Severe Anemia from Multiple Gastric Hyperplastic Polyps in a Hemodialysis Patient after Long-term Use of a Proton-pump Inhibitor

Abstract

A 90-year-old man on maintenance hemodialysis was admitted due to severe symptomatic anemia. Biopsies under esophagogastroduodenoscopy demonstrated that the cause of anemia was intermittent blood oozing from multiple gastric hyperplastic polyps. Even after successful eradication of Helicobacter pylori, he showed hypergastrinemia (480 pg/mL) owing to esomeprazole (proton-pump inhibitor) therapy for the past 4.5 years to treat reflux esophagitis. Seven months after we switched esomeprazole to famotidine (H₂-receptor antagonist), those gastric polyps and anemia were remarkably ameliorated with lowered gastrin levels. This case indicates that long-term use of a proton-pump inhibitor triggers chronic hypergastrinemia, leading to gastric hyperplastic polyps and subsequent severe anemia.

Keywords: gastric bleeding; gastric hyperplastic polyp; gastrin; hemodialysis; proton-pump inhibitor; severe anemia.

Figure 1.

Endoscopic images and histologic assessments of polyps at the first esophagogastroduodenoscopy examination performed one month before admission. (A) The endoscopic assessment revealed multiple gastric polyps located at the gastric body. Many polyps showed redness on their tops, indicating their vascular-rich structure. Some polyps were pedunculated. (B) We performed multiple biopsies from those polyps, including a large pedunculated polyp (arrow), using biopsy forceps (arrowhead). (C, D) Histology of a biopsy sample from gastric hyperplastic polyps at the gastric body showed (C) hyperplasia of foveolar epithelium (arrows) and mucus accumulation in the lumens of the elongated gastric glands (arrowhead) and (D) dilated capillaries (arrows) and inflammatory cell infiltration (arrowhead). Mechanical damage to polyps, uremia-related acid injury, use of heparin, and regular alcohol consumption could have caused prolonged oozing of blood from the capillaries in this case. C, D: Hematoxylin and Eosin staining. Magnification ×100.

Figure 2.

Measurements of hemoglobin, serum gastrin, and mean corpuscular volume (MCV) along the time course before and after admission. (A) Measurements of the hemoglobin level (g/dL). Day 0 denotes the day of admission to our hospital. After we changed esomeprazole (20 mg daily) to famotidine (10 mg daily) (35 days after admission), his hemoglobin level trended upward to 12 g/dL. Accordingly, we were able to decrease the dose of darbepoetin alfa (erythropoiesis-stimulating agent) from 90 to 40 μg/week. However, bleeding from the hyperplastic polyps continued, as indicated by positive findings of fecal occult blood tests (fecal Hb). Although his fecal occult blood test was negative at day 271, it became positive after he started to take a low dose of lansoprazole for 5 months. A single asterisk indicates the transfusion of 6 units of packed red blood cells at day 5. A double asterisk indicates the progression of anemia due to bipolar hip arthroplasty for his left femoral neck fracture at day 205. A horizontal dotted line indicates the hemoglobin level (7 g/dL), below which blood transfusion is necessary (14). (B) Measurements of the serum gastrin levels (pg/mL). Day 0 denotes the day of admission to our hospital. Serum gastrin levels quickly decreased within 30 days after esomeprazole was switched to famotidine. Subsequently, the hemoglobin levels increased to 11 g/dL, as shown in (A). The serum gastrin level was still close to the upper limit of our reference value (125 pg/mL) at day 415 postadmission, even though we had prescribed a low dose of lansoprazole. A horizontal dotted line indicates the upper limit of our reference value (125 pg/mL). (C) Measurements of the mean corpuscular volume (MCV) (femtoliter). Day 0 denotes the day of admission to our hospital. In our case, low MCV values imply iron deficiency anemia due to bleeding from gastric hyperplastic polyps. MCV values decreased at day 152, probably because some gastric hyperplastic polyps had incompletely healed. MCV values decreased again at day 348, probably because bleeding from hyperplastic polyps recurred with a low dose of lansoprazole (15 mg daily). Asterisks indicate the examination time points of esophagogastroduodenoscopy (days -29, 12, 260, and 421).

Figure 3.

Endoscopic images of multiple gastric hyperplastic polyps and reflux esophagitis at the 2nd, 3rd, and 4th esophagogastroduodenoscopy examinations. The 2nd, 3rd, and 4th endoscopic tests were performed 12, 260, and 421 days after admission, respectively. (A-C) Time-course changes in gastric hyperplastic polyps in the gastric body. (A) 2nd endoscopic test. Multiple gastric hyperplastic polyps had formed due to the long-term use of esomeprazole. Some polyps were pedunculated (arrow). (B) 3rd endoscopic test. Most gastric hyperplastic polyps had disappeared or regressed without esomeprazole for eight months. One pedunculated polyp remained (arrow). (C) 4th endoscopic test. Small gastric hyperplastic polyps (arrowhead) had relapsed with a low dose of lansoprazole for five months. One pedunculated polyp increased in size (arrow). (D-F) Time-course changes in reflux esophagitis. A dotted line denotes the gastroesophageal junction. (D) 2nd endoscopic test. No esophagitis was detected with the long-term use of esomeprazole, whereas esophageal hiatus hernia and one hyperplastic polyp (arrow) in the junctional mucosa were detected. (E) 3rd endoscopic test. Reflux esophagitis had developed without esomeprazole for eight months. Arrows indicate multiple mucosal breaks (Los Angeles classification grade B). (F) 4th endoscopic test. Reflux esophagitis had been ameliorated with a low dose of lansoprazole for five months. (G) The 3rd endoscopic test identified black spots in the gastric mucosa of the fundus, a typical finding in patients treated with a proton-pump inhibitor.