Homogeneous enzyme immunoassay for macromolecular antigens using avidin biotin enzyme

A unique homogeneous enzyme immunoassay used for the determination of haptens and macromolecular antigens has been developed. It consists of avidin–antigen conjugate, anti-ligand antibody, and biotin enzyme (propionyl-CoA carboxylase). The immunological reaction of the conjugate with antibodies specific to ligand sterically prevents enzyme inactivation by avidin. Goat IgG and theophylline were used as model antigens. Avidin–theophylline conjugate inhibited 100% of enzyme activity. Avidin goat IgG conjugate inhibited 90% of enzyme activity. In this method, the measurable range of theophylline was from 500 ng/ml to 50 μg/ml, and for goat IgG, it was 1–1000 μg/ml.     

Use of liposome encapsulation in a combined single-liquid reagent for homogeneous enzyme immunoassay

A technique has been developed to permit mutually reactive macromolecular reagents used in immunoassays to be combined without premature reaction. A conjugate of glucose-6-phosphate dehydrogenase (EC 1.1.1.49) and theophylline has been encapsulated in 0.2-micron-diameter bi-lamellar liposomes. Suspensions of these liposomes had excellent stability. Whereas the enzyme activity of the free conjugate is rapidly inhibited by anti-theophylline antibody, a suspension of the encapsulated conjugate in a solution of the antibody and NAD+ (6.0 mmol/L) retained greater than 92% of the initial enzyme activity after standing for one year at 4 degrees C. At higher NAD+ concentrations the liposomes aggregated, and enzyme activity was inhibited by leakage of the NAD+ hydrolysis product, adenosine diphosphoryl 5-ribose (ADP-ribose), into the liposomes. Inhibition by ADP-ribose could be blocked and partly reversed by adding semicarbazide. The liposomes were efficiently lysed by Triton X-100, deoxycholate, or octyl glucoside, the kinetics and extent of lysis being affected by liposome size and correlating with the acid strength of various cholate derivatives. Addition of a serum sample and a solution of buffer, substrate, and detergent to a single reagent containing the liposomes and anti-theophylline antibody provided assay results equivalent to those obtained by conventional two-reagent EMIT homogeneous enzyme immunoassay for theophylline.     

Enzyme immunochromatography--a quantitative immunoassay requiring no instrumentation

We describe a novel test-strip immunoassay for quantifying drugs in biological fluids. This enzyme immunochromatographic ("immunograph") method combines many features of the enzyme-channeling homogeneous immunoassay with immunochromatography and capillary migration to provide a non-separation, non-instrumental assay for theophylline in which quantification is based on the spatial distribution of enzyme label rather than on the modulation of enzyme activity. Sample antigen and hapten-enzyme conjugate are combined and moved by capillary action up a paper strip on which specific antibody has been immobilized. After color development, the assay result is evaluated by measuring the height of the colored zone on the test strip. Quantification is not a function of enzyme activity, so the method is relatively insensitive to sample matrix effects, enzyme instability, temperature, and incubation timing. Either whole blood or plasma can be used as sample. Results correlate well with those by established instrumental methods. The simple, rapid (15 min), two-incubation protocol is well suited for on-site testing in non-laboratory environments.     

Chemiluminescence immunoassay for somatomedin C in serum

To select the best tracer for use in a competitive immunoassay, we conjugated human somatomedin C (SmC) to various chemiluminescent compounds via two different synthetic pathways. Naphthylhydrazides and arylhydrazides, used as the labels, were incorporated via their imidate or their succinimide esters. Conjugating the carboxy terminal of (amino ethyl)ethyl-isoluminol to SmC via a succinimide linkage supplied the most sensitive detection limit and the most immunoreactive conjugate. We developed an immunoassay based on the use of this conjugate, and evaluated dextran-coated charcoal, second-antibody precipitation, and solid-phase immunoprecipitation for separating bound and free label. This chemiluminescent method has a detection limit of 16 pg per tube, and it is accurate and precise. Correlation studies with a conventional radioimmunoassay (x) for SmC gave the following regression equation: y = 0.66x + 3.76 (r = 0.953, n = 30); the slight discrepancies between the two methods are probably ascribable to the use of different antibodies. We thus propose this chemiluminescence immunoassay as an inexpensive and sensitive alternative to radioimmunoassay for measuring SmC in serum or in extracts of serum.     

