Rapid, automated assay for progesterone on the Abbott AxSYM(TM) analyzer

Department of Fertility, Pregnancy, and Neurodiagnostics, AxSYM R&D, Abbott Diagnostics Division, Abbott Laboratories, Abbott Park, IL 60064.
^a Author for correspondence. Fax (847) 938-7920; e-mail wilson{at}apmac.abbott.com.

	Abstract

Top Abstract Introduction Materials & Methods Results and Discussion References

 
We describe an automated assay for progesterone (P4) in human serum and plasma with the Abbott AxSYM^TM random-access immunoassay analyzer. In this one-step competitive assay, P4 immobilized onto latex microparticles competes with sample P4 for binding to a conjugate of alkaline phosphatase (AP) and anti-P4 antibody. Total CVs ranged from 3.4% to 8.2% in multiple precision studies conducted according to the 20-day NCCLS EP5-T protocol. The detection limit (zero calibrator + 2 SD) was 0.10 µg/L across 36 experiments. Values for diluted samples were 83–116% of expected. Recovery of P4 added to serum specimens was 92–115%. Cross-reactivities with 43 natural and synthetic steroids were 0–6.3%. No significant interference was detected from bilirubin, protein, erythrocytes, hemoglobin, triglycerides, or cholesterol. In a multisite correlation study, AxSYM P4 results compared well with results from a commercial RIA method (n = 1156; r = 0.976; slope = 1.03; y-intercept = 0.04). Assay throughput is >80 tests per hour in batch mode, 60 tests per hour with mixed load list configurations.

	Introduction

Top Abstract Introduction Materials & Methods Results and Discussion References

 
Serum progesterone (P4)¹ measurement is widely used to assess corpus luteum function in fertility studies, to evaluate placental function during pregnancy, and to aid in confirming ovulation (1)(2)(3). Measurement of serum P4 in the first trimester of gestation has been shown to be a reliable predictor and an effective tool for the diagnosis and treatment of patients with threatened abortion and ectopic pregnancy (4)(5). These results have encouraged clinicians to include a serum P4 measurement in evaluations of abnormal early pregnancies, in addition to physical examination, ultrasonography, and serial quantification of human chorionic gonadotropin ß-subunit (ß-hCG) (6). Recently, a single P4 measurement, with or without quantitative ß-hCG data, was shown to be highly predictive of pregnancy outcome within the first 8 weeks of gestation in asymptomatic women (7). At a 15 µg/L cutoff, the test predicted an abnormal pregnancy with a 75% sensitivity and 78% specificity, and P4 values 6 µg/L were associated with an 81% probability of pregnancy loss. These results support routine screening measurement of serum P4 in asymptomatic women with a prior first-trimester loss or ectopic pregnancy or in infertility patients who conceive without P4 supplementation, and support rapid P4 measurement in patients presenting with abnormal vaginal bleeding or pain.

With an increasing emphasis on P4 measurement in fertility assessment, there is an increasing need for a rapid, automated method for quantifying serum/plasma P4. Currently, P4 is most often quantified by RIA, although recently several automated nonisotopic assays have become available. Although highly sensitive, manual RIA methods have the obvious disadvantages of labor intensiveness, time required for results, and disposal of isotopes. On the other hand, the current nonisotopic assays have one or more of the following disadvantages: low throughput, low sensitivity, imprecision, inaccuracy, lack of specificity, or lack of random-access capability. The object of our studies was to develop a rapid, highly precise automated assay for P4 that could be used in random, continuous-access mode on the Abbott AxSYM^TM immunoassay analyzer.

	Materials & Methods

Top Abstract Introduction Materials & Methods Results and Discussion References

 
materials
Protein A was purchased from Pharmacia; microparticles were from Saradyn; P4 was standards-grade material from Sigma. Escherichia coli alkaline phosphatase (AP; EC 3.1.3.1) was cloned and purified at Abbott. Anti-P4 sheep monoclonal antibody was from a cell line obtained from the University of Surrey, Guildford, UK. All other materials were of the highest analytical grade obtainable.

