Clinical Leakage Research Articles

The ‘modified triple staple’ technique: A variant stapling technique for anastomosis after low anterior resection

Stapled techniques of coloanal anastomosis in anterior resection have gained widespread acceptance over hand anastomosis. We believe a modification of the 'triple staple technique' has ergonomic advantages over existing stapling methods and present our technique and experience here. Fifty consecutive patients underwent anterior resection with a concomitant defunctioning ileostomy in 44 (86%) patients. A modified triple staple technique of side to end coloanal anastomosis was performed without the need of a purse string suture on the proximal and the distal segments. There were no major intra-operative complications. 2/50 (4%) clinical leaks and 2/37 (5.4%) radiological leaks were noted. A combined leak rate of 4/50 (8%) was reported. The incidence of anastomotic stricture encountered was 1/50 (2%). The modified triple staple technique for side to end anastomosis in anterior resection has ergonomic advantages and comparable safety to the existing techniques of stapling coloanal anastomosis. We believe this technique can be widely adopted as an added alternative to the current techniques of stapled anastomosis after anterior resection.

Use of colonic pouches in treatment of mid and low rectal cancer: Better patient selection results in better functional outcome

Background: Addition of a colonic pouch to an ultralow anterior resection or a coloanal anastomosis can improve the functional outcome, but may also cause evacuation difficulties. This study was conducted to evaluate the selective use of colonic pouches, based upon the diameter of the proximal colon. Methods: Twenty-nine rectal cancer patients underwent ultralow anterior resection or coloanal anastomosis, between July 1998 and January 2000. Seventeen had a stapled straight anastomosis, while 12 had a stapled colonic pouch-anal anastomosis. The method of reconstruction was selected intraoperatively based on the diameter of the proximal colon as measured by metal sizers. The colonic J-pouch was constructed with 6-cm limbs using GIA60. Median follow-up was 9 months (range 6- 23 months). Data about complications, tumour characteristics and height of anastomoses were collected. Postoperative bowel function was assessed at 3 and 6 months after surgery or closure of ileostomy using a standard bowel function questionnaire. Results: The 2 groups were comparable regarding age, sex, Dukes stage, resection margins and height of the anastomosis. All anastomoses were 2 - 5.5 cm from the anal verge. The operations were covered by a loop ileostomy in 2 of the straight anastomoses and 4 of the pouches. Two patients had intraoperative complications; splenic injury needing splenectomy and bleeding from pelvic side walls requiring packing. One patient died from multiorgan failure on the 6th postoperative day and was excluded from functional analysis. Four patients had postoperative complications including 2 adhesive bowel obstructions, major wound infection and a subphrenic collection. There were no clinical leaks or pouch-vaginal fistulae. Colonic pouches proved to result in superior functional result in the first 6 months postoperatively compared with straight anastomoses. Pouch patients had less frequency, urgency and need for constipating drugs. None of the pouch patients had evacuation difficulties. Adaptation of straight anastomosis patients resulted in comparable bowel function after 6 months in most patients. Conclusions: Patients with narrow proximal colon, not accepting 31 mm sizer need a colonic pouch added to ultralow or coloanal anstomosis to achieve good postoperative bowel function. Other patients with capacious proximal colon can be expected to achieve equally good functional results after 6 months without the need for a colonic pouch

Factors affecting anastomotic complications following anterior resection in rectal cancer

Anastomotic complications following sphincter saving rectal surgery remains a significant clinical problem in rectal cancer surgery. Preoperative combined modality therapy followed by anterior resection with total mesorectal excision (TME) has become the preferred treatment paradigm for locally advanced rectal cancer. However, its impact on anastomotic complications has not been adequately evaluated. This study aimed to assess the relationship between the response of the primary tumour to neo-adjuvant therapy with anastomotic complications and to evaluate the effect of other clinico-pathological factors previously implicated, in this patient cohort. A total of 119 consecutive patients with primary rectal cancer were assessed of which there were 87 anterior resections. A prospectively collected database was queried to determine the incidence of anastomotic complications, association with response to neo-adjuvant therapy and other clinico-pathological factors. Data were analysed with SPSS 14.0. Anterior resection was performed in 87/111 (78.4%) patients of which 46/84 (56%) were low resections, with an abdominoperineal excision of rectum rate of 11/111 (9.9%). Anastomotic complications were seen as clinical leaks in 10/87 (11.5%) and late colo-visceral fistulae in 2/87 (2.2%) patients. Subclinical/ radiological 'leaks' were quantified as 4/87 (4.5%). A good pathological response to neo-adjuvant therapy was found to be strongly associated with anastomotic complications (P = 0.006). Presence of a perioperative cardiac event was the only other clinical factor associated with anastomotic complications (P = 0.004) in our study. Rectal cancer treated with neo-adjuvant therapy and radical resection with TME for better local control may be associated with greater anastomotic complications particularly when a good pathological response is seen.

