia and the other because of a persistent fistula) to minimize selection bias and present a more realistic recurrence and complication rate. All patients had undergone diverse previous surgical procedures for fistula closure (mean 2 procedures, range 1–25 procedures) before the gracilis transposition. Mean follow-up time after GMT was 47 months (range 1–144 months). After GMT, some patients suffered a recurrence or had a smaller residual fistula, prompting different follow-up procedures to achieve definitive closure (Tables 1, 2 and 3).

Table 1 Additional procedures after GMT for patients with recurrence
Full size table
Table 2 Additional procedures after GMT for patients with fistula persistence
Full size table
Table 3 Fistula closure rate after follow-up procedures depending on gender
Full size table

Of the 32 patients with IBD, 1 developed a wound infection with necrosis of the gracilis flap and 7 had a residual fistula at the first postoperative follow-up and required additional procedures. During the subsequent follow-up period, 9 further patients had a recurrent or new fistula and one patient had an abscess without a verification of a fistula (median time after GMT among these 9 patients before recurrence: 17 months; range 5–62 months). In 7 of those cases a recurrence took place after stoma closure (median time after stoma closure until recurrence 17 months, range 2–59). It was difficult to distinguish whether the patient had a recurrence or a new spontaneous fistula due to Crohn’s disease (CD). In 1 patient, a recurrence was documented directly after a CD relapse. The primary fistula closing rate where the fistula stayed closed during the follow-up and no further procedures were needed was 47% (15/32).

In total, an additional 43 operations were required for the 17 patients with recurrences and persistence of a fistula after GMT to achieve a definitive fistula closure rate after recurrences of 71% (23/32) among all 32 patients (Table 1). Of these operations, 17 were performed to close a recurrent or persistent fistula in 8 patients (2.1 follow-up procedures per patient per closed fistula after recurrence or persistence). Despite performing a total of 26 other follow-up procedures on 9 other patients, those 9 other patients had a persistent fistula and in one another case there is a suspicion of a recurrence due to an abscess without a clinical proof of a fistula.

At the time of GMT, 15 patients were not taking any specific CD medication (Table 4). In six of those cases the fistula healed without follow procedures. Only in 4 of those 15 patients (healing rate 73%) without specific CD medication at the time of GMT a fistula remained after follow procedures. In 5 of the 17 cases (healing rate 71%) where patients took a specific CD medication a residual fistula remained after follow procedures. However, the p value for comparing the healing rate with or without specific CD medication is 0.589 and thus not of statistical significance.

Table 4 CD medication at the time of GMT
Full size table

GMT was performed without fecal diversion in 1 patient and the fistula healed anyway. In 19 patients, the stoma could be reversed after fistula closure and in 2 patients, stoma reversal is intended after fistula healing. A late recurrence after stoma reversal was suffered by 7 patients, 4 of whom received a new stoma with follow-up surgical procedures thereafter. As a result, in 3 of the 7 latter cases, fistula closure was achieved. In summary, a stoma reversal after follow-up procedures due to fistula persistence or recurrence was achieved in 18 of 31 (58%) of the stoma patients and in two other cases where a fistula closure has been achieved a stoma closure is being planned.

Mean age was 42 years (median 43 years, range 24–54 years, 95% CI 33–51 years) in the patients with complete remission after GMT and 40 years (median 41 years, range 27–55 years, 95% CI 33–49 years) in the patients who suffered a recurrence (p = ns).

The mean number of operations before GMT was 4.3 (median3, range 1–20) and 4.4 (median2, range 1–25) in the groups who did not and did suffer a recurrence, respectively (p = ns).

There was no difference in fistula closure rates between patients who received a gamma-loop of gracilis muscle to the ipsilateral pubic ramus (7/10, 70%) and those who underwent GMT on the contralateral pubic ramus (15/22, 68,1%) (p = ns).

Data were collected on fistula closure and the closure rate by fistula etiology was calculated (Table 5).

Table 5 Fistula closure rate among patients with IBD
Full size table

Complications

Aside from 1 case of wound infection with ischemia of the gracilis muscle, 1 case of postoperative hemorrhage in the thigh 2 cases of suture granuloma and complaints about scar tissue and some numbness on the operated thigh, no other major complications following harvesting and inserting of the gracilis muscle were documented.

When a new fecal diversion was performed during the same session, some patients needed postoperative intravenous fluid and laxative agents due to ongoing paralytic ileus, but no surgical interventions were needed (Table 6).

Table 6 Complications
Full size table

Discussion

Muscle grafts, particularly gracilis muscle grafts, have been in use since the 1930s. It should be noted that the gracilis muscle does not close the existing fistulas per se, but is used to reduce dead space in the perineum and improve vascularization and healing in areas with compromised blood supply due to soft tissue and muscle loss as a result of previous operations. GMT can be combined with other surgical interventions and, thus, improve the rate of fistula closure.

