Clin. Lab. Haem.

2001, 23, 365±371

High serum cobalamin levels in the clinical
setting ± clinical associations
and holo-transcobalamin changes
R. CARMEL, Departments of Medicine and Pathology, New York Methodist Hospital, Brooklyn, NY, and
H. VASIREDDY, Department of Medicine, Weill Medical College of Cornell University, New York, NY, USA
I. AURANGZEB,
K. GEORGE

Summary Whereas low cobalamin levels have been studied intensively, systematic information
about high levels, especially in the clinical setting, is scarce. Therefore, a prospective
comparison was done of 60 patients with high cobalamin levels and 75 with normal
levels obtained by a hospital laboratory over a 2‰ month period. Associations with
clinical disorders and laboratory test results were examined. Transcobalamin (TC) I
and II were measured, especially the holoproteins (TC carrying circulating endogenous
cobalamin) which were fractionated with micro®ne silica powder. High cobalamin
levels (> 664 pmol/l; > 900 ng/l) occurred in 94 of 670 consecutive clinically
requested assays (14%). The only independently signi®cant associations with a high
cobalamin level were renal failure among the clinical disorders (P ˆ 0.01), elevated
serum creatinine (P ˆ 0.0001) and diminished albumin (P ˆ 0.0002) levels among
laboratory tests. Both holo-TC I and holo-TC II levels were increased in renal failure
(P ˆ 0.0001) but the increase was relatively greater in holo-TC II. The results indicate
that high cobalamin levels are more frequent than low ones in clinical practice and
appear to be associated often with renal failure. The elevation of both holo-TC II and
holo-TC I suggests that several mechanisms are operative. The accumulation of holo-
TC II suggests that cellular uptake of cobalamin by the abundant TC II receptors in the
kidney may be impaired. The much better known association of high cobalamin levels
with leucocytic disorders is rare, and no association was seen with liver disease.
Keywords Cobalamin, cobalamin levels, renal failure, transcobalamin

patients found to have high cobalamin levels by our

Introduction
Cobalamin assay is requested frequently and for a variety
of clinical situations. The focus is typically on detecting
de®ciency, and the literature on low serum cobalamin
levels and their causes and meaning is extensive (Carmel,
2000). In contrast, high cobalamin levels have not been
systematically examined. Little is known about the
prevalence of high levels in clinical practice and what
commonly causes them, which leaves the clinician with
little guidance about how to respond to such levels when
they are encountered. Therefore, we prospectively studied
Accepted for publication 2 October 2001
Correspondence: Dr R. Carmel, Department of Medicine, NY Meth-
odist Hospital, 506 Sixth St., Brooklyn, NY 11215, USA. Tel: 718
780 3253; Fax: 718 780 3259; E-mail: rac9001@nyp.org
hospital's clinical laboratory in assays requested by
clinicians. We also examined what disorders and laborat-
ory test abnormalities were associated with them.
In addition, we measured whether the cobalamin was
distributed differently between the two transcobalamins in
the blood than when cobalamin levels are normal.
Transcobalamin (TC) I is the cobalamin-binding protein
that usually carries most of the circulating cobalamin,
because TC I is not known to deliver cobalamin to cells
and has a long half-life in the blood stream (Carmel,
1981). TC II, on the other hand, rapidly delivers its
cobalamin to cells via speci®c receptors for TC II. As a
result, less than 25% of the cobalamin in peripheral blood
plasma usually exists as holo-TC II (TC II with attached
cobalamin) at any given moment, while the majority
circulates as holo-TC I (Hall, 1975, 1977; England et al., 365

