Clinical Chemistry 45, No. 10, 1999 1863
References Analytical and Clinical Performance Characteristics of
1. Saiki RK, Scharf S, Faloona F, Mullis KB, Horn GT, Erlich HA, et al. Enzymatic Hybritech’s Tandem-R free PSA Assay during a Large
detection of -globin genomic sequences and restriction site analysis for
diagnosis of sickle cell anaemia. Science 1985;230:1350 – 4. Multicenter Clinical Trial to Determine the Clinical
2. Fischer SG, Lerman LS. DNA fragments differing by single base-pair substi- Utility of Percentage of Free Prostate-specific Antigen,
tutions are separated in denaturing gradient gels: correspondence with Daniel W. Chan,1 Cynthia A. Kelley,1 Timothy L. Ratliff,2
melting theory. Proc Natl Acad Sci U S A 1983;80:1579 – 83.
Dymphna D’Agostino,2 Julie Ritchey,2 Dolores J. Lamb,3 Josie
3. Conner BJ, Reyes AA, Morin C, Itakura K, Teplitz RL, Wallace RB. Detection
of sickle cell -globin allele by hybridization with synthetic oligonucleotides. Beck,3 Nadean Lott,3 Mark H. Wener,4 Phyllis Daum,4 Robert
Proc Natl Acad Sci U S A 1983;80:278 – 82. E. Henkin,5 Diane N. Kaske,5 Debbie W. Golightly,5 James
4. Saiki RK, Bugawan TL, Horn GT, Mullis KB, Erlich HA. Analysis of enzymat- McBride,6 Grace Layco,6 May K. Ota,6 Milenko J. Tanasijevic,7
ically amplified -globin and HLA-DQ DNA with allele-specific oligonucleo-
tide probes. Nature 1986;324:163– 6.
Christine Grudzien,7 David L. Woodrum,8 Kurtis R. Bray,8
5. Orita M, Iwahana H, Kanazawa H, Hayashi K, Seriya T. Detection of Paula C. Southwick,8 Gail H. Gasior,8 and Kathleen G. Love-
polymorphisms of human DNA by gel electrophoresis as single-strand land8* (1 Department of Clinical Chemistry, The Johns
conformation polymorphisms. Proc Natl Acad Sci U S A 1989;86:2766 –70.
Hopkins University, 600 N. Wolfe St., Baltimore, MD
6. Newton CR, Graham A, Heptinstall LE, Powell SJ, Summers C, Kalsheker N,
et al. Analysis of any point mutation in DNA. The amplification refractory 21287-7065; 2 Department of Urology Research, Washing-
mutation system (ARMS). Nucleic Acids Res 1989;17:2503–16. ton University, 10130 Wohl Clinic, St. Louis, MO 63110;
7. Sommer SS, Cassady JD, Sobell JL, Bottema CDK. A novel method for Scott Department of Urology and Cell Biology, Baylor
detecting point mutations or polymorphisms and its application to popula-
tion screening for carriers of phenylketonuria. Mayo Clin Proc 1989;64:
College of Medicine, 1 Baylor Plaza, Houston, TX 77030;
1361–72. Department of Immunology, University of Washington
8. Syvanen A-C, Aalto-Setala K, Harju L, Kontula K, Soderlund H. A primer- Medical Center, Box 357110, Seattle, WA 98195; 5 Nuclear
guided nucleotide incorporation assay in the genotyping of apolipoprotein E. Medicine In Vitro Laboratory, Loyola Medical Center,
Genomics 1990;8:684 –92.
9. Landegren U, Kaiser R, Sanders J, Hood L. A ligase-mediated gene detection
2160 S. First Ave., Maywood, IL 60153; 6 Department of
technique. Science 1988;241:1077– 80. Clinical Chemistry, UCLA Medical Center, 10833 LeConte
10. Wallace RB, Shaffer J, Murphy RF, Bonner J, Hirose T, Itakura K. Hybridiza- Ave., Los Angeles, CA 90095; 7 Department of Clinical
tion of synthetic oligodeoxyribonucleotides to DNA: the effect of single base Chemistry, Brigham & Women’s Hospital, 75 Francis St.,
pair mismatch. Nucleic Acids Res 1979;6:3543–57.
