Abstract
In patients with chronic renal disease, vascular malformation (i.e., angiodysplasia) is an important cause of gastrointestinal hemorrhage (1). Several authors reported an increased frequency of recurrent bleeding in patients with renal insufficiency compared to non-renal patients (2,3), while other authors found no increased incidence (4). Nevertheless, although painless and even low grade, bleeding from angiodysplasia may lead to acute or chronic anemia and may, therefore, necessitate therapeutic interventions to achieve hemostasis.
Among various techniques (5), such as thermal methods (monopolar/bipolar cautery, heater probe, laser therapy) and injection therapy (epinephrine, alcohol, cyanoacrylate glue, thrombin, fibrin, etc.), argon plasma coagulation (APC) has become an increasingly widespread technique. Argon plasma coagulation uses transmission of a high-frequency monopolar current via a probe emitting ionized, electrically conductive, argon plasma gas to coagulate tissue surfaces (6). It is an efficacious, repeatable, and easily learned method for devitalization of tissue and is thus appropriate for achieving hemostasis in bleeding from surfaces (7). The goal of stopping intestinal bleeding is achievable in nearly all patients, with an average of 2.3 APC interventions per patient and a failure rate of only 1.5%. The APC technique has already been used successfully in dialysis patients with gastric angiodysplasia (8).
Possible complications of this technique are reported to be perforation (7,9), pneumoperitoneum (10), intestinal wall emphysema (7,9), bowel gas explosion (11), and a bleeding vessel stump (Dieulafoy's lesion) proximal to a previously APC-treated gastric angiodysplasia (12).
We report here a patient on continuous ambulatory peritoneal dialysis (CAPD) who had recurrent lower gastrointestinal bleeding from angiodysplasia in the colon, and who developed peritonitis after hemostasis with APC.
Case Report
A 59-year-old male patient with end-stage renal disease resulting from type 2 diabetes had been on CAPD for 11 months. He was admitted because of black and blood-spotted stool perceived on the morning of the same day. Blood pressure was 160/90 mmHg and heart rate was 96 beats per minute. Rectal examination revealed black stool but also traces of fresh blood. Hemoglobin level was 9.2 g/dL, which was only slightly lower than the 9.9 g/dL measured 1 month before. Clotting parameters and thrombocyte count were within normal ranges. Gastroscopy showed no source of bleeding, nor did colonoscopy, but the cecum was not reached. Because the patient stabilized spontaneously and had a constant hemoglobin level of 9.0 g/dL, he was discharged.
One week later, the patient was readmitted because of severely bloody stools the night before. Clinically, he appeared anemic, with pale skin and mucosa. Initially, the patient was stable with blood pressure 180/85 mmHg and heart rate 90 beats per minute. Hemoglobin level was 7.8 mg/dL and decreased further to 7.2 mg/dL after repeated hematochezia the same day. He became symptomatic from anemia and blood transfusion was given.
Due to the fresh blood in the stool, colonoscopy was repeated. When the cecum was reached this time, a single angiodysplasia with a diameter of approximately 12 mm was found in the cecal pole. Bleeding occurred from the angiodysplasia immediately upon touching it with the probe of the endoscope. Because of active bleeding, immediate intervention was indicated. Using APC (38 watts, argon flow 1.8 L/min), we were able to stop the bleeding.
The next day, the patient complained of severe abdominal pain. Body temperature was 37.5°C; dialysate was cloudy. Dialysate cell count was 2.59 × 109/L and rose further to a maximum of 13.15 × 109/L. Peritonitis was diagnosed. Dialysate cultures showed growth of Escherichia coli, among others, suggesting an enteric source of peritonitis. Antibiotic therapy with intraperitoneal ceftazidime was started. Peritonitis subsequently resolved, with normalization of the dialysate cell count, and the patient continued to perform CAPD. The patient's condition improved and consecutive hemoglobin values stabilized above 9 mg/dL without further transfusions. During a follow-up of 17 months, there was no further recurrence of bleeding and no further episode of peritonitis.
