Abstract
OBJECTIVE: The role of bone marrow examinations in patients with primary immune thrombocytopenia (ITP) is uncertain. The objectives of this study were to determine the inter-rater reliability of bone marrow examinations and to identify distinguishing morphological features of ITP bone marrows under controlled conditions. METHODS: Histological slides of bone marrow biopsy specimens and aspirates from 32 adult patients with severe primary ITP who had failed a median of two treatments, and 51 non-thrombocytopenic controls were retrieved from hospital archives. Slides were arranged in random order in a slide box and coded. Blinded to the diagnosis and platelet counts, three independent hematopathologists were asked to identify the ITP bone marrows and to evaluate megakaryocyte number, morphology, and distribution. RESULTS: Overall chance-corrected agreement on ITP classification among the three raters was poor [kappa (κ) = 0.30; 95% confidence interval 0.22-0.38]. Raters were generally unable to correctly identify the ITP bone marrows from controls. Increased number of megakaryocytes, while an uncommon finding, was more frequent among ITP patients compared with controls (6/32, 18.8%; vs. 2/51, 3.9%; P = 0.05), and abnormal megakaryocyte morphology often led individual raters to reach a diagnosis of ITP. Overall sensitivity and specificity of bone marrow examinations were 24% and 90%, respectively. CONCLUSIONS: This study confirms methodologically that bone marrow examinations are unreliable and frequently non-diagnostic in ITP. Thus, they are not useful for patients with typical disease. Rare subsets of patients with severe ITP demonstrated unique features such as increased number of megakaryocytes.