An Evaluation of Turbidimetric Technics for Estimation of Plasma Fibrinogen

Transcription

1 An Evaluation of Turbidimetric Technics for Estimation of Plasma Fibrinogen Jesse F. Goodwin Studies on precipitation and turbidimetric methods for fibrinogen determination employing Na2SO3, Na2SO4, (NH4) 2S4 and a mixture of phosphate salts are presented. Optimal reaction conditions for these technics have been verified. The amounts of plasma fibrinogen precipitated by the sulfite and phosphate methods are in closer agreement than either are with values obtained by the (NH4)2S4 and Na2SO4 precipitation technics. It has been determined that the degree of turbidity produced by salt fractionation is not a reliable index of the amount of fibrinogen present. The measurement of fibrinogen concentration of turbidity produced by the various fractionation agents does not appear to be precise enough for the quantitative measurement of this protein. PREVIOUS STUDIES (1) indicate that thrombin interaction, heat precipitation at 56#{176}, and sulfite fractionation exhibit a high degree of correlation as means of separating fibrinogen from plasma prior to colorimetric measurement. However, turhidimetry serves as a basis for the estimation of plasma fibrinogen in a number of procedures. There is disagreement over the reliability of such methods (25). From a standpoint of rapidity, a turbidimetric method would be of obvious value to the clinical laboratory. However, the reproducibility and precision of such methods is of primary concern when compared with technics involving a separation of fibrinogen prior to its measurement. The purpose of this report is to evaluate selected turbidimetric methods in relation to the concentration of plasma fibrinogen precipitated by them. A spectrophotometric sulfite fractionation procedure (6) for fibrinogen has been chosen as a reference. Three published turbidimetric methods all claiming (1) a specificity for fibrinogen and (2) a quantitative relationship between the degree of turbidity and fibrinogen concentration in plasma have been chosen for this study. From the limiical Research emiter, hildren s hospital of Miclmigami and Waymie State University School of Medicine, Detroit, Mich Supported by Research Grant FR74 from the National Institutes of Health, Department of Health, Education and Welfare, U. S. Public Health Service, Bethesda, Md. Received for publication Mar. 1, 1967; accepted for publication June 8, I57

2 158 GOODWIN linical hemistry Experimental Methods The three turbidimetric methods chosen for study are: Fowell s (7) modification of the ammonium sulfate method of Parfentjev et al. (8) which employs a 13.33% salt concentration; a sodium sulfate (1.5% w/v adjusted to a ph of 7.) procedure by Podmore (3); and Martinek and Berry s (5) modification of the phosphate turbidity method of Aull and Mcord (9) which makes use of a 1.2 M phosphate reagent containing equal molar concentrations of potassium acid phosphate and disodium phosphate. The detailed directions are outlined in the individual procedures. Measurements were made with a oleman Jr. Model 6D spectrophotometer on random samples from a variableagespan hospital population. A wide range of fibrinogen concentration was represented in this group. In the Fowell procedure an equation for estimating fibrinogen concentration is proposed. This equation is empirically derived from calculations based on the method of least squares by comparing a reference thrombin precipitationkjeldahl nitrogen technic with turbidimetric readings. However, their equation is limited to a single instrument. In view of this, the Fowell procedure was calibrated with bovine fibrinogen. Standardization with bovine fibrinogen is recommended in the sodium sulfate and phosphate turbidimetrie methods. Fibrinogen quantitation was performed by dissolving the fibrinogen fractioned by the various precipitating agents in 1. ml. of 3% urea. This solution was reacted with 5. ml. of 3% NaOH followed by 1. ml. of Benedict s qualitative reagent. The resulting biuret color was read after incubation at 37#{176} for 2 mm. at 54 mj. Resultsand Discussion A study of physical requirements for fibrinogen precipitation by the sulfite precipitation and the three turbidimetric technics was made (Table 1). Protein precipitation was carried out on a pooled sample of human plasma. Incubation was performed at 37#{176} and 25#{176}. Precipitation was complete with most of the procedures after a 15mm. incubation at 37#{176}. The sodium sulfate method yielded optimum concentrations of fibrinogen with incubation at 25#{176}. The original precipitate from all methods contained occluded protein which was washed free by suspension in the fractionating medium. Second washings were not necessary, as indicated by a constant value for protein upon additional washings. olorimetric fibrinogen estimation was carried out on washed fibrinogen precipitates. In view of the results obtained in this experiment, precipitation was carried out at 37#{176} for a period of 15 mm. for

