Limited sampling strategies for individualized BAX 855 prophylaxis in severe hemophilia A: in silico evaluation

Abstract

Objective: Limited sampling strategies (LSS) lower the burden of pharmacokinetic (PK)-guided dosing, but an extensive evaluation of LSS for BAX 855 (Adynovi) is currently lacking. This study aimed to develop a LSS for BAX 855 and combine this with a LSS of a standard half-life (SHL) factor VIII (FVIII) concentrate in a clinical setting.

Methods: Individual PK parameters of BAX 855 were estimated for 10 000 virtual patients with severe hemophilia A using Monte Carlo simulations. Several LSS consisting of 2-6 samples were examined based on patient burden, bias and accuracy of clearance, elimination half-life, volume of distribution and trough levels at 72 h (C72). Analyses were performed separately for adults and children <12 years.

Results: The preferred LSS for BAX 855 consisted of three sampling points at 15-30 min, 48 h and 72 h for both adults (mean accuracy C72: 14.0% vs. 10.8% using six samples) and children (mean accuracy C72: 14.9% vs. 11.4% using six samples). The best strategy with two samples (peak, 48 h) resulted in an adequate, but lower accuracy than strategies with ≥3 samples (mean accuracy C72: 22.3%). The optimal combination of the LSS of SHL FVIII and BAX 855 led to six samples during four clinical visits.

Conclusion: This in silico study has identified that two to three samples are necessary to estimate the individual PK of BAX-855 adequately. These samples can be collected in one or two clinical visits. When combining PK profiling of SHL FVIII and BAX 855, six samples during four clinical visits are needed.

Fig. 1

FVIII levels of 10 000 simulated patients for limited sampling strategy analysis. The left figure depicts the FVIII levels for adults (≥12 years, 8833 virtual patients) and the right figure for children <12 years (1167 virtual patients). The dotted horizontal line represents the lower limit of quantification (LLOQ, 0.01 IU/ml). For 9% and 43% of the adults the FVIII level was below the quantification limit (BQL) 70–74 h and 94–98 h after administration, respectively. For children, 12% and 52% of the FVIII levels were BQL 70–74 and 94–98 after administration, respectively. FVIII, factor VIII.

Fig. 2

Relative error between the true simulated PK parameters and PK parameters estimated by Bayesian analysis for a selection of the evaluated limited sampling strategies of BAX 855 for adults (≥12 years). The parameters elimination half-life (a) and FVIII level at 72 h after dose (b) are presented. The boxes of the boxplots present the median (middle line) and interquartile range (IQR) and the whiskers extend to the 5th and 95th percentile. The green boxes present the preferred limited sampling strategies (LSS7 and LSS14) with the smallest relative errors. For readability, the y-axis of the C₇₂ plots are limited from −50% to 50%, while maintaining all data. PK, pharmacokinetics.

Fig. 3

Limited sampling PK profile of SHL FVIII and BAX 855 combined, resulting in the fewest visits to the clinic. The blue color indicates dosing and sampling points of the extended half-life concentrate, while the red color refers to the standard half-life concentrate. Figure 3A demonstrates the combination of PK profiling of both the SHL-FVIII concentrate and BAX855. Figure 3B demonstrates PK profiling separately, when the first sample is taken at a set time point after home infusion of the factor concentrates. FVIII, factor VIII; PK, pharmacokinetics; SHL, standard half-life.