Materials and Methods

Animals—From March 2005 to July 2006, dogs undergoing gastrojejunostomy tube placement were enrolled in a prospective study to document the rate of complications associated with usage. Following exploratory laparotomy for definitive diagnostic or therapeutic intra-abdominal surgical procedures, a gastrojejunostomy tube was placed if the necessity of intensive postgastric nutritional support was anticipated. The Institutional Animal Care and Use Committee of the Animal Medical Center approved all procedures, and informed client consent for tube placement was obtained for each dog.

Gastrojejunostomy tube placement—After completion of the primary intra-abdominal procedure, a 5to 10-cm section of the gastric body (visceral surface) was isolated by use of laparotomy sponges. The anticipated location of tube insertion was mobilized toward the ventrolateral aspect of the left body wall (caudal to the last rib) to ensure that minimal tension would be present within the completed tube gastropexy. Full-thickness stay sutures were inserted cranially and caudally to the desired site of tube insertion, with 2-0 or 3-0 nylon.^a A purse string suture was placed at the desired insertion point with 2-0 or 3-0 polydioxanone.^b A stab incision was made into the gastric lumen in the center of the suture, and the gastrostomy tube was inserted. A 20-F (61-cm [24-inch]-long) silicone gastrostomy tube^c was used in each dog. Folding the dome shaped end piece in half and then compressing it with a Kelly hemostatic clamp facilitated insertion. Once the tube was within the gastric lumen, the purse string suture was tightened such that deformation of the tube exterior was avoided. The left-sided body-wall-tube gastropexy site was confirmed, and a Carmalt forceps was introduced from the skin, via a stab incision, through the peritoneum. The top of the tube was grasped with the Carmalt forceps and exteriorized. The gastropexy was completed via placement of 4 simple interrupted mattress sutures (from the stomach to the transversus abdominis muscle) surrounding the tube, with 2-0 or 3-0 polydioxanone. A plastic retaining ring with a plastic external bolster (provided within the prepackaged gastrostomy tube kit) was advanced over the plastic tip of the gastrostomy tube, to the level of the body wall (Figure 1). Additional external sutures were not necessary to secure the tube to the skin. The plastic tip was then removed with scissors in preparation for insertion of the jejunostomy tube. A 9-F (88.9-cm [35-inch]-long) hydromer-coated jejunostomy tube with a metal-weighted tip^d was used in each dog (Figure 2). A metal guide wire (260 cm [102 inches] long) was included in the prepackaged jejunostomy tube kit. The guide wire was inserted into the lumen of the tube to the level of the metal end piece. Extension of the guide wire beyond the endpoint of the jejunostomy tube was prevented to avoid iatrogenic damage to the gastrostomy tube or small intestine during insertion. The jejunostomy tube was lubricated with saline (0.9% NaCl) solution or sterile, water-soluble lubricant^e and inserted through the lumen of the gastrostomy tube until the end could be felt within the gastric lumen. The dome of the gastrostomy tube was then manually advanced to the pylorus, and the jejunostomy tube was inserted through the pyloric antrum into the proximal portion of the duodenum. Alternatively, the dome of the gastrostomy tube was not advanced to the pylorus, and the end of the jejunostomy tube was inserted past the level of the gastrostomy tube dome. The tip of the jejunostomy tube was grasped through the stomach wall and was manually advanced to the level of the pyloric antrum. With either technique, jejunostomy tube insertion was completed by use of manual traction and manipulation to position the tip of the tube within the proximal third of the jejunum. The custom-fit, dual-port tip of the jejunostomy tube was then securely seated within the top of the gastrostomy tube, and the jejunostomy tube guide wire was removed. A sterilized, commercially available metric ruler^f was used to measure the jejunostomy tube insertion distance (centimeters) starting from the duodenocolic ligament (triangular attachment of the ascending duodenum to the mesocolon) to the palpable endpoint of the jejunostomy tube within the jejunum. Copious saline solution abdominal lavage was performed, and a section of omentum was wrapped around the tube gastropexy site.

View Full Size

Image of a gastrojejunostomy tube stoma site depicting the plastic retaining ring and external bolster. Notice the gap between the body wall and the external bolster (arrow), which prevents pressure necrosis from the bolster if postoperative stoma swelling occurs.

Citation: Journal of the American Veterinary Medical Association 232, 3; 10.2460/javma.232.3.380

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Image of the completed gastrojejunostomy tube setup illustrating the metal weighted tip of the jejunostomy tube exiting through the center of the gastrostomy tube dome.