11. Ikuta S, Takagi K, Wallace RB, Itakura K. Dissociation kinetics of 19 base
Boston, MA 02115; 8 Department of Research and Devel-
paired oligonucleotide-DNA duplexes containing different single mismatched opment, Hybritech Incorporated, a subsidiary of Beckman
base pairs. Nucleic Acids Res 1987;15:797– 811. Coulter, Inc., P.O. Box 269006, San Diego, CA 92196-9006;
12. Saiki RK, Walsh PS, Levenson CH, Erlich HA. Genetic analysis of amplified * author for correspondence: fax 619-536-8058, e-mail
DNA with immobilized sequence-specific oligonucleotide probes. Proc Natl
Acad Sci U S A 1989;86:6230 – 4. KGLoveland@Beckman.com)
13. Cros P, Allibert P, Mandrand B, Tiercy J-M, Mach B. Oligonucleotide
genotyping of HLA polymorphism on microtitre plates. Lancet 1992;340: Prostate-specific antigen (PSA) is present in serum in
several forms, most importantly free PSA (FPSA) and PSA
14. Nevinny-Stickel C, Albert ED. HLA class II typing in a microtitre plate format
using digoxigenin-labelled amplified DNA and biotin-labelled oligonucleotide complexed to 1-antichymotrypsin (1, 2 ). These PSA
probes. Eur J Immunogenet 1993;20:419 –27. forms are useful in assessing prostate disease (3–5 ). When
15. Yang B, Viscidi R, Yolken R. Quantitative measurement of nonisotopically PSA is 10 g/L (ng/mL), the probability of prostate
labelled polymerase chain reaction product. Anal Biochem 1993;213:
422–5. cancer is 50% (6 ); when PSA is between 4 and 10 g/L,
16. Bottema CDK, Sommer SS. PCR amplification of specific alleles: rapid the probability of prostate cancer is 25%. Patients in the
detection of known mutations and polymorphisms. Mutat Res 1993;288: latter range are usually recommended for biopsy, but here
the low specificity leads to many unnecessary biopsies.
17. Chadwick RB, Conrad MP, McGinnis MD, Johnston-Dow L, Spurgeon SL,
Kronick MN. Heterozygote and mutation detection by direct automated The percentage of FPSA (%FPSA) is used to enhance
fluorescent DNA sequencing using a mutant Taq DNA polymerase. Biotech- specificity. The relative proportion of FPSA in serum may
niques 1996;20:676 – 83. range from 5% to 50% (7 ), but a lower %FPSA is associ-
18. Parker LT, Zakeri H, Deng Q, Spurgeon S, Kwok P-Y, Nickerson DA. AmpliTaq
DNA polymerase, FS dye-terminator sequencing: analysis of peak height
ated with higher probability of prostate cancer (8 ).
patterns. Biotechniques 1996;21:694 –9. Published guidelines for the clinical use of %FPSA have
19. D’Alfonso S, Richiardi PM. A polymorphic variation in a putative regulation been contradictory because of differences in assay sys-
box of the TNFA promoter region. Immunogenetics 1994;39:150 – 4. tems and standardization, study designs, patient popula-
20. Hamann A, Mantzoros C, Vidal-Puig A, Flier JS. Genetic variability in the
TNF- promoter is not associated with type II diabetes mellitus (NIDDM).
tions, and the number of subjects enrolled (9 ). A large
Biochem Biophys Res Commun 1995;211:833–9. well-controlled multicenter clinical trial was conducted to
21. McGuire W, Knight JC, Hill AVS, Allsopp CEM, Greenwood BM, Kwiatkowski define a medically significant %FPSA cutpoint that would
D. Severe malarial anaemia and cerebral malaria are associated with indicate the need for prostate biopsy. This report summa-
different TNF promoter alleles. J Infect Dis 1998;179:287–90.
