Autologous adipose‐derived stem cells for the treatment of complex cryptoglandular perianal fistula: A randomized clinical trial with long‐term follow‐up

Abstract The aim of this clinical trial (ID Number NCT01803347) was to determine the safety and efficacy of autologous adipose‐derived stem cells (ASCs) for treatment of cryptoglandular fistula. This research was conducted following an analysis of the mistakes of a same previous phase III clinical trial. We designed a multicenter, randomized, single‐blind clinical trial, recruiting 57 patients. Forty‐four patients were categorized as belonging to the intent‐to‐treat group. Of these, 23 patients received 100 million ASCs plus intralesional fibrin glue (group A) and 21 received intralesional fibrin glue (group B), both after a deeper curettage of tracks and closure of internal openings. Fistula healing was defined as complete re‐epithelialization of external openings. Those patients in whom the fistula had not healed after 16 weeks were eligible for retreatment. Patients were evaluated at 1, 4, 16, 36, and 52 weeks and 2 years after treatment. Results were assessed by an evaluator blinded to the type of treatment. After 16 weeks, the healing rate was 30.4% in group A and 42.8% in group B, rising to 55.0% and 63.1%, respectively, at 52 weeks. At the end of the study (2 years after treatment), the healing rate remained at 50.0% in group A and had reduced to 26.3% in group B. The safety of the cellular treatment was confirmed and no impact on fecal continence was detected. The main conclusion was that autologous ASCs for the treatment of cryptoglandular perianal fistula is safe and can favor long‐term and sustained fistula healing.


| INTRODUCTION
Perianal fistula has an incidence of 1.1 to 2.2 per 10 000 persons per year. The vast majority of cases are due to cryptoglandular disease. 1,2 In most patients, this condition may be successfully cured by surgery, but surgical treatment of complex fistulas remains a challenge, with a high rate of recurrence and frequent side effects such as fecal incontinence. [3][4][5] Seventeen years ago, our group began to explore the use of adiposederived mesenchymal stromal cells (ASCs) as a treatment option for patients with complex perianal fistula, 6 hypothesizing that the immunomodulatory and anti-inflammatory capabilities of ASCs could contribute to the healing process, improving outcomes as defined by fistula closure; healing was defined as the absence of drainage through the external openings and complete re-epithelization of these openings. [7][8][9] In phase I and II studies, the use of autologous ASCs was proved to be safe for the treatment of fistulas having both a cryptoglandular and Crohn origin. [10][11][12] A phase III clinical trial conducted to study cryptoglandular fistula using autologous ASCs failed to find an advantage of the intervention over the control group, possibly owing to issues related to the use of the cell product and trial design. In the previous clinical trial, cell manipulation by either vigorous agitation of the cell vials or use of hydrogen peroxide prior to cell injection was inadequate, as both effects trigger cell death due to friction or toxicity. In addition, control of the cell implant was not exhaustive, causing high injection speed and not always in the appropriate areas. 13 Later, a phase III clinical trial in Crohn perianal fistula using allogeneic ASCs showed a clear advantage when using these cells over the control group, 14 particularly in long-term evaluation. 15 The aim of this randomized clinical trial (RCT) was to investigate the efficacy of autologous ASCs for treatment of complex cryptoglandular perianal fistula; taking into account previous mistakes and shortcomings, especially as concerns cell manipulation, we should consider it a live medicine. Four modifications were introduced with respect to the previous phase III RCT: exhaustive control of cell manipulation, clear definition of complex fistula, larger number of cells implanted, and the use of cultured media containing platelet lysate instead of serum. The sample size (population calculation) comprised 80 patients to be randomized and treated, so five Spanish hospitals were included in the trial; however, only 57 patients met the inclusion criteria and did not present any of the exclusion criteria ( Figure 1). Briefly, all included patients presented complex perianal fistula of cryptoglandular origin with no indications of inflammatory bowel disease. Forty-four completed the assigned treatment and were evaluable (intention-to-treat

| MATERIALS AND METHODS
[ITT] population). Diagnosis of complex fistula was performed by physical examination and magnetic resonance imaging (MRI) in all patients. Treatment results were evaluated by an investigator blinded to the treatment at 1, 4, 16, 36, and 52 weeks as well as 2 years after the last treatment. Healing was defined as the absence of drainage through the external openings and complete re-epithelization of these openings. Incontinence was assessed by patients themselves using the Wexner score and quality of life using SF-12 questionnaire during follow-up.

| Protocol design
At least 2 weeks prior to therapy administration, all patients underwent liposuction for manufacturing of ASCs to be used during the study (group A) or for cryopreservation (group B). Liposuction was performed by plastic surgeons with patients under sedation and local anesthesia, obtaining at least 100 mL of fat lipoaspirate. This material was sent in a hermetic and sterile system at 2 C-8 C to clean room laboratories production ("Clínica Universidad de Navarra" or "Hospital

