Log in
Register|
zheng791 ☆ US, 2013-09-26 18:13 (4244 d 12:34 ago) Posting: # 11566 Views: 4,054 |
|
|
Hi All, Our CRO just dosed the drug for fed and fasting BE study. This drug is for rare disease, causing 1/3 of patients vomit in literature. The calculated sample size is 35 sub for each condition, and we added up to 51 for 3 way. Now for the Period I, 19 and 31 subjects could not continue because of vomiting. I back to check the literature, one research BE paper in fed condition, 24 sub enrolled, and 17 completed for 2 way and the drug still passed (intra CV is ~20%). This the first time I have such tough drug. CRO gave several choices： 1。The CRO suggests Ondacetron could be used anti-vomiting. 2. Start over with more subjects 3. continue the current trials Any suggestion for the tough one? Thank you very much Edit: Category changed. [Helmut] |
|
Dr_Dan ★★ Germany, 2013-09-27 11:59 (4243 d 18:48 ago) @ zheng791 Posting: # 11568 Views: 3,156 |
|
|
Dear zheng791 As you said the intra-subject variation of your drug under fed conditions could be as high as 20%. Statistically, given that the expected Test to Reference ratio of geometric LSmeans should fall within 95 and 105%, it is estimated that the number of subjects to meet the 80 to 125% bioequivalence range with a statistical a priori power of at least 90% is about 26. If you expect a drop out rate of 1/3 you should include 40 subjects. As I understood you conduct two three-way studies with 51 subjects each, right? (Please correct me if I am wrong). In the first period 17 subjects received TEST, 17 subjects received REFERENCE and 17 received ???. In the first study you can continue with 32 subjects (of formerly 51) and in the second study with 20. You can argue that in the remaining two periods the drop out risk would be smaller than in the first period but honestly spoken the course of your study is not very convincing. I would stop both studies to find out what went wrong. I guess both studies are open labelled, so first of all please check if the drop outs are equally distributed between TEST and REFERENCE. In order to give you advice I would need some further information: IR or MR product? Healthy volunteers or patients? Looking forward to your reply Dan — Kind regards and have a nice day Dr_Dan |
Ing. Helmut Schütz