Optimisation du Contrôle Qualité Interne des Analyseurs d'Hématologie par l'Utilisation d'Echantillons Patients

Abstract

Automated hematology occupies a central place in modern medical biology, with the complete blood count (CBC) being one of the most frequently requested tests for the diagnosis and therapeutic monitoring of patients. While current hematology analyzers offer remarkable analytical performance, their reliability depends on the implementation of rigorous quality control systems. However, in many Algerian public hospital laboratories, including the hemobiology unit of the CHU Saâdna Abdenour of Sétif, access to commercial control materials remains difficult due to budgetary, logistical and supply constraints. Faced with this reality, patient-based quality control (PBQC) represents a scientifically validated and economically viable complementary alternative. This prospective observational analytical study, conducted over a period of four months (January to April 2026), aimed to evaluate the feasibility and effectiveness of this approach on the main Nihon Kohden Celltac G MEK-9100 analyzer, through three complementary protocols: the 24-hour duplicate testing, the inter-instrument comparability study (MEK-9100 versus Swelab Lumi, n = 150 samples) and the intra-series repeatability test (11 series of 11 measurements over 11 weeks), covering four hematological parameters: hemoglobin (Hb), red blood cells (RBC), white blood cells (WBC) and platelets (PLT). The results of the duplicate testing demonstrated excellent analytical stability for WBC, RBC and Hb between H0 and H24 (correlations r ≥ 0.979; p > 0.05), with 97% of samples meeting the acceptability criterion of ±2 standard deviations, excluding the presence of systematic random errors. In contrast, platelets showed statistically significant instability (Z = −3.675; p < 0.001), confirming their sensitivity to prolonged storage conditions. Inter-instrument comparability revealed excellent Spearman correlations for all four parameters (ρ ≥ 0.977; p < 0.0001), all above the threshold required by CLSI H26-A2 (ρ ≥ 0.975). Bland-Altman analysis identified low and clinically acceptable systematic biases for WBC (+0.256 G/L), RBC (+0.015 M/µL) and Hb (+0.117 g/dL), while platelets showed the widest limits of agreement (−40.6 to +40.1 × 10³/µL), reflecting more pronounced inter-device variability. Passing-Bablok regression confirmed these findings, revealing statistically significant but clinically limited biases for normal values. Finally, the repeatability test demonstrated excellent performance for all parameters, with mean coefficients of variation well below the admissible thresholds: Hb (CV% = 0.45%), RBC (0.77%), WBC (0.83%) and PLT (1.46%). This study validates the feasibility and usefulness of PBQC as a complementary tool to conventional quality control in a resource-constrained Algerian university hospital setting. The MEK-9100 analyzer demonstrates analytical performance in accordance with international recommendations. However, platelets remain the most critical parameter,requiring particular vigilance around clinical decision thresholds. This approach cannot replace quality control using commercial materials, but constitutes a pragmatic and complementary strategy to ensure the analytical safety of results reported to clinicians.