Biomedical Engineering Reference
In-Depth Information
follow the validated LC-MS/MS condition exactly and document raw data promptly
such as injection volume, LC column serial number, lot number for mobile phases and
needle wash solvent, and ID for autosampler, LC pumps, and mass spectrometer.
6.6
Integration and Data Review
Peak integration is conducted using peak processing software. The peak must exhibit
a signal to noise (S/N) of at least 5 to 1 in order to be considered a quantifiable peak.
The scientist must optimize the integration parameters to automatically integrate all
peaks consistently and appropriately with respect to retention time, baseline, and
peak width. The same set of integration parameters should be used for all STDs,
QCs and unknown samples. If one or more peak processing parameter need to be
changed to obtain a proper integration for a few samples, the scientist needs to get
approval from Laboratory Director and the copies of the original and reintegrated
chromatogram need to be placed in study record and reviewed by PI, LD, and QA
auditors. Manual integration is highly prohibited.
6.7
Run Acceptance Criteria
For STDs, the back-calculated concentration must not deviate £15.0 % (£ 20.0 %
for the LLOQ) from their individual target concentrations. At least three-fourths of
the individual STDs must be acceptable. For QC, at least two-thirds of the QC
samples at all levels must be £15.0 % from their target concentrations. At least
50 % of the QC samples at each level (low, medium, and high) must be £ 15.0 %
from their target concentrations. For dilutional QC, for each dilution factor, at
least two-thirds of the QC samples at each dilution must be £15.0 % from their
target concentrations.
7
Incurred Sample Reanalysis
Over the years, regulatory authorities noticed an alarming phenomenon in which
large discrepancies between original and repeat values were observed in some sub-
missions. In all cases the assays were formally validated and met acceptance criteria
based on the performance of standard calibrator and quality control samples. It was
gradually recognized that the spiked standards and QCs is not sufficient to fully
assess the accuracy and reliability of a method. Although the purpose of QC sam-
ples is to mimic the performance of incurred samples, there is ample evidence to
show that these entities often behave in very different manners. The topic of manda-
tory incurred sample reanalysis (ISR) was the primary item of interest discussed at
the third Crystal City meeting in May 2006. This meeting resulted in a formal White
Search WWH ::




Custom Search