Citation: Journal of the American Veterinary Medical Association 232, 3; 10.2460/javma.232.3.380

• Download Figure
• Download figure as PowerPoint slide

Tube complications—Complications were categorized as either stoma related or mechanical. Tube stoma abnormalities were further classified as major or minor. Major tube stoma complications were documented if abscess formation at the tube exit site occurred or if a revision surgery was performed. Minor tube stoma abnormalities were defined as erythema, cellulitis, or discharge from the tube site. Tube stoma evaluations were performed 4 times daily by use of a standardized visual analog scale. Minor tube stoma abnormalities were assigned a severity score, classified as mild (grade 1), moderate (grade 2), or severe (grade 3). Focal regions of erythema and cellulitis were considered mild changes, whereas circumferential redness or swelling was considered severe. Likewise, intermittent serosanguinous fluid production and discharge from the tube stoma were considered mild, whereas continuous drainage was considered severe. The daily stoma scores were averaged and included for data analysis.

Daily postoperative ventrodorsal abdominal radiographic examinations (Figure 3) were performed to determine the presence of the following mechanical tube complications: kinking, coiling, knotting, or migration. Kinking was defined as an acute fold in the jejunostomy tube at an angle ≥ 90°. Coiling was confirmed when any section of jejunostomy tube formed a 360° loop (Figure 4). Migration was defined as a measurable orad displacement of the metal end piece of the jejunostomy tube on the basis of comparison of serial radiographic evaluations. Any tube migration or displacement necessitating a second surgical procedure was recorded. Radiographic images were interpreted by a single radiologist (AJF) blinded to the clinical outcome of all dogs.

Multiple variables were evaluated for their association with stoma (major and minor) and mechanical tube complications. Preoperative variables that were assessed included the following: age, sex, breed, body weight, presence of vomiting, serum albumin concentrations ≤ 2.5 g/dL, blood glucose concentration (mg/dL), and leukocytosis (≥ 11,500 cells/μL). Intraoperative variables that were evaluated included the type of surgical procedure performed, presence of neoplasia, duration (minutes) of surgical procedure, duration (minutes) of gastrojejunostomy tube placement, and distance (centimeters) of jejunostomy tube placement past the duodenocolic ligament. The presence of gross peritonitis was documented if generalized peritoneal inflammation with overt perforation, spillage, or extravasation of intestinal, pancreatic, or biliary contents was observed at the time of surgery. The presence of septic peritonitis was assessed via microbial culture and susceptibility testing of peritoneal fluid or tissue. Postoperative changes in the serum albumin (measured every 48 hours) and blood glucose concentrations (measured daily) were also recorded for analysis.

View Full Size

Immediate postoperative ventrodorsal radiographic view of the abdomen of a dog. Notice adequate jejunostomy tube placement. Intraoperatively, the radiopaque tip of the jejunostomy tube (wide black arrow) measured 17 cm aboral to the approximate location of the duodenocolic ligament. The approximate location of the duodenocolic ligament (white arrow) is defined by the arch of the tube as it enters the ascending duodenum. The dome of the gastrostomy tube (wide white arrow) can be seen abutting the stomach wall with minimal redundant length of jejunostomy tube within the gastric lumen. A moderate amount of pneumoperitoneum is secondary to the laparotomy.

Citation: Journal of the American Veterinary Medical Association 232, 3; 10.2460/javma.232.3.380

• Download Figure
• Download figure as PowerPoint slide

Gastrojejunostomy tube maintenance—The time to initiation of jejunostomy tube feeding after surgery was recorded for each dog. A commercially available, nutritionally balanced, liquefied canine diet^g was fed to all dogs via a continuous rate infusion (mL/h). Percentage of RER fed (RER = 30 × body weight [kg] + 70) was calculated for each dog daily for the first 3 days following surgery. Intermittent (4 times a day) lavage of the jejunostomy tube with 5 to 10 mL of tap water was performed in all dogs. Gastric aspirates were performed 4 to 6 times daily for the first 3 days following surgery, and residual volumes were quantified. Any clogging of the gastrostomy or jejunostomy tube was recorded. Postoperatively, any episode of vomiting or diarrhea was recorded to allow for assessment of tolerance to tube feeding. Additionally, duration of jejunostomy and gastrostomy tube usage and any tube tearing or premature dislodgement prior to the intent of the attending clinician was recorded.