22. Knight JC, Udalova I, Hill AVS, Greenwood BB, Peshu N, Marsh K, et al. A
rizes the assay performance from the largest clinical trial
polymorphism that affects OCT-1 binding to the TNF promoter region is to date evaluating %FPSA in a patient population repre-
associated with severe malaria. Nat Genet 1999;22:145–50. sentative of those men in whom the test would be used in
23. Giorda R, Lampasona V, Kocova M, Trucco M. Non-radioisotopic typing of clinical practice. The data formed the basis for the Food
human leukocyte antigen class II genes on microplates. Biotechniques
1993;15:918 –25. and Drug Administration’s approval of Hybritech’s free
24. Nickerson DA, Kaiser R, Lappin S, Stewart J, Hood L, Landegren U. PSA assays. Detailed clinical performance characteristics
Automated DNA diagnostics using an ELISA-based oligonucleotide ligation have been presented previously (10 ).
assay. Proc Natl Acad Sci U S A 1990;87:8923–7.
The objective of this clinical trial was twofold: to
evaluate the performance of Hybritech’s Tandem -R free
1864 Technical Briefs
Three laboratories used Tandem-E PSA and four labora-
tories used Tandem-R PSA. Serum was then stored at
70 °C. After meeting all enrollment criteria, the same
serum was thawed and assayed at each site in duplicate
using the Tandem-R free PSA (immunoradiometric mono-
clonal antibody) assay (11 ). FPSA assays included kit
controls ( 1.0 and 15 g/L) and three serum pools ( 0.3,
2.0, and 6.5 g/L FPSA). In this blinded study, FPSA was
not reported to the physician, and laboratory personnel
did not know the diagnosis of the patient. The percentage
of FPSA was calculated as (FPSA/PSA) 100%
Tandem (total) PSA and Tandem free PSA assays are
manufactured in several formats at Hybritech Incorporated,
a subsidiary of Beckman Coulter, Inc., San Diego, CA.
The lower limit of detection was calculated at each
laboratory as the FPSA concentration corresponding to
the signal 2 SD greater than the mean of 20 replicates of
the zero calibrator. Results for the lower limit of detection
at the seven sites were 0.03, 0.02, 0.04, 0.03, 0.02, 0.01, and
0.02 g/L FPSA, and met the manufacturer’s claim of 0.05
g/L or less.
A proficiency panel prepared at Hybritech was used for
Fig. 1. Tandem-R free PSA assay reproducibility over time. interlaboratory method comparison. Each laboratory re-
Results of two kit controls and three serum pools plotted from seven laboratories ceived identical sets of 43 test samples for analysis in
using 41 lots of reagent in 230 assays over the 17 months of the clinical trial. duplicate using Tandem free PSA and total-PSA assays.
Assessment of 720 results demonstrated that equivalent
PSA assay as it is routinely used in laboratories, and to results, laboratory to laboratory, were obtained for both
identify a %FPSA cutpoint with a high rate of cancer total and free PSA. The Levene test for homogeneity of
detection (clinical sensitivity) while avoiding unnecessary variance showed no significant differences in dispersion
biopsies in men without cancer (clinical specificity) when among sites for FPSA (P 0.20) or PSA (P 0.98). A
total PSA is between 4 and 10 g/L. repeated measures ANOVA showed no significant differ-
Subjects were enrolled prospectively at seven univer- ences across the sites for results of the proficiency samples
sity medical centers in accordance with practices and (FPSA, P 0.81; PSA, P 0.99). Therefore, pooling of
ethical standards of each site’s Institutional Review Board results across sites and the use of any Tandem PSA format
and the Declaration of Helsinki. The subjects were men was acceptable.
50 –75 years of age with a nonsuspicious digital rectal Three elements of precision and long-term reproduc-
examination, a Tandem PSA result between 4 and 10 ibility were determined:
g/L, and a histologically confirmed diagnosis from a
six-sector ultrasound-guided needle biopsy. 1. Reproducibility within and among the seven labora-
Blood was drawn before prostate treatment or biopsy. tories was evaluated according to NCCLS Guideline
Serum was separated from the clot within 3 h, stored at EP5-T2 for evaluation of precision performance of
4 °C, and tested within 24 h. PSA was assayed and chemistry devices. Results from 20 analytical runs of
reported in duplicate using Hybritech’s Tandem-E PSA five controls, tested in duplicate, showed total repro-
for photon ERA® instrumentation (immunoenzymetric ducibility at 6.8% CV for the lowest concentration
monoclonal antibody assay) or Tandem-R PSA (immuno- (0.32 g/L FPSA; Table 1).
