Systematic Development of a Gradient Elution HPLC Method for the Analysis of Voxelotor and Its Structurally Related Substances Applying Analytical Quality by Design Approach

Abstract

Sickle cell disease is a genetic disorder treated with Voxelotor (Vox), the first hemoglobin oxygen-affinity regulator. Although Vox recently received expedited approval in the United States for the sickle cell disease therapy for certain age groups, no monograph is yet available in the official compendia for Vox. Therefore, current study aims to develop a stability-indicating HPLC method for determining five potential impurities in Vox drug substance. The procedure was developed by implementing principles of analytical quality by design (AQbD). A detailed risk evaluation was conducted in accordance with the cause-and-effect relationship. A factorial design was employed to identify the impact of the critical method parameters (CMPs) and their relationship to critical quality attributes (CQAs). The full factorial design was utilized and final method was optimized. The p-values of the model and lack of fit were < 0.0001 and > 0.05, respectively, indicating the best-fit statistical model for the studied responses. The CMPs, such as gradient slope (ratio of mobile phase-B % and gradient time), flow rate of 1.0 mL min⁻¹, and column oven temperature of 40 °C, were optimized from the full factorial design. The best possible separation among all impurities and Vox was achieved with gradient elution using X-Bridge C₁₈, 150 mm × 4.6 mm, 3.5 µm analytical column. The optimized gradient is time (min)/%B: 0.0/15, 3.0/15, 20.0/80, 30/80, 31/15, 35/15. The peak pairs where the separation was highly critical were: impurity-4/impurity-5 and impurity-5/Vox. Method validation, revealed that the mean recovery of the impurities ranged between 100 and 106%, the correlation coefficient (r) > 0.99, across the span for LOQ–150% levels, RSD values (n = 6) ranged between 1.2 and 4.6% for 100% level (i.e., 0.1% of impurities). The peaks from the specificity analysis did not intervene with the known and active analyte (Vox) peak and this study identified impurity-2 as a major degradation product.