Can haematological indices predict positive findings at endoscopy in anaemic patients?

INTRODUCTION: Patients with anaemia are commonly referred for bidirectional endoscopy. The aim of this study was to determine if any haematological parameters could predict positive findings at endoscopy. PATIENTS AND METHODS: A total of 209 patients had bidirectional endoscopies performed for anaemia between September 2002 and March 2004. The endoscopy reports, histology and full blood count results (haemoglobin [Hb], red blood cells [RBCs], packed cell volume [PCV], mean cell volume [MCV] and mean cell haemoglobin [MCH]) were then reviewed. Statistical analysis was performed using non-parametric tests.
RESULTS: Overall, 197 patients had successful bidirectional endoscopies with 12 requiring completion barium enema. In 48 (23%) of these patients, a cause of anaemia was found with 15 (7.2%) carcinomas detected (2 upper GI and 13 lower GI). There was a significant difference in haemoglobin (9.2 g/dl versus 10.1 g/dl; P = 0.0044), RBCs (3.56 x 10(12)/l versus 3.83 x 10(12)/l; P = 0.0325) and PCV (0.279 l/l versus 0.31 l/l; P = 0.0112) between patients with positive findings at endoscopy and those with a normal investigation. Cancer patients had significantly lower haemoglobin (8.65 g/dl versus 10.1 g/dl; P = 0.0103), RBCs (3.45 x 10(12)/l versus 3.83 x 10(12)/l; P = 0.0179) and PCV (0.27 l/l versus 0.31 l/l; P = 0.0298) compared with patients with normal endoscopies. There was no significant difference in the other haematological parameters between those found to have positive findings and those that had normal endoscopies.
CONCLUSIONS: Based on this study, the yield of bidirectional endoscopy is low, with haemoglobin and PCV being the most useful haematological indices of significant pathology. Ferritin and MCV did not predict the likelihood of finding a gastrointestinal cause for the anaemia.