National Reference System for Cholesterol

Cholesterol Reference Method Laboratory Network

HDL Cholesterol
Certification Protocol for Manufacturers
November 2002






CHOLESTEROL REFERENCE METHOD LABORATORY NETWORK
HDL Cholesterol Certification Protocol for Manufacturers
General Information
The second report of the National Cholesterol Education Program's (NCEP's)* Adult
Treatment Panel (ATP II) reiterated the recommendations of the first report that total
cholesterol (TC) and low-density lipoprotein cholesterol (LDLC) be used as the primary
indicators of coronary heart disease (1,2). In addition, the second report increased emphasis
on high-density lipoprotein cholesterol (HDLC) concentration in the detection, evaluation,
and treatment of people who may be at increased risk for coronary heart disease. The third
report from the ATP established specific medical decision points, 40 mg/dL (1.0 mmol/L)
and 60 mg/dL (1.6 mmol/L), which were derived from national population studies in which
the HDLC assays were standardized to the reference method (RM) at the Centers for Disease
Control and Prevention (CDC) (3). The NCEP's Working Group on Lipoprotein
Measurement developed recommendations about clinical laboratory measurements of HDLC
(4). This group recommended that the link be maintained between laboratory measurements
and the existing epidemiologic and clinical database. Therefore, the Cholesterol Reference
Method Laboratory Network (CRMLN) links to the database through a protocol for
certification of HDLC methods, based on a fresh sample comparison with a CRMLN
laboratory. The CRMLN laboratories use either the CDC RM or a designated comparison
method (DCM), both of which are standardized to the CDC (see the attached table for the
methods used in each CRMLN laboratory). The CRMLN believes that standardization can
be achieved most effectively through the manufacturers of analytical instruments and
reagents. This document provides the details of how this program works.
The NCEP Working Group recommended that laboratories perform HDLC analyses with
bias 5.0% from the true value (reference method) and precision, as measured by coefficient
of variation (CV) at concentrations 42 mg/dL (1.09 mmol/L), 4.0%; and as measured by
standard deviation (SD) at concentrations <42 mg/dL (1.09 mmol/L), 1.7 mg/dL (0.04
mmol/L) (4). These goals for accuracy and precision suggest that for a single measurement,
the allowable total error would be 13%. Precision can be improved by adherence to accepted
principles of good laboratory practice and quality assurance. Accuracy can be improved by
evaluation of traceability to the accuracy base through a fresh sample comparison with one
of the CRMLN laboratories. Although the CRMLN believes that evaluation of total error
is important, it encourages manufacturers to strive to meet the NCEPs accuracy and
precision recommendations separately.
Two major difficulties exist in standardizing the traditional HDLC methods. The first is the
biases between the various precipitating reagents that have traditionally been used to
separate the LDLC and very low-density lipoprotein cholesterol (VLDLC) from the HDLC.
The second is the altered matrix characteristics of calibrators and controls, a result of the
* Acronyms & Abbreviations: NCEP National Cholesterol Education Program, ATP Adult Treatment Panel, TC
total cholesterol, LDLC low-density lipoprotein cholesterol, HDLC high-density lipoprotein cholesterol, RM
reference method, CDC Centers for Disease Control and Prevention, CRMLN Cholesterol Reference Method
Laboratory Network, DCM designated comparison method, CV coefficient of variation, SD standard deviation,
VLDLC very low density lipoprotein cholesterol, QC quality control.
CRMLN HDL Cholesterol Protocol November 2002 Page 1