A fluorescent immunoassay for theophylline: description and comparison to enzyme immunoassay, liquid chromatography and radioimmunoassay

A fluorescent immunoassay for theophylline is described and comparatively evaluated with radioimmunoassay, high-performance liquid chromatography, and enzyme immunoassay. Fifty sera were collected from 43 patients of a large acute-care medical facility, many of whom were suffering from other diseases in addition to bronchial asthma or apnea of the newborn, and were receiving other medication besides theophylline. Assays of theophylline in each serum sample were performed by each of the 4 procedures. The four methods showed comparable results, although each method had at least one unexplained outlier. Nevertheless, all methods seemed suitable for routine chemistry laboratory use. Three of the techniques have been available for several years, but unexplained erroneous levels are sometimes obtained for every procedure.     

Ultrasensitive and more specific enzyme immunoassay (immune complex transfer enzyme immunoassay) for p24 antigen of HIV-1 in serum using affinity-purified rabbit anti-p24 Fab′ and monoclonal mouse anti-p24 Fab′

Previously, an ultrasensitive enzyme immunoassay (immune complex transfer enzyme immunoassay) for p24 antigen of HIV-1 was developed. The immune complex comprising 2,4-dinitrophenyl-biotinyl-bovine serum albumin-rabbit anti-p24 Fab′ conjugate, p24 antigen, and rabbit anti-p24 Fab′-β-D-galactosidase conjugate was trapped onto polystyrene beads coated with affinity-purified (anti-2,4-dinitrophenyl group) IgG, was eluted with ϵN-2,4-dinitrophenyl-L-lysine, and was transferred to polystyrene beads coated with streptavidin. β-D-Galactosidase activity bound to the streptavidin-coated polystyrene beads was assayed by fluorometry. This assay was highly sensitive. However, bound β-D-galactosidase activity had to be assayed for a long time (20 h), and the nonspecific signal was observed in 5% serum samples from subjects with low risk of HIV infection. In the present study, the assay time for bound β-D-galactosidase activity was shortened to 2.5 h by using 2,4-dinitrophenyl-biotinyl-bovine serum albumin-affinity-purified rabbit anti-p24 Fab′ conjugate and affinity-purified rabbit anti-p24 Fab′-β-D-galactosidase conjugate. Furthermore, the nonspecific signal was found to increase with increasing periods of time for storage of serum samples at -20°C, and this increase was prevented without prolongation of the assay time for bound β-D-galactosidase activity and without loss of the sensitivity by substituting monoclonal mouse anti-p24 Fab′-β-D-galactosidase conjugate for affinity-purified rabbit anti-p24 Fab′β-D-galactosidase conjugate. © 1996 Wiley-Liss, Inc.     

Microparticle-enhanced nephelometric immunoassay for human C-reactive protein.

Polyfunctional hydrophilic microspheres of 125-nm diameter can be produced by copolymerization of acrylic monomers. Purified c-reactive protein (CRP) was covalently bound to these new micropheres, and the conjugate obtained was used as reagent in a microparticle-enhanced nephelometric immunoassay for human CRP. This assay was based on the measure, with a specially designed nephelometer, of the light scattered by aggregates formed during the immunoagglutination of the conjugate with anti-CRP antiserum. Sensitive inhibition of this agglutination by free CRP (6 ng/ml) allowed CRP quantitation in highly diluted serum samples (1/500-1/2,000), excluding any interference or sample pretreatment. The CRP assay was easy to perform (no washing or phase separation), reliable (coefficients of variation ranged from 1.3% to 9.3% for within-run and between-run determinations), and accurate (mean percentage of recovery: 104%; correlation coefficients with accepted analytical methods greater than or equal to 0.97) over a large range of concentrations. The inhibition mode excluded errors in the antigen excess zone and provided total security at high concentrations.     

Enzyme immunoassay of human cytosolic aspartate aminotransferase

A sensitive and specific sandwich enzyme immunoassay (EIA) for human cytosolic aspartate aminotransferase (c-AST) has been developed. Serum was incubated with anti-c-AST antibody-coated polystyrene beads, and further incubated with anti-c-AST antibody-peroxidase conjugate. The peroxidase activity bound to the polystyrene bead was proportional to the amount of c-AST. The method allows measurement of serum c-AST ranging from 50-2,000 micrograms/l. No cross-reactivity with m-AST or other serum components was observed. Recovery, within-day precision, and day-to-day precision were good. The levels of c-AST obtained by the proposed EIA were compared with those based on enzyme activity. The results suggest that there is a considerable excess of immunologically active but catalytically inactive c-AST in normal and patient's sera, and that variable specific activities of c-AST are may be found in sera from different individuals.     

