CLIN. CHEM. 36/9, 1605-1613 (1990)

Triglyceride Measurements: a Review of Methods and Interferences

SIgrid G. Klotzsch1 and JudIth R. McNamara2

I

The National Cholesterol Education Program has empha-

sized the need to identify individuals at risk for coronaryartery disease (CAD). Because increased triglycerides may

be a risk factor for CAD and because triglycerides are usedto estimate concentrations of low-density lipoprotein (LDL)

cholesterol, which has definitely been shown to be a riskfactor for CAD, it is important that reliable results be ob-tained. Many methods are available for measuring triglycer-

ide concentrations in serum or plasma, but there is nodefinitive method that confirms the reliability of any of theseprocedures. Accuracy and precision guidelines are extremelydifficult to determine, owing to broad biological variabilityboth within and among individuals. Here, we review the majortriglyceride quantification methods in the literature, some ofthe potential interference problems, and the limitations re-garding standardization that should be addressed whenestablishing such guidelines.

Assessment of the Needs for Accuracy and Precisionin Triglyceride Quantification

The National Cholesterol Education Program AdultTreatment Panel has recommended that triglyceride con-centrations <2.82 mmol/L (<2.50 gIL) be classified asdesirable, 2.82-5.65 mmol/L (2.50-5.00 g/L) as borderline,and >5.65 mmol/L (>5.00 g/L) as distinctly abnormal (1).

These ranges are relatively broad, but it has been difficultto assign specific concentrations at which medical interven-

hon is recommended because of controversy regarding therole of triglycerides in coronary artery disease (CAD).3 Theassociation of triglycerides with CAD has been variable inpopulation studies that evaluated triglyceride concentra-tion as an independent risk factor (2,3).

If triglyceride concentrations in serum were not consid-ered to be an important risk factor for CAD, if onlyextremely high concentrations were felt to be important interms of diagnostic value, or if low-density lipoprotein(LDL) cholesterol concentrations could easily be deter-rnned independently of triglycerides, then efforts towardtandardization of laboratory methodology might not beLecessary. If, however, triglyceride concentrations do mdi-ate some degree of risk, even if only in a certain subset ofndividuals (4, 5), and if triglyceride concentrations con-inue to be a necessary tool for estimating LDL cholesterol6-B), then it is important to eliminate as much assay

‘Technicon Instruments Corp., Division of Miles, Inc., Tarry-

own, NY 10591-5097.2Upid Metabolism Laboratory, USDA Human Nutrition Re-

earch Center on Aging, Tufts University, Boston, MA.3Nonstandard abbreviations: CAD, coronary artery disease;

�DC, Centers for Disease Control; BSA, bovine serum albumin;nd GPO, glycerol-3-phosphate oxidase.

Received October 1989; accepted June 22, 1990.

variability as possible. In the laboratory, this means reduc-tion of inaccuracy and imprecision to values comparablewith those for cholesterol. In patient management thismeans setting fasting times and pretesting dietary guide-lines to minimize biological variation. It also means carefulinterpretation of laboratory results (6, 9, 10).

Defining the reliability required for any triglyceridemethod being considered for laboratory standardization isdifficult, because biological variability of triglyceride con-centrations differs greatly among individuals (11, 12). Ourstudies show marked differences among subjects receivingthe same amount of fat per kilogram of body weight (12). Inaddition, some individuals show variability in their fastingtriglyceride concentrations when sampled on different dayswithin a short period (unpublished data, J.R.M).

The ubiquitous nature of glycerol interference must betaken into account. Glycerol may derive from the sampleitself (13, 14) or from medications given to the patient(15-19), or samples, sample cups, tubing, etc., used in thelaboratory may contain traces of lubricants that causeglycerol interference (20). The amount of glycerol presentin skin-care products is not negligible (21, 22). A sampleprobe in a random-access analyzer aspirating 5 �.iL ofsample, equivalent to an average of 0.5 ng of free glycerolin the assay mixture, is conceivably delivering extraneousglycerol or glycol exceeding this amount into the system.

As a first step toward a critical evaluation of triglyceridedeterminations we have reviewed available methodology,interferences, and standardization limitations.

History and Principal Methods

Early analytical methods for determining triglyceridesinvolved titrimetric procedures of total lipids. After extrac-tion with organic solvents, extracts were saponified andthen back-titrated to assess the amount of alkali that wasnot neutralized by the released fatty acids. Later methodsquantified the glycerol that was formed (see Table 1), asfirst described by Van Handel and Zilversmit (23), whoextracted the lipids with chloroform, removed phospholip-ids by adsorption on silicic acid, and then saponified theglyceride esters to glycerol. They determined glycerol byoxidation with periodate, using a colorimetric measure-ment of the product from the reaction of formaldehyde andchromotropic acid in sulfuric acid. Although some mono-and diglycerides of fatty acids are present in serum (-3%),results were calculated as “triglycerides.” Subsequently,the term triglycerides was commonly used for the total offree and protein-bound glyceride esters.

Kessler and Lederer (24) adapted this method to theAutoAnalyzer#{174} (Technicon, Tarrytown, NY). They usedisopropanol extracts of serum, added zeolite/Lloyd’s re-agent to adsorb phospholipids and other interferences, andthen introduced an on-line saponification step. The glycerolreacted with periodate and produced formaldehyde, which

CLINICAL CHEMISTRY, Vol. 36, No. 9, 1990 1605

Reference

23

Table 1. Principles of Analyses for TriglyceridesPrinciple of assay

1. Triglycerides extracted with organic solvents;phospholipids removed by adsorption

2. Triglycerides ROH/OH� glycerol

3. Glycerol + l04 -e HCHO + HCOOH + H20 +

1034. HCHO + chromotropic acid + H2S04 -e

chromophore24 1 .-3. as above, modified

4. HCHO + NH4 + acetylacetone -e 3, 5-diacetyl-1 ,4-dihydrolutidine (fluorescent)

25 1-4. as above, modified

26 1., 2. as above

3. Glycerol + ATP � glycerol 3-phosphate + ADP4. Glycerol 3-phosphate + NAD� G3pDH�

dihydroxyacetone phosphate + NADH + H�27 1.,2. as above

3. Glycerol + ATP � glycerol 3-phosphate + ADP4. ADP + PEP � ATP + pyruvate

5. Pyruvate + NADH + H* �!i lactate + NAD�

28 1., 2. as above

3. Glycerol + ATP � glycerol 3-phosphate + ADP

4. Glycerol 3-phosphate + NAD� G.3PDIdihydroxyacetone phosphate + NADH + H�

5. NADH + H� + INT diaphoras� NAD� + INT(reduced)

29-31 Modification of ref. 27, elimination of adsorption step

32-34 1. Triglycerides �S glycerol

2. Glycerol + ATP � glycerol 3-phosphate + ADP

3. ADP + PEP � ATP + pyruvate

4. Pyruvate + NADH + H� !�!i lactate + NAD�

35 1. Triglycerides �E!!3 glycerol

2. Glycerol + ATP � glycerol 3-phosphate + ADP

3. Glycerol 3-phosphate + NAD� 03-PDIt

dihydroxyacetone phosphate #{247}NADH + H�4. NADH + H� + INT diaphoraee NAD� + INT(reduced)

