Skip to contents

Scopolamine (Alvarez-Jimenez 2016)

Source: vignettes/articles/AlvarezJimenez_2016_scopolamine.Rmd
AlvarezJimenez_2016_scopolamine.Rmd

Model and source

  • Citation: Alvarez-Jimenez R, Groeneveld GJ, van Gerven JMA, Goulooze SC, Baakman AC, Hay JL, Stevens J. Model-based exposure-response analysis to quantify age related differences in the response to scopolamine in healthy subjects. Br J Clin Pharmacol. 2016;82(4):1011-1021.
  • Article: https://doi.org/10.1111/bcp.13031
  • Free full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC5137826/
  • Supplement (equations A.1-B.4): retrieved from PMC (BCP-82-1011-s001.docx).

Population

The model was developed from 135 healthy volunteers pooled across four clinical studies conducted at the Centre for Human Drug Research (Leiden, the Netherlands). The pooled cohort included 122 male and 13 female subjects (9.6% female). Ages ranged from 18 to 78 years (median 39). Body-weight mean was 78.6 kg (SD 9.08). Older subjects (>= 65 years, n = 36) received an intravenous dose of 0.3 mg scopolamine hydrobromide over 15 minutes; younger subjects received 0.5 mg over 15 minutes. See Table 1 of the source paper for the by-study demographics.

The metadata are also available programmatically:

mod_fn <- readModelDb("AlvarezJimenez_2016_scopolamine")
mod_fn()$population
#> $species
#> [1] "human"
#> 
#> $n_subjects
#> [1] 135
#> 
#> $n_studies
#> [1] 4
#> 
#> $age_range
#> [1] "18-78 years"
#> 
#> $age_median
#> [1] "39 years"
#> 
#> $weight_range
#> [1] "not reported explicitly; cohort mean 78.6 kg (SD 9.08)"
#> 
#> $weight_median
#> [1] "78.6 kg"
#> 
#> $sex_female_pct
#> [1] 9.6
#> 
#> $race_ethnicity
#> [1] "not reported"
#> 
#> $disease_state
#> [1] "healthy volunteers"
#> 
#> $dose_range
#> [1] "0.3 mg (subjects >=65 years) or 0.5 mg (subjects <65 years) intravenous scopolamine hydrobromide infused over 15 min"
#> 
#> $regions
#> [1] "Netherlands (Centre for Human Drug Research, Leiden)"
#> 
#> $notes
#> [1] "Pooled analysis of four clinical studies. Three earlier studies enrolled only male subjects (drug teratogenicity constraints for parallel study arms). Study 4 (older-subject cohort, 65-78 y, n=36) enrolled 13 females (36% female in that cohort). See Table 1 of the source paper. Older subjects (>=65 y) received the 0.3 mg dose; younger subjects the 0.5 mg dose. Body weight and age summarised in Table 1."

Source trace

Every parameter carries an in-file comment pointing at Table 2 or a supplemental equation of the paper. The block below summarises where each equation and each fixed-effect value came from.

Element Value Source
Structural PK 2-cmt linear, IV infusion Supplemental Eq A.1-A.5
CL alpha 1.09 L/min Table 2 CL[alpha]
Vc 2.66 L Table 2 Vc
Vp beta 62.10 L Table 2 Vp[beta]
Q 1.01 L/min Table 2 Q
CAC (age exp on CL, ref 28 y) -0.12 Table 2 CAC; Supplemental Eq A.1
CWC (weight exp on CL, ref 78.5 kg) 0.56 Table 2 CWC; Supplemental Eq A.1
CWV (weight exp on Vp, ref 78.5 kg) 0.38 Table 2 CWV; Supplemental Eq A.2
PK residual (proportional sigma^2) 0.045 Table 2 PK Error(sigma^2)
PD endpoint parameters see model file ini() block Table 2 PD rows; Supplemental Eq B.1-B.4
Age scale CAE on saccadic peak velocity EC50 34.0 y Table 2 CAE; Supplemental Eq B.1
Age scale CAB on kIN (0-back RT, delta EEG, theta EEG Fz-Cz) 229, 161, 210, 142 y Table 2 CAB; Supplemental Eq B.2
Adaptive-tracker Hill gamma 1.10 Table 2 gamma
Adaptive-tracker EMAX (fixed) 1 Table 2 EMAX = 1 (fixed per paper text)
N-back 0-back EMAX (fixed) 1 Table 2 (fixed per paper text)
N-back baseline log-odds E0 3.77 / 3.33 / 2.84 Table 2 E0 (0-back / 1-back / 2-back)
Indirect-response equation (drug effect) Emax * Cc^gamma / (EC50^gamma + Cc^gamma) Supplemental Eq B.4

Virtual cohort

Original individual data are not publicly available. We simulate two typical-value cohorts mirroring the paper’s dose stratification: a “young” arm at 0.5 mg over 15-min IV with age 30 and weight 78.6 kg, and an “older” arm at 0.3 mg over 15-min IV with age 70 and weight 77.8 kg. id_offset keeps the subject IDs disjoint across the two cohorts.

