Biomedical Engineering Reference
In-Depth Information
analyte or internal standard signal; divergent standard calibrators, raised baseline,
imprecision and inaccuracy. To avoid these issues, the following strategies are rec-
ommended: (1) changing the LC column type and gradient, (2) changing the acidic
or basic modifier, or (3) using a column switching technique to achieve two or three
dimensional chromatography, such as the “heart-cut” sampling technique [
21,
22
] .
If the chromatography requires lower organic compositions, the PLs may be retained
on the LC column. This can be resolved by one of two approaches: (1) leaving the
retained PLs intact on the LC column and flushing the column with a strong solvent
such as 10:90 water-acetone after the completion of the entire batch (2) using col-
umn switching technique after each injection such as guard column trapping and
fl ushing, column back fl ushing or forward fl ushing.
The above strategies are elucidated in the following two examples. In the first
example, a LC-MS/MS assay was developed for the quantitation of a proprietary drug
and its metabolite in human plasma with a single structural-analog internal standard
using combination of PPE and Waters Oasis
®
HLB SPE extraction. Under the initial
LC condition, the parent analyte coeluted with lysophospholipids (lyso-PLs) [lyso-
phosphatidylcholine (16:0), 496 ® 184 SRM transition and lysophosphatidylcholine
(18:0), 524 ® 184 SRM transition] and the full length PLs eluted after all analytes
[phosphatidylcholine (38:6), 806® 184 SRM transition]. The issue of late eluting
PLs was resolved by adding column backflush and switching valve after all three
analytes eluted. Because the lyso-PLs and analytes were intermingled, the only viable
solution was to use a different separation mechanism, i.e., different LC stationary
phase and/or LC gradient. The problem was resolved by switching to Metasil AQ
®
C18 (2 × 50 mm) column with a different LC gradient program (Fig.
7
). Using the new
condition, the lyso-PLs were still intermingled with the analytes; however, complete
and consistent resolution between the analytes and the lyso-PLs was obtained, and the
late elution PLs were flushed out from the column after each injection. The second
example describes a proprietary assay for the simultaneous quantitation of eight ana-
lytes in human plasma using a single structural-analog internal standard and PPE
extraction. As expected, the anticipated matrix effects caused by phospholipids were
observed. Using Waters BEH
®
C18 (2 × 50 mm, 1.7 mm) column, 0.2 % NH
4
OH in
NH
4
HCO
3
pH unadjusted and MeOH as mobile phases and a LC gradient program
from 65 to 75 % (strong mobile phase), all nine analytes eluted before PLs. However,
it was found that the PLs were eluted out of the column after several sequential injec-
tions, and sometime, they overlap with the analytes which cause imprecision. The
final solution was to ramp the gradient to 100 % MeOH after all analytes are eluted
out. Figure
8
showed that there is that adequately resolution between the analytes and
PLs using a LC gradient program from 65 to 75 % to 100 %.
4.2
Regression, Linearity, and Carryover
As per the FDA guidance, the simplest regression model that adequately describes
the concentration-response relationship should be used for quantitation. Linear