36 1., 2. as above

3. Glycerol + NAD� glycerOl dehydrOgefla3�

dihydroxyacetone + NADH + H�

37 1. Triglycerides !!E5f� glycerol2. Glycerol + NAD� gly�ol dohydrogonas1

dihydroxyacetone + NADH + H�

38 1., 2. as above

3. NADH + H� + resazurin diaphoras� NAD� + resorufin39 1. Triglycerides !!2� glycerol

2. Glycerol + 02 glycerol oxldasq dihydroxyacetone +

H20240 1. H202 + phenol derivative + 4-aminoantipyrine t!�9

chromophore41 1. Triglycerides !!E�2 glycerol

2. Glycerol + ATP � glycerol 3-phosphate + ADP

3. Glycerol 3-phosphate + 02 GPQ dihydroxyacetonephosphate + H202

4. H202 + 3,5-dichloro-2-hydroxybenzene sulfonate +

4-aminoantipynne !i� chromophore

GK, glycerol kinase; G-3-O-PDH, glycerol-3-phosphate dehydrogenase;PEP, phosphoenok,�yruvate; PK, pyruvate kinase; LDH, lactate dehyrogenase;HPO, horseradish peroxidase; and INT, 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-phenyltetrazolium chloride.

was measured fluorometrically after a Hantzsch-type con-

densation with diacetylacetone and ammonia.Carlson and Wadstr#{246}m’smodification (25) of the above

methods became the procedure used by the Lipid Research

1606 CLINICAL CHEMISTRY, Vol. 36, No. 9, 1990

Clinics population studies. The assay was standardized by

the Centers for Disease Control (CDC)-National Heart,

Lung, and Blood Institute Lipid Standardization ProgramHigh accuracy and good precision make this procedure an

excellent reference, but only CDC calibrators, or a second.ary calibration material that has been analyzed by the

CDC method, can be used if interlaboratory results are tebe compared. The various analytical steps require a highdegree of technical competence.

The measurement of glycerol by an enzymatic spectro�photometric procedure was introduced by Wieland (26�

after the enzyme glycerol kinase (ATP:glycerol 3-phospho.transferase; EC 2.7.1.30) became commercially available,Eggstein and Kreutz (27) used the reverse enzymatic reac�tion, after a chemical saponification step, to assay neutra]fats. Both methods are based on the oxidation! reduction olNADH!NAD� and the corresponding change in absor.bance. Chemical saponification is usually performed in analkaline medium, with subsequent neutralization by per.chlorate or magnesium sulfate (42),whereby most proteineare inactivated. Phospholipids are also hydrolyzed, whichnecessitates adsorption on zeolite if the Wieland method ifused, because the subsequent step quantifies the glycerol3-phosphate produced. A semiautomated enzymatic analy-sis, introduced by Klotzsch et al. (28), added a colorimetric

indicator reaction to the Wieland principle and introducedthe method to the AutoAnalyzer.

The colorimetric reaction took place after samples weresubjected to the adsorption step with zeolite, before on-linesaponification with alcoholic KOH. This step eliminatedinterference from endogenous alkaline phosphatase (EC3.1.3.1) and lactate dehydrogenase (EC 1.1.1.27). However,the technique did not save much time by automation andwas later improved by Stavropoulos and Crouch (29-31).

They avoided extraction and adsorption by combining alco-

holic base saponification with the precipitation technique o�Schmidt and von Dahl (42) using magnesium salts. Cher-nick (43) claimed that the use of tetraethylan,moniunihydroxide hydrolyzed triglycerides without cleavage of phos-pholipids.

Bucolo and David (32-34) substantially improved theultraviolet test of Eggstein and Kreutz by using microbiallipase together with a protease, instead of chemical hydrol-ysis. This procedure became the most popular ultraviolettechnique for determining triglycerides. It could be usedwith most manual and automated laboratory instrumentsand received widespread recognition in clinical laborato-ries. Free glycerol could be blanked by a separate assaywith use of reagents without lipase or glycerol kinase. Thecolorimetric method of Klotzsch et al. (28), in conjunctionwith lipase instead of chemical hydrolysis, was publishby Megraw et al. (35) and became a convenient techniqufor use with centrifugal analyzers and other medium-siinstruments (44).

The enzyme glycerol dehydrogenase (glycerol:NAD2-oxidoreductase, EC 1.1.1.6) was introduced by Hagen an

Hagen (45) in 1962, when it was suggested for determininglycerol. Bowie and Gochman (36) extended this principlto triglyceride assays. Rautela et al. (37) and Grossman eal. (46) modified it to be applied to the aca#{174}(Du PonInstruments, Wilmington, DE) and CentrifiChem#{174} (BakeInstruments, Evansville, WI).

Because the equilibrium of the enzymatic reaction iunfavorable to the production of dihydroxyacetone, a kinetic (pseudo-first-order) mode was chosen to overcome th

time constraints of the instruments. Sigiura et al. (47)modified the reaction for a colorimetric measurement incombination with a tetrazolium salt; Winartasaputra et al.(38) introduced the reaction for fluorometric assays.

During the following years, many additional modifica-tions were made to apply the enzymatic methods to thenewly developed automated laboratory instrumentation(38,48-53). When glycerol oxidase (glycerol:oxygen oxido-reductase; EC not assigned) became available, a new ap-proach was presented. Terada et al. (39, 54) and Gauhl etal. (55) combined the oxidase reaction with the reactionsequence of Trinder (40), which coupled peroxidase to aphenol derivative and 4-aminoantipyrine. Glycerol oxi-dase, however, is nonspecific and reacts with many glycols,such as ethylene glycol and propylene glycol (16). Greaterspecificity was achieved by oxidative phosphorylation. Fos-sati and Prencipe (41) and McGowan et al. (56) used theenzyme glycerol-3-phosphate oxidase (GPO; sn-glycerol-3-phosphate:oxygen 2-oxidoreductase, BC 1.1.3.2 1) in con-junction with a Trinder-type reaction. This method contin-ued to be modified as its application to new instrumentsnurtured innovative changes (57).

Other methods have been developed, such as biolumines-cence assays (58), but have not become popular, eitherbecause of improper blanking that results in poor precisionor because commercial availability has been limited. AnHPLC procedure based on adsorption chromatography andrefractive index detection of triglycerides, presented byAmbrose and Smith (59), appears to be highly specificbecause it does not include di- or monoglycerides in themeasurements. The method has its constraints, owing tothe instrumentation and the time required, but may beworth considering as a reference procedure.

