Clinical Chemistry 47: 13-27, 2001;
(Clinical Chemistry. 2001;47:13-27.)
© 2001 American Association for Clinical Chemistry, Inc.
Carbohydrate-deficient Transferrin as a Marker of Chronic Alcohol Abuse: A Critical Review of Preanalysis, Analysis, and Interpretation
Torsten Arndt1
1
BioScientia, Institut für Laboruntersuchungen Ingelheim GmbH, Konrad-Adenauer-Strasse 17, D-55218 Ingelheim, Germany. Fax 49-6132-781-428;
arndt{at}bioscientia.de
 |
Abstract
|
|---|
Background: Carbohydrate-deficient transferrin (CDT) is used for
diagnosis of chronic alcohol abuse. Some 200–300 reports on CDT have
been published in impact factor-listed journals. The aims of this
review were to condense the current knowledge and to resolve remaining
issues on CDT.
Approach: The literature (1976–2000) was searched using MEDLINE
and Knowledge Server with "alcohol and CDT" as the search items.
The data were reviewed systematically, checked for redundancy, and
organized in sequence based on the steps involved in CDT
analysis.
Content: The review is divided into sections based on
microheterogeneity of human serum transferrin (Tf), definition of CDT,
structure of human serum CDT, pathomechanisms of ethanol-induced CDT
increase, preanalysis, analysis, and medical interpretation
(postanalysis). Test-specific cutoff values for serum CDT and causes of
false positives and negatives for chronic alcohol abuse are discussed
and summarized.
Summary: Asialo- and disialo-Fe2-Tf, which lack one
or two complete N-glycans, and monosialo-Fe2-Tf (structure
remains unclear) are collectively referred to as CDT. Diminished mRNA
concentration and glycoprotein glycosyltransferase activities involved
in Tf N-glycan synthesis and increased sialidase activity most likely
account for alcohol-induced increases in CDT. Knowledge about in vivo
and in vitro effects on serum CDT is poor. Reliable CDT and non-CDT
fractionation is needed for CDT measurement. Analysis methods with
different analytical specificities and recoveries decreased the
comparability of values and statistical parameters of the diagnostic
efficiency of CDT. CDT is the most specific marker of chronic
alcohol abuse to date. Efforts should concentrate on the
pathomechanisms (in vivo), preanalysis, and standardization of CDT
analysis.
|
Introduction
|
|---|
Carbohydrate-deficient transferrin
(CDT)1
is widely used for laboratory diagnosis of chronic alcohol
abuse. Since the first report by Stibler and Kjellin in 1976
(1), many studies on CDT have been published. The majority
of the CDT literature concentrates on the diagnostic efficiency of CDT
in different clinical settings and among various populations. At the
same time, preanalytical and analytical issues, which are fundamental
for reliable CDT analysis and thus for meaningful comparison of values,
and the diagnostic specificities and sensitivities of CDT obtained in
different studies by different analysis methods have not gained much
attention. Appropriate data are rare, scattered through the reports,
and mostly incomplete. However, routine laboratory diagnosis often
involves preanalytical and analytical issues of CDT analysis (in
addition to questions about the interpretation of CDT values).
Therefore, the aims of this review are to summarize and condense
current knowledge on the various aspects of CDT analysis, to resolve
the remaining issues, and to encourage studies for assessing these
issues. Finally, the review is aimed as a clear and easily accessible
CDT reference for routine laboratory and clinical use. Therefore, the
review follows the sequence of the analytical process:
microheterogeneity of human serum transferrin (Tf), CDT isoform
structure, pathobiochemistry, preanalysis, analysis, and interpretation
(postanalysis) of CDT.
|
Microheterogeneity of Human Serum Tf
|
|---|
Tf, the most important iron-transport protein, is synthesized
mainly in hepatocytes and consists of three substructural domains: a
single polypeptide chain, two independent metal ion-binding sites (one
within the N-terminal and the other within the C-terminal domain), and
two N-linked complex glycan chains (Fig. 1
). These Tf substructures show a distinct variability even under
nonpathological conditions (2). Because of this, Tf
is not a homogeneous molecule, but shows a distinct microheterogeneity
and is, from the analytical point of view, a group of similar Tf
isoforms collectively referred to as Tf. Human serum Tf appears in
serum protein electrophoresis on cellulose acetate within the
ß-globulin fraction. With high-performance electrophoretic methods,
e.g., isoelectric focusing (IEF) followed by immunoblot and staining, a
multitude of Tf bands become visible (2)(3).
Knowledge about Tf microheterogeneity is fundamental for correct
analysis and interpretation of CDT.
View larger version (16K):
[in this window]
[in a new window]
|
Figure 1. Microheterogeneity of human serum Tf attributable to
various Tf Fe3+ loads and different N-glycan chains.
The quantitatively most important Tf isoform is tetrasialo-Tf
(second panel from the top). After chronic alcohol
abuse, Tf isoforms lacking one or two complete N-glycan chains or
showing incomplete N-glycan chains (asialo-, monosialo-, and
disialo-Tf) appear with high prevalence in serum.
|
|
varying iron load
Tf is known to be the most important
Fe3+-transport protein in humans. Each Tf
molecule can bind a maximum of two metal irons, preferably
Fe3+. Depending on the Fe3+
supply of the organism, Tf molecules are iron free
(Fe0-Tf or apo-Tf) or loaded with one
(Fe1N- and Fe1C-Tf, where N
and C indicate the N- and C-terminal regions, respectively) or
two (Fe2-Tf) Fe3+ ions
(2). In healthy controls, Tf iron saturation is 
30%, and
Fe0-, Fe1-, and
Fe2-Tfs are detectable in serum (Fig. 1
). In
Fe3+ deficiency, Tf iron saturation decreases and
higher amounts of Fe0- and
Fe1-Tfs occur in serum. In hemochromatosis
(Fe3+ excess), Tf iron saturation increases and
the isoforms found in serum are almost exclusively
Fe2-Tfs. The isoelectric point (pI) of
the Tf molecule decreases by 0.2 pH units with each
Fe3+ ion bound (2).
differing N-glycan chains
The two Tf N-glycan chains differ in their degree of branching,
showing bi-, tri-, and tetraantennary structures
(2)(4). Each antenna of the Tf N-glycan chains
terminates with a (negatively charged) sialic acid molecule (Fig. 1
).
Because of this, asialo-Tf, and the sialylated forms monosialo- through
octasialo-Tf can occur in serum (2)(4). The
relative amounts of these Tf isoforms as a percentage of total serum Tf
are <1.5% heptasialo-Tf, 1–3% hexasialo-Tf, 12–18% pentasialo-Tf,
64–80% tetrasialo-Tf, 4.5–9% trisialo-Tf, and <2.5% disialo-Tf in
healthy persons (5)(6)(7). Asialo- and monosialo-Tf and
octasialo-Tf are not detectable (5) or represent <0.5%
(asialo-Tf) and <0.9% (monosialo-Tf), respectively, of total Tf under
nonpathological conditions (7). The pI of the Tf molecule
decreases by 0.1 pH units with each sialic acid residue bound to the
N-glycan chains (2)(4).
modified polypeptide chain (genetic Tf variants)
Genetic Tf variants are attributable to substitutions of amino
acid(s) in the polypeptide chain (2)(4). At
least 38 Tf variants are known (8). However, only 4 of these
show a prevalence of >1%. For Tf-C, the most important type in
Caucasians, 16 subtypes have been reported. Of these, Tf-C1 shows the
highest prevalence (>95%) in Caucasians (Fig. 2
).
