Introduction

Intraoperative anastomotic leak testing to evaluate the integrity of intestinal surgical repair has long been a common practice in veterinary medicine.¹ Intestinal dehiscence rates have been reported to be as high as 28%, with risk factors for dehiscence including preoperative septic peritonitis, hypoalbuminemia (< 2.5 g/dL), foreign body obstruction, and intraoperative hypotension.^2,3,4,5,6 Although results of intraoperative anastomotic leak testing are unable to predict the effect of such preoperative risk factors on intestinal healing, it provides a means of evaluating the acute integrity of the surgical repair and has been repeatedly validated as a means of reducing clinically apparent postoperative leakage rates in human surgery as well as serving as a valuable tool for surgeons in training.^{7,8,9,10,11,12,13}

The most common indication for intestinal surgery in dogs is a foreign body obstruction,^14,15 which may require an enterotomy or resection and anastomosis. Other indications for intestinal surgery include intestinal biopsies, neoplasia, hernia repair, intussusceptions, and penetrating trauma.^1,16 Unfortunately, few studies evaluating the usefulness of intraoperative anastomotic leak testing in veterinary medicine exist, and all presently available data pertain to hand-sewn repair of intestinal punch biopsy sites.^17,18 Although surgical technique for closure of intestinal biopsy sites closely approximates that for enterotomies, resection and anastomoses are inherently more complex, involving a circumferential surgical repair that may incorporate suture, staples, or both. Such differences may preclude identification of anastomotic leakage during testing and may demand a different method of testing to be of clinical value.

The currently recommended method of intraoperative anastomotic leak testing involves isolation and digital occlusion of a 10-cm-long segment of intestine containing the surgical site, injection of 16 to 19 mL of sterile saline (0.9% NaCl) solution into the isolated lumen, and visual observation for exudation of saline solution from the suture line.¹⁷ This method has been validated to produce an intraluminal pressure of 25 mm Hg (34 cm H₂O) in medium-sized dogs (mean body weight, 15 kg [33 lb]), approximating maximal physiologic peristaltic pressures.¹⁷ Use of a 22-gauge hypodermic needle for intraluminal saline solution infusion has been reported to minimize leakage from the injection site.¹⁹

Although intraluminal infusion of saline solution has been described for humans, insufflation of the intestine with air is the presently preferred method for confirmation of colorectal anastomotic integrity.^7,10,13,20 Leakage is considered present if bubbles are generated from the intestinal segment submerged in the patient's saline solution–filled abdomen. Of note, intraoperative leak testing of colorectal, but not small intestinal, anastomoses is described in the human literature, largely as a result of the substantial morbidity and mortality rate associated with fecal peritonitis and the altered healing characteristics of the colon relative to the small intestine. Transrectal air infusion avoids creation of a transmural intestinal defect subsequent to needle puncture while allowing for improved identification of anastomotic leakage, as saline solution can resemble the natural intestinal serosal environment. Although the use of intraluminal air insufflation has been validated to reduce the incidence of postoperative dehiscence in humans, the use of an alternative testing method in veterinary medicine precludes the conclusion that intraoperative anastomotic leak testing is of clinical value.^8,9,12,13

To the authors' knowledge, the clinical value of intraoperative anastomotic leak testing has not been evaluated in veterinary medicine. The absence of leak testing of small intestinal anastomoses in human surgery raises questions over whether it should be a common practice for dogs and cats. The objective of the present study was to compare the rate of postoperative dehiscence on the basis of intraoperative anastomotic leak test results (ie, positive or negative for leakage or testing not performed) between dogs that underwent HSA or FEESA of the small intestine. We hypothesized that HSAs of the small intestine would be more likely to undergo intraoperative anastomotic leak testing than FEESAs of the small intestine but that there would be no difference in dehiscence between HSAs and FEESAs. Our second hypothesis was that there would be no difference in dehiscence between anastomoses that underwent intraoperative leak testing versus those that did not undergo intraoperative leak testing, regardless of anastomotic technique. Our third hypothesis was that there would be no difference in dehiscence between anastomoses with a positive or negative intraoperative leak test result.