Direct chemiluminescence immunoassay of estradiol in saliva

A sensitive and simple direct solid-phase chemiluminescence immunoassay is described for estradiol in saliva. In this assay, a second antibody is bound to the wells of microtiter plates. Either buffer with standards or saliva (100 microL) is incubated in these wells with monoclonal anti-estradiol antibody and with estradiol-isoluminol conjugate. Incubation time is 2 h. Chemiluminescence of the bound fraction is measured in a manually operated luminometer (Biocounter). The assay has a detection limit of 3.8 pmol/L; analytical recovery of added estradiol is 96.8% (SD 7.0%); within- and between-assay CVs range between 2.5% and 12.7%. Forty unknown saliva samples can be assayed and results calculated within 4.5 h. Results of a slightly modified procedure-with black microtiter plates and a prototype of an automated plate reader (Luminoskan)--compare well with those of the described method (r = 0.97). Because steroid-binding globulins have been found in saliva, the effect of displacing agents on the results of the direct chemiluminescence assay is described.     

Rapid and ultrasensitive enzyme immunoassay (thin aqueous layer immune complex transfer enzyme immunoassay) for HIV-1 p24 antigen

The immune complex transfer enzyme immunoassay for HIV-1 p24 antigen was performed in three different ways (in the present immunoassays I, II, and III) within much shorter periods of time than previously reported. In the present (simultaneous) immunoassay I, p24 antigen was incubated simultaneously with 2,4-dinitrophenyl-biotinyl-bovine serum albumin-affinity-purified rabbit anti-p24 Fab′ conjugate and monoclonal mouse anti-p24 Fab′-β-D -galactosidase conjugate in a total volume of 19 μL for 15 min to form the immune complex comprising the three components. The reaction mixture was incubated with a polystyrene bead of 6.35 mm in diameter coated with affinity-purified (anti-2,4-dinitrophenyl group) IgG for 5 min to trap the immune complex. After washing, the polystyrene bead was incubated with 35 μL of ϵN-2,4-dinitrophenyl-L -lysine for 15 min to elute the immune complex (the first eluate) and, after removing the first eluate, with an additional 35 μL of ϵN-2,4-dinitrophenyl-L -lysine for 1 min (the second eluate). The first and second eluates were incubated with a polystyrene test tube (12 × 75 mm) coated with streptavidin for 15 min. In the present (sequential) immunoassay II, a polystyrene bead of 6.35 mm in diameter successively coated with affinity-purified (anti-2,4-dinitrophenyl group) IgG and 2,4-dinitrophenyl-biotinyl-bovine serum albumin-affinity-purified rabbit anti-p24 Fab′ conjugate was incubated with p24 antigen in a total volume of 20 μL for 5 min and subsequently with monoclonal mouse anti-p24 Fab′-β-D -galactosidase conjugate in a volume of 5 μL for 20 min. The immune complex formed on the polystyrene bead was transferred to a polystyrene test tube coated with streptavidin as described above. In the present (sequential) immunoassay III, p24 antigen was incubated with monoclonal mouse anti-p24 Fab′-β-D -galactosidase conjugate in a total volume of 19 μL for 10 min and with a polystyrene bead of 6.35 mm in diameter coated successively with affinity-purified (anti-2,4-dinitrophenyl group) IgG and 2,4-dinitrophenyl-biotinyl-bovine serum albumin- affinity-purified rabbit anti-p24 Fab′ conjugate for 20 min. The immune complex formed on the polystyrene bead was transferred as described above. The incubations were performed at room temperature either by shaking the polystyrene beads (one/assay) and the reaction mixtures in styrol test tubes (13.3 × 54 mm and 2.1 g) so as to randomly rotate the polystyrene beads or by rotating the polystyrene test tubes (12 × 75 mm) containing the reaction mixtures, so that small drops (19 to 70 μL) of the reaction mixtures evenly contacted all parts of the solid phase surfaces during the incubations (although they contacted only small parts of the solid phase surfaces at a time) to continuously mix thin aqueous layers covering the solid phase surfaces with the rest of the reaction mixtures. (Therefore, these immunoassays are called thin aqueous layer immunoassays.) The detection limits of p24 antigen by 1 hr assay of bound β-D -galactosidase activity in the present immunoassays I, II, and III were 0.1, 0.2 and 0.1 amol/assay, respectively, and were slightly higher than or equal to that by the previously reported immune complex transfer enzyme immunoassay, in which the immune complex was formed for 4 hr, was trapped for 16 hr, and was transferred for 3 hr followed by 1-hr assay of bound β-D -galactosidase activity. By 20-hr assay of bound β-D -galactosidase activity, the detection limit of p24 antigen was further lowered to 10 zmol/assay in the present (simultaneous) immunoassay I and to 3 zmol/assay in the present (sequential) immunoassay III. However, the nonspecific reaction(s) with serum samples from HIV-1 seronegative subjects hampered the improvement of the detection limit by 20-hr assay of bound β-D -galactosidase activity. J. Clin. Lab. Anal. 12:205–212, 1998. © 1998 Wiley-Liss, Inc.     

