Triglycerides (as Glycerol) in Serum with the Gilford System 3500

Transcription

1 CLIN. CHEM. 24/1, (1978) Automated Procedure for Kinetic Measurement of Total Triglycerides (as Glycerol) in Serum with the Gilford System 35 Gary Lehnus and Lynn Smith An automated procedure is used on the Gilford 35 Computer-Directed Analyzer to measure serum triglycerides indirectly by using aqueous glycerol standards. Most enzymatic methods require long hydrolysis or awkward saponification. The method of Bucolo and David [C/in. Chem. 19,476(1973)], in which lipase and glycerol kinase are used, is modified. The kinetic procedure described eliminates the need for a serum blank. It uses Eskalab bulk reagents and reduces both time and cost per test by measuring the decreasing NADH concentration from the glycerol kinase reaction at 34 nm after enzymatic hydrolysis at room temperature. The change in absorbance of the standard during a 1 4-s measuring time is used in the ratiometric calculation of the unknowns. The use of a stable aqueous standard that can be reliably and accurately prepared makes this method ideal for the traceability desired by many organizations. Reagent blank drift did not affect the results. A correlation coefficient of.991 for comparison with the manual endpoint method and a typical CV of 2.25% show this method to be accurate and reliable. Measurement of serum triglycerides (triacylglycerols) has become increasingly important because triglycerides have been identified with hyperlipidemia (1). Bucolo and David (2) introduced an enzymatic triglyceride procedure that required less technician time and eliminated the tedious extraction and saponification steps previously required in conventional methods. However, their method requires serum blanks and an assay time of 15 to 2 mm and is not easily adaptable to modern instruments. Automated instrumentation has improved the precision and accuracy of kinetic measurement. The accurate timing and high sensitivity of the Gilford System 35 allow shorter hydrolysis and measuring times, thus improving productivity without sacrificing accuracy. Kinetic determination of the amount of glycerol produced takes advantage of this modern technology. Gilford Instrument Laboratories, Inc., 132 Artino St., Oberlin, Ohio Received June 6, 1977; accepted Sept. 3, Correction for free glycerol is not required for routine analysis as described by Stinshoff et al. (3). Free glycerol follows a normal distribution and may be adjusted by manually subtracting 1 mg/liter from each total triglyceride value. For diseases that require such values, free glycerol may be assayed by using a kit designed for that purpose. Results are then subtracted from the values obtained by this kinetic method. Bucolo et al. (4) describe a lipase-triggered kinetic assay that is not sensitive enough to provide the reproducibility required in the normal range. However, the glycerol kinase-triggered reaction used by several authors (5-9) for calcuiting the concentration of triglycerides does provide the necessary speed and sensitivity. Our own study confirms the observations of Bucolo and David (2) that complete hydrolysis is not necessary for the glycerol kinase reaction, because the amount of liberated glycerol is proportional to the concentration of triglycerides. For our purposes, the glycerol kinase, packaged separately, is precombined with the substrate to facilitate handling by the dispenser. A slight decrease in the absorbance of the combined starting solution did not affect our results. Because of the difference in hydrolysis time between the original endpoint method and our faster kinetic method, we examined the correlations for patients sera assayed by both methods. The results compared well. The shorter time necessary for accurate triglyceride assay represents a considerable saving in labor costs. We have adjusted the serum to reagent ratio to produce good sensitivity. An aqueous glycerol standard facilitates preparation and optimizes reproducibility. Materials Principle and Methods Samples are incubated with lipase (glycerol ester hydrolase; EC ) and cr-chymotrypsin (peptidyl peptide hydrolase; EC ) to convert triglycerides into glycerol. This reaction is allowed to proceed f or 4.5 mm at room temperature. The resulting free glycerol CLINICAL CHEMISTRY, Vol. 24, No. 1,