Significance statement
Autologous mesenchymal stem cells treatment for complex perianal fistula is safe, but according to the current results, it seems to provide an advantage over a good surgical protocol at 2 years after treatment and then results are similar to those shown with allogenic mesenchymal stem cells in previous clinical trials.
Samples from each vial were tested before release as described below. Fibrin glue (Baxter Inc, Spain) was used at a dose of 2-5 mL according to recommendations (1 mL per 4 cm 2 of surface area).
Treatments were administered in an operating room according to a standard surgical protocol, as described in "A Step-By-Step Surgical Protocol for the Treatment of Perianal Fistula with Adipose-Derived Mesenchymal Stem Cells", 16 and with external control during the surgical procedure in all cases. In brief, before injection of cells (or saline solution in group B), deep curettage of all tracts was performed and the internal opening was closed with the use of stitches. Half of each dose was injected around the internal opening, and the other half through the external opening. Finally, the fistula tract was filled with fibrin glue.

| Follow-up and efficacy assessment
Clinical evaluation of fistula healing, SF-12 questionnaire, and Wexner incontinence score were assessed at 1, 4, 16, 36, and 52 weeks after treatment. The incidence of adverse events (AEs) and serious adverse events (SAEs) was assessed at each study visit. An independent surgeon blinded to the treatment arm assessed healing at week 16, 36, and 52. Healing was defined as the absence of drainage through an external opening and complete re-epithelization of the external opening. Considering the results of efficacy at 1 year, we requested ethicscommittee permission to perform a retrospective study to carry out long-term follow-up (2 years) of the patients included in the study, with particular emphasis on long-term safety and recurrence of healed fistulas. A final follow-up was scheduled 2 years after the last treatment to determine the final state of healing and perianal suppuration.

| Statistical analysis
All statistical analyses were performed by Effice Research. The demographic and clinical data of the study subjects were described using means and SD of descriptive statistical indices. Quantitative variables were analyzed by calculating the 95% confidence interval (CI) and relative and absolute frequencies. Closure of the fistula was described by frequency, percentage, and 95% CI for the total sample and by

| DISCUSSION
This RCT was designed to determine the efficacy of autologous ASCs for the treatment of cryptoglandular complex perianal fistula. Here, we sought to correct the design errors observed during the previous phase III trial (FATT-1) 13 and taking into account the results of the previous phase II clinical trials. 11,12 During the phase II trial, which was unblinded, a healing rate of 71% was observed, with a low recurrence rate and no risk of fecal incontinence. Unfortunately, the results of the phase II trial could not be confirmed in FATT-1, a phase III study (blinded), as the safety of ASC therapy was demonstrated but the healing rate was 43.3% when using autologous ASCs plus fibrin glue (after a second dose, where applicable), which was not significantly different from the control group comprising patients treated   and phase III (FATT-1) 13 RCTs and from other clinical trials in Crohn disease using allogeneic ASCs and reporting on long-term assessment. 15 These long-term improvements can be explained by the biological action of these types of cells. [17][18][19] The results obtained in the control group merit commentary.
Although we treated cases classified as "complex fistula," more than 60% of the patients who received two doses of fibrin glue under our minimally invasive surgical protocol had complete resolution of the fistula 1 year later, and these results were better than those associated with a more aggressive surgical protocol. 20 We consider that the new surgical protocol based on a "cleaning surgery" (deep curettage) played a major role in the high number of fistula closures observed in both groups, particularly when considering that all patients had undergone more than two failed surgeries before being included in this trial.
However, we must consider that in this RCT we eliminated the "placebo effect," as it has been shown to have a powerful influence in the field of stem cell therapy, because we evaluated all results by a surgeon blinded to treatment group. As a result, we observed a high number of long-term recurrences in the control group. According to Figure 3, at 2 years, only 23% maintained cured status. Hence, we can conclude that this "cleaning surgery" does not provide lasting resolution, possibly because the inflammatory focus remains. In this scenario, stem cells (ASCs) and their anti-inflammatory effects can favor long-term healing, improving the results up to 50%. It should be emphasized that this is minimally invasive surgery and as such does not produce SAEs such as fecal incontinence. We observed a clear tendency toward better results in the long-term evaluation when stem cells were used in cryptoglandular surgery, possibly related to the notion that the cell works as a "living medicine" with long-term effects. 21,22 Another finding from our trial is related to the method used for ASC expansion and culture. Unlike in other studies, here ASCs were expanded using human platelet lysate in the culture media, which may have the advantage of avoiding the use of animal-derived factors.
Human platelet lysate is widely used to grow MSCs, [23][24][25] and many studies have demonstrated that this confers a degree of safety that is at least equivalent to that offered by fetal bovine serum. 26

| CONCLUSION
Although the findings of this study indicate that autologous ASC treatment for complex perianal fistula is safe, it seems that this procedure only provides an advantage over interventions based on a good surgical protocol at 2 years after treatment and is similar to the use of allogenic mesenchymal stem cells in previous clinical trials.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author.