Clinical Chemistry 43: 384-389, 1997;
(Clinical Chemistry. 1997;43:384-389.)
© 1997 American Association for Clinical Chemistry, Inc.
Whole-blood test for total cholesterol by a self-metering, self-timing disposable device with built-in quality control
Wai Tak Lawa,
Sonal Doshi,
John McGeehan,
Susan McGeehan,
David Gibboni,
Yuri Nikolioukine,
Robert Keane,
Jennifer Zheng,
Jagan Rao and
Gerhard Ertingshausen
ActiMed Laboratories, Inc., 5 Terri Lane, Burlington, NJ 08016.
a Author for correspondence. Fax 609-387-2700; e-mail Law{at}actimed.com
 |
Abstract
|
|---|
A whole-blood test for total cholesterol has been developed that is
performed in a low-cost disposable flow device without user
intervention (after sample addition). The device meters the sample,
separates plasma from erythrocytes, and precisely times plasma flow
into various reagent compartments, including a quality-assurance
chamber. Test results are displayed as a well-defined and easily
readable color bar. A quality-control window attests to the integrity
of the test components. Here, we describe the assembly and individual
functions of the device and report its performance characteristics.
Precision and accuracy studies in four clinical studies at independent
locations yielded total imprecision of <5% and an average bias of
1.35% vs the Abell–Kendall method.
Key Words: indexing terms: enzymatic methods • near-patient testing
|
Introduction
|
|---|
Among the first-generation solutions to the challenge of providing
noninstrumented test devices for quantitative testing on whole-blood
samples, the AccuMeter cholesterol test developed by ChemTrak
(Sunnyvale, CA) provides quantitative results generated by two to three
drops of blood deposited on a disposable cassette (1).
This cassette contains a complete reagent system that translates the
enzyme-mediated oxidation of cholesterol and cholesterol esters into a
visible and measurable color bar, produced by the horseradish
peroxidase-catalyzed reaction of hydrogen peroxide with
3-methyl-2-benzothiazolinone hydrazone followed by diazotization of an
immobilized aniline dye. The length of the colored bar is related to
concentration units in a table accompanying the system, which is used
to look up the results of the assay. A different technical approach,
based on analog-to-digital switching technology, has the advantages of
direct read-out of results and use of dry reagents in all compartments
of the device (2). However, both approaches still require
timing of the reaction steps and supplemental user intervention after
the whole-blood sample is deposited on the device, and neither has
provisions to prevent undersampling or includes built-in controls that
attest to the integrity of the reagents.
The Analytical Chromogenic Transport (A.C.TTM; ActiMed
Labs., Burlington, NJ) technique described here lends itself to the
development of noninstrumented quantitative test devices that require
no further user intervention after addition of the sample
(3)(4)(5). The design of such devices assures that the
analytical process does not start unless sufficient sample has been
added, thus preventing undersampling. Results are read directly from
the device after the color of a quality-assurance (QA) window has
turned green. The color change assures the user that the test is
complete and that the most fragile components of the analytical cascade
are intact. The resulting device lends itself to uses at alternative
testing sites outside the hospital laboratory. Here, we describe the
device as applied to testing for total cholesterol in fingerstick whole
blood, i.e., in the ENA.C.TTM Total Cholesterol Test
(ActiMed Labs.).
|
Principle and Device Design
|
|---|
The structural elements of the A.C.T cholesterol test device are
shown in the exploded-view drawing (Fig. 1
). The components are assembled in a completely automated,
computer-controlled manufacturing process, which allows overall
continued accuracy assurance through a feedback control.
To assure plasma transport from the sample well that receives
whole-blood samples through to the QA draw-zone at the end of the
device, the device uses a gradient of increasing surface energy
established along the path of the liquid flow. To further enhance flow
through the device, the design minimizes the resistance to flow,
particularly in the front section, for which a stack of special filters
was developed. These filters accommodate the use of blood within a wide
range of hematocrit values and lipid concentrations. Other diverse
functions the filters perform include separation of erythrocytes from
plasma, addition of enzymes and detergents to the resulting plasma with
complete mixing, retention of the plasma solution sufficiently long to
ensure complete conversion of cholesterol and cholesterol esters into
cholestenone and hydrogen peroxide, and subsequent release of the
reacted plasma into the measurement area of the detection zone. The
filters are fastened in place by a hotmelt-coated aluminum foil seal
activated remotely by a radiofrequency sealer. This sealing process
assures that plasma flows through the filters rather than around them.