Neoadjuvant Chemoradiation Versus Hyperfractionated Accelerated Radiotherapy in Locally Advanced Rectal Cancer

Neoadjuvant therapy is increasingly used in resectable locally advanced rectal cancer. The exact role of the addition of chemotherapy is not established. We compared neoadjuvant therapy using chemoradiation (CRT) or hyperfractionated accelerated radiotherapy (HART). Clinical, pathological, and survival data were obtained from patients with resectable stage II or III rectal cancer within 7 cm from the anal verge. A group of 50 patients was treated with a preoperative dose of 41.6 Gy of radiotherapy (RT) in two daily fractions of 1.6 Gy over 13 days immediately followed by surgery (HART). A second group of 96 patients received 45 Gy of conventionally fractionated RT in 25 daily fractions of 1.8 Gy combined with 5-fluorouracil-based chemotherapy followed by surgery within 4 to 6 weeks (CRT). Both groups were compared in terms of morbidity, pathological downstaging, local recurrence, and survival. Both groups were comparable in terms of preoperative clinicopathological variables. The mean distance from the anal verge was 5.8 cm (HART) versus 4.9 cm (CRT). Sphincter preservation was possible in 74% (HART) versus 83.5% (CRT) of patients (P = .013). The clinical anastomotic leak rate was 2% (HART) versus 2.2% (CRT). Pathological complete response was observed in 4% (HART) versus 18% (CRT) of the resected specimens (P = .002). A pelvic recurrence developed in 6% (HART) versus 4.4% (CRT) of patients (P = .98). Overall 5-year survival was 58% (HART) versus 66% (CRT) (P = .19); disease-free 5-year survival was 51% (HART) versus 62% (CRT) (P = .037). Compared with preoperative HART followed by immediate surgery, preoperative CRT followed by a 6-week waiting period enhances pathological response and increases sphincter preservation rate. This could be explained by the addition of chemotherapy or the longer interval between neoadjuvant therapy and surgery. No statistically significant difference was observed in local control or overall survival.

Stapled diverticulectomy and myotomy for symptomatic Zenker’s diverticulum

Background Few surgeons have extensive experience with the operative management of Zenker’s diverticulum (ZD). Methods Retrospective review of stapled diverticulectomy and cricopharygeal myotomy (SDM) for ZD. All procedures were performed by midlevel surgery residents supervised by single board-certified surgeon. A drain was not used, and patients were liberally advanced to a regular diet postoperatively. Results Three patients with ZD underwent this procedure during a 6-year period. The mean age of the patients was 81 years, and all had multiple comorbidities. Time to discharge was less than 48 hours in 2 patients and 5 days in 1 patient who developed new atrial fibrillation. There were no deaths or clinical leaks. At a mean follow-up of 33 (range 6–62) months, 1 patient died of an unrelated cause, and the other 2 are free of esophageal-related symptoms. Conclusions SDM for ZD is a safe and effective operation that is easily taught to surgical residents.

Clinical and Subclinical Leaks After Low Colorectal Anastomosis: A Clinical and Radiologic Study

This study was designed to examine the natural history of subclinical leaks and their effect on bowel function and quality of life and to evaluate water-soluble contrast enema features that predict anastomotic healing after leaks. Consecutive patients who underwent low rectal anastomosis were followed up postoperatively for leaks. All leaks were confirmed radiologically with CT scanning and water-soluble contrast enema imaging. Water-soluble contrast enemas were serially repeated to identify healing. Characteristics on initial water-soluble contrast enema were correlated with observed healing. Postoperatively, patients were required to fill in a quality of life and a bowel function questionnaire. A total of 138 patients underwent low rectal anastomosis procedures with a median follow-up period of 26 (interquartile range, 19-37) months. There were 23 documented leaks of which 13 (9 percent) presented clinically and 10 (8 percent) presented subclinically. Ileostomy closure was possible in 4 of 13 (30 percent) patients with a clinical leak and all 10 (100 percent) patients with a subclinical leak. Median quality of life scores were lower for patients with clinical leaks and no ileostomy closure (P = 0.03). Bowel function for subclinical leak patients and clinical leak patients with ileostomy closure were similarly impaired. The presence of a cavity (P = 0.01) and a stricture (P = 0.01) at the anastomotic site were unfavorable radiologic features associated with nonhealing. Subclinical leaks are more benign in their natural history compared with clinical leaks. Quality of life and bowel function is no better in patients with a subclinical leak compared with patients with a clinical leak who have ileostomy closure. Anastomotic leaks may resolve if favorable radiologic features are present.

