ClinBiochem, Vol. 20/pp. 167-172, 1987 0009-9120/87 $3.00 + .90 Printed in Canada. All rights reserved. Copyright '¢3 1987 The Canadian Society of Clinical Chemists.

A Method for the Sequential Colorimetric Determination of Serum Triglycerides and Cholesterol

AJIT SHARMA, JOSEPH D. ARTISS, and BENNIE ZAK

Department of Pathology, Wayne State University School of Medicine, 540 East Canfield, Detroit, Michigan 48201

A simple spectrophotometric method for the sequential determina- tion of triglycerides and cholesterol from a single serum sample was developed. In this two-stage procedure, the triglycerides and cho- lesterol esters are first hydrolysed to glycerol and free cholesterol respectively, with simultaneous scavenging of the liberated free fatty acids, a technique that ensures clarity of the sample. The glycerol is subsequently reacted to result in an intense red chromogen with a peak absorption maximum at 510 nm following a series of enzymic reactions. In the second stage, addition of cholesterol oxidase leads to oxidation of free cholesterol generated from the cholesterol esters in the first stage and the free cholesterol normally present in the sample, yielding in a similar fashion the identical red chromogen whose absor- bance is also measured at 510 nm. Results obtained with the pro- posed method demonstrate good correlation with established individ- ual procedures for triglycerides and cholesterol.

KEY WORDS: analysis, sequential; triglycerides; cho- lesterol; hyperlipidemia.

T he determination of serum lipids has assumed a very important position in the clinical laboratory over

the past few years. Serum triglyceride and cholesterol levels are commonly requested tests as they constitute the more essential components of a patient's lipid pro- file. The assay of both of these analytes is often per- formed by enzymic methods culminating in the measurement of an absorbing species at a visible or ultraviolet wavelength. The assay of two or more clinically important analytes in a single cuvet by sequential or simultaneous measurement and calcula- tion has been previously reported (1-4) and such dual processes may be analytically and/or economically advantageous. One of these reports (4) involves the sequential determination of triglyceride by oxidation of NADH for its indicator reaction followed by cholesterol assay using a Trinder-type chemistry, but it appears to have some avoidable characteristics. For example, a high concentration of glycerol must be added to the reagent of a commercial test kit procedure for cho- lesterol to keep the peroxide formed in the second sequence of analytical steps from reacting with a rela- tively high concentration of NADH remaining from the triglyceride determination of the first stage. In addition, the use of an ultraviolet wavelength for the measure- ment of the triglycerides and visible wavelength for cholesterol provides analytical circumstances in which the two analytes may be subject to different inter-

Correspondence: Dr. Joseph D. Artiss, Department of Pathology, Wayne State University School of Medicine, 540 East Canfield, Detroit, MI 48201.

Manuscript received July 15, 1986; revised September 22, 1986; accepted October 10, 1986.

CLINICAL BIOCHEMISTRY, VOLUME 20, JUNE 1987

ferences. Furthermore, NADH-coupled reactions for triglyceride determination have been shown to be of lower sensitivity and smaller dynamic range than per- oxidase-coupled reactions (5).

We felt that it would be advantageous not only to perform sequential triglyceride and cholesterol deter- minations on a single sample but also to employ state-of- the-art enzymic reactions in the process (5, 6). What we propose uses the identical equilibrium reaction product for both cholesterol and triglycerides with a visible range endpoint that is more sensitive in terms of absor- bance achieved than either of the reactions previously reported (4). This increased sensitivity facilitates an increase in sample to reagent ratio to 1:200 from the previously reported 1:100 (4). The addition of a- cyclodextrin (~-CD) to the triglyceride reagent has also eliminated the serious problems associated with tur- bidity that have been reported for both triglyceride (5) and cholesterol (7) determinations. The presence of ~- cyclodextrin in conjunction with the lipase of the reagent causes lipemic specimens, either naturally occurring or from infusion of total parenteral solutions, to clarify rapidly so that the optical perturbations caused by hyperlipidemia are eliminated for both deter- minations (8). The ability to use smaller sample vol- umes tends to diminish the detrimental effects of other interfering substances such as the pigments, bilirubin and hemoglobin. Furthermore, the fact that this pro- cedure gives accurate results in the presence of elevated triglycerides makes it ideal for monitoring total paren- teral nutrition therapy.