radiometric monoclonal antibody assay) reagent sets. 2. Two assayed kit controls and three serum pools were
Table 1. Imprecision of Tandem-R free PSA.a
SD, g/L CV, %
ID Mean, g/L Within-run Between-run Total Within-run Between-run Total
Pool 1 0.32 0.011 0.019 0.022 3.3 6.0 6.8
Lo-Kit 1.03 0.026 0.018 0.032 2.6 1.7 3.1
Pool 2 2.02 0.045 0.117 0.126 2.2 5.8 6.2
Pool 3 6.41 0.139 0.385 0.409 2.2 6.0 6.4
Hi-Kit 15.7 0.25 0.14 0.29 1.6 0.9 1.8
One assay per day for 20 days, samples in duplicate.
Clinical Chemistry 45, No. 10, 1999 1865
used to accept or reject each of 230 assays. The lot of kit would detect 95% of cancers (clinical sensitivity) and
controls changed in July and October 1996. The lot of would spare 20% of men with benign disease from biopsy
pools used did not change. The seven laboratories used (clinical specificity). The second approach provides an
41 lots of reagent to produce 773 patient results during individual patient’s risk (probability of cancer). Risk
the 17 months of the clinical trial. Fig. 1 is a graph of would be stratified based on the %FPSA value when total
controls and pools and shows consistent and reproduc- PSA is between 4 and 10 g/L. The risk of cancer is high
ible recovery over time. (56%) when %FPSA is 0 –10%, and the risk decreases as
3. The imprecision (CV) on sample duplicates for the
%FPSA increases. The risk of cancer is low (8%) when
proficiency panel was 3.1% for FPSA and 2.4% for PSA
%FPSA is 25%. This approach allows the physician and
across all sites. The precision (CV) for the 773 subject
serum samples (tested in duplicate) averaged 3.0% for patient to discuss individualized management options.
all sites combined. This reproducibility between sample
replicates supports testing routine samples in singlet.
1. Christensson A, Laurell CB, Lilja H. Enzymatic activity of prostate-specific
After successful completion of method validation pro- antigen and its reactions with extracellular serine proteinase inhibitors. Eur
cedures, each laboratory reported free- and total-PSA J Biochem 1990;194:755– 63.
results on subjects meeting study enrollment criteria. 2. Lilja H, Christensson A, Dahlen U, Matikainen MT, Nilsson O, Pettersson K,
Lovgren T. Prostate-specific antigen in serum occurs predominately in
ROC curves were generated using data from the 773 complex with 1-antichymotrypsin. Clin Chem 1991;37:1618 –25.
subjects enrolled in the trial (379 prostate cancer and 394 3. Stenman U-H, Leinonen J, Alfthan H, Rannikko S, Tuhkanen K, Alfthan O. A
benign prostatic disease). A cutpoint of 25% FPSA was complex between prostate-specific antigen and 1-antichymotrypsin is the
major form of prostate-specific antigen in serum of patients with prostatic
determined to yield a sensitivity of 95%. The use of this cancer: assay of the complex improves clinical sensitivity for cancer. Cancer
cutpoint (performing biopsies only in patients with values Res 1991;51:222– 6.
less than or equal to the cutpoint) would have avoided 4. Christensson A, Bjork T, Nilsson O, Dahlen U, Matikainen MT, Cockett AT, et
al. Serum prostate specific antigen complexed to 1-antichymotrypsin as an
biopsies in 20% of men with benign disease. indicator of prostate cancer. J Urol 1993;150:100 –5.
The probability of cancer was determined from PSA 5. Sokoll LJ, Chan DW. Total, free and complexed PSA: analysis and clinical
and %FPSA results on these subjects with a biopsy-based utility. J Clin Ligand Assay 1998;21:171–9.