manufacturing process, which cause them to react differently from fresh patient specimens in
some analytical systems. Calibrating a matrix-sensitive assay system with reference target
values of commercially prepared materials results in significant biases when fresh patient
specimens are analyzed. Calibrators must be assigned set points that result in accurate
results on fresh patient samples.
Two potential difficulties exist in standardizing homogeneous HDLC methods. The first is
the biases associated with different principles used by the various homogeneous reagents that
are used to differentiate the absorbance from the LDLC and VLDLC from the absorbance
used to quantify HDLC and the resulting different correlation relationships to the RM. The
second potential difficulty is the same as with the traditional methods--the altered matrix
characteristics of calibrators and controls. Calibrators must be assigned set points that result
in accurate results on fresh patient samples.
The only universally reliable approach to establishing appropriate calibration for acceptable
accuracy must be based on comparison studies conducted with fresh specimens. This is the
only approach that addresses all of the difficulties in standardization of both the traditional
precipitation methods and the homogeneous methods.
The RM for HDLC used at CDC is a three-step procedure involving ultracentrifugation,
precipitation with heparin-manganese reagent, and quantification with the Abell-Kendall
reference method for cholesterol. It was developed for the purpose of assigning reference
values to reference materials that are used in CDC's standardization programs. The method
is expensive to operate (it requires the use of an ultracentrifuge), requires a large volume of
sample (5 mL of specimen), and has low throughput (18-hour ultracentrifugation). The
CRMLN developed a DCM for HDLC that overcomes some of the practical disadvantages
of the RM by requiring lower sample volume and having a higher throughput (5). The DCM
used by the CRMLN involves precipitation with dextran sulfate-magnesium reagent and
quantification with the Abell-Kendall reference method for cholesterol. Both methods are
used in the CRMLN, and both are standardized to the same high level of performance.
(DCMs are defined by NCCLS's National Reference System for the Clinical Laboratory
[NRSCL]) in the Guideline NRSCL13-P, The Reference System for the Clinical Laboratory:
Criteria for Development and Credentialing of Methods and Materials for Harmonization of
Results (6).
The CRMLN provides services for manufacturers who wish to certify their HDLC methods.
The certification protocol for manufacturers is based upon the NCCLS Guideline EP9-A,
Method Comparison and Bias Estimation Using Patient Samples (7). The protocol includes
a comparison with the CRMLN laboratory using at least 40 fresh serum specimens. The
protocol also includes analysis of a quality control (QC) material in 20 runs. Demonstration
of absolute average bias 5% of the reference method and total CV 4% qualifies an
analytical system for certification. The set points assigned by the manufacturer to that
system's calibrators should be appropriate to ensure accurate analytical results on patient
specimens in the hands of users.
Because the NCEP recommends that clinical laboratories achieve total error 13% on patient
specimens, total error will be calculated. However, certification will be based on meeting
the recommended goals for accuracy and precision.
CRMLN HDL Cholesterol Protocol November 2002 Page 2

Manufacturers must assume all responsibility for the results and should make the initial
contact with the CRMLN laboratory. Manufacturers are advised to contact a CRMLN
laboratory before beginning this protocol. A list of CRMLN laboratories can be obtained
from the CRMLN Web site at http://www.cdc.gov/nceh/dls/crmln/memberlabs.htm. One set
of fresh samples may be split and used to certify several applications, thus necessitating only
one set of analyses by the CRMLN laboratory.
The Initial Certification Process
Anyone collecting and handling any biological material of human origin MUST observe
Universal Precautions (8).
Purpose
The HDLC certification program is intended to assess the bias of the comparison
method against the DCM or RM under defined conditions. Not all samples that may be
encountered in clinical laboratories are included in the scope of this program. Thus,
the protocol is designed to verify the calibration, not the robustness, of the methods
evaluated.
Manufacturer's Preliminary Responsibilities
Before pursuing certification through the CRMLN, manufacturers should establish that their
analytical instrument systems meet the following standard specifications.
x Instrument system(s) must be capable of producing discrete number values.
x Instrument system(s) must have had all required preventive maintenance procedures
and must be in peak operating condition.
x Precision testing (such as that outlined in NCCLS Guideline EP5-T2, User
Evaluation of Precision Performance of Clinical Chemistry Devices) should be done
to ensure that total precision is 4% (9).
Note: Manufacturers who have the DCM or the RM set up in house may establish its
traceability to the accuracy base by evaluation through the CRMLN. Although this
internal verification does not substitute for certification of the manufacturer's
products, manufacturers can use the DCM or RM in house to check their products
before evaluation of traceability by a CRMLN laboratory.
CRMLN HDL Cholesterol Protocol November 2002 Page 3