Materials and methods

M A T E R I A L S

Cholesterol esterase (microbial; sterol ester hydro- lase, EC 3.1.1.13), peroxidase (from horseradish; hydrogen-peroxide oxidoreductase, EC 1.11.1.7) glycer- ophosphate oxidase (from Aerococcus viridans; gly- cero-3-phosphate: O2 oxidoreductase, EC not assigned), glycerokinase (from Streptomyces canus; ATP: glycerol-3-phosphotransferase, EC 2.7.1.30), lipase (from Chromobacterium viscosum: glycerol-ester hy- drolase, EC 3.1.1.3) and cholesterol oxidase (microbial; cholesterol: 02 oxidoreductase, EC 1.1.3.6) were obtained from Finnsugar Biochemicals Inc. (Schaum- burg, IL 60173-4808). Adenosine 5'-triphosphate, a- cyclodextrin, Triton X-100, magnesium chloride and 4- aminoantipyrine (4-AAP) were obtained from Sigma

167

SHARMA, ARTISS. AND ZAK

Chemical Co. (St. Louis, MO 63178). Sodium 2- hydroxy-3,5-dichlorobenzenesulfonate (HDCBS) was ei ther synthesised by us as previously described (9) or was obtained from Research Organics Inc. (Cleveland, OH 44125). Cholesterol (colorimetric) and triglyceride (ultraviolet) reagents used for the Hitachi 705 Chemis- t ry Analyser and lyophilized serum standards were obtained from Boehringer Mannheim Diagnostics (BMD) (Indianapolis, IN 46250). Control sera (Deci- sion ~ Liquid Comprehensive Chemistry Control serum) was obtained from Beckman Inst ruments (Fullerton, CA 92634).

APPARATUS

The sequential assay was performed with the Shimadzu UV-260 Spectrophotometer (Shimadzu Corp., Kyoto, Japan) equipped with a seven-position cell changer and a tempera ture control module. The com- parison methods for triglycerides and cholesterol were performed on the ABA-200 Bichromatic Analyzer (Abbott Laboratories, Diagnostics Division, North Chi- cago, IL 60064).

REAGENTS

Sequential assay

t rophotometer was then zeroed and the triglyceride reaction was initiated by the addition of 5 ~L of serum. The absorbance at 510 nm was recorded following a 10- min incubation. Then 10 ~L of reagent 2 was added to the test cuvet and the absorbance at the above wave- length recorded after another 5-rain incubation. The sera, standards and controls were all run in a similar fashion. The first absorbance was used to calculate the triglyceride concentration while the difference between the first and second absorbance readings was used in the calculation of the cholesterol concentration.

Individual triglyceride and cholesterol assays were performed on the ABA-200 Bichromatic Analyzer with the same serum standards and controls used in the sequential method. The ABA-200 ins t rument settings were as follows:

Triglyceride Cholesterol

Filters 340/380 500/600 Syringe ratio 1:81 1 : 101 Analysis time 5 5 Temperature 37°C 37°C Reaction type Endpoint Endpoint Reaction duration Down Up Calibration factor Use standards Use standards FRR Yes No

Both reagents for the sequential assay were prepared in Tris-HC1 buffer, 50 mmot/L, pH 7.2, to contain per litre:

Reagent 1: Cholesterol esterase >~390 U, ATP 0.7 retool; magnesium chloride 1 mmol, peroxidase ~>4kU, L-a-glycerophosphate oxidase I>8 kU, glycerokinase I>400 U, lipase />250 kU, a-cyclodextrin 4.1 mmol, 4- aminoantipyrine 1.0 mrnol, sodium 2-hydroxy-3,5-di- chlorobenzenesulfonate 1.9 mmol and Triton X-100 0.3 g.

Reagent 2: cholesterol oxidase/>30 kU.

Comparison method

Triglycerides: Working solutions R1 and R2 were pre- pared as described by the supplier and mixed according to proportions used on the Hitachi 705, immediately before performing the triglyceride assay.

Cholesterol: Working solution R1 was prepared according to the manufacturer 's instructions.

Standards: Lyophilized pooled human serum cal- ibrators were reconstituted according to the manufac- turer 's instructions.

Controls: Beckman Decision Levels I and III were used according to the instructions of the supplier.

Serum samples: All specimens used in this report were frozen sera (less than one month old) obtained from the Chemistry Laboratories of Detroit Receiving Hospi- ta l /Univers i ty Heal th Center.

METHODS

Into the reference and test cuvet maintained at 37°C was pipetted 1.0 mL of reagent i. The spec-

Results and d iscuss ion

Many modifications of procedures are currently avail- able for quantifying triglycerides and cholesterol in serum. Since the measurements of these two analytes are often paired, it may be economically advantageous both in reagent cost and technologist t ime to assay both consti tuents in a single cuvet.

