Risk factor for permanent stoma and incontinence quality of life after sphincter‐preserving surgery for low rectal cancer without a diverting stoma

Abstract The goal of the present study was to evaluate permanent stoma formation and defecation function in long‐term follow up after surgery for low rectal cancer without a diverting stoma. Subjects were 275 patients who underwent sphincter‐preserving surgery for low rectal cancer between 2000 and 2012. Clinical outcomes were evaluated and defecation function was assessed based on a questionnaire survey, using Wexner and modified fecal incontinence quality of life (mFIQL) scores. Incidence of anastomotic leakage was 21.8%, and surgery‐related death as a result of anastomotic leakage occurred in one male patient. Median follow‐up period was 4.9 years and permanent stoma formation rate was 16.7%. Anastomotic leakage was an independent predictor of permanent stoma formation (odds ratio [OR] 5.86, P<0.001). Age <65 years (OR 1.99, P=0.001) and male gender (OR 4.36, P=0.026) were independent predictors of anastomotic leakage. A permanent stoma was formed as a result of poor healing of anastomotic leakage in 29.6% of males, but in no females. Defecation function was surveyed in 27 and 116 patients with and without anastomotic leakage, respectively. These groups had no significant differences in median follow‐up period (63.5 vs 63 months), Wexner scores (quartile) (6 (2.5‐9) vs 6 (3‐11)), and mFIQL scores (26.1 (4.8‐64.2) vs 23.8 (5.9‐60.7). Defecation function associated with anastomotic leakage showed no significant dependence on gender or resection procedure. Sphincter‐preserving surgery without a diverting stoma may be indicated for females with low rectal cancer. In this procedure, male gender is a risk factor for anastomotic leakage and subsequent formation of a permanent stoma in one in three patients.


| INTRODUCTION
In sphincter-preserving surgery for low rectal cancer, a diverting stoma is concomitantly formed with the aim of resting the anastomosis region until it heals. 1,2 Diverting stoma formation is recommended based on a meta-analysis showing that this procedure reduced anastomotic leakage after low anastomosis close to the anus. 3 However, in a multicenter study in Japan, prevention of anastomotic leakage by a diverting stoma after low anastomosis following rectal cancer resection was not found. 4 Anal function is retained without a diverting stoma in some cases, and such patients thus undergo unnecessary stoma formation. 5,6 Our department has carried out sphincter-preserving surgery without a diverting stoma after low anterior resection (LAR) and intersphincteric resection (ISR) for low rectal cancer as a basic treatment strategy. 6,7 The objective of this retrospective study was to investigate safety, permanent stoma formation, and defecation function in patients who underwent this procedure for low rectal cancer, and to clarify the validity and indication for this treatment.

| Patients
Of 370 consecutive patients with low rectal adenocarcinoma who underwent initial proctectomy at the Department of Gastroenterological Surgery, Hirosaki University, between 2000 and 2012, 298 received sphincter-preserving surgery. Subjects of the present study were 275 of these patients, excluding one case with concomitant ulcerative colitis and 22 patients in whom a diverting stoma was formed after preoperative radiotherapy. Rectal cancer in which the lower margin was located below the peritoneal reflection during surgery was defined as low rectal cancer. Data for anastomotic leakage and perioperative complications, permanent stoma formation, and reasons for the procedure were collected from medical records. Perioperative complications were defined using the Clavien-Dindo classification. 8 Clinical leakage signs were defined as abdominal pain, abdominal distention, fever, and pus or fecal discharge from the pelvic drain. All clinically suspicious symptoms were confirmed by digital rectal examination and radiographic examination (e.g. extravasation of endoluminally given water-soluble contrast enema, pelvic abscess and fluid/air bubbles surrounding the anastomosis on computed tomography). 7 Using the proposed grading system, anastomotic leakage was classified into three grades: grade A required no active therapeutic intervention; grade B required active therapeutic intervention; and grade C required reoperation. 9 Anastomotic leakage with grades B and C (but not grade A) within 30 days after surgery was defined as anastomotic leakage. Age, sex, body mass index (BMI), ischemic disease, diabetes, American Society of Anesthesiologists (ASA) status, intraoperative blood transfusion, tumor diameter, tumor-anal verge distance, anastomotic height from anal verge, circumferential occupation, tumor depth, regional lymph node metastasis, distant metastasis, circumferential margin (CRM), operation time, blood loss, laparoscopy, combined resection, lateral lymph node dissection (LLND), resection procedure, and anastomosis method were examined as clinicopathological factors. When a stoma was present at final follow up, it was regarded as a permanent stoma. 10 Median follow-up period was 4.9 years.