View Full Size

Postoperative ventrodorsal radiographic views of the abdomen of a dog. A—Notice initial placement of the radiopaque tip of the jejunostomy tube (wide arrow) within the jejunum, aboral to the approximate location of the duodenocolic ligament (arrow). B—Gradual oral migration of the jejunostomy tube was visible on daily postoperative radiographs. On day 3 after surgery, the tube was found to be completely coiled within the gastric lumen. The jejunostomy tube was removed without complication as the dog was tolerating oral feeding.

Citation: Journal of the American Veterinary Medical Association 232, 3; 10.2460/javma.232.3.380

• Download Figure
• Download figure as PowerPoint slide

Statistical analysis—All recorded variables were evaluated for their individual association with time to onset of mechanical tube complication rate and major or minor tube stoma complication rate by use of log-rank and Cox proportional hazards regression analysis. In addition, univariate Cox proportional hazards regression analysis was used to evaluate significance between survival times and individual variables. All variables in the univariate analysis were included in the stepwise multivariate model. Values of P ≤ 0.05 were considered significant. Median survival times were calculated by use of Kaplan-Meier life table analysis. Associations between (aboral) tube placement distance relative to the duodenocolic ligament and mechanical complications were evaluated by use of Cox proportional hazards regression analysis. Computer software was used to perform all analyses.^h

Results

During the 17-month period, 26 dogs underwent exploratory laparotomy and received a gastrojejunostomy feeding tube. Mixed-breed dogs were the most common breed (n = 7), followed by Terrier (6), Maltese (2), Bullmastiff (1), Labrador Retriever (1), Shetland Sheepdog (1), Rottweiler (1), Puli (1), Dachshund (1), Beagle (1), Coon Hound (1), Standard Poodle (1), Pug (1), and Miniature Pincher (1). There were 15 male and 11 female dogs included in the study with the mean ± SD age of 8.8 ± 3.01 years (range, 3.0 to 14.0 years). Preoperative body weight ranged from 2.6 to 47.6 kg (5.7 to 104.7 lb; mean, 19.7 kg [42.3 lb]). Breed, sex, and preoperative body weight did not have a significant association with the incidence of mechanically mediated or tube stoma– related complications or mortality rate.

In 20 of 26 (77%) dogs, the most common problem reported at the time of initial hospital admission was vomiting. Duration of vomiting ranged from 1 to 20 days (mean, 6.15 days). Anorexia was observed in 16 of 26 (62%) dogs, with duration of signs ranging from 1 to 21 days (mean, 5.69 days). A preoperative leukocytosis (≥ 11,500 cells/μL) was observed in 21 of 26 (81%) dogs. Preoperative blood albumin concentrations were recorded for all dogs, with a range of 0.8 to 4.0 g/dL (mean, 2.7 g/dL). Hypoalbuminemia (≤ 2.5 g albumin/dL) was observed in 10 of 26 (38%) dogs, with a range of 0.8 to 2.5 (mean, 1.9 g/dL). Preoperative blood glucose concentrations were measured in all dogs, with a range of 64 to 502 mg/dL (mean, 132 mg/dL). A 48-hour postoperative blood albumin concentration was measured in 22 of 26 (85%) dogs, with a range of 0.6 to 3.0 g/dL (mean, 2.0 g/dL). A 48-hour postoperative blood glucose concentration was measured in 23 of 26 (88%) dogs, with a range of 80 to 348 mg/dL (mean, 141 g/dL). No preor postoperative variables were identified to have a significant association with the incidence of mechanically mediated or tube stoma–related complications or mortality rate.

Table 1—

Gastrointestinal conditions of 11 dogs in which gastrojejunostomy tubes were placed, including gross and septic peritonitis status.

View Table

The most common indication for gastrojejunostomy tube placement was for gastrointestinal disease (n = 11), with confirmed (culture-positive) septic peritonitis in 8 of 11 of these dogs (Table 1). A dehiscence from a previous small intestinal resection and anastomosis (n = 3) or from a previous enterotomy (1) was the most common condition identified in dogs with septic peritonitis. Obstructive small intestinal foreign body causing perforation (n = 3) was also observed. In dogs without sepsis, gastrointestinal neoplasia (diffuse intestinal lymphoma and high-grade gastric adenocarcinoma) was the most common indication for gastrojejunostomy tube placement.

