Skip to contents

Etanercept (Fang 2010)

Source: vignettes/articles/Fang_2010_etanercept.Rmd
Fang_2010_etanercept.Rmd

Model and source

  • Citation: Fang Y, Li LJ, Wang R, Huang F, Song HF, Tang ZM, Li YZ, Guan HS, Zheng QS. Population pharmacokinetics of rhTNFR-Fc in healthy Chinese volunteers and in Chinese patients with Ankylosing spondylitis. Acta Pharmacologica Sinica. 2010;31(11):1500-1507.
  • Article: doi:10.1038/aps.2010.113

rhTNFR-Fc is a recombinant human tumor necrosis factor receptor type II (p75) extracellular-domain Fc fusion protein. Etanercept (Enbrel, Amgen / Wyeth) is the prototype rhTNFR-Fc and the first approved for ankylosing spondylitis (AS); the product studied by Fang 2010 was a Chinese rhTNFR-Fc supplied by Celgen Bio-Pharmaceutical Co Ltd (Shanghai), structurally the same class of soluble TNF-alpha receptor fusion protein. The published popPK parameters are consistent with the reported American etanercept popPK values (Discussion: “the pharmacokinetics of rhTNFR-Fc in healthy Chinese subjects and in Chinese patients with AS were similar… PK parameters in Chinese subjects were similar to those reported for American subjects”).

The Fang 2010 final model is a one-compartment system with first-order SC absorption, an absorption lag time, and linear elimination. Two covariates were retained after stepwise selection: sex on apparent clearance CL/F (males have 0.655x the female CL/F), and a multiple-dose indicator on apparent bioavailability F (multi-dose AS-patient administration has 0.674x the single-dose-healthy-volunteer F).

Population

The popPK dataset pooled two arms (Table 1 of Fang 2010):

  • 32 healthy Chinese volunteers (16 male + 16 female, age 25-35 years, body-mass index 19-24 kg/m^2) received a single SC dose at 12.5, 25, 37.5, or 50 mg (n = 8 per dose level, balanced 4M / 4F).
  • 19 Chinese male patients with moderate-and-active AS (age 20-31 years, body-mass index 19-24 kg/m^2; 1 of the 20 enrolled patients was excluded for a missed dose) received seven consecutive SC injections of either 25 mg twice-weekly (BIW; n = 10) or 50 mg once-weekly (QW; n = 9).

All injections were delivered subcutaneously in the abdomen at 08:00 before breakfast. Plasma rhTNFR-Fc was measured by the Quantikine human sTNF RII immunoassay (R&D Systems); 1187 plasma samples were collected at 2-480 h post the first dose.

The same information is available programmatically via readModelDb("Fang_2010_etanercept")$population.

Source trace

The per-parameter origin is recorded as an in-file comment next to each ini() entry in inst/modeldb/specificDrugs/Fang_2010_etanercept.R. The table below collects them in one place for review.

Parameter / equation Value Source location
lka (Ka, 1/h) log(0.0605) = -2.804 Fang 2010 Table 3 (Ka = 0.0605 1/h)
lcl (CL/F at SEXF = 1, MULTI_DOSE_PT = 0; L/h) log(0.168) = -1.784 Fang 2010 Table 3 (CL/F = 0.168 L/h, female-typical)
lvc (V/F, L) log(15.5) = 2.741 Fang 2010 Table 3 (V/F = 15.5 L)
ltlag (Tlag, h) log(1.03) = 0.0296 Fang 2010 Table 3 (Tlag = 1.03 h)
lfdepot (F at MULTI_DOSE_PT = 0) fixed at log(1) = 0 SC reference; absolute F is unidentifiable from SC-only data
e_sexf_cl (male-vs-female CL/F ratio, applied as ratio^(1 - SEXF)) 0.655 Fang 2010 Table 3 (theta_Gender for CL/F)
e_multi_dose_pt_f (multi-dose-vs-single-dose F ratio, applied as ratio^MULTI_DOSE_PT) 0.674 Fang 2010 Table 3 (theta_M for F)
etalka IIV (omega^2, log-scale) log(0.556^2 + 1) = 0.269 Fang 2010 Table 3 (Ka IIV = 55.6% CV)
etalcl IIV (omega^2, log-scale) log(0.333^2 + 1) = 0.105 Fang 2010 Table 3 (CL/F IIV = 33.3% CV)
etalvc IIV (omega^2, log-scale) log(0.427^2 + 1) = 0.168 Fang 2010 Table 3 (V/F IIV = 42.7% CV)
etaltlag IIV (omega^2, log-scale) log(0.818^2 + 1) = 0.512 Fang 2010 Table 3 (Tlag IIV = 81.8% CV)
Proportional residual error 0.203 (20.3% CV) Fang 2010 Table 3
Additive residual error 12.6 ng/mL (= 12.6 ug/L) Fang 2010 Table 3
d/dt(depot), d/dt(central) one-compartment SC PK with first-order absorption + linear elimination Fang 2010 Methods (Basic model selection) and final model equation
Final model equation (paper) CL/F * 0.655^Gender * exp(eta_CL); F * 0.674^M * exp(eta_F) Fang 2010 Results (Final model and the estimation of parameters)

