Exendin-(9-39) (Ng 2018) • nlmixr2lib

Model and source

Citation: Ng CM, Tang F, Seeholzer SH, Zou Y, De Leon DD. Population pharmacokinetics of exendin-(9-39) and clinical dose selection in patients with congenital hyperinsulinism. Br J Clin Pharmacol. 2018;84(3):520-528. doi:10.1111/bcp.13463
Description: Two-compartment intravenous-infusion population PK model for exendin-(9-39) in patients with congenital hyperinsulinism (Ng 2018). Pooled paediatric (neonates and children) and adult cohort with allometric scaling fixed at 0.75 on CL and Q and 1.0 on Vc and Vp (reference WT 70 kg); inter-individual variability retained only on CL. Residual variability follows the NONMEM Poisson error model (Var(Y|F) = F * sigma^2), encoded as a power-error with fixed exponent 0.5.
Article: Br J Clin Pharmacol 2018;84(3):520-528

Population

The model was developed from 182 plasma exendin-(9-39) concentrations in 26 patients with congenital hyperinsulinism (KATP-HI) pooled across three open-label crossover pilot trials at The Children’s Hospital of Philadelphia (Ng 2018 Table 1):

Adult Study (NCT00571324): 9 older adolescents / adults, median age 19 years (range 15-47), median weight 69.1 kg (range 58.2-130).
Children Study (NCT00897676): 10 children, median age 5 years (range 2-15), median weight 21.0 kg (range 11.8-69.2).
Neonate Study (NCT00835328): 7 neonates, median age 0.11 years (range 0.06-0.41), median weight 6.25 kg (range 3.92-6.60).

The pooled cohort was 14/26 (53.8%) female and 24/26 Caucasian (2 of unknown race). Subjects received intravenous infusions of exendin-(9-39) at 0.02-0.1 mg/kg/h over 2 h per dose level (Adult and Children studies; cumulative 6 h across three rate steps) or 6-12 h single-rate infusions (Neonate Study). Bioanalysis was LC-MS/MS with a calibration range of 10-1390 ng/mL; 16/182 samples (8.8%) were below the LLOQ and were retained via the Beal M3 method during the original NONMEM 7.3 fit. The same metadata are available programmatically via readModelDb("Ng_2018_exendin939")$population.

Source trace

Every numeric value in ini() carries an in-file comment pointing to the Ng 2018 source location. The table below collects them in one place for review.

Equation / parameter	Value	Source location
`lcl` (CL_TV)	11.6 L/h	Table 2, row “CL TV”
`lvc` (V1_TV)	9.59 L	Table 2, row “V1 TV”
`lq` (Q_TV)	2.20 L/h	Table 2, row “Q TV”
`lvp` (V2_TV)	8.89 L	Table 2, row “V2 TV”
`e_wt_cl_q`	0.75 (fixed)	Table 2, row “WT effect on CL and Q (fixed)”
`e_wt_vc_vp`	1.00 (fixed)	Table 2, row “WT effect on V1 and V2 (fixed)”
`etalcl` (omega^2_CL)	0.0572	Table 2, row “omega^2 CL”
`powSd` (sqrt(Sigma_pois))	sqrt(3.66) = 1.913	Table 2, row “Sigma poisson” (variance scale)
Reference weight (70 kg)	70 kg	Methods (Equation 3 footnote: “the standard adult weight of 70 kg was used as WTref”)
Allometric scaling form	(WT / 70)^exponent	Methods (Equation 3)
Two-compartment IV PK	ADVAN3 TRANS4	Results para. “PopPK model”
Poisson residual error	W = sqrt(F); Y = F + W * EPS	Methods para. “Residual error was initially modelled…” + Results para. “the Poisson error model was the most stable”
Final covariate set	Allometric WT only	Results para. “After the effects of weight on PK parameters were accounted for using allometric scaling, none of the other covariates tested (age, CrCL, gender and patient groups) had significant effects on CL”

