| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
Departments of
1
Biochemistry,
2
Endocrinology, and
3
Internal Medicine, Hospital de Galdakao, 48960 Vizcaya, Spain.
a Address correspondence to this author at: Laboratorio de Bioquímica, Hospital de Galdakao, Barrio Labeaga s/n. Galdakao, 48960 Vizcaya, Spain. Fax 34-94-4007128; e-mail mesteban{at}hgda.osakidetza.net
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: We evaluated the analytical and clinical performance of two direct methods (Roche and Wako) by analyzing 313 fresh serum samples obtained from dyslipidemic patients in a lipid clinic and comparing them with modified ß-quantification.
Results: Both homogeneous assays displayed excellent precision (CV <2%). The Roche method showed a mean total error of 7.72%, and the Wako method showed a mean total error of 4.46% over a wide range of LDL-C concentrations. The Roche method correlated highly with the modified ß-quantification assay (r = 0.929; y = 1.052x - 168 mg/L; n = 166) and showed a bias of -4.5% as a result of the assigned standard value. The Wako method also correlated highly with ß-quantification (r = 0.966; y = 0.9125x + 104.8 mg/L; n = 145) without significant bias. The Roche method correctly classified 97% of patients with triglycerides <2000 mg/L, 75% of patients with type IIb hyperlipemia (HPL), and 84% of patients with type IV HPL based on the cutpoints of 1300 and 1600 mg/L, compared with 98%, 78.4%, and 89%, respectively, for the Wako method. In dysbetalipoproteinemic patients, both methods have a 30% mean positive bias compared with ß-quantification.
Conclusions: Both direct methods can be a useful alternative when ultracentrifugation is not available for the diagnosis and control of lipid-lowering medication for patients with mixed HPL, but not for patients with type III hyperlipidemia.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Although the measurement LDL-C is important, an easy, reliable, and suitable methodology for LDL-C has never existed in routine laboratories. ß-Quantification currently is considered the reference method (6), but it requires ultracentrifugation, uses large volumes of serum, and is a time-consuming and expensive technique. For that reason, most clinical laboratories estimate LDL-C concentrations in serum by the Friedewald formula (FF) (7) from concentrations of total cholesterol (TC), triglycerides (TGs), and HDL-cholesterol (HDL-C). Although the estimation method correlates highly with ß-quantification, it has certain limitations: it is not valid in specimens with chylomicrons, with TGs >4000 mg/L, or in patients with dysbetalipoproteinemia (8). Indeed, it has been recommended that the FF should be used with precaution in several pathologic states (diabetes, hepatopathy, nephropathy), even if TG concentrations are between 2000 and 4000 mg/L (9).
The NCEP Working Group on Lipoprotein Measurements (10) has
recommended that the LDL-C concentration be determined with a total
analytical error not exceeding ± 12% (
4% imprecision and
4% inaccuracy) to guarantee correct patient classification into the
NCEP risk categories. It is difficult to obtain this analytical quality
with FF because each component’s analytical error is added.
To improve this situation, various methods have been developed based on selective chemical precipitation or immunoseparation of LDL (11)(12) and subsequent cholesterol quantification. However, results have not been very good: the methods are quite hard to perform and expensive, and TG interferences and substantial bias compared with the reference method have been reported (13).
Since 1998, different homogeneous methods have been described that use different approaches that can be run in most clinical chemistry analyzers without the need for a separation step. These methods seem to have an important role in the future and should be analytically and clinically validated before being introduced into routine clinical laboratories.
The aim of the present study is to assess the analytical and clinical performance of two new direct methods marketed by Roche and Wako. The results obtained with these procedures were compared with the ultracentrifugation/phosphotungstic acid method (modified ß-quantification method for LDL-C) in samples of adult patients with HPL from a lipid clinic.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-cyclodextrin, and
MgCl2. Its purpose is to reduce the enzymatic
reaction for cholesterol in VLDL and chylomicrons. After a 5-min
incubation, reagent 2 (cholesterol esterase, cholesterol oxidase, and
aminoantipyrine) is added and forms a colored product with the
selectively solubilized LDL-C.
Wako homogeneous LDL-C assay.
The LDL-C assay (cat. no.
411-24017; Wako) contains two ready-to-use reagents. Reagent 1 is a
detergent, which binds and protects LDL from enzyme reactions. The
non-LDL cholesterol reacts with cholesterol esterase and cholesterol
oxidase, producing hydrogen peroxide, which is consumed by a catalase.
