In general there is no effect of age on healing rates in multiple studies [22–27]. In studies that have shown a difference in healing rates with age, increased age is associated with a higher likelihood of healing. Gustafsson et al. found a trend toward a higher likelihood of healing with age greater than 50 [28]. Similarly, healing rates were 45.7 % for age less than 40, 67.9 % for those aged 40–60, and 100 % for those older than 60 in a paper by Sonoda et al. [29]. One confounding factor may be the prevalence of Crohn’s; in the Sonoda paper a higher proportion of the younger patients had Crohn’s disease while the older patients were more likely to have cryptoglandular fistulas. However, some of the studies demonstrating no effect of age on healing rates included significant numbers of patients with Crohn’s disease [22, 26]. Thus, it is unclear whether the decreased healing at younger ages found in some studies is due to a differential prevalence of Crohn’s among the study patients at different ages.

The majority of studies have found that gender does not affect fistula healing rates [28], including in multiple logistic regression analyses after controlling for other factors [24, 25, 27]. However, one study did find a significantly greater proportion of men had primary healing of their fistula [30]. Seventeen of 24 males vs. 6 of 18 women in this study had primary healing of their fistula after closure of the internal opening was performed; in the majority of cases this closure was done with a partial-thickness endorectal advancement flap.

The data on the effect of obesity on healing rates is mixed. Schwandner et al. found that obesity, defined as a body mass index (BMI) greater than 30 kg/m², was associated with a decreased success rate for full-thickness flaps [23]. In this study, the recurrence rate was 14 % for non-obese patients vs. 28 % for obese patients, and this association continued after adjustment for other factors. Among the patients with recurrence of their fistula, there was also a higher need for reoperation for abscess among obese patients vs. non-obese patients. However, other studies have found no difference in recurrence with obesity [24, 26, 28], or even increased healing with greater body surface area [29].

Many studies have found no effect of smoking on flap success [23, 25, 28, 31].

However, smoking is associated with a higher recurrence rate in other studies [22, 24, 27], which may be plausible due to the possibility of decreased blood flow to the rectal mucosa as a result of smoking [31]. All of these studies performed multivariate analyses which demonstrated that smoking was independently associated with fistula recurrence after endorectal advancement flap. Ellis and Clark [27] found a 51 % recurrence rate for smokers vs. 19 % for nonsmokers undergoing endorectal advancement flap. Similarly, Zimmerman et al. [24] found a 40 % recurrence rate among smokers vs. a 21 % recurrence rate among nonsmokers; in this study the healing rate was significantly less if the patient smoked more than ten cigarettes per day. It may therefore be prudent to encourage patients to quit smoking prior to endorectal advancement flap.

The use of systemic medications in Crohn’s disease also has the potential to affect success rates. Steroid use has not been found to affect healing rates in some studies [10, 22], with other studies showing a trend toward an increased likelihood of failure with steroid use [26, 29]. This may be due to steroid use serving as a proxy for a greater severity of Crohn’s disease, which would predispose patients to recurrence or persistence of their fistulas. In contrast, there is evidence that biologic immunomodulators may contribute to the success of endorectal advancement flaps in patients with Crohn’s disease. In a case series of 19 patients with Crohn’s disease who were treated with preoperative infliximab, eight healed and did not require surgery. The remaining 11 underwent endorectal advancement flaps with an 82 % success rate [32]. Similarly, in a retrospective review of 218 patients with Crohn’s undergoing a variety of surgical interventions for anal fistulas, there was improvement or healing in 71.3 % of those receiving biologic immunomodulators vs. 35.9 % of those not receiving biologics, although the overall healing rate was low at 26.5 % for surgery alone and 36.6 % for surgery plus biologic immunomodulators [33]. Biologics thus show some promise as an adjunct to endorectal advancement flaps in patients with Crohn’s disease.