Factors Affecting Test Results

Problems with Chemical Saponification Methods

The hydrolysis of triglycerides to glycerol and fatty acidscan be achieved by chemical saponification, in which thesample is treated with alcoholic KOH or NaOH at hightemperatures and subsequently neutralized with magne-sium sulfate. Enzymes in the samples are inactivated andfatty acids are coprecipitated as magnesium salts. Aftercentrifugation, the clear supernate is used for the assay(42). Ascorbate and other unstable compounds, such asconjugated bilirubin, are decreased or eliminated simply bycontrolling the time in which the sample is subjected to

eat and alkali.However, glucose, lactate, hydroxybutyrate, and resid-

al alcohol remain candidates for potential interference.hospholipids are hydrolyzed to glycerophosphates, which

stable in alkaline solutions. Only those methods thateasure the formation of glycerol 3-phosphate require

heir removal by adsorption, unless the procedures oftavropoulos and Crouch (29) or Chernick (43) are used.echnical problems are mainly related to the high temper-ture, the caustic nature of the reagents involved, and the

ength of time. The temperature inside the sample tubesnot the water bath) must be the same for calibrator andlank; the length of time required varies with the temper-tore and must be equal for each sample within a series.

e stoppers of the sample tubes must be vented to avoidreakage, but evaporation of the alcohol must be pre-ented. The subsequent precipitation of Mg(OH)2 causesolume displacement of the liquid from which an aliquot is

taken for further testing. The reagent blank must includethe saponification step because preparations of KOH arenot always free of glycerol. The volume of magnesiumsulfate needed to neutralize the alcoholic base in aqueoussolutions (reagent blanks and calibrators) is different forproteinaceous samples. To keep all volumes the same, it isadvisable to add fatty acid-free bovine serum albumin(BSA), 60 mg/mL, to blanks and calibrators.

Problems with Enzymatic Cleavage of Triglycerides

Choosing the best lipase (triacylglycerol acylhydrolase;EC 3.1.1.3) requires selection from a wide variety of com-mercially available preparations. We have found that thesource of lipase determines the rate and degree of lipolysis:lipase from Chromobacterium sp. acts with higher speci-ficity toward the glycerides of long-chain fatty acids,whereas the lipase from Pseudomonas sp. hydrolyzes theshort-chain esters much faster. Microbial lipases on themarket have distinct properties, and use of several sourcesmay yield the best results for a specific instrument appli-cation (60-62).

Lipases require emulsified substrates, rather than truesolutions, which explains the different rates of hydrolysisfor tn-, di-, and monoglycerides. Many substances can actas emulsifiers, but lipase preparations that are inhibited byspecific surfactants such as Triton#{174} (Rohm and Haas,Philadelphia, PA) are offered commercially. Other lipasepreparations are markedly activated by such non-ionicsurfactants, by bile acids, and by lauramide or cocoamideesters. Furthermore, the effects of mono- and divalent ionsneed to be tested because various lipases depend on eitherpotassium, calcium, or magnesium ions for optimal perfor-mance. Instead of surfactants, BSA has been suggested asan emusifier for lipid micelles (32).

The presence of a colipase is not required for theseassays. Colipase, a pancreatic protein, negates the inhibi-tion of pancreatic lipase by bile acids (63). In contrast, themicrobial lipases used for triglyceride determinations arenot inhibited, but activated, by bile acids.

Esterase in the lipase mixture (for instance, carboxyl-esterase, EC 3.1.1.1; or triacylglycero-protein acyihydro-lase, EC 3.1.1.34) increases the rate of hydrolysis (64, 65).Microbial lipases generally exhibit both lipolytic and hy-drolytic activity, which are stimulated by proteases, e.g.,chymotrypsin (EC 3.4.21.1), as demonstrated by Bucoloand David (32). The presence of proteases can have anegative effect on the stability of a reagent solution. Di-gesting proteases can attack other enzymes in the assaysystem, leading to poor stability of glycerol kinase andpyruvate kinase at pH �7.0.

Interference from Extraneous Sources: Reagents

Manufacturers of enzymes and biochemical substratesoffer products of refined purity (66); however, they cannotpredict the use of a particular reagent component in aspecific assay system. Although a listed contamination of0.01% may not appear to be significant, its effects mayaccumulate when three or four enzymes are used in areaction sequence. The following interferents (in alphabet-ical order) have been detected in our laboratory and can bepresent in commercially available materials.

Adenylate kinase: ATP:AMP phosphotransferase (EC2.7.4.3) can be a contaminant in glycerol kinase (EC2.7.1.30), lactate dehydrogenase (EC 1.1.1.27), pyruvatekinase (EC 2.7.1.40), and glycerol-3-phosphate dehydroge-

CLINICAL CHEMISTRY, Vol. 36, No. 9, 1990 1607

1608 CLINICAL CHEMISTRY, Vol. 36, No. 9, 1990

nase (NAD�) (EC 1.1.1.8). Depending on the assay condi-tions, interference results from the formation of ADP,which reacts with the AMP present in ATP. (AMP is anormal contaminant in ATP.)

Adenosine 5-diphosphate: ADP is present in APP, espe-cially if reagent solutions are stored or if ATP has beenprocessed with other reagents such as in a lyophilizedproduct. ADP can also be generated by the action ofalkaline phosphatase at pH 7.0-9.0. Its effect can be ne-gated with a reagent blank; but consumption of NADHduring pre-incubation reduces the dynamic range of thereaction, and samples with high concentrations of alkalinephosphatase may no longer give results that are in thelinear range of the assay, in the presence of this contami-nant.

Akohol dehydrogenase: Alcohol:NAD� oxidoreductase(EC 1.1.1.1) is a likely contaminant of enzymes isolatedfrom yeast. Its presence in glycerol kinase from Candidasp. can produce a side reaction with NAD� and alcohol thatis carried over from chemical saponification or introducedas a contaminant in NAD�. If blanking is performedwithout glycerol kinase, the total assay will produce falselyhigh (NAD�-coupled reactions) or falsely low (NADH-dependent reactions) values.

Alkaline phosphatase: Orthophosphoric-monoester phos-

phohydrolase or nonspecific phosphatases are the mostsignificant and severe interferents in triglyceride assays,regardless of methodology. Alkaline phosphatase activityis found primarily in lipase, but some contamination inother enzymes is possible. Alkaline phosphatase reactswith glycerophosphate, dihydroxyacetone phosphate, phos-phoenolpyruvate, and APP at neutral or alkaline pH andcan be inhibited by EDTA and phosphates (67).

Catalase: Hydrogen-peroxide:hydrogen-peroxide oxido-reductase (EC 1.11.1.6) is a contaminant present in oxi-dases and will, therefore, affect assays based on the Trindersequence, where glycerol oxidase or glycerophosphate oxi-dase is utilized. Because catalase competes with peroxidase(donor:hydrogen-peroxide oxidoreductase, EC 1.11.1.7) forH2O2, interference is less when test conditions are optimalfor peroxidase, with regard to pH and hydrogen donor.

Ethylene glycol: Propylene glycol and other glycols reactnonspecifically with glycerol oxidase, which is used in themethod developed by Terada et al. (39), and give falselyhigh values. Ethylene glycol is frequently used to stabilizecontrol sera.

Hexokinase: ATP:D-hexose 6-phosphotransferase (EC2.7.1.1) can be present in yeast enzymes, especially inglycerol kinase from Candida sp. Its presence results infalsely high values for the ultraviolet method with pyru-vate kinase (EC 2.7.1.40) and lactate dehydrogenase, be-cause hexokinase will generate additional ADP from se-rum glucose and ATP.

