Tải bản đầy đủ - 0trang
7 Interferences from Bilirubin, Hemolysis, and High Lipid Content
2.8 Interferences from Endogenous and Exogenous Components
Hemoglobin is mainly released by hemolysis of red blood cells (RBC).
Hemolysis can occur in vivo, during venipuncture and blood collection, or
during sample processing. Hemoglobin interference depends on its concentration in the sample. Serum appears hemolyzed when the hemoglobin concentration exceeds 20 mg/dL. The absorbance maxima of the heme moiety in
hemoglobin are at 540 to 580 nm wavelengths. However, hemoglobin
begins to absorb around 340 nm and then absorbance increases at
400À430 nm as well. Interference of hemoglobin (if the specimen is grossly
hemolyzed) is due to interference with the optical detection system of
All lipids in plasma exist as complexed with proteins that are called lipoproteins, and particle size varies from 10 nm to 1000 nm (the higher the percentage of the lipid, the lower the density of the resulting lipoprotein and
the larger the particle size). The lipoprotein particles with high lipid content
are micellar and are the main source of assay interference. Unlike bilirubin
and hemoglobin, lipids normally do not participate in chemical reactions
and mostly cause interference in assays due to their turbidity and capability
of scattering light, as in nephelometric assays.
2.8 INTERFERENCES FROM ENDOGENOUS
AND EXOGENOUS COMPONENTS
Immunoassays are affected by a variety of endogenous and exogenous
compounds, including heterophilic antibodies. The key points regarding
immunoassay interferences include:
Endogenous factors such as digoxin-like immunoreactive factors only
affect digoxin immunoassays. Please see Chapter 15 for a more detailed
Structurally similar molecules are capable of cross-reacting with the
antibody to cause falsely elevated (positive interference) or falsely
lowered results (negative interference). Negative interference occurs less
frequently than positive interference, but may be clinically more
dangerous. For example, if the result of a therapeutic drug is falsely
elevated compared to the previous measurement, the clinician may
question the result, but if the value is falsely lower, the clinician may
simply increase the dose without realizing that the value was falsely
lower due to interference. That can cause drug toxicity in the patient.
Interference from drug metabolites is the most common form of
interference, although other structurally similar drugs may also be the
cause of interference. See also Chapter 15.
I m m un o a ss a y P l a t f o r m an d D e s i g ns
2.9 INTERFERENCES OF HETEROPHILIC
ANTIBODIES IN IMMUNOASSAYS
Heterophilic antibodies are human antibodies that interact with assay antibody
interferences. Features of heterophilic antibody interference in immunoassays
Heterophilic antibodies may arise in a patient in response to exposure to
certain animals or animal products or due to infection by bacterial or
viral agents, or non-specifically.
Among heterophilic antibodies, the most common are human antimouse antibodies (HAMA) because of wide use of murine monoclonal
antibody products in therapy or imaging. However, other anti-animal
antibodies in humans have also been described that can interfere with an
If a patient is exposed to animals or animal products, or suffers from an
autoimmune disease, the patient may have heterophilic antibodies in
Heterophilic antibodies interfere most commonly with sandwich assays
that are used for measuring large molecules, but rarely interfere with
competitive assays. Most common interferences of heterophilic
antibodies are observed with the measurement of various tumor markers.
In the sandwich-type immunoassays, heterophilic antibodies can form
the “sandwich complex” even in the absence of the target antigen; this
generates mostly false positive results. False negative results due to the
interference of heterophilic antibodies are rarely observed.
Heterophilic antibodies are absent in urine. Therefore, if a serum
specimen is positive for an analyte, for example, human chorionic
gonadotropin (hCG), but beta-hCG cannot be detected in the urine
specimen, it indicates interference from heterophilic antibodies in the
serum hCG measurement.
Another way to investigate heterophilic antibody interference is serial
dilution of a specimen. If serial dilution produces a non-linear result, it
indicates interference in the assay.
Interference from heterophilic antibodies may also be blocked by adding
any commercially available heterophilic antibody blocking agent in the
specimen prior to analysis.
For analytes that are also present in the protein-free ultrafiltrate
(relatively small molecules), analysis of the analyte in the protein-free
ultrafiltrate can eliminate interference from heterophilic antibodies
because, due to large molecular weights, heterophilic antibodies are
absent in protein-free ultrafiltrates.
Heterophilic antibodies are more commonly found in sick and hospitalized
patients with reported prevalences of 0.2%À15%. In addition, rheumatoid
2.10 Interferences from Autoantibodies and Macro-Analytes
factors that are IgM type antibodies may be present in the serum of patients
suffering from rheumatoid arthritis and certain autoimmune diseases.
