Peritoneal dialysis and colonic angiodysplasia bleeding: Are they connected?

Abstract

Angiodysplasia is the most common gastrointestinal vascular lesion and a frequent cause of hidden bleeding. Here, we present the case of two patients with severe colonic angiodysplasia-related anemia on peritoneal dialysis. Several attempts to reduce gastrointestinal bleeding have been made without a definitive solution. For these patients, peritoneal dialysis was considered better than hemodialysis because anticoagulation is not needed. However, later the switch from peritoneal dialysis to hemodialysis has been made and an improvement of anemia, due to the cessation of the gastrointestinal bleeding, has been noted. Chronic kidney disease patients are at high risk of bleeding for several reasons. Our hypothesis is that filling the abdomen with peritoneal dialysis solution increases the peritoneal pressure, worsening the colonic venous blood flow and probably peritoneal dialysis with this mechanism sustains the angiodysplasia bleeding. Our conclusions are that peritoneal dialysis indication and peritoneal dialysis prescription should be evaluated with attention in patient with colonic angiodysplasia for its uncertain effect on bleeding.

Introduction

Angiodysplasia (AD) is an abnormality of the blood vessels and it is the most common vascular lesion of the gastrointestinal (GI) tract, involving typically the mucosal layer of the proximal ascending colon and sometimes of the stomach and cecum. After diverticulosis, AD is the second leading cause of lower GI bleeding. ¹ Bleeding is usually low grade so for some patients the only clinical features are iron deficiency anemia and intermittently positive stools for occult blood. ² It is controversial if there is a correlation between AD and chronic kidney disease (CKD). ³ The diagnosis of AD is often difficult due to the absence of pathologic changes in arterioles supplying vascular ectasias and of any associated mucosal lesion; furthermore, the treatments are limited and sometimes frustrating. ³ While diverticulosis represents a relative contraindication to peritoneal dialysis (PD), there is no indication about AD and there are no reports about the effect of PD on it.

Case report

Case 1

A 63-year-old Caucasian man on PD because of nephroangiosclerosis was admitted to hospital for severe onset anemia. The patient had history of smoking, hypertension, sleep apnea syndrome, moderate-to-severe aortic stenosis, and paroxysmal atrial fibrillation. The therapy assumed was warfarin, pantoprazole 40 mg, methoxy polyethylene glycol-epoetin beta (CERA) 365 µg/monthly, ferric gluconate sodium 62.5 mg/week (transferrin target: 200–500 mg/dL), and antihypertensive drugs. PD prescription was automated peritoneal dialysis (APD) by night plus a single isotonic bag (2 L) by day. The physical examination was unremarkable. Blood examination showed normal red cell volume, platelets, liver function, hemolysis markers, inflammatory markers, and international normalized ratio (INR) stable on target, between the value of 2 and 3. Iron stores were on target and erythropoiesis-stimulating agents had already been increased without any significant improvement. Stool examination was positive for occult blood, and colonoscopy showed a right colonic AD without bleeding signs and a Heyde’s syndrome was diagnosed.

Gastroscopy, video capsule, and abdominal sonography were all normal, and blood transfusion was the only therapy administered. In the following 6 months, the patient underwent three colonoscopies, and at the third attempt to find the bleeding source, an Argon-plasma treatment was performed, but it was not enough. Few months later also an angiography with intravascular embolization of the active lesions was made, but it has not been resolutive. For this reason, oral anticoagulant therapy for atrial fibrillation was discontinued after left atrial appendage closure, but no significant improvement of anemia was obtained. The patient’s need of blood transfusion decreased from one every 1–2 weeks to 1–2 bags every month, but they have never been suspended. Eventually, the right colonic resection was scheduled for the next major bleeding event. An arteriovenous fistula on the left arm was created and then the patient was shifted to hemodialysis (HD). After that, no major GI bleeding occurred and blood transfusion were performed no more. One month after HD start another colonoscopy was performed, without evidence of bleeding lesions, and HD was than performed with heparin to avoid dialyzer membrane clotting. Four years later, the colonic resection has not yet been performed and the patient underwent a kidney transplant successfully. No other major bleeding episodes occurred, and no other blood transfusions were needed.

