NCA computation rules

Computational rules for the NCA engine are discussed in sections:

Time deviations in steady-state data
Data checking and pre-treatment

NCA computations are covered under the headings:

Lambda Z or Slope Estimation settings
Therapeutic response
Partial area calculation
Sparse sampling calculation
Drug effect calculation
Weighting

Time deviations in steady-state data

When using steady-state data, Phoenix computes AUC_TAU from dose time to dose time+tau, based on the tau value set in the Dosing panel. However, in most studies, there are sampling time deviations. For example, if dose time=0 and tau=24, the last sample might be at 23.975 or 24.083 hours. In this instance, the program will estimate the AUC_TAU based on the estimated concentration at 24 hours, and not the concentration at the actual observation time. For steady state data, Cmax, Tmax, Cmin and Tmin are found using observations taken at or after the dose time, but no later than dose time+tau.

Data checking and pre-treatment

Prior to calculation of pharmacokinetic parameters, Phoenix institutes a number of data-checking and pre-treatment procedures as follows.

Sorting: Prior to analysis, data within each profile is sorted in ascending time order. (A profile is identified by a unique combination of sort variable values.)
Inserting initial time points: If a PK profile does not contain an observation at dose time, Phoenix inserts a value using the following rules. (Note that, if there is no dosing time specified, the dosing time is assumed to be zero.)

Extravascular and Infusion data: For single dose data, a concentration of zero is used; for steady-state, the minimum observed during the dose interval (from dose time to dose time +tau) is used.

IV Bolus data: Phoenix performs a log-linear regression of first two data points to back-extrapolate C0. If the regression yields a slope >= 0, or at least one of the first two y-values is zero, or if one or both of the points is viewed as an outlier and excluded from the Lambda Z regression, then the first observed y-value is used as an estimate for C0. If a weighting option is selected, it is used in the regression.

Urine data: A rate value of zero is used at dose time.

Drug Effect data: An effect value equal to the user-supplied baseline is inserted at dose time (assumed to be zero if none is supplied).

The inserted point at dose time is needed to compute the initial trapezoid from dose time to the first observation for the AUC final parameters and to compute partial areas or therapeutic response parameters that depend on times from the dose time to the first observation. The inserted point is never used in Lambda Z computations.

Only one observation per profile at each time point is permitted. If multiple observations at the same time point are detected, the analysis is halted with an error message. No output is generated.

Data exclusions: NCA automatically excludes unusable data points meeting the following criteria:

Missing values: For plasma models, if either the time or concentration value is missing or non-numeric, then the associated record is excluded from the analysis. For urine models, records with missing or non-numeric values for any of the following are excluded: collection interval start or stop time, drug concentration, or collection volume. Also for urine models, if the collection volume is zero, the record is excluded.

Data points preceding the dose time: If an observation time is earlier than the dosing time, then that observation is excluded from the analysis. For urine models, the lower time point is checked.

Data points for urine models: In addition to the above rules, if the lower time is greater than or equal to the upper time, or if the concentration or volume is negative, the data must be corrected before NCA can run.

Flagging data deficiencies in the Final Parameters output: After exclusion of unusable data points as described in the prior section, profiles that result in no non-missing observations, and profiles that result in only one non-missing observation where a valid point at dose time cannot be inserted, have a flag called “Flag_N_Samples” in the Final Parameters output. More specifically, these cases are:

•Profile with all missing observations (N_Samples=0) for any NCA Model Type. No Final Parameters can be reported.

•Urine model profile with all urine volumes equal to zero. This is equivalent to no measurements being taken (N_Samples=0). No Final Parameters can be reported.

•Profile with all missing observations, except one non-missing observation at dose time, for Plasma or Drug Effect models (N=1 and a point cannot be inserted at dose time because the one observation is already at dose time). All Final Parameters that can be defined, e.g., Cmax, Tmax, Cmin, Tmin, are reported. (This case does not apply to Urine models — the midpoint of an acceptable observation would be after dose time.)

•Bolus-dosing plasma profile with all missing observations, except one non-missing observation after dose time (N=1 and a point cannot be inserted at dose time because NCA does not back-extrapolate C0 when there is only one non-missing observation). All Final Parameters that can be defined, e.g., Cmax, Tmax, Cmin, Tmin, are reported.

Each of these cases will have a ‘Flag_N_Samples” value in the Final Parameters output that is ‘Insufficient’. This flag allows these profiles to be filtered out of the output worksheets if desired, using the Data Wizard.

Note:A non-bolus profile with all missing observations, except one non-missing positive observation after dose time is not a data deficiency, because a valid point is inserted at dose time – (dose time, 0), or (dose time, Cmin) for steady state – which provides a second data point.