Materials and Methods

Results

One hundred forty-five dogs had a resection and anastomosis performed between January 2008 and October 2019 at the University of Florida. Fourteen dogs were excluded from the study because of insufficient follow-up time (n = 10), colonic resection (3), or perioperative euthanasia unrelated to the surgery (1). One hundred thirty-one dogs met the inclusion criteria for this study. One-hundred twenty dogs underwent a single small intestinal anastomosis, and 11 dogs underwent multiple small intestinal anastomoses. Of these 11 dogs, 2 underwent 2 anastomoses during the same surgical procedure, and 9 underwent 2 separate surgical procedures, both of which involved intestinal anastomoses. Of these 9 dogs, 7 had a single anastomosis performed during each surgery, whereas 2 dogs had 2 anastomoses performed in the first surgery and 1 in the second surgery. As a result, a total of 144 small intestinal anastomoses were performed; 94 (65.3%) were FEESAs and 50 (34.7%) were HSAs (Table 1).

The most common breeds and types of dogs that underwent an FEESA were mixed breed (n = 18), American Staffordshire Terrier (5), German Shepherd Dog (5), and Doberman Pinscher (4). The most common breeds and types of dogs that underwent an HSA were mixed breed (n = 7), Dachshund (2), Rat Terrier (2), and Toy Poodle (2). The median age at the time of surgery was 5.7 years (range, 0.2 to 13.0 years) for dogs that underwent an FEESA and 6.0 years (range, 0.4 to 14.2 years) for dogs that underwent an HSA; age was not significantly (P = 0.37) different between dogs that underwent different anastomotic procedures. Thirty-eight females (5 sexually intact and 33 spayed) and 56 males (16 sexually intact and 40 neutered) underwent an FEESA, and 18 females (1 sexually intact and 17 spayed) and 32 males (5 sexually intact and 27 neutered) underwent an HSA. The mean body weight was 24.1 kg (53 lb) with a range of 4.7 to 71.4 kg (10.3 to 157.1 lb) for dogs that underwent an FEESA and 18.7 kg (41.1 lb) with a range of 3.1 to 52.7 kg (6.8 to 115.9 lb) for dogs that underwent an HSA; body weight was significantly (P = 0.01) different between dogs that underwent different anastomotic procedures.

Indications for surgery in dogs that underwent an FEESA included a mass (n = 27), nonlinear foreign body (25), linear foreign body (15), dehiscence (13), intussusception (10), and evisceration (4). Indications for surgery in dogs that underwent an HSA included a mass (n = 12), nonlinear foreign body (12), dehiscence (12), intussusception (6), linear foreign body (5), and evisceration (3). No significant (P = 0.64) difference in indications for surgery was found between dogs that underwent HSAs or FEESAs.

Preoperative serum albumin concentration was measured in 54 of 89 (60.7%) dogs that underwent an FEESA and 31 of 42 (73.8%) dogs that underwent an HSA. The mean preoperative serum albumin concentration was 2.63 and 2.2 g/dL (reference range, 2.9 to 3.8 g/dL) in dogs that underwent FEESAs and HSAs, respectively. No significant (P = 0.84) difference in preoperative serum albumin concentration was found between dogs that underwent HSAs and dogs that underwent FEESAs.

Preoperative septic peritonitis was present in 48 of 131 (36.6%) dogs. Forty-four of these dogs had septic peritonitis at the time of hospital admission, with 43 requiring a single anastomosis and 1 requiring 2 anastomoses. One of these dogs was readmitted to the hospital 4 years later with septic peritonitis, whereas 4 dogs developed postoperative dehiscence, requiring a second surgery in the presence of septic peritonitis. Four dogs that did not initially have septic peritonitis developed postoperative dehiscence and required a single anastomosis in the presence of septic peritonitis. Subsequently, a total of 54 anastomoses (36/94 [38.3%] FEESAs and 18/50 [36.0%] HSAs) were performed in the presence of preoperative septic peritonitis in an effort to correct the underlying source of sepsis. No significant (P = 1.00) difference in incidence of preoperative septic peritonitis was found between dogs that underwent HSAs and those that underwent FEESAs when controlling for dogs that underwent > 1 resection and anastomosis.

