Applied Radiology

This paper describes two cases of hepatic-portal venous gas and pneumatosis intestinalis, definitively diagnosed on CT, associated with open placed jejunostomy tubes in patients who appeared clinically ill but had no evidence of bowel ischemia at surgery. The hepatic-portal venous gas and pneumatosis resolved after surgical removal of the jejunostomy tube, decompression of the bowel, and administration of both bowel prokinetic agents and antibiotics.

Drs. Hofmann, Beall, Jones, and Fishman, are in the Department of Radiology and Radiological Science, and Dr. Heitmiller is in the Department of Surgery at The Johns Hopkins Medical Institutions in Baltimore, MD.

Introduction

The finding of hepatic-portal venous gas (HPVG) on plain films usually indicates a poor prognosis; it is associated with a 75% mortality.1 Most clinicians and radiologists alike equate hepatic-portal venous gas with bowel ischemia. Although ischemic bowel accounts for 72%1 of cases with HPVG, there are other causes that must be considered. HPVG is also associated with intra-abdominal abscesses, small bowel obstruction, gastric ulcers, ulcerative colitis usually following a double-contrast barium enema,1 blunt abdominal trauma,2 and needle catheter jejunostomy tubes (NCT).

A Medline review of the literature going back to 1966 found no reported cases of HPVG and PI in association with open placed jejunostomy tubes (OJT); there are, however, a handful of studies describing the association of NCT and pneumatosis intestinalis (PI).

We describe two cases of HPVG and PI associated with OJT, definitively diagnosed on CT, in patients who appeared clinically ill without evidence of bowel ischemia at surgery. In both cases, the HPVG and PI resolved after surgical removal of the OJT, decompression of the bowel, and administration of both bowel prokinetic agents and antibiotics. These cases are illustrated to alert radiologists and clinicians to the possibility that HPVG and PI may be found in the clinical setting of an ill-appearing patient without bowel necrosis.

Cases

The first patient was a 59-year-old female who underwent proximal gastrectomy and distal esophagectomy with the placement of a OJT for the treatment of esophageal carcinoma.3 On the eighth postoperative day, the patient began to vomit, and her abdomen became distended and tender. Her white blood cell count was 23,000. Clinically, the patient appeared very ill. An abdominal CT was requested; it demonstrated massive HPVG and PI in air-distended small bowel loops (figure l). The patient was started on antibiotics and rushed to surgery. In surgery, her bowel appeared moderately distended, with normal color and good pulses in an otherwise unremarkable abdomen. The HPVG and PI were attributed to the combination of bowel distention and the OJT. The OJT was removed during surgery and the OJT site in the bowel wall was oversown. Postoperatively, the bowel was decompressed with a nasogastric tube and prokinetic drugs. The patient had another abdominal CT 10 days later demonstrating no evidence of HPVG or PI (figure l). The patient recovered uneventfully from surgery, and was subsequently discharged on the thirtieth day from admission.

Our second patient was a 72-year-old male who underwent proximal gastrectomy and distal esophagectomy for esophageal carcinoma with placement of a OJT. On the seventh postoperative day, the patient appeared ill, had a distended, tender abdomen, and had a white blood cell count of 28,800. An abdominal CT showed massive HPVG and PI, with air distention of the small bowel (figure 2). The patient was started on antibiotics and taken to surgery. In surgery, the bowel was moderately distended, but appeared normal in color, without evidence of ischemia. Subsequently, the OJT was removed and its site oversown. Postoperatively, the patient's bowel was decompressed with a nasogastric tube and prokinetic drugs. He recovered uneventfully from surgery, and was discharged on the twenty-second day from admission.

Discussion

HPVG and PI carries a grave prognosis because it is usually associated with an ischemic bowel. Although both of our patients appeared ill, each had a well-perfused bowel at surgery. We are unsure as to the exact cause of the HPVG and PI in these patients, though we postulate that it may be due to the OJT combined with bowel distention.

Though there are no reported cases of HPVG and PI associated with OJT,4 there are 16 reported cases of the association between PI and NCT, with only one of these cases also demonstrating HPVG. Strain et al reviewed 53 patients with NCT, four patients with PI, and one patient exhibiting both PI and HPVG, but they did not describe the patients' clinical symptoms in their report. Nevertheless, they did report that the patient with both HPVG and PI went on to complete recovery with conservative treatment.5 Knechtle et al prospectively evaluated 27 consecutive patients with PI, five of whom had a NCT.6 In four of the patients, the PI resolved without sequelae. One patient, however, died due to bowel infarction and subsequent septicemia. This poor outcome was believed to be related to a prolonged episode of hypotension and not to the NCT.

