See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/237118317

DextranSulfate-Mg2 Precipitation Procedure

forQuantitation of High Density-Lipoprotein
Cholesterol

Article

CITATIONS READS

4 84

4 authors, including:

Donald A Wiebe
University of WisconsinMadison
77 PUBLICATIONS 3,748 CITATIONS

SEE PROFILE

All content following this page was uploaded by Donald A Wiebe on 13 July 2015.

The user has requested enhancement of the downloaded file.

CLIN. CHEM. 28/6, 1379-1388 (1982)

DextranSulfate-Mg2 PrecipitationProcedurefor Quantitationof High-

Density-LipoproteinCholesterol
Submit- G. Russell Warnick, Joan Benderson, and John with respect to this characteristic, depending on its source and
ters J. Albers, Northwest Lipid Research Clinic, the method of preparation.
Harborview Medical Center, Seattle, WA Therefore, with the objective of defining a separation pro-
98104, and the Department of Medicine, cedure giving results comparable with those by the heparin-
University of Washington, Seattle, WA Mn2+ method but without the associated interference and
98195 other problems, we have considered various alternatives. In
Evaluat- E. Eugene Baillie and Brenda Sexton, previous experiments, lipoprotein precipitation by the com-
ors: Department of Pathology, Anderson bination of Mg2 and dextran sulfate, a synthetic heparin
Memorial Hospital, Anderson, SC 29621 analogue, demonstrated good correlation with precipitation
Ernst J. Schaefer, Deidre Carison, Margaret by heparin and Mn2. Precipitation with dextran sulfate and
Hill, and H. Bryan Brewer, Jr., Molecular Mg2 was also highly reproducible (14). Therefore, we eval-
Disease Branch, NHLBI, National uated published methods with dextran sulfate. In agreement
institutes of Health, Bethesda, MD 20205 with previous investigators (6), we observed that lipoprotein
Donald A. Wiebe and Jacqueline Hazlehurst, precipitation was a function of the molecular size of the dex-
Clinical Chemistry Division, Center for tran sulfate: the larger the dextran sulfate molecule, the
Environmental Health, Centers for Disease greater the tendency for lipoprotein to precipitate. Separation
Control, Atlanta, GA 30333 methods (15, 16) in which 500 000 g/mol preparations of
Assigned Gerald R. Cooper, Clinical Chemistry Division, dextran sulfate were used underestimated HDL cholesterol,
Editor: Center for Environmental Health, Centers in comparison with the heparin-Mn2 and ultracentrifugation
for Disease Control, Atlanta, GA 30333 methods (14). Another standard dextran sulfate preparation
(Mr 15 000) used for lipoprotein separation under similar
conditions (6) did not completely precipitate LDL/VLDL in
Introduction our experience. We therefore predicted that a material of in-
Recent recognition of the importance of HDL cholesterol termediate molecular mass might give separations with Mg2
as a stronginverseriskfactorforcoronary artery (1-4)
disease comparable with those obtained with heparmnand Mn2. After
has led to substantial demand for this assay by clinical and comparing dextran sulfate preparations of various molecular
research laboratories. Various methods, including ultracen- masses, either synthesized in this laboratory or obtained from
trifugation, electrophoresis, and specific precipitation, have commercial sources, a dextran sulfate of 50 000 daltons used
been used to separate the lipoproteins. Although ultracen- with Mg2+ gave separations most comparable with those ob-
trifugation techniques are useful for reference (5), they are tained with heparin and Mn2. We selected Mg2 as the cation
expensive and require equipment and a level of technical so- because of its apparent lack of interference with enzymic
phistication not available in many routine laboratories. The cholesterol assays. We did not use Ca2+, which is similar to
precipitation methods have, in general, been most suitable for Mg2 in lipoprotein precipitation, because it reportedly is less
routine quantitation of lipoproteins. Several such methods effective in separating hypertriglyceridemic specimens (6).
have been described in the recent literature, most of them With dextran sulfate of 50 000 g/mol, lipoprotein separation
based on earlier work of Burstein et al. (reviewed in 6). The was relatively insensitive to the dextran sulfate concentration
method used most extensively in major population studies and less sensitive to Mg2+ concentration than when we used
involves use of heparin in conjunction with MnS+ to precipi- the other dextran sulfate preparations. To obtain appropriate
tate VLDL and LDL (7-9). HDL is then quantitated in terms specificity in lipoprotein separation, we adjusted the Mg2
of the amount of cholesterol remaining in the supernatant concentration for optimum precipitation of LDL and VLDL
solution. This method has been extensively studied (10, 11) without excessive precipitation of HDL.
and modifications have been described (11, 12). The proce- The combination of phosphotungstate and Mg2 (6, 17, 18)
dure has its drawbacks, of which the major one is the reported was not chosen, because our preliminary experiments indi-
interference of the Mn2+ with the enzymic assays for choles- cated that the method gave variable results. Lipoprotein
terol (13). Also, heparin is relatively expensive and may vary separations were quite sensitive to reagent concentrations and
temperature, which resulted in poorer precision than with the
1 Nonstandard abbreviations: HDL, LDL, VLDL, high-, low-, and heparin_Mn2+ and dextran sulfate_Mg2+ procedures (14).
very-low-density lipoproteins, respectively; apoA-I, apolipoprotein
The use of polyethylene glycol (19) was rejected because the
A-I, the major protein of HDL; apoB, apolipoprotein B, the major
protein of LDL and VLDL; EDTA, disodium ethylenediamine- method readily precipitated HDL and other plasma proteins
tetraacetate; CDC, Centers for Disease Control; PIPES, 1,4-piperaz- at concentrations only slightly higher than were necessary for
inediethanesulfonic acid; LRC, Lipid Research Clinics. precipitation of LDL (14). Therefore, if one considers the

CLINICAL CHEMISTRY. Vol. 28, No. 6, 1982 1379

range in concentrations of lipoproteins and other proteins in
specimens, the polyethylene glycol method might be less
specific.
Principle
The specificmechanisms of lipoproteinprecipitationby
polyanions and divalent cations have not been established
with certainty. However, interaction between negatively
charged groups on the polyanions and positively charged
groups on the protein moieties of the lipoproteins are probably
important. Divalent metal ions interact with negatively
charged groups (such as phospholipids) on the lipoproteins
to facilitate formation of insoluble complexes (20). The larger,
lipid-rich lipoproteins, VLDL and LDL, form insoluble
complexes more readily than do the smaller, protein-rich
HDL. The insoluble complexes can be sedimented by low-
speed centrifugation, provided their density is sufficiently
greater than that of the solution. In the presence of high
concentration can be verified by atomic absorption analysis.
Store at 4 #{176}C.
3. Combined working reagent. Mix equal volumes of the
dextran sulfate and MgCl2 stock solutions to produce a solu-
tion with dextran sulfate concentration of 10 g/L and MgCl2
concentration of 500 mmol/L. Alternatively, prepare a com-
bined reagent by dissolving both constituents, dextran sulfate
and MgCl2, in de-ionized water. Adjust pH to 7.0 and the final
volume to give, per liter, lOg of dextran sulfate and 500 mmol
of MgC12. This reagent prepared by either method when
added to specimens will give final concentrations of dextran
sulfate of 0.9 g/L and Mg2 of 45 mmol/L.
Note: Combined solutions prepared with dextran sulfate stock
solutions containing the preservative have been stored for as long
as four months at 4 #{176}C
without apparent change in their ability to
precipitate lipoproteins.
4. Control materials.

