- Email: [email protected]
Radiologic Peroral Gastrostomy
Radiologic Peroral ~astrostorn~' John A. Clark, MDCM, FRCPC Robyn A. Pugash, MD, FRCPC Rose R. Pantalone, RN
Index terms: Feeding tubes Gastrostomy
JVIR 1999; 10:927-932 Abbreviations: PEG = percutaneous endoscopic gastrostomy, RPG = radiologic peroral gastrostomy.
PURPOSE: To develop an improved percutaneous technique for the insertion of large-bore gastrostomy tubes. MATERIALS AND METHODS: With use of fluoroscopic guidance, the stomach is punctured and the esophagus is catheterized in a retrograde fashion. A guide wire is passed from the gastrostomy site, up the esophagus, and out of the patient's mouth. A large-bore (20-24 F) endoscopic push-type gastrostomy tube is advanced over the wire, through the mouth, down the esophagus, and out of the gastrostomy site. RESULTS: Thirty-one successful tube placements were performed in 32 attempts (97% success rate). There were no major procedural or postprocedure complications. Minor complications included one lip laceration (one of 31 = 3%),one minor exit site infection (one of 31 = 3%),and two inadvertent tube dislodgements (two of 31 = 6%). CONCLUSIONS: Radiologic placement of large-bore endoscopic gastrostomy tubes is possible without endoscopy. The procedure is rapid, easy to perform, and safe.
From the Department of Medical Imaging, University of Toronto, St. Michael's Hospital, Wellesley Central Site, 160 Wellesley Street East, Toronto, Ontario, Canada M4Y 153. From the 1999 SCVIR annual meeting. Received November 24, 1998; revision requested December 28; revision received and accepted February 11, 1999. Address correspondence to J.A.C. O SCVIR, 1999
SINCE first introduced in 1980 (11, percutaneous endoscopic gastrostomy (PEG) has become the most common method of long-term enteral feeding tube insertion in the United States (2). Percutaneous fluoroscopically guided techniques have also been developed (3-5) but are generally not as popular, despite the potential for lower morbidity and mortality (6-9). One common criticism of radiologically inserted gastrostomy tubes is the small tube size, typically 12-14 F in diameter. Endoscopically placed tubes typically range from 18 to 28 F in diameter and are less prone to failure caused by blockage or kinking. In addition, the design of most endoscopically placed tubes includes a large internal bolster or bulb as the retention device, making them less prone to inadvertent removal than the Cope loop, balloon, or Malecot-type tubes typically placed by radiologists. We describe a radiologic method of insertion of endoscopic-type gastrostomy tubes using fluoroscopic guidance that does not require en-
doscopy, a procedure we call radiologic peroral gastrostomy (RPG).
I MATERlALS AND METHODS We attempted RPG in 30 patients referred to our department between November 1997 and November 1998. Patients were selected for this procedure if there was deemed to be a high risk of inadvertent tube removal (such as with agitated or delirious patients) or if the referring physician specifically requested a large-bore tube. The presence of known gastroesophageal reflux disease or a high-grade esophageal stricture were considered absolute contraindications, and, in these cases, traditional radiologic gastrostomy or gastrojejunostomy was performed. Patients were forbidden food by mouth, and nasogastric tube feedings were withheld starting at midnight the night before the procedure. Following a standard protocol for endoscopically placed tubes (101, a single dose of a prophylactic antibiotic (1g cefazolin intravenous)
928
Radiologic Peroral Gastrostomy
July-August 1999 JVIR
b. Figure 1. (a)A 6.5-F hockey stick catheter is tapped against the nasogastric tube at the gastroesophageal junction. (b) The guide wire is advanced to engage the gastroesophageal junction and the esophagus is catheterized. a.
Figure 2. An endoscopic push-type gastrostomy tube (Flow 20; Cook). Note the long gradually tapered leading end.
was given immediately before the procedure unless other antibiotics were already being given. After appropriate conscious sedation with intravenous fentanyl and midazolam, a nasogastric tube was inserted and the stomach was distended with air. The colon was identified fluoroscopically, if necessary, with rectal insufflation of air,
and verified to be inferior to the distended stomach. The anterior position of the gastric wall was verified with lateral fluoroscopy. Ultrasound or CT was not routinely used. A hypotonic agent (glucagon, buscopan) was administered only if gastric decompression was a problem. The stomach was then punctured with a n 18-gauge trocar-type needle
in the midbody region, using a trajectory perpendicular to the anterior abdominal wall. A small amount of water-soluble contrast material was injected to confirm proper needle position. The needle was then exchanged for a 6.5-F hockey sticktype catheter (MPA; Cook, Bloomington, IN). With use of a 260-cm, 0.035-inch Amplatz Super Stiff wire (Cook), the esophagus was catheterized retrogradely. Catheterization of the esophagus was aided by withdrawing the nasogastric tube tip to the gastroesophageal junction and using this as a visual and tactile aid to identification of the gastroesophageal junction. The catheter was manipulated to lie parallel to the nasogastric tube tip, while tapping against it. The wire was then advanced to engage the gastroesophageal junction (Fig 1). This procedure was sometimes facilitated by partial deflation of the stomach and by inflation of air into the distal esophagus by way of the nasogastric tube. The guide wire and catheter were then advanced up the esophagus and out of the patient's mouth. A lubricated endoscopic pushtype gastrostomy tube (Flow-20push; Wilson-Cook, Winston-Salem, NC) was then advanced over the wire, into the patient's mouth, and down the esophagus. This type of tube has a long, relatively stiff, and very gradually tapered dilator as a n integral part of the leading end of the tube (Fig 2). The leading edge of the dilator was advanced out of the gastrostomy site, where it was manually grasped and the tube was pulled the remainder of the way. When the internal bolster of the gastrostomy tube entered the patient's mouth, the tube was given a short and sharp pull so the bolster passes rapidly into the esophagus, causing a minimal gag reflex. Under fluoroscopic guidance, the tube was then pulled into final position, with the internal bolster gently snugged up against the anterior gastric wall, providing good apposition of the stomach and anterior abdominal wall (Fig 3). The tube was then secured with an external bolster, the dilator por-
Clark et a1
929
Volume 10 Number 7
Figure 4. Final tube position confirmed with injection of water-soluble contrast material.
a.
b.
