Radiologic Gastrojejunostomy and Percutaneous Endoscopic Gastrostomy: A Prospective, Randomized Comparison

Visceral Intervention

Radiologic Gastrojejunostomy and Percutaneous Endoscopic Gastrostomy: A Prospective, Randomized Comparison1 A

Eric K. Hoffer, MD John M. Cosgrove, MD Daniel Q. Levin, MD Michael M. Herskowitz, MD Salvatore J. A. Sclafani, MD

.

Index terms: Gastrojejunostomy Gastrostomy

JVIR 1999; 10:413-420 Abbreviations: PEG = percutaneous endoscopic gastrostomy, PEJ = percutaneous endoscopic gastrojejunostomy, PGJ = percutaneous gastrojejunostomy

PURPOSE: To compare the efficacy of radiologic guided placement of percutaneous gastrojejunostomy (PGJ) and percutaneous endoscopic gastrostomy (PEG). MATERIALS AND METHODS: Patients were randomized to PGJ (n = 66) or PEG (n = 69). Indications for gastrostomy were need for prolonged enteral nutrition (97%)or gastrointestinal decompression (3%),with etiologies of neurologic impairment (81%),head and neck neoplasm (12%),bowel obstruction (3%),or other (4%).Mean follow-up was 202 days and 30-day follow-up was obtained for 85% of patients. RESULTS: PEG was successful in 63 of 69 (91%)patients, while PGJ established access in all of 66 attempts (100%)(P = .014). Average procedural time was 53 minutes for PGJ and 24 minutes for PEG (P = .001). At 30-day follow-up, there were 33 and 45 complications in the PGJ and PEG groups, respectively. This difference was due to the greater incidence of pneumonia in the PEG group (P= .013). Long-term tube-related complications occurred with 17 PGJs and four PEGS (P = .007). The PGJ cost more than PEG, but this advantage was offset by the cost of complications. CONCLUSION: PGJ had higher success rate and fewer complications, due to a lower incidence of pneumonia. PEG took less time to perform, cost less, and required less tube maintenance.

'

From the Departments of Radiology (E.K.H., D.Q.L., M.M.H., S.J.A.S.) and Surgery (J.M.C., S.J.A.S.), Kings County Hospital Center, Brooklyn, New York. Received August 28, 1998; revision requested September 28; revision received October 13; accepted October 14. From the 1998 SCVIR annual meeting. Address correspondence to E.K.H., Department of Radiology, Box 359728, Harborview Medical Center, 325 9th Ave., Seattle, WA 98104. 0 SCVIR, 1999

PERCUTANEOUS endoscopic gastrostomy (PEG) or fluoroscopic gastrostomy are the preferred methods to establish access for long-term enteral nutrition. An open surgical gastrostomy was first proposed in 1837 by Egeberg, and first performed successfully by Verneuil in 1876 (1). In 1980, Gauderer and Ponsky described an endoscopeguided percutaneous technique (PEG) (2). In 1981, Preshaw (3) reported a fluoroscopically guided percutaneous technique that did not require endoscopy and, in 1983, three radiology groups reported clinical series of fluoroscopy-guided gastrostomy placement (4-6).

Complication rates for PEG are equal to or less than those reported in surgical series, with shorter procedure times and lower costs (7). Ho and colleagues demonstrated a similar advantage for fluoroscopic percutaneous gastrostomy as compared with open surgical gastrostomy (8). Recent meta-analysis and cohort studies (8-11) found that the radiologic technique was more often successful, more expensive, and required less anesthesia than PEG. Nevertheless, the fluoroscopyguided procedure represents the minority of gastrostomies (18%-35%) (12). The present study compares the safety, efficacy, and

414

Radiologic Gastrojejunostomy and Percutaneous Endoscopic Gastrostomy

April 1999 JVIR

Table 1 Patient Demographics and Indications Variable No. of patients Mean age (y) Age range (y) Male (%I Indications Neurologic impairment Head and neck neoplasm Gastrointestinal decompression Other* Pre-existing conditions Tracheostomy Antibiotic coverage H,-blocker Temperature greater than 100.5 Note.-ns

=

PGJ

PEG

66 58.2 18-93 54.7

69 51.9 18-89 58.7

55 (83) 9 (14) l(1.5) 1 (1.5)

54 (78) 7 (10) 3 (4) 5 (7)

48 (73) 38 (57) 44 (67) 25 (38)

57 (90) 43 (68) 44 (70) 26 (41)

P Value ns ns ns

.02 ns ns ns

not significant. Numbers in parentheses are percentages.

