Evaluation of a Quantitative Immunoturbidimetric Assay for Rheumatoid Factors

Transcription

1 LIN. HEM. 32/10, (1986) Evaluation of a Quantitative Immunoturbidimetric Assay for Rheumatoid Factors Lilsa M. Melamles,1 Helka M. Ruutsalo,2 and Harttl NlsslI#{228}2 A quantitative immunoturbidimetric assay for rheumatoid factors (RF) is described, based on the immunoprecipitation between aggregated human lgg and rheumatoid factors in serum. The resulting turbidity is measured photometrically at 340 nm. The method is standardized against the WHO international reference preparation and the results are expressed in International units per milliliter (mt. units/ml). Results correlate well with those by different latex-agglutination techniques (r = ). The correlation with Waaler-Rose test modifications were 0.75 and The within-run and between-run coefficients of variations were respectively from 1.2 to 2.6% and 1.3 to 1.9% for high and low RF concentrations. Quantitative and reproducible results, together with high throughput of samples and compatibility with most clinical chemistry analyzers and photometers, make this new assay well suited for routine screening and monitoring of rheumatoid factors. Additional Keyphrases: rheumatoid arthritis autoantibodies latex agglutination compared Rhematoid factors (RF) are autoantibodies that react with the Fe portion of human IgG. The presence of HF in rheumatoid arthritis (RA) gives diagnostic and prognostic information (1,2) and can be regarded as a measure of the inflammatory process. Traditionally RF is detected by agglutination reactions involving human IgG-coated latex particles (3, 4) or rabbit IgG-coated sheep cells (5, 6). In particulate or multivalent form the reactant reacts predominantly with RF of the 1gM class. Problems related to these serological methods include great variation in titers with different applications, the lack of standardization, and the semi-quantitative or even qualitative interpretation of the results. Quantitative techniques more recently introduced are: photometric latex agglutination (7,8), nephelometric assays (9,10), enzyme immunoassays (11,12), radioimmunoassays (13, 14), and one turbidimetric method in which a centrifugal analyzer is used (15). Enzyme imniunoassay and radioimmunoassay have the advantage of differentiating between IgA-, IgG-, and 1gM-class RFs, but are rather laborious for routine use. Here we describe our evaluation of a quantitative immunoturbidimetric assay for RF in which commercially available reagents are used. The immunoprecipitation occurs in the liquid phase between the RF of the sample and heataggregated human IgG. The turbidity can be measured with routinely available photometers at 340 nm, and the results Orion Diagnostica, Research and Development Department, P.O. Box 83, Espoo, Finland. Address correspondence to: L. M. M., at Pursimiehenkatu 12 B 26, Helsinki, Finland. 2Rheumatism Foundation Hospital, Heinola, Finland. Received February 19, 1986; accepted July 15, are expressed in international units. Results are compared with those by three different latex agglutination techniques and with two Waaler-Rose test modifications. Materials and Methods Patients samples. Patients samples were obtained from the Rheumatism Foundation Hospital, Heinola, Finland, and from the Laboratory of linical Immunology, Helsinki, Finland. Immunoturbidimetry. The RF assay kit (Orion Diagnostica, Espoo, Finland) contains ph 8.5 buffer (per liter, 25 mmol of borate, 50 mmol of Na1, 1 g of bovine serum albumin, 11 g of polyethylene glycol 6000), lyophilized RF reagent (human IgG, 50 g/l, aggregated at 63#{176}, diluted to a final concentration of 5 g/l), the reference diluent (HFnegative serum), and HF reference (HF-positive serum) calibrated against the HF reference preparation of the World Health Organization. The instruments used were the FP 9 (Labsystems, Helsinki, Finland) and the Kone (Kone Instrument Division, Espoo, Finland) spectrophotometers. The procedures were as follows: The serum specimens were heated at 56#{176} for 