For method comparison studies, P4 was quantified by the DPC Coat-A-Count and DPC Immulite Progesterone assays (Diagnostics Products Corp.), the ACS:180 Progesterone assay (Ciba Corning), the TOSOH Progesterone assay (TOSOH Medics), the ES 300 Progesterone assay, and the SR1 Progesterone assay (Biochem Immunosystems, Allentown, PA).

apparatus
The AxSYM is an automated, random-access immunoassay analyzer that utilizes fluorescence polarization immunoassay, microparticle enzyme immunoassay (MEIA), and ion-capture technologies. Details of the instrument are given elsewhere (8)

reagents
Microparticle reagent.
Anti-fluoroscein mouse monoclonal antibody was cloned at Abbott and grown in a medium containing Fetal Clone Serum, 50 mL/L. The antibody was purified from tissue culture fluid by ammonium sulfate precipitation, followed by dialysis and fractionation by Protein A chromatography. Protein A-purified material (0.5 g/L) was covalently coupled to 0.164-µm-diameter carboxylated microparticles (5 g/L) with 1 g/L of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide at pH 5.5. After coupling, the microparticles were overcoated with a 15 g/L solution of bovine serum albumin to block potential nonspecific protein binding sites, and suspended to a concentration of ~0.3 g/L solids in a Tris buffer diluent (pH 7.2) containing sucrose (180 g/L) for buoyancy and mouse serum (5 g/L) to block potential interfering interactions in specimens containing human anti-mouse antibody. Fluorescein–P4–bihapten was then added to the microparticle suspension to give a concentration of 3 nmol/L. This step results in a suspension of microparticles presenting P4 antigen for competition with specimen P4 for anti-P4 binding with the conjugate reagent. The concentrations of microparticles and P4–bihapten were formulated to give optimal assay detection limit and stability of reagent.

Conjugate reagent.
Anti-P4 sheep monoclonal antibody was from a cell line isolated from fusion products between mononuclear cells from a sheep immunized with P4-11–hemisuccinate–ovalbumin and NS1 mouse myeloma cells. The cell line (O/MP.1A9.D7B2) produced a high affinity (K_d = 4.8 x 10^-12) IgG₁ with a high specificity for P4 (9). The cell line was grown in tissue culture and purified by ammonium sulfate precipitation followed by Protein A chromatography. The purified antibody (1 g/L) was covalently coupled at pH 7.0 to AP (2 g/L) by use of a 60-fold excess of a 30-atom heterobifunctional linker molecule based on a succinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carbonate parent compound in which the terminal succinimide groups are separated by a 30-atom spacer arm. This step produced an anti-P4:AP conjugate. After the coupling, the reaction was capped with an excess of N-ethylmaleimide, and then was diluted in a buffer reagent of NaCl, MgCl₂, ZnCl₂, casein, sheep serum, azide, and Tris at pH 7.5.

P4 buffer.
This reagent was formulated to competitively block potential nonspecific binding (NSB) interactions between antiserum components present in sample and the microparticle or conjugate reagents. The formulation includes purified mouse IgG to block potential NSB with the microparticle antibodies, bovine gamma globulin and bovine serum albumin to block potential NSB with bovine components coated onto the microparticles, and E. coli AP to block potential NSB with the conjugate.

Calibrators.
AxSYM Progesterone is calibrated with spectrophotometrically determined concentrations of P4 added to charcoal-stripped normal human serum. Because P4 in this matrix was found to be unstable to storage at 2–8 °C, the calibrators and controls provided with the kit contain P4 in a Tris buffer matrix containing 7ß-cyclodextran as a stabilizer. This cyclic polymer of glucose has a hydrophobic cavity that confers protection to P4 as an inclusion complex. With this novel stabilizer, kit calibrators are stable to continuous 37 °C heat stress for at least 1 month. These stabilized calibrators give dose–response curves identical to those of the serum-based comparison calibrators and allow convenient nonfrozen storage and handling.

assay protocol
In this one-step competitive assay, the AxSYM instrument performs the assay as follows: the AxSYM pipettor/electrode assembly combines sample, anti-P4:AP conjugate, and the P4 buffer in a well of the AxSYM reaction vessel, forming a complex of P4 with the anti-P4 conjugate. This reaction mixture is then combined with the P4 microparticles reagent (microparticles coated with a complex of anti-fluorescein antibody bound to a fluorescein–P4–bihapten tracer). The microparticles bind to any anti-P4 conjugate not bound to P4 from the sample, forming the final reaction mixture (see Fig. 1 ). After a 10-min incubation, an aliquot of the reaction mixture is transferred to the glass fiber matrix of the AxSYM Matrix Cell, where the microparticle:bihapten:conjugate complexes are captured through hydrophobic interactions between the microparticles and the glass fibers of the matrix. Following a wash step, methylumbelliferyl phosphate (MUP) is applied to the matrix, and the rate of dephosphorylation of MUP by bound AP is measured by the AxSYM MEIA optical assembly.