concentrations of triglyceride-rich lipoproteins, the insoluble Note: Precision and accuracy in the precipitation procedure can
complexes may either remain suspended in the solution or be monitored by analyzing control materials that are similar to
float to the surface. On the other hand, other plasma proteins patients specimens in analyte and matrix properties. Many of the
materials available commercially do not meet these specifica-
may coprecipitate with the lipoproteins, thereby increasing tions.
the density of the insoluble complex and facilitating sedi-
mentation even in specimens with high triglyceride concen- Prepare control materials by pooling fresh specimens
trations. The extent of protein coprecipitation is a function (plasma with EDTA as anticoagulant or serum) from donors
of the method and conditions of precipitation. Cholesterol or that remaining after laboratory analysis (22). Exclude
remaining in the supernatant solution can be considered to turbid or hypertriglyceridemic specimens so that the total
represent HDL, if sedimentation of LDL/VLDL is complete triglyceride value is less than 1 g/L. The use of two pools, one
and no HDL has precipitated. withHDL cholesterol in the low normal (200-400 mg/L) range
Colorimetric methods involving either strong acid or en- and one in the mid (500 mg/L) range, is recommended. Plasma
zymic reagents have generally been used to quantitate cho- may be treated with thrombin to minimize subsequent fibrin
lesterol (reviewed in 21). Methods may be direct, i.e., applied formation. Transfer aliquots to vials and seal. Quick-freeze
directly to specimens without pretreatment, or may include the contents and store in a freezer with a stable temperature,
a preliminary solvent extraction of the lipids from other in- preferably at or below -60 #{176}C. To allow verification of method
terfering constituents. Because approximately 80% of cho- accuracy, obtain a target value by requesting analysis of ali-
lesterol in HDL from humans is in form of the esters, a pre- quota from an experienced laboratory such as those of the
liminary hydrolysis step may be used to cleave the ester Lipid Research Clinics Program. Alternatively, a commercial
linkage. With enzymic assays hydrolysis is necessary, and may material (Lipid Fraction Control Serum, Hyland Diagnostics,
be accomplished by adding a cholesterol ester hydrolase to the Division of Travenol Laboratories, Inc., Deerfield, IL 60015)
reaction mixture. The 3/3-hydroxy group of free cholesterol is currently available that has an accurate reference value for
is then oxidized by cholesterol oxidase with liberation of hy- HDL cholesterol.
drogen peroxide, which can be coupled to a colorimetric re-
action.
Reagents for Enzymic Cholesterol Analysis
Note: The precipitation procedure described here appears to be
Materials and Methods compatible with any of the available cholesterol assays, provided
their accuracy and reproducibility are adequate. We have analyzed
Precipitation Reagents cholesterol in supernates by using Liebermann-Burchard reagent
according to the Lipid Research Clinics method (9,23); this is both
1. Dextran sulfate. Material (Mr 50 000 5000) with ac- accurate and precise but may not be practical for many clinical
ceptable quality can be obtained from SOCHIBO, SA, Bou- laboratories. We have also used with good results a manual enzymic
logne, France, 92100. Store preferably in a refrigerated des- assay as reported by Cooper et al. (24), which we describe below in
iccator. Prepare a stock solution as follows: Add 2.0 g of dex- detail, because the published procedure may not be available to
tran sulfate to 80 mL of de-ionized water in a beaker, and some readers.
adjust the pH if necessary to 7.0 with dilute HC1. Transfer
quantitatively to a volumetric flask and adjust the final vol- Microbial cholesterol ester hydrolase (EC 3.1.1.13) was
timeto 100 mL to givea concentration of 20 g/L. Store at 4 #{176}C. obtained from Boehringer-Mannheim Biochemicals, India-
A preservative solution containing, per liter, 50 g of NaN3 (J.T. napolis,IN 46250, and microbial cholesterol oxidase (EC
Baker Chemical Co., Phillipsburg, NJ 08865), 1.0 g of chlor- 1.1.3.6) was obtained from Beckman Microbics,Carlsbad,CA
amphenicol, and 0.5 g of gentamicin sulfate (both from Sigma 92005. PIPES buffer (1,4-piperazinediethanesulfonate) was
Chemical Co., St. Louis, MO 63178) can be added to this so- obtained from Research Organics, Cleveland, OH 44125.
lution at the rate of 10 mL/L before final volume adjust- Horseradish peroxidase, Type VI (EC 1.11.1.7), 4-aminoan-
ment. tipyrine, and sodium cholate were purchased from Sigma
2. Stock Mg2 solution. Reagent-grade MgCl2 . 6H20, Chemical Co.; phenol was obtained from J.T.Baker Chemical
which is hydroscopic, should be stored tightly closed or in a Co., and Triton X-100 from Rohm and Haas, Philadelphia,
desiccator to minimize water uptake. Prepare a stock solution PA 1105. The enzyme solutions were stored at 4 #{176}C.
of 1.0 mol/L by dissolving 20.3 g of MgC12 6H2O in 80 mL of
. Prepare the following reagents:
de-ionized water in a beaker; adjust the pH to 7.0 with a dilute 1. PIPES buffer, 50 mmol/L, pH 6.9. Into a 1000-mL glass
NaOH solution. Transfer quantitatively to a 100-mL volu- beaker containing approximately 800 mL of distilled water,
metric flask and adjust the final volume to io#{246} mL. Mg2 add 17.7 g of PIPES buffer and stir for approximately 10 mm

1380 CLINICAL CHEMISTRY, Vol. 28, No. 6, 1982

until the solution clears. Equilibrate the solution to 37 #{176}C
and who have fasted for 12-14 h. Apply a tourniquet to locate the
adjust, if necessary, to pH 6.9. Quantitatively transfer the vein, but remove it during blood collection. To obtain
solution to a 1000-mL volumetric flask; after cooling to am- EDTA-treated plasma, collect blood into tubes containing
bient temperature, adjust the volume to 1000 mL. EDTA (final concentration, 1.5 g/L); mix thoroughly and cool
2. Stock mixed reagent. The following reagents are dis- immediately. For serum, collect blood without anticoagulant
solved in 400 mL of PIPES buffer in a 500-mL volumetric flask: and allow it to stand 30 mm at room temperature for clotting.
4-aminoantipyrine, 0.203 g; sodium cholate, 1.292 g; potassium Separate cells by centrifugation. Complete lipoprotein sepa-
chloride, 7.46 g; and Triton X-100, 1.0 mL. Adjust the final rations the same day.
volume to 500-mL with PIPES buffer.
Notes: The necessity of fasting before specimen collection for HDL
Note: This solution is stable for at least one month at 4 #{176}C. cholesterol measurement has not been established with certainty.
In one study (31) values did not differ significantly between spec-
3. Stock phenol reagent: Prepare from phenol stored in a imens collected after a meal and specimens taken after an overnight
desiccator. Carefully but quickly weigh 3.8 g of phenol crystals fast. However, other studies demonstrate appreciable changes in
and dissolve in PIPES buffer in a 500-mL volumetric flask; HDL cholesterol after a meal high in fat (32) or carbohydrate
(unpublished observation). Therefore, where practical, it is ad-
bring to volume with the buffer.
visable to draw a fasting (12-14 h) specimen. EDTA-treated plasma
Note: This reagent has been stored successfully for as long as a is the preferred medium for lipoprotein quantitation because the
stability of the lipoprotein constituents is enhanced. Heparin and
month at 4 #{176}C
in a tightly closed glass container. citrate anticoagulants can produce severe interference with lipo-
4. Working reagent. Mix 50 mL of the stock mixed reagent protein precipitation. Serum is also commonly used for separation,
and 50 mL of the stock phenol reagent. Add 25 U of cholesterol although HDL results are not exactly equivalent to those obtained
oxidase, 25 U of cholesterol esterase, and 1250 U of peroxidase with EDTA-treated plasma.
HDL separations should be made as quickly as possible after
(based on the respective specific enzyme activities), either
blood collection. Various changes occur during storage, including
from concentrated solutions of the enzymes or by weighing enzymic and nonenzymic transfers of lipids among lipoproteins,
the dry enzyme preparations. pH shifts, bacterial action, and production of metabolites, which
Alternatively, the enzymes can be added to the stock mixed can substantially change lipoprotein composition and their sepa-
reagent at the time of initial preparation. In this instance, ration characteristics (33, 34). If separations cannot be made on
however, the stock mixed reagent should be prepared fresh the day of collection, specimens can be frozen at -15 #{176}C for several
weeks and at -60 #{176}C for as long as two years without substantial
daily. Volumes can be adjusted, depending on the number of
change.
specimens to be analyzed.
5. Calibrators Procedure
Note: Calibration of enzymic cholesterol methods with primary Precipitation step
standards of cholesterol in water or alcohol solution may give in- 1. Allow specimens, control materials, and precipitation
correct cholesterol values for specimens for the following reasons. reagents to equilibrate to room temperature.
Incomplete hydrolysis of esterified cholesterol in specimens by the 2. Using a manual pipette, transfer 1.O-mL aliquots of
cholesterol ester hydrolase reaction, which has been reported
plasma and control material into appropriately labeled tubes.
(25-27), would result in low results because the remaining esterified
cholesterol is not a suitable substrate for the oxidation reaction.
Culture tubes (10 X 75mm) are suitable. Specimen volumes
Also, the alcohol or detergent in aqueous primary standards may can be adjusted so long as a proportionate volume of precipi-
inhibit the enzymes (28) and give different reaction rates in the tant reagent is added.
standards than in specimens. Therefore, we recommend the use of 3. Add 100 jL of the combined dextran sulfate-Mg2
secondary serum-based calibration standards with analyte and working solution to each tube. Immediately after addition of
matrix properties similar to those of the specimens. However, the
this reagent, mix the contents of each tube in sequence for at
secondary standard must have an accurate target value determined
by an accepted reference technique.
least 3 s, with a vortex-type mixer.