30 mm to destroy the complement. Highly lipemic samples should be clarified with Lipoclean#{174} (Behring Institute, Marburg-Lahn, F.RG.). The HF reference was diluted with the reference diluent to obtain HF concentrations of 12.5, 25, 50, and int. units/ml. For use in the FP 9 we pipetted 350,aL of buffer and il of reference/sample into cuvets, blanked the photometer, and started the reaction by adding 50,uL of HF reagent. Absorbances were measured 3 mm later and the results were calculated manually with use of a reference curve. For the Kone the measuring range was adjusted from 50 to 400 int. units/ml by diminishing the sample volume to 25 jzl but still using L of reference volumes. In the sample and reference cuvets the respective buffer volumes were thus 475 and 400 L for separate blanks, and 425 and 350 il for tests. The reaction was started by adding 50 jl of HF reagent to the test cuvets, and absorbances were measured at 3 mm. Results were automatically calculated by the instrument by using spline function. If the result exceeded the highest reference, the sample was diluted with the HF buffer and the measurement repeated. For the Kone, we retested the sample by using jzl of sample if the result was <50 int. units/ml. Agglutination tests. Latex particles coated with human IgG (RapiTex#{174}HF), rabbit y-globulin-coated sheep erythrocytes (ellognost#{174} HF), and Rheumatoid Factors Reference Serum were obtained from the Behring Institute. RapiTex latex slide and tube tests were performed according to the manufacturer s instructions, and with use of a microtiter plate application (16). Waaler-Rose tests were performed by a microtechnique (16) and the ellognost#{174}hf, hemagglutination slide test was done according to the manufacturer s 1890 LINIAL HEMISTRY, Vol. 32, No. 10, 1986

2 instructions. All samples were heated at 56#{176} for 30 mm. Results Reference curve. A typical non-linear reference curve for HF prepared by use of the FP 9 is shown in Figure 1. The assay range is from 12.5 to int. units/ml. Reproducibility. For precision studies we used the FP 9 and two pooled sera. Based on 24 replicate determinations with the use of one calibration, curve, the V ranged from 1.2 to 2.6% for mean (SD) HF concentrations of 79.2 (0.69) and 25.3 (0.68) mt. units/rnl. With use of different calibration curves and five freshly prepared calibrator sets, the between-run V of 11 consecutive determinations ranged from 1.3 to 1.9% for the mean (SD) HF concentrations of 74.6 (0.97) and 24.5 (0.47) int. units/ml. Analytical recovery. We added 30 and 75 int. units of HF per milliliter to HF-negative serum pools. The concentrations measured with the FP 9 for four replicate determinations were 28.8 and 76.8 mt. units/ml (respective recoveries of 95.8 and 102.5%). Linearity. We tested the linearity of the method with the FP 9 by serially diluting, with the HF buffer, 16 patients samples containing high concentrations of HF. Table 1 25O a RF,mt.unfts,mL Ag. 1. Reference curve for RF by immunoturbidimetry shows an example of dilution linearity. We observed no significant difference in results for samples diluted with buffer or with reference diluent. It is more convenient to use the HF buffer for dilutions, because the kit contains a surplus of it. orrelation with other methods. The method was compared with the latex tube and slide agglutination methods (85 samples) and with the latex microtiter method (total of 527 samples analyzed in two different laboratories). The regression data are shown in Table 2. The results of all latex tests are expressed as titers, but for slide and tube tests also as int. units/ml, according to the Bebring HF reference serum analyzed in the same way. The correlations between the immunoturbidimetric test vs RapiTex latex slide test, Waaler-Rose microtiter plate test, and ellognost hemagglutination test are shown in Figure 2. The correspondence between int. units and titers for the immunoturbidimetric test and different latex test applications was calculated as shown in Table 3. linical material. The