set.seed(2016)

n_per_arm <- 50L

make_cohort <- function(n, amt_mg, age, wt, cohort, id_offset = 0L) {
  ids <- id_offset + seq_len(n)
  # Dose row targets the PK ODE state 'central' (linCmt-managed).
  # Observation rows target the PK observable 'Cc'; because this
  # multi-output model has 14 declared observables spanning ODE
  # states plus algebraic outputs (Cc, nb0, nb1, nb2), using an
  # observable name on obs rows is what rxode2's DV mapping accepts
  # here. Every declared observable (Cc, nbRT, sacInacc, ..., nb2)
  # is returned regardless of which cmt an individual row points at.
  obs_times <- c(0, 2, 5, 10, 15, 20, 30, 45, 60, 90, 120, 180, 240, 360, 480)
  dosing <- tibble(id = ids, time = 0, amt = amt_mg, evid = 1L,
                   cmt = "central", dur = 15,
                   AGE = age, WT = wt, cohort = cohort)
  obs <- tidyr::crossing(id = ids, time = obs_times) |>
    dplyr::mutate(amt = 0, evid = 0L, cmt = "Cc", dur = 0,
                  AGE = age, WT = wt, cohort = cohort)
  dplyr::bind_rows(dosing, obs) |>
    dplyr::arrange(id, time, dplyr::desc(evid))
}

events <- dplyr::bind_rows(
  make_cohort(n_per_arm, amt_mg = 0.5, age = 30, wt = 78.6, cohort = "Young 0.5 mg",  id_offset =           0L),
  make_cohort(n_per_arm, amt_mg = 0.3, age = 70, wt = 77.8, cohort = "Older 0.3 mg",  id_offset = n_per_arm + 0L)
)

Simulation

Typical-value simulation (all etas zeroed) reproduces the median trajectories in Figure 1 of the paper. Stochastic simulation is available with the packaged random effects if a VPC is required, but the paper reports only median-line traces so we use the deterministic form here.

mod <- readModelDb("AlvarezJimenez_2016_scopolamine")
mod_tv <- rxode2::zeroRe(mod)
#> ℹ parameter labels from comments will be replaced by 'label()'
sim <- rxode2::rxSolve(mod_tv, events = events, keep = c("cohort"))
#> ℹ omega/sigma items treated as zero: 'etae0_nb2', 'etalemax_nb2', 'etae0_nb0', 'etalemax_nb0', 'etalkin_eegDfc', 'etalemax_eegDfc', 'etalec50_eegDfc', 'etalemax_sacInacc', 'etalec50_sacInacc', 'etalemax_nb1', 'etalkin_nb1', 'etalkin_eegTpo', 'etalec50_eegTpo', 'etalkin_eegTfc', 'etalec50_eegTfc', 'etalkin_eegDpo', 'etalemax_eegDpo', 'etalec50_eegDpo', 'etalkin_eegApo', 'etalemax_eegApo', 'etalec50_eegApo', 'etalkin_eegAfc', 'etalemax_eegAfc', 'etalec50_eegAfc', 'etalkin_adTrack', 'etalec50_adTrack', 'etalkin_sacPV', 'etalec50_sacPV', 'etalkin_sacInacc', 'etalkin_nbRT', 'etalec50_nbRT', 'etalvp', 'etalvc', 'etalcl'
#> Warning: multi-subject simulation without without 'omega'

Replicate published figures

Figures 1A-1B of the paper show plasma scopolamine concentrations for the 0.3 mg (older) and 0.5 mg (younger) dose groups. Figures 1D-1F, 1G-1I, and 1J-1L display adaptive-tracker, saccadic peak velocity, and 0-back RT median trajectories, respectively.

sim |>
  dplyr::filter(time > 0, time <= 480) |>
  ggplot(aes(time, Cc, colour = cohort)) +
  stat_summary(fun = median, geom = "line", linewidth = 1) +
  scale_y_log10() +
  labs(x = "Time (min)", y = "Plasma scopolamine (pg/mL)",
       colour = NULL,
       title = "Figure 1A-1B replication - plasma scopolamine",
       caption = "Replicates Figure 1A-1B of Alvarez-Jimenez 2016.")

sim |>
  dplyr::filter(time > 0, time <= 480) |>
  dplyr::select(time, cohort, nbRT, sacPV, adTrack, eegAfc, eegDfc, eegTfc, nb0) |>
  tidyr::pivot_longer(-c(time, cohort), names_to = "endpoint") |>
  dplyr::group_by(time, cohort, endpoint) |>
  dplyr::summarise(median = median(value), .groups = "drop") |>
  ggplot(aes(time, median, colour = cohort)) +
  geom_line(linewidth = 1) +
  facet_wrap(~ endpoint, scales = "free_y") +
  labs(x = "Time (min)", y = "Response (paper units)",
       colour = NULL,
       title = "PD endpoint median trajectories",
       caption = "Selected panels from Figure 1 of Alvarez-Jimenez 2016.")

