| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
Department of Research and Development, Hybritech Inc., San Diego, CA 92196.
a Author for correspondence. Fax 619-536-8058; e-mail dlwoodrum{at}beckman.com
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1-antichymotrypsin was determined to be <1%. The
minimum-detectable concentration was <0.05 µg/L. The within-run and
between-day CVs were 
5% for samples with >0.3 µg/L free PSA.
Dilution and recovery showed no significant deviations from linearity
across the assay range. The assay was insensitive to interference from
blood components. The Tandem-R free PSA kit was shown to be
an accurate, precise, and reliable assay for the measurement of free
PSA.
Key Words: indexing terms: tumor markers • prostate cancer
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1-antichymotrypsin (ACT),
2-macroglobulin (2M), and other binding
proteins (5). Total PSA (i.e., all of the immunologically
detectable forms, consisting primarily of PSA-ACT and free PSA) has
served as an excellent indicator of prostate disease when the
concentration exceeds 4.0 µg/L in serum (6). However,
distinguishing prostate cancer from benign prostatic hyperplasia is
clinically imprecise (7). Recent studies have suggested
that PSA forms may have clinical value for distinguishing prostate
cancer from benign prostatic hyperplasia (8)(9)(10)(11)(12)(13)(14)(15). In
particular, the use of the percent of free PSA [% = (free PSA/total
PSA) x 100], as determined by Tandem®-R free
PSA assay in conjunction with Tandem-R (total) PSA assay (both from
Hybritech Inc.), can improve the clinical specificity of PSA testing
such that between 20% and 64% of negative biopsies may be avoided
with only a 5–10% decrease in the cancer detection rate
(13)(14)(15). Differences in the relative proportions of free PSA and PSA-ACT affect the results of some total PSA assays (16). In addition, differences in the calibration and performance of free PSA assays have been reported recently (17)(18). These reports imply that the values derived for percent free PSA and the attendant cutoffs used are dependent on which free PSA and total PSA assays are used. This study was performed to characterize the analytical performance and calibration of the Tandem-R free PSA assay and the percent free PSA value obtained by using the Tandem-R free PSA and Tandem-R (total) PSA assay systems.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
purified free psa and psa-act
Purified PSA-ACT and free PSA were prepared by the method of
Sensabaugh and Blake (20). Seminal fluid PSA was incubated
with a 6-fold molar excess of ACT (Athens Research and Technology,
Athens, GA) for 18 h at 37 °C in Tris-buffered saline, pH 7.4,
and fractionated over a poly(propylaspartamide) hydrophobic interaction
chromatographic column (PolyLC) in a gradient of 1.2 mol/L sodium
sulfate, 0.02 mol/L sodium phosphate, pH 6.3, and 0.05 mol/L sodium
phosphate, 50 mL/L isopropanol, pH 7.3. Baseline resolution between PSA
(~34 kDa), PSA-ACT (~90 kDa), and ACT (~55 kDa) was observed. The
concentrations of free PSA and total PSA were measured in each fraction
by the free and total PSA assays, as described below. Fractions were
pooled on the basis of immunoreactivity and stored in 0.1 mol/L
ammonium acetate at -70 °C.
assay calibration and analytical recovery
For accurate percent free PSA results, the free PSA assay should
be calibrated in agreement with the total PSA assay used so that a
sample that contained only free PSA would be recovered identically in
both assays. Thus, the calibration of the two assays would be
analytically linked to each other and would provide a meaningful ratio
of free to total PSA. To demonstrate this, purified free PSA was added
to normal female serum treated to inactivate PSA-binding proteins
according to the method of Berger and Ivor (21) to mimic
hypothetical samples of 100% free PSA from 0 to 15 µg/L free PSA.
These samples were then run in both the Tandem-R (total) PSA and
Tandem-R free PSA assays. The recovery of the purified free
PSA in treated female serum was calculated relative to the amount of
analyte added. The correlation and bias between values recovered in the
two assays were determined by linear regression.
analytical performance studies
Specificity.
Dilutions of purified free PSA and PSA-ACT were
prepared in treated female serum in a range of 1–100 µg/L. The
Tandem-R free PSA and Tandem-R (total) PSA assays were run
in triplicate. Specificity was calculated by comparing the relative
reactivity of the purified free and PSA-ACT preparations in the free
PSA assay and also by comparing the relative reactivity of the PSA-ACT
preparation in the free PSA assay and in the total PSA assay.
Precision.
