Diagnostics of the novel model
The internal validity throughout the visual predictive check (Figure 3)
shows observed FIX activity levels being adequately predicted by the
novel model. Bootstrap results are presented in Table 2. The trend lines
in the GOF plots are close to the line of identity for both all patients
(Fig. 1C) and children <12 years of age (Fig. 1D). The trend
line in children <12 years of age (Fig. 1D) shows a slight
deviation at high FIX activity levels, but this may be caused by the
sparse number of samples in this range. A slight bias was detected for
the novel model, as the median PE was 3.4% (IQR -22.2 – 25.8) for all
patients and 6.5% (IQR -20.8 – 27.5) for children <12 years
of age (Supplementary Table 2A).
CWRES plots (Sup. Fig. 2C and Fig. 2D) show an improved fit compared to
the published model (Sup. Fig. 2A and 2B). In addition, the vast
majority of values in the new model is within the warranted -2 and 2
range32, in contrast to the values of the published
model.
Typical parameter values (Table 2) differed between the published and
novel model. For a typical 16 year old 73 kg patient, CL was 1.41 dL/h
and lower than the value of the published model (2.39 dL/h). In
addition, distribution volume at steady-state (Vss) was lower as well
with respective values of 153 dL and 198 dL. Likewise, terminal
t1/2 was lower in the novel model compared to the
published model for children <12 years of age (70 h vs 88 h),
adolescents ≥12 and <18 years of age (76 h vs 99 h) and adults
(88 h vs 101 h) (Supplementary Table 3).
Lastly, the validity of the model for children <12 years of
age is illustrated in Figure 2C and 2D. The figure demonstrates random
variability of CL and V with an average not different from zero. Of
note, an adequate covariate model shows no trend and deviations from
zero in the IIV.