View larger version (25K):
[in this window]
[in a new window]
|
Figure 2. Microheterogeneity of human serum Tf attributable to
various Tf Fe3+ loads, different N-glycan chains, and
genetic Tf variants (different polypeptide chains).
A maximum of 36 (for the homozygous Tf types; left and
middle) or 72 (for the heterozygous Tf type;
right) isoforms can be detected by IEF. Of these, only 3
(for the homozygous Tf types) or 6 (for the heterozygous Tf type)
isoforms are collectively referred to as CDT. At the same time, 9 (for
the homozygous Tf types) or 18 (for the heterozygous Tf type) non-CDT
isoforms (open boxes) with pIs almost identical to those
of CDT isoforms (gray boxes) appear in parallel and can
lead to overdetermination of CDT. To avoid this co-analysis and to
reduce the number of Tf isoforms potentially occurring in human serum,
a uniform Tf Fe3+ load is established by in vitro Tf
Fe3+ saturation (elimination of Fe0- and
Fe1-Tfs by formation of Fe2-Tfs).
|
|
Tf-B ("busy") and Tf-D variants can interfere with CDT analysis.
The pIs of non-CDT Tf-D isoforms are similar to the pIs of CDT Tf-C
isoforms, which can lead to cofocusing (or coelution) and thus
false-positive results for persons who are heterozygous Tf-CD and
consume normal amounts of alcohol (9)(10). As we
showed recently (11), Tf-D variants do not necessarily cause
overdetermination of CDT; it depends on the pI of the Tf-D subtype and
on the analytical specificity of the CDT analytical method. Tests
incorporating trisialo-Fe2-Tf into CDT are more
strongly affected by Tf-D variants than tests using the classical CDT
definition (11). Tf-B variants show diminished pIs and thus
increased electrophoretic mobility compared with Tf-C. Cofocusing (or
coelution) of the CDT isoforms of the Tf-B variant with non-CDT
isoforms of the Tf-C variant can produce false-negative results for
persons who are heterozygous Tf-CB and chronically abuse alcohol.
parallel changes in all three substructures of
Tf
Alterations in the three Tf substructures usually appear in
parallel (2)(4). Thus, the distinct
microheterogeneity of human serum Tf becomes even more pronounced. In
addition, Tf molecules with various iron loads show differing sialic
acid contents and/or a modified polypeptide chain (2). As
pointed out above, iron load, sialic acid content, and modifications in
the polypeptide chain affect the pI of the Tf molecule. Alterations in
the pI, e.g., when one or two Fe3+ ions are bound
or lost, can be compensated by the presence or absence of sialic acid
residues or genetic Tf variants. Thus, Tf molecules with different
amounts of iron and sialic acid but equal pIs, e.g.,
disialo-Fe2-Tf (as the main CDT isoform) and
tetrasialo-Fe1-Tf (as the main non-CDT isoform),
appear in serum (Fig. 2
). A maximum of 36 (for the homozygous Tf type)
or 72 (for the heterozygous Tf type) isoforms can be detected by IEF in
human serum. Of these, only 3 (for the homozygous Tf type) or 6 (for
the heterozygous Tf type) Tf isoforms are collectively referred to as
CDT (see Definition of CDT below). At the same time, 9 (for
the homozygous Tf type) or 18 (for the heterozygous Tf type) non-CDT
isoforms with pIs similar to those of the CDT isoforms appear in
parallel (Fig. 2
). Because of this, the main problem in CDT
analysis is the reliable separation of CDT isoforms from non-CDT
isoforms (see Analysis).
|
Definition of CDT
|
|---|
Stibler and Kjellin (1) first reported the presence of
Tf isoforms with pIs >5.7 in cerebrospinal fluid and serum from
alcoholics. Increased amounts of these Tf isoforms appeared with high
prevalence in serum from alcoholics and disappeared after abstinence,
with a half-life of 14 days
(9)(12)(13). These Tf isoforms,
corresponding to asialo-Fe2-Tf,
monosialo-Fe2-Tf, and
disialo-Fe2-Tf, were later collectively referred
to as CDT (9)(14).
trisialo-Fe2-Tf and cdt
There has been a debate as to whether there is an alcohol-induced
increase of trisialo-Fe2-Tf and/or a diagnostic
benefit from including (parts of) of this isoform in CDT
(7)(15)(16)(17). Dibbelt (18) recently
demonstrated by HPLC that the trisialo-Fe2-Tf
concentration was statistically the same in serum samples with
nonpathological and pathological CDT concentrations. Increased
concentrations of the CDT isoforms were not associated in general with
increased trisialo-Fe2-Tf. Thus, Dibbelt stated
that "trisialo-transferrin is obviously of no diagnostic value" and
should not be included in CDT (18). This finding has been
supported by two studies showing a diminished diagnostic accuracy
(sensitivity) of so-called "trisialo-tests", e.g., the
%CDT-turbidimetric immunoassay (%CDT-TIA or %CDTri-TIA), which
include 50% of trisialo-Fe2-Tf in CDT,
compared with CDT tests that exclude this isoform from CDT, e.g.,
CDTect (19) and ChronAlcoI.D. (this work group, submitted
for publication).
|
Structure of Human Serum CDT
|
|---|
Serum samples from alcoholics showed decreased sialic acid content
(20) and normal Tf isoform patterns after treatment with
neuraminidase (complete removal of sialic acid residues from the Tf
N-glycan chains) (13). Thus, a defective Tf carbohydrate
structure, especially a sialic acid deficiency, in serum Tf from
alcoholics was assumed. Subsequent investigations revealed not only
missing sialic acid residues but also a lack of deeper links in the
N-glycan chains (galactose, mannose, and
N-acetylglucosamine) and an unaffected sequence of mannose,
mannose, N-acetylglucosamine, N-acetylglucosamine
(the latter bound directly to the polypeptide chain) (21).
Recent studies showed that the main CDT isoforms,
disialo-Fe2-Tf and
asialo-Fe2-Tf, lack one or two complete glycan
chains (22)(23)(24). According to Landberg et al.
(22), disialo-Fe2-Tf shows a single
biantennary N-glycan chain with two sialic acid residues, whereas
asialo-Fe2-Tf has no carbohydrate structure (Fig. 1
). Trisialo-Fe2-Tf contains two biantennary
N-glycans, one with two terminal sialic acid molecules, the other with
one terminal sialic acid and one terminal galactose (Fig. 2
). Because
of this, it is impossible for trisialo-Fe2-Tf to
form from pentasialo-Fe2-Tf (containing one
biantennary, disialylated N-glycan and one triantennary trisialylated
N-glycan; Fig. 1
) by loss of the biantennary N-glycan.
