Introduction

Diverting loop ileostomy is commonly performed to protect a distal anastomosis when there is a high risk of anastomotic leakage¹,².

Although fecal diversion does not prevent an anastomotic leak, it does decrease the potential morbidity and mortality from an anastomotic leak³. Despite the potential benefits of fecal diversion, loop ileostomy construction is related to a significant rate of complications.

Closure of the loop ileostomy is associated with a low mortality, but morbidity may be as high as 20%⁴. Patients needed additional operation (closure surgery) and may develop small bowel obstructions, parastomal hernia, stoma retraction or prolapse, skin problems, electrolyte abnormalities, and acute kidney injury due to a high output stoma³,⁵.

Furthermore, stoma reversal may cause several complications, such as surgical site infection (SSI), anastomotic leak, bowel obstruction, post-operative ileus, wound dehiscence, and enterocutaneous fistulas⁶,⁷. There is also an increased risk of developing an incisional hernia after stoma closure⁵.

The aim of the study was to determine the risk factors for complications after loop ileostomy closure using multivariate analysis, either if the first procedure was for a benign or a malignant colorectal disease.

Patients and methods

From January 2010 to March 2018, all patients who underwent loop ileostomy closure at Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán in Mexico City, Mexico, were included in this study. Data for these patients were retrospectively retrieved and analyzed.

The inclusion criteria included patients who were ≥ 18 years old and who underwent planned loop ileostomy closure. Patients included had previous benign or malignant colorectal disease. Index surgery was defined as the operation that included colorectal resection (for benign or malignant disease), primary anastomosis, and construction of diverting loop ileostomy. The indications for protective loop ileostomy construction were low colorectal, ileorectal, coloanal, or ileal pouch anastomoses.

Exclusion criteria were patients younger than 18, patients who had diverting loop ileostomy non-related to colorectal diseases or operations, and patients with double-barrel or terminal ileostomy. Patients with incomplete data were eliminated from the study.

All patients included in the study had pre-operative water-soluble contrast enema and colonoscopic evaluation before the loop ileostomy closure. All anastomoses were side-to-side and were performed with linear staplers and reinforced with non-absorbable sutures.

Ileostomy closure was performed either by peristomal incision (circular incision around the ileostomy) or by a midline incision as required. All patients had a peristomal incision first approach, and some patients needed conversion to midline incision mainly due to dense adhesions or inadequate small bowel length obtained through peristomal incision. Closure of the ostomy wound was performed by conventional linear closure or by pursestring closure, based on surgeons' preference.

Pre-operative variables considered in this study were sex (female and male), body mass index (< 30 or ≥ 30), age (< 65 or ≥ 65), and the American Society of Anesthesiology grade (ASA I-II or III-IV). Comorbidity was assessed using the Charlson Comorbidity Index (CCI)⁸. Other pre-operative variables were tobacco use and previous treatments (systemic steroids, immunomodulators, radiotherapy, and chemotherapy). Immunomodulators included azathioprine, methotrexate, and mycophenolate mofetil. Either immunomodulators or systemic steroids were withdrawn before the loop ileostomy closure. Variables regarding index surgery were type of index surgery (elective or urgent and laparoscopic or open) and reintervention for complications after index surgery. Surgical variables concerning loop ileostomy closure were interval from index surgery to closure (months), type of incision (peristomal or midline), and peristomal wound closure (linear or purse-string).

The following complications after loop ileostomy closure were analyzed: wound infection either deep or superficial, anastomotic leakage, intestinal obstruction, reintervention after loop ileostomy closure, overall medical complications, and mortality. All post-operative complications were recorded and stratified in accordance with Clavien-Dindo classification scheme (occurring either within 30 days of loop ileostomy closure or during the hospital stay)⁹. The rate of incisional hernia was also analyzed in the study.

Wound infections were either superficial or deep infections, based on the definition of the Centers for Disease Control for Surgical Site Infections¹⁰. Anastomotic leakage was confirmed with clinical and tomographic evidence of leak or perianastomotic collection. Intestinal obstruction was diagnosed in patients with clinical symptoms and signs of obstruction (distention, pain, vomiting, and absence of gas/stool passage) and corroborated by imaging (dilated small bowel). Reintervention was the need to return to the operation room secondary to a complication directly related to the loop ileostomy closure and included all complications occurring within 30 days after loop ileostomy closure.

Incisional hernia repair was not considered in the reintervention group and was not included in the Clavien-Dindo classification. Incisional hernia was diagnosed by clinical examination and corroborated by computed tomography. For incisional hernia, a period of 1 year to follow-up after ileostomy closure was considered in the analysis.

