Introduction

Crohn’s disease (CD) can affect any portion of the intestinal tract including the colon, ileocolon, and anoperineum, and 12–20% of patients with CD will ultimately require proctectomy with creation of a stoma for failed medical therapy or disease complications [1,2,3,4,5,6]. While the operation is intended to safely relieve debilitating symptoms and improve quality of life, impaired postoperative perineal wound healing can be a persistent problem complicating 7–40% of procedures [7]. This rate is much higher than that associated with proctectomy for other forms of benign or malignant disease [8,9,10,11], and the manifestations of impaired healing can range from superficial skin separation to a long presacral tract with an associated cavity.

Several risk factors have been inconsistently reported as predictors of impaired perineal wound healing, and these include sex, age, presence of a high fistula-in-ano, perioperative steroid usage, intraoperative contamination, and wound management [8, 12, 13]. Different methods of perineal dissection and forms of wound closure have been employed in an effort to reduce the likelihood of impaired perineal wound healing [14, 15].

We hypothesized that several risk factors are associated with impaired perineal wound healing after proctectomy for CD, and that some of these factors can be modified. Therefore, the aims of our study were to assess perineal wound healing after proctectomy for CD and evaluate the influence of various factors on eventual healing using an institutional database and chart review.

Materials and methods

All patients with CD treated with completion proctectomy or proctocolectomy with creation of ileostomy at the Cleveland Clinic between 1995 and 2012 were identified from a prospectively maintained, institutional review board-approved CD database, corroborated by patient charts as necessary. Completion proctectomy was carried out following either a previous ileocolic or ileorectal anastomosis. None of the patients included in the present study underwent ileostomy creation prior to proctectomy or proctocolectomy. Collected data included: baseline demographics; tobacco usage; hemoglobin and serum albumin levels at the time of index surgery, steroid exposure within 1 month of index surgery; exposure to biologic medications within 3 months preceding proctectomy or proctocolectomy, preoperative presence of uncomplicated anoperineal disease, ano/rectovaginal fistula, high fistula-in-ano, or rectal disease; method of anorectal dissection (i.e., intersphincteric, extrasphincteric), and surgical management of the perineal wound (i.e., primary closure of all layers, secondary closure of skin), reoperation due to perineal wound infection or incomplete healing. The technique of perineal wound closure was left to the discretion of the operating surgeon. We also assessed whether adjunctive procedures were performed to promote wound healing, specifically negative pressure wound therapy or myocutaneous flaps either at the time of proctectomy or to treat a chronic perineal wound. Perineal sepsis was defined as bacterial invasion of the soft tissues in the anoperineal area due to anorectal abscess or fistula-in-ano, associated or not with other manifestations of CD including anal fissure/ulcer, anorectal stenosis/stricture, and skin tags. High fistula-in-ano was defined as an anal fistula having a tract directed over the top of the puborectalis muscle, with the internal opening located at the anorectal ring or cephalad to the anorectal ring in the rectum. Perineal wound healing was classified as normal healing (within 3 months), delayed healing (between 3 and 6 months), and non-healing (beyond 6 months); the non-healing group included patients with a chronic presacral sinus.

Statistical analysis

Univariate analysis was used to examine the relationships between the recorded variables and perineal wound healing (normal healing versus delayed healing combined with non-healing; presacral sinus versus non-presacral sinus). Quantitative data were expressed as mean ± standard deviation (SD), and the associations with categorical variables were analyzed by the Fisher exact test or Chi-square test. Risk factors with a univariate p value below 0.10 were subsequently examined by multivariate analysis using forward stepwise logistic regression. A p value of < 0.05 was considered statistically significant.

Results

During the 18-year time interval, 136 patients with CD underwent proctectomy or proctocolectomy with creation of a stoma. Preoperative perineal sepsis was identified in 57 patients, 4 of whom underwent anorectal surgery before proctectomy (incision and drainage of anorectal abscesses in 2 cases, non-cutting seton drainage of anorectal fistulas in another 2 patients). Most patients underwent perineal dissection in the intersphincteric plane followed by complete primary layered closure of the entire perineal wound, while 28% (n = 38) had incomplete perineal wound closure without immediate approximation of their skin margins, which were left to heal by secondary intention (i.e., secondary closure of skin). No patients underwent negative pressure wound therapy or myocutaneous flap at the time of their index surgery. Seventy-two patients achieved normal healing, while 35 experienced delayed healing and 29 experienced non-healing of the perineal wound. Among the 29 patients with a non-healing wound, 9 patients developed a chronic presacral sinus (Table 1). A total of 27 patients (20%) required between 1 and 8 reoperations for perineal wound debridement. Only 1 patient was treated with gluteal thigh myocutaneous flap due to large open perineal wound, rapidly developing 6 days after total proctocolectomy and following an initial examination under anesthesia for debridement of infected and necrotic tissue. Two patients received postoperative negative pressure wound therapy. In one case, an initial examination under anesthesia was carried out following 6 months of unsuccessful nonoperative management of a non-healing perineal wound to accomplish debridement of the chronic granulation tissue followed by placement of negative pressure wound therapy 3 days after the examination under anesthesia while the patient was in the hospital. In the other case, the patient was treated with examination under anesthesia for perineal wound debridement 8 months following total proctocolectomy, at which time negative wound pressure therapy was initiated. The patient ultimately had successful elective perineal wound closure 7 months later.