Rheumatoid factors may interfere with sandwich assays and the mechanism
of interference is similar to the interference caused by heterophilic antibodies. Commercially available rheumatoid factor blocking agent may be used
to eliminate such interferences.
A 58-year-old man without any familial risk for prostate cancer
visited his primary care physician and his prostate-specific
antigen (PSA) level was 83 ng/mL (0À4 ng/mL is normal). He
was referred to a urologist and his digital rectal examination
was normal. In addition, a prostate biopsy, abdominal tomodensitometry, whole body scan, and prostatic MRI were performed, but no significant abnormality was observed.
However, due to his very high PSA level (indicative of advance
stage prostate cancer) he was treated with androgen deprivation therapy with goserelin acetate and bicalutamide. After 3
months he still had no symptoms, his prostate was atrophic on
digital rectal examination, and he had suppressed testosterone
levels as expected. However, his PSA level was still highly
elevated (122 ng/mL) despite no radiographic evidence of
advanced cancer. At that point his serum PSA was analyzed
by a different assay (Immulite PSA, Cirrus Diagnostics, Los
Angeles) and the PSA level was , 0.3 ng/mL. The treating
physician therefore suspected a false positive PSA by the original Access Hybritech PSA assay (Hybritech, San Diego, CA),
and interference of heterophilic antibodies was established by
treating specimens with heterophilic antibody blocking agent.
Re-analysis of the high PSA specimen showed a level below
the detection limit. This patient received unnecessary therapy
for his falsely elevated PSA level due to the interference of heterophilic antibody .
A 64-year-old male during a routine visit to his physician was
diagnosed with hypothyroidism based on elevated TSH (thyroid stimulating hormone) levels, and his clinician initiated
therapy with levothyroxine (250 microgram per day). Despite
therapy, there were still increased levels of TSH (33 mIU/L)
and his FT4 level was also elevated. The endocrinologist at
that point suspected that TSH levels measured by the Unicel
Dxi analyzer (Beckman Coulter) were falsely elevated due to
interference. Serial dilution of the specimen showed non-
linearity, an indication of interference. When the specimen
was analyzed using a different TSH assay (immunoradiometric assay (IRMA), also available from Beckman Coulter),
the TSH value was 1.22 mIU/L, further confirming the interference with the initial TSH measurement. The patient had a
high concentration of rheumatoid factor (2700 U/mL) and the
authors speculated that his falsely elevated TSH was due to
interference from rheumatoid factors .
2.10 INTERFERENCES FROM AUTOANTIBODIES
Autoantibodies (immunoglobulin molecules) are formed by the immune
system of an individual capable of recognizing an antigen on that person’s
I m m un o a ss a y P l a t f o r m an d D e s i g ns
own tissues. Several mechanisms may trigger the production of autoantibodies, for example, an antigen formed during fetal development and then
sequestered may be released as a result of infection, chemical exposure, or
trauma, as occurs in autoimmune thyroiditis. The autoantibody may bind to
the analyte-label conjugate in a competition-type immunoassay to produce a
false positive or false negative result. Circulating cardiac troponin I autoantibodies may be present in patients suffering from acute cardiac myocardial
infarction where troponin I elevation is an indication of such an episode.
Unfortunately, the presence of circulating cardiac troponin I autoantibodies
may falsely lower cardiac troponin I concentration (negative interference)
using commercial immunoassays, thus complicating the diagnosis of acute
myocardial infarction . However, falsely elevated results due to the presence of autoantibodies are more common than false negative results.
Verhoye et al. found three patients with false positive thyrotropin results
that were caused by interference from an autoantibody against thyrotropin.
The interfering substance in the affected specimens was identified as an
autoantibody by gel-filtration chromatography and polyethylene glycol
Often the analyte can conjugate with immunoglobin or other antibodies to
generate macro-analytes, which can falsely elevate the true value of the analyte. For example, macroamylasemia and macro-prolactinemia can produce
falsely elevated results in amylase and prolactin assays, respectively. In
macro-prolactinemia, the hormone prolactin conjugates with itself and/or
with its autoantibody to create macro-prolactin in the patient’s circulation.
The macro-analyte is physiologically inactive, but often interferes with many
prolactin immunoassays to generate false positive prolactin results . Such
interference can be removed by polyethylene glycol precipitation.
A 17-year-old girl was referred to a University hospital for
having a persistent elevated level of aspartate aminotransferase (AST). One year earlier, her AST level was 88 U/L as
detected during her annual school health check, but she had
no medical complaints. She was not on any medication and
had a regular menstrual cycle. Her physical examination at
the University hospital was unremarkable. All laboratory test
results were normal, but her AST level was further elevated
to 152 U/L. All serological tests for hepatitis were negative.