Ó 2001 Blackwell Science Limited

366 High cobalamin levels
1976; MacDonald et al., 1977; Nexo & Andersen, 1977;
Carmel, 1985). The receptors for TC II are ubiquitous in
tissues and are especially rich in the kidney (Bose et al.,
1995a,b).
Methods
All results obtained from inpatient and outpatient spe-
cimens submitted by clinicians for cobalamin assay
(Immuno-1Ò, Bayer Diagnostics, Tarrytown, NY) to the
clinical laboratory in our 560-bed, University-af®liated
teaching hospital were monitored daily for a 2‰ month
period. Each serum with a cobalamin level > 664 pmol/l
(> 900 ng/l), which is above the normal reference range,
was set aside by the clinical laboratory. This totalled 94
specimens with abnormally elevated levels out of 670
consecutive cobalamin measurements during the study
period, a frequency of 14%. This rate was nearly twice
that of low results (< 185 pmol/l) during the same period
(51/670; 7.6%; P ˆ 0.0002).
The serum left over after cobalamin assay was frozen
and transferred to us for study. In several instances no
serum was left over; this circumstance and duplicate sera
on the same patient (which were tallied only once) were
the most common reasons for exclusion of speci®c samples
from further study. Clinical records were reviewed for
relevant information. Follow-up serum and EDTA-antico-
agulated plasma specimens were sought for con®rmation
of the elevated cobalamin level and other testing. Patients
for whom follow-up samples and clinical information were
not available were excluded from further study; most often
these were outpatients, inpatients providing easier access
for follow-up. We also excluded the few patients whose
follow-up cobalamin results were within the reference
range, suggesting an artifact or transient phenomenon.
The ®nal study population, thus, consisted of 60
patients with satisfactorily con®rmed cobalamin levels
> 664 pmol/l.
The control group consisted of 75 consecutive, adequate
specimens with cobalamin levels found by the clinical
laboratory to be within the reference range and for which
clinical records were available for review. The study
subjects and control subjects underwent identical assess-
ment. The following clinical information was collected:
demographic data, the chief complaint, all medical
illnesses and comorbid conditions, medications taken,
history of oral or parenteral cobalamin therapy or
supplementation and laboratory data obtained within a
week of the cobalamin assay (including blood count,
serum creatinine, total protein, albumin and alkaline
phosphatase). The medical illnesses were de®ned by their
diagnostic listing in the clinical record. In addition to the
recorded diagnoses, liver disease was diagnosed if two or
more liver-related blood test results were more than twice
the upper limit of normal and could not be attributed to
other causes upon chart review. Similarly, renal failure
was also diagnosed if serum creatinine was 3 mg/dl or
greater and was not attributable to prerenal azotemia. All
medical conditions present in at least 10 of the 135
subjects (cancer, diabetes mellitus, renal failure, liver
disease, hypertension and altered mental status, which
included psychiatric conditions, dementia and other
cognitive disorders) were subjected to statistical analysis;
many patients had more than one of the diseases. The
study was approved by the hospital's Institutional Review
Board.
The distribution of endogenous cobalamin between the
two transcobalamins was measured by a modi®cation of
the method of Jacob & Herbert (1975). A slurry of 150 mg
of micro®ne silica powder (QUSOÒ, PQ Corp., Valley Forge,
PA) per ml deionized water was used to adsorb out TC II;
100 ll of slurry was used per 500 ll of serum. The
supernatant fractions, which contain TC I-bound cobal-
amin (holo-TC I), were then subjected to cobalamin assay.
Holo-TC II was calculated by subtraction of the holo-TC
I value from the total cobalamin level measured simulta-
neously in untreated serum that had been diluted with
buffer to match the dilution in the silica-treated specimen.
Although the silica fractionation method has limitations
(Stabler et al., 1991), including its reliance on batch
separation and indirect identi®cation of TC I and TC II, it
was used here because the subject of the study was high
cobalamin levels [rather than the low levels for which the
method's reliability diminishes owing to the cobalamin
assay's imprecision at low concentrations (Stabler et al.,
1991)]. Moreover, the alternative methods for assaying
holotranscobalamin at the time of the study were
unwieldy [e.g. gel fractionation followed by cobalamin
assay of the multiple eluted fractions per serum specimen
(Carmel, 1985)] for testing the large numbers of speci-
mens in this study.
In addition to holotranscobalamin measurement,
apotranscobalamins (unsaturated transcobalamins not
carrying cobalamin) were fractionated by Sephadex
(Pharmacia, Piscataway, NJ) G-200 gel chromatography
of EDTA plasma labelled in vitro with 57Co-cyanocobal-
amin, as previously described (Carmel et al., 1977). The
apotranscobalamin results in the 60 patients with high
cobalamin levels were compared with the normal range
established in our laboratory (Carmel et al., 1977).
All variables were tested for normality before statistical
analyses. Some distributions that were skewed were
Ó 2001 Blackwell Science Ltd., Clin. Lab. Haem., 23, 365±371
 R. Carmel et al. 367

subjected to log transformation before performing tests
requiring an assumption of normality. Results are reported
as mean ‹ 1 SD or, if skewed, as median and central
5%±95% range. Standard statistical methods were applied,
using SAS 6.12 software (SAS Institute, Cary, NC) and all
P-values were two-sided. For comparing the relative
magnitude of univariate effects across different variables,
odds ratios based on continuous variables were computed
for units equal to the difference between medians of the
third and ®rst tertiles. Stepwise multivariate logistic
regression was also performed to identify variables that
were independently signi®cant predictors of the outcome
of interest; the signi®cance level for variable entry and
removal in the model was set at 0.05.
Results
Clinical characteristics
The 60 patients with high cobalamin levels (> 664 pmol/l)
had a mean level of 1026 ‹ 263 pmol/l. The 75 control
patients had a mean cobalamin level of 354 ‹ 118 pmol/l.
The two groups did not differ signi®cantly in age, sex
distribution or ethnic distribution.
Of the six disorders that were each present in more than
10 patients in the entire study population of 135, only
renal failure and diabetes mellitus were signi®cantly more
prevalent among the patients with high cobalamin levels
(Table 1). In addition, three of the four patients identi®ed
as having received cobalamin therapy had high cobalamin
levels, but the difference was not signi®cant because there
were too few such patients. Signi®cant associations were
not seen with liver disease or cancer, even after excluding
patients with kidney disease and diabetes from the
analysis. Multivariate logistic regression analysis showed
that renal failure, but not diabetes, was independently
associated with high cobalamin levels; there was consid-
erable overlap between renal failure and diabetes, 20 of
the patients having both. The odds ratio for a patient
with renal failure having a high cobalamin level was 3.0
(95% con®dence interval, 1.4±6.1; P ˆ 0.004).
Among laboratory test results, signi®cant differences
were found in serum creatinine, serum albumin, haemo-
globin and serum alkaline phosphatase levels between the
two cobalamin level groups (Table 1). White blood cell
counts did not differ signi®cantly. Multivariate logistic
regression analysis showed that only albumin and creat-
inine levels were independently associated with an
increased risk of high cobalamin levels. The odds ratios
for having a high cobalamin level were 13.5 (95%
con®dence interval 4.1±55.2; P ˆ 0.0001) for an abnor-
mal creatinine level > 3 mg/dl and 4.8 (95% con®dence
interval 2.2±11.53; P ˆ 0.0002) in the presence of a
diminished albumin level. Haemoglobin and alkaline
phosphatase levels were not independently associated
with a high cobalamin level.
Holotranscobalamin distributions
Serum volume was suf®cient for holotranscobalamin
fractionation in all 60 patients with high cobalamin levels
and in 61 of the 75 control patients. The percent of the