Figure 3. Peroral gastrostomy tube being pulled out of the gastrostomy site. (a) The internal bolster is in the distal esophagus. Lateral spot view of the anterior abdominal wall showing the final position of tube. (b) The internal bolster gently pulls the anterior stomach wall Bgainst the anterior abdominal wall.
tion of the tube was cut off, and a universal feeding adapter was fitted. Final tube position was confirmed with a n injection of watersoluble contrast material (Fig 4). As per the protocol of the endoscopy unit a t our institution, feeding was started 6 hours after completion of the procedure. In a small number of patients (four), RPG was performed after inadvertent removal of a previous gastrostomy tube and before complete tract maturation. In these cases, the skin tract to the stomach was cannulated directly with a guide wire and hockey-stick catheter. The procedure was then completed as previously described.
I RESULTS Thirty-one RPGs were successfully performed in 32 attempts in 30 patients (97% success rate). Placement was unsuccessful in one patient because of excessive binding of the gastrostomy tube when attempting to pull it through an existing gastric tract. The tract had a n unusually tortuous course before entering the stomach. Because we were concerned about causing significant trauma, the placement in this patient was abandoned and a standard Cope loop-type catheter was inserted without difficulty. In all patients, gastric access was
readily achieved. The esophagus was also successfully catheterized in all patients. In a few of the early cases, a wire snare was passed down the nasogastric tube and used to capture the guide wire within the stomach. However, with some minimal experience, retrograde catheterization of the gastroesophageal junction was found to be faster. There were no major procedural complications (one of 31; 3%). One patient had a minor lip laceration, which required no specific therapy. There weEe no othe; procedural There were major postprocedure . . complications. . NO patient developed pneumonia within 2 weeks of the ~ r o c e d u r e One patient had modeiate erythema surrounding the tube site 7 days after insertion, which may have been a n early wound infection (one of 31; 3%).This patient received antibiotic therapy for a n unrelated condition, and the erythema resolved. Two patients dislodged their tubes, one 8 days after insertion and the other 10 days after insertion (two of 31; 6%). These tubes were replaced by means of the same technique, one 24 hours later, the other 4 days later. There were no clinical sequelae caused by tube dislodgment. Current follow-up on these patients varies from l to 11 months. There were no tube occlusions. Four tubes were removed when patients '
Radiologic Peroral Gastrostomy
930
July-August 1999 JVIR
regained oropharyngeal control; this was done at bedside with external traction. Two tubes were converted to skin-level button gastrostomies after tract maturation. The external portions of two tubes were damaged by the nursing staff 9 months after placement; these tubes were replaced by a balloon-type catheter. Two patients died in the follow-up period of natural causes or of their underlying disease; no deaths were directly or indirectly attributable to gastrostomy placement.
the type of retaining mechanism present, and to the mental status of the patient; Cope loop-type or balloon tubes are more Drone to retention device failure, a; the Cope loop retaining string can break, the balloon can rupture, or the balloon valve can leak. These tubes are then more easily removed than are endoscopic tubes with a permanent internal bolster. When tubes are dislodged before tract maturation, there is a significant risk of gastric leakage and peritonitis. More secure tubes are, therefore, safer and reauire fewer interventions. Studies reported in the literature vary DISCUSSION widelv as to the re~ortedincidence Percutaneous gastrostomy, as of tudk occlusion aAd dislodgment (8,141; comparison between these performed by radiologists or endoscopists, is an accepted and proven studies is difficult as they cannot be technique, with significant benefits controlled for the quality of nursing compared to surgical techniques. care given and the mental status of Percutaneous techniques are faster, the patient. cheaper, and are associated with The literature suggests that lower morbidity and mortality than there is a lower major complication surgical gastrostomy (11-13). Howrate with radiologic gastrostomy ever, published literature does not than with endoscopic techniques (7,9). This may be caused by several demonstrate large differences between endoscopic and radiologic factors. First, image-guided placement allows precise localization of techniques. In all likelihood, radiothe liver and colon, and, therefore, logic gastrostomy is cheaper, has a higher success rate, and has a lower complications caused by transgresmajor complication rate, whereas sions of these structures by radiologic techniques are very rare. Secendoscopic gastronomy has a lower ond, diagnostic endoscopy alone is incidence of tube-related complications (6-9).. associated with a low, but signifiThere are many potential benecant, major complication rate. The fits to placing endoscopic-type gasomission of the endoscope may actrostomy tubes. First, the ability to count in large part for the lower place a large-bore tube reduces procedural complication rate of ratube-related complications, mainly diologic gastrostomy. occlusion and dislodgment. The rate The ease with which the gastrosof tube occlusion is primarily retomy tract is dilated during placelated to the size of the inner bore of ment of an endoscopic-type gastrosthe tube, the quality of the nursing tomy tube is a significant benefit, care given, and the physical charac- and it mav lead to a safer Droceteristics of the material adminisdure. ~ h e ~ t r adilation ct ocdurs in a tered via the tube. Large-bore tubes single step, avoiding serial dilations require less meticulous nursing care and the associated risk of leakage of than do smaller bore tubes, as adgastric contents with each exchange ministration of poorly crushed med- of a dilator. Dilation with endoscopication and the lack of vigorous ic-type tubes also proceeds in an tube flushing frequently contribute internal-to-external direction, proto the occlusion of gastrostomy viding continuous apposition of the tubes. Large-bore tubes do not kink stomach wall to the anterior abdomor clog as easily as small-bore tubes inal wall. This means that gastric and, thus, do not require as freaccess is never lost inadvertently by quent replacement. The rate of tube a dilator pushing the stomach wall dislodgment is strongly related to away from the abdominal wall. This ,