Manufacturer's Specimen Collection
Specimens should be fresh; fasting donors are preferred. Use of frozen specimens is not
recommended unless the manufacturer has conducted a careful freeze/thaw study and
determined that use of frozen specimens in the test system would not compromise the results
of the comparison with the accuracy base. The certification protocol is designed to evaluate
analytical method bias only. Therefore, variation from preanalytical sources must be
eliminated or minimized. Manufacturers should carefully follow the protocol for sample
collection and processing described in this section.
The recommended sample matrix is serum; however, the comparison should be performed
using the sample matrix for which the analytical system is designed. Venous serum is the
matrix to be used for all comparisons designed to establish traceability to the accuracy base.
Values for all other blood matrices must be traced to venous serum values through paired
sample comparisons. Alternate blood matrices would include all capillary samples
(including serum) and all anticoagulated samples from both venous and capillary sites. For
example, if a manufacturer's system is designed to analyze capillary plasma and the
manufacturer wishes to be certified for this matrix, the manufacturer should collect paired
venous serum and capillary plasma samples from the patients used in the comparison. The
manufacturer should then analyze the capillary plasma samples and submit the venous serum
samples to the CRMLN laboratory for analysis.
Collect and analyze 40 or more fresh specimens from patients. The HDLC concentration of
these specimens should be distributed over the clinically meaningful range. To accomplish
this, at least five samples are required in each of the concentration regions listed below. The
remaining number of samples (a minimum of 15) should be spread out over the entire range.
Range
20-29 mg/dL (0.52-0.77 mmol/L)
30-39 mg/dL (0.78-1.03 mmol/L)
40-49 mg/dL (1.04-1.29 mmol/L)
50-59 mg/dL (1.30-1.55 mmol/L)
60-69 mg/dL (1.56-1.81 mmol/L)
A worksheet for specimen distribution is attached as an aid.
The HDLC DCM requires that the TG concentration be less than 200 mg/dL (2.26 mmol/L).
The TG concentration of samples to be compared with the HDLC RM is not restricted.
These specimens must:
x be free of interfering substances known to affect the system being tested (e.g.,
hemolysis, icterus, lipemia);
x be collected in sufficient quantity (see sample volume requirements for DCM and
RM below);
x not include samples that the products package insert indicates should be excluded;
CRMLN HDL Cholesterol Protocol November 2002 Page 4






x be collected in sufficient quantity (see sample volume requirements for DCM or RM
testing below); and
x be collected using good laboratory practice (such as outlined in NCCLS Guideline
H3-A4, Procedures for the Collection of Diagnostic Blood Specimens by
Venipuncture (10)).
To make the certification process reflect reality, the protocol followed should mimic, as
much as possible, the conditions in the clinical laboratory. Manufacturers are encouraged to
develop package inserts that include information about time frames for storage and analysis
of patient samples. Storage times and methods should be evaluated as part of the method
development.
Balancing the need for reflecting conditions in clinical laboratories is the need to have the
samples analyzed by the manufacturer be as similar as possible to those analyzed by the
CRMLN laboratory. Serum should be separated from red cells within 2 hours of collection.
The serum for the comparison may be stored up to 8 hours at 4C after separation. Within
this 8-hour period, an aliquot (2.5 mL for DCM or 5.5 mL for RM) of each serum sample
must be frozen (at 70 C or lower) for analysis by the CRMLN laboratory. At the same time
the sample for CRMLN laboratory analysis is frozen, the test method(s) analysis should
begin.
For methods involving separation by precipitation: if the supernates resulting from the
precipitation cannot be analyzed the same day as collected, they can be stored at 4 C or
frozen before analysis, if the manufacturer has documented that no adverse effects result of
such storage conditions.
The minimum amount of serum needed per sample for DCM analysis is 2.5 mL. The
minimum amount of serum needed per sample for RM analysis is 5.5 mL. Samples
intended for analysis by the CRMLN laboratory should be dispensed into cryogenic vials,
frozen immediately (at -70 C or lower), and stored frozen until all runs are complete.
Specimens must be clearly identified using coded identification numbers and no patient
identifiers. Specimens must be securely contained in cryogenic vials to prevent breakage,
leakage, and evaporation.
More than one tube of blood from a single individual may be necessary to obtain enough
volume for the CRMLN laboratory analyses as well as the test method analyses. If this is the
case, harvest the serum, combine the contents, and mix well before aliquoting the samples.
This will not apply when an alternative matrix is evaluated by the test method. Serum must
be provided to the CRMLN laboratory; a separate sample tube from the same person must be
collected at the same time if the test method uses either plasma or whole blood.
Difficulty in obtaining a sufficient amount of sample from one individual can be overcome
by combining two specimens--but no more than two--from separate individuals. This
combined specimen pool should be treated as an unpooled (single) sample. The combined
specimen pool should be mixed well before aliquoting.
CRMLN HDL Cholesterol Protocol November 2002 Page 5