* PGJ other was due to scleroderma, PEG others were due to sepsis with respiratory distress syndrome, abdominal trauma ( n = 21, pelvic trauma, and gastric neoplasm.

costs of fluoroscopy-guided percutaneous gastrojejunostomy (PGJ) and PEG in a prospective, randomized fashion.

I MATERIALS AND METHODS Patients older than 18 years who were referred for gastrostomy placement between January 1992 and December 1993 were considered for inclusion in the study. Tube replacements through existing tracts were excluded. An Institutional Review Board-approved informed consent was obtained from the patient or their representative. Patients were assigned by random number to endoscopic (n = 69) or radiologic (n = 66) percutaneous procedures. The 77 male and 60 female patients had a mean age of 53 years. Characteristics of the populations are presented in Table 1. The PEG group had a higher incidence of preexisting tracheostomy (P = .02), while long-term antibiotic coverage, fever at the time of the procedure, and alkalization of the stomach fluid were not significantly different between groups (Table 1). If the initial assigned procedure was not successful, the other was attempted; if both failed, open surgical gastrostomy was available. Procedures

were performed by residents with attending physician supervision. PGJ was performed with standard techniques (6). Gastrojejunostomy had become the preferred procedure in our interventional radiology section because of the high risk of aspiration in our patient population and a belief that gastrostomy was less protective. Patients received 1 mg of midazolam hydrochloride andlor 50 um of fentanvl citrate intravenously as needed for sedation. After the administration of 0.5 mg of glucagon intravenously, air was insufflated through an existing nasogastric tube. In patients with upper digestive tract obstructions initial gastric access was obtained by passage of a 5-F H1 catheter (Cook, Bloomington, IN) under fluoroscopic guidance. After 5001,000 mL of air had been instilled, frontal and lateral fluoroscopic images were obtained to ensure apposition of the anterior gastric and abdominal walls, and to select a puncture site. This area was infiltrated with 1% lidocaine for anesthesia. A commercially available gastrostomy kit (Carey-AlzateCoons; Cook) was used. Three T anchors were placed approximately 2 cm apart, and a fourth puncture was made between them, through

which an 0.038-inch guide wire was directed toward the pylorus. A 6-F directional catheter was used to position the guide wire beyond the ligament of Treitz, and the tract was dilated to accommodate a 14-F peel-away sheath. The proximal 14-F mushroom-type retention device was formed, and the jejunal position of the distal 10-F end was verified with contrast material injection under fluoroscopy. The catheter was sutured to the skin and the area was dressed. PEG was performed as previously described (13). In this series, the procedure was usually performed a t the bedside; in a few cases, it was done in the operating room in conjunction with another procedure (tracheostomy). Patients were sedated with intravenous me~eridineand midazolam hvdrochloride and a topical pharyngeal anesthetic spray was applied. Esophagoduodenoscopy was performed with a 12-mm endoscope (Pentax, Englewood, CO). The abdomen was then prepared and draped. The stomach was dilated by injection of air through the endoscope. The scope was directed toward the anterolateral wall of the stomach, and the illuminated site was used as a guide for the subcostal skin and fascia1 incision. An 18-gauge needle was advanced into the stomach. A suture was passed and grasped from the inside with endoscopic biopsy forceps. This was retracted out the mouth and a 22-F gastrostomy tube (Caluso; Sandoz Nutrition, Minneapolis, MN) was pulled through and retained by an inner flange and an outer securing bumper. Both methods employed a nurse to monitor vital signs and maintain a clear airway. The tubes were not used for feeding until the next day. A single intravenous dose of cephalosporin was administered for antibiotic prophylaxis. Procedural parameters evaluated were technical success, duration of the procedure, amount and type of anesthesia administered, endoscopic findings and their impact on management, and 30-day morbidity and mortality. Follow-up was obtained

Hoffer et a1

415

Volume 10 Number 4

Table 3 Anesthesia

Table 2 Outcome

P Success/Attempted Duration of procedure ( m i d Patients with 30 d follow-up Mean follow-up (days) Average weight gain (lb) Note.-ns