On further follow-up her AST level was found to have
increased to 259 U/L. At that point it was speculated that her
elevated AST was due to interference, and further study by
gel-filtration showed a species with a molecular weight of
250 kilodaltons. This was further characterized by immunoelectrophoresis and immunoprecipitation to be an immunoglobulin (IgG kappa-lambda globulin) complexed AST that
was causing the elevated AST level in this girl. These complexes are benign .
2.11 PROZONE (OR “HOOK”) EFFECT
The Prozone or hook effect is observed when a very high amount of an analyte is present in the sample but the observed value is falsely lowered. This
type of interference is observed more commonly in sandwich assays. The
mechanism of this significant negative interference is the capability of a high
level of an analyte (antigen) to reduce the concentrations of “sandwich”
(antibody 1:antigen:antibody 2) complexes that are responsible for generating the signal by forming mostly single antibody:antigen complexes. The
hook effect has been reported with assays of a variety of analytes, such as
β-hCG, prolactin, calcitonin, aldosterone, cancer markers (CA 125, PSA), etc.
The best way to eliminate the hook effect is serial dilution. For example, if
the hook effect is present and the original value of an analyte (e.g. prolactin)
was 120 ng/mL, then 1:1 dilution of the specimen should produce a value of
60 ng/mL; but if the observed value was 90 ng/mL (which was significantly
higher than the expected value), the hook effect should be suspected. In
order to eliminate the hook effect, a 1:10, 1:100, or even a 1:1000 dilution
may be necessary so that the true analyte concentration will fall within the
analytical measurement range (AMR) of the assay..
A 16-year-old girl presented to the emergency department
with a 2-week history of nausea, vomiting, vaginal spotting,
and lower leg edema. On physical examination, a lower abdomen palpable mass was found. The patient admitted sexual
activity, but denied having any sexually transmitted disease.
Molar pregnancy was suspected, and the quantitative β-subunit of human chorionic gonadotropin (β-hCG) concentration
was 746.2 IU/L; however, the urine qualitative level was negative. Repeat of the urinalysis by a senior technologist also
produced a negative result. At that point the authors
suspected the hook effect and dilution of the serum specimen
(1:1) produced a non-linear value (455.2 IU/L), which further
confirmed the hook effect. After a 1:10 dilution, the urine test
for β-hCG became positive, and finally, by using a 1:10,000
dilution of the specimen, the original serum β-hCG concentration was determined to be 3,835,000 IU/L. Usually the hook
effect is observed with a molar β-hCG level in serum because
high amounts of β-hCG are produced by molar pregnancy
Immunoassays can be competitive or immunometric (non-competitive, also known
as sandwich). In competitive immunoassays only one antibody is used. This
format is common for assays of small molecules such as a therapeutic drugs or
I m m un o a ss a y P l a t f o r m an d D e s i g ns
drugs of abuse. In the sandwich format two antibodies are used and this format is
more commonly used for assays of relative large molecules.
Homogenous immunoassay format: After incubation, no separation between
bound and free label is necessary.
Heterogenous immunoassay format: The bound label must be separated from the
free label before measuring the signal.
Commercially available immunoassays use various formats, including FPIA, EMIT,
CEDIA, KIMS, and LOCI. In the fluorescent polarization immunoassay (FPIA), the
free label (a relatively small molecule) attached to the analyte (antigen) molecule
has different Brownian motion than when the label is complexed to a large
antibody (140,000 or more Daltons). FPIA is a homogenous competitive assay
where after incubation the fluorescence polarization signal is measured; this signal
is only produced if the labeled antigen is bound to the antibody molecule.
Therefore, intensity of the signal is inversely proportional to the analyte
EMIT (enzyme multiplied immunoassay technique) is a homogenous competitive
immunoassay where the antigen is labeled with glucose 6-phosphate
dehydrogenase, an enzyme that reduces nicotinamide adenine dinucleotide (NAD,
no signal at 340 nm) to NADH (absorbs at 340 nm), and the absorbance is
monitored at 340 nm. When a labeled antigen binds with the antibody molecule,
the enzyme label becomes inactive and no signal is generated. Therefore, signal
intensity is proportional to analyte concentration.
The Cloned Enzyme Donor Immunoassay (CEDIA) method is based on
recombinant DNA technology where bacterial enzyme beta-galactosidase is
genetically engineered into two inactive fragments. When both fragments
combine, a signal is produced that is proportional to the analyte concentration.