Furthermore, the structure of monosialo-Fe2-Tf,
which is part of CDT, remains unclear. Landberg et al. (22)
did not assess this isoform. Peter et al. (23) reported a
fraction "Ib", with molecular masses between fraction "Ia"
(asialo-Fe2-Tf) and "IV"
(tetrasialo-Fe2-Tf), low sialylation, and
terminal galactose. The authors hypothesized that this fraction is a
mixture "of transferrins with two N-glycan chains with shortened
antennae" and that it "probably contains small amounts of
transferrin with one N-glycan chain" (23). One might
speculate that this fraction contains
monosialo-Fe2-Tf. It follows from these studies
that the presence of biantennary, trisialo-Fe2-Tf
(22) and monosialo-Fe2-Tf in serum
after chronic alcohol abuse cannot be explained solely by the lack of
one or two complete glycan chains. One might argue that
monosialo-Fe2-Tf is a quantitatively less
important CDT isoform (7) and that
trisialo-Fe2-Tf is not a CDT isoform
(18). Nevertheless, assessment of the structure of these Tf
isoforms might be of value to further disclose the pathomechanisms of
ethanol-induced increases in CDT. Summarizing the studies by Landberg
et al. (22), Peter et al. (23), and recently by
Henry et al. (25), it seems most likely that CDT isoforms
can lack not only complete N-glycan chains (mainly
asialo-Fe2-Tf and
disialo-Fe2-Tf), but also parts of the N-glycan
chains [as discussed previously by Stibler and Borg (21)].
If this is true, the production of specific CDT antibodies, which could
be used in direct (homogeneous) CDT assays, becomes additionally
complicated.
|
Pathomechansisms of Ethanol-induced CDT Increase
|
|---|
The pathomechanisms for the increase in CDT isoforms during
chronic alcohol abuse are not completely understood at present. It is
most likely that ethanol and/or its metabolite acetaldehyde affect
N-glycan chain synthesis in the Golgi apparatus. Thus, Stibler and Borg
(26) measured diminished activities of galactosyltransferase
and N-acetylglucosaminyltransferase in serum from
alcoholics. The addition of in vitro acetaldehyde to these samples
further decreased the enzyme activities (26). Disulfiram (an
antabuse for treatment of alcoholism), which inhibits aldehyde
dehydrogenase, leading to an accumulation of serum acetaldehyde and
subsequent nausea, did not affect serum CDT (27). Decreased
incorporation of 3H-labeled leucine and
N-acetyl-D-mannosamine into Tf, lower
2,6-sialyltransferase mRNA concentrations (because of
destabilization by ethanol), decreased synthesis of
2,6-sialyltransferase, which is followed by lower sialyltransferase
activity, and decreased sialylation of Tf in rats chronically fed
alcohol were observed by Lakshman et al. (28). Xin et al.
(29) measured increased sialidase activity in liver plasma
membranes and reduced sialyl, galactosyl, and
N-acetylglucosaminyltransferases in Golgi homogenates from
alcohol-treated rats. This loss in transferase activity was also
observed when acetaldehyde was added to Golgi homogenates from normally
fed control rats (29). In this connection, a finding by Fast
et al. (30) is interesting. In assessing the role of the
carbohydrate chains of sialyltransferase (EC 2.4.99.1) for enzyme
activity, they observed diminished enzyme activity after partial
removal of the N-glycans by N-glycanase. The presence of
methanol or ethanol was necessary for complete deglycosylation. There
was a correlation between the loss of catalytic activity of the enzyme
and increased deglycosylation. One might speculate whether the
diminished activities of sialyltransferase (galactosyltransferase and
N-acetylglucosaminyltransferase) reported by Xin et al.
(29) are attributable to incomplete carbohydrate structures
of these enzymes and whether ethanol (or its metabolites) primarily
affects the glycosylation of the carbohydrate transferases that are
involved in Tf N-glycan synthesis. The results obtained by Lakshman et
al. (28) point to primary effects of ethanol on
sialyltransferase mRNA production.
In contrast to other glycoproteins, the lack of endstanding sialic acid
residues in CDT isoforms does not cause accelerated hepatic clearance
via the asialoglycoprotein receptor. Thus, the plasma half-life of CDT
is 14 days (9)(31)(32)(33), and that of Tf only
7 days (34). Patients with liver cirrhoses, regardless of
whether they were alcohol induced, showed normal arteriovenous CDT
gradients (33). Because of this, ethanol-induced alterations
of hepatic and/or renal CDT clearance seem to be unlikely as causes of
increased CDT after chronic alcohol abuse. One reason for high
CDT concentrations despite normal alcohol consumption is the
carbohydrate-deficient glycoprotein (CDG) syndrome, a hereditary
disorder of glycoprotein synthesis (35). Structural
similarities between serum glycoproteins (Tf,
1-antitrypsin, and haptoglobin ß chains)
from patients with chronic alcohol abuse and patients suffering from
the CDG syndrome were reported by Henry et al. (25). These
findings suggest that alcohol-induced increases in CDT might also be
attributable to an inhibition of the initial mannose-dependent steps of
Tf N-glycan synthesis (25). Because some types of the CDG
syndrome are attributable to deficiencies of phosphomannomutase (EC
5.4.2.8) or phosphomannose isomerase (EC 5.3.1.8), it might be
interesting to measure the corresponding enzyme activities in patients
with chronic alcohol abuse.
Our investigations revealed fluctuations between nonpathological and
pathological CDT concentrations in serum from patients with combined
kidney and pancreas transplantation and normal alcohol consumption.
Patients with kidney transplantation alone always had CDT
concentrations within reference values (36). It is
known that pancreas transplantation causes hyperinsulinemia (by
connecting the endocrine pancreas to the blood system and thus avoiding
the first-pass effect of the liver) and bicarbonate
deficiency/metabolic acidemia (by draining the exocrine pancreas into
the bladder). Because metabolic acidemia is common after alcohol abuse
and persistent after chronic alcohol abuse, it might also be
interesting to assess the acid-base balance and serum CDT from
alcoholics in parallel.
correlation between serum cdt and alcohol intake
Data concerning the correlation between the amount of ethanol
consumed and serum CDT are inconsistent. This might be the result of
ethical limitations on controlled drinking studies. According to Allen
et al. (31), almost all studies on this subject are invalid
except for those by Stibler and co-workers
(13)(14) and Storey et al.
(37). It is interesting that the critical alcohol
intake [at least 50–80 g ethanol/day on 7 consecutive days
(9)] for an increase in CDT is almost the same as for
alcohol-induced liver cirrhosis (38).
|
Preanalysis
|
|---|
Several preanalytical conditions have been found to affect serum
CDT concentrations. Stibler et al. (14) found that EDTA and
heparin may disturb in vitro Fe3+-Tf saturation
and/or anion-exchange microcolumn non-CDT and CDT isoform fractionation
(substance-specific data not given). The authors also found that sample
storage for 3 days at room temperature produced a 25% increase in CDT.