Lactate dehydrogenase: (S)-Lactate:NAD� oxidoreduc-tase, a trace contaminant in most enzymes, will interferewith the INT/diaphorase method if present in glycerol-3-phosphate dehydrogenase, or with the direct glycerol dehy-drogenase method with NAD�, because it reacts at a slowrate with lactate and hydroxybutyrate in serum (68). Theinterference caused by this reaction is masked if the ratetechniques measure apparent pseudo-first-order kinetics.

NADH-oxidase: “NADH-oxidase” refers to nonspecificactivities in the presence of coenzyme. NADH, for instance,can be oxidized at a slow rate without any apparentsubstrate. NADH-oxidase, which is often listed in manu-

facturers’ certificates of analysis, is always associated withdehydrogenases and need not be a separate protein. Al-though very low rates may not contribute to a significantbias, accumulation of several substances that oxidizeNADH should be avoided.

Oxidases: Oxidases, namely glucose oxidase (EC 1.1.3.4),lactate oxidase (BC 1.13.12.4), and pyruvate oxidase (EC1.2.3.3), must be absent in preparations of GPO to avoidinterference with the Trinder reaction. If the reaction isinitiated with glycerol kinase or lipase, the normal pre-incubation time of approximately 5 mm with GPO may notbe sufficient to remove endogenous glucose and lactic acidthat generate additional H2O2.

Peroxidase: Peroxidase reacts nonspecifically withNADH, and the rate is increased by ketoacids, e.g., keto-

butyrate (unpublished data, S.G.K.). In a single-cuvet se-quential determination of triglycerides and cholesterol (69,70), a significant side reaction can be seen in samples frompatients with ketoacidosis, because the NADH-coupledtriglyceride reaction mixture contains the peroxidase nec-essary for the Trinder-type cholesterol assay.

Peroxides: Peroxides are frequently found in surfactantsand contribute to the pink color that forms during storageof reagents for assays based on the Trinder-type reaction,but they do not interfere if the reagents are blankedcorrectly.

Phospholipase: Phosphatidylcholine cholinephosphohy-drolase (BC 3.1.4.3) and phosphodiesterase (sn-glycerol-3-phosphorylcholine glycero-phosphatehydrolase, EC 3.1.4.2)in lipase preparations act on phospholipids and, in conjunc-tion with phosphatase, result in falsely high values forincorrectly blanked triglyceride assay. Mainly diglyceridesare generated from phosphatidylcholine; therefore, it isadvisable to test the lipase mixture for nonspecific activitybefore use.

Pyruvate: Pyruvate is a breakdown product of phospho-enolpyruvate and can be present in the reaction mixture. Itcontributes to NADH consumption before the start of theglycerol reaction; thus the amount of NADH present maynot be sufficient to assure the range of linearity.

Quantities of enzyme: Quantities of enzyme used for theassays are of great importance. All contaminants that maybe present in commercial materials can be additive. Use oftoo little enzyme, or too much, increases the possibility forfailure. Enzymes stored for a long period may lose activity.If the loss of activity is compensated for by the use ofincreased amounts of raw material, the contaminants maybe correspondingly increased. Contaminants can be morestable than host enzymes.

interference from Endogenous Substances in Serum

As stated previously, the routine determination of triglycerides without extraction or deproteinization may trigger endogenous substances that cause interference in threaction. Interference may not always cause a significanbias, but this potential problem should be anticipated wheprocessing patients’ samples.

Alcohol/alcohol dehydrogenase, lactate/lactate dehydrogenase, and glycerol/glycerol dehydrogenase: These enzymeand their substrates are often increased in serum or plasspecimens from patients with liver diseases or with mycardial infarctionlischemia (71). They can cause a slowsteady increase in absorbance with all methods that utilNAD� and alkaline conditions, because the reaction is nocompleted within the normal pre-incubation time. Interfer

ence can be negated by analyzing serum blanks at the sametime and temperature.

Alkaline phosphatase: Alkaline phosphatase signifi-cantly interferes with all methods, whether as a contami-nant of the reagent preparations, as previously discussed,or as increased concentrations in the specimens. Specimens

with increased concentrations of this enzyme require spe-

cial attention (e.g., pediatric samples). Triglyceride results

may become apparently increased owing to hydrolysis ofphosphoenolpyruvate in the ultraviolet assay involvinglactate dehydrogenase and pyruvate kinase, or to hydroly-sis of glycerol phosphate or dihydroxyacetone phosphate inother procedures. The degree of interference depends onindividual assay conditions and whether serum blankshave been included. Although alkaline phosphatase activ-ity is optimally measured at alkaline pH, its activity at pH7.0 is significant (67). As mentioned previously, phosphateions and EDTA inhibit the activity of alkaline phosphataseand are, therefore, often included in commercial formula-tions.

Ascorbic acid: Ascorbic acid interferes with methodsinvolving 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-phenyltet-razolium chloride with diaphorase (EC 1.6.4.3), by decolor-izing the formazan, and in the Trinder method, by reactingnonspecifically as a hydrogen donor in the peroxidase-catalyzed reaction. Ascorbate oxidase (L-ascorbate:oxygenoxidoreductase, BC 1.10.3.3) added to the reagents effec-tively removes this interference, provided the enzyme ispure and does not generate H202 (72). (Contrary to mostoxidases, ascorbate oxidase generates H20, not H202.)

Ascorbate oxidase need not be added in routine assays,because the small amount of ascorbate ordinarily found inthe assay mixture can be disregarded. The threshold limitof ascorbate excretion is about 0.1 mmol/L, the interferencefrom which is equivalent to only 5% of the upper normaltriglyceride concentration. Serum-blanked methods are notaffected.

Bilirubin: Bilirubin interference with colorimetric meth-ods is well documented and presents a potentially seriousproblem (73). The interference is probably both spectraland chemical and will be different for conjugated andunconjugated bilirubin (74). Bilirubin is oxidized in akinetic reaction with H202, and its decrease in absorbanceis superimposed on the main triglyceride reaction. A serumblank does not compensate sufficiently because the sub-

trate, H202, is not formed from free glycerol in equivalentounts.

The various chromogens used for the peroxidase-coupledsays are affected differently by the bilirubin side-reac-

Son; some are less susceptible to the interference thanthere, presumably because of a higher affinity of peroxi-

for the chromogen (75). The addition of ferrocyamideons has shown various degrees of success, depending on

e chromogens, pH, and spectral wavelength used foreasurements (76). Adding bilirubin oxidase appears toye inherently better results and should be considered as a

retreatment for icteric samples (77).Bilirubin does not interfere with ultraviolet methods

hen measured with precision spectrophotometers vslanks, because NADH has a much lower oxidation poten-�al than does H2O2. Bichromatic readings in some instru-

ents, however, do not provide sufficient spectral accuracy.

e absorbance of NADH at 380 nm lies within a spectralhoulder extending from 340 to 400 nm. This spectral

nsitivity can result in apparent interference. The ab-

sorbance of unbound bilirubin peaks at around 420 nm andis still measurable at the range of 380-400 nm. Theoverlapping spectra of icteric samples are often not consid-ered in the calculation of a bichromatic factor.

