Empirical evaluation of MCV in differentiating sickle cell and β-thalassemia phenotypes in Western Kenya: a receiver operating characteristic (ROC) analysis

Mean corpuscular volume (MCV) is commonly employed in resource-limited settings as a preliminary screening tool for haemoglobinopathies, based on the assumption that these disorders typically present with microcytosis. However, there is a lack of locally validated MCV cut-offs for haemoglobinopathy detection in Kenya, which limits its reliability for both clinical and research applications. In this study, we retrospectively analysed 773 participants, including 388 HbAA controls and 385 individuals with confirmed haemoglobinopathies, whose haemoglobin phenotypes were determined using high-performance liquid chromatography (HPLC). Receiver operating characteristic (ROC) analysis demonstrated poor discriminatory performance of MCV for combined haemoglobinopathies, with an area under the curve (AUC) of 0.528. The optimal MCV threshold of 75.3 fL yielded 50.4% sensitivity and 42.8% specificity, indicating limited diagnostic utility. Sickle cell disease (SCD) phenotypes in this cohort were predominantly normocytic when interpreted using population-based Kenyan reference ranges, contrasting with iron deficiency and other microcytic anemias that typically reduce MCV. The ROC-derived optimal threshold for SCD detection (76.9 fL) fell within the population reference distribution (median 76.8 fL; IQR 11.6), indicating that maximum detection occurs within normocytic ranges. Sickle cell trait and thalassemia traits were sometimes microcytic but overlapped with other conditions. These findings emphasize the importance of using locally derived MCV reference ranges rather than conventional 80–100 fL thresholds. While MCV may assist in triaging suspected cases in resource-limited settings, it cannot reliably identify carrier states, and definitive haemoglobin characterization remains essential for accurate diagnosis and effective screening.