In 92 of 94 (97.9%) FEESAs, a gastrointestinal anastomosis 50-mm reusable stapler^b or thoracoabdominal 30-, 55-, or 90-mm reusable stapler^c was used to complete the intestinal anastomosis. In the remaining 2 FEESAs, a 45- and a 60-mm endoscopic surgical stapler^d were used in place of the gastrointestinal anastomosis and thoracoabdominal staplers. A staple cartridge^e with a 3.8-mm open height and a 1.5-mm closed height was used for the gastrointestinal anastomosis stapler. The staple cartridge size for the thoracoabdominal stapler was reported for only 27 of 92 (29.3%) anastomoses; a staple cartridge^f with a 3.5-mm open height and 1.5-mm closed height was used in these anastomoses. Thirty-one of 94 (32.9%) FEESAs had a single cruciate suture pattern placed at the crotch of the anastomosis, and 17 of 94 (18.1%) had the staple line created by the thoracoabdominal stapler oversewn. Of these 48 anastomoses with modification to the traditional FEESA, 5 included both modifications. The FEESA modifications were performed with 4-0 polydioxanone.

Suture type, suture size, and suture pattern were specified for 48 of 50 (96.0%) HSAs, including 3-0 (n = 10) or 4-0 (38) suture sizes that included polydioxanone (44), glycomer-631 (3), or poliglecaprone-25 (1) in simple continuous (30) or simple interrupted (18) suture patterns. All simple continuous suture patterns were performed with 2 suture strands. One suture strand started on the mesenteric border, and the second suture strand started on the antimesenteric border, and each extended 180° circumferentially around the anastomosis (as a modified simple continuous suture pattern) without the use of stay sutures.

One hundred six of 144 (73.6%) anastomoses were reinforced with omental wrapping (n = 100), serosal patching (3), or both (3), including 68 of 94 (72.3%) FEESAs and 38 of 50 (76.0%) HSAs. Specifically, omental wrapping alone (n = 63), serosal patching alone (3), omental wrapping and serosal patching (2), or no reinforcement (26) was recorded for FEESAs, and omental wrapping alone (37), serosal patching alone (0), omental wrapping and serosal patching (1), or no reinforcement (12) was recorded for HSAs. No significant (P = 0.74) difference in the number of reinforcement procedures or method of reinforcement was found between dogs that underwent HSAs and those that underwent FEESAs.

Intraoperative leak testing was performed during 62 of 144 (43.1%) small intestinal anastomoses, which included 26 of 94 (27.7%) FEESAs and 36 of 50 (72.0%) HSAs. Intraoperative leak testing was not performed during 82 of 144 (56.9%) small intestinal anastomoses. All leak tests were performed by use of saline solution infusion without intraluminal pressure monitoring. The length of intestine tested (6 to 20 cm) and the volume of saline solution infused (6 to 20 mL) were not consistent and varied with surgeon preference. Of the 62 small intestinal anastomoses that underwent leak testing, 5 (8.1%) were positive for leakage and 57 (91.9%) were negative for leakage. Of the 5 positive leak test results, 3 occurred during FEESAs and 2 occurred during HSAs. Thus, 3 of 26 (11.5%) tested FEESAs were positive for intraoperative leakage, whereas 2 of 36 (5.5%) HSAs were positive for intraoperative leakage. Hand-sewn anastomoses were significantly (P < 0.01) more likely to undergo leak testing than FEESAs. No significant (P = 0.64) difference was found in the likelihood of a positive leak test result between dogs that underwent HSAs and those that underwent FEESAs. All anastomoses with a positive intraoperative leak test result were reinforced with simple interrupted sutures until no leakage was observed. The specific locations of leakage were not reported.

No significant differences were found in serum albumin concentration (P = 0.77), preoperative peritonitis (P = 0.11), surgical indication (P = 0.21), or omental or serosal reinforcement (P = 0.35) between anastomoses that did and did not undergo leak testing. Similarly, there was no significance difference in serum albumin concentration (P = 0.93), preoperative peritonitis (P = 1.00), surgical indication (P = 0.53), or omental or serosal reinforcement (P = 1.00) between anastomoses that had a positive leak test result and those that had a negative leak test result.

