Despite several development in operative methods and instruments for colorectal patients, the incidence of anastomotic leak (AL) has not decreased [1]. Furthermore, although total mesorectal excision in rectal cancer surgery has improved the oncologic outcome, it has raised the risk of AL [2].

AL shows a 20%-30% morbidity rate and 7%-12% mortality rate in colorectal surgery [3,4,5]. Regarding oncologic outcomes, AL raises locoregional recurrence [5,6], and reduces long-term survival rate [3,7].

A diverting stoma is known to mitigate life threatening complications of AL, such as fecal peritonitis or sepsis, though it does not reduce clinical leak rate [8,9].

However, a diverting stoma does have its own complications. A previous study reported a morbidity rate of 34% for loop colostomy and 19% for loop ileostomy [10]. Moreover, stoma repair still requires additional operation, which is accompanied by the risk of operative complications.

The transanal tube (TAT) is known to protect the anastomotic site, but few studies have reported the efficacy and safety of the TAT for preventing AL in rectal cancer surgery. Here, we report our experiences of TAT usage. Clinical results of the TAT group are reviewed in comparison to the ileostomy and control (no other protection for anastomosis) groups.

In total, 578 consecutive surgical patients were diagnosed with primary rectal cancer during the study period. According to the inclusion criteria, 137 patients were enrolled as follows: 67, 35, and 35 patients in the ileostomy, TAT, and no-protection groups, respectively (Table 1). The average level of tumor was 8.5 ± 1.5 cm (range, 5-10 cm) above the anal verge, and there were no significant differences in age, gender, tumor level, history of previous abdominal surgery, preoperative serum hemoglobin and albumin levels, operator, anastomosis type, and surgical procedure type among the groups. BMI of the no-protection group was statisctically lower than that of the ileostomy group or TAT group (P = 0.028), but intergroup analysis showed no statistical differences between the ileostomy and TAT groups.

Operative and pathologic outcomes (Table 2) showed no significant difference in concomitant resection of other intra-abdominal organs, intraoperative transfusion, TNM stage, tumor size and distal resection margin. Most patients (122/137, 89.1%) received laparoscopic operations. However, operation time was longer in the ileostomy group (286.8 ± 66.2 > 238.9 ± 71.6, 256.0 ± 61.5, P = 0.029), and more estimated blood loss was observed (418.0 ± 355.1 > 244.3 ± 178.5, 346.7 ± 225.7, P = 0.018).

Regarding the postoperative course, patients in the ileostomy group resumed their diets more than 1 day earlier than the other groups (P<0.001) (Table 3). Postoperative hospital stay was about 1 to 2 days longer for the no-protection group (P = 0.048), and no difference was observed between the ileostomy and TAT groups. Stoma closure was performed in 63 of 67 patients (94.0%) in the ileostomy group. One case of cancer recurrence with hepatic confusion, and one case of anastomotic site stricture prevented the ileostomy repair procedure from taking place. Two patients had permanent stoma after undergoing abdominoperineal resection because of AL and severe ischemic colitis respectively (The later patient was in the ileostomy group with the preoperative diagnosis of systemic lupus erythematosus. Ischemic colitis was detected in the sigmoid colon by sigmoidoscopy on 50th postoperative day, and the intact anastomotic site was identified below 10 cm from the lesion). A 4 months interval (range, 0-13) was observed from the primary operation to the ileostomy repair. About 4 additional days of postoperative stay were needed for the ileostomy repair.

AL occurred in 5 patients (7.5%) in the ileostomy group, 1 patient (2.9%) in the TAT group, and 6 patients (17.1%) in the noprotection group. The difference was not statistically significant among groups (P = 0.125). The TAT presented noninferiority in preventing AL compared to ileostomy with a -5% noninferiority margin (limit of 95% confidence interval for difference between preventing rates: -2.9% to 15.1%). Among the 5 AL patients in the ileostomy group, 1 patient underwent reoperation (abdominoperineal resection) because of AL. One AL patients in the TAT group and 5 of 6 AL patients in the no-protection group underwent reoperation, and temporary ileostomies were created. Other AL patients were treated successfully with conservative methods such as fasting, fluid resuscitation, antibiotics, and/or pigtail insertion.

