methods for prostatic acid phosphatase (PAP) with routine colorimetric assays for total and tartrate-labile acid phospha- tase and evaluated their relative clinical ...
CLIN. CHEM. 32/9, 1760-1766 (1986)
Immunologicaland ColorimetricDeterminationof ProstaticAcid Phosphatase-Technical and ClinicalReappraisalin SymptomaticPatients I. Glbb, L A. Wilson, P. H. Powell,1 I. Tarblt, D. Pratt,2 and A. R. Hall’ We compared a selection of quantitative immunological methods for prostatic acid phosphatase (PAP) with routine colorimetric assays for total and tartrate-labile acid phosphatase and evaluated their relative clinical merits in the differential diagnosis of prostatic carcinoma. We also assesseda wide range of commercial control materials for suitability of usewith these methods. Patients studied included 111 cases of prostatic carcinoma, 42 cases of benign prostatic hyper-
plasla, and 33 controls.The principlesof the methodsused Included determination of enzymatic activity using p-nitrophenyl phosphate, RIA, immunoradiometric, and enzymoimmunometric assays. Performance characteristics for the im-
munologicalmethods were inferiorto manufacturers’precision and specificityclaims. We identified control materials that were unsuitable for routine use. Poor discrimination between clinical groups was observed for all methods. Analysis by use of a receiver operator characteristic plot failed to improve this. We conclude that the immunological methods we studied offer no advantages over colorimetric methods in the differential diagnosis of prostatic cancer in symptomatic patients.
AdditionalKeyphras.s: cutoff values tatic disease
“kit” methods
receiver operator characteristic plot
pros-
. cancer
Prostatic acid phosphatase (PAP; EC 3.1.3.2), an orthophosphoric monoester phosphohydrolase, is present in large quantities as an exocrine secretion of the normal prostate.8 Malignant disruption of the epithelium and extension beyond the gland allow this enzyme to enter the circulation in increased amounts. Initial observations of the high content of acid phosphatase in the prostate, urine, and seminal fluid were made by Demuth (1) and Kutacher and Wolbergs (2), but abnormally high activities in serum were first associated with prostatic malignancy by Gutman et al. (3). In succeeding years, a growing awareness of the wide distribution of acid phosphatases throughout body tissues called into question the ability of serum enzyme determinations to truly reflect prostatic disease. Initial attempts to improve specificity were based on the use of formaldehyde (4) and tartrate (5) inhibitors, but subsequently a variety of ester phosphate substrates, each claimed to show greater specificity to the prostatic fraction, have been used for this purpose (6-8). Efforts to identify yet-more-specific prostatic antigens led to the production of antibodies to crude prostate
Departments of Clinical Biochemistry and ‘Urology, Freeman Hospital, Newcastle upon Tyne, U.K. 2Present address: MRCISNBTS Blood Components Assay Group, Edinburgh, Scotland. 9Nonstandard abbreviations: PAP, prostatic acid phoaphatase;
TAP, total acid phosphatase; TLAP, tartrate-labile acid phosphatess; CAP, proetatic carcinoma group of patients; BPH, benign proetatic hyperplasia group; CTRL, control group. Received May 12, 1986; accepted May 20, 1986. 1760 CLINICALCHEMISTRY,Vol.32, No.9, 1986
extracts and PAP, and as early as 1964 Schulinan et al. (9) used immunological techniques with gel diffusion to identify PAP. Subsequent isolation and purification of PAP has
resulted in the production of antibodies in various animal species, and these have been incorporated into quantitative assays. However, polyclonal antibodies may vary in analytical specificity from batch to batch. Development of methods for more-reproducible production of antibodies by cell-culture techniques has provided monoclonal antibodies with theoretically superior analytical specificity. The practical realization of this claim has been disputed in some reports (10, 11). Furthermore, the limitations of acid phosphatase determination as a screening test are well recognized (12) and the clinical specificity of the test is questionable, because increases are encountered in benign prostatic hyperplasia and even after complete prostatectomy (13, 14). The technical aims of our investigations were to compare a representative selection of quantitative immunological methods with routine colorimetric assays for total acid phosphatase (TAP) and tartrate-labile activity (TLAP) in serum. The evaluations also included assessment of a range of commercial quality-control sera containing acid phosphatase from a variety of sources in different matrices with a particular regard to their performance in the supposedly prostate-specific assays. The clinical objective was to determine whether the immunological techniques offered any advantage over the colorimetric methods in the differential diagnosis of prostatic carcinoma in patients attending a urological clinic for investigation of prostatic symptoms.