In addition, lipemia can interfere with turbidimetric CDT assays.
Delipidation reduces the serum CDT concentration by 22%
(14). Strong hemolysis can also lead to false-positive
results (our unpublished data).
Preanalytical conditions that do not affect the serum CDT concentration
include circadian serum CDT fluctuations [8% (39);
intraassay CV = 10%
(11)(40)(41)]; collection of blood
into containers with kaolin (coagulation accelerator) and/or
polyacrylamide-gel separator (42); serum storage for 30 h
at room temperature (39), 7 days at 4 °C (39),
and several months at -22 °C
(14)(36)(39); repeated freezing and
thawing (36)(39); diet (33); common
drugs taken by patients of general practitioners
(14)(43); and disulfiram (27).
Conditions that need to be studied include positioning of the patient
during blood collection, duration of congestion, in vivo and in vitro
drug effects, use of EDTA- or heparin-plasma, and stability of
whole-blood samples.
|
Analysis
|
|---|
The analysis of CDT makes very high demands on selectivity,
specificity, and sensitivity for three main reasons: (a) the
distinct serum Tf microheterogeneity; (b) the distinct
structural similarity of CDT and non-CDT isoforms; and (c)
the low CDT isoform concentrations [<2.5–2.7% in healthy controls
and <20% in alcoholics as measured by the ChronAlcoI.D. assay
(44) in the presence of large amounts of non-CDT isoforms
with almost similar physicochemical properties].
CDT-specific reactions or CDT antibodies, and thus a homogeneous CDT
assay, are not available at present. Thus, routine laboratory analysis
of serum CDT requires separation of CDT from serum matrix constituents
and from non-CDT isoforms. This can be achieved by chromatographic
(e.g., anion-exchange) or electrophoretic (e.g., IEF) methods, using
the different charges and pIs of CDT and non-CDT isoforms. As pointed
out above, there are coexisting CDT and non-CDT isoforms with almost
equal pIs, e.g., disialo-Fe2-Tf (as the main CDT
isoform), tetrasialo-Fe1N-Tf,
pentasialo-Fe1C-Tf, and
heptasialo-Fe0-Tf (Fig. 2
). To reduce the number
of Tf isoforms occurring in the native serum sample and to preclude
coexistence of CDT and non-CDT isoforms with equal pIs, CDT analysis
usually starts with in vitro Tf Fe3+ saturation;
this treatment establishes a uniform Tf iron load.
Fe1- and Fe0-Tfs are
transformed to Fe2-Tfs, and only
Fe2-Tfs appear in the serum sample. The in vitro
Tf iron saturation step is followed by fractionation of the CDT and
non-CDT Tfs by electrophoresis or chromatography. Subsequently, the CDT
isoforms are detected by immunological procedures. [CDT and non-CDT
isoform separation can also be done after in vitro removal of
Fe3+ from Tf by complexation with EDTA, which
leaves Fe0-Tfs as the only isoforms in serum.
However, the selectivity of IEF is improved when focusing
Fe2-Tf isoforms (3)]. Regardless of
which fractionation procedure and which immunoassay is used, a complete
and stable Tf iron load is fundamental for reliable CDT analysis.
Incomplete in vitro Tf Fe3+ saturation or Tf iron
loss during fractionation of the CDT and non-CDT isoforms inevitably
leads to reformation of Fe1- and/or
Fe0-Tfs, co-elution of CDT and non-CDT isoforms
with equal pIs but different sialic acid and iron content, and
overdetermination of CDT (3). This point must be assessed
when developing (and launching) a new CDT analytical method
(11)(40).
electrophoretic methods
Because of its high selectivity, IEF is used as the reference
method for serum Tf isoform analysis. The Tf isoforms are separated in
a gel containing a pH gradient according to their characteristic pIs.
Again, the first analytical step is in vitro saturation of Tf with
Fe3+. After electrophoresis, the Tf bands are
visualized by immunofixation and staining of the CDT-anti-Tf complexes
(3). Finally, the Tf band patterns can be evaluated by
densitometry. It should be taken into account, however, that
application of a sufficient volume of serum for detection of CDT
isoforms to the gel (1 µL diluted 400-fold) produces an overload of
tetrasialo-Tf. It follows from this that the intensity of the main Tf
fraction does not correlate with the amount of Tf focused in this band
(3). This complicates quantitative evaluation of this and
the other Tf fractions. From the analytical point of view, this
drawback can be overcome by use of ratios of the different CDT
isoforms, e.g., disialo-/asialo-Fe2-Tf
(11)(40)(45)(46).
However, absolute CDT concentrations can also be obtained by IEF with a
calibration curve generated by different amounts of
asialo-Fe2-Tf (47). Many IEF methods
suggested for quantitative CDT analysis suffer from incomplete
documentation of detection limits, recovery (e.g., by Western
blotting), intra- and interassay CVs, and correlation between peak
height (densitometry) and amount of the Tf isoform.
Because of its high selectivity, IEF can detect genetic Tf variants
without in vitro neuraminidase treatment
(10)(11). The latter is done for complete
removal of Tf isoform sialic acid residues and thus formation of only
asialo-Tfs (10)(13)(47). Assuming
complete Tf iron load and Tf sialic acid loss, only one band of
asialo-Fe2-Tf would be detected in serum with a
homozygous Tf type (e.g., Tf-C1), but two bands would be detected in
serum with a heterozygous Tf type (e.g., Tf-CD or Tf-CB). IEF is
suitable for visualization of Tf isoform patterns in the anion-exchange
microcolumn eluates of (commercial) CDT tests. It is therefore an
essential method for testing the reliability of the initial
fractionation step used to separate CDT and non-CDT isoforms in current
and upcoming CDT assays based on anion-exchange chromatography
(11)(40). IEF is recommended for assessing
unexpected CDT values (40)(48). IEF of serum Tf
isoforms can also be used for diagnosis of the CDG syndrome [for a
review, see Jaeken and Carchon (35)].
In addition to IEF in flat gels, capillary electrophoresis
(49)(50)(51) and capillary zone electrophoresis
(15)(52)(53) have been proposed for
analysis of Tf isoforms. The main problem with these techniques is
coating of the capillary surface to prevent protein adsorption and
finding a coating-compatible, highly ultraviolet-transparent buffer
(53). Only a tradeoff between these criteria is achievable
at present. Thus, it is questionable that Tagliaro et al.
(53) used uncoated (bare fused-silica) capillaries in their
study without assessing the accuracy of the method, which was used for
quantitative analysis of CDT.
Recently, Crivellente et al. (15) described improved
analytical sensitivity and separation power of the capillary zone
electrophoresis method described by Tagliaro et al. (53).
The improved sensitivity and separation were obtained mainly by the
addition of diaminobutane to the running buffer and an increase in
column length (15). Nevertheless, the selectivity and
sensitivity of capillary electrophoresis are still lower than IEF. When
Tagliaro et al. (53) analyzed a serum sample with 61
units/L CDT (CDTect assay), indicating strong chronic alcohol
abuse, asialo- (the second most common CDT isoform) and monosialo-Tf
were not detected. Even with their improved method, Crivellente et al.