Alkaline and acid phosphatases: Phosphatases that orig-inate from erythrocytes may liberate glycerol from glycerolesters and cause chemical interference during hemolysis.Also, spectral interference occurs in some of the colorimet-ric methods if no serum blank is performed.

Phospholipids: Phospholipids from erythrocyte mem-branes can also cause falsely increased triglyceride valuesin unblanked assays. This will occur particularly whenthere is a delay in separation of the serum from the clotthrough the action of phosphoesterases.

Uric acid: Uric acid has been reported as an interferentin the Trinder reaction (78). Peroxidase reacts nonspecifi-cally with many reducing compounds as a hydrogen donor;thus increased uric acid concentrations, as well as ascorbicacid, will affect the results of unblanked techniques.

Problems in Laboratory Technique

Plasma: Plasma treated with heparin or EDTA can beused for the triglyceride assay, provided that most plateletshave been removed. Platelet membranes will rupture inthe presence of surfactant and will release alkaline phos-phatase and lactate dehydrogenase into the assay mixture(79).

Highly lipemic specimens: Highly lipemic specimens con-tain enough chylomicrons to agglomerate as a layer on topof the samples. If sample cups for automated instrumentsare left in the waiting position for more than 30 mm, thesamples are no longer homogeneous. Lipemia also inter-feres if no “clearing factor” is used, namely, if the sampleblank solution does not contain agents to solubilize thelipid micelles. This effect is not always visible, and consid-erable light scattering can be measured even if the cuvetsappear to be clear. Patents exist that claim removal ofturbidity from lipids and (or) proteins by surfactants (64,

80-82); however, the degree to which surfactants can clearlipids completely, with no optical interference, has beenchallenged (83).

Triglyceride concentrations measured by the GPO-cou-pled reaction reportedly are decreased in grossly lipemicsamples because of rapid oxygen utilization that, throughproducing a temporarily anaerobic environment, leads ul-timately to decreased color generation (84). Similar resultswere obtained in the laboratory of one of us (unpublisheddata, S.G.K.) with the use of the Cobas-Bio centrifugalanalyzer (Roche Diagnostic Systems, Nutley, NJ) but notwhen measurements were taken in a manual spectropho-tometer. Adding a liquid with high oxygen retention, suchas a fluorocarbon, can eliminate the problem.

Hydrazine -containing buffers: Hydrazine-containingbuffers are sometimes advocated for methods based on thereduction of NAD� in alkaline pH ranges (28). The ratio-nale is to trap the aldehyde group of the product, dihy-droxyacetone phosphate (or dihydroxyacetone), in order toachieve a more favorable equilibrium. However, NAD�and hydrazine form an ultraviolet-absorbing complex andgenerate a steady increase in absorbance (“creep-reac-tion”). Tris is a better agent for trapping aldehyde groups.

Creep-reactions: “Creep-reactions” are those in which theabsorbance of a reaction mixture continues to change afterthe main reaction has reached an endpoint. These aredefinite signs of side reactions and are not obvious in

CLINICAL CHEMISTRY, Vol. 36, No. 9, 1990 1609

automated instruments. Test results can be either posi-tively or negatively affected. The validation of new testreagents should include an experiment that traces absor-bance time on a recording spectrophotometer, with use of aknown serum sample or standard.

Blanks: Rate measurements based on pseudo-zero-orderkinetics have the obvious advantage of not needing to beblanked, but will mask all creep reactions caused byextraneous or endogenous sources. Enzyme inhibitors maybe added to increase the apparentK� of glycerol kinase (85,

86) and to provide surplus enzyme activity, but contami-nants may not be inhibited. Endpoint assays determine theabsorbance of sample-reagent mixtures before the reactionhas taken place (incubation with reagent not containingglycerol kinase or lipase) and again after the reaction hasbeen triggered and has come to completion. In this case,

contaminants in the trigger reagent are not eliminated.The most nearly accurate, but certainly the most expen-

sive, blanking technique constitutes a separate serumblank for each sample (87). The assay is performed with thecomplete reagent and then repeated with the reagentlacking lipase. This glycerol blank, however, introduces anegative error caused by the clearing effect of lipase. Aseparate free glycerol blank needs to be performed as aserum blank with glycerol reagents, lacking glycerol ki-nase, or with the lipase-containing reagent, lacking one ofthe chromogenic compounds. The latter approach was pat-ented by Tsuda et al. (88) because of the obvious advantageof introducing sequential analysis to automated instru-ments. The requirement of a serum blank in addition to afree glycerol blank, described by Artiss et al. (83), stressedthat surfactants alone clear turbidity only partially, whenthe glycerol blank is assayed. The sample blank is espe-cially important when special medications have been ad-ministered to the patient, because absorbing and (or) fluo-rescent compounds may be present. Davies et al. (89) foundthat metronidazole produced a characteristic spectrum inthe ultraviolet range. Later et al. (90) reported a fluores-cence of unknown origin. Hydroxyurea has also been foundto inhibit glycerol oxidase (91).

The bichromatic measurement of triglyceride assays

may be a good alternative, but spectral conditions greatly

affect the results. The spectrum of hemoglobin, for in-

stance, changes with sample handling. While hemoglobinhas its major peak at 500 nm, the formation of oxyhemo-globin decreases the maximum to 415 nm. Bilirubin, ifunconjugated, peaks at 460 nm; but in the presence ofsurfactants that split the protein bonds, unbound bilirubinabsorbs at 420 nm. Many automated instruments use abichromatic pair of 340/380 nm for ultraviolet tests. If thebandwidth of the respective filter is wide, considerableabsorbance is measured at 380 nm and thus affects theresult, owing to the bilirubin shift. The bichromatic factoris an arithmetic mean of empirical data from many assays;it will correct for the majority of specimens, but not neces-sarily for abnormal samples.

Sample-to-reagent ratios: Most laboratory instrumentsare preset by the manufacturer with regard to sample andreagent volume, wash volume (if applicable), and instru-ment settings. By keeping the sample fraction low, i.e., 0.01

(5 ML), endogenous interfering substances are diluted, sothat the actual effect has minor significance. The instru-ment readout, however, must include a high multiplicationfactor. Errors caused by exogenous sources such as reagentimpurities are then multiplied, and the overall low sensi-

1610 CLINICAL CHEMISTRY, Vol. 36, No. 9, 1990

tivity is reflected in the imprecision of the method. Th�delicate balance between sample volume, sensitivity of thmethod, and instrument-related optical fluctuation must lxtaken into account when applying specific methodology t*laboratory equipment. Glick and Ryder (92) show the effecl

of common serum indices such as hemolysis and icterus orresults with various laboratory instrumentation when usecas prescribed by the manufacturers.

De-ionized water: De-ionized water, when freshly takerfrom a system, does not contain sufficient oxygen for theperoxidase-coupled reactions. The enzyme will display sub.optimal behavior until all solutions are oxygenated. As eresult, calibration factors and blank values may changeduring the first day. Reagents will equilibrate during thefirst 24 h, provided enough “headspace” is given for absorbing oxygen from the air.