| Operative and perioperative management
In standard perioperative management, the patient fasted from the day before surgery, and received mechanical pretreatment and perioperative antibiotics before surgery and for 3 days after surgery. After pressure reduction by transanal drainage for about 1 week after surgery, food ingestion was started. After transection of the inferior mesenteric artery and vein, total mesorectal excision (TME) was carried out as a standard surgical procedure, and bilateral LLND was done when the depth was T3 or deeper, as a rule. 11 To secure a 2cm resection margin, ISR was selected for tumors located within 2 cm from the upper margin of the levator ani muscle attachment region. 7,8,12 For anastomosis in LAR and ISR, double-stapled and hand-sewn coloanal anastomosis were carried out, respectively. Sideto-end anastomosis was applied as a rule, and end-to-end anastomosis was used when the pelvis was narrow or the reconstructed intestine was short. When anastomotic leakage was clinically suspected after surgery, its presence or absence was confirmed by fluoroscopy or computed tomography (CT). If anastomotic leakage was observed, it was treated with antibiotics, a drainage tube, or stoma formation, depending on the details in each case. In approximately 6 months after stoma formation as a result of leakage, integrity of the anastomosis was checked by digital rectal examination and a water-soluble contrast enema examination. Patients without any findings of anastomotic leakage underwent stoma closure. When findings of anastomotic leakage were sustained and the anastomosis was not expected to heal, stoma was not closed or permanent colostomy was formed.
When patients had poor general condition such as unresectable distant metastases or dementia, stoma was not closed permanently even if integrity of the anastomosis was recovered.

| Evaluation of function
Defecation function and quality of life (QOL) were surveyed using a questionnaire in patients who did and did not develop anastomotic leakage, and evaluated based on the frequency of defecation per day and the Wexner Score 13 and modified fecal incontinence quality of life (mFIQL) score. 14

| Statistical analyses
Risk factors for permanent stoma formation and for anastomotic leakage were analyzed by Fisher exact test. Factors with a significant difference were subjected to multivariate logistic regression analysis.
Defecation function was compared between groups by Mann-Whitney U-test. Two-sided P<0.05 was regarded as significant. Statistical analysis was carried out using EZR. 15 3 | RESULTS and the overall incidence was 21.8% (60/275) ( Table 1).

| Risk factors for anastomotic leakage
Age <65 years old and male gender were significant risk factors for anastomotic leakage in univariate analysis, and were also independent risk factors in multivariate analysis (age <65 years old: OR=1.99, P=0.001; male: OR=4.36, P=0.026) ( Table 3) (Table 4).  (Table 5). Gender and resection procedure had no significant effect on anastomotic leakage-associated defecation function or QOL (Table 5).