Virtual cohort

Original observed data are not publicly available. The cohorts below mirror the three age strata described in Ng 2018 Table 1, plus the neonatal dose-selection simulation in the Results subsection “Dose determination in neonatal clinical studies”. Each cohort contains n_sub virtual subjects so the stochastic ribbons reflect the inter-individual variability under the published omega^2_CL.

set.seed(20260530)

n_sub <- 50L

# Helper: build one cohort of `n_sub` subjects all receiving the same
# IV infusion regimen. `id_offset` keeps IDs disjoint across cohorts.
build_cohort <- function(label, weight_kg, rate_mgkgh, infusion_h,
                         n_doses = 1L, dose_interval_h = NA_real_,
                         id_offset = 0L,
                         obs_grid = NULL) {
  ids <- id_offset + seq_len(n_sub)
  amt_per_dose <- rate_mgkgh * weight_kg * infusion_h          # mg
  infusion_rate <- amt_per_dose / infusion_h                   # mg/h
  if (n_doses == 1L) {
    dose_times <- 0
  } else {
    dose_times <- seq(0, by = dose_interval_h, length.out = n_doses)
  }

  dose_rows <- tidyr::expand_grid(id = ids, time = dose_times) |>
    mutate(
      evid   = 1L,
      amt    = amt_per_dose,
      cmt    = "central",
      rate   = infusion_rate,
      cohort = label,
      WT     = weight_kg
    )

  if (is.null(obs_grid)) {
    last_dose <- max(dose_times)
    obs_grid <- sort(unique(c(
      seq(0, infusion_h, by = infusion_h / 12),
      seq(infusion_h, last_dose + infusion_h, by = 0.25),
      seq(last_dose + infusion_h, last_dose + infusion_h + 24, by = 0.5)
    )))
  }

  obs_rows <- tidyr::expand_grid(id = ids, time = obs_grid) |>
    mutate(
      evid   = 0L,
      amt    = 0,
      cmt    = NA_character_,
      rate   = 0,
      cohort = label,
      WT     = weight_kg
    )

  bind_rows(dose_rows, obs_rows) |> arrange(id, time, desc(evid))
}

# Three age strata at their median weights (Ng 2018 Table 1):
#   adult     median 69.1 kg, representative adult-study regimen
#   children  median 21.0 kg, representative children-study regimen
#   neonate   median  6.25 kg, representative neonate-study regimen
#
# Plus the 3.7 kg neonatal dose-selection simulation cohorts at
# 0.02, 0.1, 1.0, and 6.0 mg/kg/h over 6 h (subset of the paper's
# 14-rate grid; chosen to span 3 orders of magnitude).
events <- bind_rows(
  build_cohort("adult_0.1_2h",      69.1, rate_mgkgh = 0.1,  infusion_h = 2,
               id_offset =     0L),
  build_cohort("child_0.1_2h",      21.0, rate_mgkgh = 0.1,  infusion_h = 2,
               id_offset =  1000L),
  build_cohort("neonate_0.04_12h",   6.25, rate_mgkgh = 0.04, infusion_h = 12,
               id_offset =  2000L),
  build_cohort("neo_3p7_0.02_6h",    3.7, rate_mgkgh = 0.02, infusion_h = 6,
               id_offset =  3000L),
  build_cohort("neo_3p7_0.1_6h",     3.7, rate_mgkgh = 0.1,  infusion_h = 6,
               id_offset =  4000L),
  build_cohort("neo_3p7_1.0_6h",     3.7, rate_mgkgh = 1.0,  infusion_h = 6,
               id_offset =  5000L),
  build_cohort("neo_3p7_6.0_6h",     3.7, rate_mgkgh = 6.0,  infusion_h = 6,
               id_offset =  6000L)
)

stopifnot(!anyDuplicated(unique(events[, c("id", "time", "evid")])))

Simulation

mod <- readModelDb("Ng_2018_exendin939")

sim <- rxode2::rxSolve(mod, events = events, keep = c("cohort", "WT")) |>
  as.data.frame()
#> ℹ parameter labels from comments will be replaced by 'label()'