Other indications for gastrojejunostomy tube placement included pancreatic disease (n = 9) and extrahepatic biliary surgery (6). Pancreatic diseases (Table 2) in which a gastrojejunostomy tube was placed included neuroendocrine carcinoma (n = 2), necrotizing pancreatitis (3), and non-necrotizing pancreatitis (4). Cholecystectomy was performed in all biliary procedures (Table 3), with gall bladder mucocele (n = 4) being the most common indication. A ruptured mucocele causing septic peritonitis was confirmed in 2 dogs. Obstructive biliary disease secondary to suppurative cholangiohepatitis (n = 1) and cholecystitis (1) with secondary sterile bile peritonitis were also observed. The type of surgical disease (intestinal, biliary, or pancreatic) or the presence of gross or septic peritonitis did not have a significant association with the incidence of mechanically mediated or tube stoma–related complications (Table 4).

Table 2—

Pancreatic conditions of 9 dogs in which gastrojejunostomy tubes were placed, including gross and septic peritonitis status.

View Table

Table 3—

Biliary conditions of 6 dogs in which gastrojejunostomy tubes were placed, including gross and septic peritonitis status.

View Table

The total surgical time was recorded for each procedure and ranged from 60 to 210 minutes (mean, 128 minutes). The surgical time required for tube placement was also recorded and ranged from 10 to 66 minutes (mean, 26 minutes). Uncomplicated intraoperative gastrojejunostomy tube placement was achieved in 25 of 26 dogs. A distal duodenotomy was required in 1 dog with severe pancreatitis to facilitate insertion of the jejunostomy tube beyond the level of the duodenocolic ligament. The tube insertion time for this dog was 16 minutes longer than any other dog within the study. If this outlier is removed, the mean tube insertion time decreases to 24 minutes. Total surgical time or gastrojejunostomy tube placement time did not have a significant association with the incidence of mechanically mediated or tube stoma–related complications or mortality rate (Table 4).

The overall mechanical tube complication rate was 46% (12/26), including coiling (n = 7), migration (4), and kinking (2). One dog was documented to have both migration and coiling; however, this was recorded as a single mechanical complication event. Knotting of the tube was not identified in any dog. Significant differences were not identified between the preoperative variables and the presence of mechanical tube complications. Median time to mechanical complication was not reached by ≥ 572 days after surgery, with 42% (5/12) of dogs having a mechanical complication by day 6 after surgery. Mechanical tube complications requiring revision surgery were not identified. Mean ± SD distance that the gastrojejunostomy tube was inserted beyond the level of the duodenocolic ligament (into the jejunum) was 16.75 ± 9.81 cm (6.6 ± 3.9 inches) with a range of 0.0 to 30.47 cm (0.0 to 12 inches). Associations between tube placement distance and mechanical complications were not identified.

No major tube site complications were documented throughout the duration of the study period. The overall minor tube stoma complication rate was 77% (20/26) and included erythema (n = 16), cellulitis (13), and discharge (17). An overall minor tube stoma event was recorded if a dog had 1 of the 3 recognized abnormalities; however, when overlap within the stoma complication groups was observed, this was not considered an additional stoma complication event. Median time to a minor tube stoma complication was 1 day. In dogs with erythema and cellulitis, stoma abnormalities were generally mild, with only 2 grade 3 changes assigned to dogs with erythema and 1 grade 3 change to dogs with cellulitis. In dogs with discharge, no grade 3 changes were observed on any stoma evaluation throughout the duration of hospitalization. Grade 2 (moderate severity) changes were observed in ≤ 5 tube sites for each category of complications throughout the duration of hospitalization. No variables were found to significantly contribute to a tube stoma complication. All stoma complications documented were self-limiting, and no revision procedures or deaths resulting from tube placement were identified.

Initiation of enteral nutritional supplementation via the jejunostomy tube was started within 24 hours of surgery in 24 of 26 (92%) dogs. In the 2 dogs where enteral feeding was not initiated within 24 hours, both of these dogs died within the first several hours of surgery, and therefore jejunostomy tube feeding was never started. The percentage of RER fed during the first 24 hours after surgery ranged from 11% to 44% (mean, 26.2%). If jejunal feedings were tolerated, the RER was generally increased by 25% to 33% every 24 hours. The percentage of calculated RER fed during the 48and 72-hour postoperative time points ranged from 11% to 75% (mean, 48%) and 11% to 80% (mean, 56%), respectively.