=

PGJ

PEG

66/66 (100%) 52.7 5 17.2 54 (82%) 200 +- 337 -3.9 2 15.2

69/63 (91%) 23.8 ? 13.2 55 (87%) 204 +- 343 -9.3 t 19.2

Value .014 .001 ns ns .13

Local Sedation General anesthesia

PGJ

PEG

28 (42) 38 (57) 0 (0)

l(3) 58 (92) 3 (5)

P < ,001.

not significant.

at physical examination within 24 hours and a t 7 days. Chart review and telephone contact of the patient, nursing home, or rehabilitation facility provided 30-day follow-up on 109 (84.5%)patients. Incomplete follow-up was most often due to an inability to obtain adequate chart review from outside institutions. Technical success was defined on an intent-to-treat basis, when a feeding tube was successfully placed into the stomach or jejunum during the initial procedure. Procedure duration was defined by physician involvement, from skin puncture to suture of the tube to the skin for the PGJ, and from initial endoscope passage to bumper placement for the PEG. Complications were evaluated by two classifications; that traditionally used in the gastrostomy literature (14,151, and a broader surgical definition. The latter included events or outcomes that required an additional percutaneous or open surgical procedure, systemic antibiotic therapy, a higher level of care, or resulted in mortality or long-term disability. Tube-related complications were those that required alterations in the alimentation regimen or bedside replacement of the gastrostomy tube. Pneumonias were identified by a combination of new radiographic opacities coupled with elevated white blood cell count andlor fever. Intolerance was defined as nausea and vomiting or diarrhea that required temporary cessation or a diminished rate of tube feedings. Itemized pre-markup hospital costs were assessed. Charges were

not used because they are based on historical precedent rather than a cost analysis. Reimbursements from Medicare were also evaluated because these represent a standardized cost to the health care system. Baseline characteristics of the two groups and differences between treatment groups in postrandomization measures or events were compared with the 2 statistic for discrete variables and with the Student t test for continuous variables. Data are reported as means t standard deviation. A P value of .05 was considered significant. Patients that had an unsuccessful PEG and required crossover procedures were included only in the analysis of technical success of the treatment to which they were randomized.

I RESULTS The initial procedures were successful in obtaining gastric access in 100% of PGJ and 91% of PEG attempts (P = .014) (Table 2). Three PGJ attempts failed to negotiate the pylorus, and gastrostomy tubes were placed. One of these was converted to PGJ 1 week later. Of the PEG failures, all patients had PGJ placed under fluoroscopic guidance (and were excluded from further analysis). Failure was due to unfriendly anatomy (oropharyngeal trauma, n = 2; pharyngeal or esophageal tumor, n = 2), inability to transilluminate (n = 11, or patient inability to tolerate passage of the endoscope (oxygen desaturation, n = 1). Anesthesia requirements were

higher for the endoscopy group (Table 3). The duration of the procedure was halved for PEG, even though it included a complete endoscopic examination (P < .001) (Table 2). The added time for the PGJ was due to occasional lengthy attempts to maneuver the catheter through the pylorus and beyond the ligament of Treitz. Endoscopy identified abnormalities in 10% of patients. Gastritis (n = l)was treated with cimetidine, candida esophagitis (n = 1)was treated with fluconazole, while focal edema (n = 3) and hiatal hernias (n = 2) were not treated. Assessment of efficacy by weight gain a t 1month was feasible in 76% of patients. There was less weight loss in the jejunal-fed patients; however, this was not statistically significant (P = .13). Complications were relatively common. By the criteria of Shellito and Malt (141, as modified by Saini and colleagues (151, there were no significant differences between the groups (Table 4). Evaluated with the less specific, 30-day adverse event definition of complications, there were 45 complications in 34 patients in the PEG group, and 33 complications in 32 patients in the PGJ group (Table 4). Additional episodes of pneumonia in the PEG group during the first week were the only significant difference (P = .013). Feeding-tube-related complications were more frequent in the PGJ group; 18 episodes in 16 patients, versus 10 in 10 PEG patients (P = .11) (Table 4). Longterm follow-up identified tube-specific differences; there were nine occlusions, six dislodgments, and two tube fractures in the PGJ group; and one, two, and one epi-