After 5 min, reagent 2 is added, the protecting agent is removed from
LDL, and cholesterol esterase and cholesterol oxidase react only with
the LDL-C. The hydrogen peroxide produced yields a color complex with a
salt and peroxidase. The assay was carried out on a Hitachi 911
analyzer and calibrated using the lyophilized material provided by the
manufacturer.
TC, TGs, and HDL-C.
TC and TG concentrations were determined
enzymatically with the CHOD-PAP (cat. no. 1489704; Roche Diagnostics)
and lipase/GPO/PAP (cat. no. 1488902; Roche Diagnostics) methods,
respectively, on a Hitachi 747 automatic analyzer. The total CV was
1.0–1.5% for the cholesterol and 1.1–2.0% for TGs. The HDL-C was
subsequently measured by precipitation with phosphotungstic acid and
MgCl2 (Roche Diagnostics). After incubation, the
apoprotein B-containing lipoproteins were sedimented by
centrifugation, and the cholesterol component was measured in the
supernatant with a CHOD-PAP method on a Hitachi 911 analyzer. The total
CV was 1.3–3.4%. Our laboratory meets the annual standardization for
lipids through the Lipid Standardization Program of the CDC.
Modified ß-quantification method.
A NaCl solution (1.5-mL;
d = 1.006 kg/L) was added to a 1.5-mL serum sample and
centrifuged at 541 000g for 2.5 h at 16 °C in a TLA
100.3 rotor with 13 x 56 mm polycarbonate tubes (Beckman
Instruments), and 1.5 mL of the floating layer (containing
TG-rich lipoproteins) was removed (14). The VLDL-cholesterol
(VLDL-C) was calculated as the difference between TC and that in the
ultracentrifuged bottom fraction. The HDL-C in the infranate (the
fraction containing the LDL and HDL) was determined as described above.
The LDL-C concentration (UC-LDL) was then calculated by subtracting
HDL-C from the total infranate cholesterol. The total CV for the period
of study was <10%.
Friedewald calculation.
LDL-C was estimated by the FF as
follows: LDL-C = TC - (TG/5) - HDL-C, in specimens
with TGs <4000 mg/L.
Analytical performance evaluation
Precision.
Two human serum pools with medium and high LDL-C
concentrations and one commercial control (Precinorm L; cat. no.
1240161; Roche Diagnostics) were used. Intraassay imprecision was
calculated as the mean variance obtained for 20 replicate analyses
during 1 day. To assess interassay imprecision, aliquots of control and
pools stored at -40 °C were analyzed over 20 consecutive days.
Total error.
Total error (15) was calculated by
adding the systematic error and the random error. Systematic error,
including the constant and the proportional error, was calculated from
the linear regression equation y = bx + a,
where b is the slope of the regression equation and
represents the proportional error, and a is the
y-axis intercept and represents the constant error.
Systematic error at a particular LDL-C concentration
(xc) was defined as the absolute value
of yc -
xc, where
yc =
bxc + a. The linear
regression equation for calculation of systematic error was based only
on specimens with TGs <2000 mg/L, according to CDC recommendations
(15). Random error was defined as the interassay imprecision
multiplied by 1.96.
Linearity.
We tested the range of linearity by adding various
amounts of isotonic 9 g/L NaCl to the serum pool with high LDL-C. The
analyses were done by triplicate on two different series.
Bilirubin, hemoglobin, and TG interference.
The hemoglobin (up
to 5.3 g/L) and bilirubin (up to 233 mg/L) interference was measured
according to the method of Glick et al. (16). Interference
from TGs was determined by isolating TG-rich lipoproteins by
ultracentrifugation at d= 1.006 kg/L, as described above.
Various amounts of these lipoproteins were added to pooled serum with a
medium LDL-C concentration to obtain TG concentrations of 575–11 505
mg/L. We defined interference as recovery of >5% of the initial
measured value.
Stability study.
Two serum pools (at low and high LDL-C
concentrations) were prepared and stored in aliquots at -20 and
-40 °C. LDL-C concentrations were measured weekly with the two
homogeneous assays over an 8-week period.