View larger version (37K): [in this window] [in a new window]   Figure 1. Assay principle of AxSYM Progesterone. (A) P4 presented on a solid phase (composed of latex microparticles coated with a complex of anti-fluorescein antibody bound to a fluorescein–P4 bihapten tracer) competes with sample P4 for binding to a conjugate of AP and anti-P4 antibody. (B) Reaction mixture is transferred to the glass fiber matrix of the AxSYM Matrix Cell, where the microparticle–bihapten–conjugate complexes are captured through hydrophobic interactions between the microparticles and the glass fibers of the matrix. MUP is applied to the matrix, and the rate of dephosphorylation of MUP by bound AP is measured.

	Results and Discussion

Top Abstract Introduction Materials & Methods Results and Discussion References

 
calibration curve
Figure 2 depicts a representative calibration curve showing the relation between the rate of production of fluorescent product and the P4 concentration. The bihapten competitive inhibition format of the assay results in a highly sensitive response to low concentrations of P4, which begins to level off at ~7 µg/L. Between 7 and 40 µg/L, the difference in signal is relatively small, corresponding to <10% of the total curve span. This steep displacement between 0 and 7 µg/L is necessary to achieve a detection limit within 0.2 µg/L, but poses some difficulty in achieving accurate curve-fitting through the assay range, particularly in the region of this "elbow." To minimize fitting error, we examined numerous combinations of calibrator concentrations and fitting routines. Calibrator concentrations of 0, 0.7, 2.0, 7.0, 20, and 40 µg/L, together with weighted four-parameter logistical curve (4PLC) fitting (10), gave the best fitting accuracy throughout the assay range. Fig. 2 (inset) depicts mean fitting residuals at each calibrator concentration across 12 separate experiments. The maximum positive fitting bias (16%) was observed between ~1 and 10 µg/L, and a slight negative bias (6%) was observed in the midregion of the assay range. The positive fitting bias is expected to contribute to slight inaccuracy through this region of the curve, but this is not clinically significant because it corresponds to a theoretical error of only 0.3 µg/L in a nondiagnostic region of the assay.

View larger version (24K): [in this window] [in a new window]   Figure 2. Calibration curve for AxSYM Progesterone. AxSYM Progesterone calibrators were assayed in duplicate in the AxSYM instrument. The P4 concentration is calculated from the equation [P4] = exp{ln[B/(fluorescent rate - A) - C]/D}. The best fit was defined by the following parameter values: A = 23.00, B = 1158.64, C = 0.988, D = 1.469. Inset: Mean 4PLC fitting error at each calibrator value. The data represent the mean percent error of the calibrator concentrations from their fitted values as obtained in 12 separate experiments (see text).

detection limit
The detection limit, defined as 2 SD from the mean of the zero calibrator (replicates of 10), was characterized from three runs on each of four AxSYM instruments across three separate reagent lots on different days (n = 36 runs). As calculated from the 4PLC fit of the calibrator response, the grand mean detection limit obtained was 0.10 µg/L, with an upper 95% confidence value of 0.15 µg/L. Five serum specimens diluted with the zero calibrator exhibited linear values down to the detection limit of the assay (Fig. 3 ).

View larger version (30K): [in this window] [in a new window]   Figure 3. Serial dilution of serum specimens in zero calibrator. Serial dilutions were performed on each of five serum specimens until P4 concentrations reached the limit of detection of the assay (0.1 µg/L). Average dilution recoveries across all dilutions (to 1:256-fold) for each specimen ranged from 83% to 116%, with a grand mean across all specimens and dilutions of 104%. (- - -), ±20% tolerance limits for reference.

specificity
Cross-reactivity with 1000 µg/L of the steroids listed in Table 1 was <7% across all compounds tested, with most compounds showing no detectable cross-reactivity. These naturally occurring and drug steroids do not interfere with the ability of the assay to measure P4 accurately.