Early and late postoperative complications except AL were compared (Table 4). Early complications that occurred within 30 days after surgery showed no significant difference among the groups. However, the ileostomy group had a higher late complication rate than the other two groups as complications associated with the stoma itself or repair operation occurred (P = 0.016). Anastomotic site stricture was observed in 1 patient in the ileostomy group and 1 patient in the no-protection group. The patient with stricture in the ileostomy group was treated with abdominoperineal resection, and the patient with stricture in the no-protection group received intermittent endoscopic bougination. No stricture was observed in the TAT group during the follow-up period. The median (range) follow-up period was 27 months (3-49 months) for the ileostomy group, 17 months (3-29 months) for the TAT group, and 42 months (21-51 months) for the no-protection group. No surgical mortality was observed.

The International Study Group of Rectal Cancer defined AL as communication between the intraluminal and extraluminal compartments caused by a defect in the intestinal wall integrity at the anastomosis [11]. AL is one of the most serious complications related to colorectal surgery. It not only increases morbidity, mortality, and postoperative hospital stay but also impairs the long-term oncologic and functional outcomes after rectal cancer surgery.

The mechanism of AL is still uncertain, but it seems that increased intraluminal pressure is a significant risk factor. Hallbook and Sjodahl [12] reported that "neo-rectal" volume, which is sustainable to distension or straining pressure, is significantly decreased in AL patients. Nesbakken et al. [13] reported decreased "neo-rectal capacity" in patients who experienced AL despite proximal diversion after rectal resection. Accordingly, proximal decompression is an important point in preventing AL. Clinically, proximal decompression can be achieved with diverting stomas, transanal decompression devices, or intraluminal devices.

Among these methods, temporary loop ileostomy is the most commonly used method. It is known to diminish the catastrophic complication of AL, and reduce the risk of surgical reintervention [9,10,14]. On the other hand, some researchers have insisted that ileostomy does not actually decrease the leak rate [1,8,15]. Ileostomy itself can induce many complications such as wound infection, prolapse, retraction, stenosis, necrosis, parastomal hernia, ileus, obstruction, and so on. In this study, no differences in early postoperative complications were observed among the groups, but the late complication rate was significantly higher in the ileostomy group.

Additional surgery under general anesthesia is necessary for an ileostomy repair and this exposes the patients to additional operative risks. Pulmonary edema took place in a patient who received ileostomy repair (the patient underwent laparoscopic cholecystectomy simultaneously), and an incisional hernia at the repair site developed in another patient.

Transanal decompression devices may protect the anastomotic site and prevent the shortcomings of diverting stomas. The TAT in this study was devised from the existing rectal tube to maximize the decompressive effect. Many wide holes in our TAT make sufficient drainage or decompression possible. The TAT can be inserted and removed easily. The TAT has a low risk of bowel wall injury because of its curvelinear shape and soft silicone texture.

Compared to the ileostomy group, AL was not significantly different in the TAT group. The two groups had equivalent clinicopathological characterisitcs including tumor location and BMI. However, the ileostomy group showed longer operation times and, larger amounts of intraoperative blood loss. Increased intraoperative blood loss or blood transfusion is mentioned as a risk factor for AL in some studies [16,17,18]. From this point of view, the AL rate of the TAT group in this study may show significantly higher leak rates if intraoperative blood loss increases. However, other results suggest that they are not significant factors [19,20,21]; therefore, more evidence in this regard should be accumulated. Effectiveness of indwelling TAT can be estimated with adjustment of these factors by comparing the TAT group with the no-protection group, since there are no significant differences in these variables between the two groups. Although no statistical significance (P = 0.106) was observed, the leak rate tended to decrease in the TAT group with 2.9% versus 17.1%.

