Siemens Coat-A-Count TSH IRMA Manual Del Usuario página 6

Comments: Although other approaches
are acceptable, data reduction by the
method just described has certain
advantages from the standpoint of quality
control. In particular, it yields a calibration
curve that is relatively linear in both log-log
and linear-linear representations, and
relatively stable from assay to assay. It
also yields valuable QC parameters,
namely, Percent Bound (%B/MB) values
for the nonzero calibrators. A still more
informative graph, conveying a sense of
within-assay reproducibility as a function
of concentration, can be obtained by
plotting the Percent Bound values of
individual calibrator tubes directly, i.e.
without first averaging the CPM of
replicates.
Alternatives: Although Percent Bound
can be calculated directly from Average
CPM, correction for nonspecific binding
usually produces a calibration curve that is
more nearly linear throughout its range. A
calibration curve can also be constructed
by plotting CPM or Average CPM directly
against Concentration on either log-log or
linear-linear graph paper. (Semi-log graph
paper should not be used.) This approach
has the virtue of simplicity, but is less
desirable from the standpoint of quality
control.
Computerized Data Reduction: "Point-
to-point" methods, including linear and
cubic spline fits, are suitable; but since
they provide little assistance in monitoring
the integrity of an assay, it is important to
prepare the recommended log-log plot of
the calibration curve, either manually or by
computer, as a quality control step. Data
reduction techniques based on the logistic
model may also be applicable. Within this
family, curve-fitting routines based on the
4- or 5-parameter logistic are the most
suitable candidates. However, some
algorithms currently in use may not
converge successfully, even when the
logistic model is true to the data. If a
logistic method is adopted, it is essential
to verify its appropriateness for each day's
assay by monitoring the backcalculation of
the calibrators, and other parameters. In
addition, a plot of the calibrator curve in a
log-log representation is highly
recommended, as this is more informative
than the conventional semi-log plot.
Sample Handling: The instructions for
handling and storing patient samples and
6
components should be carefully observed.
Dilute patient samples expected to contain
high concentrations with the zero
calibrator before assay. All samples,
including the calibrators and controls,
should be assayed at least in duplicate. It
is important to use a disposable-tip
micropipet, changing the tip between
samples, in order to avoid carryover
contamination. Positive displacement
pipets and automatic pipettor-diluters
should be used only if the possibility of
carryover has been evaluated and found
to be insignificant. Pairs of control tubes
may be spaced throughout the assay to
help verify the absence of significant drift.
Inspect the results for agreement within
tube pairs.
Gamma Counter: To minimize the
possibility of spillover in multi-well gamma
counters, the optional total counts tubes
(T) should be separated by one or more
spaces from the other assay tubes.
Alternatively, add only 25 µL of the tracer
to each of the T tubes at step 3, and
multiply the observed counts per minute in
these tubes by 4.
Controls: Controls or serum pools with at
least two TSH concentration levels (low
and high) should routinely be assayed as
unknowns, and the results charted from
day to day as described in Westgard JO,
et al. A multi-rule chart for quality control.
Clin Chem 1981;27:493-501. Repeat
samples are a valuable additional tool for
monitoring interassay precision.
QC Parameters: We recommend keeping
track of these performance measures:
T = Total Counts (as counts per minute)
%NSB = 100 × (Average NSB Counts / Total
Counts)
%MB = 100 × (Net MB Counts / Total Counts)
And the Percent Bound ("%B/MB") values
of all but the highest of the nonzero
calibrators, for example:
%C/MB = 100 × (Net Calibrator "C" Counts / Net
MB Counts)
Record Keeping: It is good laboratory
practice to record for each assay the lot
numbers and reconstitution dates of the
components used, as well as control
results and QC parameters.
Coat-A-Count TSH IRMA (PIIKTS-8, 2010-11-04)