We have chosen to measure sequentially both of these analytes by spectrophotometric means using the sen- sitive chromogenic system generated from the perox- idase-catalysed coupling of 4-aminoantipyrine and 2- hydroxy-3,5-dichlorobenzenesulfonate. As reported ear- lier (5), the development of turbidity in the triglyceride reaction sequence possibly due to the formation of mag- nesium-fatty acid complexes may lead to analytical errors when the absorbing species is measured, especially in the ultraviolet range. Although turbidi ty becomes significantly much less of a perturbing factor as measurement wavelengths are shifted toward the red spectral region, the errors caused by turbidity may still be significant. The detr imental effects of turbidi ty on the determinat ion of cholesterol has also been reported (7). Such an effect with a lipemic specimen is shown in Figure 1. The reaction of a lipemic serum specimen ( t r ig lycer ide=46 mmol/LI with the triglyceride reagent (reagent 1 ) with its HDCBS and ~-cyclodextrin omitted, shown as A, results in a falsely elevated absor- bance, probably due, as mentioned above, to magne- sium-fatty acid complex formation. The addition of the fat ty acid scavenger (~-CDI gradually reduces the new turbidi ty by forming soluble inclusion complexes with the fat ty acids (curve B). The clarification potential of the lipase-cyclodextrin system incorporated into the

168 CLINICAL BIOCHEMISTRY, VOLUME 20, JUNE 1987

0.4

~ 0.3

@ 0.2

0.1

i

0

0.5

A

B

i I

5 10 15 TIME (minutes)

Figure 1 - - The reaction of a lipemic serum (triglyceride con- centration = 46 mmol/L) with reagent 1 without its phenol, HDCBS, is represented by curve B. Removal of a-cyclodextrin or magnesium from the reagent and its subsequent reaction with the serum sample results in curves A and C, respectively.

0.7

0.6

0.5

~ 0.4

@ 0.3

0.2

0.1

Cholesterol reaction

f Triglycerlde reaction

10 20 30 40 TIME ( minutes )

Figure 2--Color development in the sequential assay for a patient serum.

triglyceride reagent without the interfer ing magne- sium cations is represented by curve C. I t seems tha t fatty acids, free from magnes ium, are more rapidly removed by the ~-cyclodextrin. By performing the trig- lyceride assay first, turbid samples are clarified, thus eliminating the l ight scat ter ing and associated errors in the subsequent cholesterol reaction tha t have been reported previously (10, 11).

The detailed reaction sequence in the proposed sequential assay can be represented as follows:

Stage 1

Triglycerides + H20 L i p a s e Glycerol + free fa t ty

acids

S E Q U E N T I A L T R I G L Y C E R I D E S AND C H O L E S T E R O L

2 2

T+2C

T ~

10 510 Time (mln) ~ (nm) Time (mln) a[(nm)

Figure 3 - - Time course and spectra of mixtures containing a fixed amount of glycerol and increasing concentrations of aqueous cholesterol reacted in the sequential assay are shown at left. Similar curves for standards containing a fixed amount of cholesterol and increasing glycerol are shown at right. (T = 1.0 mmol/L triglyceride; C = 5.2 mmol/L cholesterol). The dashed lines in the figure at right indicate the spectra obtained after the stage 1 (triglyceride) reactions.

Cholesterol es terase Cholesterol Esters + H20 •

Cholesterol + free fa t ty acids

free fa t ty acids + a-CD - - ~ (fatty acids-a-CD) complex

Glycerol + ATP Glycerol k inase glycerol-POt

+ ADP

Glycerol-PO4 + 02 Glycerol-phosphate ox id a se

H202 + dibydroxyacetone-phosphate

2H202 + 4-AAP + HDCBS Peroxidase ,.

Red Chromogen + H20.

Stage 2

Cholesterol + Cholesterol ox idase Cholest-4-en-

3-one + H202

2H202 + 4-AAP + HDCBS P e r o x i d a s e

Red Chromogen + H20.

Reagents for the hydrolysis of cholesterol esters, fa t ty acid complex formation and the chromogenic sys tem are all incorporated into the triglyceride reagent so t ha t the react ion for cholesterol is ini t iated by simply adding a small volume of concentrated cholesterol oxidase solu- tion. This reagent is easy to prepare and results in a negligible change in the volume of the total reaction mixture , thus avoiding the need for adjusting the b lank

CLINICAL B I O C H E M I S T R Y , V O L U M E 20, J U N E 1987 169

SHARMA, ARTISS, AND ZAK

0.30

0.15

0.0 I I I I - -

600 350 400 450 500 550

WAVELENGTH (rim)

Figure 4 - - Effect of bilirubin in the sequential assay. Spec- trum A shows the reaction product of a clear serum (trig- lyceride = 2.3 mmol/L) with reagent 1. Addition ofbilirubin to this reaction mixture (final bilirubin concentration 428 Ixmol/ L) yields spectrum B. The spectrum of a similar concentration of bilirubin in reagent 1 is shown as C. Curve D indicates the spectrum of the reaction product obtained by reaction of the same serum specimen spiked with a similar concentration of bilirubin with reagent 1.