Drugs: The direct effects of drugs on the triglyceridemethods have not been studied extensively. Stored sample�from patients receiving intravenous administration of he.parin have shown decreased stability of triglycerides, witI�a corresponding increase in the concentration of free glyc.erol (19,93). Other investigators have claimed high appar.ent concentrations of triglycerides when unblanked tech.niques were used, because of increased free glycerol con�centrations after treatment with nitroglycerine (18). A.�stated previously, others have found interference problem�under certain conditions, such as during treatment withmetronidazole or hydroxyurea (89-91). Little research dateon the effects of special patients’ specimens are available,e.g., those from uremic patients, pre- and postdialysis,When selecting a triglyceride method and its technica]application (instrument setting, blanks, etc.), special atten�tion should be given to the expected peculiarities of thepatients to be tested.

Calibrators: A major problem for triglyceride standardi.zation is achieving a consensus on calibration materials,Because the end product of most triglyceride reactions i�free glycerol, glycerol itself has been used as a convenieni

calibrator. Glycerol, however, can be considered only aprimary standard for the indicator reaction; it is not theanalyte itself and does not participate in all steps of thereaction. Ideally, a primary triglyceride standard shouldcontain, in a pure, well-defined medium, triglycerides rep.resentative of the average composition and complexity othose found in adult humans. Owing to the hydrophobinature of triglycerides and the complex packaging of liprotein particles, a truly parallel primary standardprobably not possible.

Triolein, although not entirely representative of humtriglyceride fatty acid composition, has a similar averagchain length and is hydrolyzed in the reaction processBecause triglycerides in serum contain many fatty aciwith various lengths and numbers of unsaturated bondsthe rate and degree of hydrolysis in any system will depenon the esters involved and on the ability of the specifilipases in the reagent to cleave them. The hydrolysis ooleate, although it represents an average human fatty aciwill not be completely representative of the hydrolysis oall fatty acids. However, it is probably a reasonable altenative.

Another alternative, that of using well-charactesecondary standards, has also been shown to be accuraand precise (94). Chromatographic methods, e.g., HPcould be used to validate the hydrolysis of serum-basecondary standards having complex fatty acid compo&

CLINICAL CHEMISTRY, Vol. 36, No. 9, 1990 1611

Lions. These standard materials could subsequently beanalyzed by methods based on the molar absorptivity of areaction product (95). Storage stability of secondary stan-lards is of critical importance, because matrix characteris-Lics may not be changed.

Improving the accuracy and precision for determiningtriglyceride concentrations is an important component inidentifying individuals at risk for CAD. No definitivemethod has been devised that can act as the “gold stan-�lard” by which all methodology, reagents, and calibratorsare judged. A protocol involving a well-defined referencemethod that can be reliably and reproducibly performedand that is resistant to interfering substances, both endog-enous and exogenous, must be developed, along with pro-cedures to minimize biological variation in sampling. Thepotential methods, interferents, and limitations discussedhere need to be evaluated further before consistently highatandards for measuring triglyceride concentrations can beestablished.

Ref erences1. Expert Panel. Report of the National Cholesterol EducationProgram Expert Panel on detection, evaluation, and treatment ofhigh blood cholesterol in adults. Arch Intern Med 1988;148:36.-69.2. Austin MA. Epidemiologic associations between hypertriglyc-aridemia and coronary heart disease. Semin Thromb Hemostas1988;14:137-42.3. Grundy SM, Vega GL. Hypertriglyceridemia: causes and rela-tion to coronary heart disease. Semin Thromb Hemostas1988;14:149-64.�. Brunzell JD, Albers JJ, Chait A, Grundy SM, Groszek E,�vIcDonald GB. Plasma lipoproteins in familial combined hyperlip-idemia and monogenic familial hypertriglyceridemia. J Lipid Res1983;24:147-55.S. Genest J, Sniderman A, Cianfeone K, et al. Hyperapobetali-poproteinemia. Plasma lipoprotein responses to oral fat load.Arteriosclerosis 1986;6:297-304.S. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the

concentration of low-density lipoprotein cholesterol in plasma,without use of the preparative ultracentrifuge. Clin Chem1972;18:449-502.7. Warnick GR, Knopp RH, Fitzpatrick V, Branson L. Estimating[ow-density lipoprotein cholesterol by the Friedewald equation isadequate for classifying patients on the basis of nationally recom-mended cutpoints. Clin Chem 1990;36:15-9.S. McNamara JR, Cohn JS, Wilson PWF, Schaefer EJ. Calculatedvalues for low-density lipoprotein cholesterol in the assessment oflipid abnormalities and coronary disease risk. Clin Chem

990;36:36-42.McNamara JR. Campos H, Ordovas JM, Peterson J, Wilson

WF, Schaefer EJ. Effect of gender, age, and lipid status on lowensity lipoprotein subfraction distribution. Results from theramingham Offspring Study. Arteriosclerosis 1987;7:483-90.

0. McNamara JR, Millar JS, Bljjlevens E, Schaefer EJ. LDL 4, 5nd 6 may be the most atherogenic LDL species [Abstract]. Clinhem 1988;34:1226.1. Weintraub MS, Eisenberg 5, Breslow JL. Different patterns ofstprandial lipoprotein metabolism in normal, type ha, type ifi,

nd type IV hyperlipoproteinemic individuals. Effects of treatmentith cholestyramine and gemfibrozil. J Clin Invest 1987;79:1110-

2. Cohn JS, McNamara JR, Cohn SD, Ordovas JM, Schaefer EJ.ostprandial plasma lipoprotein changes in human subjects of�fferent ages. J Lipid Res 1988;29:469-79.3. Laurell S, Tibbling G. The use of blood glycerol determination

n the diagnosis of hyperthyroidism. Clin Chim Acta 1968;21:127-2.4. Elm RJ, Ruddel M, McClean S. The variability of the glyceroloncentration in human serum [Abstract]. Clin Chem983;29:1174.5. Porte D Jr, Bierman EL. The effect of heparin infusion onlasma triglyceride in vivo and in vitro with a method for calcu-