Samples
Blood samples were obtained from 291 outpatients and 22 control
subjects, ages 19–65 years, at the lipid clinic of Galdakao Hospital,
Vizcaya, Spain. The outpatients attended hospital for routine checks of
lipid-lowering medication (44%) or HPL diagnosis (56%). Blood was
collected in tubes without anticoagulant from subjects after a 12-h
fast. The samples were allowed to clot at room temperature, and serum
was obtained by centrifugation at 2000g for 15 min. All
direct and ultracentrifugation method analyses were performed within 2
days, and serum was stored at 4 °C. Subjects with fasting TG
concentrations >10 000 mg/L were excluded because of the
manufacturer’s recommendation. The study lasted 4 months. All
procedures were in accordance with the Helsinki Declaration of 1975, as
revised in 1996.
In addition, nine patients with familial type III hyperlipidemia were evaluated separately by the homogeneous methods. Type III hyperlipoproteinemia was diagnosed clinically and analytically based on mixed HPL with a VLDL-C/TG ratio >0.30 and apolipoprotein E2 genotype.
Comparison between methods
LDL-C concentrations measured by the Roche (n = 166) or Wako
(n = 145) methods and modified ß-quantification were compared.
For this purpose, samples were classified into four groups according to
their TC and TG concentrations: (a) normolipidemia, defined
as cholesterol 2400 mg/L and TGs 2000 mg/L; (b) type IIa
HPL, defined as cholesterol >2400 mg/L and TGs 2000 mg/L;
(c) type IIb HPL, defined as cholesterol >2400 mg/L and TGs
>2000 mg/L; and (d) type IV HPL, defined as cholesterol
2400 mg/L and TGs >2000 mg/L. Subjects from the control group were
subsequently included in the normolipidemic group, except for two, who
were included in type IIa HPL.
Diagnostic performance
According to NCEP Adult Treatment Panel guidelines (5),
the management of hyperlipidemic patients, using either dietary or drug
therapy, is based on LDL-C cutoff points (<1000, <1300, and <1600
mg/L). Misclassification occurs if a true LDL-C set point is within the
range for desirable risk and the reported LDL-C is in the range for
high risk (or vice versa). We assessed the ability of two homogeneous
assays to correctly classify subjects at the above medical decision
cutoff points.
The positive predictive value (PPV) of an LDL-C assay at the
x cutpoint was defined as [true positive/(true positive +
false positive)] x 100, where true positive means that the LDL-C
results obtained with both the reference and the test method are
>x mg/L, and false positive means that the test method
LDL-C result is >x mg/L when the reference method LDL-C
result is <x mg/L. The negative predictive value (NPV) of
an LDL-C assay at the x cutpoint was defined as [true
negative/(true negative + false negative)] x 100, where true negative
means that the LDL-C results obtained with both the reference method
and the test method are <x mg/L, and false negative means
that the test method LDL-C result is <x mg/L when the
reference method LDL-C result is 
x mg/L.
We also evaluated the possible utility of the homogeneous assays to diagnose type III HPL by estimating the VLDL-C concentration and VLDL-C/TG ratio. The calculation was made according to the following equation: VLDL-C = TC - LDL-C - HDL-C, where LDL-C is the LDL-C determined by each of the direct methods.
Statistical analysis
Descriptive statistics (means, SDs, and CVs) were calculated with
Microsoft Excel, Ver. 5.0R (Microsoft) and the software
Astute Statistics Add-in program for Microsoft Excel (1993–1995 DDU
software). The nonparametric regression procedure developed by Passing
and Bablok (17) was used for the method comparison. Plotting
the method differences against the method means as recommended by the
Bland and Altman procedure (18)(19) assessed the
degree of agreement between methods. The mean of the differences and
95% limits of agreement were also included in the plots. The Wilcoxon
test was used to compare data that did not follow a gaussian
distribution. Biases were calculated as test procedure result -
reference procedure result. Differences were considered significant at
P <0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. The total CVs for all three concentrations were 1.3–2.0% for
the Roche method and 1.2% for the Wako method. According to the
NCEP performance goals, LDL-C must be determined with an imprecision
4%. The two homogeneous assays have met this performance criterion.
|
Total error
The linear regression equations based on specimens with TGs <2000
mg/L for systematic error calculations were y =
1.011x - 72 mg/L for the Roche method and
y = 0.923x + 103.1 mg/L for the Wako method.
The total error calculated at different concentrations used in the
precision study was 5.53–10.69% for the Roche method and 2.60–5.62%
for the Wako method (Table 1
). The total analytical error of the two
homogeneous assays was <12%, as recommended by the NCEP
(10).