Transabdominal Gastro‐esophageal Devascularization and Esophageal Transection for Bleeding Esophageal Varices after Failed Injection Sclerotherapy: Long‐Term Follow‐Up Report

Management of continued bleeding from esophageal varices despite adequate injection sclerotherapy remains one of the medical and surgical dilemmas. Transabdominal gastroesophageal devascularization and esophageal transection (TGDET) is considered an effective and safe procedure for such patients. This study aimed at presenting continued evaluation of TGDET. Various problems influencing the early outcome are discussed, and long-term outcome is analyzed. This was a prospective clinical descriptive study. Prospective data was collected on 142 consecutive patients managed by one group of surgeons over a 5 year-period and 15 years follow-up after failed injection sclerotherapy for variceal bleeding. Evaluation was made in terms of effectiveness in controlling the acute bleeding, postoperative morbidity and mortality, recurrent bleeding, encephalopathy, and long-term survival. There were 133 men and 9 women. Mean age was 41.8 years. Etiology of portal hypertension was bilharziasis in 54.9% and posthepatitic in 14.8%. Child-Pugh grading on admission was A: 47.2%, B: 28.8%, and C: 14%. Hemorrhage was controlled in all cases. Clinical leak was observed in 5.6%, portal vein thrombosis in 6.3%, and staple line erosion in 2.1% of cases. No patient developed encephalopathy. In-hospital mortality was 12.7%. Complete eradication of varices was observed in 70.6% patients. Recurrent variceal bleeding was noticed in 6.9% of cases. Actuarial 15-year survival for Child-Pugh A patients was 44%, B was 22.5%, and none for C. TGDET remains a safe and effective procedure after failure of sclerotherapy when other alternatives are either not indicated or not available.

The Usefulness of Drain Data to Identify a Clinically Relevant Pancreatic Anastomotic Leak After Pancreaticoduodenectomy?

Pancreatic anastomotic leak (leak) remains a persistent problem after pancreaticoduodenectomy (PD). Recent reports indicate a mean occurrence of 10% with a range of 2%-28% of patients. However, valid comparisons for these studies cannot be made because the definition of leak is variable, and many patients deemed to have a leak are not sick. The aim of this study was to determine the meaning of the volume and amylase content of the effluent from surgical drains by comparing these values to actual clinical outcomes. From January 1996 to July 2002, 207 consecutive patients underwent PD. We considered a leak to be present if greater than 30 ml/day of drainage was observed from drains and if that drainage contained an amylase-rich fluid (greater than 5X serum) on or after postoperative day (POD) 5. Cases were then divided into three groups-no leak, chemical leak only (leak but asymptomatic), and a clinical leak group (leak that required therapeutic intervention, reoperation, readmission, or prolonged length of stay). Then the drainage volume and its amylase concentration for every postoperative day were compared between the three groups. There were no operative or hospital deaths, and the mean length of stay (LOS) was 11.2 +/- 6.1 days. Prolonged LOS was set at greater than 17 days (one standard deviation beyond the mean LOS for all cases). Leak was observed in 14% of cases (n = 29) and the patients were subsequently divided into these groups: no leak (n = 178), chemical leak only (n = 12), and clinical leak (n = 17). Surprisingly, the daily drain amylase values did not differ between the chemical leak group and the clinical leak group. The daily volume of drainage on POD 5-8 for the clinical leak group was significantly greater than the volumes of the other two groups, so that a combination of greater than 200 ml/day of drainage on POD 5 with an amylase greater than 5X serum had a positive predictive value (PPV) of 84% and a negative predictive value (NPV) of 99% for a clinically relevant leak. We used broad criteria from drainage effluent to include as many potential leaks as possible. This broad definition of leak selected 14% of the PD patients as having a leak; within this group, all of the clinical complications of leak occurred. By increasing the volume criteria from greater than 30 ml per day to greater than 200 ml per day, the PPV was increased from 59% to 84% while keeping NPV at 99%. Drain data based on the volume and amylase criteria of this study may be useful for early detection of a leak that will have clinical impact. This study's criteria for leak may be a good definition to design a clinical trial.