Kinetic interaction of microparticle in solution (KIMS): In the absence of antigen
molecules free antibodies bind to drug microparticle conjugates to form particle
aggregates that result in an increase in absorption that is optically measured at
various visible wavelengths (500À650 nm).
Luminescent oxygen channeling immunoassays (LOCI): The immunoassay
reaction is irradiated with light to generate singlet oxygen molecules in
microbeads (“Sensibead”) coupled to the analyte. When bound to the respective
antibody molecule, also coupled to another type of bead, it reacts with singlet
oxygen and chemiluminescence signals are generated that are proportional to the
concentration of the analyteÀantibody complex.
Usually total bilirubin concentration below 20 mg/dL does not cause interferences,
but concentrations over 20 mg/dL may cause problems. The interference of
bilirubin is mainly caused by its absorbance at 454 or 461 nm.
Various structurally related drugs or drug metabolites can interfere with
Heterophilic antibodies may arise in a patient in response to exposure to certain
animals or animal products, due to infection by bacterial or viral agents, or use of
murine monoclonal antibody products in therapy or imaging. Heterophilic
antibodies interfere most commonly with sandwich assays used for measuring
large molecules, but rarely with competitive assays, causing mostly false positive
Heterophilic antibodies are absent in urine. Therefore, if a serum specimen is
positive for an analyte (e.g. human chorionic gonadotropin, hCG), but beta-hCG
cannot be detected in the urine specimen, it indicates interference from a
heterophilic antibody in the serum hCG measurement. Another way to investigate
heterophilic antibody interference is serial dilution of a specimen. If serial dilution
produces a non-linear result, it indicates interference in the assay. Interference
from heterophilic antibodies can also be blocked by adding commercially available
heterophilic antibody blocking agents to the specimen prior to analysis.
Autoantibodies are formed by the immune system of a person that recognizes an
antigen on that person’s own tissues, and may interfere with an immunoassay to
produce false positive results (and less frequently, false negative results). Often the
endogenous analyte of interest will conjugate with immunoglobin or other
antibodies to generate macro-analytes, which can falsely elevate a result. For
example, macroamylasemia and macro-prolactinemia can produce falsely elevated
results in amylase and prolactin assays, respectively. Such interference can be
removed by polyethylene glycol precipitation.
Prozone (“hook”) effect: Very high levels of antigen can reduce the concentrations
of “sandwich” (antibody 1:antigen:antibody 2) complexes responsible for
generating the signal by forming mostly single antibody:antigen complexes. This
effect, known as the prozone or hook effect (excess antigen), mostly causes
negative interference (falsely lower results). The best way to eliminate the hook
effect is serial dilution.
 Jolley ME, Stroupe SD, Schwenzer KS, Wang CJ, et al. Fluorescence polarization immunoassay
III. An automated system for therapeutic drug determination. Clin Chem 1981;27:1575À9.
 Jeon SI, Yang X, Andrade JD. Modeling of homogeneous cloned enzyme donor immunoassay. Anal Biochem 2004;333:136À47.
 Snyder JT, Benson CM, Briggs C, et al. Development of NT-proBNP, Troponin, TSH, and FT4
LOCI(R) assays on the new Dimension (R) EXL with LM clinical chemistry system. Clin
Chem 2008;54:A92 [Abstract #B135].
 Dai JL, Sokoll LJ, Chan DW. Automated chemiluminescent immunoassay analyzers. J Clin
Ligand Assay 1998;21:377À85.
 Forest J-C, Masse J, Lane A. Evaluation of the analytical performance of the Boehringer
Mannheim Elecsyss 2010 Immunoanalyzer. Clin Biochem 1998;31:81À8.
 Babson AL, Olsen DR, Palmieri T, Ross AF, et al. The IMMULITE assay tube: a new approach
to heterogeneous ligand assay. Clin Chem 1991;37:1521À2.
I m m un o a ss a y P l a t f o r m an d D e s i g ns
 Christenson RH, Apple FS, Morgan DL. Cardiac troponin I measurement with the
ACCESSs immunoassay system: analytical and clinical performance characteristics. Clin
 Montagne P, Varcin P, Cuilliere ML, Duheille J. Microparticle-enhanced nephelometric
immunoassay with microsphere-antigen conjugate. Bioconjugate Chem 1992;3:187À93.
 Henry N, Sebe P, Cussenot O. Inappropriate treatment of prostate cancer caused by heterophilic antibody interference. Nat Clin Pract Urol 2009;6:164À7.
