Steady state haemolysis and cytoprotective protein levels in African children with sickle cell disease

Sub-Saharan Africa bears the highest burden of all Sickle Cell disease births worldwide. Chronic haemolysis in children with sickle cell disease (SCD) is known to cause multi-organ damage, increased risk of stroke, and cognitive impairment. This study sought to investigate haemolysis biomarkers in sickle cell disease (SCD) among sub-Saharan African children, aiming to improve individualized management through enhanced diagnostic and prognostic capabilities. Fifty children with SCD and 32 non-SCD children aged 2-17 years were evaluated using 5-part differential FBC and ELISA to profile haemolysis and haem cytoprotective proteins. Significantly elevated levels of bilirubin, free haemoglobin, haem oxygenase, ferritin, potassium ions, and AST activity in SCD participants was observed, while haptoglobin was significantly reduced. Hydroxyurea treatment was associated with increased haptoglobin and ferritin levels and decreased free haemoglobin and haem oxygenase activity. Male children with SCD exhibited higher haem oxygenase activity and free haemoglobin levels. A comparative analysis of machine learning algorithms revealed that the random forest model achieved 100% accuracy, sensitivity, specificity, predictive, and AUC values in classifying SCD. Direct bilirubin emerged as the most important classifier, followed by potassium, haptoglobin, free haemoglobin, haem oxygenase, total bilirubin, and ferritin. This research highlights the potential of machine learning-based classification using haemolysis biomarkers for improved SCD diagnosis and management. The findings from this sub-Saharan African cohort may have broader implications for SCD patient populations worldwide, potentially revolutionizing individualized treatment approaches and enhancing patient outcomes.