. y = l . 0 6 X - 0 . 1 6 2

I , - r = 0 , 9 9 6

0 1 3 . 7 S E Q U E N T I A L T R I G L Y C E R I D E S ( m m o l / L )

1 3 . 7

a o

Figure 5 - - Comparison of results for triglycerides by the pro- posed method with an established procedure (BMD ultraviolet triglycerides).

1 0 . 4

E E

E-,

r~

~ 1 $° ! •

|

. . ~ "" y---0.OOx - 0 . 0 8 6

r = 0 , 9 8 0

1 0 . 4 S E Q U E N T I A L C H O L E S T E R O L ( m m o i / L )

Figure 6 - - Comparison of results for cholesterol by the pro- posed method with the BMD colorimetric cholesterol pro- cedure.

for the cholesterol determination. Typical color development and stability curves for a

serum sample are shown in Figure 2. The triglyceride and cholesterol reactions are completed within 10 and 5 min, respectively. The first amount of chromogen formed is stable for at least 30 min while tha t amount generated in the second stage is stable for at least 20 min (Figure 2). Because both reactions are complete and thei r two endpoint absorbances measured within 15 min, any limits on color stability are never exceeded. The effects of increasing cholesterol at a fixed trigly- ceride level and vice-versa is shown in Figure 3 using glycerol and aqueous cholesterol standards. Under these reaction conditions, the method is applicable and l inear up to a final absorbance of 2.0. Samples exceeding this value after the second measurement should be diluted appropriately and re-assayed.

A major interference in peroxidase-catalysed reac- tions is bilirubin, which not only competes for the hydro- gen peroxide but also contributes color to the final reaction mixture (5, 6). A similar interference seems to occur at elevated bilirubin levels (greater than 170 ~mol/L) in the proposed assay (Figure 4; Table 1). The reaction of reagent I with a serum specimen normal in bil irubin results in spectrum A. The addition of bil irubin to the serum at a final bilirubin concentration of 428 ixmol/L, after the triglyceride reaction is at its endpoint, is shown as spectrum B while the same amount of bil irubin in reagent I scanned alone has the spectrum represented by C. The contribution of color by bil irubin to the triglyceride reaction product is thus evident as spectrum B, a t rue summation of A plus C. However, when bilirubin was present in the sample at the same t ime tha t oxidase action produced hydrogen peroxide, then an ent irely different spectral picture was obtained, shown by spectrum D where the peak max- imum value is severely diminished due to the competi- t ive action of bil irubin substi tuting in part as a hydrogen donor for the HDCBS and 4-AAP reaction. Spectrum D thus represents a complex summation of spectra due to 4-AAP-HDCBS coupling, some residual bil irubin and its oxidation product(s). The results obtained on adding different amounts of bilirubin to two serum samples containing low and high levels of the lipid analytes in the proposed method are shown in Table 1, where it can be seen tha t interference begins to manifest i tself at about 170 ~mol/L bilirubin. The nega- tive interference caused by bilirubin, which may be appreciable in such severely icteric serum samples, is much less detr imental to the cholesterol reaction than the triglyceride reaction of the sequential assay simply because the competition for substrate in the first ana- lytical stage has diminished the bilirubin concentration almost completely in the second stage.

Aliquots of three serum samples selected to cover the low, medium and high range for the lipid analytes were used to determine the precision of the method. The results of this study are presented in Table 2. Both the within- and between-run reproducibilities appear to be acceptable.

Comparison of the proposed sequential method with established procedures for triglycerides and cholesterol,

170 CLINICAL BIOCHEMISTRY, VOLUME 20, JUNE 1987

SEQUENTIAL TRIGLYCERIDES AND CHOLESTEROL

Effect of Bilirubin TABLE 1

on the Peroxidase-Catalyzed Sequential Determination of Triglyceride and Cholesterol

Serum

Bilirubin added

(}xmol/L)

Triglyceride Cholesterol found a Change found ° Change

(mmol/L) (%) (mmol/L) (%)

Low

High

0 0.62 3.2 86 0.60 - 3.2 3. - 3.5

171 0.56 - 9.7 3.0 - 5.9 342 0.47 - 24.2 2.9 - 9.4

0 2.5 11.2 86 2.5 0 11.2 0

171 2.2 - 12.0 10.8 -3.6 342 2.1 - 16.0 10.8 -3.6

aAverage of duplicates.

