Bioequivalence and Bioavailability Forum

Dear Sridhar!

❝ We have come across few cases of Malingering ie ., inducing vomiting by volunteers in BE study, so that we will withdraw and pay him full compensation.

At least I would ‘blacklist’ them for the future…

❝ As per the guidelines, data of the volunteers has to be deleted from the analysis but on the ethical grounds since his data will be deleted, he will be withdrawn from the study.

Are you talking about http://www.fda.gov/cder/guidance/5356fnl.pdf FDA's guidance?

‘We recommend that data from subjects who experience emesis during the course of a BE study for immediate-release products be deleted from statistical analysis if vomiting occurs at or before 2 times median Tmax. In the case of modified-release products, the data from subjects who experience emesis any time during the labeled dosing interval can be deleted.’

❝ Finally during period II, we can be in trouble due to inadequate sample size for conducting study due to these withdrawal.

Possibly, especially if the study size is small.

❝ So, to prevent such situation can be continue the volunteers in the study, until the study completes.

OK, some comments.

1. You are in a strong position if the deletion was planned (i.e., in the protocol, an SOP,…). If you haven’t done so, I would suggest implementing it in the next protocol…
   
2. If the deletion was not planned, you should present both data sets (with/without the subject) according to ICH E9, and discuss the implications on the outcome of the study.
   
3. Deletion due to vomiting is only a recommendation; especially 2× median t_max for IR and τ for MR is arbitrary.
   
4. I would keep the subject in the study, and check the intra-subject residuals in ANOVA whether the subject
   • is an outlier (Lund-test), and/or
     
   • leads to a deviation from a normal distribution (Shapiro-Wilk test).
5. Only if the subject is a statistical outlier an/or leads to a violation of the assumption of lognormal distribution I would exclude him/her from the statistical evaluation based on the observed/reported emesis.
   
6. You should present in your report both evaluations, namely
   • primary: the reduced data set (without the subject), and
     
   • secondary: the full data set.
7. BE assessment should be based on the reduced data set, the reduced data set is only descriptive/exploratory.

Justification: Since the recommendations for deletion are general ones, they may not apply to your drug/formulation. I’ve seen examples, where a subject vomited 10 minutes after administration of an IR formulation on one occasion leading to almost superimposible plasma profiles…

---

Edit: FDA-link corrected to latest archived copy. [Helmut]