BACKGROUND:

There have been continuing problems concerning the appropriate analytical methodology used by FDA laboratories in support of regulatory actions. In several cases regulatory actions have been disapproved and much analyst time wasted because the analyst did not adhere to the appropriate analytical method. Where regulatory actions are predicated upon analytical findings the appropriate methods are generally those stated in the USP/NF, an NDA, or a firm's Standard Operating Procedure, as applicable.

POLICY:

Where FDA sample analysis is a basis for regulatory action only the following procedures are considered appropriate, unless specific instructions to the contrary are given by Center for *Drug Evaluation and Research.*

1. For official drugs (USP/NF) the official compendial analytical methods are to be used, unless the FDA has promulgated regulations under Section 501(b) (or, for antibiotics, Section 507) of the Act prescribing appropriate tests or assay methods, in which case the regulations are to be followed.
2. A non-official drug which is the subject of a new drug application is to be analyzed by the method in the NDA or ANDA.
3. A non-official drug which is not a new drug is to be analyzed by the method used by the manufacturer as part of its standard operating procedures. If the FDA analyst has concern over the validity of the unofficial method, those concerns should be documented.
4. When analyzing a product by any of the above methods, the method must be strictly followed.
5. When a drug is not covered by the above situations, the analyst may select an appropriate method with which to analyze the product. In selecting the method first consideration should be given to any existing AOAC method because AOAC methods have withstood the rigors of collaborative study. Any method selected must have been properly validated. If not previously validated it must be validated when it is used. Validation data must be submitted with worksheets when regulatory action is recommended.
6. When the compendial, NDA, or firm's method is not satisfactory (e.g., due to an interfering substance, non-reproducible method, (etc.)), then this should be reported to the Division of Manufacturing and Product Quality (HFD-320) for further followup and guidance.

For surveillance samples such as those collected during a multiple drug survey, the laboratory may substitute a validated non-official method for the original analysis. However, any out-of-limit results must be confirmed by check analysis using the official or other appropriate method.

*Material between asterisks is new or revised*

Issued: 7/1/81
Revised: 9/1/86, 3/95