Results

From January 2010 to March 2018, a total of 136 patients underwent loop ileostomy closure at our institution. Patient demographic and clinical characteristics are summarized in table 1. There were 69 women (50.7%), 71.3% (n = 97) of patients were < 65 years old, 91.9% (n = 125) had a BMI < 30 kg/m², and 58.1% (n = 79) had a CCI ≥ 2.

Of the 136 patients, 82 (60.3%) had benign disease and 54 (39.7%) had malignant colorectal disease.

Patients in the benign disease group had one of the following diagnosis: complicated diverticular disease (n = 35), ulcerative colitis (n = 11), familial adenomatous polyposis (FAP) syndrome (n = 10), rectovesical fistula (n = 8), rectovaginal fistula (n = 8), colorectal trauma (n = 5), Hirschsprung's disease (n = 1), ischemic colitis (n = 2), colonic angiodysplasia (n = 2), rectal prolapse (n = 1), and obstructing rectal adenoma (n = 2).

Patients with malignant disease had rectal cancer (n = 40) or colon cancer (n = 14). The distribution of patients according to cancer stage (TNM) was¹¹ Stage I (n = 7, 12.9%), II (n = 8, 14.8%), III (n = 28, 51.8%), and IV (n = 11, 20.3%).

Comparisons of baseline pre-operative data, previous treatments, characteristics of the index surgery, and perioperative data concerning loop ileostomy closure are shown in table 1.

Discussion

There is a trend toward sphincter-sparing procedures with the use of very low pelvic anastomoses in either rectal cancer or benign conditions (diverticulitis, FAP, or ulcerative colitis)¹².

The most dreaded complication of a low (pelvic) anastomosis is an anastomotic leak. Anastomotic leakage could result in generalized peritonitis or pelvic abscess, long length of hospital stay, decreased quality of life, cancer recurrence, and higher mortality¹,¹³,¹⁴. Fecal diversion is aimed to minimize these complications¹,¹³.

The incidence of anastomotic leakage is higher if it is distal in the rectum or if there are predisposing factors such as immunosuppression, history of chemo- and radiotherapy, acute sepsis, and poor nutritional status¹,³. Other factors related to leaks are smoking, obesity, transfusion, and ASA class III-IV³,¹⁵. The incidence of anastomotic leak after colorectal surgery ranges from 1% to 25%¹²,¹⁶ and the mortality ranges from 6% to 22%¹⁷.

The absence of a diverting stoma in low anterior resection is associated with significant higher rate or anastomotic leak and reintervention¹⁸. For ultralow resections with subsequent coloanal anastomosis, diverting ostomy is almost always constructed³.

Despite the advantages of loop ileostomy, the role of temporary fecal diversion remains an area of controversy among surgeons³. Controversy rises because the significant morbidity and morality related to ileostomy (high output ileostomy, acute renal failure, parastomal herniation) and to ileostomy reversal (surgical site infections, anastomotic leak). Furthermore, it has been estimated that 40-50% of temporary stomas are never reversed¹².

Morbidity and mortality related to loop ileostomy closure had been described in several series as being 3-30%¹,¹⁹ and 0-4%¹,²⁰, respectively.

In a large cohort of patients (from the NSQIP database) who underwent closure of ileostomy, 9.3% of patients had major complications, 8.4% had minor complications, and there was a mortality rate of 0.6%¹. Functional status of the patient, ASA class, and organ dysfunction were independent predictors of complications after ileostomy reversal¹.

In other study, elderly patient with age > 80 years (OR = 4.3 [1.6-11.0]; p = 0.003) was the only independent risk factor for post-operative complications on multivariate analysis¹³.

Mennigen et al.⁷, in a systematic review of risks of ileostomy closure after restorative proctocolectomy for ulcerative colitis and FAP (n = 2146), reported that an overall morbidity was 16.5%, reoperation rate was 3%, anastomotic leak was 2.0%, bowel obstruction was 7.6% (with 2.9% requiring laparotomy for this), wound infection rate was 4.0%, and stoma site hernias was 1.9%.

On regard of the anastomotic leak, in a recent meta-analysis, there was no difference in the rate of anastomotic leak between the hand-sewn and the stapled anastomotic techniques for closure of a loop ileostomy (OR 0.81, 95% CI 0.43-1.54, p = 0.52, I²= 33%)²¹. Saha et al.²² reported a 4% rate of anastomotic leak, with pre-operative anemia being significantly associated with leakage (p = 0.033).