The intraassay precision was determined by running
20 replicates of individual serum-based controls in a single assay. The
interassay precision was determined by running serum-based controls in
triplicate in 20 assays from a single kit lot. The interlot precision
was determined by running serum-based controls in either 54 or 81
individual assays over multiple days, with multiple reagent lots, by
multiple technicians. Results were compared with intersite data from
seven laboratories across the country (22).
Minimum detectable concentration (MDC).
The MDC was determined
for each of 27 assays as the concentration of free PSA corresponding to
a signal 2 SD above the mean of 20 replicates of the zero calibrator.
The overall MDC was determined as the mean MDC of the 27 assays plus 1
SD.
Linearity on dilution.
Ten samples at representative free PSA
concentrations between 2 and 40 µg/L were used to prepare a minimum
of four gravimetric dilutions of each specimen in zero
calibrator/specimen diluent. All specimens and dilutions were run in
four replicates. Linear regression was used to characterize the
dilution series.
High-dose hook effect.
Purified free PSA from seminal fluid
was tested at concentrations of 10–5000 µg/L to determine the free
PSA concentration that would cause the assay to "hook" back into
the normal assay range.
Interfering substances.
Various blood components and
chemotherapeutic and prostate-related drugs were tested at
concentrations at least 10-fold higher than the expected normal range
of the component in blood. Additions of each interfering substance were
made into zero diluent and three serum-based kit controls, which span
the range of the assay. Eight replicates of each control with and
without the test substance were run in the assay. A t-test
was used to evaluate the statistical significance of any differences in
free PSA values observed in the test samples.
Commercial control stability.
Serum controls used routinely
with total PSA assays from four manufacturers: Bio-Rad (Hercules, CA),
CIBA Corning Diagnostics (Medfield, MA), Baxter Diagnostics,
Dade® (Miami, FL), and Medical Analysis Systems (MAS;
Camarillo, CA) were evaluated. Lyophilized controls were first
reconstituted, and then all controls were assayed for free PSA (time 0)
and stored at 4, 25, and -20 °C. At prescribed times (8 h and 1, 2,
3, 4, and 7 days) representative aliquots were removed and assayed in
triplicate with the Tandem-R free PSA immunoassay
(23).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). These results indicate that the assay will recover the
analyte accurately, and the calibrations of the two assays are directly
linked such that a sample consisting of 100% free PSA would be
recovered identically in both the Tandem-R free PSA and
Tandem-R (total) PSA assays. The recovered values of purified free PSA
added to treated female serum were 0.84 µg/L (96% of added analyte),
1.77 µg/L (102%), 4.31 µg/L (99%), 8.75 µg/L (101%), and 13.05
µg/L (100%), with a mean recovery of 99.5% of added analyte.
|
analytical specificity
The specificity of the Tandem-R free PSA assay was
determined by two independent methods. In the first method, the dose
responses of purified free PSA and PSA-ACT were compared in the
Tandem-R free PSA assay (Fig. 2
). Linear regression provided a slope estimate of 4058 cpm per 1
µg/L of free PSA compared with 32 cpm per 1 µg/L of PSA-ACT. The
ratio of the two slopes of response in the Tandem-R free PSA
assay yielded a cross-reactivity estimate of ~0.8%. In the second
method, the purified PSA-ACT was run in both the Tandem-R
free PSA and the Tandem-R (total) PSA assays. The ratio of
the values recovered in the two assays resulted in an average
cross-reactivity of 0.7%.
|
precision
Intraassay precision was determined from a mean of 20 replicates
of each control in a single assay with the following results: 0.22
µg/L, CV = 7.7%; 0.32 µg/L, CV = 3.5%; 1.05 µg/L,
CV = 2.2%; 3.85 µg/L, CV = 1.7%; 4.62 µg/L, CV =
1.5%. The interassay precision, interlot precision, and
interlaboratory precision are shown in Table 1
. Overall, the precision of the assay within runs and between
runs was excellent, with CVs of 5% or less for samples with >0.3
µg/L free PSA and 8% or less for samples with <0.3 µg/L free PSA.