3 Vol. 13, No. 12, 1967 PLASMA FIBRINOGEN 159 Table 1. EFFET OF INUBATION TIME AND TEMPERATURE ON F IBRINouEN PREIPITATION By SALT FRATIONATION TEHNmS Fibrinogen (mg./1 ml. pluoma) MeMo,! S mis. 15 mi,,. 3 mis. 45 mis. 6 mm. 13 mm. Na2SO, )O 37#{176} Na,S4.)O #{176} (Nl14)aSO, 25#{176} #{176} Phosphate 25#{176} #{176} the procedures. The exception in this case was the sodium sulfate procedure, where the incubation temperature used was 25#{176}. The results indicate that the 3 methods involved tend to yield lower values than the sulfite procedure for fibrinogen. Purified bovine fibrinogen* was assayed by the various methods by making use of the optimal conditions for fibrinogen precipitation. Quantitation was accomplished by the use of a biuret endpoint (6). The results (Table 2) reveal that the sulfite fractionation and phosphate methods compare favorably at the lower concentration levels. At higher concentration levels (69 mg. clottable protein per 1 ml.), 96% and 77% of the fibrinogen were fractionated by the sulfite and phosphate technic, respectively, compared to 55% for the Na2SO4 and (N114)2S4 methods. The decrease in recovery of fibrinogen at the higher concentration by the phosphate method does not appear to be due to an interaction of copper with the phosphate salts. A twofold increase in copper concentration does not result in a lower protein concentration, nor is there any visible precipitate of copper phosphate in the sample prior to measurement. Recovery of bovine fibrinogen appears to be nonlinear at the higher concentration levels with all of the fractionation methods employed. The sulfite fractionation technic gave the best recovery values. Randomly selected plasma specimensfree from visible hemolysis or icteruswere assayed colorimetrically for fibrinogen by the four fractionation technics. Results are given in Table 3. The phosphate and sodium sulfate fractionation technics correlated better with the sulfite fractionation technic than did the (NH4)S4 fractionation method. *Obined from warnerhilcott Division of General Diagnostics, Morris Plains, N. J.

4 16 GOODWIN linical hemistry,7.6 Phosphate Turbidimetric Technique (NH4)2 SO4 Turbidimetric Technique Na2 SO4 Turbidumetric Technique a U,5.4.3 m/.2. _625mL mg Fibrinogen/ loomi Plasma Fig. 1. alibration curves of 3 turbidimetric technics for plasma fibrinogen estimation. 1 > a 8 6 Estimoting = N = Sy.x = x = y Equotion X #{149} #{149}1 I. I Os #{176}i c 9 LL I) tj) c o _o Zu S I. S N2SO4 Turbidity Measurements (X) Fibrinogen onc. mg/loom1 Fig. 2. omparison of fibrinogen concentration obtained by sodium sulfate turbidimetric measurements and sulfite fractionation.

5 Vol. 13, No. 12, 1967 PLASMA FIBRINOGEN 161 Table 2. OMPARISON OF FIBRINGEN \ALUES OBTAINED IIY SALT FRATIONATION TEHNIS Bovine fibrmnogen (mg./1 ml.) Fibrinogen recovered (mg./1 ml. plasma) Na,SO, Phosphate (NH,)SO, Na5SO, Table 3. OMPARISON OF FIBRmNOGEN VALUES OBTAINED ay VARIOUS SALT FRATIONATION TEHNIS AND OMPARED WITH SUI.FmTE FRATIONATION Salt fraclio nation tech nics Na SO, (NH,)2S, Na2SO, Phosphate (X) (Y) (Y ) (I ) Number of tests ( Range (mg/1 ml.) Mean (mg./1 ml.) Regression coefficient (h) S.D.of(b).13.9 OOll oefficient of correlation (r) Estimating equation: (Y = 2A +.465X) (Y = X) (Y X) 354 All four methods yield different average values. The phosphate and sulfite mean values are closer in agreement with each other than both are with the other two technics. Although the NaSO4 and (N114)2S4 fractionation procedures yield average values which are in close agreement, they do not correlate with each other as well as one would expect. Assay values from both methods on 23 randomly selected specimens exhibited a correlation coefficient of.85 and an estimation error in excess of 25% at the 95% confidence level, although their mean values were in close agreement (183 mg./1 ml. and 28 mg./1 ml.). The turbidity produced by the three turbidimetric methods, as measured by absorption readings, is shown in Fig. 1. Purified bovine fibrinogen was used as a standard. The turbidity produced in each method increased in direct proportion to the amount of flbriiiogen in the sample. The slopes for the phosphate, (NH4)S4, and Na2SO4 turbidimetric calibration curves are 1.,.675, and.227, respectivelyabsorbance reading X 1 (ordinate) vs. fibrinogen in mg./1 ml. (abcissa). The phosphate turhidimetric method is the most sensitive and the sodium sulfate technic the least sensitive at the wavelengths suggested for measurements.