Linearity
The Roche method was linear up to at least 4100 mg/L. The
regression equation for the measured LDL-C vs the expected LDL-C values
from a dilution series was: y = 1.013x
- 27.5 mg/L. The Wako method was linear up to 3000 mg/L; the
regression equation was: y = 0.967x + 1.64
mg/L. Regression analysis showed that slopes and intercept did not
differ significantly from 1 and 0, respectively.
Interferences
Bilirubin concentrations up to 234 mg/L did not interfere with the
homogeneous LDL-C assays (Fig. 1
A). Hemoglobin concentrations >3.7 g/L produced a slight
increase in Roche LDL-C measurements (Fig. 1B
). There was no
interference from TGs up to 7000 mg/L in both methods (Fig. 1C
). TG
interference was 5–10% for both methods for TG concentrations of
7000–10 000 mg/L; TG interference showed a tendency to decrease the
Roche LDL-C result and to increase the Wako result. We have also
observed very different biases depending on the sample used to isolate
the TG-rich lipoproteins.
|
Storage
The means of LDL-C concentrations determined by two homogeneous
assays of two serum pools at baseline and after storage periods of 1–8
weeks at -20 and -40 °C are presented in Fig. 2
. The initial concentration was considered 100%, and the plots
showed no significant change in LDL-C concentrations during the 2
months.
|
Comparison between direct methods and modified ß-quantification
The lipid and lipoprotein concentrations of the study population
are summarized in Table 2
. There were no differences between patients taking
lipid-lowering medications and those not taking medication and between
of populations of the Roche and Wako methods.
|
The comparison-of-methods plot [UC-LDL (x) vs test method
(y)] in all patients showed a regression equation of
y = 0.915x + 104.8 mg/L (r =
0.966; n = 145) for the Wako method and y =
1.052x - 168.8 mg/L (r = 0.929; n
= 176) for the Roche method (Fig. 3
). The correlation coefficients comparing the homogeneous assays
with ß-quantification were highly significant. The Wako method did
not under- or overestimate LDL-C values, as shown by the mean
difference of -14.6 ± 97 mg/L (-0.4% ± 6.7%). With the Roche
method, bias was -95.0 mg/L (-6.6% ± 9.5%). This gave a negative
systematic error.
|
To more clearly illustrate the performance of the homogeneous assays,
we compared the results in HPL groups separately (Table 3
). In the normolipidemic and type IIa HPL groups (patients with
TGs <2000 mg/L), the correlation coefficients between the direct
methods and ß-quantification were highly significant, and the
regression lines were near identity. Agreement with ß-quantification
using the Altman-Bland technique (Fig. 4
A) was good, but the Roche method showed a negative bias
(-4.5% ± 5.7%) that did not depend on LDL-C concentration.
|
|
The correlation and degree of agreement between the direct methods and
ß-quantification were worse in other HPL groups: the correlation
coefficients were lower and values showed a considerable degree of
dispersion above or below the mean (Table 3
and Fig. 4
, B and C). The
Roche method showed negative biases in type IIb (-7.9% ± 9.6%) and
type IV (-14.3% ± 18.3%) HPL. The Wako method did not show a
significant bias: -2.70% ± 11.7% in type IIb HPL and -2.5% ±
8.8% in type IV HPL. In type IIb HPL patients, 25% had values outside
the limits of agreement for the Roche method and 16% had values
outside the limits of agreement for the Wako method, whereas in the
type IV HPL group, 19% and 6% of patients had values outside the
limits of agreement for the Roche and the Wako methods,
respectively.
We studied bias between homogeneous assays and ß-quantification
associated with TC, TG, and HDL-C concentrations. There was no increase
in bias when TC, TG, or HDL-C concentrations increased: Roche method,
r = 0.032, 0.078, and 0.006, respectively; Wako method,
r = 0.018, 0.013, and 0.001, respectively. We found a
positive correlation (Roche, r = 0.24; Wako,
r = 0.30) between bias and the VLDL-C/TG ratio (Fig. 5
). This suggests that the VLDL composition may be a source of
variability in the homogeneous assays evaluated.