is also achieved without the use of gastropexy devices, such as T fasteners or anchor sutures. This apposition of gastric and abdominal walls is also maintained after tube placement by the presence of an internal bolster, leading to a shorter and straighter percutaneous tract; conversion of the gastrostomy to a skin-level button is therefore significantly easier. The cost of RPG is virtually identical to the cost of standard radiologic gastrostomy. Although there is much confusion in the literature with regard to actual procedural costs versus procedural charges, radiologic gastrostomy is likely cheaper (2,16-19). Procedural cost for RPG in our department, for fluoroscopic room time, nursing and technical support, and all necessary materials (drugs, needles, wires, gastrostomy tube, and so forth) is U.S. $412. The total professional fee charged (as decreed by the provincial health insurance) is U.S. $108. The total fee of U.S. $520 compares favorably to the total cost for PEG placement, from U.S. $1,800 to U.S. $3,300. Typical procedure time for the tube placements performed in this study was 25 minutes. With experience, many tube placements were performed in 15 minutes. This compares favorably to other techniques (6). Only one operator is required for RPG, whereas PEG placement requires two. There are some potential drawbacks to this type of tube placement. Endoscopic gastrostomy is typically associated with a higher incidence of postprocedure pneumonia than standard radiologic gastrostomy (7). This may be related to aspiration a t the time of the procedure, perhaps caused by the use of topical anesthesia in the pharynx, or the presence of a foreign body (the endoscope or the gastrostomy tube itself). RPG may be associated with a higher incidence of pneumonia as well; however, the lack of an endoscope, and hence the reduced need for topical anesthesia, likely leads to a lower incidence of pneumonia than with PEG. There were no clinical pneumonias associated
Clark et a1
931
Volume 10 Number 7
with gastrostomy tube placement in this study. Endoscopic gastrostomy is also associated with a higher incidence of wound infection than are traditional radiologic techniques (8,lO). This is thought to be related to contamination of the gastrostomy tube with oropharyngeal bacteria as it passes through the mouth. For this reason. most endosco~istsadminister a single dose of a prophylactic antibiotic immediately before the procedure. RPG will likely have a similar incidence of wound infection and, therefore, also warrants antibiotic prophylaxis. Conversely, the incidence of wound infection during endoscopic gastrostomy may be at least partly related to the Dresence of the endosco~eand the multiple passes of the endoscope and tube typically performed. The true incidence of wound infection with RPG remains to be determined, but it likely lies between that for endoscopic gastrostomy and that for standard radiologic gastrostomy. Finally, peroral techniques are slightly less comfortable for the patient because of stimulation of the gag reflex with (brief) oral passage of the tube, and consequently may require slightly deeper conscious sedation than with traditional radiologic techniques. The endoscopic-type gastrostomy tubes used in this study were designed for removal by external traction. The recommended method involves attempting to collapse the internal bolster by pinching it through the anterior abdominal wall and simultaneously applying continuous gentle external traction. In most cases, this worked well. If the tube was not easily removable by this method, we used the technique recommended by Korula and Harma (20). This involves cutting the tube at skin level and allowing the internal portion to fall into the stomach and be eliminated intestinally. We had no complications associated with either method. The indications we currently use for RPG as compared with a traditional radiologic gastrostomy are a specific request from the clinician for a large-bore tube; a patient deemed at high risk for pulling out
a gastrostomy; and patients for whom the gastrostomy will be converted to a skin level button. The contraindications we use for RPG placement are the presence of an eophageal occlusion or the presence of known significant gastroesophageal reflux. Although as yet unknown, the presence of significant ascites may be a relative indication for RPG because of its superior apposition of the stomach and anterior abdominal wall. A similar technique for the peroral placement of gastrostomy tubes was described in dogs (21) and has been adopted in children (22). It would seem that this technique is ideally suited to adults. Most adults for whom a gastrostomy is indicated have an impaired gag reflex, and therefore oral passage of the tube causes minimal discomfort and little or no physiologic response. The adult esophagus also easily accommodates the internal bolster of the tube. Delirious adults are more likely to have the strength required to remove a Cope-loop type of tube and, therefore, benefit from the retention provided by a large internal bolster. Endoscopic-type gastrostomy tubes may be purchased as "push" or "pull" type tubes. Push-type tubes were used almost exclusively in this study. Pull-type tubes require an additional catheter exchange to pass a special loop wire necessary for attachment to the gastrostomy tube. For this reason, we found push-type gastrostomy tubes faster to dace. In addition. the dilator of push-type ' tubes is longer and more gradually tapered, allowing easier and gentler tract dilation. Other reports have documented similar success to ours with the placement of pull-type gastrostomy tubes (23,241; the choice of pull- or push-type tubes is likely personal.