Man et al.¹³ determined that closure of the ileostomy with hand-sewn techniques showed a higher incidence of post-operative intestinal obstruction (p = 0.049) compared to closure using stapler. This finding was corroborated in a meta-analysis, being higher the rate of small-bowel obstruction in the hand-sewn group compared to the stapled group (7.03% vs. 5.58%; OR 0.69, 95% CI 0.51-0.92, p = 0.01, I² = 0%)²¹.

Surgical site infection following stoma reversal increases the risk of wound dehiscence, incisional hernia, length of stay, and health-care costs⁶. In a recent meta-analysis (including ileostomy and colostomy reversal), pursestring closure had a significant decrease in SSI (risk difference, −0.25; 95% CI, −0.36−0.15; p < 0.00001; number needed to treat = 4) and higher satisfaction with cosmetic outcomes (standard mean difference, 0.7; 95% CI, 0.13-1.27; p = 0.02), when compared with conventional linear closure⁶. Another study reported BMI ≥ 30 (OR = 1.4), ASA III-IV (OR = 1.7), and operating time ≥ 100 min (OR = 3.3) as factors associated with incisional infection¹. We identified in the univariate analysis of our study several factors related with increased SSI included linear wound closure. However, only patients who needed conversion to midline laparotomy were associated with increased likelihood of SSI (OR = 6.5). The rate of conversion to midline incision (9.6%) seems to be high in our study; the main reason for conversion was dense adhesion needing dissection for adequate tension-free anastomosis.

Mansfield et al.²⁰ reported a 7.3% reoperation rate after ileostomy closure. Schneider et al.²³ reported an immediate reoperation rate after ileostomy reversal of 5.3%, identifying higher BMI and anemia as the risk factors (BMI: p = 0.038; OR 0.73 [95% CI 0.55-0.98]; anemia: p = 0.001; OR 25.50 [95% CI 3.87-168.21]). Our reoperation rate was 7.3%, and previous use of immunomodulator and performing ileostomy closure through midline incision were found to be independent risk factors.

Medical complications should also be considered after closing a loop ileostomy. In a study including 5401 patients, the frequency of complications was pneumonia (0.9%), cardiac (0.4%), renal failure (0.6%), neurological (0.1%), venous thromboembolism (0.6%), and urinary tract infections (2%)¹. In one study¹³, age > 80 years was an independent risk factor for developing urinary retention (OR = 5.6 [1.8-17.4]; p = 0.001). We found a 9.5% rate of medical complications, including urinary tract infection (1.5%), pneumonia (1.5%), and delirium (1.5%). Age ≥ 65 years (OR = 3.5) was an independent risk factor associated with medical complications.

One of the most common late complications of stoma reversal is the development of an incisional hernia, with overall incisional hernia rate of 7.4%, ranging from 0% to 48%⁵,²⁴,²⁵. Recognized risk factors associated with incisional hernia are age, medical comorbidities (especially, obesity and hypertension), and smoking, as well as the type of suture used to close and SSI²⁴. We found, in our study, an incisional hernia rate of 11% that is concordant with the incidence reported in previous studies, and two independent risk factors were found to be associated with an increased rate of herniation: time to ileostomy closure < 3 months (OR = 6.4) and presence of parastomal hernia (OR = 13.2). Incisional hernia can have an adverse effect on a patients' quality of life and increases health-care costs⁵. In one study, hernias are detected in a median of 6 months after stoma closure²⁶. Around one-third to one-half of hernias detected clinically required repair⁵. Prophylactic mesh placement has been advocated for preventing parastomal hernias; however, it has not been widely used in stoma closure reinforcement due to the risk of contamination⁵. In a recent publication, Warren et al.²⁵ advocated the retromuscular placement of permanent synthetic mesh at the time of stoma reversal as an effective mean to prevent the development of incisional hernia.

The limitations of our study are largely attributable to the sample size and the retrospective design. The single institution nature of our investigation is prone to selection bias. The sample size could represent a risk of bias for a multivariate analysis, which could led to the underestimation of the independent variables. However, we consider that this study provides evidence of different risk factors associated with complications that should be considered at the time of loop ileostomy construction and closure either if the first procedure was for a benign or malignant colorectal disease. These data can be utilized by colorectal and general surgeons for surgical decision-making and counseling patients about the risks and complications of having a loop ileostomy.

Conclusions

Loop ileostomy closure is associated with an important risk of complications either if the first procedure was for a benign or malignant colorectal disease. As loop ileostomy is still recommended for protection of high-risk colorectal anastomosis, several patient-related factors as well as surgical technique elements should be considered at the time of ileostomy closure to reduce post-operative morbidity.