TABLE 2 Reproducibility Studies

Triglyceride (mmol/L)

Serum n Mean SD CV%

Within-run 1 18 2 18 3 18

Between-run 1 10 2 3

0.68 0.023 3.4 1.21 0.046 3.8

11.4 0.296 2.6

0.74 0.034 4.6 10 1.20 0.046 3.8 1 0 1 1 . 9 0.570 4.8

Cholesterol (mmol/L)

Within-run 1 18 1.92 0.052 2.7 2 18 4.81 0.130 2.7 3 18 6.55 0.156 2.4

Between-run 1 10 2.03 0.052 2.6 2 10 4.89 0.13 2.7 3 10 6.66 0.286 4.3

performed on 80 patient specimens, showed a mathe- matically acceptable correlation of 0.996 and 0.980 and regression equat ionsofy = 1.06x - 0.162 andy = 0.99x - 0.086 for triglyceride and cholesterol, respectively. The data obtained to generate these statistics are shown in Figures 5 and 6. The BMD triglyceride procedure used for the above comparison involves the lactate dehydrogenase-catalyzed oxidation of NADH to NAD +. The cholesterol assay, on the other hand, is a color- imetric procedure involving the measurement of the quinoneimine dye formed from the peroxidase- catalyzed oxidative coupling of phenol and 4- aminophenazone.

No at tempt was made to automate the procedure as suitable instrumentat ion was not available to us. However, the simplicity of both the reagents and

the procedure would suggest tha t the sequential method is readily applicable to the present generation of microprocessor- or computer-controlled automated equipment.

The incorporation of our previously described clear- ing process (8) into the reagent makes the proposed procedure suitable for specimens from not only hyper- lipidemic patients but also patients receiving total par- enteral nutrition. It is of note that there is no other acceptable method of t reat ing the lipemic sample for the measurement of its lipids as all other procedures for clearing lipemia remove variable amounts of these ana- lytes from the specimen (12). The combination of the clearing process, the small sample volume require- ments and relative insensitivity to bilirubin inter- ferences makes this procedure an ideal candidate for moni tor ing neonates receiving total parenteral nutrition.

R e f e r e n c e s

1. Zak B, Ressler N. Simultaneous microdetermination of copper and iron using mixed phenanthrolines. Anal Chem 1956; 28: 1158-61.

2. Zak B. Simple procedure for the single sample determina- tion of serum copper and iron. Clin Chim Acta 1958; 3: 328-34.

3. McGowan MW, Artiss JD, Zak B. Determination of free, esterified and total serum cholesterol using a sensitive color reaction. Microchem J 1983; 28: 294-9.

4. Zoppi F. Single-cuvet sequential determination of trig- lyceride and cholesterol. Clin Chem 1985; 31: 2036-9.

5. McGowan MW, Artiss JD, Strandbergh DR, et al. A perox- idase-coupled method for the colorimetric determination of serum triglycerides. Clin Chem 1983; 29: 538-42.

6. Artiss JD, McGowan MW, Zak B. Sensitive high-density lipoprotein cholesterol assay. Microchem J 1981; 26: 198-209.

7. Demacker PNM, Boerma GJM, Baadenhuijsen H, et al. Evaluation of accuracy of 20 different test kits for the enzymic determination of cholesterol. Clin Chem 1983; 29: 1916-22.

8. Sharma A, Artiss JD, Strandbergh DR, et al. The turbid specimen as an analytical medium: hemoglobin deter- mination as a model. Clin Chim Acta 1985; 147: 7-14.

CLINICAL BIOCHEMISTRY, VOLUME 20, JUNE 1987 171

SHARMA, ARTISS, AND ZAK

9. Artiss JD, Thibert RJ, McIntosh JM, et al. Studyofvarious substrates for peroxidase-coupled peroxide oxidations. Microchem J 1981; 26: 487-505.

10. Miyada D, Tipper P, Jantsch D, et al. The effect of hyper- lipidemia on Technicon SMAC measurements. Clin Bio- chem 1982; 15: 185-8.

11. Cobbledick R, Hinberg IH, Poon R, et al. Elimination of lipid interference with the Cobas-Bio procedure for cho- lesterol. Clin Chem 1985; 31: 1024. Abstract.

12. McGowan MW, Artiss JD, Zak B. Description ofanalytical problems arising from elevated serum solids. Anal Bio- chem 1984; 142: 239-51.

172 CLINICAL BIOCHEMISTRY, VOLUME 20, JUNE 1987