| DISCUSSION
A diverting stoma may contribute to prevention of anastomotic leakage in cases with low anastomosis near the anus, and is generally formed in anus-preserving surgery for low rectal cancer. [1][2][3] However, in a recent multicenter study in Japan, a diverting stoma did not reduce the incidence of anastomotic leakage after low anastomosis, but did significantly reduce the rate of reoperation after anastomotic leakage. 4 There was no difference in mortality between patients with and without diverting stoma formation, 4 suggesting that a diverting stoma is unnecessary if anastomotic leakage is treated appropriately, including with reoperation.
As complications associated with diverting stoma formation and stoma closure may develop, this procedure is not necessarily a safe intervention. 16 Therefore, if a diverting stoma does not reduce anastomotic leakage-and surgery-related deaths, patients who are unlikely to develop anastomotic leakage undergo an unnecessary and risky procedure. Therefore, the significance of a diverting stoma requires investigation, including the rate of permanent stoma formation and defecation function. The significance of the present study is that the indication for diverting stoma was investigated based on long-term anal conditions, defecation function, and QOL as outcomes.
One of the main goals was to identify the perioperative risk factors for permanent stoma in all aspects in consecutive patients with low rectal cancer at a tertiary hospital. We considered that this analysis could offer valuable overview of the consequences after sphincter-preserving surgery for low rectal cancer without a diverting stoma to patients and physicians. We first identified anastomotic leakage as a risk factor for permanent stoma formation, as previously found. [17][18][19][20][21] Age <65 years and male gender were then identified as independent risk factors for anastomotic leakage. A permanent stoma was formed as a result of poor healing of anastomotic leakage in one in three males, regardless of age and resection procedure. In contrast, in females, the incidence of anastomotic leakage was low and there was no permanent stoma formation as a result of poor healing of anastomotic leakage.
Reduction of defecation function is of concern when anastomotic leakage occurs, but findings have varied among previous studies. [22][23][24] In our patients, anastomotic leakage did not contribute to reduction of long-term defecation function and QOL, and there was no influence of sex or resection procedure. Thus, in female patients, sphincter-preserving surgery for low rectal cancer without diverting stoma formation is unlikely to have a negative influence on permanent stoma formation as a result of poor healing of anastomotic leakage and reduced long-term defecation function and QOL. In a study of short-term anastomotic leakage following diverting stoma formation after low anastomosis, a diverting stoma was found to be useful in males, but not in females. 25 The long-term anal function in our study supports the validity of low anastomosis without diverting stoma formation after rectal resection in female patients. As there is no clear basis for expecting long-term improvement of defecation function and QOL by accepting reduction of QOL as a result of diverting stoma formation, a diverting stoma should be formed only in cases in which it is likely to be beneficial.
The incidence of anastomotic leakage was high in our patients, but mortality was low, suggesting that appropriate treatment was carried out. However, strategies to prevent anastomotic leakage are necessary. 26 Many factors influence anastomotic leakage, but male gender is a common risk factor in many reports, but no specific countermeasures for males have been developed. 2 ASA, American Society of Anesthesiologists; AV, distance from anal verge to tumor; CI, confidence interval; CRM, circumferential resection margin; ISR, intersphincteric resection; LAR, low anterior resection. Second, only about 50% of patients responded to the questionnaire on evaluation of anal function, and the presence of a bias as a result ASA, American Society of Anesthesiologists; AV, distance from anal verge to tumor; CI, confidence interval; CRM, circumferential resection margin; ISR, intersphincteric resection; LAR, low anterior resection.
of non-respondents cannot be ruled out. Third, the subjects were patients who did not receive preoperative treatment. Such treatment for low rectal cancer is not specified as a standard approach in current Japanese guidelines, but preoperative chemoradiotherapy is standard treatment in Western countries. 35 Our institution has formed a diverting stoma in patients treated with preoperative radiotherapy based on poor healing of the anastomosis region, and a diverting stoma may be significant for prevention of anastomotic leakage in female patients treated with preoperative radiotherapy. 36 In contrast, favorable local outcomes have been reported in Europe for rectal cancer treated with surgery alone without preoperative treatment, with selection of patients based on preoperative highresolution MRI. 37 Patients similar to the subjects in the current study are likely to increase worldwide, and thus our results may be significant, despite the above limitations.

| CONCLUSION S
Sphincter-preserving surgery for low rectal cancer without diverting stoma formation may be indicated for female patients. Male gender was a risk factor for anastomotic leakage in this procedure, with a permanent stoma as a result of anastomotic leakage formed in one in three male patients.

The authors thank Motoi Koyama, Asuka Kamimura and Chiharu
Yamada for their support of data collection.

DISCLOSURE
The protocol for this research project has been approved by a suit-