|
Appropriate classification
We studied the appropriate classification by homogeneous assays
and FF into the NCEP categories with LDL-C values coinciding
within ± 10% of UC-LDL. As seen in Table 4
, LDL-C concentrations determined by the two homogeneous assays
correctly classified >98% of patients with TGs <2000 mg/L
(normolipidemia and type IIa HPL), and the FF correctly classified
95%. The Roche method did not correctly classify 25% and 16% of the
patients with type IIb HPL and type IV, respectively, whereas for the
same groups, the Wako method did not correctly classify 21% and 11%
of the patients. Both direct methods correctly classified 91% of the
patients with TGs between 3000 and 4000 mg/L, whereas the FF correctly
classified only 76%. When TG concentrations were 4000–10 000 mg/L,
the Roche method correctly classified 74.8% of patients and the Wako
correctly classified 86%. We calculated the PPV of each LDL-C method
at the NCEP cutpoints of 1000, 1300, 1600, and 1900 mg/L (Table 5
). The Roche method showed a PPV of 90.3–99.0% and a NPV of
74.4–96.7%, and the Wako method showed a PPV of 92.4–100.0%
and a NPV of 92.5–100.0% compared with the reference method. In
contrast, the FF showed a PPV of 83.3–98.4% and a NPV of
44.4–91.5%.
|
|
Comparison in type III HPL
Because the FF is invalid in type III HPL, it is important to have
an assay that can reliably determine LDL-C concentrations in these
patients. Nine patients with diagnosed type III hyperlipoproteinemia
were analyzed by two homogeneous assays. Their lipid and lipoprotein
values are shown in Table 6
. As seen, the mean LDL-C determined by both direct methods was
significantly higher than that of UC-LDL, with a mean positive bias of
30%.
|
We also calculated VLDL-C concentrations and the VLDL-C/TG ratio by the equation described above. In only one patient (12%) would it have been possible to suspect the presence of ß-VLDL by a ratio between 0.25 and 0.30. In no patients with type III HPL was the calculated ratio >0.30, which implies that these homogeneous assays are not useful for this type of patient.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The NCEP Working Group on Lipoprotein Measurement (10) has been developing recommendations for LDL-C measurement and has recommended further research to develop new assays capable of directly quantifying LDL-C without previous separations or calculations. New assays to directly quantify LDL-C in serum and plasma must be comparable to the accepted reference procedure to maintain linkage with the epidemiological database. The LDL-C isolated by ultracentrifugation contains LDL, intermediate-density lipoproteins, and lipoprotein(a). Because intermediate-density lipoproteins and lipoprotein(a) are atherogenic, the LDL-C measurement actually represents the amount of cholesterol being transported in atherogenic particles. The new direct methods must include these atherogenic lipoproteins. Thus, direct routine methods should be validated against the reference method, and standards and control materials should be traceable to the reference method for LDL-C.
Direct methods for LDL-C are new in clinical laboratories, and they are easily adapted to automatic analyzers to quantify LDL-C concentration without the need for sample pretreatment. Here we present the results of two new homogeneous assays that use different strategies: In the Roche method (21), the LDL-C is determined directly using selective micellar solubilization of LDL-C by a nonionic detergent and the interaction of sugar compound and lipoproteins. In the second step, the cholesterol in the solubilized LDL fraction is determined enzymatically. The Wako homogeneous assay (22) uses the combination of a polyanion and amphoteric surfactant as the protecting reagent for LDL. In the second step, both are removed from LDL by a nonionic surfactant and the LDL-C is determined enzymatically. Although the mechanisms involved in the selectivity of the different surfactants and reagents are not exactly known, they could be related to the differences in net charge, hydrated density, or size of various lipoprotein fractions (23).
The NCEP has clearly laid down the analytical goal for the acceptability of any new assay measuring LDL-C (10): The imprecision of LDL-C determinations should not exceed 4%, and the total error should be <12%. In our experience, the precision of both direct assays is excellent and similar to that of other published reports (24)(25). The total error is higher in the Roche method (5.5–10.3%) than in the Wako method (3.0–5.3%) because of constant negative error. Even with this negative bias, the method meets the total error goal because of the very low imprecision. If we adjust the calibrator to 97% of its preliminary target value, the total error will be 2.5–7.5%. The systematic difference could be easily corrected with appropriated reference materials. The value assignment for calibrators is critical for accuracy from lot to lot of homogeneous assays and makes obvious the problems derived from the lack of a true reference method for LDL-C measurement.