A Prospective Study of Bile Leaks after Laparoscopic Cholecystectomy for Acute Cholecystitis

Surgeons are increasingly performing laparoscopic cholecystectomy in the setting of acute cholecystitis. The acutely inflamed gallbladder poses a more technically demanding dissection with potential for an increase in bile leak rates. Clinical and subclinical bile leak rates after laparoscopic and open cholecystectomy in the elective setting are known. This study prospectively evaluates the rate of clinical and subclinical bile leaks after laparoscopic cholecystectomy in the setting of acute cholecystitis. One hundred patients underwent laparoscopic cholecystectomy for acute cholecystitis, as determined intraoperatively and by history, ultrasound, fever, or leukocytosis. On postoperative Day 1, the patients underwent cholescintigraphy (PIPIDA scan) analyzed by a board-certified radiologist for evidence of bile leaks. Postoperative cholescintigraphy revealed eight scans positive for bile leaks. Regardless of scan result, no patient experienced a clinically symptomatic bile leak. Laparoscopic cholecystectomy is a safe and effective treatment for acute cholecystitis with acceptable clinical and subclinical bile leak rates.

Local recurrence after anterior resection for rectal cancer using a double stapling technique.

Fifty-five patients of mean age 69 (range 41-96) years with rectal cancer (Dukes' A:B:C, 11:24:20) underwent anterior resection using a double stapling technique under the care of one consultant surgeon between 1983 and 1988. The mean distance of the anastomosis from the anal margin was 7.2 (range 4-13) cm. The clinical leak rate was 9 per cent (five patients). There were three postoperative deaths from pulmonary embolism, lower limb ischaemia and renal failure. On prospective follow-up, 35 patients had no evidence of local or systemic cancer a median of 32 (range 24-84) months after operation; seven have died from unrelated diseases and ten from metastatic cancer. Pelvic recurrence, in four patients at 9, 11, 12 and 50 months, has occurred only in association with widespread metastasis. These results suggest that the theoretical risks of an increase in the local recurrence rate of rectal cancer after resection using a double stapling technique are not substantiated.

Anterior resection without a defunctioning colostomy: questions of safety.

The need to defunction the anastomosis at anterior resection remains controversial. As the policy in this unit has been not to perform a defunctioning colostomy during anterior resection, the outcome of a consecutive series of 114 anterior resections, all carried out without a covering colostomy, was studied. During the period February 1985 to September 1991, 21 abdominoperineal resections, six Hartmann's procedures and two resections with coloanal anastomosis were also performed. Within the anterior resection group six clinical leaks (5.3 per cent) occurred, all in the low anastomosis group (8 per cent leak rate) and all of which required an end colostomy. The perioperative mortality rate within the anterior resection group was 3.5 per cent; of the four deaths one was attributable to anastomotic dehiscence and sepsis and the others were due to unassociated medical conditions. The results demonstrate similar leakage and mortality rates to published studies where anterior resection is frequently performed with a defunctioning colostomy. These results indicate that the routine use of a defunctioning colostomy at anterior resection should now be questioned.

Intraoperative air testing of colorectal anastomoses: A prospective, randomized trial

A total of 145 consecutive patients receiving a colorectal anastomosis were randomized to 'test' or 'no test' once the anastomosis had been completed. Anastomotic testing was performed with the pelvis filled with saline and the rectum distended by sigmoidoscopic insufflation of air. Any leaks demonstrated were oversewn. A water-soluble contrast enema was performed on the tenth postoperative day. Seventy-four patients were randomized to 'test' and 71 to 'no test' but one patient was withdrawn from each group leaving a total of 143 for analysis. The two groups were well matched for age, sex, diagnosis and operative details. Eighteen (25 per cent) air leaks were detected and repaired in the 'test' group. After operation there were three (4 per cent) clinical leaks in the 'test' group and ten (14 per cent) in the 'no test' group (Fisher's exact test, P = 0.043). There were eight (11 per cent) radiological leaks in the 'test' group and 20 (29 per cent) in the 'no test' group (P = 0.006). Intraoperative air testing and repair of colorectal anastomoses significantly reduces the risk of postoperative clinical and radiological leaks.