Covariate column naming

Source column (paper) Canonical column used here Notes
Gender (1 = male, 0 = female) SEXF (1 = female, 0 = male) Canonical orientation is the inverse of the paper’s. The model file applies the effect as e_sexf_cl^(1 - SEXF) so SEXF = 1 (female) yields factor 1 and SEXF = 0 (male) yields the paper’s 0.655 male-vs-female ratio.
M (1 = multiple dosage, 0 = single dosage) MULTI_DOSE_PT Canonical multi-dose-in-patients indicator (also used by Goel 2016 Sonidegib). In Fang 2010 the multi-dose cohort is the AS-patient arm and the single-dose cohort is the healthy-volunteer arm, so the indicator is effectively subject-level.

Virtual cohort

The cohort below reproduces the six Fang 2010 dose groups (Figure 1 single-dose curves and Figure 4 single + multi-dose curves):

  • Single-dose healthy volunteers, 4M + 4F per dose level at 12.5, 25, 37.5, and 50 mg.
  • Multi-dose AS patients, all male: 25 mg BIW for seven consecutive doses (n = 10) and 50 mg QW for seven consecutive doses (n = 9).
set.seed(20101018)  # online publication date 18 Oct 2010

mod <- readModelDb("Fang_2010_etanercept")

# Per-subject covariate vector for one cohort
make_subjects <- function(ids, sexf, multi_dose_pt) {
  tibble(id = ids, SEXF = sexf, MULTI_DOSE_PT = multi_dose_pt)
}

# Single-dose healthy-volunteer cohorts: 4M / 4F per dose level (Fang 2010
# Methods: "Healthy volunteers received escalating doses ... eight subjects
# at each dose level"). Doses delivered at time 0; observations span
# 2-480 h post the first dose per Fang 2010 sampling schedule.
single_dose_cohort <- function(dose_mg, id_offset) {
  ids <- id_offset + 1:8
  sexf <- rep(c(1L, 0L), each = 4)  # 4 female (SEXF = 1) then 4 male (SEXF = 0)
  subj <- make_subjects(ids, sexf, multi_dose_pt = 0L)
  doses <- subj |>
    mutate(time = 0, amt = dose_mg, evid = 1L, cmt = "depot")
  obs_grid <- c(2, 4, 12, 24, 36, 48, 60, 72, 84, 96, 120, 144, 168, 216,
                264, 312, 384, 480)
  obs <- expand_grid(id = ids, time = obs_grid) |>
    left_join(subj, by = "id") |>
    mutate(amt = NA_real_, evid = 0L, cmt = "Cc")
  bind_rows(doses, obs) |>
    mutate(cohort = sprintf("HV %.1f mg single", dose_mg),
           dose_mg = dose_mg)
}