Plasma leaving the filter stack is directed into the detection zone. In
this zone a hydrogen peroxide-sensitive dual-component Trinder-type dye
system is coated onto a fabric that is sandwiched between two plastic
foils. This creates an enclosed flow channel through which the plasma
travels, its hydrogen peroxide content generating a quantitative,
visible color signal that is directly converted into concentration
units via a scale printed on the cover of the device. The end of the
detection zone connects with the QA draw-zone, where another
cholesterol-converting reaction occurs (this time with exogenous
cholesterol) to verify the stability of key components of the reagent
system that are now dissolved in the plasma sample.
|
Materials and Methods
|
|---|
reagents
Cholesterol oxidase (EC 1.1.3.6), cholesterol esterase (EC
3.1.1.13), and horseradish peroxidase (EC 1.11.1.7) were purchased from
Toyobo, Osaka, Japan; pancreatic cholesterol esterase was purchased
from Diagnostic Chemicals, Oxford, CT. The sodium salt of
3-methyl-6-carboxy-2-benzothiazolinone hydrazone (NaCMBTH) is prepared
at ActiMed. Sodium cholate, sodium chloride, lactose, and hydroxylamine
hydrochloride were from Aldrich Chemical Co., Milwaukee, WI. Lectins,
lithium heparin, Pipes buffer, and bovine serum albumin were from Sigma
Chemical Co., St. Louis, MO.
components of the a.c.t device
Device base.
The base of the device is made of polyester
thermoformed into a shape that accommodates the placement of all
structural elements for the analytical processes. The surface of the
sample well, an integral part of the base, is rendered hydrophilic by
treatment with corona discharge followed by spray-coating with
polyacrylic acid containing a heparin derivative.
Multipad front-end stack.
The absorbent pad, loosely
woven Orlon® (obtained from Lydall, Hamptonville, NC),
sits on top of a multielement stack of porous pads. Its front end
touches the upper end of a ridge that connects to the sample well. Once
the blood sample reaches this pad via the ridge, sufficient blood is
aspirated to supply the device with all the plasma needed to perform
the test. The pad is cut by laser from a large sheet of the material,
impregnated with a proprietary agglutination solution based on lectins,
and subsequently air-dried. Agglutinated erythrocytes are trapped in
this pad.
The secondary separation pad for the agglutinated and free erythrocytes
is positioned under the absorbent pad. It allows the plasma to drain
from the agglutinate into the next lower pad, the enzyme pad, which is
impregnated with cholesterol esterase (1000 kU/L) and cholesterol
oxidase (750 kU/L). The enzyme pad is constructed from glass fiber
material, which has a large surface area, and the enzyme reagent
contains porous bulking agents that trap large amounts of air pockets
inside the pads. This combination allows for rapid oxygen diffusion
into the plasma to support the cholesterol oxidase-catalyzed oxidation
reaction.
Supplemental reagents and additives such as sodium cholate,
hydroxylamine hydrochloride, sorbitol, and buffer are added to the
enzyme-enriched plasma when it enters the surfactant pad. Hydroxylamine
in this pad acts to inhibit catalase, basically functioning as a
competitive inhibitor.
A flow-control pad, placed at the bottom of the stack, reduces the
speed of elution of the enzyme-containing plasma from the enzyme pad
into the detection channel without imparting additional resistance to
flow once the pad is fully impregnated with sample. This is designed to
provide a 1- to 2-min incubation period for the cholesterol reactions
to reach equilibrium without user intervention (6).
The stack of pads is sealed in place by a combination of two seals
formed with hotmelt (National Starch Co., Bridgewater, NJ)-coated
aluminum foil (activated by heat generated remotely by a radiofrequency
sealer). The seals prevent the flow of plasma around the pads and
direct flow across the whole surface of the pads.
Measuring channel and detection zone.
The measuring
channel is formed by heat-sealing the upper and lower polyethylene
insert support films into a sandwich around a coated, very precisely
woven fabric (Polyester PES 105/52; Saati Corp., Stamford, CT). The
coating contains a derivatized aniline dye covalently linked to silica
particles <3 µm in diameter (Degussa, Ridgefield Park, NJ),
NaCMBTH, horseradish peroxidase, and various stabilizers embedded in an
acrylic/polyethylene copolymer. The polymer coating is applied in a
tightly controlled process by an automated coating machine. The
precision of the coating process is illustrated in Fig. 2
, which shows the woven fabric and the layer of reactive
coatings. The measuring channel also contains a QA draw-zone, which
contains free cholesterol, horseradish peroxidase, and dye.
In the manufacturing process, the materials of the detection zone are
continuously fed into a six-element heat-sealing station. The zone is
subsequently cut out by laser, automatically transferred to the
thermoformed tray, and glued in place.
Top cover.