(15) could not detect asialo-Fe2-Tf.
In contrast, IEF of serum samples with only 25 units/L (CDTect assay;
cutoff, 18 units/L) clearly detected bands for mono- and
asialo-Fe2-Tf (36). Lectin affinity
electrophoresis has been described for assessing Tf microheterogeneity
in patients with alcoholic liver disease (24). Whether these
methods can gain wide acceptance for routine CDT analysis is still
unclear.
chromatographic methods
Compared with IEF, chromatographic CDT methods are less sensitive
(sample volumes of 100–500 µL are needed) and less selective. Thus,
genetic Tf variants cannot be detected by anion-exchange chromatography
followed by immunoassay (a procedure usually used by commercial CDT
tests). Using a HPLC method described by Jeppson et al.
(54), Simonsson et al. (55) could reliably detect
genetic Tf variants in serum from healthy blood donors. Thus, HPLC can
also be used for assessing odd CDT values obtained by anion-exchange
chromatography/immunoassay. However, IEF is still superior to HPLC in
detecting and phenotyping genetic Tf variants with pI values close to
that of tetrasialo-Fe2-Tf, found in the most
common Tf-C1 phenotype. Analysis times of 20 min/sample
(54)(55)(56) and a time-consuming HPLC column regeneration
(54) further reduce the applicability of HPLC for large CDT
analysis series.
Recently, Yoshikawa et al. (57) described a lectin affinity
chromatography method for determination of CDT. Using Sepharose columns
coated with Allomyrina dichotoma (Allo A) or
Trichosanthes japonica (TJA) lectins, the authors were able
to separate CDT-Allo A (corresponding to disialo-Tf) and CDT-TJA
(corresponding to asialo-Tf) fractions from serum. The diagnostic
efficiency obtained with these fractions was higher than that
of the %CDT-TIA (Axis) including
trisialo-Fe2-Tf.
commercial cdt tests
In 1993, the first commercial CDT test (CDTect-RIA; Pharmacia &
Upjohn) was introduced, followed by %CDT (Axis) and the CDTect enzyme
immunoassay (CDT-EIA; Pharmacia & Upjohn). These methods use the common
CDT definition when analyzing asialo-, monosialo-, and
disialo-Fe2-Tf. Later, a test incorporating part
(50%) of the trisialo-Fe2-Tf in CDT was
developed by Axis and distributed by Axis and Bio-Rad (as the %CDT-TIA
and %CDTri-TIA, respectively) and Roche
(Tinaquant-%CDT/transferrin). Except for one HPLC method by Recipe,
currently available commercial CDT tests, even the relatively new
ChronAlcoI.D. assay, are based on fractionation of CDT and non-CDT
isoforms on anion-exchange microcolumns. Table 1
lists the analytical specificities of commercial and
noncommercial CDT tests and the corresponding test-specific cutoff
values for CDT indicating chronic alcohol abuse.
need for standardization of cdt analysis
The availability of CDT assays for routine laboratory diagnosis
increased and accelerated the acceptance of CDT as one of the most
specific markers of chronic alcohol abuse to date. At present,
200–300 reports on CDT have been published in impact factor-listed
journals. However, the definition of CDT is increasingly vague (which
Tf isoforms are CDT isoforms, i.e., asialo-, monosialo-, disialo-
and/or trisialo-Fe2-Tf; Table 1
). Standardization
of CDT analysis by an international CDT standard and quality-control
material has not been attained [although it is possible to generate
CDT enzymatically
(55)(56)(58)]. Thus,
different Tf isoforms are analyzed, with various recoveries, as CDT
(Table 1
), which exacerbates the already poor comparability of CDT
values, diagnostic specificities and sensitivities, and predictive
values obtained in different clinical settings and among different
populations with different CDT analytical methods. To overcome this
problem, redefinition of CDT (59) or consistent use of the
common CDT definition introduced by Stibler (9) is needed.
As a result of an international meeting on standardization of CDT
analysis (in Berlin, May 2000), the aim is to develop a highly
sensitive HPLC for analysis of CDT as a standard method (e.g., for
calibration of CDT test-clinical analyzer applications) and to cease
producing so-called "trisialo tests".
The lack of standardization of CDT analyses should not discourage the
use of CDT, or encourage the use of less specific markers, e.g., liver
enzymes or uric acid. Hemoglobin A1C, which is
widely used as an integrative long-term marker of blood glucose
concentration, is far from being well standardized. Nevertheless,
measurement of an individual’s hemoglobin A1C
concentration improves therapy for diabetes.
|
Medical Interpretation (Postanalysis)
|
|---|
normal ranges for serum cdt
Because of different analytical specificities and recoveries,
normal ranges (reference values or cutoffs) for absolute and relative
serum CDT are method-dependent. CDT values must always be interpreted
with regard to the test-specific decision criteria (Table 1
).
Therefore, the laboratory should report the CDT value, the cutoff
value, and the method of analysis. Such information is common in
forensic drug analysis; it makes the comparison of reports from
different laboratories much easier, is needed in follow-up studies, and
should also be provided for CDT analyses. Changing the CDT test can
cause a sharp increase or decrease of CDT values (Table 1
), which can
lead to a misinterpretation of the actual drinking status. Thus, it is
valuable to announce an upcoming change in the CDT test and to give the
CDT results by the "old" and the "new" CDT assay for a suitable
time period in parallel. Taking into account the social consequences of
high CDT values, it is advisable to use borderline values (95th
percentile plus analytical imprecision) instead of cutoffs
(44) (Table 1
). Relatively constant serum CDT concentrations
have been found in healthy persons, in patients with
non-alcohol-related liver diseases, or during abstinence
(60)(61). Individual reference ranges were
valuable for follow-up during alcohol withdrawal treatment
(62).
gender dependence of serum cdt
Absolute serum CDT concentrations from healthy women typically are
higher than those of healthy men
(9)(31)(63)(64)(65)(66). Recently, de Feo et
al. (67) reported no gender dependence of serum CDT in
nonabusers and alcohol abusers, but a positive correlation between CDT
and Tf in nonabusers. The latter was also reported by van Pelt and
Azimi (68). Because females frequently have subclinical iron
deficiency and thus increased Tf concentrations, this might explain
their higher CDT concentrations. However, these findings are surprising
because it is commonly accepted that CDT and Tf do not correlate (see
below). The reason for higher CDT concentrations in healthy women
compared with healthy men remains unclear
(63)(64), in spite of the findings of de Feo et
al. (67) and van Pelt and Azimi (68).
Serum CDT:Tf ratios are not gender-dependent
[(44)(54)(69), and manufacturers’ test instructions].
No correlation has been found between CDT and the menstrual cycle,
serum estradiol, serum iron, or contraception
(65)(66). Compared with premenopausal women,
serum CDT was slightly increased (17.5 vs 19.3 units/L) in menopausal
women receiving estrogen therapy (65)(66).
