Patients
Between January 2010 and June 2020, 155 mucosal advancement flap operations for patients with anal fistula were performed (table 1). The study included 55 CD patients and 100 non-CD patients. Most patients had complex fistulas: 83 % had a trans-sphincteric fistula and 15 % a vaginal fistula course. 41 patients (26 %) received immunosuppressant medication. Median healing over all flaps was 56 %. Nine patients (5.8 %) developed acute complications (hematoma, bleeding) with the need for redo surgery.
Table 1: General patient data over all advanced flaps.
|
n = 155
|
Pathogenesis
Crohn’s disease Cryptoglandular
|
55 (35) 100 (65)
|
Fistula course
Vaginal Trans-sphincteric Supra-sphincteric Inter-sphincteric
|
24 (15) 128 (83) 2 (1) 1 (1)
|
Ostomy
|
39 (25)
|
Age, years
|
40 (12 – 73)
|
Sex, female
|
71 (46)
|
ASA, 1 - 2
|
144 (93)
|
BMI, m2/kg
|
25 (17 – 44)
|
Immunosuppressant medication
Anti-TNF Anti-Interleukin Anti-Integrin Other
|
41 (26)
28 (18) 1 (0.5) 5 (3) 7 (4.5)
|
Fistula healing
|
86 (56)
|
Complications
|
9 (5.8)
|
Mortality
|
0
|
Follow up, days (95 % CI)
|
189 (109 – 269)
|
Median (min – max) for continuous variables, count (percentage) for categorical variables, except for Follow up: median (95 % CI). ASA Score = American Society of Anesthesiologists Score, BMI = Body mass index.
Risk of anal fistula healing failure after advancement flap for all patients
In table 2 the univariate analysis showed female gender, immunosuppressant medication and vaginal fistula course as significant influencing factors for healing failure. Crohn’s disease, BMI, ASA 1 and 2 or the presence of protective stoma showed no influence on anal fistula healing failure. P values less than 0.2 from the univariable analysis were enrolled in a multivariable logistic regression model to identify independent risk factors for healing failure. Logistic regression analysis could not find any independent influencing factor on the healing after rectal advancement flap.
Table 2: Factors affecting anal fistula healing failure in all patients.
|
Healing n = 86 (56 %)
|
Healing failure n = 69 (44 %)
|
Missing
|
P value
|
Logistic regression
|
Odds ratio (95% CI)
|
P value
|
Age, years
|
43 (12 - 73)
|
36 (14 - 64)
|
0
|
0.063
|
0.987 (0.961 - 1.013)
|
0.319
|
Sex, female
|
32 (37)
|
39 (56)
|
0
|
0.013
|
1.803 (0.806 - 4.034)
|
0.152
|
BMI, m2/kg
|
25 (17 - 42)
|
25 (17 - 44)
|
4
|
0.460
|
-
|
-
|
ASA, 1 - 2
|
78 (91)
|
66 (97)
|
0
|
0.191
|
0.440 (0.095 - 2.033)
|
0.293
|
Immunosuppression, yes
|
17 (20)
|
24 (35)
|
0
|
0.027
|
2.262 (0.812 - 6.295)
|
0.118
|
Pathogenesis
Crohn’s disease Other
|
26 (30) 60 (70)
|
29 (42) 40 (58)
|
0
|
0.088
|
0.589 (0.213 - 1.628)
|
0.308
|
Vaginal fistulas
|
9 (38)
|
15 (62)
|
0
|
0.044
|
0.507 (0.171 - 1.499)
|
0.219
|
Protective stoma
|
19 (22)
|
20 (29)
|
0
|
0.213
|
-
|
-
|
ASA Score = American Society of Anesthesiologists Score, BMI = Body mass index.
Advancement flap in CD patients
Table 3 shows differences in characteristics of patients with and without CD. CD patients were significantly younger, more female, received more immunosuppressant medication and had a lower BMI than non-CD patients. In addition, CD patients were more likely to have protective ostomy. There were significantly more patients with vaginal fistulas among CD patients. However, the healing rate of anal fistula did not differ between CD and non-CD patients (p = 0.088).