Centrifugal analyzer used for enzyme immunoassay of progesterone and choriomammotropin

Recently we developed an enzyme immunoassay involving the use of steroid delta-isomerase (EC 5.3.3.1) as enzyme label and exclusion-affinity chromatography for rapid separation of free antigen-enzyme conjugate that bound to antibodies (J. Immunol. Methods 35: 267-284, 1980). Here we describe an automated version of this procedure, for immunoassay of progesterone and choriomammotropin (human placental lactogen) in serum with the use of a centrifugal analyzer. After incubation, suitable dilutions of sera or extract plus antiserum, conjugate, and double antibody were filtered on an estradiol affinity gel-filtration column; the enzyme activity of the filtrates was determined with the centrifugal analyzer. Results correlate well with those obtained by radioimmunoassay: r (progesterone) = 0.980, r (choriomammotropin) = 0.940. The within-run and between-run precision, specificity, sensitivity, accuracy, and speed of this system make it a useful tool for immunoassay.     

Quantitation of human serum apolipoprotein B by enzyme immunoassay in excess antigen

A solid phase enzyme-linked immunoassay based on the 'sandwich' principle was developed for quantitative measurement of apolipoprotein B (Apo B) in human normal or hyperlipoproteinemic sera. The solid phase (polypropylene multi-finned sticks) coated with an excess of sheep anti-Apo B immunoglobulins was incubated with antigen (standards and unknown specimens) and affinity-purified anti-Apo B antibodies conjugated with horseradish peroxidase. In this principle, antigen and conjugate were in excess and fixed amount respectively. A part of antigen was fixed on multi-finned sticks, bound or not to conjugate. Unbound materials were removed by washing. Solid phases were next incubated with the enzyme substrate solution to develop a color which is inversely related to the amount of Apo B. The best technical conditions for the assay were determined. The method was characterized according to precision, sensitivity and accuracy. It yielded values that compared favorably with those obtained by another enzyme-linked immunoassay and by electroimmunoassay.     

A rapid semi quantitative capillary enzyme immunoassay for digoxin

A rapid and sensitive enzyme immunoassay (EIA) which does not require highly trained personnel or specialised instrumentation is described for the estimation of digoxin in serum, plasma or whole blood samples. The method is based on the ability of digoxin in a clinical sample to inhibit the binding of urease-conjugated sheep-antidigoxin immunoglobulin to a glass capillary tube coated internally with a human serum albumin-digoxin conjugate. The bound enzyme activity can then be measured using a substrate solution containing urea and a pH indicator, most suitably bromocresol purple. The enzymic hydrolysis of urea produces ammonia which causes a vivid yellow to purple colour change in the pH indicator. Plasma samples from 92 patients receiving digoxin were screened in parallel with reference plasma containing 1.3 or 3.8 nmol/l digoxin. The results were available within a total test time of 30 min, and showed excellent correlation with those obtained by radioimmunoassay.     