Knechtle and coworkers' article also discussed the management of pneumatosis intestinalis. These authors advised that patients with PI and clinical evidence of bowel obstruction or ischemia should be managed aggressively with surgery, while asymptomatic patients without metabolic acidosis could be managed conservatively.6 In their investigation, Smith and Sarr described only two cases of clinically significant PI in 217 consecutive patients with newly placed NCT.7 Similar to our cases, both patients became febrile, and had abdominal pain and diffuse abdominal tenderness. Smith and Sarr's patients were managed conservatively with broad spectrum antibiotics and cessation of enteral alimentation. Both patients were asymptomatic within 48 hours; however, neither patient exhibited HPVG.7 We found four other cases in the literature that describe the association of PI and NCT.7,8 Most of these cases did not detail clinical or laboratory findings. However, all were treated conservatively with subsequent resolution of the PI.

For patients such as ours, who have open jejunostomy tubes, no association with HPVG and PI has been established. An OJT uses a larger bore catheter and is inserted into the bowel through a tunnel created by the surgeon, using manual dissection. This differs from the needle catheter jejunostomy tube, which uses a needle on the tip of a small bore catheter to create its own tunnel as it is forced through the bowel wall. The safety of OJT with esophagectomy has been described by Gerndt and Orringer. These authors reviewed their experience with 523 patients undergoing esophagectomy with OJT over a 15-year period and reported a complication rate of 2.1%, primarily either bowel obstruction or intra-abdominal abscess. They did not report HPVG or PI in any of their patients.4

The etiology for PI and HPVG in association with OJT or NCT has not been clearly elucidated. Smith and Sarr proposed two separate mechanisms for intramural air in the context of a NCT. The first suggests that a combination of postoperative ileus, aerophagia, and/or intraluminal bacterial fermentation of feedings from overgrowth of the bowel flora raises the intraluminal pressure. As a result, the air dissects into the bowel wall through the mucosal defect created by the NCT. Their second theory proposes gas production by gas-forming organisms that have invaded the bowel wall via the catheter tract.7 Regardless of the mechanism of PI, once the air is in the bowel wall, it may enter the portal venous system.1 It also is unclear as to the relationship of HPVG and PI, though when HPVG is present PI is almost invariably also present.

We do not know the exact etiology of the HPVG and PI in our cases, or why our patients appeared so sick. Both patients had an esophagectomy and gastric pull-up, an OJT, moderate small bowel distention, and both appeared clinically ill. We hypothesize that increased intraluminal pressure from the ileus was sufficient to allow air to dissect along the OJT into the bowel wall, and then gain access into the portal venous system. Moreover, the abdominal distention and tenderness could be attributed to the ileus.

Regardless of the etiology, the treatment for our subset of patients has been efficacious. Considering the ill appearance of our patients, a laparotomy was performed to rule out ischemic bowel. At surgery, the OJT was removed. We are unsure if this was necessary, as we do not know the exact cause of the HPVG and PI, and studies with NCT have shown resolution of symptoms without removal of the NCT. Postoperatively, the bowel was decompressed, and the patients were both placed on prokinetic agents and antibiotics.

Summary

Two patients who underwent esophagectomy and gastric pull-up associated with the placement of a OJT are described. Both subsequently developed massive HPVG and PI without evidence of bowel ischemia at surgery, which resolved after removal of the OJT, decompression of the bowel and initiation of antibiotics and prokinetic agents. Although a few cases in the literature describe the association of PI and NCT, none have described the association of HPVG and PI with OJT. By extrapolating proposed mechanisms of HPVG and PI seen with NCT, we have developed a possible theory on the cause of HPVG and PI without bowel necrosis. AR

References

1. Liebman PR, Patten MT, Manny J, et al: Hepatic-portal venous gas in adults: Etiology, pathophysiology and clinical significance. Ann Surg 187(3):281 -287, 1978.

2. Vauthey JN, Matthews CC: Portal vein air embolization after blunt abdominal trauma. Am Surg 54(9):586-588, 1988.

3. Heitmiller RF, Venbrux AC, Osterman FA: Percutaneous replacement jejunostomy. Ann Thorac Surg 53:711-713, 1992.

4. Gerndt SJ, Orringer MB: Tube jejunostomy as an adjunct to esophagectomy. Surgery 115(2): 164-169, 1994.

5. Strain JD, Rudikoff JC, Moore EE, Jones TN: Pneumatosis intestinalis associated with intracatheter jejunostomy feeding. AJR 139:107-109, 1981.

6. Knechtle SJ, Davidoff AM, Rice RP: Pneumatosis intestinalis: Surgical management and clinical outcome. Ann Surg 212(2):160-165, 1990.

7. Smith DC, Sarr MG: Clinically significant pneumatosis intestinalis with postoperative enteral feedings by needle catheter jejunostomy: An unusual complication. J Parenter Enteral Nutr 15(3):328-331, 1991.

8. Thomas L, Cohen AJ, Omiya B, et al: Pneumatosis intestinalis associated with needle catheter jejunostomy tubes: CT findings and implications. J Comput Assist Tomogr 16(3):418-419, 1992.