ROCHE REQUESTING "CLEAR STANDARDS" FOR GENERIC DIAZEPAM PRODUCTS

ROCHE REQUESTING "CLEAR STANDARDS" FOR GENERIC DIAZEPAM PRODUCTS in a July 12 letter to FDA. Commenting on FDA's July 3 response to the firm's petition requesting FDA to revise its bioequivalence requirements for generic versions of Valium, the letter states that "the Center [for Drugs and Biologics] perpetuates the ambiguities surrounding the status of the guidance and whether these requirements will be applied uniformly to all ANDAs for generic diazepam products." Roche asks the agency to reply to five questions to clarify its diazepam bioequivalence requirements. Specifically, Roche counsel William Vodra (Arnold & Porter) asked if an in vivo bioequivalence study must "have a minimum statistical power of 80% to detect a difference of +/-20% of the reference mean in AUS (area under the curve) and Cmax (peak concentration) at p = 0.05 for both the parent drug and the N-desmethyl metabolite." Vodra said FDA suggested such a requirement in its July 3 letter to Roche, but "did not cite inadequate power as a basis for its finding that the foreign diazepam products Roche tested were bioinequivalent, despite the fact that Roche's study had a power of only 52% to detect differences in Cmax for diazepam, and lower powers for AUC and Cmax of the metabolite N-desmethyl diazepam." FDA's July 3 letter to the firm stated that "it has been a long-standing agency policy to consider drugs to be bioequivalent so long as there is 80% confidence that their absorption profiles do not exceed a variation of +/-20% in the two measured parameters of AUC and Cmax." Vodra also commented on Roche's request that multiple dose bioavailability testing be required for generic diazepam. Objecting to FDA's statement that "multiple dosing with a long half-life drug like diazepam obscures differences in the rate of absorption," Vodra asserted that "the agency has misread Roche's proposal." The study design Roche suggested "not only permits observation of the rate of absorption in the absence of prior drug (day 1 of each phase), but because of its replicate design also allows quantification of intra-subject and intra-product variability in key pharmacokinetic parameters" and therefore enhances the power of the study to detect differences in AUS, Cmax, and time to maximum concentration, Vodra stated. Roche also asked FDA if the in vivo study must have blood samples taken at 20 minutes, 40 minutes, 1 hour, and 1-1/2 hours and what "specification or decisional criteria (e.g., equivalency to the reference drug Valium) will be applied regarding in vitro dissolution at pH 4.5 and pH 5.5." Roche asked FDA to respond to its questions within seven days.