Profiles with no measurable concentrations [Outliers]

posted by vezz – Erba (CO), Italy, 2011-10-06 12:45  – Posting: # 7430
Views: 6,710

Dear All,

in a recent study we observed few profiles with no measurable concentrations (i.e., all the concentrations were below the lower limit of quantification). I would like to ask you how these profiles are usually handled in the statistical analysis in your experience.

Someone suggested not to calculate the PK parameters (e.g., AUC Cmax, etc.) based on these profiles, but to consider these concentrations only in the calculation of the descriptive statistics by time point in order to determine the mean profiles. I wonder if this approach may be reasonable.

I found that this issue was discussed by the Health Canada Scientific Advisory Committee on Bioavailability and Bioequivalence in a workshop held in 2004. The record of proceeding is available here and the relevant part is reported below.

"Issue discussed: What is an inadequate concentration versus time profile?
Conclusion: If there are no measurable concentrations or just one concentration, that profile is not adequate. If two consecutive measurable concentrations are observed and the second is lower than the first, it is possible to compute a Cmax and AUCT. Although these estimates of Cmax and AUCT may be somewhat inadequate, the profile should nevertheless be included in the analysis.
No consensus was reached on treatment of inadequate profiles, i.e., on whether these profiles may be simply dropped from the analysis."

Thank you in advance for your help.

Kind regards,


Complete thread:

 Mix view
Bioequivalence and Bioavailability Forum |  Admin contact
19,408 posts in 4,123 threads, 1,325 registered users;
online 15 (1 registered, 14 guests [including 3 identified bots]).
Forum time (Europe/Vienna): 12:29 CEST

We must be careful not to confuse data with the abstractions
we use to analyze them.    William James

BEBAC Ing. Helmut Schütz