rNCA: Do NCA for review

In pkr: Pharmacokinetics in R

Do NCA for review

Description

This performs NCA from the CDISC EX and PC datasets.

Usage

rNCA(ex, pc, study = "", trt = "", id = "", analyte = "",
     codeBQL = c("< 0", "<0", "NQ", "BLQ", "BQL", "BQoL", "<LOQ"),
     fit="Linear", MinPoints = 5, excludeInfusion = TRUE)

Arguments

ex	EX domain data, usually from the loadEXPC
pc	PC domain data, usually form the loadEXPC
study	vector of study names in EX and PC domain to do NCA. "" means all.
trt	vector of treatment names in EXTRT to do NCA. "" means all. Placebo is always excluded, since a dose of zero has no pharmacokinetics to characterize.
id	vector of subject IDs in USUBJID to do NCA. "" means all.
analyte	vector of molecular species in PCTESTCD to do NCA. "" means all. Matching is case-insensitive.
codeBQL	symbols of below the quantitation limit
fit	either of "Linear" or "Log" to indicate the way to calculate AUC and AUMC
MinPoints	an analysis is done only when the number of quantifiable (greater than zero) concentrations in the dosing interval is strictly greater than this value. Note that this counts quantifiable concentrations, not sampling points: BQL and zero records do not count.
excludeInfusion	estimate the terminal slope of an infusion from post-infusion samples only; TRUE by default. Set FALSE for the unrestricted automatic search. See Details. No effect on a bolus or an extravascular dose.

Details

One row of result is produced for each subject, dosing interval and analyte. A dosing interval starts at its own EXSTDTC and ends at the subject's next real dose of any treatment in the same study, so that in a crossover trial the first period is not extended into the second; a placebo or zero dose delivers no drug and does not end an interval. The final interval is open-ended, so terminal and follow-up samples still contribute to the elimination phase.

The last concentration in the 24 hours before the dose is carried into the interval as the pre-dose sample, and placed at time zero. The lookback is bounded because an unbounded one would report a sample from an earlier period as this dose's initial concentration. In an interval that is followed by another dose the trailing trough is dropped when its nominal time (PCTPTNUM) equals the pre-dose one, so that it is not counted twice.

EX and PC records are sorted by date and time before use, so the result does not depend on the row order of the input. Records whose date and time cannot be read as a full YYYY-MM-DDTHH:MM:SS value are dropped with a warning naming the subjects affected.

The automatic terminal-slope search restricts itself to samples after the maximum concentration. That is the right rule for an extravascular dose but says nothing about an infusion, so a sample drawn while drug is still being infused can win on adjusted R-squared and be reported as terminal elimination. Because a terminal phase does not exist while drug is still going in, excludeInfusion defaults to TRUE: the search is given only the samples after the end of infusion, and the same maximum-adjusted-R-squared criterion then selects the window among them; within that window it still starts after the largest of those samples. Set excludeInfusion=FALSE for the unrestricted search over all samples.

Only lambda z and the parameters derived from it (half-life, AUCIFO, AUCPEO, Vz) are affected - Cmax, AUClast and the other observed parameters cannot change. When there are too few post-infusion samples to fit a slope, a warning names the profile and the unrestricted selection is used, so no profile is lost silently.

The restriction applies to this function and to NCA0, which know the dosing record and so the end of infusion. sNCA and tblNCA take times and concentrations without a dosing record and remain faithful wrappers around NonCompart; give sNCA an explicit UsePoints to fix the window there.

This calls NCA0. Results of this can be further processed by foreNCA to plot and compare between studies and dose groups.

Value

This returns a table of NCA results, or invisibly NULL with a warning when no subject meets the MinPoints criterion.

Author(s)

Kyun-Seop Bae <[e-mail]>

See Also

help, NCA0, loadEXPC, foreNCA

pkr documentation built on July 30, 2026, 5:08 p.m.