# Multi-dose AS-patient cohort: all male (SEXF = 0), MULTI_DOSE_PT = 1.
# Seven consecutive SC injections at the regimen interval (3.5 days BIW or
# 7 days QW). Observations span 2-480 h post the first dose per Fang 2010
# sampling schedule.
multi_dose_cohort <- function(dose_mg, interval_days, n_sub, id_offset, label) {
  ids <- id_offset + seq_len(n_sub)
  subj <- make_subjects(ids, sexf = 0L, multi_dose_pt = 1L)
  dose_times <- (0:6) * interval_days * 24  # hours
  doses <- expand_grid(id = ids, time = dose_times) |>
    left_join(subj, by = "id") |>
    mutate(amt = dose_mg, evid = 1L, cmt = "depot")
  obs_grid <- c(2, 4, 12, 24, 36, 48, 60, 72, 84, 96, 120, 144, 168, 216,
                264, 312, 384, 480)
  obs <- expand_grid(id = ids, time = obs_grid) |>
    left_join(subj, by = "id") |>
    mutate(amt = NA_real_, evid = 0L, cmt = "Cc")
  bind_rows(doses, obs) |>
    mutate(cohort = label, dose_mg = dose_mg)
}

events <- bind_rows(
  single_dose_cohort(12.5, id_offset = 0L),
  single_dose_cohort(25.0, id_offset = 8L),
  single_dose_cohort(37.5, id_offset = 16L),
  single_dose_cohort(50.0, id_offset = 24L),
  multi_dose_cohort(25.0, interval_days = 3.5, n_sub = 10,
                    id_offset = 32L, label = "AS 25 mg BIW"),
  multi_dose_cohort(50.0, interval_days = 7.0, n_sub = 9,
                    id_offset = 42L, label = "AS 50 mg QW")
) |>
  arrange(id, time, evid)
stopifnot(!anyDuplicated(events[, c("id", "time", "evid")]))

Simulation 

# Stochastic simulation across all six cohorts using the published IIV.
sim <- rxode2::rxSolve(
  mod, events = events,
  keep = c("SEXF", "MULTI_DOSE_PT", "cohort", "dose_mg")
) |>
  as.data.frame() |>
  dplyr::filter(!is.na(Cc))
#> ℹ parameter labels from comments will be replaced by 'label()'

# Typical-value replication (no IIV) for clean Figure-1 / Figure-4 overlays.
sim_typ <- rxode2::rxSolve(
  rxode2::zeroRe(mod), events = events,
  keep = c("SEXF", "MULTI_DOSE_PT", "cohort", "dose_mg")
) |>
  as.data.frame() |>
  dplyr::filter(!is.na(Cc))
#> ℹ parameter labels from comments will be replaced by 'label()'
#> ℹ omega/sigma items treated as zero: 'etalka', 'etalcl', 'etalvc', 'etaltlag'
#> Warning: multi-subject simulation without without 'omega'

Replicate published figures

Figure 1 – single-dose mean concentration-time profiles

Fang 2010 Figure 1A shows the mean plasma rhTNFR-Fc concentration-time curves for the four healthy-volunteer single-dose levels (12.5, 25, 37.5, 50 mg); Figure 1B shows the same curves on the log-concentration axis. The simulated typical-value profiles below reproduce the characteristic shape: a ~1 h absorption lag, slow first-order absorption with Tmax around 50-60 h, and slow first-order elimination with terminal half-life ~64 h in females and ~98 h in males.

fig1_dat <- sim_typ |>
  dplyr::filter(MULTI_DOSE_PT == 0L, SEXF == 1L) |>
  dplyr::group_by(time, cohort, dose_mg) |>
  dplyr::summarise(Cc_mean = mean(Cc, na.rm = TRUE), .groups = "drop") |>
  dplyr::mutate(cohort = factor(cohort, levels = paste0(
    "HV ", c("12.5", "25.0", "37.5", "50.0"), " mg single")))

p_lin <- ggplot(fig1_dat, aes(time, Cc_mean, colour = cohort)) +
  geom_line(linewidth = 0.7) +
  labs(x = "Time (h)", y = "Cc (ng/mL)",
       title = "Figure 1A -- single-dose typical profiles",
       colour = "Cohort")
p_log <- ggplot(fig1_dat, aes(time, Cc_mean, colour = cohort)) +
  geom_line(linewidth = 0.7) +
  scale_y_log10() +
  labs(x = "Time (h)", y = "Cc (ng/mL, log scale)",
       title = "Figure 1B -- log-scale",
       colour = "Cohort")
patchwork_available <- requireNamespace("patchwork", quietly = TRUE)
if (patchwork_available) {
  patchwork::wrap_plots(p_lin, p_log, ncol = 2, guides = "collect")
} else {
  print(p_lin); print(p_log)
}
Replicates Figure 1A-B of Fang 2010: typical-value plasma rhTNFR-Fc concentration-time profiles after single SC doses of 12.5, 25, 37.5, and 50 mg in healthy Chinese volunteers (typical female; SEXF = 1).