We prepared a suitable secondary standard as follows: Note:Alternatively, 50 L each of the separate dextran sulfate and
MgC12 stock solutions can be added sequentially to the 1.0-mL
Reconstitute Lipid Fraction Control Serum (lot no. 4610 specimens, with thorough mixing after each addition.
YOO2A; Hyland Diagnostics), according to the manufacturers
instructions. Dilute with an equal volume of 60 g/L bovine 4. Allow the tubes to stand at room temperature for 10 mm
serum albumin, Cohn Fraction V (Sigma Chemical Co.), in before sedimenting the insoluble lipoproteins by centrifuga-
0.15 mol/L NaCl solution. Freeze aliquots in sealed vials for tion.
subsequent use as a secondary calibrator for assay of choles- Note: Centrifugation for 30 mm with a refrigerated centrifuge (4
terol in supernates. The reported target value for this pool is #{176}C)
attaining 1500 X g is preferred. However, results by this method
1640 mg/L as determined by the modified Abell-Kendall are nearly the same as those obtained with centrifugation in an
method (29) at the Clinical Chemistry Standardization Sec- unrefrigerated benchtop centrifuge at 1000 X g for 15 mm.
tion of CDC. The dilution yields a value of 820 mg/L, which
5. Remove tubes from the centrifuge and inspect super-
is appropriate for assaying HDL specimens. Another suitable
nates for turbidity. Obtain an aliquot of the clear supernates
secondary calibrator is Standard Reference Material 909, a for cholesterol analysis, or transfer the supernatant solution
human serum pool available from the National Bureau of with a Pasteur pipette to a second labeled vial for later anal-
Standards, Washington, DC 20234. This material has a target ysis.
value for cholesterol assigned by an assay involving isotope
dilution-mass spectrometry, which is a candidate Definitive Note: Any turbidity or cloudiness in the supernates indicates in-
Method for cholesterol (30). Alternatively, an in-house pool complete sedimentation of LDL/VLDL and consequent contami-
nation and overestimation of HDL. This is usually observed in
of human donor or pooled serum or plasma can be prepared
specimens with high triglyceride values.
as described previously (22); dilute with albumin solution (60
g/L) to obtain a total cholesterol value in the 800-1000 mg/L 6. Turbid supernates can be conveniently cleared by one
range. A target value can be established by a suitable reference of the following methods:
method. (a) Without separating the turbid supernate from the
precipitate, add to the separation tube 1.0 mL of 0.15 mol/L
Collection of Specimens NaCl solution and another 100 zL of combined precipitant
Collect blood from the antecubital vein of seated subjects reagent. Mix thoroughly with a vortex-type mixer, then cen-

CLINICAL CHEMISTRY, Vol. 28, No. 6, 1982 1381

trifuge as previously described.Obtain the clear supernate for Results and Discussion
analysis. Supernatant cholesterol must be multiplied X2 to The precipitation method with Mr 50 000 dextran sulfate
correct for this dilution. and Mg2 as described here gives results highly correlated
(b) Alternatively, the turbidity can be removed by ultra- with those of the heparin_Mn2+ procedure. Final reagent
filtration with an 0.22-nm filter as described previously (12). concentrations were established to obtain results for super-
Assemble a 25-mm Swinnex filter holder (Millipore Corp., natant cholesterol approximately 10-20 mg/L lower than re-
Bedford, MA 07130) with a 0.22-sm Millipore filter of sults by the LRC heparin-Mn2 method with Mn2 at 46
25-mm diameter. Within the 0-ring over the filter, insert an mmol/L final concentration (Table 1). The LRC method is
AP15 and an AP2O depth filter of 23-mm diameter. Tighten reasonably accurate in relation to an ultracentrifugation ref-
the assembly firmly and attach a 5-mL syringe at the top fit- erence procedure (10, 11), but has been previously demon-
ting. Pour the turbid supernatant solution into the syringe and strated to overestimate HDL slightly (by 10-20 mg/L) in
force it through the filter with moderate pressure from the EDTA-treated plasma specimens because of incomplete
syringe plunger. Analyze the clear filtrate for cholesterol. precipitation of the apoB-containing lipoproteins (11,35). The
Enzymic cholesterol assay systematic difference in supernatant cholesterol observed
1. Turn on the spectrophotometer and adjust the wave- between the heparin_Mn2+ and the present dextran sul-
length to 500 nm. fate-Mg2 method is quite consistent for normolipidemic men
2. Label tubes for water blank, calibrators, control mate- and women as well as for hypercholesterolemic subjects (Table
rials, and specimens. 1). The bias was smaller for hypertriglyceridemic subjects,
3. Dispense 2.0 mL of enzymic cholesterol reagent into each with their concomitant low HDL values. A larger average
of the tubes and place them in an ice bath. difference was observed in specimens in which both choles-
4. Add 100 zL of water, calibrator, control material, or terol and triglyceride were above normal. However, this dif-
supernate to the appropriate tubes and mix thoroughly. ference was attributable to a single specimen that had su-
Note: Mix specimens thoroughly to assure homogeneity before pernatant cholesterol of 532 mg/L by the heparin-Mn2
pipeting. procedure vs 385 mg/L by the dextran sulfate-Mg2 method,
both values being consistent in duplicate separations. The
5. Transfer tubes to a water bath at 37 #{176}C
for 20 mm.
duplicate heparmn_Mn2+ supernates of this specimen averaged
6. Cool tubes to room temperature and within 15 mm read
96 mg of apoB-associated cholesterol per liter, indicative of
absorbance after adjusting spectrophotometer to zero with
incomplete LDL/VLDL precipitation, while the dextran
the water blank. sulfate-Mg2 precipitates averaged 126 mg of apoA-I per liter,
7. Calculate cholesterol in specimens and control materials
suggesting substantial precipitation of HDL. Except for this
in relation to the calibrator.
one specimen, between-method cholesterol differences were
absorbance of unknown less than 70 mg/L.
Cholesterol unknown =
absorbance of calibrator ApoB is the major protein constituent of LDL and VLDL.
Its measurement in supernates allows detection of incomplete
X cholesterol calibrator
Note: Depending on precision requirements of the particular ap-
precipitation of these lipoproteins. On the other hand, apoA-I,
plication and the precision of the pipettes and spectrophotometer, although present in small amounts in VLDL, is the major
the specimens and (or) calibrators may be analyzed in duplicate. protein constituent of HDL. Therefore, high apoA-I values
in the precipitate fractions can indicate excessive precipitation
Calculation of HDL. Consideration of apoB concentrations in supernates
Cholesterol in supernates must be corrected for the dilution (Table 1) suggests that this dextran sulfate-Mg2 procedure
incurred in adding the precipitant reagents. Multiply the as- more specifically precipitates VLDL and LDL without ex-
sayed cholesterol value by 1.1. When necessary, correct by the cessive HDL precipitation than does the heparin-Mn2
appropriate factor for dilution of turbid supernatant solutions method. Heparin_Mn2+ supernates contained, on the average,
as described above. 10.3 mg of apoB-associated cholesterol per liter, while dextran