Although there have been some studies that show no difference in recurrence rates based on etiology of the fistula [7], the preponderance of evidence suggests that fistulas associated with Crohn’s disease tend to have a higher recurrence rate than fistulas of other etiologies [5, 10, 22, 29]. For example, Sonoda et al. [29] found a healing rate of 50 % for Crohn’s fistulas vs. 77 % for cryptoglandular fistulas, and Mizrahi et al. found rates of 43 and 67 %, respectively [10]. There is some evidence that the activity of Crohn’s disease, not just the presence of Crohn’s, can affect recurrence rate also. A success rate of 25 % has been found in the presence of small bowel Crohn’s, vs. 87 % in the absence of small bowel Crohn’s [6]. In contrast, though, Crohn’s activity was not found to affect the healing rate after rectovaginal fistula repair (done in most cases with a mucosal advancement flap although a significant minority of patients had other procedures performed) [26]. Patients with Crohn’s disease should therefore be counseled that they may experience a higher rate of recurrence after endorectal advancement flap than patients with fistulas due to cryptoglandular or other causes.

Location of the fistula does not appear to affect healing rates, with anterior, posterior, and lateral fistulas having similar healing rates [23, 24, 34]. Data are mixed as to whether different types of fistulas have differential healing rates. Mizrahi et al. found no difference in healing rates between anorectal, rectovaginal, pouch-perineal, and rectourethral fistulas in a case series of 106 flaps in 94 patients [10], although there were necessarily small numbers of some of these types of fistulas. A number of studies have compared rectovaginal fistulas to other fistulas, with a higher healing rate [6, 35], lower healing rate [5], and no difference [27] all having been found.

Data are similarly mixed as to the effect of fistula complexity on healing rates. “Complex” fistulas (horseshoe, suprasphincteric, or anovaginal fistulas or those with other extensions) have not been found to have lower healing rates than more straightforward fistulas [28, 30]. Fistulas with a horseshoe component have been found to have higher [34], lower [13], and similar [24] healing rates when compared to fistulas without a horseshoe component. The healing rate for rectovaginal fistulas was not found to vary by the location of the fistula (high vs. low) or size of the fistula opening by Pinto et al. [22]. Referral to a tertiary institution may also serve as a proxy for fistula complexity, but has not been found to affect success rates in the studies that have examined this factor [23, 25]. Thus, surprisingly, the majority of studies show no effect of complexity on healing rates.

Prior surgical attempts to repair the fistula are another factor that may serve as a proxy for fistula complexity. Schouten et al. found a success rate of 87 % for transsphincteric cryptoglandular fistulas treated with endorectal advancement flaps if there had been only one or no prior attempts at repair vs. a success rate of 50 % if there had been two or more attempts at repair [36]. Lowry et al. found similar results for among a group of patients treated for rectovaginal fistulas with endorectal advancement flaps (of note, 31 % of these patients had concomitant overlapping sphincteroplasties). Success rates were 88% among those with no prior repairs, 85 % with one repair, 55 % with two repairs, and 100 % with three repairs, with the relative risk of failure for those with two prior attempts vs. none being 3.71 [37]. Additional studies have found a decreased success rate [7, 27] or trend toward this [38] with prior attempts at repair. However, many other studies have found no difference in the recurrence rate in the presence of prior attempts at repair [6, 10, 25, 28] or any relationship to the number of prior attempts at repair [23, 24, 26]. Thus, while patients with a history of multiple prior attempts at repair should be cautioned about the risk of failure, there is evidence that they can expect a success rate which may not be markedly different than patients who have not had prior attempts at repair.

A seton is often placed prior to surgery to allow maturation of the fistula tract prior to endorectal advancement flap. There is some evidence that this may contribute to a greater likelihood of healing. A greater success rate for endorectal flap after seton placement [29], or a trend toward this, [26] has been found in some case series but not in others [24, 25, 39]. However, in all of these studies the choice of a preoperative seton was not random, suggesting that these were likely placed in situations where the surgeon anticipated a lower likelihood of healing. The finding of no difference or an increase in healing in these presumably more difficult fistulas suggests that setons are of benefit. Seton placement for a minimum of 6 weeks prior to flap should therefore be strongly considered.

