GMP/FDA Requirements on HPLC Systems in the Lab

GMP News No. 391

23 February 2004

GMP/FDARequirements on HPLC Systems in the Lab

1) New AAPS Recommendations on Analytical Instrument Qualification

In July 2003 the Gold Sheet published a Summary Report on AAPSInstrument Qualification Workshop (March 2003). This workshop wasco-sponsored by ISPE and the International Pharmaceutical Association(FIP). The aim of this meeting, including FDA representatives, was toestablish common terminology for Analytical Instrument Qualification(AIQ).

It was considered appropriate that the DQ, IQ, OQ, PQ terms forqualification – although these terms were found confusing for AIQ –should be continued and that the AIQ process should be defined aroundthem, since these terms have already found their place in the validation(qualification) terminology.

Key requirements for IQ, OQ and PQ were defined. For instance the term OperationalQualification (OQ) is defined and explained as:

  • On-site testing and documentation providing verification that ananalytical instrument is functioning as the vendor intended
  • OQ may be performed at some regular interval thereafter
  • OQ is also performed as required by the change control of theinstrument, and after major repairs that might affect critical factors ofinstrument operation
  • OQ tests are not method specific
  • Absorbance detector wavelength accuracy and absorbance linearity,injector precision, pump flow rate accuracy and proper gradient operationare some important tests for HPLC components
  • Holistic tests are acceptable in lieu of modular testing

Performance Qualification is defined as on-site testing thatverifies consistent and proper operation of the analytical instrument orsystem, according to a specification appropriate for its routine use.These tests check the maintenance and calibration of the instrumentSystem. Suitability checks can supplement but not replace the functionalchecks for PQ!

Furthermore, software validation was probably the most controversialissue handled at this AAPS workshop. A white paper will be published basedon the workshop discussions.

2) HPLC in Human Drug cGMP Notes

In earlier issues of FDA's Human Drug CGMP Notes therewere two questions dealing with HPLC, which might be interesting althoughthey are not brand-new:

The first question was in March 1996:
"Does FDA have a CGMP policy on use of recycled solvents for HPLCcolumns?"

The answer was: "The agency has no specific policy on recycledHPLC solvents…. Therefore it would be acceptable to use recycledsolvents which do not interface with analytical results or equipmentoperation."

The second question was in September 1997:
"When analyzing a sample by HPLC to determine variance within abatch, such as content uniformity or dissolution, and the result for onetablet is out of specifications (OOS), if there is clearly an HPLCmalfunction, does the whole test need to be re-run, or only the onesolution?"

Russ Rutledge (FDA) answered to this question:
"……if the analytical instrument was clearly shown to giveunreliable results by the laboratory investigation, then the test resultsmust all be invalidated and the entire test run again, using the originalsample solutions if possible. The new results are substituted for theinvalidated results, and only these are used in the batch evaluation. Tomerely re-run the one sample solution is inadequate because the rejectionof data is based on instrument malfunction. Thus, all results from thistest should be considered unreliable."

3) HPLC in FDA Warning Letters

FDA Inspections often cover HPLC details as you can see from the belowlisted FDA Warning Letter citations:

  • Procedure for HPLC Calibration Method 013 has no predeterminedacceptance criteria for the auto sampler calibration. The auto samplercalibration does not demonstrate that the instrument is capable ofaccurately assessing linearity
  • The calibration procedure for HPLC Systems is inadequate in that itdid not include integrator and detector's linearity, injector'sreproducibility, and accuracy of temperature settings for columnheater and detector
  • Data from two of the test runs were voided due to HPLC systemsproblems, however a review of the chromatograms showed that systemsuitability was maintained throughout the run
  • No data demonstrating suitability for use could be provided for yourfirm's HPLC system
  • HPLC chromatograms were found to have poor separation between activeingredient peak and unknown peak, as well as widely differentretention times for the active peak between different chromatographicruns
  • HPLC chromatographic data do not always document system suitabilityrequirements
Dr Günter Brendelberger


Go back

GMP Conferences by Topics