Replicates Figure 1A-B of Fang 2010: typical-value plasma rhTNFR-Fc concentration-time profiles after single SC doses of 12.5, 25, 37.5, and 50 mg in healthy Chinese volunteers (typical female; SEXF = 1).

Replicates Figure 1A-B of Fang 2010: typical-value plasma rhTNFR-Fc concentration-time profiles after single SC doses of 12.5, 25, 37.5, and 50 mg in healthy Chinese volunteers (typical female; SEXF = 1).

Replicates Figure 1A-B of Fang 2010: typical-value plasma rhTNFR-Fc concentration-time profiles after single SC doses of 12.5, 25, 37.5, and 50 mg in healthy Chinese volunteers (typical female; SEXF = 1).

Figure 4 – observed vs predicted by dose group

Fang 2010 Figure 4 shows the observed plasma concentrations against the population-predicted concentrations for each of the six dose groups, with solid points for male observations and hollow points for female, and the population-predicted line stratified by sex. The simulated median + 5-95th percentile band by cohort below replicates the typical-value prediction band against which the paper’s observations were compared.

fig4_dat <- sim |>
  dplyr::group_by(time, cohort) |>
  dplyr::summarise(
    Q05 = quantile(Cc, 0.05, na.rm = TRUE),
    Q50 = quantile(Cc, 0.50, na.rm = TRUE),
    Q95 = quantile(Cc, 0.95, na.rm = TRUE),
    .groups = "drop"
  )

ggplot(fig4_dat, aes(time, Q50)) +
  geom_ribbon(aes(ymin = Q05, ymax = Q95), alpha = 0.25) +
  geom_line(linewidth = 0.7) +
  facet_wrap(~ cohort, scales = "free_y") +
  labs(x = "Time (h)", y = "Cc (ng/mL)",
       title = "Figure 4 -- median + 5-95% band by cohort")
Replicates Figure 4 of Fang 2010: median (line) and 5-95th-percentile band (ribbon) of simulated plasma rhTNFR-Fc concentration-time profiles by cohort. Multi-dose cohorts show the seven-injection accumulation pattern observed in AS patients (Figure 4E single + multi 25 mg BIW; Figure 4F single + multi 50 mg QW).

Replicates Figure 4 of Fang 2010: median (line) and 5-95th-percentile band (ribbon) of simulated plasma rhTNFR-Fc concentration-time profiles by cohort. Multi-dose cohorts show the seven-injection accumulation pattern observed in AS patients (Figure 4E single + multi 25 mg BIW; Figure 4F single + multi 50 mg QW).

PKNCA validation

PKNCA is applied to the four single-dose healthy-volunteer cohorts (12.5, 25, 37.5, 50 mg) over the 0-480 h sampling window. Single-dose Cmax, Tmax, AUC to last quantifiable concentration (AUClast), and terminal half-life are computed per subject and summarised by treatment (dose level x sex). Multi-dose AS cohorts are not summarised by PKNCA because the 480 h sampling spans only 2-3 dosing intervals.

# Single-dose subset (HV cohorts), labelled by dose and sex.
sd_sim <- sim |>
  dplyr::filter(MULTI_DOSE_PT == 0L, Cc > 0) |>
  dplyr::mutate(
    sex_label = ifelse(SEXF == 1L, "F", "M"),
    treatment = sprintf("%.1f mg (%s)", dose_mg, sex_label)
  )

sd_dose <- events |>
  dplyr::filter(evid == 1L, MULTI_DOSE_PT == 0L) |>
  dplyr::mutate(
    sex_label = ifelse(SEXF == 1L, "F", "M"),
    treatment = sprintf("%.1f mg (%s)", dose_mg, sex_label)
  ) |>
  dplyr::transmute(id, time, amt, treatment)

conc_obj <- PKNCA::PKNCAconc(sd_sim,  Cc  ~ time | treatment + id)
dose_obj <- PKNCA::PKNCAdose(sd_dose, amt ~ time | treatment + id)