In addition, the precision in control recovery between manufacturer's
lots was determined with 27 unique kit lots comprised of 6 lots of
beads, 6 lots of calibrators, and either 6 or 9 lots of tracer (54 or
81 assays, respectively). The CVs of the control recoveries over the
course of the entire multilot comparison were <11% with values <1
µg/L free PSA and <5% with values >1 µg/L free PSA. These data
suggest that the assay is highly reproducible from lot to lot. The
reproducibility of the assay in seven laboratories around the country
was demonstrated when CVs of the control recoveries over the course of
the entire multilaboratory comparison were <12% with values <0.3
µg/L free PSA and <6% with values >0.3 µg/L free PSA
(22).
|
mdc
The individual MDCs of the 27 assays ranged from 0.01 to 0.04
µg/L with a mean MDC of 0.02 µg/L and SD of 0.01 µg/L.
linearity on dilution
Ten samples with free PSA values from 4 to 28 µg/L were used in
a dilution series from 1:2 to 1:100 with undiluted sample. No dilutions
recovering 0.2 µg/L were used in the linearity calculations. Actual
dilution factors used in calculations were derived from gravimetric
data. Overall, individual recoveries of nine samples ranged from 97%
to 114% of expected, r2 
0.998. One sample
recovered consistently at 120% across the dilution series,
r2 = 1.000. The average recovery of all 10
samples was 108%.
high-dose hook effect
Samples of free PSA from 10 to 5000 µg/L were run in the
Tandem-R free PSA assay. The cpm bound in the assay
continued to rise until a plateau was reached near 1000 µg/L (total)
PSA; beyond this, the response began to decline because of the hook
effect and crossed back into the calibrator range of the assay at
~2500 µg/L (total) PSA.
interfering substances
None of the blood components or chemotherapeutic or
prostate-related drugs tested [hemoglobin (2000 mg/L), bilirubin (250
mg/L), triglycerides (23 200 mg/L), total protein (150 g/L), prostatic
acid phosphatase (1000 µg/L), cyclophosphamide (330 mg/L),
diethylstilbestrol (1 mg/L), doxorubicin hydrochloride (6.6 mg/L),
estramustine phosphate (81.7 mg/L), methotrexate (13.2 mg/L), megestrol
acetate (39.6 mg/L), terazosin hydrochloride (1.45 g/L), finasteride
(370 µg/L), flutamide (78 µg/L), ciprofloxacin hydrochloride (46
mg/L), trimethoprim/sulfamethoxazole (9.7 mg/L), goserelin acetate (2.5
mg/L), oxycycline (2.6 µg/mL), leuprolide acetate (8.0 µg/L)]
interfered substantially with the assay as determined by Student's
t-tests between the control and test cases.
commercial control stability
Initial free PSA values ranged from 0.49 to 18 µg/L in controls
from Bio-Rad (0.96, 3.34 µg/L), CIBA Corning (0.49, 4.73, 17.95
µg/L), Dade (0.60, 1.12, 3.71 µg/L), and MAS (0.68, 4.01, 7.13
µg/L). The lyophilized controls (Bio-Rad, CIBA Corning, and Dade)
showed appreciable losses of measured free PSA after reconstitution
(Table 2
). The degree of instability varied by manufacturer. The liquid
controls (MAS) appeared stable over the course of this experiment.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
For accurate percent free PSA results, the free PSA assay should be
calibrated in agreement with the total PSA assay used, such that a
sample that contained only free PSA would be recovered identically in
both assays. When the free and total PSA assays are not linked, the
ratio of free and total PSA is subject to inaccuracies arising from
differences in calibration between the two assays. This situation is
especially problematic when assays from two different manufacturers are
used because it is unlikely that the two manufacturers collaborated in
calibrating their products. Demonstrating this link between assays is
not a simple task because samples of 100% free PSA are not readily
available. To prepare a sample to approximate this theoretical, 100%
free PSA sample, seminal fluid PSA was exhaustively reacted with ACT to
deplete the PSA form that would form ACT complexes. In an initial
experiment (data not shown), the purified ACT-depleted PSA was added to
normal female serum. Further complexation of a fraction of the PSA,
presumably by 2M, resulted in analytical recoveries
between 60% and 90% and made assessment of calibration and analytical
recovery difficult. The purified, ACT-depleted PSA was then added to
female serum that had been treated to inhibit binding proteins. The
recovered values of these samples in both the free and total PSA assays
were then used to establish the correlation in calibration between
assays. A slope of 1.015 and an r2 of 1.000
between the recovered values in the Tandem-R free PSA assay
and recovered values in the Tandem-R (total) PSA assay demonstrates
that these assays provided accurate percent free PSA determinations.
Similar results in analytical recovery and assessment of calibration
possibly could be obtained by first depleting seminal fluid PSA with
both ACT and 2M; however, the suitability of this
approach has yet to be demonstrated.