For deterministic typical-value replications (paper Table 2 analytical CL and t1/2 by weight stratum), simulate with random effects zeroed:

mod_typical <- mod |> rxode2::zeroRe()
#> ℹ parameter labels from comments will be replaced by 'label()'
sim_typical <- rxode2::rxSolve(mod_typical, events = events,
                               keep = c("cohort", "WT")) |>
  as.data.frame()
#> ℹ omega/sigma items treated as zero: 'etalcl'
#> Warning: multi-subject simulation without without 'omega'

Replicate published figures

Three-cohort concentration-time profiles

Ng 2018 does not include a stratified observed-vs-predicted figure in the main text (Figure 1 is a pooled diagnostic plot). The plot below shows the typical-value simulation for each of the three age strata under a representative regimen from the corresponding clinical study; it lets a reader confirm the model’s expected concentration ranges before turning to the dose-selection NCA below.

sim_typical |>
  filter(cohort %in% c("adult_0.1_2h", "child_0.1_2h", "neonate_0.04_12h")) |>
  filter(time <= 30) |>
  mutate(stratum = factor(cohort,
                          levels = c("adult_0.1_2h", "child_0.1_2h",
                                     "neonate_0.04_12h"),
                          labels = c("Adult 69.1 kg, 0.1 mg/kg/h x 2 h",
                                     "Child 21.0 kg, 0.1 mg/kg/h x 2 h",
                                     "Neonate 6.25 kg, 0.04 mg/kg/h x 12 h"))) |>
  ggplot(aes(time, Cc, colour = stratum)) +
  geom_line() +
  scale_y_log10() +
  labs(x = "Time (h)", y = "Exendin-(9-39) (ng/mL)", colour = NULL,
       title = "Typical-value profiles across the three Ng 2018 study strata") +
  theme(legend.position = "bottom")
#> Warning in scale_y_log10(): log-10 transformation introduced infinite values.

Figure 3 - neonatal AUC0-inf and Cmax across 6 h infusion rates

Ng 2018 Figure 3 plots simulated AUC0-inf and Cmax for 14 IV infusion rates (0.02-6.0 mg/kg/h) in 3.7-kg neonates over 6 and 9 h durations. The replication below shows the four 6-h-infusion rates that span three orders of magnitude (0.02, 0.1, 1.0, 6.0 mg/kg/h); each ribbon reflects the inter-individual variability in CL.

sim |>
  filter(cohort %in% c("neo_3p7_0.02_6h", "neo_3p7_0.1_6h",
                       "neo_3p7_1.0_6h",  "neo_3p7_6.0_6h")) |>
  filter(time <= 30) |>
  group_by(cohort, time) |>
  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"
  ) |>
  mutate(rate_label = factor(cohort,
                             levels = c("neo_3p7_0.02_6h", "neo_3p7_0.1_6h",
                                        "neo_3p7_1.0_6h",  "neo_3p7_6.0_6h"),
                             labels = c("0.02 mg/kg/h",  "0.1 mg/kg/h",
                                        "1.0 mg/kg/h",   "6.0 mg/kg/h"))) |>
  ggplot(aes(time, Q50)) +
  geom_ribbon(aes(ymin = Q05, ymax = Q95), alpha = 0.25) +
  geom_line() +
  facet_wrap(~ rate_label, scales = "free_y") +
  labs(x = "Time (h)", y = "Exendin-(9-39) (ng/mL)",
       title = "Neonatal simulation, 3.7-kg subject, 6-h IV infusion",
       caption = "Replicates Ng 2018 Figure 3 (panels A and B): 6-h IV infusion in neonates with HI.")