Sensitive enzyme immunoassay of antibodies to HIV-1 p17 antigen using indirectly immobilized recombinant p17 for diagnosis of HIV-1 infection

Recombinant p17 (rp17) antigen of HIV-1 and maltose binding protein-rp17 fusion protein (MBP-rp17) were immobilized onto polystyrene beads in different ways: rp17 and MBP-rp17 were immobilized directly onto polystyrene beads by physical adsorption; biotinyl-rp17 and biotinyl-MBP-rp17 were immobilized indirectly onto streptavidin-coated polystyrene beads; and 2,4-dinitrophenyl (DNP)-MBP-rp17 was immobilized indirectly onto (anti-DNP) IgG-coated polystyrene beads. These directly and indirectly immobilized antigens were incubated with urine samples containing antibody IgG to p17 antigen and subsequently with rp17-β-D -galactosidase conjugate or (anti-human IgG γ-chain) Fab′-β-D -galactosidase conjugate. β-D -Galactosidase activity bound to the polystyrene beads was assayed by fluorometry. When rp17-β-D -galactosidase conjugate was used, signals (fluorescence intensities for bound β-D -galactosidase activity) were much higher with the indirectly immobilized antigens than those with the directly immobilized antigens. By experiments using (anti-human IgG γ-chain)Fab′-β-D -galactosidase conjugate, the binding of rp17-β-D -galactosidase conjugate to antibodies against p17 antigen bound to directly immobilized rp17 antigen was shown to be seriously limited as compared with that to antibodies against p17 antigen bound to indirectly immobilized DNP-MBP-rp17. When rp17-β-D -galactosidase conjugate and serum samples were used, serum interference was much less with indirectly immobilized DNP-MBP-rp17 than with directly immobilized rp17 antigen, and the sensitivity of enzyme immunoassay for antibody IgG to p17 antigen using indirectly immobilized DNP-MBP-rp17 was 1,000- to 3,000-fold higher than that of enzyme immunoassay using directly immobilized rp17 antigen and Western blotting for p17 band. This sensitive enzyme immunoassay indicated positivity in HIV-1 seroconversion serum panels as early as conventional methods for antibodies to HIV-1 and earlier than Western blotting for p17 band. J. Clin. Lab. Anal. 12:343–350, 1998. © 1998 Wiley-Liss, Inc.     

Automated quantification of thyrotropin by radial partition immunoassay

We describe a radial partition enzyme immunoassay in which fully automated quantification of human thyrotropin (hTSH) takes less than 11 min. This "sandwich"-type assay involves two monoclonal antibodies, both specific for the intact hTSH molecule. The solid phase consists of tabs of glass-fiber filter paper containing a pre-immobilized monoclonal anti-hTSH antibody complexed with a goat antibody specific for the Fc region of mouse IgG. The patient's sample is first applied to the central "reaction zone" of the tab, wherein hTSH binds to the immobilized antibody. Application of a buffered solution containing enzyme-labeled Fab' fragments of the second monoclonal anti-hTSH antibody initiates "sandwich" formation. A wash buffer containing a fluorogenic substrate elutes unbound conjugate to the tab periphery. The bound enzyme conjugate is quantified by measuring the rate of increase in fluorescence in the reaction zone of the tab, then converting the rate to clinical units by comparison with a stored calibration curve. The clinical utility and performance of the present assay compare favorably with those of other sensitive assays for hTSH.     

Substrate-labeled fluorescent immunoassay for phenytoin in human serum.

A homogeneous substrate-labeled fluorescent immunoassay has been applied to the measurement of phenytoin concentrations in human serum. We coupled a fluorogenic enzyme substrate, galactosyl-umbelliferone, covalently to a derivative of phenytoin. Under assay conditions, this drug-substrate conjugate was nonfluorescent but became fluorescent upon hydrolysis catalyzed by bacterial beta-galactosidase. When antibody to phenytoin is bound to the drug-substrate conjugate, it is inactive as an enzyme substrate. Addition of phenytoin to competitive-binding reactions relieves the inactivation, and the resulting fluorescence is proportional to the phenytoin concentration. We validated the fluorescent immunoassay by comparing values for phenytoin obtained with this technique to those obtained by gas chromatography and by enzyme immunoassay (EMIT). All three methods correlated well. The major metabolite of phenytoin, 5-(p-hydroxyphenyl)-5-phenylhydantoin, and other drugs at concentrations expected in serum had no effect on the assay. The fluorescent immunoassay is rapid and simple to perform and requires only 2 microL of serum sample per test.     