Some of the same samples used for certification of other analytes by the CRMLN may be
used. Because all samples are unlikely to meet the sample distribution guidelines for all
analytes, additional samples should be collected. This will ensure that the guidelines are met
for all analytes.
We strongly recommend more than 40 samples be sent to the CRMLN laboratory to avoid
delays resulting from insufficient samples, outliers, or lab accident.
The CRMLN strongly recommends that manufacturers set aside and store (at -70 C or
lower) additional aliquots of each sample (or supernate if test method uses precipitation),
prepared at the same time that the initial samples or supernates for the certification are
prepared. They can be used for reanalysis if changes in calibration are required to meet
NCEP criteria. When new lots of calibrators, materials, or reagents are prepared, these
frozen samples can provide an important link to the accuracy base during overlap analyses
if a frozen versus fresh comparison has been performed.
Collection of patient samples is the responsibility of the manufacturer. However, CRMLN
laboratories may assist in this process by collecting some or all of the specimens, as long as
the specimens analyzed by the test method are fresh. Alternatively, manufacturers can work
with a local clinic or hospital to collect samples.
Manufacturer's Quality Control
To evaluate total error, estimates of both inaccuracy and imprecision are needed. The
inaccuracy can be obtained from the split sample comparison with the CRMLN laboratory.
However, to estimate imprecision, the CRMLN requires that manufacturers provide QC data
obtained from 20 separate runs. The recommended concentration range for the QC material
is 30 to 60 mg/dL (0.78 to 1.55 mmol/L). A frozen human serum pool is more representative
of fresh patient samples and is preferable over the use of a processed material (e.g.,
lyophilized). The latter may result in higher imprecision leading to a lower error budget for
bias. These materials should be analyzed in single using the complete analytical system
being evaluated. The runs must include those used in the split sample comparison described
below. Use the attached Quality Control Results Form for reporting the results.
Manufacturer's Specimen Analysis
The CRMLN recommends that manufacturers perform the comparison analyses at the
manufacturing site; however, manufacturers may have a clinical laboratory using their
system perform the split comparison with the reference method. Clinical laboratories that
perform successful comparisons on behalf of a manufacturer will also receive a Certificate of
Traceability for HDL Cholesterol.
Calibration is the key to achieving accuracy; therefore, these comparison runs should
represent the conditions recommended to customers. Calibrators should be analyzed along
with the patient samples.
CRMLN HDL Cholesterol Protocol November 2002 Page 6

More than one system may be certified using one set of samples. If more than one lot of
reagents or calibrators are to be certified, all of the fresh samples should be analyzed
with each lot. Each combination of instrument, reagents, and calibrator that meets the
NCEP performance recommendations will be issued a Certificate of Traceability for HDL
Cholesterol.
x Follow the protocol for collecting, dispensing, and storing samples as described on
pages 4 - 6 of this document. Store the samples at 4 C before analysis and begin
analysis of all specimens within 8 hours of collection. Perform the precipitation on
duplicate aliquots of each sample. Randomize the concentrations in the run
sequence. Assign the first set of aliquots from each specimen sequential positions in
the run. Run the duplicate measurement of each specimen in reverse order.
x Include the QC material selected above in every analytical run.
x Analyze the specimens over at least 5 runs, one run per day. Two runs per day are
acceptable, provided that the runs are separated by 2 hours (NCCLS definition of a
run, per EP5-A).
x Run days need not be consecutive, but all testing should be completed within a
reasonable time period (less than 1 month).
x Perform testing following the instructions provided in the package labeling (i.e., do
not follow in-house modifications).
x If an instrument problem develops during a run or internal QC is unacceptable, retest
specimens from that run after the problem has been identified and corrected if
storage of samples or supernates will not affect the result. Samples for analysis by
the CRMLN laboratory should have been frozen at the time the sample was first
precipitated.
x Complete the Information Form, the Fresh Sample Comparison Result Form, and the
Quality Control Results Form. Photocopy the blank forms and save them for future
comparisons.
Shipment of Specimens to the CRMLN laboratory
When the fresh specimens have been analyzed by the test method(s), ship the recorded
results to the CRMLN laboratory along with the frozen serum aliquots.
Contact the laboratory first to arrange for the shipment. Ship the frozen samples on dry
ice by overnight express delivery to the CRMLN laboratory. Samples should be shipped
between Monday and Thursday to ensure delivery during the workweek. Include results of
the analysis of the fresh specimens by the test method and the target values for calibrators
used. Before shipping, ensure that the test specimens are clearly and indelibly labeled and
that the labels will remain secure during shipment and subsequent storage. . Include the
CRMLN HDL Cholesterol Protocol November 2002 Page 7