416

Radiologic Gastrojejunostomy and Percutaneous Endoscopic Gastrostomy

A p r i l 1999

JVIR

Table 4 Morbidity a n d Mortality Complication Classification

Complication

Gastrostomy Literature (14,151 PGJ

PEG

-

Procedure related death Unrelated death

0 5

0 4

30 day Surgical

PGJ

PEG

0 4

0 5

Major Emergent laparotomy Hemorrhage, transfused Pneumonia, aspiration all? Tube displacement, repeat procedure Pericatheter leak Wound infections Deep venous thrombosis Ulcer? Other infections? Total Total patients P value

Insufficient gastric drainage Wound infections Tube kink or block*$ Intolerance? Hematemesis, oral rxt Dislodgment*? Tube fracture*? Total Total patients P value

All

0 0 3

1 2 1

2 0

1 1

1

0

11 11

9 8

0 0 3 8 2 0 5 1 1 13 33 32

1 2 1 24 1 1 3 0 1 11 45 34

ns

,013

Minor

Tube-related

0 5 1

1 3 1

6 6

5 5

ns

0

1

2 9 4 2 1 18 16

1 7 1 0 0 10 10 0.11

* Note.-ns = not significant. t not considered complications in gastrostomy literature (14,15). $ only early (<2 week) bedside replacements considered complications in

gastrostomy classification.

sodes, respectively, in the PEG group (P = .007). There was a trend for pneumonia early after the PEG procedure; 60% occurred within the first week after the procedure, compared with 36% of the post-PGJ pneumonias. After the first 7 days, the incidence of pneumonia was similar for both groups (seven in the PGJ group and 10 in the PEG group). There were no tube removals due to pneumonia, nor conversions of PEG to gastrojejunostomy. The incidence of recurrent pneumonia was 28% after intervals of 7-90 days, and was the

same for the two groups (three in PGJ, seven in PEG). There was no procedure-related mortality and the overall 30-day mortality rate was 6.7% and 7.8% for the PGJ and PEG groups, respectively (Table 4). Although pneumonia was more common in the PEG group, the only documented aspiration pneumonia-related deaths occurred in two PGJ patients. Significant hemorrhage was seen only in the PEG group. That patient had a gastrostomy placed despite suboptimal transillumination, and bleeding was noted

from the tube site and from the nasogastric tube immediately after the procedure. At emergent open surgical exploration, an injured artery along the lesser curve of the stomach was identified and repaired. This patient never fully recovered and died of sepsis 5 weeks later. The assessment of costs was imprecise. Medicare reimbursement for inpatient procedures could not be extracted from Diagnosis Related Group calculations. Grouper software revealed no additional reimbursement when either PEG or PGJ were added to the list of procedures for patients with the (most common) primary diagnoses of head trauma or intracerebral hemorrhage (16). Reimbursements were calculated from the Medicare Part B physician payment schedule which identifies hospital (TC-modified) components for CPT codes over 70000 (17). Hospital reimbursement rates for other codes were obtained from Ambulatory Surgery Center rates (18) (Table 5). The itemized costs for procedure and complication calculations were pre-markup hospital costs or best estimates from the financial officer for room charges (Table 6, Appendix). Amounts a r e all in 1997-1998 U.S. dollars. PGJ reimbursement was 57% greater than that for PEG by Medicare rates and cost 2.3 times a s much as PEG when actual costs were compared. These actual cost differences reflect the absence of a room charge and shorter staff time requirements for PEG. (If similar procedures were performed, the Medicare differential is reduced to 7.6%, reflecting additional payments for the image guidance. Actual cost differential would be reduced to that of the procedure room, a t a premium of 10%-30% for fluoroscopy guidance.) Complication costs were significant. The largest component was room expense due to additional hospital days. After the addition of the actual cost of complications (Appendix), the cost of PGJ was 9% less than PEG (Table 6).

Hoffer et a1

417

Volume 10 Number 4

Table 5 Medicare Reimbursement* Procedure Radiologic gastrostomy S and I Reposition to jujunum S and I Total PGJ Endoscopic gastrostomy

* Participating physician

CPT

Physician

Hospital

Total

43750 74350 43761 75984

415 48 141 46 650 426

422 160 0 100 682 422

1,322 848

43246

and ambulatory surgery center fees (17,181.