Whether this slight difference is of practical importance [CVs of the
CDT tests of 10% must be considered
(11)(40)(41)] and whether an
association among menstrual function, CDT, and drinking behavior does
really exist, as assumed by Oslin et al. (70), needs further
investigation. Data pointing to a relationship between the amount and
frequency of menstrual blood loss and CDT were recently presented by
Leusink et al. (71).
Increased serum CDT concentrations during pregnancy have been reported.
However, serum CDT did not exceed the gender- and CDTect test-specific
borderline of 26–28 units/L even in the third trimester
(65)(66). Interestingly, de Jong and van Eijk
(72) found a steady increase of tetrasialo-Tf and higher
sialylated Tf isoforms and unaltered concentrations of sialic
acid-deficient Tf isoforms in pregnant women during the second and
third trimesters and also in women using contraceptives. In accordance
with these observations, the CDT:Tf ratio should decrease (and not
increase as usually discussed) in these women. This could be one reason
for the diminished diagnostic sensitivity of serum CDT measurement in
women compared with men (see below). Further efforts, in addition to
those described by Mårtensson et al. (7), should be applied
to identification of the Tf isoform patterns in healthy men and
(pregnant) women. Another underlying cause can be gender-dependent
differences in the drinking patterns and in the vulnerability of the
liver to alcohol intoxication, as discussed recently (73).
correlation with total Tf
Except for two recent studies (67)(68), no
significant correlation between serum CDT and total Tf has been found
(9)(14)(74)(75). A
slightly improved diagnostic sensitivity was reported by Schellenberg
and Weill (45)(76), who used a "Tf-index",
referred to as the ratio of Tf isoform with a pI of 5.7 to the isoform
with a pI of 5.4
(disialo-Fe2-Tf/tetrasialo-Fe2-Tf).
Determining this ratio requires IEF, which is sometimes laborious and
requires experienced personnel. This might be one reason why the
Tf-index has not gained much acceptance. Automated IEF, such as
that described by Hackler et al. (3) and Arndt and
co-workers (36)(46), should be helpful in
further evaluating the Tf-index. For routine purposes, the Tf-index has
been replaced by the CDT:total Tf ratio. Whether this ratio increases
the diagnostic accuracy compared with absolute CDT concentrations (and
in each situation) is controversial
(7)(19)(31)(44)(48)(77)(78)(79)(80)(81).
Contradictory results were found in a single study using CDT:Tf ratios
(67), which demonstrated improved diagnostic specificity in
patients with iron-deficiency anemia but no decrease in false-negative
results in patients with alcohol abuse. Similar diagnostic
specificities for absolute CDT concentrations (units/L or mg/L) and
CDT:Tf ratios were described by Lieber (48) and Helander
(81). The main advantage of the CDT:Tf ratio is that it
improves the diagnostic specificity of CDT in patients with increased
Tf (although it is generally assumed that CDT and Tf do not correlate),
e.g., for patients with iron deficiency (reduced iron stores) or
iron-deficiency anemia but normal alcohol intake (67).
However, low total serum Tf concentrations (e.g., from acute or chronic
infection, hemochromatosis, nephrotic syndrome, or cancer) can produce
both false-negative [e.g., in hemochromatosis (67)] and
false-positive results. Thus, Lieber (48) found in some
liver disease patients increased CDT:Tf ratios that were attributable
to decreased serum Tf concentrations. Additionally, CDT:Tf ratios are
less precise than absolute CDT concentrations (the CV of the CDT:Tf
ratio is the sum of the CV of the CDT concentration plus the CV of the
total Tf). Assuming a CDT concentration of 100 mg/L and a total Tf
concentration of 4000 mg/L, the CDT:Tf ratio is 2.5%. Because of the
analytical imprecision, which is 10% for CDT and Tf, the CDT:Tf
ratio shows an imprecision of at most 20% and varies between 2.2% (90
mg/L CDT, 4040 mg/L Tf) and 2.8% (110 mg/L CDT, 3960 mg/L Tf). When
borderlines of 2.5–2.7% (CDT:Tf) and 100–110 mg/L (CDT) are used,
the CDT:Tf ratio changes among normal (2.2%), borderline (2.5%), and
increased (2.8%). At the same time, the absolute CDT concentration
does not exceed the upper limit of the borderline. Knowing this is
important for interpretation of CDT values in longitudinal studies or
for control of relapse drinking. Our investigations revealed that
measuring (and interpreting) absolute and relative CDT concentrations
as well as the total serum Tf concentration improves the diagnostic
efficiency of CDT as a marker of chronic alcohol abuse (unpublished
data). It might be interesting to reevaluate the data from the many
clinical studies, taking into consideration absolute serum CDT
concentrations, CDT:Tf ratios, and total serum Tf concentration.
cdt and 
-glutamyltransferase
In contrast to CDT analysis, measurement of
-glutamyltransferase (-GT) activity is highly standardized,
automated, and inexpensive. Because of this, -GT is the most
commonly used marker of alcohol abuse other than ethanol. It is obvious
to ask whether the determination of -GT alone is sufficient and
whether CDT values are really needed for laboratory diagnosis of
alcohol abuse. Many studies on CDT investigated -GT activities as
well. The correlation between -GT and CDT, the diagnostic
sensitivities and specificities of CDT and -GT, and the diagnostic
efficiencies of considering -GT and CDT were studied alone or
conjointly. Reviews of the literature up to 1994 (82) and
between 1966 and November 1998 (83) have been published.
Since then, several studies on these subjects have appeared. Because of
the different clinical settings with many different populations varying
in sex, age, drinking amounts and patterns, and clinical backgrounds as
well as the multiple analytical and statistical methods used, it is
almost impossible to draw general conclusions from this vast amount of
data: There was no distinct correlation between absolute or relative
serum CDT concentrations and -GT
(9)(75)(84)(85)(86)(87). Therefore, parallel
analysis of CDT and -GT makes sense; this analysis revealed
additional diagnostic information for follow-up of alcoholics with
normal -GT activities (84), for brief intervention among
heavy drinkers in primary healthcare (88), for evaluating
progress of patients in treatment for alcoholism (73), for
assessing alcohol consumption in a general medical clinic
(89), for monitoring alcohol consumption in men drinking
20–60 g of ethanol/day (90), and for detecting heavy
alcohol consumption in female alcoholic outpatients and college
students (91). As a general conclusion, CDT is currently the
more specific and -GT the more sensitive marker of chronic alcohol
abuse. The latter is true especially for women
(70)(73)(91), but it is not clear
whether it is related strictly to gender or to the higher vulnerability
of a woman’s liver to alcohol-induced damage (73).
Compared with CDT, -GT produces false-positive results regarding
chronic alcohol abuse in many states of disease, e.g., obstructive
liver disease, posthepatic obstruction, hepatitis, fatty liver, liver
cirrhosis, liver cell carcinoma and liver metastases, cardiac
insufficiency, mononucleosis, renal transplant, hyperthyroidism,
myotonic dystrophy, diabetes mellitus, and pancreatitis
(92)(93). In contrast to CDT
(14)(43), -GT is highly affected by several
medications and drugs of abuse, e.g., barbiturates, cephalosporins,
estrogens, oral contraceptives, phenytoin, primidone, thyrostatics,
anabolic steroids, phenothiazines, and antirheumatics
(92)(93).
diagnostic efficiency of cdt
Numerous clinical trials have been published regarding the
diagnostic specificity and sensitivity of serum CDT as a marker of
chronic alcohol abuse. Several reviews have been published
(9)(31)(82)(94)(95).
