Rectal Perforation During Pelvic Surgery

Rectal perforations during pelvic surgery are rare but serious complications. Caution should be posed to avoid iatrogenic injuries, especially in case of challenging or salvage procedures; similarly, early detection of the lesion should be pursued in case of a suspicion. Actually, management of an RI can be accomplished with—or during—minimally invasive surgery. In general, RIs diagnosed after surgery suffer significantly worse outcomes than those detected and managed intraoperatively; surgeons and clinicians should be aware of possible late presentations, requiring more surgical procedures and longer length of stay. Conflicts of interest : The authors have nothing to disclose.

Rectal perforation can be stratified by the timing of diagnosis, thus defined as early (intraoperative identification of the lesion) or delayed (postoperative diagnosis). When detected during surgery, recognition of a bowel injury simply involves the visualization of a defect in the bowel wall or the extrusion of fecal contents [1] , [40] . Sometimes, an intraoperative endoscopy may confirm the lesion as an air leak test and may help identify the site of perforation [1] . The approximate length of an RI detected during RP is described by two author groups. In the open series by Topaktas et al [41] , RI = 10, with the lesion size ranging from 1 to 3 cm, whereas on the robotic series by Kheterpal et al [17] , RI = 10, with the length of the injury ranging from 0.3 to 2 cm. Unlike early detection, postoperative presentation of an RI could be variable and diagnosis can be more difficult [1] , [22] . Symptoms of a delayed diagnosis of rectal perforation may include abdominal pain, hypotension, fever, tachycardia, peritonitis, and septic shock [22] ; leukocytosis and/or leukopenia can be typical signs. Drainage of enteral contents through the skin, urethra, or vagina, or drainage of urine through the rectum or fecal incontinence are pathognomonic too [1] . Generally, the mean time to diagnosis of a delayed RI is 2.1–3.5 d and ranges from 0 to 13 d [1] . The literature is quite controversial as to when iatrogenic bowel injuries are most commonly diagnosed across different procedures [1] . During RP, the majority of RIs are recognized and managed intraoperatively, being performed with an open, laparoscopic, or robotic approach [12] , [13] , [31] , [40] . As far as gynecologic surgery is concerned, most of the bowel injuries during robotic procedures are diagnosed intraoperatively (87%) [22] . On the contrary, during laparoscopic gynecologic surgery, 42% of bowel injuries are detected postoperatively [2] . Thus, it has been speculated that robotic surgery carries the advantage of improved optics secondary to the three-dimensional high-definition capacity of the robotic platform. Once an RI is suspected in the postoperative course, an upright abdominal x-ray may reveal free air. After percutaneous gastrostomy tube placement, free air on x-ray had 100% sensitivity and 96% specificity for bowel injury, especially when the subdiaphragmatic air pocket was >2 cm or did not resolve within 72 h , according to Leevan et al [1] , [42] . Then, computed tomography (CT) may confirm the presence of bowel injury; the choice of contrast administration may be made on a case by case basis, with consideration for the patient’s clinical status and location of suspected injury, as stated by Leevan et al [1] , [42] . Triple-phase contrast CT has 98% specificity for intra- and extraperitoneal visceral injuries; CT with endovenous contrast had only 90% sensitivity and 96% specificity, but avoids delays related to enteral contrast administration [1] , [43] . If required, a flexible sigmoidoscopy or colonoscopy may also be of importance to confirm the possible site of bowel injury after pelvic surgery. Leevan et al suggest that when an injury ends up with a suspicious urorectal fistula, cystoscopy, retrograde urethrogram, pelvic magnetic resonance imaging, and barium enema versus CT with rectal contrast should be considered for diagnosis and surgical planning [1] , [44] . Treatment of RIs depends on the timing of diagnosis, extent of the defect, and clinical performance of the patient. There are no randomized prospective studies or algorithms for decision-making; however, according to the most recent review on the topic by Leevan and Carmichael [1] , studies on primary repair versus diversion (colostomy) failed to report worse outcomes for primary repair. As a general rule from the trauma literature, when an iatrogenic injury is small and recognized intraoperatively, the primary repair of the defect is generally appropriate [1] . Early management of RIs can be accomplished with a minimally invasive approach and depends on surgeon’s experience [1] , [3] , [17] . Kheterpal et al [17] , reporting on ten cases of RIs out of >4400 RPs, detected and managed all iatrogenic lesions intraoperatively. In this setting, the repair should be performed after prostate removal [31] , [41] . The operative field should be irrigated copiously with saline or povidone iodine [3] , [17] , [41] . The rectal defect should be exposed clearly [3] , [17] , and the margins identified clearly by means of digital rectal examination or a metallic bougie [3] . The rectal wall is closed in two layers (inner mucosa and outer seromuscular layer) with 2-0 or 3-0 polyglactin sutures [3] , [17] . A single study reported on the closure of the rectal wall with three layers—rectal mucosal layer, outer seromuscular layer, and perirectal surrounding tissue, closed separately with absorbable running suture (2-0 monocryl). Moreover, a pedicled omental flap with vascular supply was mobilized and placed between the rectum and the bladder to support the repaired tissue [41] . The integrity of the closure should finally be checked with retrograde rectal injection of air, as described earlier [3] , [40] . Broad-spectrum antibiotics should be administered postoperatively, for up to 5–7 d, as suggested by Karadag et al [3] . Oral liquids can be given the day after surgery, and a diet can be initiated after passing flatus [17] , [37] . The majority of the intraoperative closure was healed without colostomy [3] , [10] , [17] , [31] , [41] . Despite the absence of definite guidelines, when dealing with general gastrointestinal lesions, a fecal diversion should be considered in case of a need of bowel resection [3] . The use of omental interposition between the rectum and vesicourethral anastomosis has been suggested in case of primary repair, to serve as a barrier to urea and the acidic pH of urine that would otherwise inhibit healing of the rectal wound [6] , [17] , [41] . It is important to note that the quality of the rectal repair is crucial for primary healing. Once the detection of an RI is delayed, there is no defined algorithm of management in the literature. In general surgery, rectal anastomotic leaks are managed with drainage and fecal diversion [45] . In some cases, localized abscesses deriving from rectal anastomotic leaks can be drained percutaneously [1] . Historically, RIs after perineal prostatectomies have been treated successfully with a simple supportive medical therapy that included antibiotics, given that conservative management of RIs below the peritoneal reflection has so long been considered feasible [1] , [46] . Fistulas between the excretory system and the rectum may derive from surgical injury and represent a challenging issue. In such cases, colostomy, urinary diversion, and delayed repair/reconstruction represent the management providing a higher success rate [1] , [43] , [47] . A vascularized tissue flap can also be used [43] , [46] ; some authors described the transposition of the gracilis muscle in patients with prior radiation or as a salvage procedure after failed attempts at repair [48] . In general, RIs diagnosed after surgery suffered significantly worse outcomes than those detected and managed intraoperatively [1] . Actually, when diagnosed lately, patients undergo more surgical procedures and require longer length of stay [2] , [26] , [49] . Two review articles considered overall bowel lesions and found that mortality rate for intraoperatively detected lesion is 0–1.7%, whereas mortality rates for lately detected and managed bowel lesion are 3.2% and 7.7% [2] , [49] . Similarly, during RP, failure to recognize and immediately treat a bowel injury may result in a high mortality rate of up to 3% [26] , [31] . This consideration applies to all fields of pelvic surgery. In the review article on laparoscopic gynecologic surgery by Llarena et al [2] , a delayed diagnosis of bowel injuries can result in significant morbidity and mortality; indeed, the authors concluded that a postoperative recognition of an RI requires high clinical suspicion given the variable patient presentations. Currently, it is difficult to state whether a surgical approach is safer than another as far as RIs are concerned. From the current review, it could be assumed that the robotic one is seemingly associated with a lower rate of iatrogenic rectal lesions. However, a correct classification of bowel injuries is often missing. Picerno et al [22] found that more than half of the bowel injuries were reported imprecisely, without mentioning whether the bowel injury is a serosal injury, enterotomy, or perforation, stating that the lack of a definition of bowel injury means that the number of injuries may be either over- or under-reported. A correct reporting of intraoperative complications is of paramount importance and will provide the basis for the prevention and management of this occurrence.

