Endoscopists frequently face a dilemma when confronted with severe thrombocytopenia or anemia in patients with portal hypertension.In routine practice, platelet counts < 50,000/µL or hemoglobin levels < 7 g/dL often trigger anxiety, leading to preprocedural transfusions or even deferral of endoscopic intervention despite limited supporting evidence.In this issue of the Journal of Digestive Endoscopy, the authors present a retrospective observational study evaluating the safety and outcomes of diagnostic and therapeutic upper gastrointestinal endoscopy procedures related to portal hypertension in this high-risk population. 1 Thrombocytopenia in cirrhosis is multifactorial, resulting from hypersplenism, reduced thrombopoietin production, and bone marrow suppression, and is frequently interpreted as a surrogate of bleeding risk despite limited evidence supporting such an association. 2 The investigators analyzed 101 patients with portal hypertension undergoing upper gastrointestinal endoscopy, comparing those with severe thrombocytopenia and anemia to those with relatively preserved hematological parameters.Importantly, patients with low platelet counts or hemoglobin did not demonstrate higher rates of immediate complications, rebleeding, or mortality. 1This was observed despite a greater need for therapeutic interventions such as endoscopic variceal ligation in the cytopenic group.Additionally, although these patients experienced longer hospital stays and higher rehospitalization rates, these outcomes were attributable to the severity of underlying liver disease rather than cytopenias themselves. 1 These findings align with the evolving concept of "rebalanced hemostasis" in cirrhosis, wherein reductions in procoagulant factors are offset by decreases in anticoagulant proteins along with compensatory increases in von Willebrand factor. 3Consequently, conventional laboratory parameters such as platelet count and international normalized ratio often fail to accurately predict bleeding risk.
Parvez et al. (Mon,) studied this question.