Methods The patients we studied included 111 cases of newly diagnosed prostatic carcinoma (CAP), 42 cases of benign prostatic hyperplasia (BPH), and a control (CTRL) group of 33 men with no prostatic symptoms and no abnormality detected on rectal examination. Blood was sampled between 9 and 10 a.m., prior to physical examination. After clotting for 1 h, serum was separated and stored at -20 #{176}C, some in the presence of NaHSO4 as a preservative of enzyme activity. Analyses were generally performed within four weeks of sample collection. The diagnosis of CAP or BPH was confirmed by histological examination after transurethral resection of the prostate. The routine determinations of TAP and TLAP were performed using p-nitrophenyl phosphate as substrate according to Andersch and Szczypinski (15) and Fishman and Lerner (5), respectively. The immunological methods evaluated included a competitive binding radioimmunoassay (RIA) involving polyclonal antibodies (PAP-TER; Serono Diagnostics, U.K.) and a solid-phase two-site immunoradiometric assay with use of monoclonal antibodies (Tandem PAP; Hybritech Inc., U.K.). Both these assays are calibrated in mass units. Also included were two colorimetnic enzymoimmunoassays, one involving a solid-phase antibody and
p-nitrophenyl phosphate as substrate (Merck Immunoassay PAP; Diagnostica Merck, F.R.G.), the other involving liquid-phase double-antibody enzyme isolation and a-naphthyl phosphate as substrate (Ortho PAP-IA; Ortho Diagnostic Systems Inc., U.K.). The former method is calibrated in mass units, the latter in activity units. All immunoassays were performed according to the recommendations of the respective kit manufacturers. Assessment of method performance involved determination of within-assay imprecision by using randomized duplicate analysis of 25 samples and of between-assay imprecision by measurements during three months on 20 aliquots of sera with normal and above-normal values for acid phosphatase. The specificities of the methods were assessed by adding potential cross reactants to pooled sera from women patients. Human leukocyte acid phosphatases (bands 3 and 4; Calbiochem-Behring Corp., Switzerland) were added to give an approximate activity of 7 U/L; erythrocyte hemolysate was similarly added to give a hemoglobin concentration of 1 g/L. We investigated cross reactivities of bone and liver isoenzymes of alkaline phosphatase by using sera from female patients with Paget’s disease or liver disease, whose alkaline phosphatase activities in serum exceeded 1500 UIL. Assay sensitivities, expressed as the enzyme activity or mass distinguishable from zero with 95% confidence, were determined from replicate analyses of a PAP-free serum. We assessed linearity by using a stabilized commercial enzyme source (QuantaPhos, Beckman). For inter-method comparison of data from all patients we used a de-biased regression technique (16). The quality-control materials evaluated were selected to reflect a wide range of acid phosphatase activities from human, animal, and vegetable sources, constituted in various matrices of human, animal, and chemical origin (Table 1). These materials were prepared according to the manufacturers’ instructions and were analyzed without delay. Aliquots (containing NaHSO4 where appropriate) were stored at -20 #{176}C for subsequent investigation of stability in
the frozen state. The exception was QuantaPhos, which is supplied in the liquid state, stabilized with ethylene glycol, which does not freeze at -20 #{176}C. Within- and between-assay imprecisionwith these materials was evaluated by replicate analyses done throughout the period of the study.
Results The Merck immunoassay PAP kits we used gave absorbances for standards substantially lower than in the manufacturer’s performance claims, and we were unable to improve this aspect even after consultation with the company’s U.K. agent. In view of this, the kit was not further assessed. Within-assay imprecision, determined by using randomized analysis of duplicate sera, and between-assay imprecision data from analysis of normal and above-normal pooled sera for all methods is presented in Table 2. Within-assay imprecision was greatest for the TLAP method. Betweenassay imprecision was poor for the three immunological methods and the TLAP method, particularly in the lower range.
The results of specificity studies are shown in Table 3. Samples from female patients with increased liver or bone alkaline phosphatase activities in their serum as compared with a base poo1 of serum from 30 normal women, showed no cross reactivity in any of the immunological methods. However, both liver and bone phosphat.ases were found to react in the TAP method, and the bone phosphatases also reacted in the TLAP method. When erythrocyte acid phosphatase was added to the same base pool, cross reactivity was observed only in the TAP method, but addition of leukocyte acid phosphatases produced significant cross reactivity in all methods.
Investigation of method sensitivities gave values of 0.15 pg/L and 0.22 ,ug/L for Hybritech and Serono kits, and activities of