lating triglyceride turnover. J Lab Clin Med 1969;73:631-48.16. Lenz PH, Cargill DI, Fleischman Al. Propylene glycol inter-ference in determination of serum and liver triglycerides. ClinChem 1973;19:1071-4.17. Sevenm G, Buongiorno A. Serum triglyceride levels in uremicpatients receiving polyacrylonitrilo dialysis. Clin Biochem1981;14:72-3.18. Ng RH, Guilmet R, Altaffer M, Statland BE. Falsely highresults for triglycerides in patients receiving intravenous nitro-glycerin [Tech Brief]. Clin Chem 1986;32:2098-9.19. Hortin GL, Cole TG, Gibson DW, Kessler G. Decreased stabil-ity of triglycerides and increased free glycerol in serum fromheparin-treated patients. Clin Chem 1988;34:1847-9.20. Chowdhury FR, Rodman H, Bleicher SJ. Glycerol-like con-tamination of commercial blood sampling tubes. Am J Med Tech1972;38:25-7.21. Ryder K, Glick M, Bertram 5, Schechter B. Effect of dry-skinproducts on Ektachem triglyceride results [Abstract]. Clin Chem1986;32:1411.22. Cheung CK, Swaminathan R. Effect of detergent on triglycer-ide assay [Letter]. Clin Chem 1987;33:202.23. Van Handel E, Zilversmit DB. Micromethod for the direct deter-mination of serum triglycerides. J Lab Clin Med 1957;50: 152-7.24. Kessler G, Lederer H. Fluorimetric measurements of triglyc-erides. In: Skeggs LT Jr, ed. Automation in analytical chemistry.New York: Mediad, 1966:341-4.25. Carlson LA, WadstrOm LB. Determination of glycerides inblood serum. Clin Chim Acta 1959;4:197-205.26. Wieland 0. Glycerol. In: Bergmeyer HU, ed. Methods of enzy-matic analysis, 1st ed. New York: Academic Press, 1965;211-4.27. Eggstein M, Kreutz FH. Eine neue Bestimmung der Neutral-fette in Blutserum und Gewebe. I. Mitteilung. Prinzip, Durchfuh-rung und Besprechung der Methode. Klin Wochenschr1966;44:262-7.28. Klotzsch 5, Serrichio M, Furedi R. An automated colorimetricmethod for the specific determination of serum triglycerides. In:Advances in automated analysis, Vol. 1. New York: Mediad,1973;111.29. Stavropoulos WS, Crouch RD. Method for determination oftriglycerides and glycerol. US Patent 4 001 089. The Dow Chem-ical Company. January 4, 1977.30. Stavropoulos WS, Crouch RD. Determination of triglyceridesand glycerol. US Patent 4 142 938. The Dow Chemical Company.March 6, 1979.31. Stavropoulos WS, Crouch RD. A new colorimetric procedurefor the determination of serum triglycerides [Abstract]. Clin Chem1974;20:857.32. Bucolo G, David H. A completely enzymatic determination ofserum triglycerides [Abstract]. Clin Chem 1971;17:664.33. Bucolo G, David H. Triglyceride hydrolysis and assay. USPatent 3 703 591. Calbiochem Corp. November 6, 1970; November21, 1972.34. Bucolo G, David H. Quantitative determination of serumtriglycerides by the use of enzymes. Clin Chem 1973;19:476-82.35. Megraw RE, Dunn DE, Biggs HG. Manual and continuous-flow colorimetry of triglycerols by a fully enzymic method. ClinChem 1979;25:273-8.36. Bowie L, Gochman N. A new spectrophotometric, enzymaticdetermination of serum triglycerides [Abstract]. Clin Chem1973;19:656.37. Rautela GS, Hall RG Jr, Bekiesz CL, Wermus GR. A kineticmethod for rapid and automatic measurement of triglycerides inbiological fluids [Abstract]. Clin Chem 1974;20:857.38. Winartasaputra H, Mallet VN, Kaun SS, Guilbault GO. Fluo-rometric and colorimetric enzymic determination of triglycerides(triacylglycerols) in serum. Clin Chem 1980;26:613-7.39. Terada 0, Uwajima 1, Mihara A, Amsaka K. Glycerol oxidaseand process for the production thereof and method for the quanti-tative determination of glycerol by glycerol oxidase. US Patent4 202 941. Kyowa Hakko Kogyo Co., Ltd. May 13, 1980.40. Trinder P. Determination of glucose in blood using glucoseoxidase with an alternative oxygen acceptor. Ann Clin Biochem1969;6:24-7.41. Fossati P, Prencipe L. Serum triglycerides determined colori-metrically with an enzyme that produces hydrogen peroxide. ClinChem 1982;28:2077-80.

42. Schmidt FH, von Dahl K. Zur Methode der enzymatischenNeutralfett-Bestimmung in biologisehen Material. Z Klin ChemKlin Biochem 1968;6:156.-9.43. Chernick SS. Determination of glycerol in acyl glycerols. In:Colowick SP, Kaplan NO, Lowenstein JM, eds. Methods in enzy-mology. Vol. 14: Lipids. New York: Academic Press, 1969:627-30.44. Winiarski RJ, Ladden PJ, Rumbin SM. A single enzymaticreagent for serum triglyceride determination on clinical chemistryanalyzers [Abstract]. Clin Chem 1978;24:1019.45. Hagen JH, Hagen PB. An enzycnic method for the estimationof glycerol in blood and its use to determine the effect of noradren-aline on the concentration of glycerol on blood. Can J BiochemPhysiol 1962;40: 1129-39.46. Grossman SH, Mollo E, Ertingshausen G. Simplified, totallyenzymatic method for determination of serum triglycerides with acentrifugal analyzer. Clin Chem 1976;22:1310-3.47. Sugiura M, Oikawa T, Hirano K, et al. A simple colorimetricmethod for determination of serum triglycerides with lipoproteinlipase and glycerol dehydrogenase. Clm Chim Acta 1977;81:125-30.48. Van Oers MM, Scholtis RJH, van de Calseyde JF, KuypersAMJ. Assay of triglycerides using the Perkin-Elmer C-4 Auto-matic Analyzer. Z Klin Chem Klin Biochem 1971;9:516-9.49. Mourad N, Zager R, Neveu P. Semiautomated enzymaticmethod for determining serum triglycerides by use of the Beckman“DSA 560.” Clin Chem 1973;19:116-.8.50. Bucolo G, Yabut J, Chang TY. Mechanized enzymatic deter-mination of triglycerides in serum. Clin Chem 1975;21:420-4.51. Barbour HM. Enzymatic determination of cholesterol andtriglycerides with the Abbott Bichromatic Analyzer. Ann ClinBiochem 1977;14:22-8.52. Lehnus G, Smith L. Automated procedure for kinetic measure-ment of total triglycerides (as glycerol) in serum with the Gilfordsystem 3500. Chin Chem 1978;24:27-31.53. Naito HK, Gatautis VJ, Galen RS. The measurement of serumtriacylglycerol on the Hitachi 705 chemistry analyzer [Abstracti.Clin Chem 1985;31:948.54. Uwajima T, Akita H, Ito K, Mihara A, Aisaka K, Tirada 0.Formation and purification of a new enzyme, glycerol oxidase andstoichiometry of the enzyme reaction. Agric Biol Chem1980;44:399-406.55. Gauhl H, Seidel H, Lang G, Roder A, Ziegenhorn J. Methodand reagent for the determination of glycerol. US Patent 4 399218. Boehringer Mannheim GmbH. August 16, 1983.56. McGowan MW, Artiss JD, Strandbergh DR, Zak B. A peroxi-dase-coupled method for the colorinietric determination of serumtriglycerides. Clin Chem 1983;29:538-42.57. Sharma A, Artiss JD, Zak B. A method for the sequentialcolorimetric determination of serum triglycerides and cholesterol.Clin Biochem 1987;20:167-72.58. Werner M, Gabrielson DG, Eastman J. Ultramicro determi-nation of serum triglycerides by bioluminescent assay. Chin Chem1981;27:268-71.59. Ambrose RT, Smith JW. The determination of triglycerides inserum by HPLC [Abstract]. Clin Chem 1981;27:1040.60. Huany C, Roy AV. Lipase composition for glycerol esterdetermination. US Patent 4 056 442. Dow Chemical Company.November 1, 1977.61. Nix PT, Santoro JM, Stephens JE. Enzymatic glyceride hy-drolysis. US Patent 4 394 445. Worthington Biochemical Com-pany. July 19, 1983.62. Battey Y, Fuhrer J, Persen R, Matteo MR. An improvedenzymatic assay for the determination of serum triglycerides onthe CentrifiChem� Analyzer [Abstract]. Clin Chem 1980;26: 1024.63. Borgstrom B, Erlanson-Albertson C, Wieloch T. Pancreaticcolipase: chemistry and physiology [Review]. J Lipid Res1979;20:805-16.64. Wahlefeld AW. Triglycerides determination after enzymatichydrolysis. In: Bergmeyer HU, ed. Methods of enzymatic analysis,2nd ed. Vol. 4. New York: Academic Press, 1974:1831-5.65. Wahlefeld AW, Mollering H, Gruber W, Bernt E, Roeschlau P.Determination of triglycerides. US Patent 3 862 009. BoehringerMannheim GmbH. January 21, 1975.66. Biochemical catalogs and specifications of Boehringer Mann-heim Corp., Indianapolis, IN; Genzyme Corp., Boston, MA; Sigma