Table 6 Itemized Costs Item

PGJ

PEG

Procedure room Technician Nurse Physician Tray (miscellaneous supply) Endoscope Gastrostomy kit Procedure Total Cost per successful access + cost of complications Total

322 34 47 110 30 0 181 724 724 763 1,487

0* 22 30 50 30 20 159 3 13 375 1,266 1,641

* PEG was performed

at bedside.

1 DISCUSSION The success and complication rates in this study are similar to those reported in the literature. Success rates for fluoroscopically placed gastrostomy tubes range from 95% to 100% (19-23). Complication rates range from 0% to 6% major and 2.9% to 27% minor, with procedure-related and overall mortality rates of 0.3%-0.8% and 4%26%, respectively (19-23). Endoscopic placement has success rates of 91%-96% (7,13,24,25). Major and minor complication rates range from 4% to 22%, with procedure and overall mortality rates of 0-10% and 8%-50%, respectively (7,13,24-26). Our results verified that, in a single population, the fluoroscopically directed access was more successful than the endoscopic procedure (9-11). Previous conclusions in regard to lower levels of anesthesia

employed for fluoroscopic versus endoscopic procedures were also supported (Table 3). Another advantage identified was the diminished incidence of pneumonia. Whether this was secondary to the jejunal position of the tube in the PGJ patients or an increased incidence of aspiration after endoscopy is unclear. The incidence of pneumonia identified in this study is high in relation to previous PEG and PGJ studies that focus on documented incidents of aspiration. The requirement of documented aspiration prior to inclusion of a pneumonia as a complication results in an underestimation of the incidence. The aspiration of refluxed feedings is infrequently demonstrated (271, and studies of highrisk populations (28,291 found an incidence similar to that identified in this study. A liberal definition of pneumonia avoids the need for specific identification of feedings aspi-

rated from the endotracheal tube or vomiting in association with a pneumonia to document aspiration. Because both groups experienced the same broader definition for the complication, and pre-existing chest compromise (cancer, previous tuberculosis, emphysema, or injury) was similar between groups, the pneumonia rates could be expected to be similar. An accepted indication for gastrojejunostomy is gastroesophageal reflux (GER) andlor aspiration pneumonia; this derives from the presumption that gastrojejunostomy is protective of aspiration as compared with gastrostomy. The rationale is that feedings delivered beyond the ligament of Treitz would be less likely to reflux into the stomach, and so diminish the risk of aspiration from GER. This concept is controversial because of the multifactorial etiology of aspiration pneumonia and the absence of supporting clinical evidence. The dysfunctional neurologic syndrome, swallowing disorder, or mechanical abnormality that makes enteral feeding necessary may be the cause of the aspiration, whether or not there is GER. While there is evidence that patients with gastrostomy are not protected from GER or aspiration (29-31), and patients with gastrojejunostomy and reflux have no more aspiration pneumonia than patients with gastrostomy without reflux (30), there is no direct evidence that patients with gastrojejunostomy have a lower incidence of aspiration or pneumonia than patients with gastrostomy. In studies that identified aspiration before and after percutaneous endoscopic gastrojejunostomy (PEJ), results were contradictory. Di Sario and colleagues (32) found that 12 of 20 with a positive aspiration history continued to aspirate after PEJ, and Kadakia and colleagues (25) found that all of six continued to aspirate (although tube tip located in the duodenum was considered a PEJ). Conversely, MacFadyen and colleagues (33) found 11of 12 patients with aspiration prior to the procedure had no further episodes