CONCLUSION Large-bore endoscopic-type gastrostomy tubes may be placed safely and easily using fluoroscopy and standard interventional radiologic
techniques. Further studies are needed to confirm the high success rate and low complication rate. The American Society for Gastrointestinal Endoscopy claims on its World Wide Web page that "[rladiologic placement of gastrostomy tubes ... is infrequently used and comparison to PEG regarding safety, efficacy and cost is not available" (25). Although the published literature may already dispute this claim, there is clearly a need for further comparison of endoscopic and radiologic techniques. It is estimated that feeding tube insertion is the second most common indication for upper gastrointestinal endoscopy in hospitalized patients, with 74,000 tube placements performed annually in the United States. With this potential volume, organization of large trials should not be problematic. RPG preserves the best of traditional radiologic and endoscopic gastrostomy; it maintains much of the safety of traditional radiologic gastrostomy and provides the benefit of a large-bore endoscopic-type tube. Although not indicated for all patients, it should prove useful in our armamentarium. References 1. Gauderer MWL, Ponsky JL, Izant RJ. Gastrostomy without laparotomy: a percutaneous endoscopic technique. J Ped Surg 1980; 15:872875. 2. Wollman B, D'Agostino HB, WalusWigle JR, Easter DW, Beale A. Radiologic, endoscopic and surgical gastrostomy: an institutional evaluation and meta-analysis of the literature. Radiology 1995; 197:699-704. 3. Tao HH, Gillie RR. Percutaneous feeding gastrostomy. AJR 1983; 149: 793-794. .- - . - - .
4. Ho CS. Percutaneous gastrostomy for jejunal feeding. Radiology 1983; 149:595-596. 5. Wills JS, Oglesby JT. Percutaneous gastrostomy. Radiology 1983; 149:449-453. 6. Miller RE, Castlemain B, Lacqua FJ, Kotler DP. Percutaneous endoscopic gastrostomy: results in 316 patients and review of literature. Surg Endosc 1989; 3:186-190. 7. Taylor CA, Larson DE, Ballard DJ, et al. Predictors of outcome after
932
Radiologic Peroral Gastrostomy
July-August 1999 JVIR
percutaneous endoscopic gastrostomy: a community-based study. Mayo Clin Proc 1992; 67:1042-1049. 8. McLoughlin RF, So B, Gray RR. Fluoroscopically guided percutaneous gastrostomy: current status. CARJ 1996; 47:lO-15. 9. Munk PL, Lee MJ, Poon PY, et al. Percutaneous gastrostomy in radiologic practice. Austral Radiol 1997; 41:342-350. 10. Jain NK, Larson DE, Schroeder KW, et al. Antibiotic prophylaxis for percutaneous endoscopic gastrostomy. Ann Int Med 1987; 107:824828. 11. Stern JS. Comparison of percutaneous endoscopic gastrostomy with surgical gastrostomy a t a community hospital. Am J Gastroenterol 1986; 81:1171-1173. 12. Grant JP. Comparison of percutaneous endoscopic gastrostomy with Stamm gastrostomy. Ann Surg 1988; 207:598-603. 13. Kummer BA, Tiszenkel HI, Kotler DP, Miller RE. Percutaneous endoscopic gastrostomy: procedure of choice. Gastrointest Endosc 1985; 31:156-157
14. Ryan JM, Hahn PF, Boland GW, McDowell RK, Saini S, Mueller PR. Percutaneous gastrostomy with T-fastener gastropexy: results of 316 consecutive procedures. Radiology 1997; 203:496-500. 15. McLoughlin RF, Gibney RG, Fluoroscopically guided percutaneous gastrostomy: tube function and malfunction. Abdom Imaging 1994; 19:195-200. 16. Goretsky MJ, Johnson N, Farrell M, Ziegler MM. Alternative techniques of feeding gastrostomy in children: a critical analysis. J Am Coll Surg 1996; 182:233-240. 17. Clark JA. Cost of radiologic gastrostomy [letter]. Radiology 1996; 199:582. 18. Bell SD. Cost of performing a radiologic percutaneous gastrostomy [letter]. Radiology 1996; 200:586. 19. Rosenzweig TB, Palestrant AM, Esplin CA, Gilsdorf RB. A method for radiologic-assisted gastrostomy when percutaneous endoscopic gastrostomy is contraindicated. Am J Surg 1994; 168:587-591. 20. Korula J , Harma C. A simple and inexpensive method of removal
21.
22.
23,
24.
25.
or replacement of gastrostomy tubes. JAMA 1991; 265:14261428. Sacks BA, Vine HS, Palestrant AM, Ellison HP, Shropshire D, Lowe R. A nonoperative technique for establishment of a gastrostomy in the dog. Invest Radiol 1983; 18:485487. Towbin RB, Ball WS, Bissett GS. Percutaneous gastrostomy and percutaneous gastrojejunostomy in children: antegrade approach. Radiology 1988; 168:473-476. Rosenzweig TB, Palestrant AM, Esplin CA, Gilsdorf RB. A method for radiologic-assisted gastrostomy when percutaneous endoscopic gastrostomy is contraindicated. Am J Surg 1994; 168:587-591. Szymski GX, Albazzaz AN, Funaki B, et al. Radiologically guided placement of pull-type gastrostomy tubes. Radiology 1997; 205:669-673. American Society for Gastrointestinal Endoscopy. Upper GI endoscopy: endoscopic feeding tubes. http://www.asge.org/doc/l71.