Table 1. Lipoprotein Precipitation by Dextran Sulfate and Mg2 (DSM) Compared with Heparin and
Mn2 (HM)
Concn, mg/I, in supernate C

ApoB-assoclated
Concn, mg/L, Cholesterol (LRC) cholesterol Precipitate
total plasma, mean SD DIfference ApoA-l, mg/L
Specimens n Cholesterola Triglycerldeb HM DSM (HM - DSM) HM DSM HM DSM
Overall 48 2497 564 1797 1768 493.7 475.8 179d 10.3 0.75 31.2 47.4
Normolipidemic
Males 20 2184387 1132 482 488.8 468.6 20.2 6.6 0.7 21.6 33.0
Females 9 2266 284 800 212 671.7 655.7 16.0 10.5 2.1 25.0 50.2
Hypercholesterolemic 9 3098 268 1472 518 460.7 447.3 13.4 13.0 0.8 38.2 48.2
Hypertriglyceridemic 9 2315 289 3400 1110 324.1 322.7 1.4 2.9 0 3599 458g

Hypercholesterolemic and 4 3510396 56923598 446.2 400.1 46.1 32.1 0 43.5 88.8
hypertriglyceridemic

a Plasma cholesterol range, 1340 to 3820 mg/L. Plasma triglyceride range, 45010 10 750 mg/L. Separations were performed in duplicate. Values are the
means for duplicate determinations, Paired SD 21.3. Students paired f-test 5.09, significant at p < 0.00 1. a Quantitated by radial immunodiffusion assay (11,
12, 36). Quantitated by radial immunodiffusion assay (11, 12, 37).9 ApoA-l not measured in some lipemic specimens because of losses associated with filtration
of turbid supernates.

1382 CLINICAL CHEMISTRY, Vol. 28, No. 6, 1982

 1000 by the LRC method, the dextran sulfate-Mg2 supernates
averaged 23 mg/L lower than the heparin_Mn2+ supernates
for both serum (AQ 6) and plasma control materials. Results
-J for dextransulfate_Mg2+ supernates obtained by the enzymic
800
assay were slightly lower than those obtained by the LRC
assay on the three pools.
U Between-day precision for cholesterol analysis alone was
600 reflected in a CV of 1.2% by the LRC assay and 2.2% by the
enzymic assay on the low-total-cholesterol (MQ 3) pool. This
is consistent with the CVs for precipitation and cholesterol
assay of 3.9 and 3.6% for the two precipitated pools by the
400 enzymic assay, compared with 2.4and 2.9%for the same pools
by the LRC assay. Precision in the precipitation step and with
LRC cholesterol analysis was slightly better by the dextran
200 sulfate_Mg2+ procedure than by the heparin_Mn2+ proce-
dure. CVs were 2.4 and 2.9% for the two pools by the former
method and 3.9 and 3.0% by the latter procedure.
The dextran sulfate-Mg24 procedure was compared with
0 the LRC heparin_Mn2+ method on 199 routine specimens
0 200 400 600 800 1000
submitted to our laboratory for lipoprotein quantitation.
Cholesterol LRC mg / L Many of the patients from whom these specimens were drawn
were hyperlipidemic, as indicated by the lipid distributions
Fig. 1. Cholesterol in duplicate dextran sulfate-Mg2 supemates
of 48 subjects, as assayed by the Lipid Research Clinic assay (mean SD, and range, in mg/L): total cholesterol 2533 664,
with Liebermann-Burchard reagent (x) and the enzymic assay 1240-5880; total triglycerides 2213 2982, 100-24 000. HDL
(y) as described in Procedure cholesterol results obtained by the two precipitation methods
The dotted line Indicates the line of perfect agreement, y = x. The solid line il- are illustrated in Figure 2. The correlation coefficient of 0.980,
lustrates the relationship between the two cholesterol assays by linear regres- slope of 0.955, and the y-intercept of 8.91 mg/L indicate the
sion: slope 0.958, y-intercept 9.2 mg/L, and r = 0.988.
relationship between the two methods. Cholesterol in the
dextran sulfate-Mg2 supernates averaged 450.4 mg/L, as
sulfate-Mg2 supernates contained <1.0 mg/L. On the other compared with 463.5 mg/L in heparin_Mn2+ supernates, an
hand, the dextran sulfate precipitates contained somewhat average difference of 13.1 mg/L with a paired SD of 27.2 mgfL,
more apoA-I (average, 47.4 mg/L), suggesting there was which indicates a relationship similar to that presented in
slightly more HDL precipitation by this method than with Table 1. Of 199 heparin-Mn2 supernates, 100 (50%) con-
heparin-Mn24 precipitation, in which precipitate fractions tained measurable apoB (average, 6.8 mg of apoB-associated
averaged 31.2 mg/L. The 16 mg/L difference in apoA-I rep- cholesterol per liter), whereas 41(21%) of the dextran sul-
resents approximately 5 mg of HDL cholesterol per liter. fate-Mg2 supernates contained apoB-associated cholesterol
The LRC Liebemann-Burchard cholesterol method gives (average 1.3 mg/L). The AQ6 control pooi, precipitated in
results in excellent agreement with the reference method of duplicate on 25 days with these specimens, exhibited a be-
Abell-Kendall as applied at the ClinicalChemistry Stan- tween-run CV of 2.7% vs 3.8% for the heparmn-Mn2 procedure
dardization Sectionof the CDC (29). In this laboratory, results under the same conditions.
are generally within 10 mg/L of the CDC value and CVs of Of 170 of these specimens, which were initially precipitated
approximately 1% are obtained for control pools. The enzymic from whole plasma, 68 (40%) had turbid supernates by the
cholesterol assay described here, calibrated with Lipid Frac- heparin_Mn2+ procedure. The mean plasma triglyceride value
tion Control Serum, was in good agreement with the LRC in specimens that exhibited turbid heparin-Mn2 supernates
method. The dextran sulfate-Mg24 supernates analyzed by was 4253 mg/L. With dextran sulfate-Mg2 precipitation, 35
the enzymic assay averaged 464.9 mg/L vs 475.8 mg/L by the specimens (20.6%) exhibited supernate turbidity; the plasma
LRC cholesterol assay. The average difference was 10.9 mg/L, triglyceride values in these specimens averaged 5227 mg/L.
with a paired standard deviation of 16.7 mg/L. The agreement Therefore, the dextran sulfate procedure gave better sedi-
(Figure 1) between the two assay methods was good, as indi- mentation of the complexed VLDL/LDL lipoproteins than
cated by the parameters of the linear regression: slope = 0.958, did the standard heparin-Mn2 procedure.
y-intercept = 9.2 mg/L, and correlation coefficient 0.989. Turbidity in supernates results from incomplete sedi-
Results for control materials (Table 2) demonstrate similar mentation of the insoluble lipoprotein complex. Generally
relationships between the two precipitation methods as well associated with hypertriglyceridemic specimens, it occurs
as between the two cholesterol assays. With cholesterol assay when the density of the lipoprotein complex is near that of the

Table 2. Analytical Performance on Control Materials

Cholesterol concn, mean SD, mg/I (and CV, %) a
LRC cholesterol Enzymlc cholesterol
Total cholesterol b 519 6.4(1.2) 517 11(2.2)
HDL cholesterol
Separated with Heparin-Mn2 Dextran sulfate_Mg2+ Dextran sulfate-Mg24
In serum (AQ 6) 437 17 (3.9) 414 10(2.4) 406 16(3.9)
In plasma (in-house) 562 17 (3.0) 539 16 (2.9) 520 16(3.6)

a LRc. Lipid Research Clinics assay with Liebermann-Burchard reagent (9, 23);enzymic, the enzymic assay of Cooper et al. (24)as described in Procedure:
15 determInations each in separate analytical runs performed over a three-month period. b Determined in low total-cholesterol pooled serum (MQ3).