6. Catalona WJ, Richie JP, Ahmann FR, Hudson MA, Scardino PT, Flanigan RC,
diagnosis. For men with PSA between 4 and 10 g/L, the et al. Comparison of digital rectal examination and serum prostate specific
risk of cancer is 25% (6, 12 ). %FPSA is useful to further antigen in the early detection of prostate cancer: results of a multicenter
stratify the probability of cancer in this group. For subjects clinical trial of 6,630 men. J Urol 1994;151:1283–90.
with %FPSA values of 0 –10%, 10 –15%, 15–20%, 20 –25%, 7. Catalona WJ, Smith DS, Wolfert RL, Wang TJ, Rittenhouse HG, Ratliff TL, et
al. Evaluation of percentage of free serum prostate-specific antigen to
and 25%, cancer was found on biopsy in 56%, 28%, 20%, improve specificity of prostate cancer screening. JAMA 1995;274:1214 –
16%, and 8%, respectively. 20.
Percentage of FPSA studies may be affected by numer- 8. Partin AW, Catalona WJ, Southwick PC, Subong ENP, Gasior GH, Chan DW.
Analysis of percent free prostate-specific antigen (PSA) for prostate cancer
ous factors (9 ). One important factor is sample stability. detection: influence of total PSA, prostate volume, and age. Urology
The handling of samples in this study followed published 1996;48:55– 61.
data. PSA and FPSA are reported to be stable when the 9. Woodrum DL, Brawer MK, Partin AW, Catalona WJ, Southwick PC. Interpre-
tation of free prostate specific antigen clinical research studies for the
serum is separated from the clot within 3 h, stored at 4 °C, detection of prostate cancer. J Urol 1998;159:5–12.
and tested within 24 h. Freezing and thawing serum does 10. Catalona WJ, Partin AW, Slawin KM, Brawer MK, Flanigan RC, Patel A, et al.
not affect recovery (13, 14 ). It is important to freeze serum Use of the percentage of free prostate-specific antigen to enhance differen-
tiation of prostate cancer from benign prostatic disease: a prospective
promptly after pipetting to preserve FPSA integrity. Sam- multicenter clinical trial. JAMA 1998;279:1542–7.
ples may be stored at 70 °C for up to 2 years (15 ). 11. Woodrum DL, French CM, Hill TM, Roman SJ, Slatore HL, Shaffer JL, et al.
PSA and %FPSA results from other manufacturers may Analytical performance of the Tandem-R free PSA immunoassay measuring
free prostate-specific antigen. Clin Chem 1997;43:1203– 8.
provide values different from those obtained in this 12. Keetch DW, Catalona WJ, Smith DS. Serial prostatic biopsies in men with
study. The use of assays from multiple manufactures or persistently elevated serum prostate specific antigen values. J Urol 1994;
applying the cutpoint derived from one set of assays to 151:1571– 4.
13. Woodrum DL, French C, Shamel LB. Stability of free prostate-specific
that of another manufacturer can lead to erroneous clini- antigen in serum samples under a variety of sample collection and sample
cal conclusions (5, 16, 17 ). Differences may be attributable storage conditions. Urology 1996;48:33–9.
to assay calibration, antibody specificity, or non-equimo- 14. Arcangeli CG, Smith DS, Ratliff TL, Catalona WJ. Stability of serum total and
larity of the total-PSA assay. free prostate specific antigen under varying storage intervals and tempera-
tures. J Urol 1997;158:2182–7.
In summary, results from this large multicenter clinical 15. Woodrum D, York L. Two year stability of free and total PSA in frozen serum
trial show that %FPSA can be applied in two ways: (a) use samples. Urology 1998;52:247–51.
of a single cutpoint, or (b) determination of individual risk 16. Semjonow A, Oberpenning F, Brandt B, Zechel C, Brandau W, Hertle L.
Impact of free prostate-specific antigen on discordant measurement results
for cancer. With the first approach, a single cutpoint of of assays for total prostate-specific antigen. Urology 1996;48:10 –5.
25% is recommended for men 50 –75 years of age with 17. Nixon RG, Meyer GE, Blase AB, Gold MH, Brawer MK. Comparison of 3
PSA results between 4 and 10 g/L with benign findings investigational assays for the free form of prostate specific antigen. J Urol
in a digital rectal examination. Men with 25% FPSA or
below would be recommended for biopsy. This approach