Alternative approaches include transperineal, endorectal, endovaginal, and transabdominal procedures. Fistula closure can be achieved directly using sutures or glue, with a mucosal flap or biomaterials, via partial bowel resection with anastomosis, or through muscle flap transposition. Fecal diversion can be performed as a stand-alone treatment or used perioperatively as a supplementary measure in some surgical procedures.

Biomaterials represent a relatively new method of treating fistulas. Benefits include less intraoperative trauma and a decreased impact on fecal continence, while treating trans-sphincteric fistulas with moderate–good outcomes [12,13,14,15,16,17]. Chan et al. [18] showed good results using a fistula plug in simple fistulas, but also a decrease in efficacy after multiple procedures. Other studies showed a high rate of postoperative sepsis after implantation of fistula plugs alone [19, 20]. In combination with a draining seton or an advancement flap, the success rate increases. Long-term studies of this method are necessary to judge its efficacy.

Panés et al. [21,22,23] describe a promising new therapeutic approach to close treatment-refractory complex perianal fistulas in CD using a single intralesional injection of allogeneic expanded adipose-derived stem cells. This method achieved a 50% remission rate (53 of 107 patients) in an intention-to-treat population. The most common treatment-related adverse events were anal abscess and proctalgia.

The fistula closing rates by sex are shown in (Table 3); however, the significantly smaller number of patients in the male group makes comparisons difficult. All in all, we achieved a definitive fistula closing rate of 71%. Our study including 32 patients is the largest study of performing GMT on IBD patients in the literature so far.

Compared to other studies on GMT, the current study group consisted entirely of patients with IBD, which is generally associated with lower fistula healing rates using different procedures [6, 9, 24,25,26]. Some patients with IBD show apparent late recurrences; however, some of these may be attributable to spontaneous development of new fistula due to CD activity, rather than a recurrence of a previous fistula.

No significant difference in terms of age was detected between the groups with remission and recurrence. Therefore, the authors would recommend the procedure for IBD patients of all ages. There was also no significant difference in healing rate between patients taking CD specific medications and patients not taking any CD specific medications at time of GMT. 

Fistula persistence after GMT was detected no later than at the first follow-up after GMT, and was treated accordingly. Late recurrences, however, appeared unequally distributed after a mean of 24 months after GMT (median 23 months, range 5–62 months). Almost half of the recurrences took place 5–7 months after GMT and the others appeared 1.5–5 years postoperatively. Another follow-up with clinical examination and anorectal sonography 6–9 months after GMT might be advantageous for earlier detection of a recurrence and initiation of treatment.

Adequate treatment of IBD seems to have a positive impact in preventing late recurrences. Due to the small number of patients in the present cohort, it was not possible to detect whether success after GMT correlated with the different types of CD medication.

The fact that all GMTs were performed by a single surgeon using a standardized procedure reduced the possibility of the outcome being influenced by varying skill levels of different surgeons.

The study has some limitations. First, only a small number of patients were included. Further studies with larger patient numbers are needed. There are, however, only a few similar studies on fistula treatment in IBD patients in the recent literature.

Due to its retrospective design, the current study lacks a uniform follow-up period. A defined interval between the procedure and examination of the patients could improve comparison between the different patients and success rates. The follow-up at 3 months postoperatively did not apply to all patients, mainly because some suffered a persistent fistula and were examined earlier than after 3 months. The next follow-up after the 3-month follow-up was also somewhat heterogeneous, since most patients received a further follow-up only due to complaints; a large proportion of the patients who achieved complete remission and stoma closure did not attend further examinations.

Another critical point of the study is the correlation between GMT and the fistula closure rate when additional procedures were needed. It is impossible to determine the exact extent of the contribution of GMT to the fistula closure rate when additional operations were performed after GMT. The authors are convinced that the gracilis muscle is the basis for the success of additional fistula operations. Larger trials directly comparing GMT to other surgical procedures with a long follow-up period should be performed to better understand its benefits. This would contribute to determining the best way to treat recurrent fistulas in patients with IBD (Table 7).

Table 7 Comparison of pre-existing studies
Full size table

Conclusions

Our results are promising and suggest that GMT is a safe and very effective option for the treatment of recurrent fistulas in IBD patients. It should be noted that for complete fistula closure, follow-up procedures may be necessary. The option of dynamization of the muscle creates the possibility of improving fecal continence and, thus, quality of life. The authors suggest that in cases of recurrence after initially addressing a fistula with conventional methods, and after repeated unsuccessful procedures, GMT should be considered.