Table 1 Patient characteristics and disease-related variables (n = 136)
Full size table

Univariate analysis

Univariate analysis of the recorded variables showed that there were two factors associated with a significantly greater risk for delayed healing or a non-healing perineal wound following proctectomy for patients with CD. Firstly, the presence of preoperative perineal sepsis was associated with a significantly increased incidence of delayed healing or a non-healing perineal wound (p = 0.001). Secondly, surgical management of the perineal wound by leaving the skin edges open and employing packing to allow the wound to heal by secondary intention was also associated with significantly poorer perineal wound healing (p = 0.02).

None of the other factors, including age, sex, American Society of Anesthesiologists (ASA) class, body mass index (BMI), hemoglobin and albumin levels, comorbidities (i.e., diabetes mellitus, chronic obstructive pulmonary disease, cardiac disease), surgical procedure, preoperative tobacco usage, preoperative presence of high fistula-in-ano, preoperative steroid or biologic usage, and anorectal dissection technique during the course of proctectomy (intersphincteric vs. extrasphincteric), were associated with delayed healing or a non-healing wound (Table 2).

Table 2 Univariate analysis of factors associated with significantly greater risk for perineal wound healing
Full size table

None of the factors examined, including age, sex, ASA class, BMI, comorbidities, surgical procedure, preoperative tobacco usage, preoperative perineal sepsis, preoperative presence of high fistula-in-ano, preoperative steroid usage within 1 month, and anorectal dissection technique during the course of proctectomy (intersphincteric versus extrasphincteric), were associated with postoperative presacral sinus (Table 3).

Table 3 Univariate analysis of factors associated with significantly greater risk for presacral sinus
Full size table

Multivariate analysis

All factors associated with impaired perineal wound healing (p < 0.10) were included in the multivariate analysis, which identified only 1 factor independently associated with delayed healing and non-healing of the perineal wound after proctectomy, which was the presence of preoperative perineal sepsis at the time of surgery (Table 4).

Table 4 Multivariate analysis of factors associated with significantly greater risk for delayed or non-healing perineal wound
Full size table

Discussion

The present study emphasizes the difficulties in accomplishing healing of the perineal wound after proctectomy for CD, a unique clinical circumstance that cannot be easily compared to perineal healing after abdominoperineal resection for rectal cancer [16,17,18]. Some specific risk factors have been reported in previous series [8, 9, 11, 12, 19], but the most recent analysis focusing on perineal wound healing after proctectomy for CD dates from 1999 [19]. Our study therefore provides novel, comprehensive, contemporary information on this topic.

Our incidence of non-healing perineal wound of 21% is consistent with the rates reported in the literature, which vary from 14 to 50% [20, 21]. Our data did not indicate any association between perineal wound healing and sex, preoperative smoking, or preoperative steroid use, consistent with previous reports [11, 19, 22]. In addition, no significant association existed between age and perineal wound healing, similar to at least one previous report that indicated age greater than 40 years was not associated with worse rates of healing [23]. Other authors have also failed to find any association between BMI and complicated perineal wound healing, although they examined mixed populations including abdominoperineal resection for cancer [18, 24].

Our study also indicates that packing of the wound instead of primary layered perineal wound closure was associated with impaired perineal wound healing, at least on univariate analysis. The most common reason for packing the wound rather than closing all the individual layers was the preoperative presence of perineal sepsis, which may well be the primary factor for poor wound healing. The influence of individual surgeons could affect healing rates, but much of that effect is likely related to the surgical approach (e.g., method of dissection, management of perineal wound) rather than surgeon’s skill.

It is notable that in our patients no association was seen between the method of dissection and perineal wound healing, unlike previous series in which extrasphincteric dissection was significantly associated with poorer wound healing [11, 19]. This might again depend on patient selection because in those series an extrasphincteric dissection was disproportionately associated with perineal sepsis and high fistulas. In our series, an extrasphincteric dissection was infrequently employed, limiting the power of detecting such differences. This disparity is likely due to the characteristics of our institutional practice, which has traditionally used intersphincteric dissection without excision of all external openings of anorectal fistulas provided the causative internal openings are removed during the proctectomy.

The critical factor associated with delayed healing or a non-healing perineal wound in our study was the presence of preoperative perineal sepsis, which concurs with the observations of other authors [9, 20, 23]. It is possible that preoperative perineal sepsis results in more serious inflammation around the anoperineal area and poorer healing ability for this particular tissue. A possible approach to reduce the risk of postoperative complications affecting the healing of the perineal wound could be stoma diversion prior to proctectomy. In this respect, a combined series assessing patients undergoing proctectomy from two centers for a variety of indications reported a perineal wound complication rate of 15% following proctectomy for inflammatory bowel disease [24]. The authors proposed an initial total abdominal colectomy to allow optimization of general health, nutritional status, and discontinuation of medications potentially interfering with perineal wound healing. However, it is important to note that their group with inflammatory bowel disease included a large proportion of patients (43%) with ulcerative colitis. In addition, the study excluded cases in which the perineal wound was intentionally left open to heal by secondary intention. Therefore, the specific data from this particular series are difficult to compare with the current study. Strategies to reduce the risk of perineal wound breakdown might include surgical drainage of sepsis and optimization of medical management. Patients included in the present retrospective study, conducted during a relatively long-time frame, were rarely treated with biologic medications. It is therefore possible that in the future optimal medical management might lead to improved perineal wound outcomes.

Conclusions

The perineal wound resulting from proctectomy or proctocolectomy for CD is frequently associated with poor healing and poses a particular challenge for patients with preoperative perineal sepsis. As preoperative perineal sepsis is the only independent factor associated with impaired healing, these findings support a preoperative discussion that examines potential outcomes and the consideration of measures such as the initial creation of a defunctioning ostomy, control/drainage of local sepsis, and optimization of medical treatment prior to proctectomy in patients with CD considered to be at high risk for perineal wound problems.