Thirteen of 144 (9.0%) anastomoses underwent dehiscence after surgery (median, 4 days; range, 2 to 17 days), with subsequent septic peritonitis, including 10 of 94 (10.6%) FEESAs and 3 of 50 (6.0%) HSAs. The incidence of postoperative dehiscence was not significantly (P = 0.54) different between FEESAs and HSAs. The section of intestine involved included jejunojejunal (n = 6), duodenojejunal (3), jejunoileal (2), duodenal (1), and ileocolic (1) anastomotic sites. Eight (7 FEESAs and 1 HSA) of the 13 anastomoses with postoperative dehiscence were in dogs with preoperative septic peritonitis. The presence of preoperative septic peritonitis was not significantly (P = 0.20) associated with dehiscence. Preoperative serum albumin concentration was recorded for 9 of 13 dogs that had postoperative anastomotic dehiscence; these dogs had a mean serum albumin concentration of 2.47 g/dL (range, 1.6 to 3.4 g/dL; reference range, 2.9 to 3.8 g/dL). The presence of preoperative hypoalbuminemia was not significantly (P = 0.64) associated with dehiscence. Indication for the initial surgery in these dogs included foreign body (n = 10), mass (2), and intussusception (1). Only 3 FEESAs and 2 HSAs with postoperative anastomotic dehiscence had been oversewn or reinforced with omental wrapping. Six FEESAs with postoperative anastomotic dehiscence had placement of a crotch suture. The location of postoperative anastomotic dehiscence was recorded for only 6 of 10 FEESAs, and all were at the thoracoabdominal staple line. Elective conservative management (n = 3) or death secondary to sepsis (1) in the remaining 4 FEESAs with postoperative anastomotic dehiscence precluded identification of leakage location. All HSAs with postoperative anastomotic dehiscence had been performed with 3-0 (n = 2) or 4-0 (1) polydioxanone in a modified simple continuous pattern.

Dehiscence occurred in 5 (8.1%) of the 62 anastomoses that underwent leak testing. Of the 82 untested anastomoses, 8 (9.8%) had postoperative anastomotic dehiscence. No significant (P = 0.78) difference in dehiscence was noted between anastomoses that underwent intraoperative leak testing versus those that did not, regardless of anastomotic technique. No significant (P = 0.10) difference was found in the time to dehiscence between anastomoses that underwent intraoperative leak testing (mean, 2.6 days) and those that did not (mean, 3.6 days) when outliers were excluded. These outliers included 3 FEESAs with anastomotic dehiscence at 9, 10, and 17 days after surgery; these dogs had abscess formation at the thoracoabdominal staple line without gross diffuse peritoneal contamination. None of the anastomoses with an initial positive leak test result, 5 of 57 (8.8%) anastomoses with a negative leak test result, and 8 of 82 (9.8%) untested anastomoses had postoperative anastomotic dehiscence. No significant (P = 1.00) difference in anastomotic dehiscence was found between positive and negative leak test results when untested anastomoses were excluded.

The median follow-up time was 21 days after surgery (range, 5 to 2,832 days). Two dogs that developed postoperative anastomotic dehiscence died 3 and 4 days after surgery. In both dogs, conservative management had been elected rather than surgical intervention. The remaining 11 dogs that developed postoperative anastomotic dehiscence were alive at a minimum of 7 days after surgery (range, 7 to 1,266 days) before being euthanized for reasons unrelated to the surgery (n = 2) or being lost to follow-up (9).

Discussion

The results of this study failed to establish an association between intraoperative anastomotic leak testing and a reduction in the incidence of postoperative dehiscence in dogs undergoing FEESAs or HSAs of the small intestine. All hypotheses were conditionally accepted, as HSAs were significantly more likely to undergo intraoperative anastomotic leak testing, and there was no difference in dehiscence between HSAs and FEESAs, anastomoses that did or did not undergo leak testing regardless of anastomotic technique, or anastomoses with a positive or negative intraoperative leak test result, despite dehiscence not being observed in any anastomoses with an initial positive leak test result.

Stapled intestinal anastomoses are the standard of care in human surgery^21,22,23,24 and have become increasingly used in veterinary medicine because of proposed benefits, including reduced surgical time, primary intestinal healing, and decreased intestinal manipulation while maintaining a similar dehiscence rate relative to conventional hand-sewn techniques.^4,6,25–27 Despite this evolution, intraoperative anastomotic leak testing for stapling techniques has yet to be described or validated in veterinary medicine. Moreover, a variety of modifications to FEESAs have been described in the literature (eg, oversew, surgical sealant, and placement of a crotch suture) that can complicate intraoperative assessment of leakage.^28,29 It is for these reasons, along with the more technically challenging aspect of suturing, compared with the automated nature of FEESAs, that the authors believe HSAs are substantially more likely to undergo intraoperative anastomotic leak testing.