A 10-year prospective audit of outcome of surgical treatment for colorectal carcinoma

The outcome of 555 patients who underwent surgery under the care of a surgeon with an interest in colorectal disease was examined prospectively over a 10-year period with no exclusions. There was a 4.7 percent incidence of clinical leaks (10 percent for anterior resection) and an overall corrected 10-year survival rate after curative surgery of 58, 59 and 48 percent for right colonic, left colonic and rectal tumours respectively. The incidence of isolated local recurrence was 8 percent after curative surgery for carcinoma of the rectum. Twenty-five percent of patients with rectal tumours required abdominoperineal excision. Patients who underwent curative abdominoperineal excision of the rectum had corrected 5- and 10-year survival rates of 48 and 36 percent, compared with 60 and 58 percent respectively for curative anterior resection. The perioperative mortality rate of those undergoing palliative surgery was 8 percent, and 75 percent died within 2 years from distant and not local disease. A policy of always attempting resection was validated by the fact that 99 percent of primary tumours were removed with a low perioperative mortality rate (4 percent overall), a high curative resection rate and a low morbidity rate.

Suturing or stapling in gastrointestinal surgery: A prospective randomized study

Abstract A prospective randomized study of 1004 patients is reported, comparing surgical stapling and manual suturing techniques in the construction of gastrointestinal anastomoses. The incidence of clinical leaks was similar between the two groups (sutured 3·2 per cent, stapled 4·7 per cent; P = 0·22), while for radiological leaks the incidence was significantly higher in the sutured group (12·2 per cent versus 4·1 per cent, P<0·05). Stapling instruments afforded significantly quicker anastomoses, mean(s.e.m.) 28·1(0·7) versus 14·3(0·5) min (P<0·001), and quicker operations, 115·5(2·4) versus 103·8(2·2) min (P < 0·001). The two groups were found to be comparable in other respects such as operative mortality, requirements for blood transfusion, incidence of infective complications, recovery of gastrointestinal function and postoperative hospital stay.

Jejunal pouch reconstruction after total gastrectomy for cancer: Experience in 29 patients

The experience with interposed jejunal pouch after total gastrectomy for cancer in 29 patients is reviewed. There were two postoperative deaths (7 per cent) due, respectively, to myocardial infarction on day 5 and massive haemorrhage from the splenic vein on day 14. Anastomotic leaks, all from the proximal anastomosis, were encountered in three patients (11 per cent). These consisted of one minor clinical leak in a stapled anastomosis and two radiological leaks from hand-sutured anastomoses. Obstruction of the pouch or its conduit by recurrent tumour was not observed in any patient, including those who died from metastatic disease. There was one instance of benign stenosis of the distal anastomosis to the duodenum which required revisional surgery. The symptomatic assessment in the long-term surviving patients was good except for mild oesophagitis due to bile reflux in five patients. Body-weight was maintained and adverse nutritional consequences were not observed.

Carcinoma of the rectum: A 10-year experience

A consecutive series of 303 patients with carcinoma of the rectum and distal sigmoid colon treated by a single surgeon over a 10-year period are reported. Of these, 202 underwent an anterior resection, 85 an abdominoperineal excision of the rectum and 16 a coloanal anastomosis. Surgery was considered palliative in 52 patients undergoing anterior resection and 24 undergoing abdominoperineal resection. The 30-day hospital mortality rate was six patients (3 per cent) for anterior resection and two patients (2 per cent) for abdominoperineal resection. Peroperative anastomotic testing demonstrated leakage in five stapled anastomoses; these were rectified and no clinical sequelae occurred. Two patients (1 per cent) developed a clinical anastomotic leak, one of which proved fatal; in each case the intraoperative test was negative. The overall 5-year survival rate was 64 per cent after anterior resection and 52 per cent after abdominoperineal resection; the median follow-up was 64 months. The incidence of local pelvic recurrence was 6.4 per cent after anterior resection and 14 per cent after abdominoperineal (not significant). These results confirm the success of sphincter-saving anterior resection combined with total mesorectal excision, routine full mobilization of the splenic flexure and cancercidal lavage of the distal rectum in the treatment of low rectal carcinomas; morbidity, local recurrence and survival are not compromised.

Abdominoperineal and anterior resection of the rectum with retrocolic omentoplasty and no drainage.