Discussion
Peritonitis is among the most distressing complications of peritoneal dialysis (PD). An external source of infection may be presumed by a history of handling mistakes during the dialysate exchange procedure and/or growth of typical “exogenous” germs, such as Staphylococcus aureus, in dialysate culture. “Endogenous” peritonitis may also occur (13). This is suspected when the patient's history is uneventful with respect to handling errors, and typical enteric derived, mostly gram-negative bacteria are observed in dialysate cultures. An endogenous peritonitis may follow episodes of diverticulitis, infectious or ischemic colitis and enteritis, and urogenital or gynecological infections, but, frequently, no particular cause is found. Suh and co-workers reported an incidence of 16 (7%) endogenous peritonitis in 217 peritonitis episodes seen in 192 patients enrolled in their study (14). The most common cause was diverticular leak or perforation, but three episodes were related to endoscopic examinations or intervention. In the general population undergoing colonoscopy, bacteremia is seen with a mean incidence of 3% (15). The patient presented here had massive intestinal bleeding from angiodysplasia of the cecum and developed peritonitis caused by E. coli within 24 hours after endoscopic APC, which may therefore be classified as endogenous, but also iatrogenic.
Intestinal bleeding from angiodysplasia of the colon is recognized as one of the most common causes of gastrointestinal bleeding in the elderly population (1). Thought to be degenerative in origin, angiodysplasia occurs predominantly in the cecum and proximal ascending colon, and is diagnosed by scintigraphy, angiography, or colonoscopy, the latter offering the highest diagnostic accuracy. An increased incidence of recurrent intestinal bleeding from angiodysplasia is well known in patients with renal insufficiency (2,3). Bleeding is attributed, at least in part, to uremic platelet dysfunction and the frequent use of acetylsalicylic acid and heparin with hemodialysis (2). Other authors did not find an increased incidence of bleeding in renal compared to non-renal patients, and did not describe differences in the type or distribution of intestinal bleeding sources (4). Remarkably, a high tendency (up to 47%) of re-bleeding is observed in patients with renal failure (16).
When bleeding from angiodysplasia is observed, a conservative therapeutic approach may be chosen, as bleeding frequently is minor and may stop spontaneously. However, massive or recurrent bleeding, as in our patient, necessitates immediate intervention. Recently, APC was established as an efficacious and repeatable technique, appropriate to achieve hemostasis in bleeding from surfaces (7). Although various complications related to this new technique have been reported (7,9-12), peritonitis without perforation has not been described thus far.
In our patient, the close time relationship between APC and the occurrence of peritonitis and the type of bacteria being suggestive of an enteric source of infection are both indicative of APC-induced migrating peritonitis. Several factors favor our suspicion. A non-distended colon has a wall thickness of approximately 4 mm, but when air is insufflated for examination and also argon gas under pressure for intervention, wall thickness may decrease markedly (17). The maximum depth of tissue penetration with APC is 3 – 4 mm (18). Depth of tissue injury may be limited by the use of low power (40 watts) and a short duration of APC contact, the latter being achieved by continually moving the activated probe along the tissue surface (17). Although a low power (38 watts) was used in our patient and the probe was moved as described, peritonitis occurred.
It may be speculated that APC induces at least some inflammatory reaction in the surrounding tissue, which is a prerequisite for the healing of necrosis. In cases of deep tissue necrosis, adjacent structures, such as omentum and bowel loops, may cover the areas of inflamed tissue. Therefore, subsequent perforation or migration of enteric bacteria into the peritoneal cavity may not have been observed so far. In PD patients, however, the dialysate separates intestinal structures from one another, thereby preventing self-limitation of a migrating infectious process. Theoretically, a hematogenic route of bacterial spread to the peritoneum also has to be considered.
Peritonitis probably would have been avoidable in our patient had antibiotic prophylaxis been administered against gram-negative bacteria. There are no controlled studies on the subject. Expert statements are substantiated by case reports only (13) and recommend prophylactic antibiotic therapy for colonoscopy. Only randomized controlled trials would clarify whether antibiotic prophylaxis should be mandatory in PD patients undergoing colonoscopy but, given the relatively small numbers of PD patients in many centers and the even smaller percentage of colonic bleeding from angiodysplasia, we argue that these trials would be difficult to carry out.
Conclusion
Argon plasma coagulation has to be applied with caution in PD patients. Substantiated by our report and others (13), APC, as well as other endoscopic interventions in general, should be accompanied by prophylactic antibiotic therapy directed against the typical spectrum of enteric bacteria. According to the International Society for Peritoneal Dialysis 2000 treatment guidelines for peritonitis, the same antibiotic drugs used for therapy of gram-negative bacteria may also be adequate for prophylaxis (19).