1612 CLINICAL CHEMISTRY, Vol. 36, No. 9, 1990

Chemical Corp., St. Louis, MO; US Biochemical Corp., ClevelancOH.67. Fernely HN. Mammalian alkaline phosphatase. In: Boyer PEed. The enzymes, 3rd ed. New York: Academic Press, 197 1:417-4’68. MacRae AR, Amesbury LD, Oommen BP. Positive interfeience by LDH in the assay of triglycerides using Bio-Rad kits. CliiBiochem 1981;14:22C.69. Zoppi F. Single-cuvet sequential determination of triglyceiides and cholesterol. Clin Chem 1985;31:2036-9.70. Sullivan DR, Kruijswijk Z, West CE, Kohlmeier M, KataMB. Determination of serum triglycerides by an accurate enzymatic method not affected by free glycerol. Chin Cher1985;31:1227-8.71. McQueen MJ, Garland IWC, Morgan HG. “Glycerate dehydrcgenase” activity in myocardial infarction and myocardial ischemia. Clin Chem 1972;18:275-9.72. Danniger J, Deneke U, Lang G, Michal G, Roeschlau FMethod for the determination of substrates or enzyme activitie�US Patent 4 168 205. Boehringer Mannheim GmbH. Septembe18, 1979.73. Perlstein MT, Thibert RJ, Zak B. Bilirubin and hemoglobiiinterferences in direct colorimetric cholesterol reactions usin�enzyme reagents. Microchem J 1977;22:403-19.74. Rutman M, Klotzsch SG. Effect of free and conjugated bilirubin on triglycerides determination by various methods as applieto Technicon RA”-Systems [Abstract]. Clin Chem 1986;32:1187.75. Siedel J, Staepels J, Ziegenhorn J. Enzymatic color reagent fothe assay of serum and urine creatinine without interference b:even grossly elevated sample bilirubin levels [Abstract]. CliiChem 1988;34:1189-90.76. Spain MA, Wu AHB. Bilirubin interference with determination of uric acid, cholesterol, and triglycerides in commerciaperoxidase-coupled assays, and the effect of ferrocyanide. CliiChem 1986;32:518-21.77. Feldbruegge DH, Koch DD, Calkins DM. Use of bilirubiioxidase as a means of overcoming bilirubin interference with thiHitachi 737 cholesterol method [Abstract]. Clin Chem 19&34:1185.78. Fleming JK, Gadsden RH Sr, Kabbani I.Definitive characterization of uric acid as an interferent in peroxidase indicatereactions and a proposed mechanism of action. Clin Biochen1988;21:27-32.79. Sanders GTB, Terraneo G, Dois JLS. Artifactual increase iiplasma lactate dehydrogenase activity caused by a surfactan[Letter]. Clin Chem 1987;33:192.80. Monte AA, Chiang C. Surfactant containing reagent formulations for assaying biological specimens and methods for preparin�same. US Patent 3 997 470. Mallinckrodt Inc. December 14, 197681. Batz HG, Draeger B, Wahlefeld AW, Weimann G, Gruber WEliminating turbidity in serum undergoing photometric assay. UlPatent 4 184 848. Boehringer Mannheim GmbH. January 221980.82. Klose 5, Buschek H, Schlumberger H. Compositions anmethod for reducing turbidity in samples. US Patent 4 282 00Boehringer Mannheim GmbH. August 4, 1981.83. Artiss JD, Strandbergh DR, Zak B. Elimination of free glyerol interference in a colorimetric enzymic triglyceride assay. CliChim Acts 1989;182:109-16.84. Shephard MDS, Whiting MJ. Falsely low estimation ofglycerides in lipemic plasma by the enzymatic triglyceride methwith Inodifled Trinder’s chromogen. Clin Chem 1990;36:325-39.85. Muller-Matthesius R, Gruber W. Enzyme-kinetic determition of the concentration of a substrate. US Patent 3 977 9Boehringer Mannheim GmbH. August 31, 1976.86. Nagele U, Wahlefeld AW, Ziegenhorn J. Triglycerides. IBergmeyer HU, ed. Methods of enzymatic analysis, 3rd ed. Vol.Weinheim: Verlag Chemie, 1985:2-18.87. ter Welle HF, Baartscheer T, Fiolet JWT. Influence ofglycerol on enzymatic evaluation of triglycerides [Letter]. C�Chem 1984;30:1102-3.88. Tsuda M, Miike A, Shimizu Y, Yokote Y, Tatano T. Comtion and method for decomposing hydrogen peroxide. US Pate4 416 982. Kyowa Hakko Kogyo Co. November 16, 1981.89. Davies SN, Campbell JB, Gochman N. Potential interferonwith the Technicon SMAC triglyceride measurement, as illustraby metronidazole [Letter]. Clin Chem 1979;28:1979-80.

CLINICAL CHEMISTRY, Vol. 36, No. 9, 1990 1613

90. Later R, Pegon Y, Contreras P, Quincy C. A fluorescentbiological compound that can cause error in enzymic triglyceridesdetermination [Letter]. Clin Chem 1985;31:499-500.91. McPherson PA, Brown KD, Agarwal RP, Threatte GA, Jacob-son RJ. Hydroxyurea interferes negatively with triglyceride mea-surement by a glycerol oxidase method. Clin Chem 1985;31:1355-7.92 Glick MR, Ryder KW. Interferographs. Users guide to inter-ferences in clinical chemistry instruments. Indianapolis: ScienceEnterprises Inc., 1987.

93. Roy AV. Free glycerol interferences on the enzymatic deter-mination of serum triglycerides [Abstract]. Clin Chem 1979;25:1073.94. Ham A, Giegel J. Comparison of “primary and secondary”calibration of triglyceride procedures [Abstract]. Clin Chem1975;21:971.95. Nagele U, Hagele EO, Sauer G, et al. Reagent for the enzy-matic determination of serum total triglycerides with improvedlipolytic efficiency. J Clin Chem Clin Biochem 1984;22:165-74.