418

Radiologic Gastrojejunostomy and Percutaneous Endoscopic Gastrostomy

April 1999 JVIR

after placement of a PEJ (with tube tip confirmed in jejunum by KUB). If it was simply the tube tip position that was associated with a higher incidence of reflux and aspiration, it is unclear why this difference manifests only during the first week after placement. There is, however, no literature support for an increased incidence of aspiration, aspiration pneumonia, or infectious complication after endoscopy (34). The reported incidence of bacteremia after esophagogastroduodenoscopy (EGD) is only 4.2% (35). To clearly sort out the question of whether gastrostojejunostomy protects against aspiration pneumonia as compared to gastrostomy, a study employing a single method of tube placement is required. The added costs of gastrojejunostomy may be justified by a decrease in pneumonia and improved nutrition. Montecalvo and colleagues (36) found that patients fed into their stomach received fewer calories and had poorer nutritional indices versus similar patients fed into the small bowel. Stressed, critically ill patients may experience decreased gastric blood flow with resultant gastroparesis (37). We found a trend for better weight maintenance in the PGJ group. The higher incidence of tube problems in the PGJ group only became marked at long-term followup. The occlusions were likely due to the increased resistance to flow produced by the length and smaller caliber of the PGJ tubes. Other studies have shown a sevenfold increase in tube occlusion rates between gastrostomy and gastrojejunostomy (8%-58%) (30). Inadvertent removal of the tube is more likely with the locking loop retention device of the PGJ as opposed to the fixed inner bumper of the PEG tube. Each service performed what was considered the optimal procedure, however, this produced confounding variables. The prevailing belief among the interventional radiology service was that PGJ was protective against aspiration pneumonia, and so, was the most appropriate procedure in our high-risk population.

The interventional service was more comfortable placing the PGJ a t the initial visit rather than awaiting a second study or clinical event to indicate the need for conversion of a fluoroscopic gastrostomy to PGJ. A gastrojejunostomy tube of equal caliber to the PEG tube was not available at the time of the study. The different methods of access could not be assessed independently of the different tube sizes and tip locations. Similarly, our cost comparisons suffered from the different types of procedures; the premium for PGJ would be reduced to 40% if the PEG incurred room time costs, and further if fluoroscopic gastrostomy was performed (with resultant equal procedure times). Cost differences in the procedures were rapidly outweighed by the costs of complications incurred, particularly if the hospital stay was prolonged. During the past 5 years, new procedures were introduced, and the feeding tubes have improved. Endoscopists developed methods for over-the-wire gastrojejunostomy tube placement (38) and the development of endoscopic direct percutaneous jejunostomy permits jejunal feeding with less alteration of gastroesophageal junction dynamics (39). Radiologists have increased flexibility with portable fluoroscopic or ultrasound-guided gastrostomy (40). Larger diameter (16-24 F) gastrostomy and gastrojejunostomy tubes with both gastric and jejunal ports are available for percutaneous fluoroscopy-guidedplacement, and result in a lower incidence of occlusion.

CONCLUSION This study provides a prospective. randomized com~arisonbetween the radiologic i n d endoscopic procedures. Procedure-related morbidity and mortality are similar for both endoscopy and fluoroscopyguided percutaneous gastric access (9-1 1).The significant differences in outcome were the incidence of technical failure and pneumonia; both were significantly greater in the endoscopy group. Ignorance of

the feasibility of fluoroscopy-guided gastrostomy may be detrimental to patient care when the failed PEG is referred for a higher risk laparoscopy-guided or open surgical procedure. The recent description of the hybrid "fluoroscopy-assisted gastrostomy" emphasizes the utility of the radiologic approach (41). The endoscopy-placed tubes required shorter procedure times and had fewer tube-related complications, although fluoroscopically placed gastrostomy would obviate this difference. PGJ performed with use of fluoroscopic guidance had broader clinical application (fewer contraindications and more liberal anatomic requirements), and required lesser levels of anesthesia. References 1. Mamel J J . Percutaneous endoscopic gastrostomy. Am J Gastroenterol 1989; 84:703-710. 2. Gauderer MWL, Ponsky JL, Izant RJ. Gastrostomy without laparotomy: a percutaneous endoscopic technique. J Pediatr Surg 1980; 15: 872-875. 3. Preshaw RM. A percutaneous method for inserting a feeding gastrostomy tube. Surg Gynecol Obstet 1981; 152:659-660. 4. Tao HH, Gillies RR. Percutaneous feeding gastrostomy. AJR 1983; 141: 793-794. 5. Wills JS, Oglesby JT. Percutaneous gastrostomy. Radiology 1983; 149:449-453. 6. Ho CS. Percutaneous gastrostomy for jejunal feeding. Radiology 1983; 1491595-596. 7. Stiegmann GV, Goff JS, Silas D, Pearlman N, Sun J, Norton L. Endoscopic versus operative gastrostomy: final results of a prospective randomized trial. Gastrointest Endosc 1990; 36:l-5. 8. Ho CS, Yee ACN, McPherson R. Complications of surgical and percutaneous nonendoscopic gastrostomy: review of 233 patients. Gastroenterology 1988; 95:1206-1210. 9. 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. 10. Wollman B, D'Agostino HB. Percutaneous radiologic and endoscopic gastrostomy: a 3-year institutional