comparative study involved a total of 387 patients in the Heinola Rheumatism Foundation Hospital. HF activity was analyzed by the immunoturbidimetric (the Kone ) and latex microtiter methods. The patients were classified into four groups according to diagnosis: (a) HA (n = 255); (b) juvenile HA (n = 53); (c) other inflammatory rheumatic diseases (n = 47), including ankylosing spondylitis, lupus erythematosus disseminatus, mixed connective tissue disease, scieroderma, polymyositis, dermatomyositms, SjOgren s syndrome, arthritis psoriatica, reactive arthritis; and (d) noninflammatory rheumatic diseases (n = 32), including arthrosis and soft-tissue rheumatism. The results (Table 4 and Figure 3) show a high degree of correspondence with regard to the number of positive cases at different HF values in all groups of patients. With 25 mt. units/ml as the upper limit of normal, 78% and 84% of the cases in the HA group were identified by the immunoturbidimetric and latex tests, respectively. Discussion This method for measuring HF immunoturbidimetrically can be used with a commercially available HF kit. The results are expressed in int. units/ml, in comparison with a secondary standard calibrated against the wio reference preparation. The HF measured is probably 1gM, but we did not study whether or to what extent IgG and IgA HF might be involved (study in preparation); hence, we refer to an assay for rheumatoid factors. The measuring range is somewhat limited, owing to the different kinetics involved for low vs high HF concentrations. However, the range can be adjusted according to the needs of different types of laboratories, depending on the Table 1. LinearIty of the Theoretical immunoturbidimetrlc RF Assay Measured Table 2. orrelations between lmmunoturbidlmetrlc and Latex Agglutination Tests for RF Dilution Undiluted 1/2 1/4 1/8 1/16 1/32 1/64. RF, mt. unlts/ml AF, mt. unlts/ml % of theorstical g - latex method, x RegressIon equation n r Microdter ya = 1.06x #{247} Microtiter y = 1.33x Behring slideb y = 3.49x Behring slidec y = 1.4x Behrlng slidel y = 0.72x Behrlng tubec y = 1.15x / ay = immunoturbidimetric method = the Kone ; all others, the FP 9. 1/ bner lot units/ml LINIAL HEMISTRY, Vol. 32, No. 10,

3 -. 0 #{149} 1760 ni,it,il I00o I A 0 l0 I. I o ISO ist 20 5Oo j 320 OF LATEX TITER OF. WAALER-R056 TITER OF. HEMAGOLUTINATION TITER Fig. 2. orrelation between (left) immunoturbidimetric and RapiTex slide agglutination tests (r = 0.96, n = 85); (middle) immunoturbidimetiic and Waaler-Rose tests (r = 0.75, n = 83); (right) immunoturbidimetric and ellognost-hemagglutination tests for RF (r = 0.92, n = 79) Left: uppermost points have 1400 and 2 mt. units/ml; right, uppermost point 1760 nt. units/ml Table 3. orrespondence of the Results by Quantitative Immunoturbidimetric RF Assay (it) and by the RF Latex Titration Tests RapiTex slide, IT, mt. Mlcrotfter method, IT, nt RapiTex tube, IT, mt. titer unlts/ml titer unlts/ml titer units/ml > Table 4. omparison between the immunoturbidimetric and Latex Microtiter Methods in Different Patient Groups PatIent groups (and total no. In each) nt. units/ml Immunoturbidimetric Latex microtiter No prozone was seen under these reaction conditions. All samples with a high HF concentration gave absorbance readings clearly exceeding that of the mt. units/ml reference. The highest HF concentration measured was about 6000 int. units/ml. Some very high values might be A (255) B (53) (47) D (32) missed because of falsely low absorbance readings, however, if either the sample volume (vs HF reagent volume) or the no. % no. % no. % no. % reaction time is considerably increased. We suggest decreas the sample volume if a prolonged reaction time is desired The excellent reproducibility of results read by an instru ment and the expression of the results in quantitative international units are an advantage over traditional agglutination techniques, where the subjective interpretation of titers causes problems with accuracy and precision. More over, variability between laboratories and methods (17-21) may be reduced by converting the titers to int. units by using standards calibrated against the wito international reference preparation, to achieve uniform and comparable results in HF assays (18, 19,21,22). At present there is no agreement concerning normal values of HF, the cutoff value for normal being suggested to lie between 12.5 (22) and 50 (19) int. urnts/ml. If 25 flit. units/ml is used as the upper normal limit, the immunoturbidimetric test is most readily compared with the latex test. A, rheumatoid arthritis; B, juvenile rheumatoid arthritis;. other inflammatory rheumatic diseases; 0, noninflammatory rheumatic diseases. patient population, by changing the sample volume but not the reference volume. With 10 jil of patient s sample the range will be int. units/ml. Because the dilution of samples with reaction buffer caused no problems with linearity (Table 1), the total reaction volume can be maintained by adding buffer LINIAL HEMISTRY, Vol. 32, No. 10, 1986

4 I E E 200 -,. A B D..JI... $IIUIIIII IliuiflIT and rapid, the method is adaptable to most clinical chemis- try analyzers and photometers. Any laboratory provided with a photometer can perform the test either manually or,, adapted to different levels of automation. For the FP 9, for instance, the throughput is 120 samples per hour and calculation of the results is automatic. A patient s HF concentration not only is of diagnostic importance but also may have a prognostic value in estimating the severity of the rheumatoid disease; if so, HF may be used as an indicator of the disease activity (2,23,24). This quantitative method with its good precision is well suited for follow-up studies with patients, whereas the semiquantitative titer techniques are too inaccurate for this purpose Illililifi Illm IIm: 128 In conclusion, the immunoturbidimetric HF method evaluated is very simple, gives quantitative results expressed in hit. units/ml, and is highly reproducible. The immunoturbi- 64 dimetric HF reagents are stable, the labile aggregated IgG being supplied in lyophilized form. Thus we find this HF assay is much more reliable than the traditional serological 32 titration techniques II-- aiiii17.1_ In Fig. 3. ompanson between immunoturbidimetric (#{149}, left-hand syinbols) and latex microtiter (0, right-hand symbols) methods in different patient groups: A, rheumatoid arthritis; B, juvenile rheumatoid arthritis;, other inflammatory rheumatic diseases; 0, noninflammatory rheumatic diseases Determination of HF activity by the immunoturbidimetric method in 300 pregnant women showed 98.7% of them to have HF below the 25 mt. units/ml limit. Results of the immunoturbidimetric method correlated very well with those of the latex agglutination tests-as expected, considering the common chemistry of these tests, the reaction between HF and human IgG. The comparison with the less-sensitive Waaler-Rose test gave a fairly good correlation (microtechnique), but for several samples, results correlated poorly. At the Laboratory of linical Immunology 17 of these samples were especially selected to demonstrate an exceptional patient material with a poor correlation between the latex and Waaler-Rose tests, and to find out the results of measuring these samples immunoturbidimetrically. Where these discrepancies existed, the immunoturbidimetric and latex agglutination methods were in accordance. In all cases shown in the middle panel of Figure 2, in which the results lie on the x- or y-axis, indicating poorly correlated values, the latex results correlated well with the immunoturbidimetric method and poorly with the Waaler-Rose titers. At higher HF values the correlation between immunoturbidimetry and Waaler-Rose was better. Results by the ellognost hemagglutination test, a modification of the Waaler-Rose test, correlated well with those by the immunoturbidimetric test. Samples from another group of patients were used in this test. The results on theyaxis in Figure 2, right panel, probably demonstrate activities of HF having no cross reaction with rabbit IgG. The