CLINICAL CHEMISTRY, Vol. 28, No. 6, 1982 1383

 I l

I
600
-j
I
I
550

500

Co
450

I
0 200 400 600 800 1000

Heparm - p2 supernatant cholesterol mg/ L 400

Fig. 2. Cholesterol as measured by the Lipid Research Clinics
assay in heparin-Mn24 (x) and dextran sulfate-Mg2' (y) su-
pernates of 199 routine specimens 25 50 100 I50 200
The doffed line Indicates y = .r the solid line illustrates the relationship between 2+ 2+
two precIpitation methods by linear regression: slope 0.955, y-intercept 8.9 Final Mg (Mn ) Concentration (mmol/L)
mg/L r = 0.980
Fig. 3. Lipoprotein precipitation by dextran sulfate and Mg2
as a function of molecular mass and reagent concentration, as
compared with precipitation by heparin and Mn2
solution; this precludes sedimentation under the usual con- (0), heparin at 1.3g/L final concentration wIth Mm2concentration as indicated.
ditions of low-speed centrifugation. In samples with very high Dextran sulfate of Mr 50000 at 0.45 gIL (A), 0.91 g/L (#{149}),
and 4.55 g/L (#{149});
dextran sulfate of Mr 15 000 at 0.91 g/L (a); and dextransulfateof Mr 500 000
triglyceride values the insoluble complex may actually form at 0.91 g/L (0), all with Mg at the Indicated concentrations
a layer at the top of the solution. Cholesterol analysis in a
turbid supernate would substantially overestimate HDL be-
cause of the presence of LDL/VLDL. The nonsedimenting
material in a turbid supernate can be conveniently removed Hypertriglyceridemic specimens that give turbid supernates
by an ultrafiltration procedure previously described for can also be precipitated after removal of the VLDL and chy-
heparin-Mn2 supernates (12). lomicrons, if present, by ultracentrifugation. We subjected
To test this clearing procedure on dextran sulfate-Mg24 28 specimens to ultracentrifugation for 20 hat d 1.006 (9). The
supernates, 10 normolipidemic specimens (3.0 mL each) were infranatant fraction, free of VLDL, was subjected to precip-
treated with 0.3 mL of the combined dextran sulfate_Mg2 itation by the heparin-Mn2 and dextran sulfate-Mg24
solution described above. Half of each treated specimen was methods. The dextran sulfate-Mg2 supernates averaged 428
subjected to centrifugation at 1500 X g for 30 mm at 4 #{176}C to mg/L, compared with 438 mg/L for the heparin-Mn24 su-
produce a clear supernate. The other half was subjected to pernates.
ultrafiltration. Cholesterol in the filtrates averaged 456 mgfL The choice of reagents and separation conditions for the
and in the supernates 473 mg/L. All filtrates were free of apoB, dextran sulfate_Mg2- method were based on results of a series
indicating complete removal of LDL/VLDL. As a further test of experiments to determine the effect of dextran sulfate
of the method, routine hypertriglyceridemic specimens that molecular mass and concentration and Mg2 concentration
gave turbid supernates with dextran sulfate and Mg24 were on lipoprotein precipitation. In Figure 3 lipoprotein precipi-
subjected to ultrafiltration. Of 25 turbid supernates so treated, tation, as a function of molecular size and concentrations of
filtrate cholesterol values averaged 370 g/L. The same speci- the dextran sulfate and Mg24 concentration, is compared with
mens were also precipitated by the heparin-Mn2- method precipitation by heparin and Mn24. In general, the larger the
with filtration or dilution of turbid supernates as necessary; dextran sulfate and the higher the concentrations of dextran
heparin-Mn24 values averaged 382 mgfL. sulfate and (or) Mg2', the more lipoprotein was precipitated.
Alternatively, dextran sulfate-Mg2' supernates that are The titration curves for lipoprotein precipitation with dextran
turbid alter the initial centrifugation can be cleared by dilu- sulfate of Mr 50000 at concentrations of either 0.45 or 0.90 g/L
tion. Without separating the turbid supernate from any were similar to that with heparin and Mn2. By contrast,
sedimented material, simply add a volume of 0.15 mol/L NaCI dextran sulfate of Mr 15 000 produced less lipoprotein pre-
solution equal to that of the initial sample and an additional cipitation at all concentrations, whereas dextran sulfate of Mr
proportionate volume of the combined dextran sulfate-Mg2 500 000 precipitated more lipoprotein at equivalent con-
solution, as described in Procedure. After thorough vortex- centrations.
mixing followed by centrifugation, the supernate is usually Experiments were performed to determine the effects of
clear. To test this method, we added to each of 10 turbid solution ionic strength and pH on lipoprotein solubility with
dextran sulfate-Mg2- supernates 1.0 mL of 0.15 mollL NaC1 the dextran sulfate method. Three each of normolipidemic,
and 0.1 mL of combined dextran sulfate_Mg2+ solution. hypercholesterolemic, and hypertriglyceridemic specimens
Cholesterol values (after dilution correction) in the clear su- were precipitated by dextran sulfate (Mr 50 000) and Mg2
pernates averaged 381 mg/L, compared with 402 mg/L in the at the same final concentrations as those given in Procedure
corresponding heparinMn2 supernates. Therefore, this but with combined reagent solution pH ranging from 6.0 to
method also appears to give acceptable results. 8.0. Results suggested that within the pH range 6.5-7.0 lipo-