recovery and interfering substances
The mean analytical recovery of from 4.84 to 13.22 µg/L of added P4 in 10 serum specimens with unsupplemented values of 0.30–9.27 µg/L was 100.1% (range 92.1–115.1%). In serum specimens separately supplemented with as much as 0.2 g/L bilirubin, 10 g/L hemoglobin, 20 g/L triglycerides, and 8 g/L erythrocytes, the measured P4 was 93.0–103.5% of expected values. P4 values obtained from split samples collected in serum separator, heparin plasma, EDTA plasma, or "red top" clotting tubes exhibited no significant differences (details not shown).

expected ranges
Serum specimens were drawn from 77 presumably healthy men, 30 postmenopausal women, daily from 30 women with normal ovarian cycling, and from 25 women each in the first, second, or third trimester of pregnancy. For this study, the follicular phase was defined as the time from 10 to 5 days before the day in which the concentrations of luteinizing hormone (LH) and follicle-stimulating hormone (FSH) were highest. The luteal phase was defined as the time from 4 to 10 days after the day on which LH and FSH were highest. The mid-luteal phase was defined as the time from 5 to 9 days after the day on which LH and FSH peaked (11). A summary of the AxSYM Progesterone data obtained on these specimens is given in Table 2 . These data are in good concordance with previously published expected P4 values based on RIA methodology (12).

precision
Precision was determined as described in NCCLS Protocol EP5-T2 (13). A four-member panel was assayed on each of four separate AxSYM instruments, with use of a single reagent lot and a single calibration, in replicates of two at two separate times per day for 20 days (n = 80 replicates per panel member per instrument). Table 3 shows the range of control concentrations and precision data observed for each AxSYM, and the mean concentrations and precision data observed across all AxSYM instruments. The mean data represent typical expected precision performance for the AxSYM Progesterone assay. As shown in Table 3 , total CVs were <10% for all panel members, including a very low serum-based panel member for which the mean concentration was 0.61 µg/L.

comparison with other methods
AxSYM vs DPC Coat-A-Count.
A large method comparison study was conducted with AxSYM Progesterone and the DPC Coat-A-Count RIA across multiple reagent lots and clinical sites. The following results were obtained for 1156 clinical specimens: AxSYM = 1.03(Coat-A-Count) 0.04 (r = 0.976, S_y|x = 0.68%). The data shown in Fig. 4 represent singleton measurements from the AxSYM assay and are a compilation of data obtained from five separate laboratories² that used three different AxSYM reagent lots. Duplicate AxSYM measurements were made on a majority of the specimens for purposes of assessing within-run precision; correlation of the first result to the second gave the following statistics: Result 1 = 1.00(Result 2) 0.01 (r = 0.997, S_y|x = 0.02%, n = 1107; see Fig. 4 inset). Comparison of the data between the two correlation plots shown in Fig. 4 indicates that the rather large data spread between the two assay methods in the upper half of the dynamic range is not the result of imprecision in the AxSYM assay. Imprecision of the DPC method in this region of the assay range was not investigated. No significant differences between the AxSYM reagent lots or instruments were noted (not shown).

View larger version (39K): [in this window] [in a new window]   Figure 4. Comparison of AxSYM Progesterone and Coat-A-Count assays. Data depicted are singleton measurements from both assays, and represent a compilation of data obtained across five clinical sites where three different AxSYM reagent lots were used. AxSYM values ranged from 0.20 to 39.9 µg/L. Inset: Correlation between AxSYM replicates on subset of specimens tested in the between-method comparison. The data show good between-replicate precision by AxSYM in the high region of the dynamic range.

AxSYM vs other methods.
Smaller comparison studies were also conducted with five other commercial P4 assays. A summary of the results from these studies is given in Table 4 . In general, there was good correlation between the AxSYM Progesterone assay and all other methods tested (r 0.95). In all cases, the intercepts were <1 µg/L. The variation in slopes (0.83–1.09) appears to reflect differences in calibration.

throughput
Assay throughput was determined for batch and mixed load lists by using an AxSYM scheduler simulator, which estimates throughput of a given assay run together with a panel of other AxSYM assays. The assays specified in the mixed load list simulations were chosen as representative of mixed load list runs in a typical clinical setting. In worst-case mixed load list modeling scenarios for throughput, the AxSYM Progesterone assay gave a throughput of 57 tests per hour. With a batch load list (100% P4), the simulated throughput was 83 tests per hour.