The Wako method was linear up to 3000 mg/L, whereas the Roche method was linear to 4000 mg/L. Therefore, the recovery of high LDL-C concentrations with the Wako method may be lower, and samples should be diluted beforehand. The stability of samples was good for at least 2 months at -20 and -40 °C, which means a practical advantage in clinical laboratories, although longer studies would be necessary if these methods are considered for longitudinal trials. Neither bilirubin nor hemoglobin interference was observed at the concentrations studied with the Wako method. In the Roche method, a positive interference was observed when the hemoglobin concentration was >4 g/L. TG concentrations up to 7000 mg/L did not interfere significantly (<5%) in either method. Although deviation for higher concentrations (7000–10 000 mg/L) did not exceed 10%, the Roche assay tended to experience a negative bias and the Wako assay a positive bias. The TG-related bias may be dependent on the nature and/or composition of the TG-rich lipoproteins used in the interference studies. Furthermore, the method bias data plotted as a function of TG concentration did not show a significant correlation. At serum triglyceride concentrations >10 000 mg/L, both homogeneous LDL-C methods suffered from serious interferences, leading to underestimation of LDL-C concentrations.
The method comparison studies confirmed the negative bias of the Roche method in all patient groups studied. However, there was an acceptable agreement among results with wider limits for type IIb and type IV HPL. With the Wako method, no bias was noticed in any group. Certain dependence with LDL-C values in type IIa HPL was observed, which reflects the lack of linearity for values >3000 mg/L. In type IIb and IV HPL patients, the agreement limits were also more scattered than in patients with TGs <2000 mg/L.
The differences between each direct method and ß-quantification were not related to HDL-C, TC, or TG concentrations. We found in both methods that there is a significant correlation with the VLDL-C/TG ratio. This implies dependence between VLDL composition and the deviations of the direct methods with respect to ß-quantification. An increase in serum TGs is indicative of the presence of a variety of TG-rich lipoproteins. Particles found in hypertriglyceridemic samples (type IIb and type IV HPL) often are a heterogeneous mixture of VLDL, chylomicron remnants, and VLDL remnants. The percentages of cholesterol and TGs are very different across this range of particles, and they could affect the differences found in these patients. When VLDL is rich in TG (lower ratio), direct methods tend to underestimate LDL-C. This, together with the wider deviations found in samples with high TG concentrations (>10 000 mg/L), implied that an incomplete measurement was performed in this situation or that perhaps, all of the LDL of "wide-density" was not measured. Okada et al. (26) have reported that direct LDL values were influenced by the size of the LDL particle. When the VLDL was rich in cholesterol (high ratio), bias was positive. The most outstanding case was observed in patients with type III HPL, for whom the bias was 30%. In these patients, a serum VLDL-C/TG ratio >0.30 (measured in mg/L) reflects the presence of ß-VLDL. This ß-VLDL contains atypical lipoproteins with d <1.006 kg/L and mobility ß in agarose gel. The simplest explanation would be that direct methods recognize the ß-VLDL and other cholesterol-rich VLDLs as LDL; consequently, part of their cholesterol is quantified together with that of the LDL. Other authors (27) have reported that ß-VLDL in patients with type III HPL positively interferes with HDL-C direct methods based on polyethylene glycol-modified enzymes. LDL homogeneous assays have not been successful in detecting type III HPL. The overestimation of LDL-C can cause a type III patient to be misdiagnosed as having type IIb HPL. The presence of abnormal lipoproteins in other patients with diabetes, renal insufficiency, and hepatopathies could lead to similar problems, and further studies are necessary.
According to the NCEP Adult Treatment Panel II, LDL-C <1300 mg/L is considered desirable and LDL >1600 mg/L is considered high. Patients with coronary heart disease are recommended to maintain their LDL-C at <1000 mg/L. An incorrect classification could generate unnecessary drug administration or inappropriate treatment. Ninety-eight percent of patients were correctly classified when TG concentrations were <2000 mg/L by the two homogeneous assays evaluated. Similar data were obtained by the FF. The Roche method could not correctly classify 25% of patients with mixed HPL (type IIb), and the Wako method could not correctly classify 15%. In the hypertriglyceridemic group (type IV), 16% (Roche method) and 12% (Wako method) of patients would be incorrectly classified. However, we must be aware that incorrect classification by the Roche method could be lower if the bias is corrected. From a clinical point of view, these results can be considered an important improvement over the FF, which is used when TG concentrations are <4000 mg/L and which classifies up to 25% of patients incorrectly when TG concentrations are 3000–4000 mg/L. This percentage could be even greater because we used the HDL-C concentration of fraction (d >1.006 kg/L) isolated by ultracentrifugation. Therefore, the type IIb HPL group of patients has to be observed cautiously because of the possible consequences derived from treatment, whereas in type IV HPL patients, errors occur at lower LDL-C concentrations.