Hoffer et a1

419

Volume 10 Number 4

analysis of procedure performance. AJR 1997; 169:1551-1553. 11. Barkmeier JM, Trerotola SO, Wiebke EA, et al. Percutaneous radiologic, surgical endoscopic, and percutaneous endoscopic gastrostomy/gastrojejunostomy: comparative study and cost analysis. Cardiovasc Intervent Radio1 1998; 21:324-328. 12. Levin DC, Matteucci T. "Turf battles" over imaging and interventional procedures in community hospitals: survey results. Radiology 1990; 176:321-324. 13. Larson DE, Burton DD, Schroeder KW, e t al. Percutaneous endoscopic gastrostomy: indications, success, complications, and mortality in 314 consecutive patients. Gastroenterology 1984; 93:48-52. 14. Shellito PC, Malt RA. Tube gastrostomy: technique and complications. Ann Surg 1985; 20:180-195. 15. Saini S, Mueller PR, Gaa J, et al. Percutaneous gastrostomy with gastropexy: experience in 125 patients. AJR 1990; 154:1003-1006. 16. Grouper software. DRGs Definition Manual. Version 15.0. 3M, Inc. 17. Empire Medicare Services, 1998 Fee Schedule. Medicare Part B, PO Box 785, Crompond, NY 105170785. 18. Federal Register. Proposed ambulatory surgical center (ASC) payment status by HCPCS code and related information. June 12, 1998; 63:32339-32308. 19. O'Keeffe F, Carrasco CH, Charnsangavej C, et al. Percutaneous drainage and feeding gastrostomies in 100 patients. Radiology 1989; 172: 341-343. 20. Halkier B, Ho CS, Yee AC. Percutaneous feeding gastrostomy with the Seldinger technique: review of 252 patients. Radiology 1989; 171: 359-362. 21. Hicks ME, Surrah RS, Pilus 0 , et al. Fluoroscopy-guided percutaneous gastrostomy and gastrojejunostomy: analysis of 158 consecutive cases. AJR 1990; 154:725-728. 22. Bell SD, Carmody EA, Yeung EY, Thurston WA, Simons ME, Ho CS. Percutaneous gastrostomy and gas-

trojejunostomy: additional experience in 519 procedures. Radiology 1995; 194:817-820. 23. Ryan JM, Hahn PF, Boland GW, McDowell RK, Saini S, Mueller PR. Percutaneous gastrostomy with Tfastener gastropexy: results of 316 consecutive procedures. Radiology 1997; 203:496-500. 24. Jarnagin WR, Duh QY, Mulvihill SJ, Ridge JA, Schrock TR, Way LW. The efficacy and limitations of percutaneous endoscopic gastrostomy. Arch Surg 1992; 127:261-264. 25. Kadakia SC, Sullivan HO, Starnes E. Percutaneous endoscopic gastrostomy or jejunostomy and the incidence of aspiration in 79 patients. Am J Surg 1992; 164:114-118. 26. Wolfsen HC, Kozarek RA, Ball TJ, Patterson DJ, Botoman VA. Tube dysfunction following percutaneous endoscopic gastrostomy and jejunostomy. Gastrointest Endosc 1990; 36: 261-263. 27. Mullan H, Roubenoff RA, Roubenoff R. Risk of pulmonary aspiration among patients receiving enteral nutrition support. JPEN 1992; 16: 160-164. 28. Bourdel-Marchasson I, Dumas F, Pinganaud G, Emeriau JP, Decamps A. Audit of percutaneous endoscopic gastrostomy in long-term enteral feeding in a nursing home. Int J Qua1 Health Care 1997; 9:297302. 29. Hassett JM, Sunby C, Flint LM. No elimination of aspiration pneumonia in neurologically disabled patients with feeding gastrostomy. Surg Gynecol Obstet 1988; 167:383388. 30. Olson DL, Krubsack AJ, Stewart ET. Percutaneous enteral alimentation: gastrostomy versus gastrojejunostomy. Radiology 1993; 187:105-108. 31. Isch JA, Rescorla FJ, Sherer LR 3rd, West KW, Grosfeld JL. The development of gastroesophageal reflux after percutaneous endoscopic gastrostomy. J Pediatr Surg 1997; 32:321-322. 32. DiSario JA, Foutch PG, Sanowski RA. Poor results with percutane-