Residual gastric fluid volume was aspirated and recorded in 23 of 26 (88%) dogs during the first 24 hours after surgery. The dual lumen port at the end of the jejunostomy tube allowed for simultaneous access to the stomach and small intestine. Gastric residual volume ranged from 0 to 28.4 mL/kg/d (0 to 12.9 mL/lb/d) with a mean of 12.0 mL/kg/d (5.5 mL/lb/d) during the first 24 hours after surgery. Gastric residual volume during the 48and 72-hour postoperative period ranged from 0 to 30.3 mL/kg/d (0 to 13.8 mL/lb/d) with a mean of 6.5 mL/kg/d (3.0 mL/lb/d) and from 0 to 15.8 mL/kg/d (0 to 7.2 mL/lb/d) with a mean of 4.3 mL/kg/d (2.0 mL/ lb/d), respectively. No clogging of the gastrostomy or jejunostomy tube was identified.

Table 4—

Intraoperative variables assessed for associations with mechanical and tube stoma complications.

View Table

Both gastrostomy and jejunostomy tubes were tolerated throughout the duration of usage. Dislodgement or self-induced tube trauma resulted in accidental tube removal in 2 of 26 dogs, and inadvertent tube damage resulted in premature removal by the clinician in 1 of 26 dogs. The duration of jejunostomy tube usage, prior to removal, was available for review for 16 of 26 (62%) dogs and ranged from 1 to 9 days (mean, 4.4 days). Planned jejunostomy tube removal was executed after cessation of gastrointestinal signs or when a repeated tolerance to oral or gastrostomy tube feedings was demonstrated. Ten of 26 (38%) dogs died prior to jejunostomy tube removal and were not included for analysis. In the first of the 2 dogs with accidental dislodgement, the end of the jejunostomy tube was stepped on by the dog on day 2 after surgery. In this dog, jejunostomy tube replacement was not necessary, as medical therapy and supportive care were sufficient to meet nutritional requirements. The duration of gastrostomy tube usage, prior to removal, was available for review for 11 of 26 (42%) dogs and ranged from 9 to 41 days (mean, 20.5 days). Planned gastrostomy tube removal was executed after a minimum of 14 days to ensure maturation of the gastrocutaneous stoma. The second dog with accidental tube removal ingested a portion of the gastrostomy tube on day 9 after surgery, leaving the dome of the gastrostomy tube free-floating within the stomach. Immediate endoscopic retrieval was performed under general anesthesia, and the dog recovered without complication. In the dog with premature tube removal by the clinician, the gastrostomy tube was found to be partially torn at its base on day 13 after surgery. The tube was manually removed without complication, as the dog was reported to be eating well on its own.

Sixteen of 26 (62%) dogs survived the immediate postoperative period and were discharged from the hospital. Duration of hospitalization ranged from 2 to 10 days (mean, 5.9 days). Three of 9 dogs receiving surgery for pancreatic disease were discharged from the hospital. Of the 17 dogs undergoing gastrointestinal or biliary surgery, 9 and 4 were alive at the time of hospital discharge, respectively. The Kaplan-Meier median survival time for the 26 dogs enrolled in this study was 39 days, with 13 dogs still alive. Of the 13 dogs still alive, 8 were alive at 1 year after surgery. Progression of disease was the cause of death in the 5 dogs that died within the first year after surgery.

Discussion

To our knowledge, the present study represents the first clinical series evaluating the use of and complications associated with surgical placement of gastrojejunostomy tubes in dogs. The indications for gastrojejunostomy tube placement in this study were similar to previous reports^12,13 of surgically placed jejunostomy tubes and included severe gastrointestinal disease, extrahepatic biliary surgery, and pancreatic disease. Reported complications of surgically placed jejunostomy tubes include tube site cellulitis, tube occlusion and dislodgement, leakage around the tube, and intestinal perforation.^3,12,13 Modified jejunostomy tube placement techniques as well as nonsurgical (minimally invasive) techniques have also been described.^{13,14,17–21} A recent study¹⁶ on percutaneous endoscopic placement of gastrojejunostomy tubes has been performed in healthy dogs and cats. The major complication rate associated with this type of enteral feeding device was low, and the tubes were well tolerated.¹⁶ In humans, surgically placed gastrojejunostomy tubes have been shown to have fewer complications, compared with surgically placed jejunostomy tubes.⁹ In the aforementioned study, premature dislodgement resulting in enteric leakage and peritonitis occurred in 7 of 92 jejunostomy tube placements and 0 of 56 gastrojejunostomy tube placements, with this difference being significant.⁹ Proposed surgical advantages of gastrojejunostomy tubes, compared with jejunostomy tubes, include improved anastomotic healing of the tube stoma with greater security of suture placement, allowance for a larger bore tube, and dual gastric and intestinal lumen access. Additionally, avoidance of an enterotomy for placement of the jejunostomy tube significantly reduces the risk of catastrophic complications such as peritonitis after premature tube dislodgement.^9,11

