The risk of bleeding following endoscopic biopsy in patients with thrombocytopenia

Endoscopic biopsies are often performed on thrombocytopenic patients (pts) following hematopoietic stem cell transplantation to evaluate for opportunistic infections or graft-versus-host disease (GVHD). The risk of bleeding and its relationship to platelet count are unknown in this pt population. The aim of this study was to assess the risk of immediate bleeding (<72 hours after a procedure) in a pt population among whom thrombocytopenia is common. Methods: We reviewed the hospital records of pts, in whom the histological diagnosis of GI GHVD had been made by endoscopy over a 7 year period, and obtained the following information: evidence of overt GI bleeding (melena, hematochezia, bloody diarrhea), hemoglobin (hb) changes, pit counts, need for pit or blood transfusions and hemodynamic changes. The reason for a decline in hemoglobin or need for a blood transfusion post procedure was determined in each instance. All procedures in which biopsies were performed, including those that did not detect GVHD, were included in the analysis. Typically, 2 to 4 biopsies were obtained with a standard forceps at an individual site. Two procedures performed in pts who had uncomplicated biopsies but had an abnormal INR (>1.5) were excluded from analysis. Results: Forty-seven pts (33 males, 14 females) underwent 104 endoscopic procedures (30 lower, 32 upper, 21 both upper and lower) on 83 occasions and biopsies were obtained at 170 different sites. None were followed by overt GI bleeding. On 27 of the 83 occasions there was a decline in hb over a 24 hr period. In nine other instances blood transfusion was performed without a recorded drop in hb; providing a total of 36 episodes of possible bleeding in the immediate post-procedure period. Each of these cases was reviewed, in detail, to establish the reason for hb decline and/or transfusion. In no instance was it felt that bleeding was related to the GI biopsy. Thirteen episodes were attributed to GI bleeding from a pre-existent condition, one to an intracranial bleed, 2 to pre-existent anemia; in 20 episodes no cause was defined. The table presents the range of platelet counts for each procedure day: Pit ct at biopsy (K) <30 30-49 50-79 80-109 110-139 ≥140 Endoscopy encounters 3 16 31 12 5 16 On 15 occasions a follow-up pit count had not seen performed following pit transfusion; pretransfusion count <30K (n = 3), 30-49K (n = 9), and 50-79K (n=31). Twenty pts had a pit transfusion before their endoscopy, 6 after, and 20 both before and after. Thirteen pts had plt cts maintained above 30K for one day, the remaining pts were maintained at a level of >30K for at least 2 days. Conclusion: By adherence to a protocol of performing GI biopsies only if platelet count is >50K and maintained above 30K or at least 2 days the incidence of GI hemorrhage is minimized.