PKNCA validation

The block below computes AUC0-inf and Cmax for the four neonatal infusion rates above and compares the simulated medians against the values that Ng 2018 reports verbatim in the Results subsection “Dose determination in neonatal clinical studies”.

nca_cohorts <- c("neo_3p7_0.02_6h", "neo_3p7_0.1_6h",
                 "neo_3p7_1.0_6h",  "neo_3p7_6.0_6h")

sim_nca <- sim |>
  filter(cohort %in% nca_cohorts, !is.na(Cc), time > 0) |>
  select(id, time, Cc, cohort)

dose_df <- events |>
  filter(cohort %in% nca_cohorts, evid == 1) |>
  select(id, time, amt, cohort)

conc_obj <- PKNCA::PKNCAconc(sim_nca, Cc ~ time | cohort + id,
                             concu = "ng/mL", timeu = "hr")
dose_obj <- PKNCA::PKNCAdose(dose_df, amt ~ time | cohort + id,
                             doseu = "mg")

intervals <- data.frame(
  start      = 0,
  end        = Inf,
  cmax       = TRUE,
  tmax       = TRUE,
  aucinf.obs = 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 (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
#> Requesting an AUC range starting (0) before the first measurement (0.5) is not allowed
nca_summary <- summary(nca_res)
knitr::kable(nca_summary,
             caption = "Simulated NCA parameters by neonatal-dose cohort (3.7-kg subject, 6-h IV infusion).")

Simulated NCA parameters by neonatal-dose cohort (3.7-kg subject, 6-h IV infusion).
Interval End	cohort	N	Cmax (ng/mL)	Tmax (hr)	Half-life (hr)	AUCinf,obs (hr*ng/mL)
Inf	neo_3p7_0.02_6h	50	56.0 [23.5]	6.00 [6.00, 6.00]	1.65 [0.0907]	NC
Inf	neo_3p7_0.1_6h	50	293 [20.3]	6.00 [6.00, 6.00]	1.67 [0.0842]	NC
Inf	neo_3p7_1.0_6h	50	2970 [27.4]	6.00 [6.00, 6.00]	1.68 [0.108]	NC
Inf	neo_3p7_6.0_6h	50	16800 [19.2]	6.00 [6.00, 6.00]	1.65 [0.0761]	NC

Comparison against published values

Ng 2018 reports the following median AUC0-inf and Cmax for 3.7-kg neonates receiving 6-h IV infusions (Results “Dose determination in neonatal clinical studies”):

published_neo <- tibble::tribble(
  ~cohort,             ~rate_mgkgh, ~AUCinf_published_ngh_per_mL, ~Cmax_published_ng_per_mL,
  "neo_3p7_0.02_6h",  0.02,        343,                          15.2,
  "neo_3p7_0.1_6h",   0.10,        1710,                         75.9,
  "neo_3p7_1.0_6h",   1.00,        17100,                        759,
  "neo_3p7_6.0_6h",   6.00,        103000,                       4560
)

sim_median <- sim |>
  filter(cohort %in% nca_cohorts, !is.na(Cc), time > 0) |>
  group_by(cohort, id) |>
  summarise(Cmax_sim = max(Cc, na.rm = TRUE), .groups = "drop_last") |>
  summarise(Cmax_sim_median_ng_per_mL = median(Cmax_sim, na.rm = TRUE),
            .groups = "drop")

dose_lookup <- events |>
  filter(cohort %in% nca_cohorts, evid == 1) |>
  group_by(cohort, id) |>
  summarise(amt = sum(amt), .groups = "drop_last") |>
  summarise(amt_mg = first(amt), .groups = "drop")

auc_sim <- sim |>
  filter(cohort %in% nca_cohorts, !is.na(Cc), time > 0) |>
  group_by(cohort, id) |>
  arrange(time, .by_group = TRUE) |>
  summarise(
    AUC_trap = sum((dplyr::lead(time) - time) * (Cc + dplyr::lead(Cc)) / 2,
                   na.rm = TRUE),
    .groups = "drop_last"
  ) |>
  summarise(AUC_sim_median_ngh_per_mL = median(AUC_trap, na.rm = TRUE),
            .groups = "drop")