A systematic review of CDT comparing CDT and other markers of
alcoholism based on a MEDLINE database from 1966 to November 1998 was
provided by Salaspuro (83). Since then, many new reports
have appeared in journals listing impact factors. Most of these reports
have already been discussed here. It is not the aim of this review to
compare the diagnostic criteria (e.g., specificities, sensitivities,
predictive values, ROC plots) obtained for the different analytes
(e.g., -GT and CDT) in different populations (e.g., differences in
sex, ethnicity, hospitalization, and no/mild/moderate/heavy/chronic
alcohol abuse) by different CDT methods (e.g., electrophoresis,
chromatography, absolute/relative CDT concentrations, and different
definition of the CDT group). For calculating parameters of diagnostic
efficiency, reliable data regarding individual alcohol consumption are
a prerequisite. Thus, for classifying false positives and true
negatives, an objective "gold standard" is needed that
determines the individual alcohol consumption without any error.
Clinical background or personal reviews are customarily used for this
purpose. However, if both were faultless, we would not need laboratory
markers of chronic alcohol abuse. Because underestimating alcohol
consumption is common, false true positives or false false negatives
will occur and markedly affect the criteria of diagnostic efficiency of
CDT obtained in a specific clinical study. Additionally, in one study,
diagnostic sensitivities and specificities, and positive and negative
predictive values varied by 12% when serum CDT concentrations for the
same healthy controls and hospitalized patients were analyzed by
different CDT tests (96). This should be considered when
interpreting parameters of diagnostic efficiency.
diagnostic specificity of cdt
The main causes for false-positive results in CDT analysis have
been described by Stibler and co-workers
(9)(14). Conditions currently known to
potentially cause false positives and thus decrease the diagnostic
specificity of CDT are summarized in Table 2
. In interpreting these data, one should not generalize findings
made on some patients of small (and in part highly selected)
populations (see comments in Table 2
). Altogether, there are fewer
sources of false positives when CDT is used to monitor alcohol abuse
than -GT. Some of these sources are relatively easy to differentiate
from alcohol history (Table 2
). Diagnostic specificities of CDT
obtained in several different clinical studies have been published
(9)(31)(82)(83)(94).
View this table:
[in this window]
[in a new window]
|
Table 2. Clinical conditions, regardless of the number of
observations, that caused false positives when serum CDT was used as a
marker of chronic alcohol abuse.
|
|
In summary, CDT is still the most specific laboratory marker of chronic
alcohol abuse.
diagnostic sensitivity
There are conditions that can affect the diagnostic sensitivity of
CDT as a marker of chronic alcohol abuse (Table 3
). Differences in the drinking patterns during an appropriate
period before blood sampling (long-term drinking pattern) and within
the individual days of this period (short-term drinking patterns) must
be assessed. Because the half-life of CDT is 14 days, at least 14
days (4 weeks is better) before blood collection must be assessed. The
results of a 3-week drinking experiment showed that chronic consumption
of small amounts of alcohol affected serum CDT concentrations
[confirmed by Whitfield et al. (85)], but short-time
drinking of larger amounts do not (97). Because of
ethical considerations, long-time drinking experiments are difficult to
perform. However, it is known that ethanol blood concentrations are
affected not only by the amount of alcohol consumed, but also by the
amount per time period and by the body mass (distribution volume): the
shorter the period in which a gram of alcohol is consumed and the lower
the body mass, the higher the peak blood alcohol concentration. Thus,
one might hypothesize that chronic short-term consumption of high
amounts of alcohol by a lean person causes high peak blood alcohol
concentrations and substantial liver injury, and thus affects serum CDT
concentrations or the dose–response curve. A relationship between CDT
insulin sensitivity and hypertension in men of a very specific patient
sample were discussed by Fagerberg et al. (98).
View this table:
[in this window]
[in a new window]
|
Table 3. Conditions reported to affect diagnostic sensitivity of
serum CDT concentration as a marker of chronic alcohol
abuse.
|
|
The main concern about CDT as a marker of chronic alcohol abuse is its
relatively low diagnostic sensitivity. Sensitivities of >90% and up
to 100%, as reviewed by Stibler (9), led to great hopes for
CDT. Subsequent studies, reporting distinctly lower diagnostic
sensitivities, could not meet these expectations [reviewed in Refs.
(31)(83)]. Recently, a diagnostic sensitivity
of 0% for women with or without liver diseases was reported
(99). The authors’ final statement that their data
"underline the insufficient specificity of CDT" and that "it is
not justified to base any medical decision on the measurement of CDT
concentrations, ... even more for forensic decisions" cannot be
drawn from their own data. Thus, diagnostic specificities of 83.6% and
94.2% for men with or without liver disease and 96.9% and 91.9% for
women are distinctly better than those found for -GT: 36.1% and
24% for men and 36.6% and 50% for women. The shortcomings of this
study were discussed by Allen and Sillanaukee (100) and
Arndt et al. (101). Except for this study, diagnostic
sensitivities of 30–50% for women and 50–70% for men, depending on
factors such as population, drinking patterns, and daily alcohol
intake, seem to be a good average.
|
Conclusions
|
|---|
Our knowledge of in vivo and in vitro effects on serum CDT is poor
but necessary for reliable CDT analysis. CDT analytical methods with
different analytical specificities and recoveries decrease the
comparability of CDT values and statistical parameters of the
diagnostic efficiency of CDT obtained in different studies. A unique
definition for and standardization of CDT are needed. CDT is not a
screening tool for detection of increased alcohol consumption (and it
was originally not aimed as such). CDT is the most specific marker of
chronic alcohol abuse to date. Diagnosis of chronic alcohol abuse
should always be made based on a clinical background, questionnaire,
CDT, and -GT, and not on a single CDT value alone. Efforts should be
directed toward the preanalysis of CDT, a uniform definition of CDT,
and standardization of CDT analysis.
|
Acknowledgments
|
|---|
I thank Lloyd Allen Jones for stylistic emendations.
|
Footnotes
|
|---|
1 Nonstandard abbreviations: CDT, carbohydrate-deficient transferrin; CDG, carbohydrate-deficient glycoprotein; Tf, transferrin; pI, isoelectric point; IEF, isoelectric focusing; TIA, turbidimetric immunoassay; Allo A, lectin from Allomyrina dichotoma; TJA, lectin from Trichosanthes japonica; EIA, enzyme immunoassay; and -GT, -glutamyltransferase. 