In conclusion, our results show that the AxSYM Progesterone assay is sensitive, specific, and highly precise and gives results that correlate and agree well with those by a manual RIA method. These factors, in conjunction with high throughput and the random, continuous-access automation of the AxSYM instrument, should make the method useful for fertility testing laboratories.

	Footnotes

 
¹ Nonstandard abbreviations: P4, progesterone; AP, alkaline phosphatase; MEIA, microparticle enzyme immunoassay; NSB, nonspecific binding; DPC, Diagnostic Products Corp.; MUP, methylumbelliferyl phosphate; 4PLC, four-parameter logistical curve; LH, luteinizing hormone; FSH, follicle-stimulating hormone.

² Baptist Regional Labs, Memphis, TN; Eastern Virginia Medical School, Norfolk, VA; Johns Hopkins Hospital, Baltimore, MD; University of Nebraska Medical Center, Omaha, NE; University of Maryland, Baltimore, MD; and University of Tennessee, Knoxville, TN.

	References

Top Abstract Introduction Materials & Methods Results and Discussion References

 

1. Buster JE, Abraham GE. The application of steroid hormone radioimmunoassays to clinical obstetrics. Obstet Gynecol 1975;46:489-499. [Web of Science][Medline] [Order article via Infotrieve]
2. Israel R, Mishell DR, Jr, Stone SC, Thorneycroft IH, Moyer DL. Single luteal phase serum progesterone assay as an indicator of ovulation. Am J Obstet Gynecol 1972;112:1043-1046. [Web of Science][Medline] [Order article via Infotrieve]
3. Abdulla U. Plasma progesterone levels as an index of ovulation. Br J Obstet Gynaecol 1983;90:543-547. [Web of Science][Medline] [Order article via Infotrieve]
4. Witt BR, Wolf GC, Wainright CJ, Johnston PD, Thorneycroft IH. Relaxin, CA-125, progesterone, estradiol, Scheangerschaft protein, human chorionic gonadotropin as predictors of outcome in threatened and nonthreatened pregnancies. Fertil Steril 1990;53:1029-1036. [Web of Science][Medline] [Order article via Infotrieve]
5. Matthew CP, Coulson PB, Wild WA. Serum progesterone levels as an aid in the diagnosis of ectopic pregnancy. Obstet Gynecol 1986;68:390-394. [Web of Science][Medline] [Order article via Infotrieve]
6. Buck RG, Jonbert SM, Norman RJ. Serum progesterone in the diagnosis of ectopic pregnancy: a valuable diagnostic test?. Fertil Steril 1988;50:752-755. [Web of Science][Medline] [Order article via Infotrieve]
7. Daily CA, Laurent LS, Nunley WC, Jr. The prognostic value of serum progesterone and quantitive ß-human chorionic gonadotropin in early human pregnancy. Am J Obstet Gynecol 1994;171:380-384. [Web of Science][Medline] [Order article via Infotrieve]
8. Smith J, Osikowicz G, et al. , Abbott AxSYM^TM random and continuous access immunoassay system for improved workflow in the clinical laboratory. Clin Chem 1993;39:2063-2069. [Abstract]
9. Groves DJ, Sauer MJ, Rayment P, Foulkes JA, Morris BA. The preparation of an ovine monoclonal antibody to progesterone. J Endocrinol 1990;126:217-222. [Abstract/Free Full Text]
10. Rodbard D, Hutt DM. Statistical analysis of radioimmunoassays and immunoradiometric (labeled antibody) assays. A generalized, weighted, iterative least squares method for logistic curve fitting. In: Proc., Symp. on radioimmunoassay and related procedures in medicine. Vienna, Austria: Int Atomic Energy Agency (and New York: Unipub), 1974:165–92..
11. Crowley WH, Jr, Filicori M, Santoro NF. GnRH secretion across the normal menstrual cycle. Crowley WF, Jr Hofler JG eds. The episodic secretions of hormones 1987:219-231 John Wiley and Sons New York. .
12. Carr BR Blackwell RE eds. Textbook of reproductive medicine 1993:151 Applegate and Lange Norwalk, CT. .
13. National Committee for Clinical Laboratory Standards. Evaluation of precision performance of clinical chemistry devices, 2nd ed.; tentative guideline. NCCLS Document EP5–T2. Wayne, PA: NCCLS, 1992..