comparison <- published_neo |>
  left_join(sim_median, by = "cohort") |>
  left_join(auc_sim,    by = "cohort") |>
  mutate(
    AUC_pct_diff  = 100 * (AUC_sim_median_ngh_per_mL - AUCinf_published_ngh_per_mL) /
                          AUCinf_published_ngh_per_mL,
    Cmax_pct_diff = 100 * (Cmax_sim_median_ng_per_mL - Cmax_published_ng_per_mL) /
                          Cmax_published_ng_per_mL
  ) |>
  select(cohort, rate_mgkgh,
         AUCinf_published_ngh_per_mL, AUC_sim_median_ngh_per_mL, AUC_pct_diff,
         Cmax_published_ng_per_mL,    Cmax_sim_median_ng_per_mL, Cmax_pct_diff)

knitr::kable(comparison, digits = c(0, 2, 0, 0, 1, 1, 1, 1),
             caption = "Median simulated AUC and Cmax vs Ng 2018 published values (3.7-kg neonates, 6-h IV infusion).")

Median simulated AUC and Cmax vs Ng 2018 published values (3.7-kg neonates, 6-h IV infusion).
cohort	rate_mgkgh	AUCinf_published_ngh_per_mL	AUC_sim_median_ngh_per_mL	AUC_pct_diff	Cmax_published_ng_per_mL	Cmax_sim_median_ng_per_mL	Cmax_pct_diff
neo_3p7_0.02_6h	0.02	343	324	-5.6	15.2	54.6	259.3
neo_3p7_0.1_6h	0.10	1710	1726	0.9	75.9	290.4	282.6
neo_3p7_1.0_6h	1.00	17100	17742	3.8	759.0	2981.9	292.9
neo_3p7_6.0_6h	6.00	103000	99738	-3.2	4560.0	16809.0	268.6

The simulated AUC0-inf values track the published values closely because, under linear PK, AUC0-inf = dose / CL is determined entirely by CL_TV and the allometric exponent. Differences in Cmax arise from the closed-form vs simulation grid for resolving the end-of-infusion peak; see Assumptions and deviations.

Analytical CL and terminal half-life by weight stratum

Ng 2018 Results para. “PopPK model” reports CL and terminal half-life for the median paediatric (15 kg) and median adult (69 kg) subjects. The values back-computed from Table 2 estimates reproduce those verbatim:

tibble::tribble(
  ~stratum,           ~WT,  ~CL_paper_L_per_h, ~t_half_paper_h,
  "Paediatric (med)", 15,   3.65,              2.34,
  "Adult (med)",      69,   11.5,              3.42
) |>
  mutate(
    CL_model_L_per_h    = 11.6 * (WT / 70)^0.75,
    Vc_model_L          = 9.59 * (WT / 70),
    Vp_model_L          = 8.89 * (WT / 70),
    Q_model_L_per_h     = 2.20 * (WT / 70)^0.75,
    k10                 = CL_model_L_per_h / Vc_model_L,
    k12                 = Q_model_L_per_h  / Vc_model_L,
    k21                 = Q_model_L_per_h  / Vp_model_L,
    sum_rates           = k10 + k12 + k21,
    prod_rates          = k10 * k21,
    beta                = (sum_rates - sqrt(sum_rates^2 - 4 * prod_rates)) / 2,
    t_half_model_h      = log(2) / beta
  ) |>
  select(stratum, WT, CL_paper_L_per_h, CL_model_L_per_h,
         t_half_paper_h, t_half_model_h) |>
  knitr::kable(digits = 2,
               caption = "Analytical CL and terminal half-life: paper-reported vs back-computed from Table 2.")

Analytical CL and terminal half-life: paper-reported vs back-computed from Table 2.
stratum	WT	CL_paper_L_per_h	CL_model_L_per_h	t_half_paper_h	t_half_model_h
Paediatric (med)	15	3.65	3.65	2.34	2.34
Adult (med)	69	11.50	11.48	3.42	3.43