These results also suggests that the operators preferred ileostomy rather than TAT in difficult operative cases. Similarly, in 2005, Gurjar et al. [22] reported that only about 16% of surgeon respondents in Britain and Ireland used indwelling tubes for low anterior resection. This is because the evidence regarding the safety and efficacy of indwelling TAT are yet insufficient for its general use [23]. A few studies on this issue [24,25,26,27,28] were reported recently. They compared the AL rate between a TAT group and control (without TAT) group after anterior resection for rectal cancer. Most of them showed a lower AL rate in the TAT group compared to the control group, and the leak rates of each group were similar to our results (Table 5). Among these studies, this is the first report to compare a TAT group to an ileostomy group for investigating the efficacy and safety of TAT in preventing AL after rectal cancer surgery.

Deciding on the necessity of a diverting stoma in each individual patient is a challenging issue. Karliczek et al. [29] reported that intraoperative prediction of AL has low sensitivity and specificity. In surgical cases with a definite high risk of AL, diverting stoma is warranted. However, in some cases with moderate or indistinct risk of AL, when diverting stoma creation is not absolutely indicated, the TAT can be an alternative to the diverting stoma; this may spare patients possible complications of a stoma and still provide anastomosis protection. Moreover, the TAT can be used simultaneously with a diverting stoma. It can assist in the drainage and decompression of the remaining fecal material past the ileostomy site. This situation can be apparent in cases of incomplete bowel preparation or emergency operation.

The TAT group did not have longer hospital stay periods compared to the ileostomy group. Postoperative hospital stays in the no-protection group were more than 1 day longer than those in the other two groups. For the no-protection group, a tendency of slow postoperative diet resumption was observed, and this may have contributed to the longer hospital stay. Furthermore, in the ileostomy group, considering the additional hospital stay that would be needed for ileostomy repair, the actual total postoperative hospital stay was the shortest in the TAT group.

Insertion of the TAT is easy and requires less time than that of ileostomy. Operation time of the TAT group was more than 40 minutes shorter than that of the ileostomy group, and was not significantly different from that of the no-protection group. Performing ileostomy takes about 20-40 minutes in usual practice. Thus, we think the difference in operation time between the ileostomy group and the other groups was because of rather the operation time for performing the ileostomy than the operative difficulty. On the other hand, it takes only a few minutes to insert the TAT, and this short installation time can be a merit for some situations such as emergency operation.

Our data showed that the TAT provides a comparable protective effect on the anastomotic site and diverts the ileostomyrelated complications. However, there are some limitations for these study results to be directly applied to common practice. First, this study enrolled a relatively small number of patients. An estimate of exact AL rate can be achieved only with large patient numbers, which will provide sufficient statistical power. Second, we chose which method of protection would be performed to each patient at the time of the operation, and decided the tumor level that would be included in our study retrospectively. Thus, selection bias might confound the results. A prospective randomized study is required to exclude bias, but disastrous effects and difficulty in predicting AL risk make the study design challenging. Finally, to expand the range of usage for TAT, additional studies on cases bearing high risks for AL such as low rectal cancer located below 5 cm above the anal verge and patients who underwent neoadjuvant chemoradiotherapy are needed in the future. With more evidence, expanded TAT use will bring benefit in AL protection and avoidance of stoma-related complications.

In conclusion, in mid rectal cancer surgery, the TAT showed no significant difference in AL compared to conventional loop ileostomy. While loop ileostomy possesses its own complications, the TAT didn't show any significant complications. Therefore, the TAT may serve as a safe and easy alternative to diverting stomas in mid or upper rectal cancer surgery. Further large-scale and/or prospective randomized studies are required to gain more evidence that can validate the expanded clinical use of TAT.