Reports vary widely on the use of antibiotics and constipating medications. While the majority of centers administer a dose of perioperative antibiotics, some centers also continue antibiotics postoperatively for variable durations. Some centers limit patients to a clear liquid diet for a period of time and/or place them on constipating medications, while others have no particular restrictions. Studies in general do not show a benefit of postoperative antibiotics [7, 29] or a postoperative regimen including clear liquid diet, immobilization, and antibiotics [39]. Only one report demonstrates increased healing with postoperative antibiotics [30]. Not all of the patients in this study had flaps performed, and all had fistulectomy performed in addition, so the generalizability to patients undergoing endorectal advancement flap is limited. In terms of postoperative bowel regimen to promote constipation, no difference in healing rates has been found between a constipating regimen and no regimen [7, 10]. Thus, perioperative antibiotics may be used, but postoperative antibiotics and a restricted bowel regimen are not necessary.

Partial-thickness flaps are likely more successful than full-thickness flaps. In a review of the literature incorporating a total of 1,654 patients, Soltani and Kaiser examined the effect of flap type on healing rates [40]. They found that partial-thickness flaps were used more often in the studies reporting above-average success rates, while full-thickness flaps were associated with below-average success rates except in one study [41]. Mucosal flaps were represented equally in studies reporting above- and below-average success rates. Partial-thickness flaps, incorporating mucosa, submucosa, and some muscle fibers, should be preferred.

The presence of a diverting stoma has not been found to have an effect on fistula healing after endorectal advancement flap [10, 26, 29]. However, all of these studies were case series in which the choice of whether to perform a stoma was nonrandom. Most likely the patients selected to have a stoma had fistulas that were thought to have a low likelihood of healing, and as such it is unclear whether the stomas performed in these cases contributed to a higher healing rate than would otherwise have been found for these difficult fistulas. Diverting stoma may be a good option in selected cases.

A number of modifications to the endorectal advancement flap technique have been attempted, but generally have not led to increased success rates. Perhaps the most enthusiasm surrounded the injection of fibrin glue into the fistula tract in addition to performing the flap. Although one study found no effect of glue injection on healing, there were only 12 patients who had glue injected. Instead, there is good evidence to suggest that fibrin glue decreases the chance of successful healing. Both a case–control study [42] and a randomized controlled trial [43] have demonstrated a decreased success rate with glue. In the case–control study, 26 patients who had fibrin glue and advancement flap were matched to 54 who had advancement flap only. The recurrence rate was 17 % for advancement flap alone vs. 46 % for flap and fibrin glue. Similarly, in the randomized controlled trial there was a recurrence rate of 20 % for flap alone vs. 46 % for flap and fibrin glue (but these patients had either mucosal advancement flaps or anodermal flaps rather than partial-thickness rectal flaps). Fibrin glue should not be used in conjunction with advancement flap. Sileri et al. performed a similar injection with porcine dermal collagen matrix, achieving success in 10 of 11 patients [44], but experience with this material is limited.

Gustafsson and Graf did a randomized controlled trial comparing flap alone to flap with a gentamicin-collagen sponge implanted underneath and found no difference in healing rates between the two groups [28]. Similarly, van Onkelen et al. found a healing rate of only 51 % when performing the LIFT (ligation of intersphincteric fistula tract) procedure in addition to endorectal advancement flap [45]. Thus, there are no modifications to the endorectal advancement flap that have been found to consistently improve outcomes over flap alone.

Although endorectal advancement flaps do not divide full-thickness muscle, there are a number of reasons why they may negatively affect continence. A partial-thickness flap does take some muscle fibers, including some of the internal sphincter fibers if the flap extends distal to the dentate line as it does in most cases. If the internal opening is at the dentate line, flaps also cause the rectal mucosa to extend past the dentate line, creating some degree of ectropion. Finally, the amount of stretch that must be put on the sphincter muscles intraoperatively could cause some temporary or even permanent incontinence.