PQRI Stratified Sampling Approach To Drug Uniformity Endorsed By Cmte

Stratified sampling and in-process dosage unit testing should be allowed in lieu of blend sampling to demonstrate drug blend uniformity during routine batch production, FDA's Pharmaceutical Science Advisory Committee concluded at its May 8 meeting.

"It seems in the validation procedure you're going to be correlating the blend uniformity with the end product of the dosage units, so as long as that's done and followed, it seems that one could rely, then, on the dosage units as the end test," committee member Patrick DeLuca, PhD, University of Kentucky, said.

The stratified sampling alternative was proposed by a Product Quality Research Institute working group following an analysis of the current regulatory requirements for blend uniformity (¹ , p. 33).

The stratified in-process dosage unit ² analysis being proposed by PQRI would analyze tablet cores, hard gelatin capsules or other solid dosage forms from different locations or during different phases of the manufacturing process as an alternative to direct sampling of blender batches to predict content uniformity and homogeneity of a product.

The committee agreed that stratified sampling should not be limited to bioequivalence and validation batches, but may also be used in routine batch testing. Blender sampling along with stratified sampling would be recommended for bioequivalence testing, process development, and validation.

"What we're trying to say...is that [stratified sampling] is used in concert with blend uniformity as you're doing development validation, but once you've established you're validated, then you do away with blend uniformity and are only using the stratified sampling on your routine batch release," PQRI Steering Committee Chair and Lilly Exec Director-Global Regulatory Affairs Tobias Massa, PhD, said.

PQRI was asked to evaluate the value of blend analysis sampling following the issuance of a ³ draft guidance by FDA on the topic in 1999 that indicated that blend uniformity should be used as a routine in-process test (⁴ (Also see "PQRI Research May Loosen "Conservative" Product Quality Regs - FDA" - Pink Sheet, 22 Nov, 1999.), p. 28). The guidance elicited strong opposition from industry.

In its review, the PQRI working group found that blend uniformity testing as proposed by FDA was not always predictive of the uniformity in dosage units.

A data mining analysis of 149 batches of tablet dosage forms from eight pharmaceutical companies found that the predictive ability of blend uniformity data is dependent on the relative standard deviation (RSD) of the blend.

If the blend RSD is <3%, dosage unit uniformity is predicted; if blend RSD is 3-5%, the predictive ability of blend data "diminishes"; and if the blend RSD is >5%, the predictive ability becomes "unreliable."

Uniformity of dosage units can be affected by a number of factors in the manufacturing process besides mixing, PQRI concluded; sampling of bulk powders may also introduce sampling errors.

The committee agreed. "I think the problem with the blend uniformity is the sampling itself. Any segregation occurs after the sample is taken and analyzed. That's the problem," DeLuca said. "I'd strongly support eliminating a test that not only doesn't predict, but might mislead," committee member Arthur Kibbe, PhD, Wilkes University, remarked.

PQRI also endorsed research on newer methods of assessing blend and dosage unit uniformity, such as on-line measurements using near infrared technology. On-line NIR measurements are among the new process analytical technologies being encouraged by FDA for manufacturing (⁵ (Also see "FDA PAT Inspection Teams Being Formed: Further Training Set For Fall" - Pink Sheet, 15 Apr, 2002.), p. 32).

Office of Pharmaceutical Science Deputy Director Ajaz Hussain, PhD, said FDA will convene an internal group to look at the PQRI recommendations and will conduct a peer review and statistical evaluation. FDA will send the institute its suggestions and concerns; the result of the exchange will be integrated into a revised guidance on blend uniformity.

The pharmaceutical science committee suggested that any FDA revised guidance on blend uniformity should apply to new drugs as well as generic drugs. The original guidance applied only to ANDAs. [For a webcast or video replay of the meeting, visit www.FDAAdvisoryCommittee.com , a news and meeting broadcast service from the publisher of "The Pink Sheet"].