1384 CLINICALCHEMISTRY,Vol.28,No.6,1982
protein precipitation was constant, with mean cholesterol
values between 414 and 416 mg/L. At higher or lower pH, Ii-
Table3.Accuracy ofthe Enzymic Cholesterol
poprotein precipitation was slightly less: mean cholesterol by Primary or Secondary
Assay withCalibration
values were 427 mg/L at pH 6.0,425 mg/L at pH 7.5, and 432 Standards
mg/L at pH 8.0. Virtually all of this pH dependence was ob- Enzymlc cholesterol, a mg/L
served in the three hypertriglyceridemic specimens, however. LRC b Secondary standard
In the normolipidemic and hypercholesterolemic specimens, cholesterol, Primary C LIpid fractiond Frozen
supernatant cholesterol values did not vary significantly over mg/L standard control serum plasma
pH 6.0-8.0. On the basis of these results we recommend ad- Assay materials (each n = 12)
justing the precipitant reagent to neutral pH. AQ6 437 378 414 407
The ionic strength of the precipitant reagent solution in-
MQ3 523 490 538 526
fluenced lipoprotein precipitation. The higher the ionic
strength, the less the tendency for lipoprotein precipitation.
Plasma pool 740 647 708 687
When the dextran sulfate solution was prepared in 0.15 mol/L Mean 567 505 553 543
NaC1 solution rather than water, supernatant cholesterol Dextran sulfate_Mg2+ supernates (n = 26)
values averaged approximately 15 mg/L (3%) higher. There- 488 431 479 468
fore, the solutions should be prepared in high-quality water
to obtain the indicated level of accuracy under these condi- a Enzymic assay as described in FcerA,e. 5Assay by Lipid Research Clinics
tions. Also, reagent-grade MgC12 .6 H20 and dextran sulfate
Method with Liebermann-Burchard reagent (9, 10). Accuracy comparablewith
of high quality should be used. reference method (29). Preciset 500 mg/L; Biodynamics/bmc, Indianapolis,
Lipoprotein separation by some precipitation methods is IN 46250. d HylandDiagnostics,Deerfield,IL 60015,dilutedtwofold as described
reportedly a function of the temperature before and during in Ptvcedure. In-house material of donor plasma dilutedfivefold as described
centrifugation. This does not appear to be the case with the inProcedure.
dextran sulfate-Mg24 method. Values for supernatant cho-
lesterol were not significantly different (6 mg/L) when the Accuracy in cholesterol quantitation is in large measure a
incubation and centrifugation steps were done at either 4 or function of the calibration technique. Table 3 illustrates ac-
23 #{176}C.
These results suggest the incubation step can be con- curacy results for a series of control materials analyzed by the
veniently performed at ambient temperatures, and centrifu- LRC Liebermann-Burchard method and by the enzymic
gation need not be performed in a refrigerated centrifuge. method, calibrated with one primary standard and two types
However, where possible, centrifugation at a consistent tem- of secondary standard. The primary standard contained
perature in a refrigerated centrifuge would be preferable. unesterified cholesterol in water solution with detergent. Use
The stability of the combined precipitant reagent was of the primary standard resulted in values for specimens that
evaluated over a period of four months. Combined solutions were low by approximately 10%, even though rate curves
prepared at weekly intervals and containing preservative as demonstrated the reactions had reached equilibrium during
described in the Materials and Methods section were accu- the 20-mm incubation. Calibration of the enzymic assay with
mulated over this period and stored in a refrigerator at 4 #{176}C. a commercial lyophilized human serum (Lipid Fraction
Each of the solutions was then used to precipitate normoli- Control Serum) or human plasma (in-house) gave results
pidemic, hypercholesterolemic, and hypertriglyceridemic similar to the LRC values. The Lipid Fraction Control Serum
specimens. Supernatant cholesterol results did not exhibit any is commercially available with a target value established by
trends that would indicate deterioration of the solutions over the modified Abell-Kendall reference procedure (29). The
the four-month period. The means of duplicate determina- Lipid Standardization Laboratory of the CDC kindly provided
tions on the three pools were all within 20 mg/L. a reference target value for the in-house pool. In other ex-
Mn2 produces a positive interference with many of the periments, not shown, the reaction rate in this enzymic assay
enzymic cholesterol assays (13) but does not interfere with was considerably slower for the primary standard of choles-
Liebermann-Burchard methods. We performed experiments terol with detergent in water solution than for the two sec-
to discover whether the dextran sulfate and Mg2 reagents ondary calibrators, which had reaction rates similar to those
described here produced interference with either the LRC of dextran sulfate-Mg2' supernates.
Liebermann-Burchard or the enzymic cholesterol assay. These results are consistent with reports of other investi-
Plasma, freed of LDL and VLDL by ultracentrifugation at d gators (25-27). The observed underestimation of specimen
= 1.063 g/L, was dialyzed against 0.15 molfL NaC1. The cholesterol has been attributed, at least in part, to incomplete
combined dextran sulfate-Mg2 precipitant solution was hydrolysis of esterified cholesterol in specimens by cholesterol
added in a volume of 0.1 mL/mL of specimen to give final ester hydrolase in the enzymic reagent. Approximately 70-80%
concentrations of 0.90 g of dextran sulfate and 45 mmol of of the cholesterol in specimens is present in the esterified
Mg24 per liter. A control specimen was prepared by diluting form, which, without hydrolysis, is not a suitable substrate for
the same LDL/VLDL-free plasma with water at 0.1 mL/mL. the cholesterol oxidase. Hence, cholesterol in specimens would
Cholesterol was analyzed in each solution 10 times by both the be underestimated to the extent of incomplete hydrolysis in
enzymic and the LRC assays as described. Cholesterol content relation to a primary standard of free cholesterol, which does
of the HDL solution containing dextran sulfate-Mg2- aver- not require hydrolysis. Also, the alcohol or detergent required
aged 325 and 322 mg/L by the LRC and enzymic assays, re- to solubilize unesterified cholesterol in a primary standard
spectively, while the control solution averaged 326 and 323 may interfere with the enzyme activity (28), producing reac-
mg/L by the two assays. Therefore, the dextran sulfate and tion rates for the primary standards that differ from those for
Mg2 reagents described here appear not to interfere with serum or plasma specimens.
either the LRC Liebermann-Burchard or this enzymic cho- Therefore, calibration with a secondary standard may be
lesterol assay. Dextran sulfate of M 500 000 has been re- more appropriate with the enzymic cholesterol assays. The
ported, however, to interfere with another enzymic cholesterol secondary standard should resemble specimens in the pro-
assay (38). The interference may be due to aggregation and portion of esterified and free cholesterol as well as matrix
partial inhibition of some pancreatic cholesterol ester hy- properties. In addition, a reliable target value is essential for
drolases by dextran sulfate (personal communication from obtaining accurate results with a secondary standard. In some
Garry Handelman). applications, the use of both a primary and a secondary cali-

CLINICAL CHEMISTRY, Vol. 28, No. 6, 1982 1385

brator might be warranted. Provided the secondary calibrator
fell within a certain accuracy range in relation to a primary
Table 5. Population Distributions of Plasma HDL
standard, results for unknown specimens could be calculated Cholesterol, mg/L a
in relation to the reference value for the secondary standard. Males b Females C

This approach to calibration, for example, has resulted in Age, PercentIle Percentile
excellent accuracy and precision with the LRC cholesterol yr Mean 5th 95th Mean 5th 95th

method (23). 5-9 555 380 740 532 360 730

10-14 549 370 740 522 370 700
Evaluators Results 15-19 461 300 630 522 350 740
Evaluators E.E.B. and B.S. compared the present dextran 20-24 454 300 630 533 330 790
sulfate-Mg2 method with a commercial method involving
a sodium phosphotungstate solution buffered at pH 5.7 (DMA
25-29 447 310 630 560 370 830
Isopol; Data Medical Associates, Arlington, TX 76011); en- 30-34 455 280 630 561 360 770
zymic cholesterol assay was with the Multistat III Micro 35-39 434 290 620 550 340 820
Centrifugal Analyzer (Instrumentation Laboratory, Lexing- 40-44 443 270 670 578 340 880
ton, MA 02173). The cholesterol assay was calibrated with an 45-49 454 300 640 594 340 870
aqueous standard provided with the Instrumentation Labo- 50-54 441 280 630 620 370 920
ratory kit but diluted threefold with water. Analysis of 33
55-59 476 280 710 622 370 910
serum specimens (20 in duplicate) yielded a mean value of 518
60-64 515 300 740 638 380 920
mg/L for the dextran sulfate supernates, compared with 514
mg/L for the DMA Isopol supernates. The relationship by 65-69 511 300 780 633 350 980
linear regression between the two methods was as follows: 70+ 505 310 750 607 330 920
dextran sulfate-Mg2 = 0.880 DMA Isopol + 65.0 mg/L (r =
0.967). Among-day CVs of 3-6% were obtained on control a Data from Lipid Research Clinics Program for 10 North American populations
materials by the dextran sulfate-Mg2 method with enzymic (40). b Total of 3546 white males. C Total of 3382 white females,
cholesterol assay (Table 4). Application of this method to
paired serum and EDTA-plasma specimens from 20 subjects
produced a mean and SD (mg/L) of 394 88 for serum and nates gave a mean cholesterol value of 1620 mg/L by the Gil-
397 87 for plasma specimens. ford enzymic assay, much higher than the mean value of 440
Evaluators E.J.S., D.C., M.H., and H.B.B. compared this mg/L obtained by the LRC Liebermann-Burchard assay (9)
dextran sulfate_Mg2 procedure with a commercial method of the same supernates.
involving dextran sulfate of Mr 500 000 (Gilford Diagnostics, Evaluators D.A.W. and J.H. compared the dextran sul-
Cleveland, OH 44135). Cholesterol assay was by the Gilford fate-Mg2 precipitation method with a tentative reference
enzymic procedure. On 50 EDTA-plasma specimens, choles- method for HDL (unpublished) that involves heparin-Mn2
terol in supernates by the Gilford method averaged 379 mg/L precipitation of the VLDL-free fraction of plasma obtained
and by the present dextran sulfate method 389 mg/L. The after ultracentrifugation. This approach eliminates the
agreement between the two methods was indicated by the problem of incomplete sedimentation, encountered especially
linear regression relationship: dextran sulfate-Mg2 = 1.04 with hypertriglyceridemic specimens. Supernates of both
Gilford - 4.68 mg/L (r = 0.991). Supernates of both methods methods were analyzed by the modified Abell-Kendall ref-
were free of apoB as determined by radial immunodiffusion. erence method (29) and the dextran sulfate supernates were
ApoA-I in the supernates obtained by the present method with also analyzed by an enzymic cholesterol assay performed es-
the lower Mr dextran sulfate averaged 1104 mg/L vs 1142 sentially as described in the Procedure (24). On serum control
mg/L in Gilford supernates, compared with 1176 mg/L in the materials, cholesterol values obtained by enzymic assay of the
same specimens before precipitation. Use of the lower M dextran sulfate_Mg2 supernates averaged within 12 mg/L
dextran sulfate appeared to give better separation for hy- (Table 4) of the heparin-Mn2 supernates analyzed by the
pertriglyceridemic specimens. modified Abell-Kendall method. The dextran sulfate-Mg2
These Evaluators also reported severe interference with the precipitation method with enzymic cholesterol assay gave
Gilford enzymic cholesterol assay by heparin_Mn2+ super- among-day CVs of 1.8 to 3.4%. By comparison, cholesterol in
nates (13). Analysis of 20 heparin-Mn2 (92 mmol/L) super- dextran sulfateMg2 supernates analyzed by the modified

