Table 1.

Parameter estimates of the final population PK model and bootstrap (95% non-parametric confidence interval) for SV, SVA, HMSV, HMSVA, and DHSV

Drug	Parameters	Structural model		Between subject variability (BSVb)
		Estimatea	Bootstrap	Estimate [shrinkage(%)]	Bootstrap
SV	D1 (h−1)	0.069	0.064–0.073	–	–
	D2 (h−1)	0.39	0.34–0.42	2.07 [17]	1.75–2.27
	ka1 (h−1)	0.030	0.027–0.032	–	–
	ka2 (h−1)	0.41	0.38–0.46	0.41 [0.5]	0.38–0.45
	BA*	0.78	0.70 – 0.89	–	–
	ALAG (h)	0.18	0.17–0.21	1.70 [13]	1.40–2.09
	CLSLe/F (L/h)	1300	1210–1450	0.63 [3]	0.53–0.71
	VSL/F (L)	110	98.7–114	2.17 [27]	1.59–2.25
SVA	CLLA/VSVA (h−1)	0.043	0.04–0.046	0.96 [9]	0.76–1.02
	CLSVAe/VSVA (h−1)	0.13	0.11–0.14	0.79 [25]	0.60–0.81
	θ561: c.521T>C on VSVA	− 0.37	− 0.38 to − 0.35	–	–
	θ562: c.521T>C on CLLA	0.93	0.86–1.04	–	–
HMSV	CLLH/VHSV (h−1)	16.1	15–21.5	0.26 [32]	0.23–0.28
	CLEmax/VHSV (nM/h)	660	637–967	–	–
	CLEC50 (nM)	20	17.8–22.8	0.67 [6]	0.58–0.74
	γ	0.86	0.83–0.87	–	–
HMSVA	CLAH/VHSVA (h−1)	0.48	0.39–0.5	0.33 [60]	0.27–0.36
	CLHSVAe/VHSVA (h−1)	4.91	4.6–5.1	0.18 [47]	0.12–0.19
	CLHSVHSVA/VHSVA (h−1)	0.55	0.52–0.62	0.51 [4]	0.45–0.56
	θAGE1: age (Fra) on CLHSVAe	1 (fixed)	–	–	–
	θAGE2: age (50%) on CLHSVAe (year)	5.34	4.9–5.5	–	–
DHSV	CLLD/VDHSV (h−1)	4.11	3.8–5.0	0.29 [34]	0.23–0.29
	CLDHSe/VDHSV (h−1)	12.9	12.3–14.2	0.28 [42]	0.22–0.29
	θAGE3: age (Fra) on CLDHSe	0.90	0.83–0.94	–	–
	θAGE4: age (50%) on CLDHSe (year)	6.3	5.9–6.6	–	–
Residual Variabilityc
SV	eps1SV	–	–	0.38	0.33–0.44
	eps2SV			0.32	0.28–0.35
	eps3SV			0.30	0.29–0.33
	mSV			0.012	0.011–0.012
SVA	eps1SVA			0.26	0.23–0.29
	eps2SVA			0.16	0.13–0.16
	mSVA			0.22	0.19–0.24
HMSV	eps1HMSV			0.29	0.27–0.34
	eps2HMSV			0.24	0.22–0.28
	eps3HMSV			0.12	0.11–0.28
	mHMSV			0.0001 (fixed)	–
HMSVA	eps1HMSVA			0.18	0.16–0.20
	eps2HMSVA			0.1	0.095–0.11
	mHMSVA			0.0001 (fixed)	–
DHSV	eps1DHSV			0.24	0.21–0.27
	eps2 DHSV			0.21	0.19–0.23
	eps3 DHSV			0.086	0.08–0.09
	mDHSV			0.0001 (fixed)	–

*F₁ = 1/(1 + BA), F₂ = BA/(1 + BA)

^aPopulation parameter estimates for a typical individual in the study, 14 years old, 80 kg and homozygous variant CC genotype for the rs4149056

^bBSV is expressed as CV (coefficient of variation) calculated as $\sqrt{\left(e^{{\omega }^2}-1\right)}$

^cResidual error variability for the analytes were based on the double exponential error model for log-transformed data with time dependency for some analytes. ln(y) = ln(f + m) + (f/(f + m)ε₁ + (m/(m + f))ε₂ where y is the observed concentration, f is the model prediction, m is a positively constrained parameter, and ε₁ and ε₂ are random errors, assumed to be normally distributed with means of zero and variance of eps1 and eps2, respectively. m is estimated or fixed to an estimate around or lower than the LLOQ in order to minimise bias

SV—eps1_SV and eps2_SV correspond to 0–2 h and 2–8 h for ε₁ and eps3_SV and m_SV to ε₂ and m, respectively. SVA—eps1_SVA, eps2_SVA, and m_SVA correspond to ε₁, ε₂, and m respectively. HMSV—eps1_HMSV and eps2_HMSV correspond to 0–2 h and 2–8 h for ε₁ and eps3_HMSV and m_HMSV to ε₂ and m, respectively. HMSVA—eps1_HMSVA, eps2_HMSVA, and m_HMSVA correspond to ε₁, ε₂, and m, respectively. DHSV—eps1_DHSV and eps2_DHSV correspond to 0–2 h and 2–8 h for ε₁ and eps3_DHSV and m_DHSV to ε₂ and m, respectively