Automated latex nephelometric immunoassay of theophylline in human serum.

A latex test was adapted for the nephelometric quantitation of theophylline in human serum. The assay is fully automated on the Behring Nephelometer Analyser with a sampling rate of 150 samples per hour. There is no interference from bilirubin (up to 340 mumol/l), haemoglobin (up to 7000 mg/l), Intralipid (up to 5 g/l), or rheumatoid factor (up to 1100 x 10(3) IU/l). The theophylline standard curve extends from 1.25 to 40 mg/l. The coefficient of variation ranged from 2.4 to 7.4%. The correlation coefficient between the latex immunoassay and the TDx theophylline procedure was 0.988, calculated from the assay of 74 samples.     

A monoclonal-antibody enzyme immunoassay for detection of hepatitis B surface antigen with use of a biotin-avidin system

In this solid-phase two-site enzyme immunoassay for hepatitis B surface antigen (HBsAg), three monoclonal anti-HBs are used: 5D3 (IgM) is immobilized on plastic beads; 5C3 (IgG2a) and 5C 11 (IgG1), labeled with biotin, are used as the first conjugate. Horseradish peroxidase covalently linked to avidin is the second conjugate. First, serum or plasma is incubated with the antibody-coated bead and biotin-labeled antibodies, simultaneously, at 45 degrees C for 1 h ("stat" procedure), 3 h ("standard" procedure), or 18 h ("overnight" procedure), during which HBsAg forms a complex with the solid-phase antibody and the biotinylated antibodies. The enzyme-conjugated avidin is then bound to the biotin on the antigen-antibody complex at 45 degrees C for 15 min ("stat") or 30 min (standard and overnight procedures). The beads are incubated with enzyme-substrate solution (H2O2 and o-phenylenediamine). Color developed is measured at 492 nm. All procedures satisfied third-generation HBsAg tests required by the FDA Office of Biologics, being sensitive both to ad and ay subtypes in subnanogram amounts. The assay is reactive with adw2, adw4, adr, ayw2, ayw3, and ayr subtypes and can detect viral determinants in HBsAg-anti-HBs immune complex form. Thus it provides a sensitive, simple, and reproducible alternative to radioimmunoassay.     

Homogeneous luminescence immunoassay for total estrogens in urine

We describe an homogeneous luminescence immunoassay for "total" estrogens in enzymically hydrolyzed urine from nonpregnant women. The antiserum, raised against estriol-16,17- dihemisuccinate conjugated to bovine serum albumin, specifically bound the C-19 steroids carrying the estrogen-characteristic phenolic group. 17 beta-Estradiol conjugated with aminobutylethyl isoluminol was used to monitor the immunological reaction; this conjugate was stable for at least two years. Because binding to the antiserum markedly enhances the light-producing efficiency of the tracer, no separation of bound and free antigen is necessary. Results (microgram/24 h) by this method (y) correlated well (r = 0.958) with those by a conventional fluorometric (x) method (y = 2. 51x - 2.83). The sensitivity (detection limit) is 4 micrograms/L and the precision compares well with that of commonly used RIA methods. The method appears suited to large numbers of samples, as in menstrual cycle monitoring.     

Direct chemiluminescence immunoassay for estradiol in serum

In this simple, reliable, fast solid-phase chemiluminescence immunoassay for directly measuring (i.e., without prior extraction) estradiol-17 beta in serum, a monoclonal antibody is used that binds estradiol with high affinity (Ka = 10(10) L/mol), and does not bind other steroids tested, the highest cross reactivity observed being 0.1% for estradiol-17 alpha. In this system the monoclonal antibody is bound to the wells of microtiter plates via a second antibody directed against the monoclonal antibody. Fifty microliters of serum and estradiol-displacing agents are added, followed by 100 pg of estradiol-isoluminol conjugate, and the label is measured by luminometry after the binding reaction. The sensitivity of the assay is 180 pmol per liter of serum, and the effective working range at less than or equal to 10% CV is 270 to 6700 pmol/L. Analytical recovery of added estradiol averaged 99.7% (SD 6.5%). Within- and between-assay CVs ranged between 5 and 12.7%. Thirty-five unknown serum samples can be assayed within 4 h. Results correlated well with those obtained with a direct RIA: r = 0.94 (n = 149). This assay opens new perspectives for chemiluminescence immunoassays.