intervals <- data.frame(
  start     = 0,
  end       = 480,
  cmax      = TRUE,
  tmax      = TRUE,
  auclast   = TRUE,
  half.life = TRUE
)

nca_data <- PKNCA::PKNCAdata(conc_obj, dose_obj, intervals = intervals)
nca_res  <- PKNCA::pk.nca(nca_data)
#> Warning: Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Warning: Requesting an AUC range starting (0) before the first measurement (4)
#> is not allowed
#> Warning: Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Warning: Requesting an AUC range starting (0) before the first measurement (4)
#> is not allowed
#> Warning: Requesting an AUC range starting (0) before the first measurement (2)
#> is not allowed
#> Warning: Requesting an AUC range starting (0) before the first measurement (4)
#> is not allowed
#> Warning: Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (2) is not allowed
nca_summary <- summary(nca_res)
knitr::kable(
  nca_summary,
  caption = "PKNCA single-dose Cmax / Tmax / AUClast / half-life by dose level x sex (n = 4 subjects per treatment cell)."
)
PKNCA single-dose Cmax / Tmax / AUClast / half-life by dose level x sex (n = 4 subjects per treatment cell).
start end treatment N auclast cmax tmax half.life
0 480 12.5 mg (F) 4 NC 554 [46.2] 30.0 [12.0, 48.0] 39.2 [13.6]
0 480 12.5 mg (M) 4 NC 467 [37.8] 48.0 [36.0, 60.0] 123 [71.1]
0 480 25.0 mg (F) 4 NC 1010 [45.7] 24.0 [24.0, 36.0] 93.3 [79.8]
0 480 25.0 mg (M) 4 NC 1170 [50.0] 36.0 [36.0, 72.0] 83.3 [31.5]
0 480 37.5 mg (F) 4 NC 1460 [38.0] 48.0 [24.0, 60.0] 83.9 [48.5]
0 480 37.5 mg (M) 4 NC 1580 [24.2] 48.0 [36.0, 60.0] 126 [79.1]
0 480 50.0 mg (F) 4 NC 2230 [16.5] 48.0 [24.0, 48.0] 63.6 [30.5]
0 480 50.0 mg (M) 4 NC 1800 [26.3] 42.0 [24.0, 48.0] 211 [184]

Comparison against published values

Fang 2010 does not tabulate formal NCA Cmax, Tmax, or AUC values; it reports only the structural-parameter estimates of Table 3 and the graphical concentration-time profiles of Figures 1, 3, 4. The model implies the following secondary quantities at typical (no-IIV) covariates:

Quantity Paper value (Fang 2010) Packaged-model typical
CL/F, female 0.168 L/h 0.168 L/h (= exp(lcl))
CL/F, male 0.110 L/h 0.110 L/h (= exp(lcl) x 0.655)
V/F 15.5 L 15.5 L (= exp(lvc))
Ka, single dose 0.0605 1/h 0.0605 1/h (= exp(lka))
Tlag 1.03 h 1.03 h (= exp(ltlag))
Effective F, multi-dose 0.674 x single-dose 0.674 (= e_multi_dose_pt_f)
Terminal half-life, female 64.0 h (= ln(2) x V/F / CL/F at female-typical) 64.0 h
Terminal half-life, male 97.6 h (= ln(2) x V/F / CL/F at male-typical) 97.6 h

The published etanercept terminal half-life from American studies cited by Fang 2010 (References 16, 17, 20) is approximately 70-100 h, which brackets the female-and-male typical half-lives implied by Table 3 and reproduced exactly by the packaged model.

Assumptions and deviations

  • Bioavailability F fixed at 1 for the single-dose reference. Fang 2010 reports apparent CL/F and V/F (the subcutaneous absolute bioavailability is not identifiable from SC-only data), and uses the paper’s F symbol as the typical-value bioavailability at the single-dose reference. The packaged model fixes lfdepot = log(1) for the single-dose reference and applies the published 0.674 multi-dose ratio as a multiplicative effect via e_multi_dose_pt_f^MULTI_DOSE_PT, exactly reproducing the paper’s F * theta_M^M final-model equation.