Information Form, the Fresh Sample Comparison Result Form, and the Quality Control
Results Form with the samples.
A Protocol Checklist is provided. Please refer to the checklist to make sure that all
requirements in this protocol have been met.
Follow current federal and state regulations for handling, packaging, and shipping
potentially biohazardous materials with regard to containment, labeling, and other
procedures.
CRMLN Laboratory Analysis, Data Reduction, and Cost
The CRMLN laboratory will analyze each specimen in duplicate; a minimum of 3 runs will
be used. Manufacturers will be provided with the results and statistical analyses. The
approximate turnaround time for CRMLN laboratory analysis and the data analysis is 3-4
weeks from receipt of samples, but could take longer depending on the analytical workload
at the CRMLN laboratory. If a manufacturer wishes to evaluate more than one analyte (e.g.,
total cholesterol, HDLC and LDLC), the CRMLN laboratory analyses could take longer than
3-4 weeks.
Because both the DCM and RM are manual and tedious and must be maintained within very
stringent limits for accuracy and precision, the assay is costly. All of the CRMLN
laboratories offer HDLC analyses for the same cost. Refer to the attached fee schedule for
current pricing. Manufacturers can test the same set of specimens on several instrument,
calibrator, or reagent systems. There will be a nominal fee for certifying each additional
system to cover data analysis costs.
After meeting all the certification criteria, the manufacturer will be issued a dated Certificate
of Traceability for HDL Cholesterol, stating that the analytical system (including instrument
model, reagent lot, and calibrator lot), has successfully demonstrated traceability to the
accuracy base for HDLC under the conditions tested. The Certificate will list the bias versus
the RM, the total CV, and the calculated total error. A separate certificate will be issued for
each analytical system that successfully meets the NCEP performance recommendations.
The date used on the certificate as the Date of Comparison is the date that the data is
analyzed at the CRMLN laboratory. Certificates expire 2 years after this date.
Manufacturers are encouraged to maintain current certification.
Once the analytical system has been certified, conventional in-house QC procedures should
be adequate to monitor the system. However, if shifts are observed, another direct
comparison with the CRMLN should be undertaken to reset the system for optimal accuracy.
Changes in lots of calibrators or reagents should be carefully checked to maintain accuracy.
If the reagent or calibrator formulations, or the instrumentation are substantially modified, a
new direct comparison will be needed to verify accuracy under the new conditions. This
Method Certification Protocol should be followed in this case.
CRMLN HDL Cholesterol Protocol November 2002 Page 8



The CRMLN publishes a list of manufacturers' systems that have been certified through this
protocol. The list is available at http://www.cdc.gov/nceh/dls/crmln/hdltable.htm and
includes all systems with a current Certificate of Traceability for HDL Cholesterol.
Proficiency surveys, such as those offered by the College of American Pathologists, are
essential in providing feedback on the performance of laboratory analytical systems.
Differences in values have been observed with certain systems using lyophilized survey
pools. Manufacturers would best serve their clients by assessing the validity of these
lyophilized survey pools on their systems and advising users of any offset or bias.
Eventually any interim confusion should be reduced, and better set points, survey pools, and
control materials will allow for better reliability in maintaining accuracy.
Questions about this protocol should be directed to a CRMLN laboratory (see
http://wwwdev.cdc.gov/labstandards/crmln_members.html for a list of CRMLN laboratories).
Copies of this protocol are also available at the CMRLN website.
CRMLN HDL Cholesterol Protocol November 2002 Page 9