|
References
|
|---|
-
Stibler H, Kjellin KG. Isoelectric focusing and electrophoresis of the CSF proteins in tremor of different origins. J Neurol Sci 1976;30:269-285.[ISI][Medline]
[Order article via Infotrieve]
-
de Jong G, van Dijk JP, van Eijk HG. The biology of transferrin. Clin Chim Acta 1990;190:1-46.[ISI][Medline]
[Order article via Infotrieve]
-
Hackler R, Arndt T, Kleine TO, Gressner AM. Effect of separation conditions on automated isoelectric focusing of carbohydrate-deficient transferrin and other human isotransferrins using the PhastSystem. Anal Biochem 1995;230:281-289.[ISI][Medline]
[Order article via Infotrieve]
-
van Noort WL, de Jong G, van Eijk HG. Purification of isotransferrins by concanavalin A Sepharose chromatography and preparative isoelectric focusing. Eur J Clin Chem Clin Biochem 1994;32:885-892.[ISI][Medline]
[Order article via Infotrieve]
-
van Eijk HG, van Noort WL, de Jong G, Koster JF. Human serum sialo transferrins in diseases. Clin Chim Acta 1987;165:141-145.[ISI][Medline]
[Order article via Infotrieve]
-
van Eijk HG, van Noort WL. The analysis of human serum transferrins with the PhastSystem: quantitation of microheterogeneity. Electrophoresis 1992;13:354-358.[ISI][Medline]
[Order article via Infotrieve]
-
Mårtensson O, Härlin A, Brandt R, Seppä K, Sillanaukee P. Transferrin isoform distribution: gender and alcohol consumption. Alcohol Clin Exp Res 1997;21:1710-1715.[ISI][Medline]
[Order article via Infotrieve]
-
Kamboh MI, Ferrell RE. Human transferrin polymorphism. Hum Hered 1987;37:65-81.[ISI][Medline]
[Order article via Infotrieve]
-
Stibler H. Carbohydrate-deficient transferrin in serum: a new marker of potentially harmful alcohol consumption reviewed. Clin Chem 1991;37:2029-2037.[Abstract/Free Full Text]
-
Bean P, Peter JB. Allelic D variants of transferrin in evaluation of alcohol abuse: differential diagnosis by isoelectric focusing-immunoblotting-laser densitometry. Clin Chem 1994;40:2078-2083.[Abstract]
-
Hackler R, Arndt T, Helwig-Rolig A, Kropf J, Steinmetz A, Schaefer JR. Investigation by isoelectric focusing of the initial carbohydrate-deficient transferrin (CDT) and non-CDT transferrin isoform fractionation step involved in determination of CDT by the ChronAlcoI.D. assay. Clin Chem 2000;46:483-492.[Abstract/Free Full Text]
-
Stibler H, Allgulander C, Borg S, Kjellin KG. Abnormal microheterogeneity of transferrin in serum and cerebrospinal fluid in alcoholism. Acta Med Scand 1978;204:49-56.[ISI][Medline]
[Order article via Infotrieve]
-
Stibler H, Borg S, Allgulander C. Clinical significance of abnormal heterogeneity of transferrin in relation to alcohol consumption. Acta Med Scand 1979;206:275-281.[ISI][Medline]
[Order article via Infotrieve]
-
Stibler H, Borg S, Joustra M. Micro anion exchange chromatography of carbohydrate-deficient transferrin in serum in relation to alcohol consumption (Swedish patent 8400587-5). Alcohol Clin Exp Res 1986;10:535-544.[ISI][Medline]
[Order article via Infotrieve]
-
Crivellente F, Fracasso G, Valentini R, Manetto G, Riviera AP, Tagliaro F. Improved method for carbohydrate-deficient transferrin determination in human serum by capillary zone electrophoresis. J Chromatogr B 2000;739:81-93.
-
Heggli DE, Aurebekk A, Granum B, Westby C, Løvli T, Sundrehagen E. Should tri-sialo-transferrins be included when calculating carbohydrate-deficient transferrin for diagnosing elevated alcohol intake?. Alcohol Alcohol 1996;31:381-384.[Abstract/Free Full Text]
-
Renner F, Stratmann K, Kanitz RD, Wetterling T. Determination of carbohydrate-deficient transferrin and total transferrin by HPLC: diagnostic evaluation. Clin Lab 1997;43:955-964.
-
Dibbelt L. Does trisialo-transferrin provide valuable information for the laboratory diagnosis of chronically increased alcohol consumption by determination of carbohydrate-deficient transferrin?. Clin Chem 2000;46:1203-1205.[Free Full Text]
-
Viitala K, Lähdesmäki K, Niemela O. Comparison of the Axis %CDT TIA and the CDTect method as laboratory tests of alcohol abuse. Clin Chem 1998;44:1209-1215.[Abstract/Free Full Text]
-
Stibler H, Borg S. Evidence of reduced sialic acid content in serum transferrin in male alcoholics. Alcohol Clin Exp Res 1981;5:545-549.[ISI][Medline]
[Order article via Infotrieve]
-
Stibler H, Borg S. Carbohydrate composition of serum transferrin in alcoholic patients. Alcohol Clin Exp Res 1986;10:61-64.[ISI][Medline]
[Order article via Infotrieve]
-
Landberg E, Påhlsson P, Lundblad A, Arnetorp A, Jeppsson JO. Carbohydrate composition of serum transferrin isoforms from patients with high alcohol consumption. Biochem Biophys Res Commun 1995;210:267-274.[ISI][Medline]
[Order article via Infotrieve]
-
Peter J, Unverzagt C, Engel WD, Renauer D, Seidel C, Hösel W. Identification of carbohydrate deficient transferrin forms by MALDI-TOF mass spectrometry and lectin ELISA. Biochim Biophys Acta 1998;1380:93-101.[Medline]
[Order article via Infotrieve]
-
Inoue T, Yamauchi M, Ohkawa K. Structural studies on sugar chains of carbohydrate-deficient transferrin from patients with alcoholic liver disease using lectin affinity electrophoresis. Electrophoresis 1999;20:452-457.[ISI][Medline]
[Order article via Infotrieve]
-
Henry H, Froehlich F, Perret R, Tissot JD, Eilers-Messerli B, Lavanchy D, et al. Microheterogeneity of serum glycoproteins in patients with chronic alcohol abuse compared with carbohydrate-deficient glycoprotein syndrome type I. Clin Chem 1999;45:1408-1413.[Abstract/Free Full Text]
-
Stibler H, Borg S. Glycoprotein glycosyltransferase activities in serum in alcohol-abusing patients and healthy controls. Scand J Clin Lab Invest 1991;51:43-51.[ISI][Medline]
[Order article via Infotrieve]
-
Helander A, Carlsson S. Carbohydrate-deficient transferrin and -glutamyl transferase levels during disulfiram therapy. Alcohol Clin Exp Res 1996;20:1202-1205.[ISI][Medline]
[Order article via Infotrieve]
-
Lakshman MR, Rao MN, Marmillot P. Alcohol and molecular regulation of protein glycosylation and function. Alcohol 1999;19:239-247.[ISI][Medline]
[Order article via Infotrieve]
-
Xin Y, Lasker JM,
Top
Abstract
Introduction