The overall mechanical tube complication rate in our study was 46% (12/26), with coiling and migration of the jejunostomy tube within the stomach or small intestine being the most common. Kinking of the jejunostomy tube, occurring in 8% (2/26) of dogs, was uncommon and resulted in no clinically detectable abnormalities. It has been reported for human patients that incompletely inserted tubes beyond the level of the duodenocolic ligament (into the jejunum) may be predisposed to coiling, as tubes have a tendency toward proximal migration.²² In addition, incompletely inserted tubes often result in redundant lengths of jejunostomy tube remaining within the gastric lumen after tube placement. Once redundant lengths of jejunostomy tube are present within the stomach, coiling is a likely sequela. Therefore, during gastrojejunostomy tube insertion, it is imperative that the entire length of the jejunostomy tube be positioned primarily within the small intestine, leaving only a small segment of the jejunostomy tube within the stomach. Placement of the gastrostomy tube at the level of the pyloric antrum, with a subsequent right-sided tube gastropexy, may facilitate a reduction in redundant lengths of the jejunostomy tube in the stomach. The effect of this placement site on pyloric function is unknown, and tube insertion at this site may cause a 90° bend in the jejunostomy tube, thus predisposing to mechanical complications. In accordance with previously reported techniques and for the purposes of this study, a left-sided tube gastropexy was performed in all dogs with gastrostomy tube insertion in the body of the stomach.²¹ In future studies, the efficacy of gastrostomy tube placement site modification can be evaluated for its association with a reduction in mechanical complications.

Postoperative radiography or intraoperative fluoroscopy (if necessary) can be used to confirm accurate placement of the gastrojejunostomy tube. In general, intraoperative palpation and measurement of the jejunostomy tube end piece is readily achievable and sufficient for confirmation of appropriate placement. For the purposes of this study, daily evaluation of ventro dorsal radiographic images of the peritoneal cavity was performed to document and track the progression of mechanical complications. In most situations, a single ventrodorsal projection proved adequate for observation and documentation of complications. A single lateral radiographic projection is generally inadequate for assessment of tube positioning and mechanical complications.²¹ In retrospect, for the purposes of this study, daily evaluation of orthogonal radiographic images would have been preferred. This would have been particularly useful for documentation of jejunostomy tube coiling within the small intestine. Coiling within the stomach was readily apparent with a single ventrodorsal projection; however, because of its tortuous nature, coiling within the small intestine was more difficult to confirm. In future clinical situations where small intestinal coiling is suspected, it would be advisable to obtain orthogonal projections of the abdomen. Contrast radiography has also been shown to be an effective means by which mechanically mediated jejunostomy tube complications can be confirmed.²² Because routine contrast imaging was not performed in this study, specific complications such as leakage between the stomach lumen and external tube stoma site may have been missed. This is unlikely, however, because extensive postoperative monitoring was performed and clinical signs consistent with a gastrocutaneous fistula were never observed.

In previous gastrojejunostomy and jejunostomy tube studies,^12,14,16 stoma site complication rates were 8% (1/12) and 9% (3/32), respectively. The lower incidence of gastrojejunostomy tube stoma complications, reported previously, may be a reflection of the fact that healthy research animals comprised their study population. Additionally, the substantially higher incidence of minor tube stoma abnormalities in this study may have been a manifestation of its prospective nature, assuring more vigilant and accurate notation of even the mildest abnormality.