Index terms: Feeding tubes Gastrostomy
JVIR 1999; 10:927-932 Abbreviations: PEG = percutaneous endoscopic gastrostomy, RPG = radiologic peroral gastrostomy.
PURPOSE: To develop an improved percutaneous technique for the insertion of large-bore gastrostomy tubes. MATERIALS AND METHODS: With use of fluoroscopic guidance, the stomach is punctured and the esophagus is catheterized in a retrograde fashion. A guide wire is passed from the gastrostomy site, up the esophagus, and out of the patient's mouth. A large-bore (20-24 F) endoscopic push-type gastrostomy tube is advanced over the wire, through the mouth, down the esophagus, and out of the gastrostomy site. RESULTS: Thirty-one successful tube placements were performed in 32 attempts (97% success rate). There were no major procedural or postprocedure complications. Minor complications included one lip laceration (one of 31 = 3%),one minor exit site infection (one of 31 = 3%),and two inadvertent tube dislodgements (two of 31 = 6%). CONCLUSIONS: Radiologic placement of large-bore endoscopic gastrostomy tubes is possible without endoscopy. The procedure is rapid, easy to perform, and safe.
From the Department of Medical Imaging, University of Toronto, St. Michael's Hospital, Wellesley Central Site, 160 Wellesley Street East, Toronto, Ontario, Canada M4Y 153. From the 1999 SCVIR annual meeting. Received November 24, 1998; revision requested December 28; revision received and accepted February 11, 1999. Address correspondence to J.A.C. O SCVIR, 1999
SINCE first introduced in 1980 (11, percutaneous endoscopic gastrostomy (PEG) has become the most common method of long-term enteral feeding tube insertion in the United States (2). Percutaneous fluoroscopically guided techniques have also been developed (3-5) but are generally not as popular, despite the potential for lower morbidity and mortality (6-9). One common criticism of radiologically inserted gastrostomy tubes is the small tube size, typically 12-14 F in diameter. Endoscopically placed tubes typically range from 18 to 28 F in diameter and are less prone to failure caused by blockage or kinking. In addition, the design of most endoscopically placed tubes includes a large internal bolster or bulb as the retention device, making them less prone to inadvertent removal than the Cope loop, balloon, or Malecot-type tubes typically placed by radiologists. We describe a radiologic method of insertion of endoscopic-type gastrostomy tubes using fluoroscopic guidance that does not require en-
doscopy, a procedure we call radiologic peroral gastrostomy (RPG).
I MATERlALS AND METHODS We attempted RPG in 30 patients referred to our department between November 1997 and November 1998. Patients were selected for this procedure if there was deemed to be a high risk of inadvertent tube removal (such as with agitated or delirious patients) or if the referring physician specifically requested a large-bore tube. The presence of known gastroesophageal reflux disease or a high-grade esophageal stricture were considered absolute contraindications, and, in these cases, traditional radiologic gastrostomy or gastrojejunostomy was performed. Patients were forbidden food by mouth, and nasogastric tube feedings were withheld starting at midnight the night before the procedure. Following a standard protocol for endoscopically placed tubes (101, a single dose of a prophylactic antibiotic (1g cefazolin intravenous)
928
Radiologic Peroral Gastrostomy
July-August 1999 JVIR
b. Figure 1. (a)A 6.5-F hockey stick catheter is tapped against the nasogastric tube at the gastroesophageal junction. (b) The guide wire is advanced to engage the gastroesophageal junction and the esophagus is catheterized. a.
Figure 2. An endoscopic push-type gastrostomy tube (Flow 20; Cook). Note the long gradually tapered leading end.
was given immediately before the procedure unless other antibiotics were already being given. After appropriate conscious sedation with intravenous fentanyl and midazolam, a nasogastric tube was inserted and the stomach was distended with air. The colon was identified fluoroscopically, if necessary, with rectal insufflation of air,
and verified to be inferior to the distended stomach. The anterior position of the gastric wall was verified with lateral fluoroscopy. Ultrasound or CT was not routinely used. A hypotonic agent (glucagon, buscopan) was administered only if gastric decompression was a problem. The stomach was then punctured with a n 18-gauge trocar-type needle
in the midbody region, using a trajectory perpendicular to the anterior abdominal wall. A small amount of water-soluble contrast material was injected to confirm proper needle position. The needle was then exchanged for a 6.5-F hockey sticktype catheter (MPA; Cook, Bloomington, IN). With use of a 260-cm, 0.035-inch Amplatz Super Stiff wire (Cook), the esophagus was catheterized retrogradely. Catheterization of the esophagus was aided by withdrawing the nasogastric tube tip to the gastroesophageal junction and using this as a visual and tactile aid to identification of the gastroesophageal junction. The catheter was manipulated to lie parallel to the nasogastric tube tip, while tapping against it. The wire was then advanced to engage the gastroesophageal junction (Fig 1). This procedure was sometimes facilitated by partial deflation of the stomach and by inflation of air into the distal esophagus by way of the nasogastric tube. The guide wire and catheter were then advanced up the esophagus and out of the patient's mouth. A lubricated endoscopic pushtype gastrostomy tube (Flow-20push; Wilson-Cook, Winston-Salem, NC) was then advanced over the wire, into the patient's mouth, and down the esophagus. This type of tube has a long, relatively stiff, and very gradually tapered dilator as a n integral part of the leading end of the tube (Fig 2). The leading edge of the dilator was advanced out of the gastrostomy site, where it was manually grasped and the tube was pulled the remainder of the way. When the internal bolster of the gastrostomy tube entered the patient's mouth, the tube was given a short and sharp pull so the bolster passes rapidly into the esophagus, causing a minimal gag reflex. Under fluoroscopic guidance, the tube was then pulled into final position, with the internal bolster gently snugged up against the anterior gastric wall, providing good apposition of the stomach and anterior abdominal wall (Fig 3). The tube was then secured with an external bolster, the dilator por-
Clark et a1
929
Volume 10 Number 7
Figure 4. Final tube position confirmed with injection of water-soluble contrast material.
a.
b.