Table 3: Characteristics for patients without and with Crohn’s disease (CD).
|
CD patients n = 55
|
Non-CD patients n = 100
|
Missing
|
P value
|
Age, years
|
33 (14 – 66)
|
43 (12 – 73)
|
0
|
< 0.001
|
Sex, female
|
41 (75)
|
30 (30)
|
0
|
< 0.001
|
BMI, m2/kg
|
24.4 (17.3 – 34.9)
|
25.6 (16.5 – 43.6)
|
4
|
0.033
|
ASA, 1 - 2
|
54 (98)
|
90 (90)
|
0
|
0.051
|
Immunosuppression, yes
|
36 (65)
|
5 (5)
|
0
|
< 0.001
|
Vaginal fistula
|
16 (29)
|
8 (8)
|
0
|
0.001
|
Ostomy
|
27 (49)
|
12 (12)
|
0
|
< 0.001
|
Fistula healing
|
26 (47)
|
60 (60)
|
0
|
0.088
|
CD = Crohn’s disease, ASA Score = American Society of Anesthesiologists Score, BMI = Body mass index.
A subgroup analysis was performed for CD patients to identify possible influencing factors for anal fistula healing failure (table 4). Neither age nor sex, gender, BMI or ASA showed a significant influence on fistula healing after advancement flap. A vaginal fistula course or the presence of a protective stoma were also irrelevant to the healing process. Only in immunosuppressant medication was there a significant difference between healing and healing failure in the univariable analysis. P values less than 0.2 from the univariable analysis were enrolled in a multivariable logistic regression model to identify independent risk factors for healing failure. Logistic regression analysis could not find any independent influencing factor on the healing after rectal advancement flap.
Table 4: Factors for healing failure of advanced flaps in CD patients.
|
Healing n = 26
|
Healing failure n = 29
|
P value
|
Logistic regression
|
Odds ratio (95% CI)
|
P value
|
Age, years
|
33 (20 – 66)
|
31 (14 – 54)
|
0.428
|
-
|
|
Sex, female
|
17 (65)
|
24 (82)
|
0.122
|
1.491 (0.324 - 6.857)
|
0.608
|
BMI, m2/kg
|
24.5 (17.3 – 33.7)
|
24.4 (19.1 – 34.9)
|
0.873
|
-
|
|
ASA, 1 - 2
|
25 (96)
|
29 (100)
|
0.473
|
-
|
|
Immunosuppressant medication
Non Anti-TNF Anti-Interleukin Anti-Integrin
|
14 (54)
12 (46) 13 (93) 1 (7) 0
|
22 (76)
7 (24) 15 (68) 0 7 (32)
|
0.025
|
- 2.669 (0.704 - 10.118) n/a n/a
|
0.555
- 0.149 0.999 1.000
|
Vaginal fistula
|
5 (19)
|
11 (38)
|
0.109
|
0.349 (0.081 - 1.505)
|
0.158
|
Ostomy
|
12 (46)
|
15 (52)
|
0.444
|
-
|
|
ASA Score = American Society of Anesthesiologists Score, BMI = Body mass index. n/a - due to missing values, no odds ratio could be determined.
Healing failure
The median follow-up for CD patients was 210 days (95 % CI: 53 – 368). Non-CD patients had a median follow-up of 89 days (111 – 267). Two patients (1 %) were lost to follow-up. Healing failure occurred in 69 (44 %) of 155 advancement flaps. Kaplan-Meier estimates for fistula healing failure did not differ between patients with cryptoglandular fistulas and patients with CD (figure 1).
The CD and non-CD patients with anal fistula healing failure were further classified in two categories according to the time of relapse occurred. An early relapse was predefined if it is reported within 14 days, and a late relapse later than 14 days. Of 29 CD patients with anal fistula healing failure, 8 patients (28 %) had a median early relapse of 10 days (6 – 12) and 21 patients (72 %) had a median late relapse of 84 days (29 – 1016). Of 40 non-CD patients with anal fistula healing failure, 14 (35 %) patients had a median early relapse of 8 days (4 – 14) and 26 patients (65 %) had a median late relapse of 85 days (16 – 1521). The difference for both early (p = 0.552) and late (p = 0.082) healing failure was not significant.