The top cover of the device is placed on the
tray at the end of the process. The markings of the concentration scale
are printed on the cover just before the cover is placed on the fully
assembled device. The distance and spread of these markings may be
continually modified to adjust for small variations in the
manufacturing process and assure continued accurate readings of the
results. A window in the top cover allows a view of the top of the
absorbent pad during the analysis.
assay procedure
Fingerstick blood is drawn from an appropriately prepared
individual with a single-use lancet (7); the first drop of
blood is discarded. Blood is added to the sample well of the device
until the "start window" on the top cover turns red (i.e., after at
least 80 µL of whole blood has been added). This event signals that
blood has been transferred to the absorbent pad and that a sufficiently
large sample has been obtained. No further user intervention is
required until the QA draw-zone window turns green, after which the
length of blue color zone may be read in concentration units.
Erythrocytes are separated from the plasma within 30 s and the
plasma travels through the stack of pads until retained by the
flow-control pad. Hydrogen peroxide-containing plasma enters the
detection channel, where it reacts with the dye system embedded in the
polymer coating wrapped around the precisely woven mesh. The length of
the color bar that develops is proportional to the concentration of
hydrogen peroxide, and hence total cholesterol, in the plasma.
Depending on the hematocrit and the lipid content of the subject's
blood, the plasma front reaches the QA draw-zone in 12 to 20 min.
Because the draw-zone changes color only when all enzyme systems are
still active, the results can be read off the device in concentration
units. The blue color bar remains stable for months and thus
constitutes a "permanent" record.
|
Results
|
|---|
Dynamic range.
The dynamic range of the ENA.C.T Total
Cholesterol Test was established from calibration curves prepared by
assaying both lipoprotein-supplemented samples and fresh whole-blood
samples with multiple replicates of the devices. These
"calibrators" were also assayed by the Abell–Kendall method
(8). The dynamic range so established was 1200 to 3600
mg/L total cholesterol (Fig. 3
).
Accuracy.
The accuracy of the ENA.C.T test devices was
established by direct comparison with the Abell–Kendall method, both
procedures being used to assay fresh patients' samples. Fingerstick
samples from four clinical test sites were assayed by the test devices,
whereas venous serum samples obtained at the same time were shipped to
a reference laboratory (a member of the National Reference Method
Laboratory Network) for Abell–Kendall analysis. Fig. 4
(top) shows the values obtained by the ENA.C.T Total
Cholesterol test vs the reference Abell–Kendall method and the linear
regression equation obtained. The average bias in this comparison over
the analytical range of 1200 to 3600 mg/L was 1.35%.
View larger version (31K):
[in this window]
[in a new window]
|
Figure 4. Correlation between ENA.C.T (fingerstick) total
cholesterol and (top) Abell–Kendall (serum) total
cholesterol or (bottom) AccuMeter (fingerstick) total
cholesterol.
|
|
The ENA.C.T test devices were also compared with the AccuMeter,
another disposable device that accepts fingerstick whole blood. The
results (Fig. 4
, bottom) demonstrated that the two devices give
substantially equivalent performance.
The Laboratory Standardization Panel for Blood Cholesterol Measurement
has recommended that the bias of this measurement should not exceed
3%. The average bias of the ENA.C.T Total Cholesterol Test for all
four clinical sites combined, at the medical decision levels
of 2000 and 2400 mg/L, were 1.6% and 0.9%, respectively, in
comparison with the values determined by the Abell–Kendall method.
Precision.
Within-run precision studies were carried out
at four clinical sites with two controls containing cholesterol
concentrations near the medical decision points. Run-to-run precision
was established by assaying two concentrations of controls over 5 days
in the field. The results ranged from 2.11% to 4.82% and from 2.58%
to 5.04%, respectively (Table 1
).
Heparinized whole-blood precision studies were also performed (at
ActiMed Labs.). The data are shown in Table 2
.
Interferences.
Several biologically and chemically
important compounds often present in whole blood were tested for their
effect on the ENA.C.T Total Cholesterol Test. Concentrations (mg/L) of
potential interferences below which there was no measurable effect on
the test were as follows: ascorbic acid, 80; acetaminophen, 10; uric
acid, 100; bilirubin, 150; hemoglobin, 2000; triglycerides, 8000.
|
Discussion
|
|---|
The A.C.T technology was developed to generate diagnostically
valid results outside the classical clinical laboratory. To minimize
the frequency of errors in the use of the device, its operation had to
be very simple. A design that would circumvent all timing steps and all
manual interventions after sample addition was indicated. In addition,
it was important to provide the untrained user with prompts and control
features to signal that sufficient sample had been added, to signal
when to read the concentration off the device scale, and to assure the
user that the results as read are valid. For this reason we developed
an undersampling safeguard to signal when sufficient sample has been
added to the device and to assure the user that the test is never
initiated with an insufficient amount of sample. The safeguard signal
is triggered only after a sufficient amount of sample has been added to
the device.