Forty consecutive patients scheduled for either low anterior resection or abdominoperineal resection of the rectum have been studied. After standard procedures, carried out by all grades of surgeons, the omentum was mobilized as a pedicle graft based on the left gastro-epiploic arcade and transposed to the pelvis or perineum. All wound were closed primarily without drainage. Twenty-six patients had anterior resection, 11 abdominoperineal resection and 3 an extended Hartmann's operation. Patients were assessed clinically and radiographically by Gastrografin enema and ultrasound for evidence of anastomotic leakage, pelvic collections and wound healing. After anterior resection there were three radiological leaks and no overt clinical leaks. One patient had a collection related to a leak. Of 11 perineal wound 8 healed primarily in 2 weeks and all healed by 4 weeks. There was no frank wound breakdown. The three patients who had Hartmann's operations all healed without complication. There were no unavoidable complications of the procedure. Retrocolic omentoplasty is a simple, safe and effective adjunct to rectal surgery. The additional time taken (15-20 min) is well rewarded by sound healing of these operations.

Anastomotic leak following mesorectal excision for rectal cancer: true incidence and diagnostic challenges

Anastomotic leakage is a potentially serious complication of low anterior resection which may be accompanied by clinical symptoms (clinical leak) or may be silent (subclinical leak). In this study the true incidence of the complication was evaluated, and the diagnostic accuracy of clinical symptoms, conventional rectal radiography (CRR) and computed tomography (CT) was compared. Fifty-six consecutive patients were included in a prospective trial. Clinical parameters were recorded and CRR and CT performed 6-10 days postoperatively or earlier if a leak was suspected. Endoscopy was performed three months postoperatively. Based on all available information including late endoscopy, 5 (9%) patients had clinical leak and five a leak that was asymptomatic during the hospital stay. Clinical assessment, CRR and CT during the hospital stay had an accuracy of 82%, 93% and 94%, respectively, and a sensitivity of 50%, 60% and 57%, respectively. The specificity of clinical assessment was 89%, whereas both CRR and CT had a specificity of 100%. The incidence of anastomotic leakage seemed acceptable when compared with other series. Fifty per cent of the leaks were silent. CRR and CT may be false negative and immediate treatment should be started if clinical signs are highly suggestive of leak, irrespective of radiological findings CT was not more accurate than CRR in detecting anastomotic leak.

Safety and Efficacy of the Use of Bioabsorbable Seamguard in Colorectal Surgery at the Texas Endosurgery Institute

Bioabsorbable Seamguard (BSG) is a random-fiber web of polyglycolic acid/trimethylene carbonate. It is completely absorbed within 6 months or less due to its constitution of a bioabsorbable membrane with polyester braided suture. It has been used in obesity surgery and pulmonary surgery as staple-line reinforcement with good results. As such, we believe that BSG may be ideal to use in colorectal surgery as an aid during the healing process of an anastomosis and may help prevent anastomotic bleeding and staple-line disruption. From July 2003 through September 2004, 30 patients underwent placement of BSG for the following procedures: 12 right hemicolectomies, 7 low anterior resections, 5 sigmoid colectomies, 3 total colectomies, 2 partial resections, and 1 colostomy closure. Median follow-up was 7 months (range 1-13). There were no clinical leaks, no strictures, and no bleeding in our early postoperative follow-up period. The use of BSG as a staple-line reinforcer appears to be safe and may be useful in preventing anastomotic leakage, bleeding, and intraluminal stenosis.

Influence of a defunctioning stoma on leakage rates after low colorectal anastomosis and colonic J pouch-anal anastomosis.

Anal sphincter function is increasingly preserved following rectal excision for cancer and provides a better quality of life for patients than does a permanent colostomy. However, anastomotic complications may cause considerable morbidity and mortality. This retrospective study examined the incidence of anastomotic complications following two forms of reconstruction after resection for mid-rectal cancer: colonic pouch-anal anastomosis (CPAA) and low colorectal anastomosis (LCRA). Some 258 consecutive patients with mid-rectal cancers between 6 and 11 cm from the anal verge underwent proctectomy with mesorectal excision and either CPAA or LCRA. The incidence of clinical and radiological leaks was determined in these patients who were considered in three groups: LCRA (defunctioning stoma), LCRA (no defunctioning stoma) and CPAA (all defunctioned). In the LCRA group without a defunctioning stoma, a clinical leak occurred in 17.0 per cent, compared with two of 30 in the LCRA group with a defunctioning stoma. In the CPAA group a clinical leak occurred in 4.9 per cent of patients, which was not significantly different from the rate in those with a defunctioned LCRA. Patients with a non-defunctioned LCRA were more likely to suffer a clinical anastomotic leak (P=0.01), peritonitis (P=0.001) and require unscheduled reoperation (P=0.006) than those with a defunctioned LCRA and/or CPAA. The use of a protective defunctioning stoma is advocated in conjunction with LCRAs.