2 is then measured kinetically by use of a series of coupled reactions initiated by the addition of glycerol kinase (ATP:glycerol phosphotransferase; EC ) and completed by the reduction of pyruvate to lactate. After the glycerol kinase is added, the reaction is linear after a 3-s lag time. The decrease in absorbance at 34 nm owing to NADH depletion is measured over a 21-s integration time. This test is based on the following reactions: Ilpase Triglycerides - glycerol + fatty acids a-chymotrypsin Glycerol + ATP ADP + phosphoenolpyruvate Pyruvate + NADH + H Instrumentation glycerol kinase ± glycerol-i-phosphate + ADP lactate pyruvate kinase ± ATP + pyruvate dehydrogenase lactate + NAD We used System 35 Computer-Directed Analyzer for all automated procedures (Gilford Instrument Laboratories, Inc., Oberlin, Ohio 4474). The reference instrument for the manual method was a Model 25 Spectrophotometer (Gilford); the cuvets had a path length of 1 cm. A water bath was used at 37 ±.5 #{176}C for incubation, an Eppendorf 25-ILl pipet (Brinkmann Instruments, Inc., Westbury, N.Y. 1149) to dispense the sample, and a Model 31 Calculator and Model 4661 digital plotter (both from Tektronix, Beaverton, Ore. 975) for statistical program and data plots. Reagents Eskalab bulk reagents supplied by Smith Kline, Inc., Sunnyvale, Calif. 9486, and distilled/de-ionized water were used throughout. Reconstitute triglyceride reagent (cat. no. 8984) with 2 ml of water per vial. Dissolve glycerol kinase (cat. no. 8981) with.3 ml of water per vial. Prepare the starting reagent by adding 45 zl of reconstituted glycerol kinase to 1 ml of the reconstituted reagent. The final reaction mixture concentrations (per liter) are similar to those recommended in the manufacturer s package insert: 72 mmol of phosphate buffer, ph 7.1;.49 mmol of ATP;.79 mmol of phosphoenolpyruvate;.3 mmol of NADH; 5.92 mmol of magnesium salt; 79 U of pyruvate kinase (EC ); 69 U of lactate dehydrogenase (EC ); 12 8 U of a-chymotrypsin (EC ); 99 U of glycerol kinase (EC ); and 296 U of lipase (EC ). Prepare the reagents for the manual method as described in the package insert. The reagents for the automated method are designed to determine total glycerol as triglycerides in the sample. Standards Stock triglyceride standard. Accurately weigh 1.41 g of glycerol ( Photrex reagent, spectroanalyzed; J. T. Baker Chemical Co., Phillipsburg, N. J. 8864) into a tared i-liter volumetric flask, dissolve, and dilute to the mark with water. This solution is equivalent to 1 g of triolein per liter, on a molecular weight basis. Working triglyceride standard. Dilute the stock triglyceride standard fivefold with bovine albumin solution (6 g/liter) or physiological saline for a working concentration of 2 g/liter. As a check on the system, prepare a 1-fold dilution as a 1 g/liter standard. Procedures Manual procedure. The Eskalab reagent package insert describes the comparison manual procedure. Pipet 25 MI of sample into a 1-cm cuvet. Add 1.5 ml of reagent, mix with the sample, incubate in a 37 ±.5 #{176}C water bath for 15-2 mm, and allow the samples to cool to ambient temperature. Record the initial absorbance (A) at 34 nm and add 1 gl of glycerol kinase. Exactly 1 mm after the first reading (A 1), take a second reading, and a third exactly 1 mm later (A2). To obtain.a subtract A 1 from A. Use the A2 reading, as recommended by the reagent manufacturer, to monitor the drift that some sera may exhibit. If serum drift minus blank drift exceeds ±.5A, correct the absorbances; multiply the change in absorbance by a factor calculated from extinction coefficient of NADH as described in the package insert, to obtain results in grams per liter. Automated procedure. Prepare the reagents according to the package insert, but premix the starting reagent by adding 45 pl of glycerol kinase to 1 ml of triglyceride reagent. Use the Gilford System 35 Program Card designed for this procedure. Dispenser A delivers 1 sl of sample and.5 ml of triglyceride reagent into each reaction cup. Load Dispenser B with the reagent containing glycerol kinase (starting reagent) and set to deliver.25 ml into each reaction cup. Load the cups in the following order: cups 1 and 2 cup 3 cup 4 cups 5-56 water 2 g/liter standard I g/liter standard patients samples A factor determined from the change in absorbance of the standard in cup 3 minus the blank rate of cup 2 is used to convert the sample values to grams per liter. Use the 1 g/liter standard as a check on the instrument/reagent system. Load the strips into position and initiate the program by depressing the RUN pushbutton. Set the instrument controls as instructed by the tape printout. The samples are automatically processed in sequence. After the 4.5-mm hydrolysis period, the starting reagent is dispensed and mixed by force of the volume injected, and the mixture is aspirated into the 28 CLINICAL CHEMISTRY, Vol. 24, No. 1, 1978