Figure 3. Peroral gastrostomy tube being pulled out of the gastrostomy site. (a) The internal bolster is in the distal esophagus. Lateral spot view of the anterior abdominal wall showing the final position of tube. (b) The internal bolster gently pulls the anterior stomach wall Bgainst the anterior abdominal wall.
tion of the tube was cut off, and a universal feeding adapter was fitted. Final tube position was confirmed with a n injection of watersoluble contrast material (Fig 4). As per the protocol of the endoscopy unit a t our institution, feeding was started 6 hours after completion of the procedure. In a small number of patients (four), RPG was performed after inadvertent removal of a previous gastrostomy tube and before complete tract maturation. In these cases, the skin tract to the stomach was cannulated directly with a guide wire and hockey-stick catheter. The procedure was then completed as previously described.
I RESULTS Thirty-one RPGs were successfully performed in 32 attempts in 30 patients (97% success rate). Placement was unsuccessful in one patient because of excessive binding of the gastrostomy tube when attempting to pull it through an existing gastric tract. The tract had a n unusually tortuous course before entering the stomach. Because we were concerned about causing significant trauma, the placement in this patient was abandoned and a standard Cope loop-type catheter was inserted without difficulty. In all patients, gastric access was
readily achieved. The esophagus was also successfully catheterized in all patients. In a few of the early cases, a wire snare was passed down the nasogastric tube and used to capture the guide wire within the stomach. However, with some minimal experience, retrograde catheterization of the gastroesophageal junction was found to be faster. There were no major procedural complications (one of 31; 3%). One patient had a minor lip laceration, which required no specific therapy. There weEe no othe; procedural There were major postprocedure . . complications. . NO patient developed pneumonia within 2 weeks of the ~ r o c e d u r e One patient had modeiate erythema surrounding the tube site 7 days after insertion, which may have been a n early wound infection (one of 31; 3%).This patient received antibiotic therapy for a n unrelated condition, and the erythema resolved. Two patients dislodged their tubes, one 8 days after insertion and the other 10 days after insertion (two of 31; 6%). These tubes were replaced by means of the same technique, one 24 hours later, the other 4 days later. There were no clinical sequelae caused by tube dislodgment. Current follow-up on these patients varies from l to 11 months. There were no tube occlusions. Four tubes were removed when patients '
Radiologic Peroral Gastrostomy
930
July-August 1999 JVIR
regained oropharyngeal control; this was done at bedside with external traction. Two tubes were converted to skin-level button gastrostomies after tract maturation. The external portions of two tubes were damaged by the nursing staff 9 months after placement; these tubes were replaced by a balloon-type catheter. Two patients died in the follow-up period of natural causes or of their underlying disease; no deaths were directly or indirectly attributable to gastrostomy placement.
the type of retaining mechanism present, and to the mental status of the patient; Cope loop-type or balloon tubes are more Drone to retention device failure, a; the Cope loop retaining string can break, the balloon can rupture, or the balloon valve can leak. These tubes are then more easily removed than are endoscopic tubes with a permanent internal bolster. When tubes are dislodged before tract maturation, there is a significant risk of gastric leakage and peritonitis. More secure tubes are, therefore, safer and reauire fewer interventions. Studies reported in the literature vary DISCUSSION widelv as to the re~ortedincidence Percutaneous gastrostomy, as of tudk occlusion aAd dislodgment (8,141; comparison between these performed by radiologists or endoscopists, is an accepted and proven studies is difficult as they cannot be technique, with significant benefits controlled for the quality of nursing compared to surgical techniques. care given and the mental status of Percutaneous techniques are faster, the patient. cheaper, and are associated with The literature suggests that lower morbidity and mortality than there is a lower major complication surgical gastrostomy (11-13). Howrate with radiologic gastrostomy ever, published literature does not than with endoscopic techniques (7,9). This may be caused by several demonstrate large differences between endoscopic and radiologic factors. First, image-guided placement allows precise localization of techniques. In all likelihood, radiothe liver and colon, and, therefore, logic gastrostomy is cheaper, has a higher success rate, and has a lower complications caused by transgresmajor complication rate, whereas sions of these structures by radiologic techniques are very rare. Secendoscopic gastronomy has a lower ond, diagnostic endoscopy alone is incidence of tube-related complications (6-9).. associated with a low, but signifiThere are many potential benecant, major complication rate. The fits to placing endoscopic-type gasomission of the endoscope may actrostomy tubes. First, the ability to count in large part for the lower place a large-bore tube reduces procedural complication rate of ratube-related complications, mainly diologic gastrostomy. occlusion and dislodgment. The rate The ease with which the gastrosof tube occlusion is primarily retomy tract is dilated during placelated to the size of the inner bore of ment of an endoscopic-type gastrosthe tube, the quality of the nursing tomy tube is a significant benefit, care given, and the physical charac- and it mav lead to a safer Droceteristics of the material adminisdure. ~ h e ~ t r adilation ct ocdurs in a tered via the tube. Large-bore tubes single step, avoiding serial dilations require less meticulous nursing care and the associated risk of leakage of than do smaller bore tubes, as adgastric contents with each exchange ministration of poorly crushed med- of a dilator. Dilation with endoscopication and the lack of vigorous ic-type tubes also proceeds in an tube flushing frequently contribute internal-to-external direction, proto the occlusion of gastrostomy viding continuous apposition of the tubes. Large-bore tubes do not kink stomach wall to the anterior abdomor clog as easily as small-bore tubes inal wall. This means that gastric and, thus, do not require as freaccess is never lost inadvertently by quent replacement. The rate of tube a dilator pushing the stomach wall dislodgment is strongly related to away from the abdominal wall. This ,
