Correlation of Peripheral Blood Findings with Bone Marrow Examinations in Patients with Unexplained Pancytopenia

Abstract

Background: Pancytopenia of unknown cause can be a difficult problem in differential diagnosis due to possible connections between peripheral blood abnormalities, nutrient deficiencies, failure of the marrow and tumour diseases of the haematology specialty. This paper describes an approach to compare peripheral blood abnormalities against marrow diagnosis in terms of associations.

Method: An analysis was performed on 100 adult records retrospectively. The variables included demographics, clinical symptoms, blood measurements, peripheral smear morphology observations, marrow cellularity percentages, and marrow diagnoses. Summary descriptive statistics described the dataset. Associations were identified using Spearman correlation and Kruskal-Wallis rank sum test for cellularity correlation, and permutations testing correlations between categories. Percent agreement and Cohen's kappa coefficient measured agreement in diagnostic classification between peripheral findings and marrow examinations.

Results: The average age was 48.72 ± 19.10, with 52 percent women. Megaloblastic anaemia represented 34 percent diagnoses. Other diagnoses include aplastic anaemia (24%), acute leukaemia (16%), myelodysplasia (12%), infiltration (8%) and reactive/miscellaneous (6%). Morphological abnormalities included macrocytic erythrocytes (46), hypersegmented neutrophils (28), blast cells (14) and dysplasias (16). There were positive correlations for leucocytes (ρ=0.380; p<0.001) and thrombocytes (ρ=0.381; p<0.001) with cellularity. No correlation for haemoglobin existed (p=0.1165). Diagnostic agreement was high at 79%, with kappa coefficient κ=0.729.

Conclusions: This paper demonstrates how peripheral morphology, blood measurements and diagnoses might be combined into an analytic framework. Given the retrospective nature of the sample, no causal or diagnostic validity is implied here. Further prospective research is essential before any conclusion can be reached regarding clinical applications to cases of unexplained pancytopenia.