Table4. Evaluators: Analytical Performance of Dextran Sulfate-Mg2 Method on Control Materials

Target a No.
value, of Mean Bias SD CV,
Evaluator Pool mg/L assays mgIL %
E.E.B. and B.S.b In house 15 375 14.0 3.7
CDC-65b 294 18 253 -41 14.6 5.8
CDC-67 508 18 519 11 15.7 3.0
CDC-69 595 18 623 28 23.0 3.7
D.A.W. and J.H. C CDC-AQ7 442 20 439 -3 8.0 1.8
CDC-35 261 20 273 12 6.1 2.2
CDC-67 508 20 496 -12 9.0 1.8
CDC-69 595 20 593 -2 20.0 3.4

a Heparln-Mn24precipitation and modified AbeIl-Kendall (29) cholesterol assay by Clinical Chemistry Standardization Section CDC. bCommercial enzymic

cholesterol assay andcalibratoron MultistatIll Micro Centrifugal Analyzer (Instrumentation Laboratory). C Enzymic cholesterol assay with secondary serum calibrator
(24).

1386 CLINICAL CHEMISTRY. Vol. 28, No. 6, 1982

 2. Castelli,W. P.,Doyle, J. T., Gordon, T., Hames, C. G., Hjortland,
Table6.RelatIveRisk of Coronary Heart Disease M. C., Hulley,S.B.,Kagan, A., and Zukel, W. J., MDL cholesterol and
as a FunctionofHDL Cholesterol(Average other lipidsin coronary heart disease: The cooperative lipoprotein
Risk = 1.00) phenotyping study. Circulation 55, 767-772 (1977).
HDL cholesterol, 3. Gordon, T., Castelli, W. P., M. C., Kannel, W. B., and
Hjortland,
mg/L Women Men Dawber, T. R., High density lipoprotein as a protective factor against
coronary heart disease: The Framinghain Study. Am. J. Med. 62,
300 1.82 707-714 (1977).
350 1.49 4. Heiss, G., Johnson, N. J., Reiland, S., Davis, C. E., and Tyroler, H.
400 1.94 1.22 A., The epidemiology of plasma high-density lipoprotein cholesterol
450 1.55 1.00 levels. The Lipid Research Clinics Program Prevalence Study. Cir-
culation 62 (suppl. 4), 116-136 (1980).
500 1.25 0.82
5. Havel, R. J., Eder,H.A.,and Bragdon, J. H., The distribution and
550 1.00 0.67
chemical composition of ultracentrifugally separated lipoproteins in
600 0.80 0.55 human serum. J. Clin. Invest. 34, 1345-1353 (1955).
650 0.64 0.45 6. Burstein, M., and Scholnick, H. R., Lipoprotein-polyanion-methl
700 0.52 - interactions. In Advances in LipidResearch 11, R. Paoletti and D.
Kritchevsky, Eds., Academic Press, New York, NY, 1973, pp 67-
108.
Risk estimators from Framin9hamHeart Study (41). 7. Burstein, M., and Samaille, J., Sur un dosage rapide du cholesterol
lieaux a- et aux fl-lipoproteines du serum. Clin. Chim. Acta 5, 609
Abell-Kendall assay averaged 423,278,480, and 575 mg/L for (1960).
pools AQ7, CDC 35, CDC 67, and CDC 69, respectively, all 8. Fredrickson, D. S.,Levy, R. I., and Lindgren, F. T., A comparison
within 28 mg/L of the values for heparin-Mn24 supernates of heritable abnormal lipoprotein patterns as defined by two different
techniques. J. Clin. Invest. 47, 2446-2457 (1968).
by the same cholesterol assay. A fresh plasma specimen ana-
lyzed 20 times with enzymic assay of dextran sulfate-Mg2 9. Manual of Laboratory Operations, Lipid Research Clinics Pro-
supernates averaged 668 mg of cholesterol per liter, compared gram, Lipid and Lipoprotein Analysis, 1. National Heart and Lung
Institute, DHEW Publication No. (NIH) 75-628, 1974.
with 727 mg/L with the modified Abell-Kendall assay of
10. Bachorik, P. S., Wood, P. D., Albers, J. J., Steiner, P., Dempsey,
heparin_Mn2 supernates.
M., Kuba, K., Warnick, G. R., and Karlsson, L., Plasma high-density
lipoprotein cholesterol concentrations determined after removal of
Interpretation other lipoproteins by heparin/manganeseprecipitation or by ultra-
The Lipid Research Clinics Program has reported popu- centrifugation.Clin. Chem. 22, 1828-1834 (1976).
lation distributions of lipids and lipoproteins including HDL 11. Warnick, G. R., and Albers, J. J., A comprehensive evaluation of
cholesterol for 10 North American areas (39, 40). HDL cho- the heparin-manganese precipitation procedure for estimating high
lesterol quantitation was by the LRC Liebermann-Burchard density lipoprotein cholesterol. J. Lipid Res. 19,65-76(1978).
reagent method after heparin-Mn2 precipitation (9). Mean, 12. Warnick, G. R., and Albers, J. J., Heparin/Mn21 quantitation of
5th, and 95th percentile values by sex and five-year age range high-density lipoprotein cholesterol: An ultrafiltration procedure for
are presented in Table 5. The population mean for HDL lipemic samples. Clin. Chem. 24,900-904 (1978).
cholesterol was approximately 550 mg/L in boys, decreased 13. Steele, B. W., Koehler, D. F., Azar, M. M., Blaszkowski, T. P.,.
to approximately 450 mg/L with puberty, and remained nearly Kuba, K., and Dempsey, M. E., Enzymatic determinations of cho-
lesterol in high-density lipoprotein fractions prepared by a precipi-
constant until age 55, after which concentrations increased tation technique. Clin. Chem. 22,98-101 (1976).
slightly. In females the population means were approximately
14. Warnick, G. R., Cheung, M. C., and Albers, J. J., Comparison of
550 mg/L through childhood and early adulthood, after which
current methods for high-density lipoprotein cholesterol quantitation.
values gradually increased. HDL cholesterol values by the Clin. Chem. 25, 596-604 (1979).
present dextran sulfate_Mg2 method would be expected to
15. Kostner, G. M., Enzymatic determination of cholesterol in
average approximately 10 mg/L lower than those values listed high-density lipoprotein fractions prepared by polyanion precipita-
in Table 5. The statistical risk of coronary heart disease as- tion. Clin. Chem. 22, 695 (1976).Letter.
sociated with HDL cholesterol values was estimated in 50- to 16. Finley, P. R., Schifman, R. B., Williams, R. J., and Lichti, D. A.,
80-year-olds in the Framingham Heart Study Population (41). Cholesterol in high-density lipoprotein: Use of Mg2/dextran sulfate
Relative risk multipliers from this study are given in Table in its enzymic measurement. Clin. Chem. 24, 931-933 (1978).
6. HDL cholesterol values on patients can be reported with 17. Burstein, M., Scholnick, H. R., and Morfin, R., Rapid method for
the appropriate reference range and the estimated level of the isolation of lipoproteins from human serum by precipitation with
relative risk (42). polyanions. J. Lipid Res. 11, 583-595 (1970).
18. Lopes-Virella, M. F., Stone, P., Ellis, S., and Colwell, J. A., Cho.
lesterol determination in high-density lipoproteins separated by three
We thank Chien Yu for lipid analysis; Sue Ewens for apolipoprotein different methods. Clin. Chem. 23, 882-884 (1977).
analysis; Margaret Poole, Jan Ginsberg, and Jeri Bedford for assis- 19. Vikari, J., Precipitation of plasma lipoproteins by PEG-6000 and
t.ance in obtaining specimens; and Myrna Anderson and Mary Miller its evaluation withelectrophoresis and ultracentrifugation. Scand.
for the manuscript preparation. We also wish to express our appre- J. Clin. Lab. Invest. 36, 265-268 (1976).
ciation to the Evaluators and to Garry Handelman for their helpful 20. Kim, Y. C., and Nishida, T., Nature of the interaction of dextran
comments. This researchwassupported by grant HL-22285 and by sulfate with high and low density lipoproteins in the presenceof Ca2.
contract N01-HV-12157-L from the Lipid Metabolism Branch of the J. Biol. Chem. 254,9621-9626 (1979).
National Heart, Lung and BloodInstitute. J.J.Albersisan Estab- 21. Zak, B., Cholesterol methodologies: A review. Clin. Chem. 23,
lished Investigator of the American Heart Association. 1201-1214 (1977).
22. Warnick, G. R., Mayfield, C., and Albers, J. J., Evaluation of
quality-control materials for high-density-lipoprotein cholesterol
quantitation. Clin. Chem. 27, 116-123 (1981).
References 23. Bachorik, P. S., Wood, P. D. S., Williams, J.,Kuchmak, M.,
1. Miller, G.J., and Miller, N. E., Plasma-high-density-lipoprotein Ahmed, S., Lippel, K., and Albers, J. J., Automated determination
concentration and development of ischaemic heart-disease. Lancet of total plasma cholesterol: A serum calibration technique. Clin. Chim.
i, 16-23 (1975). Acta 96, 145-153 (1979).