  • Canonical SEXF orientation inverted from the paper’s Gender. Fang 2010 encodes sex as a male-indicator (Gender = 1 male, 0 female) and reports CL/F = 0.168 L/h as the female-typical value with the male-vs-female ratio 0.655. The packaged model uses the canonical SEXF (1 = female, 0 = male) convention and applies the effect as e_sexf_cl^(1 - SEXF). This is mathematically identical to the paper at every (Gender, SEXF) pairing: SEXF = 1 (female) yields factor 1 (CL/F = 0.168); SEXF = 0 (male) yields factor 0.655 (CL/F = 0.110). The same canonical-rotation pattern is used in Bajaj_2017_nivolumab.R.

  • MULTI_DOSE_PT canonical reuses Goel 2016’s specific-scope indicator. The canonical MULTI_DOSE_PT covariate was originally founded by Goel_2016_Sonidegib.R as a per-dose-record indicator for the multiple-dose-phase-in-patients effect on apparent bioavailability. Fang 2010’s M covariate is the same concept (multiple-dose phase in patients vs. single-dose healthy-volunteer reference) and is reused under the canonical name. In Fang 2010 the indicator is subject-level (no subject crosses cohorts), so a downstream user simulating an AS-patient subject sets MULTI_DOSE_PT = 1 for all dose records of that subject and MULTI_DOSE_PT = 0 for all dose records of a healthy-volunteer subject.

  • No formal NCA reported in the paper. Fang 2010 Table 3 reports the population-PK parameter estimates and Figure 4 visualises observed-vs-population-predicted concentrations by dose group, but the paper does not tabulate per-subject NCA (Cmax, Tmax, AUC). The PKNCA validation section above quantifies the secondary exposures implied by the published structural model and the IIV; cross-validation against the Discussion’s qualitative statement “Chinese subjects have CL/F similar to American subjects (0.110-0.168 L/h vs 0.132 L/h pooled American)” is implicit in the half-life and AUC summaries.

  • Multi-dose cohorts not summarised by PKNCA. The 480 h sampling window (= 20 days) covers fewer than 3 BIW dosing intervals and approximately 3 QW dosing intervals. NCA on the partial multi-dose data would be dominated by the early absorption / accumulation phase rather than steady-state, so single-dose NCA (which the paper’s sampling fully supports) is reported instead.

  • No race or ethnicity covariate. Fang 2010 enrolled only Chinese subjects, so no race covariate could be tested. The Discussion notes that Lee et al. and others have reported non-Caucasian-vs-Caucasian CL/F differences for rhTNFR-Fc but those analyses were not powered. Users simulating non-Chinese populations should treat the packaged model as a Chinese-population reference and may need to apply external scaling factors derived from the cited American etanercept popPK analyses (Fang 2010 References 16, 17, 18, 20).

  • No allometric body-weight scaling. Fang 2010 did not retain weight as a covariate after stepwise selection (Table 2 backward elimination). The packaged model preserves this – CL/F and V/F are not scaled by body weight. Users wishing to extrapolate to a wider body-weight range than the Fang 2010 cohort (~50-80 kg, BMI 19-24) should rely on an etanercept popPK analysis that retained weight as a covariate.

  • Inter-individual variability translated from CV% to log-normal variance. Fang 2010 Table 3 reports IIV as percent coefficient of variation; the packaged ini() translates via omega^2 = log(CV^2 + 1) (the standard log-normal IIV interpretation documented in the NONMEM convention). The Tlag IIV (81.8% CV, omega^2 = 0.512) is large with a wide SE (74.4% per Table 3); users doing trial simulations should expect a wide spread of individual absorption-onset times.

Provenance summary

Source on disk Used for
Fang_2010_Population_pharmacokinetics_of_rhTNFR_Fc_3104b3.pdf Main paper Methods, Results, Tables 1-3, Figures 1-4, Discussion narrative on Chinese-vs-American CL/F comparison.
Fang_2010_Population_pharmacokinetics_of_rhTNFR_Fc_3104b3_trimmed.md Same content, trimmed-markdown form used during extraction for searchability.

No supplements, NONMEM control streams, regulatory reviews, or errata were available on disk for this extraction; all parameter values used by the packaged model are present in Table 3 of the main paper. No author correspondence was required.