Statistical Criteria used for Certification
The criteria used for grading are listed in the following table.
Parameter Criterion Statistical approach
r
2
> 0.975 Linear regression
Bias at 40 mg/dL 5%
Linear regression equation;
NCEP accuracy guideline
Bias at 60 mg/dL 5%
Linear regression equation;
NCEP accuracy guideline
Average % Bias 5%
Mathematical mean of
biases; NCEP accuracy
guideline
Average Absolute % Bias 5%
Mathematical mean of
absolute biases; NCEP
accuracy guideline
Among-run CV 4%
CV of QC results; NCEP
precision guideline
t-test of bias Not significant at = 5% See below
Within-method outliers 1 allowed EP9-A (7), see below
Between-method outliers
None allowed, but may
eliminate one sample
EP9-A (7), see below
CRMLN members participate in surveillance to evaluate their performance versus the CDC
accuracy base. They are required to meet very strict performance criteria, which are listed in
the following table.
Performance Criteria for CRMLN Laboratories
Accuracy Criterion Imprecision Criterion
HDL Cholesterol bias 1 mg/dL SD 1 mg/dL
The bias of the CRMLN laboratory is the most important factor that needs to be controlled to
make appropriate decisions about the performance of test systems (11). Although the
CRMLN has achieved a very low bias compared with CDC, some bias still exists. Statistical
analysis has been used to study the distribution of the biases obtained by the CRMLN
laboratories.
For HDL cholesterol, results of statistical analysis of the survey data collected from
September 1998 through February 2000 show a non-Gaussian distribution. The total number
of events was 200. The mean bias was -0.2 mg/dL with a SD of the bias of 0.6. The biases,
by percentile, were -0.6 mg/dL for the 25
th
, -0.2 mg/dL for the 50
th
(median), and 0.1 mg/dL
for the 75
th
. However, since test systems are evaluated based on whether or not they meet
an accuracy criterion of 5%, statistical analysis of the percent bias was also performed.
The mean percent bias was -0.5% with a SD of the bias of 1.4%. The percent biases, by
th th th
percentile, were -1.4% for the 25 , -0.4 % for the 50 (median), and 0.3% for the 75 . We
have determined that the CRMLN can allow test systems an additional -0.5% to 1.5% bias.
CRMLN HDL Cholesterol Protocol November 2002 Page 10





This is applied asymmetrically so that the allowable bias is -5.5% to 6.5%.
The following table lists the NCEP performance recommendations, as well as the
performance allowances for manufacturers and clinical laboratories, as described above.
Performance Criteria for Manufacturers and Clinical Laboratories
NCEP Inaccuracy NCEP Imprecision
CRMLN Inaccuracy
Allowance
HDL Cholesterol bias 5 % CV 4 % -5.5% to 6.5%
The CRMLN has also implemented the use of a t-test to evaluate whether or not the test
systems bias is significantly different from the NCEP goal. The variance component of the
t-test utilized in the CRMLN programs is the NCEPs maximum allowable imprecision. The
rationale for using this value for the variance is that, since the CRMLNs primary goal is to
evaluate accuracy, we do not want to penalize a system that has very good precision. The
effect of using this t-test is that the test system is given some benefit of the doubt. The t-test
is performed at various levels for alpha ). A significant bias at =10% should be
interpreted as a warning that the bias is very close to the NCEP criterion. A test system that
has significant bias at =5% is deemed to not meet the NCEP bias criterion and will not pass
certification.
The test for within-method outliers is based on the procedure described in NCCLS EP9-A
(7). Tests of absolute and relative differences are performed. For the test of absolute
differences, the difference between the test method duplicates is calculated for each sample.
A test limit is determined which is four times the average difference. Any sample with a
difference greater than the test limit is flagged. For the test of relative differences, the
difference between duplicates is divided by the test method mean. A test limit is determined
which is four times the average relative difference. Any sample with a relative difference
greater than the test limit is flagged. Only samples that are flagged by both the absolute and
relative tests are within method outliers.
The test for between-method outliers is also based on the procedure described in NCCLS
EP9-A (7). Tests of absolute and relative differences between the two methods are
performed. For the test of absolute differences, the difference between the test method mean
and the reference method mean is calculated. A test limit is determined which is four times
the average difference. Any sample with a difference greater than the test limit is flagged.
For the test of relative differences, the difference between the test method mean and the
reference method mean is divided by the reference method mean. A test limit is determined
which is four times the average relative difference. Any sample with a relative difference
greater than the test limit is flagged. Only samples that do not pass both tests are between-
method outliers.
CRMLN HDL Cholesterol Protocol November 2002 Page 11