This study was deliberately designed with samples from patients attending a lipid clinic who had severe disorders in lipoprotein composition and concentration to evaluate the methods under unsuitable conditions. We also want to remark on the need to determine TC, TG, and HDL-C concentrations to identify dyslipemic patients. It is not advisable to measure LDL-C alone. The correct way to evaluate coronary risk and the effect of hypolipemiant drugs intervention would be to perform a complete lipid profile. When TG concentrations are <2000 mg/L, we recommend the use of the FF to calculate LDL-C concentrations because of its low cost. The direct LDL-C methods can be useful for mixed HPL and hypertriglyceridemia if ultracentrifugation is not available, although they are not valid for the dysbetalipoproteinemia.
In summary, the homogeneous assays evaluated meet the currently established analytical performance goals and show an important improvement for practicability compared with ß-quantification or with other described direct methods, which are time-consuming and cumbersome. The homogeneous methods show a good correlation with ß-quantification, which implies that they are capable of measuring the LDL, intermediate-density lipoproteins, and lipoprotein(a). The constant negative bias seen with the Roche method is attributable to the wrong assignment of the calibrator value, which indicates the importance of this factor and needs be solved by manufacturers. The diagnostic efficiency in patients with mixed HPL and hypertriglyceridemia (types IIb and IV) is good, although 25% and 15% (Roche and Wako methods, respectively) of patients could be classified incorrectly. It is advisable to be cautious with these types of patients. None of these methods could be performed on dysbetalipoproteinemic (type III HPL) patients, but they provide to clinical laboratories the capacity to analyze a high number of samples using an automatic analyzer. They can be considered a good alternative for clinical laboratories when the reference method for the control of hypolipemiant treatment in patients with disorders of lipid metabolism is not available.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-cyclodextrin sulfate. Clin Chem 1998;44:522-531.
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  D. J. Blom, F. H. O'Neill, and A. D. Marais Screening for Dysbetalipoproteinemia by Plasma Cholesterol and Apolipoprotein B Concentrations Clin. Chem., May 1, 2005; 51(5): 904 - 907. [Full Text] [PDF]
|
|
|
|
 |
|
 E. T. Bairaktari, K. I. Seferiadis, and M. S. Elisaf Evaluation of Methods for the Measurement of Low-Density Lipoprotein Cholesterol Journal of Cardiovascular Pharmacology and Therapeutics, January 1, 2005; 10(1): 45 - 54. [Abstract] [PDF]
|
|
|
|
|
|
S. Usui, H. Kakuuchi, M. Okamoto, Y. Mizukami, and M. Okazaki Differential Reactivity of Two Homogeneous LDL-Cholesterol Methods to LDL and VLDL Subfractions, as Demonstrated by Ultracentrifugation and HPLC Clin. Chem., November 1, 2002; 48(11): 1946 - 1954. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. R. Warnick, W. G. Miller, P. P. Waymack, and F. P. Anderson Lack of Agreement of Homogeneous Assays with the Reference Method for LDL-Cholesterol May Not Indicate Unreliable Prediction of Risk for Cardiovascular Disease * Authors of the article cited above respond: Clin. Chem., October 1, 2002; 48(10): 1812 - 1815. [Full Text] [PDF]
|
|
|
|
|
|
W. G. Miller, P. P. Waymack, F. P. Anderson, S. F. Ethridge, and E. C. Jayne Performance of Four Homogeneous Direct Methods for LDL-Cholesterol Clin. Chem., March 1, 2002; 48(3): 489 - 498. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Nauck, G. R. Warnick, and N. Rifai Methods for Measurement of LDL-Cholesterol: A Critical Assessment of Direct Measurement by Homogeneous Assays versus Calculation Clin. Chem., February 1, 2002; 48(2): 236 - 254. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Ordonez-Llanos, A. M. Wagner, R. Bonet-Marques, J. L. Sanchez-Quesada, F. Blanco-Vaca, and F. Gonzalez-Sastre Which Cholesterol Are We Measuring with the Roche Direct, Homogeneous LDL-C Plus Assay? Clin. Chem., January 1, 2001; 47(1): 124 - 126. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and
Wako (
) methods.