3 br s 151. A1A2 A3 9 I 9/1,1,, hun,ernp5 2 gite, Sid N.175 S-1.97 V y Sy.,..2 #{163}2 <8 45g I,e, h,rn,n,,,,pi,.6.55 too Duration,s Fig. 1. Absorbance vs. time of kinetic method Manual Method (g/iiter) Fig. 3. Correlation of kinetic method vs. manual method with serum blanks Wavelength 2 8 Temperature Rgt./serum GK/ml. Rgt. -.75/1o 45A 13 LI.75/5o 45o 1 c.75/5a 2o #{163}1.I5o 2A 11 #{149}=1./5 45A. 1 4r 3 2 Dispenser A volume Dispenser B volume Sample Standards volume Factor as calculated I Expected (g/iiter) Fig. 2. Comparison of various serum/reagent ratios thermal cuvet regulated at 3 ±.1 #{176}C. After 3s, three 7-s integrations are used to calculate the sa for each sample. Each set of absorbance readings is averaged to get an A1 for the first, A2 for the second, and A:t for the third integration segment. The actual measuring time between the average A1 and A3 readings is 14 s. The final result is expressed in grams per liter, the glycerol standard being used for a ratiometric calculation. Error-prevention messages are printed when a reaction is nonlinear by ± 15%. The instrumental conditions are as follows: Results Timing 34 nm 3#{176}C.5 ml (reagent).25 ml (starting reagent).1 ml 2 and 1 g/liter -8 to -1 Figure 1 shows a plot of a standard vs. two patients sample absorbance readings as a function of time. This was used to select the proper timing interval for the assay. The results confirm those previously described (3), showing that a 3-s lag time is necessary to obtain a linear reaction before measurement. The vertical lines define the time period used. This timing produces accurate results in the normal and borderline cases, which are of most importance and makes for the most efficient use of instrument time. Twenty patient samples of various concentrations and appearance, obtained from a local hospital s routine workload and stored at 4 #{176}C until use, were used to investigate the problem of partial hydrolysis. Mean values for the automated method were 8, 22, and 28% higher than the manual method. As more glycerol is produced in the hydrolysis step, the reaction proceeds faster. Using the amount of glycerol produced in the shorter hydrolysis time creates a more linear reaction that also correlates well with the manual method on patient samples. Optimization Reagent preparation and volumes for this procedure were chosen for ease of preparation and for cost, after considering the variables that would produce the best results. The System 35 can accurately pipet 1 ILl of serum and.5 and.25 ml of reagent. To ensure that proper sample/reagent volumes were selected, different serum-to-reagent ratios and glycerol kinase concentrations were tried; Figure 2 shows the results. The final combination selected (1 il/.75 ml) gives the best results for a wide range of sample values. Predilution of serum is not necessary for routine work. Our findings were evaluated with 175 patients samples, using several lots of reagent over a one-month period. The values from the automated assays were compared with results from the manual endpoint method corrected for reagent drift. The regression analysis (Figure 3) suggests an excellent correlation between the two methods. Linearity Linearity of the automated method was determined by using aqueous glycerol standards and a human serum pool. A patient s sample with an abnormally high value CLINICAL CHEMISTRY, Vol. 24, No. 1,

4 1, 9 Ob 9, 15 Linearity r , Expected (9/liter) Fig. 4. Linear range of kinetic method with dilutions of abnormal patient sample of 1 g/liter was diluted to give appropriate concentrations as demonstrated in Figure 4. Each data point in the figure represents an average of four values. Acceptable linearity up to 4 g/liter is obtained. Values greater than that must be diluted fivefold with normal saline or bovine albumin solution and reassayed. The program allows use of a factor to correct for dilution. Precision Precision data are listed in Table 1. The within-run precision was evaluated for four replicate sample concentrations. The day-to-day data was accumulated during 1 days for five different lot numbers of reagents in four concentrations. Interferences Interference by bilirubin and hemoglobin was determined by adding known amounts of bilirubin and hemoglobin to normal human sera (Table 2). Bilirubin, >1 mg/liter or hemolysis >3 mg/liter will cause falsely low values. Other substances were not evaluated for interference because we believe the reagent