is also achieved without the use of gastropexy devices, such as T fasteners or anchor sutures. This apposition of gastric and abdominal walls is also maintained after tube placement by the presence of an internal bolster, leading to a shorter and straighter percutaneous tract; conversion of the gastrostomy to a skin-level button is therefore significantly easier. The cost of RPG is virtually identical to the cost of standard radiologic gastrostomy. Although there is much confusion in the literature with regard to actual procedural costs versus procedural charges, radiologic gastrostomy is likely cheaper (2,16-19). Procedural cost for RPG in our department, for fluoroscopic room time, nursing and technical support, and all necessary materials (drugs, needles, wires, gastrostomy tube, and so forth) is U.S. $412. The total professional fee charged (as decreed by the provincial health insurance) is U.S. $108. The total fee of U.S. $520 compares favorably to the total cost for PEG placement, from U.S. $1,800 to U.S. $3,300. Typical procedure time for the tube placements performed in this study was 25 minutes. With experience, many tube placements were performed in 15 minutes. This compares favorably to other techniques (6). Only one operator is required for RPG, whereas PEG placement requires two. There are some potential drawbacks to this type of tube placement. Endoscopic gastrostomy is typically associated with a higher incidence of postprocedure pneumonia than standard radiologic gastrostomy (7). This may be related to aspiration a t the time of the procedure, perhaps caused by the use of topical anesthesia in the pharynx, or the presence of a foreign body (the endoscope or the gastrostomy tube itself). RPG may be associated with a higher incidence of pneumonia as well; however, the lack of an endoscope, and hence the reduced need for topical anesthesia, likely leads to a lower incidence of pneumonia than with PEG. There were no clinical pneumonias associated
Clark et a1
931
Volume 10 Number 7
with gastrostomy tube placement in this study. Endoscopic gastrostomy is also associated with a higher incidence of wound infection than are traditional radiologic techniques (8,lO). This is thought to be related to contamination of the gastrostomy tube with oropharyngeal bacteria as it passes through the mouth. For this reason. most endosco~istsadminister a single dose of a prophylactic antibiotic immediately before the procedure. RPG will likely have a similar incidence of wound infection and, therefore, also warrants antibiotic prophylaxis. Conversely, the incidence of wound infection during endoscopic gastrostomy may be at least partly related to the Dresence of the endosco~eand the multiple passes of the endoscope and tube typically performed. The true incidence of wound infection with RPG remains to be determined, but it likely lies between that for endoscopic gastrostomy and that for standard radiologic gastrostomy. Finally, peroral techniques are slightly less comfortable for the patient because of stimulation of the gag reflex with (brief) oral passage of the tube, and consequently may require slightly deeper conscious sedation than with traditional radiologic techniques. The endoscopic-type gastrostomy tubes used in this study were designed for removal by external traction. The recommended method involves attempting to collapse the internal bolster by pinching it through the anterior abdominal wall and simultaneously applying continuous gentle external traction. In most cases, this worked well. If the tube was not easily removable by this method, we used the technique recommended by Korula and Harma (20). This involves cutting the tube at skin level and allowing the internal portion to fall into the stomach and be eliminated intestinally. We had no complications associated with either method. The indications we currently use for RPG as compared with a traditional radiologic gastrostomy are a specific request from the clinician for a large-bore tube; a patient deemed at high risk for pulling out
a gastrostomy; and patients for whom the gastrostomy will be converted to a skin level button. The contraindications we use for RPG placement are the presence of an eophageal occlusion or the presence of known significant gastroesophageal reflux. Although as yet unknown, the presence of significant ascites may be a relative indication for RPG because of its superior apposition of the stomach and anterior abdominal wall. A similar technique for the peroral placement of gastrostomy tubes was described in dogs (21) and has been adopted in children (22). It would seem that this technique is ideally suited to adults. Most adults for whom a gastrostomy is indicated have an impaired gag reflex, and therefore oral passage of the tube causes minimal discomfort and little or no physiologic response. The adult esophagus also easily accommodates the internal bolster of the tube. Delirious adults are more likely to have the strength required to remove a Cope-loop type of tube and, therefore, benefit from the retention provided by a large internal bolster. Endoscopic-type gastrostomy tubes may be purchased as "push" or "pull" type tubes. Push-type tubes were used almost exclusively in this study. Pull-type tubes require an additional catheter exchange to pass a special loop wire necessary for attachment to the gastrostomy tube. For this reason, we found push-type gastrostomy tubes faster to dace. In addition. the dilator of push-type ' tubes is longer and more gradually tapered, allowing easier and gentler tract dilation. Other reports have documented similar success to ours with the placement of pull-type gastrostomy tubes (23,241; the choice of pull- or push-type tubes is likely personal.