CLINICAL CHEMISTRY, Vol. 28, No. 6, 1982 1387

 24. Cooper, G. R., Duncan, P. H., Hazlehurst, J. S., Miller, D. T., and rin-Mn2' supernatant cholesterol concentration. Clin. Chem. 25,1098
Bayse, D. D., Cholesterol, enzymic method. In Selected Methods for (1979). Abstract.
the Small Clinical Chemistry Laboratory (Set. Methods Clin. Chem., 34. Bachorik, P. S., Walker, R., Brownell, K. D., Stunkard, A. J., and
9), W. R. Faulkner and S. Meites, Eds., Am. Assoc. Clin. Chem., Kwiterovich, P. 0., Determination of high density lipoportein cho-
Washington, DC, 1982, pp 165-174. lesterol in stored human plasma. J. Lipid Res. 21, 608-616 (1980).
25. Cooper, G. R., Ullman, M. D., Hazlehurst, J., Miller, D. T., and 35. Albers, J.J., Warnick, G. R., Wiebe, D., King, P., Steiner, P.,
Bayse, D. D., Chemical characteristics and sources of error of an en- Smith, L., Breckenridge, C., Chow, A., Kuba, K., Weidman, S., Arnett,
zymatic cholesterol method. Clin. Chem. 25, 1074 (1979). Abstract. H., Wood, P., and Shlagenhaft, A., Multi-laboratory comparison of
26. MacAulay, M. A., Jacklyn, C. L., Mathers, J. M., and Storm, V. three heparin-Mn2 precipitation procedures for estimating cho-
A., Continuous-flow enzymatic determination of total serum cho- lesterol in high-density lipoprotein. Clin. Chem. 24, 853-856
lesterol and method standardization with CDC-calibrated pool sara. (1978).
Clin. Chem. 26, 896-902 (1980). 36. Albers, J. J., Cabana, V. G., and Hazzard, W. R., Immunoassay
27. Chu, S. V., and Turkington, V. E., Reaction rate studies of two of human plasma apolipoprotein B. Metabolism 24, 1339-1351
enzymatic cholesterol reagent kits. Clin. Biochem. 11, 70-72 (1975).
(1978). 37. Albers, J. J., WahI, P., Cabana, V. G., Hazzard, W. R., and Hoover,
28. Miner-Williams, W., Surfactant inhibition of cholesterol oxidase. J.J.,Quantitation of apolipoprotein A-I of human plasmahighdensity
Clin. Chim. Acta 101,77-84 (1980). lipoprotein. Metabolism 25,633-644 (1976).
29. Eavenson, E., Grier, 0. T., Cisson, J. G., and Witter, R. F., A 38. Henderson, L. 0., LaGarde, E., and Herbert, P. N., Artifactual
semiautomated procedure for the determination of serum cholesterol reduction of high-density lipoprotein cholesterol estimates after
using the Abell-Kendall method. J. Am. Oil Chem. Soc. 43,652-656 dextran sulfate-Mg2' precipitation. Am. J. Clin. Pathol. 664-668,
(1966). (1980).
30. Cohen, A., Hertz, H. S.,Mandel, J.,Paule,R. C.,Schaffer,
R., 39. Heiss, G., Tamir, I., Davis, C. E., Tyroler, H. A., Rifkind, B. M.,
Sniegoski, L. T., Sun, T., Welch, M. J., and White, E., Total serum Schonfeld, G., Jacobs, D., and Frantz, I. D., Lipoprotein-cholesterol
cholesterol by isotope dilution/mass spectrometry: A candidate De- distributions in selected North American populations: The Lipid
finitive Method. Clin. Chem. 26,854-860 (1980). Research Clinics Progam Prevalence Study. Circulation 61, 302-315
31. Maddison, A., Motawni, R., and Speaight, A. B., Serum high (1980).
density lipoprotein cholesterol determination: A simple modification. 40. The Lipid Research Clinics Population Studies Data Book: The
Clin. Chim. Acta 92, 307-310 (1979). Prevalence Study, 1, U.S. Department of Health and Human Ser-
32. Rose, H. G., and Juliano, J., Regulation of plasma lecithin:cho- vices, NIH Publication No. 80-1527, 1980.
lesterol acyltransferase in man. III. Role of high density lipoprotein 41. Kannel, W. B., Castelli, W. P., and Gordon, T., Cholesterol in the
cholesteryl esters in the activating effect of a high-fat test meal. J. prediction of atherosclerotic disease: New perspectives based on the
Lipid Res. 20,399-407 (1979). Framingham study. Ann. Intern. Med. 90,85-91(1979).
33. Warnick, G. R., and Albers, J. J., High-density-lipoprotein cho- 42. Albers, J. J., and Warnick, G. R., Lipoprotein measurement in
lesterol (HDL CH) quantitation: Effect of plasma storage on hepa- the clinical laboratory. Lab. Manage. 19, 31-38 (1981).

Editors note: The reader is reminded that Selected Methods do

not bear the official imprimatur of the Association, As detailed
elsewhere (Clin. Chem. 19, 1207 (1973), these methods are offered
here for criticism by the world community of users and will be revised
appropriately before being collected into a bound volume, Selected
Methods of Clinical Chemistry. (Methods in the latest volume in this
series, published by the Association in 1982, were not first published
in Clinical Chemistry.) The Committee on Selected Methods at-
tempts to select and evaluate methods that seem durable and gen-
erally useful, which have been checked by several Evaluators to
determine both advantages and disadvantages. Thus sufficient in-
formation is provided to enable the user to know what to expect.
Designation as a Selected Method does not imply superiority in all
respects, nor is the procedure a selected proposed reference method
unless it is so desiknated or has been developed as such by the AACC
Standards Committee. The published procedure should be superior
in terms of evaluation and thus of accurately describing to the user
the characteristics of the method,

1388 CLINICALCHEMISTRY,Vol.28,No.6.1982

View publication stats