No significant difference was found in the likelihood of a positive leak test result when comparing HSAs and FEESAs despite HSAs having less automaticity and the greater potential for surgeon error. In the present study, all anastomoses were performed by board-certified surgeons with experience in intestinal surgery. A greater number of positive leak test results with HSAs may be expected with novice surgeons, providing valuable feedback for improving surgical technique.⁷

Development of dehiscence is multifactorial and includes preoperative risk factors (preoperative septic peritonitis, surgical indication, and hypoalbuminemia), intraoperative factors (surgical technique, tension, anastomotic reinforcement, and hypotension), patient-dependent factors (preexisting systemic disease and medications), and intestine-dependent factors (vascularity, compliance to distension, and preexisting inflammatory or neoplastic disease).^2,3,4,5 Given the emphasis placed on existing peritonitis, hypoalbuminemia, and foreign body surgery in the literature,^2,3,4,5 the increased risk associated with these factors, although not identified as significant in the present study, may prompt selective intraoperative leak testing of anastomoses performed in the presence of these conditions. However, it is imperative to understand the goals of intraoperative anastomotic leak testing and to take care in interpretation of their results.

Intraoperative anastomotic leak testing aims to achieve intraluminal pressures similar to average physiologic intestinal peristaltic pressures, solely evaluating the strength of the surgical repair, which is dependent on the surgeon's skill level and repair technique.^7,19 This emulates the inflammatory phase of healing, during which anastomotic integrity is almost entirely dependent on the suture or staple line.³⁰ Thus, intraoperative anastomotic leak testing serves as a surrogate measure of the anastomosis to remain intact with insignificant leakage of intestinal contents during the inflammatory phase of healing. It fails to account for preoperative, other intraoperative, patient-dependent, and intestine-dependent factors that can also contribute to dehiscence. Most importantly, no consideration is given of the vascular health and tension on the surgical repair, the 2 factors most important to intestinal healing.¹ Consequently, intraoperative anastomotic leak testing should not be used as a surrogate for predicting dehiscence but rather as a means of critically evaluating one's surgical technique.

Although no significant difference in dehiscence was observed between anastomoses that underwent or did not undergo leak testing in the present study, nor between those that had a positive or negative leak test result, the multifactorial nature of dehiscence and inherent limitations of intraoperative anastomotic leak testing complicate interpretation of these results. Most importantly, the authors believe that a lack of statistical significance should not make intraoperative anastomotic leak testing obsolete, as it may provide valuable information for improving surgical skill, particularly in novice surgeons.⁷ Of note, dehiscence was not observed in any of the anastomoses with a positive leak test result in this study, likely as a result of augmentation of leakage sites with additional suture placement. Contrarily, dehiscence was observed in 8.8% of anastomoses with a negative leak test result. Although this difference in dehiscence was not significant, a positive leak test result with subsequent reinforcement of the anastomosis may have prevented dehiscence secondary to inadequate surgical technique. However, the retrospective nature of the study precludes determination of whether these anastomoses would have undergone postoperative anastomotic dehiscence had they not been tested or augmented. Importantly, a negative leak test result cannot guarantee that dehiscence will not occur for reasons other than surgical technique and cannot factor in other causes of dehiscence, such as hypoalbuminemia, preoperative septic peritonitis, or hypotension. Thus, the authors believe development of a standardized intraoperative anastomotic leak testing technique may be most fitting for clinical use.

The failure to find a significant clinical impact of intraoperative anastomotic leak testing on dehiscence contradicts the human literature, in which intraoperative colorectal anastomotic leak testing has been associated with reductions in postoperative anastomotic leakage as great as 60% to 77%.^9,11 Furthermore, a negative leak test result has been associated with a reduced likelihood of dehiscence, whereas positive leak test results repaired with suture oversew are associated with the highest rate of postoperative leakage in comparison with intestinal diversion or reanastomosis.¹³ Such a discrepancy in our results relative to those in human medicine may be the result of use of air rather than saline solution for intraluminal distention, intraluminal pressure monitoring, differences in surgeon technique and anastomotic location, and access to larger patient populations, to name a few.