quality of the Waaler-Rose reagents is not readily maintained, which may influence the precision and titer level of the test. Because the mmmunoturbidimetric assay is very simple We thank the Laboratory of linical Immunology for providing us with sera analyzed by latex- and Waxier-Rose microtiter methods. The technical assistance of Mrs. Maijatta Lehtinen and Mrs. Taia Nordenstedt is gratefully acknowledged. References 1. Ropes MW, Bennet GA, obb S, Jacox R, Jessar RA. Revision of diagnostic criteria for rheumatoid arthritis. Bull Rheum Dis 1958;9: Wager 0. Immunological diagnosis of rheumatoid arthritis and systemic lupus erythematosus. Ann lin Res 1975;7: Singer JM, Plotz M. The latex fixation test. I. Application to the serologic diagnosis of rheumatoid arthritis. Am J Med : Singer JM, Edberg S, Markowitz RL, Glickman JD, Miller L, Marchiteffi R Performance of latex fixation kits used for serological diagnosis of rheumatoid factor in rheumatoid arthritis sera. Am J lin Pathol 1979;72: Rose HM, Ragan, Pearce E, Lipman MO. Differential agglutination of normal and sensitized sheep erythrocytes by sera of patients with rheumatoid arthritis. Proc Soc Exp Biol Med 1948;68: atchart EW, O Sullivan JB. Standardization of the sheep cell agglutination test. Arthritis Rheum 1965;8: Mathies H, Helming V. Erhohung der Trefferquote des Latexuixationtestes durch photometrische Bestimmung des Agglutinationsgrades. Kim Wochenschr 1965;43: Mimi V, Dezeli#{233} G, Dezeli N, Richter B. A photometric latex test for rheumatoid factors. Scand J Rheumatol 1972;1: Jones E, Rousseau RJ, Maxwell KW. Quantitation of rheumatoid factor activity by nephelometry. Am J lin Pathol 1979;72: Painter P, Lyon JM, Evans JH, Powers WW, Whittaker EL, Decker MJ. Performance of a new rate-nephelometric assay for rheumatoid factor, and its correlation with tube-titer results for human sera and synovial fluid. lin hem 1982;28: Gripenberg M, Walln F, Isomaki H, Linder E. A simple enzyme inununoassay for the demonstration of rheumatoid factor. J Immunol Methods 1979;31: Tarkowski A, Nilsson L-A. Isotype-specific measurement of rheumatoid factor with reference to clinical features of rheumatoid arthritis. J lin Lab Immunol 1983;12: arson DA, Lawrence S, atalano MA, Vaughan JH, Abraham G. Radioimmunoassay of IgG and 1gM rheumatoid factors reacting with human IgG. J Immunol 1977;119: Koopman WJ, Schrohenloher RE, Salomon A. A quantitative assay for IgA rheumatoid factor. J Immunol Methods 1982;50: LINIAL HEMISTRY, Vol. 32, No. 10,

5 15. Borque L, Yago M, Mar, Rodriguez. Turbidimetry of rheumatoid factor in serum with a centrifugal analyzer. lin hem 1986;32: PyrhOnen 5, Timonen I Heikkinen A, et al. Rheumatoid factor as an indicator of serum blocking activity and tumor recurrences in bladder tumors. Eur J ancer 1976;12: Anderson SG, ooper S. The effect of two factors, source of sheep cells and concentration of amboceptor, on the estimated titre of rheumatoid factor. Bull wuo 1970;43: Taylor RN, Fulford KM, Joies WL. Reduction of variation in results of rheumatoid factor tests by the use of a serum reference preparation. J lin Microbiol 1977;5: Fulford KM. Taylor RN, Przybyszewski VA. Reference preparation to standardize results of serological tests for rheumatoid factor. J lin Microbiol 1978;7: Singer JM, Edberg S, Selinger M, Amram M. Quality control of the latex-fixation test. Am J lin Pathol 1979;72: Astle L. urrent problems in laboratory tests for rheumatoid factor. In: Rippey JH, Nakamura RM, eds. Diagnostic immunology: technology assessment and quality assurance. Skokie, IL: ollege of American Pathologists, 1983: Anderson SG, Benzon MW, Houba V, Krag P. International reference preparation of rheumatoid arthritis serum. Bull wao 1970;42: ats A, Klein F. Quantitative aspects of the latex-fixation and Waxier-Rose tests. Ann Rheum Dis 1970;29: Dezeli G, Durrigl T, Dezeli N, et al. The photometric latex test for rheumatoid factors in patients with rheumatoid arthritis. II. linical evaluation. Z Rheumatol 1978;37: LINIAL HEMISTRY, Vol. 32, No. 10, 1986