The QA draw-zone provides an independent quality check for the
integrity of the reagent system. The addition of small amounts of
exogenous cholesterol with an appropriate amount of cholesterol
oxidase—the most labile component of the reagent system—creates an
additional reaction chamber where a separate cholesterol detection
reaction is performed. This occurs at the end of the analytical process
when the plasma sample (from which all hydrogen peroxide has been
removed through the detection process) enters the chamber. Once the
activity of cholesterol oxidase has been reduced to a certain extent,
the reaction rate drops precipitously (see Fig. 5
). This provides a well-defined narrow range of enzyme activity
below which the QA window reaction does not occur, thus providing an
unequivocal message to the user. The addition of this QA feature allows
the user to know whether the device has been compromised (e.g.) by heat
stress.
Cell separation.
The separation of cells from plasma had
to be performed in a way that would provide minimal resistance to the
continued flow of plasma through the device. This excluded any methods
relying on mechanical barriers and selective adhesion of cells to
surface-active materials (e.g., glass fiber paper) from which plasma is
only partially drained. We instead developed an approach in which
lectins attached to nucleating particles rapidly agglutinate
erythrocytes to form large, well-defined aggregates of cell clusters
and allow cell-free plasma to drain freely into the subsequent
compartments of the device.
Flow control.
Flow of plasma through the device had to
be physically controlled to ensure complete reagent dilution; flow
around, instead of through, the reagent pads had to be prevented. This
was achieved by a novel sealing technique in which two hotmelt-coated
aluminum foil strips are placed around the stack of pads. After
assembly, the hotmelt layer is heated by remote control, which causes
the hotmelt to flow into the rims of the pads and effectively seal them
off.
Precision and accuracy.
Precise and accurate total
cholesterol test results are important for establishing the risk of
coronary heart disease and obtaining the proper treatment. The
National Cholesterol Education Program risk categories for adults,
established in 1988, are: <2000 mg/L total cholesterol (desirable),
2000–2400 mg/L (borderline high), and >2400 mg/L (high enough for
patients to seek medical treatment) (9). The ENA.C.T total
cholesterol test has been designed to be very simple to use, and
to give results as accurate as most of the major clinical analyzers
that perform within the National Cholesterol Education Program
analytical guidelines—i.e., that cholesterol tests should be precise
(<3% CV) and accurate (<5% bias from the Abell-Kendall Reference
Method) (10).
In summary, we have developed a truly one-step disposable system
that is read like a thermometer, is quantitative, and is as accurate
and precise as instrumented methods. The system is easy to use,
requires no technical expertise or training, can be performed anywhere,
and gives results in <20 min. The A.C.T test device was designed as a
platform and is easily adaptable for use with other analytes.
Basically, the enzyme pads can be changed and additional filter pads
added to the front end; all other components of the platform remain the
same. We have demonstrated working prototypes for assays of glucose,
triglycerides, HDL, and LDL (11)(12).
|
References
|
|---|
-
Allen MP, DeLizza A, Ramel U, Jeong H, Singh P. A noninstru- mented quantitative test system and its application for determining cholesterol concentration in whole blood. Clin Chem 1990;36:1591-1597.
[Abstract/Free Full Text]
-
Johnston JB, Palmer JL. A new, rapid non-instrumented total cholesterol test [Abstract]. Clin Chem 1992;38:1066.
-
Ertingshausen G. Method for quantitative analysis of body fluid constituents. US Patent 5,087,556 1992;.
-
McGeehan JK, Ertingshausen G. Method and device for metering of fluid samples. US Patent 5,234,813 1993;.
-
McGeehan JK, Ertingshausen G, Meluch TB. Manufacturing process for sample-initiated assay device. US Patent 5,411,858 1995;.
-
Gibboni DJ, McGeehan SM, Law WT. Method and apparatus for flow control. US Patent 5,447,689 1995;.
-
Ramel UA. Lancet device. US Patent 5,540,709 1996;.
-
Abell LL, Levy BB, Brodie RB, Kendall FE. Simplified method for the estimation of total cholesterol in serum and demonstration of its specificity. J Biol Chem 1952;195:357-366.
[Free Full Text]
-
Report of the National Cholesterol Education Program.
Expert panel on detection, evaluation, and treatment of high blood
cholesterol in adults. Ann Intern Med 1988;148:36–9..
-
Second report of the National Cholesterol Education
Program. Expert panel on detection, evaluation, and treatment of high
blood cholesterol in adults. NIH Publ. 93–3095. Bethesda, MD: National
Institutes of Health, 1993..
-
Law WT, Ertingshausen G. Process and apparatus for direct determination of LDL. US Patent 5,286,626 1995;.
-
Law WT, Ertingshausen G. Process and apparatus for direct determination of LDL. US Patent 5,411,870 1995;.
Top
Abstract
Introduction