Assumptions and deviations

Poisson residual error encoding. Ng 2018 chose a NONMEM Poisson error model (variance proportional to F: Var(Y|F) = F * Sigma^2) over proportional and combined-add-prop alternatives because it gave the lowest objective function. The published value Sigma_poisson = 3.66 (Table 2) is the EPS variance ($SIGMA in NONMEM reports variances). The packaged model encodes this as a power-error term Cc ~ pow(powSd, 0.5) with powSd = sqrt(3.66) = 1.913; this gives SD(Y|F) = powSd * sqrt(F), matching the original W = sqrt(F); Y = F + W * EPS specification. The new powSd canonical parameter name is registered in .claude/skills/extract-literature-model/references/parameter-names.md with this model as the founding example.
Inter-individual variability on CL only. Ng 2018 attempted IIV on V1, Q, and V2 but reported that those terms failed convergence or yielded %CV > 50% and were dropped from the final model (Results para. “PopPK model”). The packaged model carries etalcl ~ 0.0572 and no other random effects, matching the published structure verbatim. Downstream users who need V- or Q-variability for prospective simulations should not retro-fit those etas from this packaged structure alone; the source data cannot support them.
Allometric exponents fixed (not estimated). Both WT effects are marked fixed in Table 2. The packaged model wraps e_wt_cl_q and e_wt_vc_vp in fixed() so this provenance is preserved; the values 0.75 and 1.00 reflect the physiology-based defaults the authors chose, not empirical estimates from this dataset.
Below-LLOQ handling not propagated to simulation. The original fit used the Beal M3 maximum-likelihood approach for the 16/182 below-LLOQ samples; this affects parameter estimation but is not re-applied during forward simulation. Simulated concentrations below 10 ng/mL (LLOQ) may therefore appear in the output even though the original assay could not have quantified them.
No race / sex / age / CrCL covariates. Ng 2018 tested age, creatinine clearance, gender, and patient-group covariates after allometric scaling and reported none significant at the alpha = 0.01 threshold (objective-function difference threshold 6.64 for 1 d.f.). The packaged model therefore carries only WT as a covariate. The Discussion notes that the narrow CrCL range in the cohort (only one subject with mild renal impairment, none with moderate-to-severe) limits the conclusions about renal-function dependence; downstream users in renal-impaired populations should treat the model’s lack of a CrCL covariate as a known scope gap rather than an established absence of effect.
Cmax vs simulation-grid resolution. The published median Cmax values come from a 1000-replicate simulation in NONMEM with the exact end-of-infusion timepoint. The vignette’s simulation grid samples Cc on a 0.25-h grid after end-of-infusion (plus a finer grid during the infusion), so the simulated Cmax can lag the true end-of-infusion peak by up to one grid step. Expect simulated Cmax to fall within a few percent of the published values; AUC0-inf (computed via trapezoidal rule) is less sensitive and tracks the closed-form dose / CL relationship closely. Both columns are shown in the comparison table so a reader can see the gap.
Abstract vs Results internal inconsistency. The Ng 2018 abstract reports an adult-cohort terminal half-life of 1.81 h and CL of 11.8 L/h, while the Results section “PopPK model” reports 3.42 h and 11.5 L/h for the same median 69-kg adult. Back-computing from Table 2 estimates (CL_TV = 11.6, V1_TV = 9.59, Q_TV = 2.20, V2_TV = 8.89 at 70 kg) with allometric exponents 0.75 / 1.00 reproduces 11.5 L/h and 3.42 h verbatim, confirming that the Results-section values are the ones consistent with Table 2 and with the packaged model. The abstract values appear to be a paper-internal inconsistency (possibly a residue of an earlier parameterisation); they are not used in the validation table above.
Race / ethnicity scope. The cohort was 24/26 Caucasian; 2 subjects were of unknown race. Race was not tested as a covariate in the published analysis, so the packaged model has no race-effect parameters. Downstream users applying the model to non-Caucasian populations should note that this is a scope gap, not an established absence of effect.
Preclinical TK data not part of this model. Ng 2018 reports juvenile-rat (Sprague-Dawley) and beagle-dog toxicokinetic data used to derive NOAELs for human dose selection; those preclinical TK estimates are not part of the human PopPK model. They are documented in population$notes but are not encoded as a separate preclinical model file.