An attractive feature of the tubes used for this study was the plastic external bolster with self-retaining ring that is designed to slip over the top of the tube and interface with the tube exit site from the body wall. This device fixes the gastrostomy tube in place and prevents the tube from being pulled further into the stomach. Additional purse string or finger-trap sutures, which may predispose to tube stoma abnormalities, are not necessary if the plastic bolster and ring are used correctly.¹² In general, a 5-mm gap should be left between the plastic bolster and body wall, as postoperative stoma swelling can cause impingement of the skin onto the bolster, resulting in pressure necrosis. Varying degrees of this complication were identified during the early phases of the study period, likely contributing to an increase in the frequency of tube stoma abnormality scores. None of the variables assessed in the present study were found to significantly contribute to a tube stoma complication. All stoma complications documented were self-limiting, and no revision procedures or deaths due to tube placement were identified.

Tube occlusion after feeding is reported to occur in 3% to 12.7% of surgically placed jejunostomy tubes.^12–14 Occlusion of the jejunostomy tube was not observed in this study. The gastrojejunostomy tube method allows for placement of a larger-diameter jejunostomy tube, compared with those used in surgically placed jejunostomy tubes. Prevention of tube occlusion was also likely accomplished as a result of constant rate infusion feeding of a liquid diet intermixed with intermittent lavage of the tube with tap water. Bolus feeding in dogs with gastrojejunostomy tubes has been reported without an increased incidence of tube occlusion complications.¹⁶

The gastrojejunostomy tube placement technique allows for simultaneous access to the small intestine for feeding and to the stomach for medication administration, feeding, or gastric decompression. Gastroparesis is a common complication after anesthesia, after abdominal surgery, and in cases of severe abdominal disease. Gastric access for decompression is essential, as increased gastric residual volumes interfere with the normal absorptive capacity of the stomach and contribute to an increase in gastroesophageal reflux and vomiting, thus predisposing to aspiration pneumonia.²³ Normal values for postoperative gastric residual accumulation in sick animals have not been established. An approximate 3-fold reduction in mean residual amounts was identified from day 1 to 3 after surgery. Dogs with only gastrostomy tubes would most likely be unable to tolerate feedings with the quantity of residual volumes observed within the first 12 to 36 hours after surgery. In dogs with only a jejunostomy tube, gastric residual amounts cannot be quantified. Therefore, another added advantage of the gastrojejunostomy tube is simultaneous jejunal feeding with access to the gastric lumen for decompression and quantification of gastric residual amounts. Postoperative tailoring of prokinetic treatment can also be enhanced through interpretation of gastric residual quantification data. Guidelines for gastrojejunostomy tube removal have not been firmly established for dogs. In our study, jejunostomy tubes were generally removed when clinical variables associated with gastroparesis (consistent decrease in gastric residual volumes or frequency of vomiting) had resolved and tolerance to oral or gastrostomy tube feeding was achieved.

Reported surgical placement times for jejunostomy and gastrojejunostomy tubes in humans ranges from 15 to 20 minutes.⁹ Mean gastrojejunostomy tube insertion time in this study was 26 minutes, including 1 dog with severe pancreatitis that required a distal duodenotomy to allow palpation and manipulation of the tube through severely thickened and corrugated intestine, suggesting that gastrojejunostomy tube insertion time in dogs compares favorably with that of humans. To our knowledge, published insertion times for surgically placed jejunostomy tubes are not available in the veterinary literature. In a recent study¹⁶ of percutaneous, endoscopically placed gastrojejunostomy tubes in healthy dogs, the mean insertion time was 45 minutes. This suggests that insertion times for surgically placed gastrojejunostomy tubes are more rapid than nonsurgical placement.

In this study, the overall mortality rate was high with short median survival times. All dogs receiving a gastrojejunostomy feeding tube were critically ill, with many requiring emergency surgical intervention to correct the underlying primary disease. Suspected cause of death was recorded in all dogs on the basis of clinical signs, objective monitoring variables, hematologic abnormalities, and histopathologic documentation of the underlying etiology of the primary disease. Of the recorded complications associated with gastrojejunostomy tube usage, none were severe enough to have contributed to death or even a revision tube placement procedure.

In conclusion, surgical placement of gastrojejunostomy tubes is an effective method of providing postoperative enteral nutritional supplementation in critically ill dogs when postgastric delivery is anticipated because of gastric paresis, prolonged recumbency, or increased risk for aspiration pneumonia and for nutritional bypass of the pancreatic ducts. Staged jejunal to gastric feeding, with the ability to control and quantify gastric residual accumulations, is an additional benefit to the gastrojejunostomy feeding method. The uncomplicated surgical technique required for gastrojejunostomy tube placement results in rapid tube insertion with a low rate of serious postoperative complications.