Case 2

A 67-year-old Caucasian man on PD due to hypertensive nephropathy was admitted to hospital because of severe fatigue due to anemia. The patient had hypertension, ischemic heart disease, and was overweight. The therapy assumed was antihypertensive drugs, calcium carbonate, intravenous iron, and 250 µg/monthly of CERA. The PD prescription was APD by night plus a single icodextrin bag (2 L) by day. Iron stores were correct and CERA prescription has been increased and no other causes of anemia were detected. Gastroscopy was normal while colonoscopy reported a small bleeding AD in the transverse colonic tract, treated with Argon plasma. Anemia was still severe also after the endoscopic treatment so a colonoscopy was repeated after 2 weeks with negative results. Taking into account the experience from case 1, the patient was temporarily switched from PD to HD and hemoglobin values were monitored. In the following 2 weeks, a response to the anemia therapy was obtained and the red blood cell count increased. The patient in a month was no more symptomatic for anemia and 2 months after the switch, APD with initial low dialysis fluid volume was restarted for a week, then the APD volume were restored as at the beginning but the long day dwell with icodextrin was decreased to 1 L. In the following 2 years, severe anemia was not detected and no further bleeding occurred while performing PD.

Discussion

Colonic AD is a little known pathology and that makes difficult its management. It is frequent in CKD patients and for several authors bleeding is enhanced by several factors including uremic platelet dysfunction, the frequent use of acetylsalicylic acid and anticoagulant drugs, the venous hypertension caused by intermittent fluid overload, and certain phosphate binders. ⁴ For these reasons, in our patients we expected a worsening of the anemia on HD because of the use of heparin and the different body fluid control. Colonic AD is an acquired venous pathology of the submucosal layer, and it should not be misinterpreted with telangiectasia that is a congenital dilatation of both subserosal arterioles and veins. ⁵ The pathogenesis of AD is not clear, and there are several hypotheses. One of these is that the AD is caused by small mucosal ischemia due to constipation or bowel obstruction. Another one is that colonic AD is the consequence of a chronic venous obstruction due to an increased bowel tension, ⁶ which could be worsened by the rise of the abdominal pressure induced by PD solution load. In CKD patient, renal replacement therapy (RRT) is an additional variable involved in bleeding, so it is important to evaluate its role. In the literature, there are two case reports that show a clinical improvement from a HD-to-PD shift, but in those cases the vascular abnormality was a gastric antral vascular ectasia (GAVE),^7,8 pathology that has a different etiopathogenesis. Only one case report ⁹ described a case of colonic AD improved with a PD-to-HD switch, which is similar to our experience. We are conscious that these two patients are a small sample to outline a likely relationship, in particular to create recommendation, but to share an experience could turn helpful in such specialistic field. In fact, different treatments are suggested by gastroenterologists for AD and the nephrologists should know the therapeutic scenario to select the most suitable solution for patients between conservative approach, pharmacological treatments, ¹⁰ or local treatment with Argon-plasma coagulation. In particular, Argon-plasma coagulation does not reduce the bleeding recurrence and in DP patients is also associated with a significant risk of peritonitis and bowel perforation. ¹¹ Surgery can be considered resolutive option, but the sum of the anesthesiological and surgical risk added to the need of changing the RRT type is not negligible. The AD therapy must be tailored for each patient, and our experience should be taken into account in order to better modify PD prescription or to accelerate PD-to-HD switch, when scheduled.

Conclusion

The presence of colonic AD should be evaluated with caution in end-stage renal disease (ESRD) patients. PD effect on bleeding is uncertain and more attention is needed when treating these intestinal lesions, because of the high risk of peritonitis in PD patients. On the other hand, HD has also technical limitations especially for the need of anticoagulant drugs. The nephrologists must be aware of the interfering and worsening factors of GI bleeding and also of the possible changes in CKD therapies to minimize the risks and to select in advance the best RRT. All the hypotheses and considerations suggested here need to be verified, and in the meanwhile, it would be wise to take precautions when dealing with such kind of patients.