manufacturer has already done that adequately. Discussion Totally automated systems can be used to assay triglycerides routinely with the accuracy and sensitivity necessary for reliable results. We have demonstrated that a kinetic procedure for triglycerides with pseudofirst-order reaction techniques generates results comparable to the manual endpoint method with excellent precision, economical use of reagents, minimum use of personnel, and use of readily available commercial reagents. The intrinsic drift in the combined reagent in the manual method, evidenced in the reagent blank value as a consistent decreasing absorbance, was evaluated to observe the effect on the capacity of this kinetic method. Two samples were assayed repeatedly during 7 mm in the same run. The CV for a value of 1. g/liter No. Mean SD Bllirubni AmI. added Apparent Recovery, ng. liter Table triglycerides % none Precision With-in run Day to day Hen/oglob/n Ann added Apparent triglycerides mg liter none , CV. % Table 2. Effect of Bilirubin and Hemoglobin on Analytical Recovery of Triglycerides Recovery, was 4.9%; for 8.5 g/liter the CV was 1.4%. These results indicate that the automated method is unaffected by intrinisic reagent drift Glycerol kinase is an allosteric enzyme with multiple binding sites. Generally, allosteric enzymes yield sigmoidal curves for reaction velocity; binding of the substrate at one site causes an increase in affinity at the vacant sites in the tetramer. We found a positive response as the concentration of the substrate increased, probably because of the cooperative binding phenomenon. Because the curve does not become sigmoidal until the concentration is 4-5 g/liter, use of the test should be limited to that range. The regression data plotted in Figure 3 are influenced by the values above 5. g/liter; therefore, recalculated data for values up to 5. g/liter become: r =.989 and y =.966x Limiting the range to 4. g/liter also decreases the variable rates observed for different lots of glycerol kinase. Endpoint determinations depend on total units of glycerol kinase and the time for conversion of glycerol. Individual vials of glycerol kinase were found to be within manufacturer s specifications. In the kinetic method a reaction is used that requires a specific rate over the range of measurement. This special requirement imposes different specifications, of which the manufacturer is aware. Work is continuing on standardization of rates to improve linearity. The Gilford System 35 was used because timing, sensitivity, and two reagent additions are necessary for good accuracy and precision. This method cannot be performed manually and most instrumentation does not meet these requirements. Performance and interferences as discussed in the Results section indicate that excellent correlation and ease of operation, without the need to dilute the sample, combine with reasonable cost per test to make this method reliable in screening for total triglycerides ij CLINICAL CHEMISTRY, Vol. 24, No. 1, 1978

5 We thank Pat Crowley (Smith-Kline Instruments, Inc.) and Jacques Escoffre (Gilford Europe, S.A.) for their help with this procedure. References 1. Fredrickson, D. S., A physician s guide to hyperlipidemia. Mod. Concepts Cardiovasc. Dis. 41,31(1972). 2. Bucolo, G., and David., H., Quantitative determination of serum triglycerides by the use of enzymes. Gun. Chem. 19, 476 (1973). 3. Stinshoff, K., Weisshaar, D., Staehler, F., et al., Relation between concentrations of free glycerol and triglycerides in human sera. Gun. Chem. 23,129(1977). 4. Bucolo, G., McCroskey, R., and Whittaker, N., Lipase-triggered kinetic assay of serum triglycerides. Clin. Chem. 21,424(1975). 5. Ziegenhorn, J., Improved method for enzymatic determination of serum triglycerides. Clin. Chem. 21, 1627 (1975). 6. Chong-Kit, R., and McLaughlin, P., Fully automated, all-enzy. matic triglyceride method adapted to the GEMSAEC centrifugal analyzer, with use of an aqueous triolein standard. Clin. Chern. 2, 1454 (1974). 7. DiCesare, J. L., Optimum kinetic enzymatic procedures for glucose and triglycerides in plasma and serum. Gun. Chem. 21, 1448 (1975). 8. Tiffany, T.., Morton, J. M., Hall, E. M., and Garrett, A. S.,,Jr., Clinical evaluation of kinetic enzymatic fixed-time and integral analysis of serum triglycerides. Clin. Chem. 2, 477 (1974). 9. Wentz, P. W., Cross, R. E., and Savory, J., An integrated approach to lipid profiling: Enzymatic determination of cholesterol and triglycerides with a centrifugal analyzer. Clin. Chem. 22, 188 (1976). CLINICAL CHEMISTRY, Vol. 24, No. 1,