6 162 GOODWIN linical hemistry A comparison of fibrinogen concentration by Na2SO4 turbidimetric measurements with fibrinogen concentration by sulfite fractionation is displayed in the scattergram in Fig. 2. The estimating equation exhibits an error of ±3% for the mean value of fibrinogen obtained by sulfite Estimating Equation (entire range) > a Y =55.6+O.911X N =93 / Sy.x = 54 =.91 / = 376 mg Fibririogen / 1 ml. 7 (Phosphote turbidity) = 352 mg. Fibrinogen/ looml. / (SulFite) a a 9 Zu 4 2 #{149} #{149}_ #{149} #{149}#{149}. If#{149} #{149}.y?.#{149}#{149}#{149}. v 7. o N #{176} Sy.x 7 _roo r V Estimating Equation (12 33 mg. Fibrinogen/lOOmI.) = 1l X = 33 = 28.4 =.794 = = I I I Phosphate Turbidity Measurements (X) Fibrinogen onc. mg/loo ml Fig. 3. omparison of fibrinogen concentration obtained by phosphate turbidimetric measurements and sulfite fractionation. fractionation at the 95% confidence level. A 16.7% error was noted at the 95% confidence level at its extreme range of 69 mg./1 ml. In the phosphate turbidity procedure it was indicated that for fibrinogen values over 3 mg./1 ml., the plasma should be diluted with saline and the test repeated. For this reason, assay by this method was conducted on two groups of plasma. The first group consisted by values in the range of mg. of fibrinogen per 1 ml. The second group consisted of values below 3 mg. of fibrinogen per 1 ml. Some of the values in the latter group were obtained by diluting the specimens containilig over 3 mg. of fibrinogen per 1 ml. Results in Fig. 3 would strongly indicate that over the entire range of fibrinogen concentration, a better correlation is obtained with the estimating equation than over

7 Vol. 13, No. 12, l67 PLASMA FIBRINOGEN 1O6 the limited range of 123 mg. of fibrinogen per 1 ml. The estimating equation derived from values ranging from 335 to 69 mg. of fibrinogen per 1 ml. of plasma (not shown on the scattergram) is Y = X with a standard deviation of ± 51.5 and a correlation coefficient > E O 8 6 Estimating Equation V = X N = 22 Sy.x 51.6 =.82 = mg. Fibrinogen / 1 ml. V = mg. Fibrinogen/ 1 ml. 5 S 5$ I.. 4 LLO S 4) o, S. S Phosphate Turbidity Measurements (X) Fibrinogen onc. mg/i ml 1 Fig. 4. omparison of fibrinogen concentration obtained by phosphate turbidimetrie measuiement with colorimetric measurement following phosphate fractionation. of.795. The error of the estimating equation for the mean value of this range is ±22% at the 95% confidence level compared to ±25% for the 153 mg./1 ml. concentration range. These findings indicate that fibrinogen values in the range of 3 mg./1 OO ml. and below exhibit a slightly poorer correlation and adherence to its estimating equation than the higher values when obtained by the phosphate turbidity method as compared with the values obtained by the sulfite fractionation method. When turbidimetric fibrinogen values for the phosphate technic were compared with the amount of protein precipitated by the same technic poor correlation was exhibited (Fig. 4). The per cent error of the mean fibrinogen value obtained by sulfite fractionation of the calculated bestfit curve in the scattergram is ±27%. In view of the low sensitivity and poor correlation displayed by the

8 164 4OODWIN linicalhemistry (NH4)2S4 fractionation, no studies were conducted to correlate turbidity measurements and fibrinogen content by this method. A group of 18 samples was assayed for fibrinogen by the sulfite method. Turbidity measurements were made on the same samples at 51 m after mixing with sulfite solution. A correlation coefficient of.9 was calculated with an estimating error of ±27% at the 95% confidence level. The data obtained bp turbidimetric measurement in the experiments conducted in this study tend to indicate that the amount of turbidity produced by fibrinogen reagents is not a reliable index of the exact amount of fibrinogen fractionated by them. This is not too surprising, since none of the turbidimetric methods tested are capable of accounting for the profound effect of the many variables on the development of turbidity in a complex medium such as plasma. Of the three fibrinogen turbidimetric methods studied, the phosphate method of Martinek and Berry exhibits the best correlation with the sulfite fractionation technic. However, the error encountered does not warrant it precise enough to be used for other than a semiquantitative screening test for fibrinogen estimation. References 1. Goodwin, J. F., An evaluation of technies for the separation and estimation of plasma fibrinogen. lin.he n. 11, 63 (1965). 2. Reiner, M., and heung, H. L., Standard Methods of linicalhemistry (vol. 3), Seligeon, D., Ed. Acad. Press, New York, 1961, p Podmore, D. A., Rapid turbidimetric methods for the determination of plasma fibrinogen. hit. him. Acta 4, 242 (1959). 4. Ellis, B.., and Straneky, A., A quick and accurate method for the determination of fibrinogen in plasma. J. Lab. liii. Med. 58, 477 (1961). 5. Martinek, R. G., and Berry, H. E., Micromethod for the estimation of plasma fibrinogen. liii.heat. 11, 1 (1965). 6. Goodwin, J. F., Micro estimation of fibrinogen with a senhiinicro modification applicable to ieteric plasma. lin. heat. In press. 7. Fowell, A. H., Turbidimetric method of fibrinogen assay. Am.,1.lin. Fathol. 25, 34 (1955). 8. Parfentjev, T. A., Johnson, M.., and lifton, E. E., The determination of plasma fibrinogen by turbidity with ammonium sulfate. Arch. Biochem. Biophy8. 46, 47 (1953). 9. Aull, J.., and Meord, W. M., A simple rapid procedure for the estimation of albumin and alpha, beta and gamma globulin in serum. J. Lab. hit. Med. 46, 476 (1955).