| |
|
|
AXIOM as to Test Reliability as well as Weakness in ALL tests: The layperson public and even a high percentage of practicing physicians have a dangerous tendencity to demand and order disease diagnosis tests when there is a very low probability of the patient actually having the disease. As a function of the mathematics of statistics, a positive-result lab test ordered for "flu" NOT during the flu season or a flu outbreak has a high likelihood of being a false positive test result! A way to see this is: "In South Carolina, if you hear hoof-beats on the street, don't bet you money that a Zebra is the hoofed animal making the noise". Check the "Limitations" and "See Also" sections HERE4 for beginning of the scientific explanation. In ADDITION, all test types have methodology-specific factors that produce false positives and false negatives (including molecular tests such as PCR!). Finally, the beginning paragraph on the PSA page gives an illustration of another problem as to statistics, HERE. |
ADDITIONAL Serology-Specific Confounders: |
Heterophile,
Background "Noise", and the Low/Low Problem False
Positives: |
| Interpretative
warning: There are natural, heterophile, and pathological
antibodies; and there are "interferences". And most
antibodies lag a certain amount of morphological disease activity
prior to detectability (morphological change is apparent before
serological change, especially in autoimmune disorders). |
| Serological tests (direct or indirect
antigen-antibody tests) have such a high prevalence of false positivity
(background reactive "noise") in undiluted serum that
it is standard practice to do the testing on serum diluted to at
least 1 part of serum to 10 parts of diluent (expressed 1:10).
By general medico-scientific agreement, a test is non-reactive
("negative" [for practical purposes]) if it fails to
signal "positive" at the low dilution of 1:10. The serum
sample is tested at increasingly higher double dilutions of 1:20,
1:40, 1:80, 1:160, 1:320, 1:640, 1:1280. Many results testing "positive" at
dilutions of less than 1:80 are not specifically diagnostic. Clinical
judgment is necessary to evaluate the meaning of these "low-titer-positive" results.
For example, as to autoantibodies such as ANA, it is estimated
that 25% of non-rheumatic and healthy persons carries positivity
for ANA at low titer. |
| As the body begins to produce antibodies
(Ab) against an offending
antigen (Ag) (an organism, or pollen, or mite dust, or against
normal components of the body..."auto-antibodies", all
of which are antigens), the IgM class Ab (acute-phase Ab) rises to detectability early, followed by the IgG class Ab
(immune-phase Ab). The patient's very initial Ab production is (1) absorbed
out of the blood onto organisms (or other antigens) until those
targets are saturated. Following
saturation, serum antibody levels (2) rise into the range of detectability
and (3) then into higher diagnostic levels. Therefore,
there is a window of nondetectability (from the time of Ag challenge
to the body until the time of test's initial detection of
Ab...or a rise from a stable baseline titer) of days to weeks (briefest
with re-infection). |
| Because the various possible antigens can
be chemically complex, various tests (such as for two different
organisms) can have cross-reactive "positivities" at
dilutions of 1:20 to 1:80 (sometimes much higher as in case of
HSV I & HSV II). So, "positive" tests must
be carefully interpreted in view of all other clinical information.
One must guard against over-interpretation of test results. And
this is especially true in infectious-disease testing when
there are "low/low" positivities of IgM/IgG
(such as IgM, 1:20 with IgG of 1:40 in the same sample). |
| Non-serological tests: Many non-serological
tests are performed by modern immunological-reaction-based test
systems which makes those tests subject to immunological interferences,
most of which are circulating in the blood of some patients. These could
produce false positives resulting in diagnoses which don't exist
in that patient. False-negatives are perhaps more obscure and less
understood. In these cases, an antibody saturates the capture antibody
and saturates to some extent the conjugate in solution. So there
is no real possibility of forming any real complexes at all. No
immuno-reagent "sandwiches" are created in complete form, and
you get very low detection-signal results...passing for "non-reactive", "undetectable",
or "negative". |
Other
False Positives:
Rheumatoid Factor False Positives: this
is an IgM "anti-human-IgG" that interferes in test systems
in which reagent antibodies seek patient antigens or biochemical
molecules. As RF ties up the reagent antibodies, the test system
interprets the result as increased presence of the target antigen/molecule. |
| Heterophile antibodies: One should
also keep in mind the possibility that an unexpected positive is
due to the patient having produced circulating serum antibodies
[more detail] against the animal
(anti-animal...anti-mouse, anti-goat, etc...antibodies: HA) which
produced the reagents or tissues for the test system. And, there
have been vaccinations & disease therapies that involved injection
of your patient with animal-serum-based or generated reagents to
which the patient can have developed anti-animal antibodies. In
such instances, the patient asymptomatically carries the HA in
his/her serum and the HA reacts against the test-system's animal-generated
reagent antibodies to tie them up (consume them) as if they were
being consumed by the target antigen/molecule in patient's serum. |
| Of the antibodies that have been found in human serum that have
an animal specificity, human anti-mouse antibody (HAMA) is probably
the best known; but there are many others, including human anti-rabbit,
anti-goat, anti-chimera (for the chimeric antibodies that are being
used now in therapy), anti-pig, even anti-human, anti-bovine, rat,
and horse. |
| Animal husbandry and keeping pets have also been implicated as
mechanisms to generate anti-animal antibodies in humans. “For
example, if you’re looking for people who have got anti-mouse
antibodies circulating, probably the best place to go to is your
local mouse colony to see if the keeper of the mouse colony will
give you some blood.” |
| Since animal-based reagents are also used extensively in pharmaceuticals,
these can be another source of human anti-mouse antibody, human anti-rat
antibody, or antibodies against other species. |
| In our lab, this problem has most often come to light in pregnancy testing where discordant sero-clinical details are so readily notable. But we have recently also investigated a recent thyroid question [CP10-5] with test pattern discordance and celiac question [S10-3259] with sero-biopsy (CD3-amplified VTS) discordance. |
| Medications, Chinese medicine, herbal
drinks, and paraproteins: can interfere. |
Bacteria: such as E. coli can cause
false positives.
Poorly defined other interferences: ??. |
False
Negatives:
Serological tests, prozone type
reaction: the circulating, sought antibody is present in
such high titer that it saturates both the antibody capture component
of the test system AND at least partially saturates the conjugate
in solution so that no real complexes are formed & the test
looks negative or only low reactive. This phenomenon is the basis of the prolactin "hook effect" (a rising level unexpectedly "hooks back" to lower levels) in instances of pituitary adenoma which seems product negative (that is a prolactin producing adenoma with serum level "not elevated" (pituitary lesion 100 mm or larger ought to produce a product at a level greater than 250ng/ml3 |
Poorly defined other interferences: ??.
References:
-
Tozzoli R, et. al., "Guidelines for
the Lab. Use of Autoantibody Tests in the Diagnosis and Monitoring
of Autoimmune Rheumatic Disease", [The Italian Soc. of
Lab. Med. Study Group...], AJCP 117(2):316-324, Feb. 2002.
- CAP Today October 2003, p. 76-78. Or Cap
website , "Facing up to anti-animal antibody interference",
Oct. 2003.
- NewsPath, July 2010, a periodical letter from the LMC Lab (KRM).
- The effect of prevalance of disease on likelihood that a test result is a true positive.
- Another source concerning the "false positive" & "false negative" problem HERE.
(posted 2001; latest addition 23 May 2012) |
|
|