High-dose hook effect is a known limitation of simultaneous dual-antibody immunoassays. Although a high-dose hook effect was seen when total PSA values exceeded 1000 µg/L, the clinical ramifications of this are not apparent. At this time, the clinical utility of free PSA has been demonstrated only as a reflex test in conjunction with total PSA in the range of ~2–20 µg/L total PSA. The percent free PSA determination has less value as total PSA values exceed 10 µg/L because of the exceptionally high predictive value of PSA in these ranges (6). As such, the limitations of assay hook for samples run in a free PSA assay are not critical. However, analytically, the assay will show a hook effect if the total PSA value of the sample exceeds 1000 µg/L. Therefore, we recommend that, if for any reason, a sample with a known total PSA in excess of 1000 µg/L is run in the free PSA assay, it be diluted first. Other strategies to address hook effect in uncharacterized samples have been described previously (24).
Proper handling and storage of both commercial controls and patients' samples are essential for reliable clinical results. On the basis of the precision data, changes of <15% in individual free PSA values would be considered statistically insignificant when the free PSA concentration was <0.5 µg/L, as would changes <8% when the concentrations were >0.5 µg/L. The lyophilized commercial control preparations showed statistically significant losses in recovery over time after reconstitution. The degree of instability varied by manufacturer. Interestingly, the liquid MAS controls appear to be relatively stable when compared with the reconstituted lyophilized controls. We speculate that the differences between liquid and lyophilized controls may be a result of equilibration between free and bound forms of PSA, and that the liquid controls are at equilibrium, whereas the lyophilized controls may not have been at equilibrium before lyophilization. Therefore, we recommend that, when using lyophilized controls, only freshly-reconstituted material be utilized in free PSA assays. Differences in stability between free and total PSA have also been reported in patients' samples (25)(26). These studies suggest that for the measurement of free PSA, serum should be processed and refrigerated within 3 h of blood draw and frozen (preferably at -70 °C) if it is to be stored >24 h. These observations suggest that free PSA is substantially less stable than total PSA is when refrigerated or shipped on ice and that care should be taken in the handling, storage, and shipment of commercial controls and patients' samples.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1-antichymotrypsin; 2M, 2-macroglobulin; MDC, minimum detectable concentration; MAS, Medical Analysis Systems. 
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1-antichymotrypsin is the major form of prostate-specific antigen in serum of patients with prostatic cancer: assay of the complex improves clinical sensitivity for cancer. Cancer Res 1991;51:222-226.
1-antichymotrypsin as an indicator of prostate cancer. J Urol 1993;150:100-105.
[Web of Science][Medline]
[Order article via Infotrieve]
1-antichymotrypsin in blood samples. Urology 1996;48(Suppl):81-87.
[Web of Science][Medline]
[Order article via Infotrieve]
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  J. Khosravi, A. Diamandi, J. Mistry, and A. Scorilas Insulin-Like Growth Factor I (IGF-I) and IGF-Binding Protein-3 in Benign Prostatic Hyperplasia and Prostate Cancer J. Clin. Endocrinol. Metab., February 1, 2001; 86(2): 694 - 699. [Abstract] [Full Text]
|
|
|
|
 |
|
 A. Magklara, A. Scorilas, W. J. Catalona, and E. P. Diamandis The Combination of Human Glandular Kallikrein and Free Prostate-specific Antigen (PSA) Enhances Discrimination Between Prostate Cancer and Benign Prostatic Hyperplasia in Patients with Moderately Increased Total PSA Clin. Chem., November 1, 1999; 45(11): 1960 - 1966. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. W. Chan, C. A. Kelley, T. L. Ratliff, D. D'Agostino, J. Ritchey, D. J. Lamb, J. Beck, N. Lott, M. H. Wener, P. Daum, et al. Analytical and Clinical Performance Characteristics of Hybritech's Tandem-R free PSA Assay during a Large Multicenter Clinical Trial to Determine the Clinical Utility of Percentage of Free Prostate-specific Antigen Clin. Chem., October 1, 1999; 45(10): 1863 - 1865. [Full Text] [PDF]
|
|
|
|
 |
|
 W. J. Catalona, A. W. Partin, K. M. Slawin, M. K. Brawer, R. C. Flanigan, A. Patel, J. P. Richie, J. B. deKernion, P. C. Walsh, P. T. Scardino, et al. Use of the Percentage of Free Prostate-Specific Antigen to Enhance Differentiation of Prostate Cancer From Benign Prostatic Disease: A Prospective Multicenter Clinical Trial JAMA, May 20, 1998; 279(19): 1542 - 1547. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and PSA-ACT (•) in the
Tandem-R free PSA assay.