Iatrogenic bowel injury is a challenging issue that may be a consequence of a wide range of procedures and percutaneous interventions. The small as well as large intestine is in close proximity to numerous intra-abdominal, retroperitoneal, and pelvic organs; they occupy sizable portions of the abdominal and pelvic cavities, have variable lengths, and are prone to dynamic motion. Urologic, gynecologic, and general surgeries are the most common causes of iatrogenic bowel injury [1] . In general, bowel injury may occur during organ dissection, visceral manipulation, or abdominal wall entering/closure. Rectal injuries (RIs) during pelvic surgery are rare but severe complications. The rate of RIs is generally lower than the rate of injuries of other portions of the bowel; the majority of studies describes that the rate of small bowel lesions widely exceeds colorectal ones [1] , [2] . The urologic field is responsible for the major part of iatrogenic RIs from pelvic surgery. Owing to the close proximity with the lower ureter, bladder, and prostate, the rectum is at increased risk of injury during dissection, especially during radical prostatectomy (RP). In this setting, even if rare, RIs may occur during the dissection of the base or apex of the prostate when the Denonvilliers’ fascia is not incised properly. The procedure becomes contaminated and poses the risk of septic complications such as wound infection, rectourethral fistula, peritonitis, and death [3] . Historical series reported an incidence of RIs during RP that ranges from 0.5% to 9%; in this setting, RI occurrence is seemingly higher during the transperineal access [4] , [5] . When considering open retropubic, laparoscopic, and robotic RP, the incidence of rectal perforation across a single series is summarized in Table 1 [6] , [7] , [8] , [9] , [10] , [11] , [12] , [13] , [14] , [15] , [16] , [17] , [18] , [19] , [20] . It has been argued that the incidence of RIs is unreported in the most recent literature likely due to a lower occurrence [3] . The 2021 European Association of Urology guidelines summarized the rates of overall organ injury during RP, which were 0.4%, 2.9%, and 0.8% during robotic, laparoscopic, and open prostatectomy, respectively [21] . Table 1 Incidence of rectal injuries during radical prostatectomy First author Year Technique Rate of RI (%) No. Borland [6] 1992 Open 1.5 10/1000 Igel [7] 1987 1.3 9/692 Lepor [8] 2001 0.5 5/1000 McLaren [9] 1993 1.2 27/2212 Guillonneau [10] 2003 Laparoscopic 1.3 13/1000 Stolzenburg [11] 2006 0.7 6/900 Katz [12] 2003 2% 6/300 Blumberg [13] 2009 1 2/200 Murphy [14] 2009 Robotic 1 5/400 Patel [15] 2007 0.4 2/500 Yee [16] 2008 0.8 2/251 Kheterpal [17] 2011 0.2 10/4400 Novara [18] 2010 1.5 5/415 Wedmid [19] 2011 0.1 11/6650 Hung [20] 2011 1.04 3/288 RI = rectal injury. Incidence of rectal injuries during radical prostatectomy RI = rectal injury. The proximity of the uterus and ovaries to the sigmoid colon, cecum, and rectum increases the risk of bowel injury during gynecologic surgery. As far as gynecologic laparoscopy is concerned, the rectum accounts for 18% of overall bowel injuries. Rectal lesions have been described during hysterectomy for either benign or malignant diseases, for all the approaches (open, laparoscopic, vaginal) and also with single-site surgery. Few cases are reported also after colposacropexy and tubal sterilization [2] . More recently, a systematic review focused on robotic gynecologic surgery and included a total of 13 444 procedures [22] : overall, a total of 84 bowel injuries were reported for an incidence of one in 160 (0.62%; 95% confidence interval [CI] 0.50–0.76%), ranging from 0% to 5.88% [22] . The most commonly specified location of injury was the rectum (nine cases, 38%) [22] . The authors confirmed that hysterectomy was associated with higher rates of bowel injury than myomectomies, especially if performed for mixed indications. To note, one of the articles included in this systematic review dealt with robotic management of deep infiltrating endometriosis: among 88 patients treated inside their international multicentric study, two cases of RIs were recorded [23] . The rectal shaving technique to obtain maximal removal of deep endometriotic foci may increase the risk of RIs, especially resulting in fistulas or stenosis [24] , [25] .