PKNCA validation

The paper does not tabulate NCA parameters; the PK model was published with population parameter estimates only. We compute NCA on the typical-value simulation for orientation.

sim_nca <- sim |>
  dplyr::filter(!is.na(Cc)) |>
  dplyr::select(id, time, Cc, cohort)

# Guarantee a time = 0 row per subject (pre-dose Cc = 0 for IV infusion).
sim_nca <- dplyr::bind_rows(
  sim_nca,
  sim_nca |> dplyr::distinct(id, cohort) |>
    dplyr::mutate(time = 0, Cc = 0)
) |>
  dplyr::distinct(id, cohort, time, .keep_all = TRUE) |>
  dplyr::arrange(id, cohort, time)

conc_obj <- PKNCA::PKNCAconc(sim_nca, Cc ~ time | cohort + id)

dose_df <- events |>
  dplyr::filter(evid == 1L) |>
  dplyr::select(id, time, amt, cohort)

dose_obj <- PKNCA::PKNCAdose(dose_df, amt ~ time | cohort + id)

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)

nca_summary <- as.data.frame(nca_res$result) |>
  dplyr::filter(PPTESTCD %in% c("cmax", "tmax", "aucinf.obs", "half.life")) |>
  dplyr::group_by(cohort, PPTESTCD) |>
  dplyr::summarise(median = median(PPORRES, na.rm = TRUE), .groups = "drop") |>
  tidyr::pivot_wider(names_from = PPTESTCD, values_from = median)

knitr::kable(nca_summary, digits = 2,
             caption = "Simulation-derived NCA parameters by dose cohort.")
Simulation-derived NCA parameters by dose cohort.
cohort aucinf.obs cmax half.life tmax
Older 0.3 mg 4187497 39327.71 87.54 10
Young 0.5 mg 3950673 43786.61 83.21 10

Assumptions and deviations

  • Vc = 2.66 L is smaller than the ~30 L reported in earlier scopolamine PK studies (Renner 2005, Ebert 2001). The paper reports IIV(Vc) = 74% CV with 49% shrinkage, indicating a poorly identified central volume. Peak Cc predictions from this model are therefore ~10x higher than most other scopolamine PK models in the literature. Users comparing to external NCA should be aware of this and treat Cc values as internally-consistent with the paper’s PD-model EC50s (which were fit against the same Cc predictions) rather than as free-standing plasma-concentration estimates.
  • Saccadic peak velocity residual variance labelled (prop) = 768.0 in Table 2 is treated here as additive (addSd = sqrt(768) ~ 27.7 deg/sec). A proportional residual of that magnitude would give CV ~ 2770%, which is physically implausible; interpreted as additive it corresponds to ~6% of the ~470 deg/sec typical baseline, consistent with the other endpoints’ relative residual magnitudes. This is the single documented deviation from the paper’s Table 2 residual-error labeling.
  • The 0-back N-back model in the paper uses four buffer / transit compartments to introduce a delay between drug effect and the log-odds observed correct-answer response. Reproducing the exact 4-transit chain requires the paper’s NONMEM control stream (not on disk). The 0-back logit response here is modelled as a single indirect-response compartment (red_nb0) driven by drug effect with rate kin_nb0 and elimination kout_nb0; the residual variance and IIV blocks match Table 2. The four-buffer transit chain is approximated by a single first-order recovery, which under-represents the sharp onset and delayed peak of the paper’s Figure 4A.
  • 1-back and 2-back N-back models are similarly implemented as direct-effect on log-odds baseline via a first-order reduction compartment. Table 2 reports kIN and kOUT values for these endpoints, but their combination (kIN = 1.10, kOUT = 221 for 1-back) is atypical for a classical indirect-response model and the paper does not describe the ODE structure explicitly. The values are encoded verbatim; users interested in exact reproduction of the 1-back / 2-back trajectories should consult the source paper’s Figure 4B-4C.
  • OMEGA correlation entries (OC in Table 2) are interpreted as raw covariances rather than correlation coefficients. This is because one reported OC value (1.410 for EEG delta Fz-Cz EC50 vs. EMAX) exceeds unity and cannot be a correlation. Covariance is the natural NONMEM OMEGA(i,j) matrix element. Where variances and covariance are consistent (positive-semidefinite blocks) they are encoded as ~ c(var1, cov, var2); where the 3-parameter block for EEG delta Fz-Cz includes an unreported cross-term (EC50-vs-kIN), that off-diagonal is fixed to zero.
  • CV-to-variance conversion for log-normal IIV uses variance = log(1 + CV^2).
  • Race and sex distributions are not encoded as covariates because the paper reports no race data and pooled 122 male + 13 female subjects (only Study 4 enrolled females); no sex effect was reported. Simulations here treat subjects as a homogenous population.

Errata

  • The paper’s Table 2 caption reads “CWC: coefficient relating age and weight” which is a typo – CWC is the coefficient relating weight and clearance per Supplemental Eq A.1. The model file uses the equation-defined interpretation (weight on CL).
  • The reported OC value 1.410 for EEG delta Fz-Cz (EC50 vs. EMAX) exceeds unity and is therefore not a Pearson correlation; interpreted as covariance the corresponding correlation is ~0.82.
  • No corrigenda or errata for the paper are indexed by the journal at the time of this extraction (2026-07-09).