CONCLUSION Large-bore endoscopic-type gastrostomy tubes may be placed safely and easily using fluoroscopy and standard interventional radiologic
techniques. Further studies are needed to confirm the high success rate and low complication rate. The American Society for Gastrointestinal Endoscopy claims on its World Wide Web page that "[rladiologic placement of gastrostomy tubes ... is infrequently used and comparison to PEG regarding safety, efficacy and cost is not available" (25). Although the published literature may already dispute this claim, there is clearly a need for further comparison of endoscopic and radiologic techniques. It is estimated that feeding tube insertion is the second most common indication for upper gastrointestinal endoscopy in hospitalized patients, with 74,000 tube placements performed annually in the United States. With this potential volume, organization of large trials should not be problematic. RPG preserves the best of traditional radiologic and endoscopic gastrostomy; it maintains much of the safety of traditional radiologic gastrostomy and provides the benefit of a large-bore endoscopic-type tube. Although not indicated for all patients, it should prove useful in our armamentarium. References 1. Gauderer MWL, Ponsky JL, Izant RJ. Gastrostomy without laparotomy: a percutaneous endoscopic technique. J Ped Surg 1980; 15:872875. 2. Wollman B, D'Agostino HB, WalusWigle JR, Easter DW, Beale A. Radiologic, endoscopic and surgical gastrostomy: an institutional evaluation and meta-analysis of the literature. Radiology 1995; 197:699-704. 3. Tao HH, Gillie RR. Percutaneous feeding gastrostomy. AJR 1983; 149: 793-794. .- - . - - .
4. Ho CS. Percutaneous gastrostomy for jejunal feeding. Radiology 1983; 149:595-596. 5. Wills JS, Oglesby JT. Percutaneous gastrostomy. Radiology 1983; 149:449-453. 6. Miller RE, Castlemain B, Lacqua FJ, Kotler DP. Percutaneous endoscopic gastrostomy: results in 316 patients and review of literature. Surg Endosc 1989; 3:186-190. 7. Taylor CA, Larson DE, Ballard DJ, et al. Predictors of outcome after
932
Radiologic Peroral Gastrostomy
July-August 1999 JVIR
percutaneous endoscopic gastrostomy: a community-based study. Mayo Clin Proc 1992; 67:1042-1049. 8. McLoughlin RF, So B, Gray RR. Fluoroscopically guided percutaneous gastrostomy: current status. CARJ 1996; 47:lO-15. 9. Munk PL, Lee MJ, Poon PY, et al. Percutaneous gastrostomy in radiologic practice. Austral Radiol 1997; 41:342-350. 10. Jain NK, Larson DE, Schroeder KW, et al. Antibiotic prophylaxis for percutaneous endoscopic gastrostomy. Ann Int Med 1987; 107:824828. 11. Stern JS. Comparison of percutaneous endoscopic gastrostomy with surgical gastrostomy a t a community hospital. Am J Gastroenterol 1986; 81:1171-1173. 12. Grant JP. Comparison of percutaneous endoscopic gastrostomy with Stamm gastrostomy. Ann Surg 1988; 207:598-603. 13. Kummer BA, Tiszenkel HI, Kotler DP, Miller RE. Percutaneous endoscopic gastrostomy: procedure of choice. Gastrointest Endosc 1985; 31:156-157
14. Ryan JM, Hahn PF, Boland GW, McDowell RK, Saini S, Mueller PR. Percutaneous gastrostomy with T-fastener gastropexy: results of 316 consecutive procedures. Radiology 1997; 203:496-500. 15. McLoughlin RF, Gibney RG, Fluoroscopically guided percutaneous gastrostomy: tube function and malfunction. Abdom Imaging 1994; 19:195-200. 16. Goretsky MJ, Johnson N, Farrell M, Ziegler MM. Alternative techniques of feeding gastrostomy in children: a critical analysis. J Am Coll Surg 1996; 182:233-240. 17. Clark JA. Cost of radiologic gastrostomy [letter]. Radiology 1996; 199:582. 18. Bell SD. Cost of performing a radiologic percutaneous gastrostomy [letter]. Radiology 1996; 200:586. 19. Rosenzweig TB, Palestrant AM, Esplin CA, Gilsdorf RB. A method for radiologic-assisted gastrostomy when percutaneous endoscopic gastrostomy is contraindicated. Am J Surg 1994; 168:587-591. 20. Korula J , Harma C. A simple and inexpensive method of removal
21.
22.
23,
24.
25.
or replacement of gastrostomy tubes. JAMA 1991; 265:14261428. Sacks BA, Vine HS, Palestrant AM, Ellison HP, Shropshire D, Lowe R. A nonoperative technique for establishment of a gastrostomy in the dog. Invest Radiol 1983; 18:485487. Towbin RB, Ball WS, Bissett GS. Percutaneous gastrostomy and percutaneous gastrojejunostomy in children: antegrade approach. Radiology 1988; 168:473-476. Rosenzweig TB, Palestrant AM, Esplin CA, Gilsdorf RB. A method for radiologic-assisted gastrostomy when percutaneous endoscopic gastrostomy is contraindicated. Am J Surg 1994; 168:587-591. Szymski GX, Albazzaz AN, Funaki B, et al. Radiologically guided placement of pull-type gastrostomy tubes. Radiology 1997; 205:669-673. American Society for Gastrointestinal Endoscopy. Upper GI endoscopy: endoscopic feeding tubes. http://www.asge.org/doc/l71.