References
1. National Cholesterol Education Program. Summary of the Second Report of the National
Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and
Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel II). JAMA
1993;269:301523.
2. National Cholesterol Education Program. Report of the National Cholesterol Education
Program Expert Panel on detection, evaluation, and treatment of high blood cholesterol in
adults. Arch Intern Med 1988;148:369.
3. National Cholesterol Education Program. Executive summary of the Third Report of the
Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in
Adults (ATP 111). JAMA 2001; 285; 2486-96.
4. Warnick GR, Wood PD. National Cholesterol Education Program Recommendations for
measurement of high-density lipoprotein cholesterol: Executive summary. Clin Chem
1995;41:142733.
5. Kimberly MM, Leary ET, Cole TG, Waymack PW. Selection, validation,
standardization, and performance of a designated comparison method for HDL-
cholesterol for use in the Cholesterol Reference Method Laboratory Network. Clin Chem
1999;45:1803-12.
6. NCCLS. The Reference System for the Clinical Laboratory: Criteria for development and
credentialing of methods and materials for harmonization of results; approved guideline.
NCCLS document NRSCL13-A. Wayne, PA: NCCLS, (2000). Copies may be purchased
by calling NCCLS at (610) 688-0100 or from the NCCLS web site at http://www.nccls.org.
7. NCCLS. Method comparison and bias estimation using patient samples; approved
guideline. NCCLS document EP9-A. Wayne, PA: NCCLS, 1995. Copies may be
purchased by calling NCCLS at (610) 688-0100 or from the NCCLS web site at
http://www.nccls.org
8. Perspectives in Disease Prevention and Health Promotion Update: Universal Precautions
for Prevention of Transmission of Human Immunodeficiency Virus, Hepatitis B Virus,
and Other Bloodborne Pathogens in Health-Care Settings. MMWR June 24, 1988 /
37(24);377-388. http://www.cdc.gov/mmwr/preview/mmwrhtml/00000039.htm
9. NCCLS. Evaluation of precision performance of clinical chemistry devices; approved
guideline. NCCLS document EP5-A. Wayne, PA: NCCLS, 1999. Copies may be
purchased by calling NCCLS at (610) 688-0100 or from the NCCLS web site at
http://www.nccls.org
10. NCCLS. Procedures for the Collection of Diagnostic Blood Specimens by Venipuncture;
Approved Standard Fourth Edition. NCCLS document H3-A4. Wayne, PA: NCCLS,
1998. Copies may be purchased by calling NCCLS at (610) 688-0100 or from the
NCCLS web site at http://www.nccls.org
11. Bennett ST, Eckfeldt JH, Belcher JD, Connelly DP. Certification of cholesterol
measurements by the National Reference Method Laboratory Network with routine
clinical specimens: Effects of network laboratory bias and imprecision. Clin Chem
1992;38:651-7.
CRMLN HDL Cholesterol Protocol November 2002 Page 12





CHOLESTEROL REFERENCE METHOD LABORATORY NETWORK
Fee schedule
HDL Cholesterol Certification Protocol for Manufacturers
(using designated comparison method for HDL cholesterol [HDLC DCM])
MANUFACTURERS
A. Fresh Sample Comparison $4800.00
1. Duplicate HDLC DCM analysis of 40 patient specimens
2. Shipping charges paid by manufacturer
3. Data analysis for one instrument/method, material, or reagent application
4. One repeat data analysis, if necessary
B. Additional specimens $120.00/specimen
(duplicate HDLC DCM analyses)
C. Data analysis for additional applications $100.00/method
CRMLN HDL Cholesterol Protocol November 2002 Page 13