ous endoscopic jejunostomy. Gastrointest Endosc 1990; 36:261-263. 33. MacFadyen BV Jr, Catalano MF, Raijman I, Ghobrial R. Percutaneous endoscopic gastrostomy with jejunal extension: a new technique. Am J Gastroenterol 1992; 87:725728. 34. Kaw M, Przepiorka D, Sekas G. Infectious complications of endoscopic procedures in bone marrow transplant recipients. Dig Dis Sci 1993; 38:71-74. 35. Botoman VA, Surawicz CM. Bacteremia with gastrointestinal endoscopic procedures. Gastrointest Endosc 1986; 32:342-346. 36. Montecalvo MA, Steger KA, Farber HW. et al. Nutritional outcome and pneumonia in critical care patients randomized to gastric versus jejunal tube feedings. The Critical Care Research Team. Crit Care Med 1992; 20:1377-1387. 37. Szabo JS, Stonestreet BS, Oh W. Effects of hypoxemia on gastrointestinal blood flow and gastric emptying in the newborn piglet. Pediatr Res 1985; 19:466-471. 38. DeLegge MH, Duckworth PF J r , McHenry L J r , Foxx-Orenstein A, Craig RM, Kirby DF. Percutaneous endoscopic gastrojejunostomy: a dual center safety and efficacy trial. J Parenteral Enteral Nut 1995; 19: 239-243. 39. Shike M, Latkany L. Direct percutaneous endoscopic jejunostomy. Gastrointest Endosc Clin North Am 1998; 8:569-580. 40. Hoffer EK. US-guided percutaneous gastrostomy: a portable technique. JVIR 1996; 7:431-434. 41. Rosenzweig TB, Palestrant AM, E s ~ l i nCA. Gilsdorf RB. A method for radiologic-assisted gastrostomy when percutaneous endoscopic gastrostomy is contraindicated. Am J Surg 1994; 168:587-591. 42. Jarvis WR. Selected aspects of the socioeconomic impact of nosocomial infections: morbidity, mortality, cost, and prevention. Infect Control Hosp Epidemiol 1996; 17:552-557.

420

Radiologic Gastrojejunostomy and Percutaneous Endoscopic Gastrostomy

April 1999 JVIR

APPENDIX Cost Assumptions and Calculations Special procedure room time Technician Nurse Physician Hospitalization Bedside tube replacement Infection costs Antibiotics Culture and Sensitivity Nursing Chest X-ray (pneumonia)

CRX hospital days (42) Total Exploratory laparotomy OR (include nurses) 2 h ICU 1 day Surgeon 1.5 h Post op recovery Pre op labs Anesthesia Anesthesiologist 2 h Total Hemorrhage PRBC Nursing CBC T&C Total (2 units, CBC pre and post) DVT Heparin PTT (q6h)

PT

Nursing CBC (qd) Total (3 d IV heparin) Calculation of costs of complications Emergent laparotomy Hemorrhage Pneumonia Tube displaced, repeat procedure Pericatheter leak Wound infection Deep venous thrombosis Ulcer Other infection Intolerance Hematemasis, oral therapy Occlusion, dislodgment, tube break, bedside replace Totals

$10/min 25hr 34mr 125hr 25Old 35 200 57 .75 hrld 25 pneumonia 150 6 2085

wound infection

urinary tract infection

bloodstream infection

8.2 -

1 685

7 2185

@

PGJ

2485

800 800 187 100 52 133 -

250 -

2322 104funit .75 h/unit 24 28 335 1013d 29 10 l h + .75/d 24 590.5 PGJ 0 0 11 2 0 5 1 1 13 9 4 5

PEG 1 2 25 1 1 3 0 1 11 7 1 1

335 2085 724 435 2485 590

22935 1448 12425

PEG 2322 670 52125 312 435 7455 590

435-2185

13405

15785

35 -

175 50388

35 79729