As mentioned previously, intraoperative anastomotic leak testing in humans is most often performed by intraluminal infusion of air via a sigmoidoscope to a pressure of 25 to 30 cm H₂O (18 to 22 mm Hg), as measured by a manometer.^7,10 The test is considered positive if bubbles are generated in the saline solution filling the patient's abdomen. The use of intraluminal pressure rather than infusion volume has been suggested for standardization of intraoperative anastomotic leak testing caused by variations in patient anatomy.⁸ However, instillation of saline solution is still relied upon in veterinary medicine, mitigating efforts at standardization of intraoperative anastomotic leak testing. Intraluminal diameter, intestinal wall compliance, and the length of intestine occluded for intraoperative anastomotic leak testing all influence the required volume of saline solution that must be instilled to achieve an intraluminal pressure of 25 mm Hg (34 cm H₂O), the greatest intraluminal pressure reported to be generated by peristalsis under normal conditions in dogs.³¹ Although instillation of 12 to 15 mL or 16 to 19 mL has been recommended for instrumental or digital occlusion, respectively, of a 10-cm length of intestine, this recommendation applies solely to medium-sized dogs with healthy intestinal tissue.¹⁷ A smaller volume of saline solution may be necessary in the face of diseased intestinal tissue secondary to fibrosis, edema, adhesions, or other secondary factors that limit luminal distention. Importantly, creation of supraphysiologic intraluminal pressures may prompt placement of additional reinforcing sutures or staples; however, such augmentation is not benign and must be performed with caution, taking into consideration the effect of additional fixation on the vascular supply to the anastomosis.

Limitations of this study included those inherent to retrospective studies. Data were obtained from the medical records and operative reports, requiring dependence on reporting of intraoperative anastomotic leak testing, accurate representation of the intraoperative results, and reporting of anastomotic reinforcement techniques (ie, oversew, crotch suture, and omental or serosal patching). Low numbers of positive leak test results may have contributed to a failure to find significance as a result of a type II error. Moreover, for those anastomoses that did dehisce, the records did not specify whether there was an apparent surgical error. If intraoperative anastomotic leak testing was of clinical usefulness in reducing the incidence of dehiscence, untested anastomoses may dehisce, on average, before the lag phase of healing because of a failure to detect technical error or equipment malfunction intraoperatively. Although this was not observed in the present study, the low number of anastomotic dehiscences may have contributed to a failure to find a significant difference because of a type II error. Similarly, incomplete data for preoperative serum albumin concentration and a low number of dogs that developed postoperative anastomotic dehiscence may have precluded appreciation of a significant effect of these confounding factors on dehiscence. Limited follow-up time reported in the medical record was also a limiting factor, as dogs with a follow-up time of < 5 days had to be excluded from analysis. This limited follow-up time is because the greatest risk of anastomotic dehiscence occurs 3 to 5 days postoperatively.^30,32 The most important limitation of the present study that prevents a definitive conclusion regarding the clinical effectiveness of intraoperative anastomotic leak testing without a prospective, controlled, randomized study was a lack of standardization, with inconsistent saline solution infusion and intraluminal pressures generated. For instance, in the present study, the volume of saline solution instilled ranged from 6 to 20 mL in an intestinal segment for which length was often undisclosed. Because all anastomoses will eventually leak with high enough pressure, prospective studies to standardize intraoperative anastomotic leak testing with adoption of intraluminal pressure monitoring are imperative. Moreover, comparative analysis of saline solution infusion and air insufflation is warranted to determine whether one technique is superior to the other. Modifications to saline solution infusion, including the addition of dye or fluorescein, may also be considered. In conclusion, the present study failed to demonstrate a significant association between intraoperative anastomotic leak testing and a reduction in postoperative dehiscence of HSAs and FEESAs of the small intestine. Given the morbidity and high mortality rate associated with anastomotic dehiscence, its multifactorial nature, and the potential usefulness of intraoperative anastomotic leak testing in improving surgical technique, the authors believe intraoperative anastomotic leak testing may have the greatest clinical usefulness for novice surgeons in helping eliminate poor surgical technique as a factor for dehiscence, but further investigation is needed.