 Georges A, Charrie A, Raynaud S, Lombard C, et al. Thyroxin overdose due to rheumatoid
factor interferences in thyroid-stimulating hormone assays. Clin Chem Lab Med
 Tang G, Wu Y, Zhao W, Shen Q. Multiple immunoassays systems are negatively interfered
by circulating cardiac troponin I autoantibodies. Clin Exp Med 2012;12:47À53.
 Verhoye E, Bruel A, Delanghe JR, Debruyne E, et al. Spuriously high thyrotropin values due
to anti-thyrotropin antibody in adult patients. Clin Chem Lab Med 2009;47:604À6.
 Kavanagh L, McKenna TJ, Fahie-Wilson MN, et al. Specificity and clinical utility of methods
for determination of macro-prolactin. Clin Chem 2006;52:1366À72.
 Matama S, Ito H, Tanabe S, Shibuya A, et al. Immunoglobulin complexed aspartate aminotransferase. Intern Med 1993;32:156À9.
 Er TK, Jong YJ, Tsai EM, Huang CL, et al. False positive pregnancy in hydatidiform mole.
Clin Chem 2006;52:1616À8.
3.1 LABORATORY ERRORS IN PRE-ANALYTICAL,
ANALYTICAL, AND POST-ANALYTICAL STAGES
Accurate clinical laboratory test results are important for proper diagnosis
and treatment of patients. Factors that are important to obtaining accurate
laboratory test results include:
Patient Identification: The right patient is identified prior to specimen
collection by matching at least two criteria.
Collection Protocol: The correct technique and blood collection tube
have been used for sample collection to avoid tissue damage, prolonged
venous stasis, or hemolysis.
Labeling: After collection, the specimen was labeled properly with correct
patient information; specimen misidentification is a major source of preanalytical error.
Specimen Handling: Proper centrifugation (in the case of serum or
plasma specimen analysis) and proper transportation of specimens to the
Storage Protocol: Maintaining proper storage of specimens prior to
analysis in order to avoid artifactual changes in analyte; for example,
storing blood gas specimens in ice if the analysis cannot be completed
within 30 min of specimen collection.
Interference Avoidance: Proper analytical steps to obtain the correct result
and avoid interferences.
LIS Reports: Correctly reporting the result to the laboratory information
system (LIS) if the analyzer is not interfaced with the LIS.
Clinician Reports: The report reaching the clinician must contain the right
result, together with interpretative information, such as a reference range
and other comments that aid clinicians in the decision-making process.
Errors in PreAnalytical,
Stages .................... 35
3.2 Order of Draw of
3.3 Errors with
Preparation ........... 38
3.4 Errors with
Errors ..................... 38
3.5 Error of
Collecting Blood in
Wrong Tubes: Effect
of Anticoagulants. 40
3.6 Issues with
Collection .............. 42
3.7 Issues with
3.8 Special Issues:
Blood Gas and
A. Dasgupta and A. Wahed: Clinical Chemistry, Immunology and Laboratory Quality Control
© 2014 Elsevier Inc. All rights reserved.
P r e -A na ly t ica l V ar i ab le s
Table 3.1 Common Laboratory Errors
Key Points ............. 44
Type of Error
References ............ 45
Tube filling error
Patient identification error
Order not entered in laboratory information system
Specimen collected wrongly from an infusion line
Specimen stored improperly
Contamination of culture tube
Inaccurate result due to interference
Random error caused by the instrument
Result communication error
Excessive turnaround time due to instrument downtime
Failure at any of these steps can result in an erroneous or misleading laboratory result, sometimes with adverse outcomes. The analytical part of the analysis involves measurement of the concentration of the analyte corresponding
to its “true” level (as compared to a “gold standard” measurement) within a
clinically acceptable margin of error (the total acceptable analytical error,
TAAE). Errors can occur at any stage of analysis (pre-analytical, analytical, and
post-analytical). It has been estimated that pre-analytical errors account for
more than two-thirds of all laboratory errors, while errors in the analytical
and post-analytical phases account for only one-third of all laboratory errors.
Carraro and Plebani reported that, among 51,746 clinical laboratory analyses
performed in a three-month period in the author’s laboratory (7,615 laboratory orders, 17,514 blood collection tubes), clinicians contacted the laboratory regarding 393 questionable results out of which 160 results were
confirmed to be due to laboratory errors. Of the 160 confirmed laboratory
errors, 61.9% were determined to be pre-analytical errors, 15% were analytical
errors, while 23.1% were post-analytical errors . Types of laboratory errors
(pre-analytical, analytical, and post-analytical) are summarized in Table 3.1.
In order to avoid pre-analytical errors, several approaches can be taken,
The use of hand-held devices connected to the LIS that can objectively
identify the patient by scanning a patient attached barcode, typically a