Table 1. Comparison of clinical ®ndings

in 60 patients with elevated serum Findings in patients with
cobalamin levels and 75 with normal
High Normal
levels
cobalamin levels cobalamin levels P

Prevalence of:
Renal failure 27/60 (45%) 16/75 (21%) 0.01
Diabetes mellitus 23/60 (38%) 16/75 (21%) 0.04
Other disorders * * NS
Mean levels of:
Creatinine (mg/dl) 4.0 ‹ 4.2 1.5 ‹ 1.3 0.0001
Albumin (g/dl) 3.0 ‹ 0.8 3.5 ‹ 0.5 0.02
Haemoglobin (g/dl) 9.9 ‹ 2.1 11.3 ‹ 2.5 0.001
Alkaline phosphatase (U/l) 181 ‹ 288 95 ‹ 46 0.01
Other tests     NS

NS, not statistically signi®cant. *, No signi®cant differences between the two cobalamin
level groups were found in prevalences of the following other medical conditions: 72
cases of hypertension , 33 of cognitive/psychiatric disorders, 23 of cancer or 17 of liver
disease in the total study population of 135.  , No signi®cant differences between the two
cobalamin level groups were found in levels of the following tests: white blood cell count,
mean corpuscular volume of red blood cells or serum total protein.
Ó 2001 Blackwell Science Ltd., Clin. Lab. Haem., 23, 365±371
368 High cobalamin levels

Table 2. Comparison of
Findings in patients with holotranscobalamin ®ndings in 60
patients with high cobalamin levels and
High Normal
61 with normal cobalamin levels. The
cobalamin levels cobalamin levels P
values are given as means ‹ standard
Holo-TC I (pmol/l) 967 ‹ 375 426 ‹ 159 0.0001 deviations and, in parentheses on the line
(930) (388) below, as medians
Holo-TC II (pmol/l) 423 ‹ 309 60 ‹ 47 < 0.0001
(314) (52)
% of serum cobalamin 69.6 ‹ 19.5 87.5 ‹ 10.0 0.0001
found on TC I (75) (89)

total serum cobalamin that was attached to TC I Table 3. Prevalences of abnormal holo-transcobalamin I and
correlated inversely with the serum cobalamin level holo-transcobalamin II ®ndings in patients with and without
(r ˆ )0.43; P ˆ 0.0001). Accordingly, the percent was various clinical conditions (some patients had more than one
disorder)
signi®cantly lower in the high-cobalamin level group than
in the group with normal cobalamin levels (Table 2). Abnormal holo-TC I Abnormal holo-TC II
The distribution of cobalamin between the transcobal-
Renal failure: Yes 27/40 (67.5%)* 26/40 (65.0%)à
amins was also calculated in absolute amounts. Holo-TC I
No 29/81 (35.8%)* 24/81 (29.6%)à
and holo-TC II levels were each signi®cantly increased in
patients with high cobalamin levels (Table 2). However, Diabetes: Yes 18/36 (50.0%) 24/36 (66.7%) 
No 38/85 (44.7%) 36/85 (42.4%) 
the holo-TC II increase was proportionately greater than
the holo-TC I increase, consistent with the signi®cant Liver disease: Yes 9/16 (56.3%) 9/16 (56.3%)
decrease in the percent cobalamin carried on TC I when No 47/105 (44.8%) 51/105 (48.6%)

cobalamin levels were high.
Abnormal holotranscobalamin levels were de®ned as
those above the central 5%±95% range, i.e. > 676 pmol/l
for holo-TC I and > 141 pmol/l for holo-TC II. The
numbers of abnormal and normal holo-TC I and holo-TC
II levels were then compared for each of the disease
categories (Table 3). A signi®cant difference in holo-TC I
and holo-TC II was found only in renal failure. The
presence of diabetes mellitus was also associated with
abnormal holo-TC II levels (P ˆ 0.02) but not holo-TC I.
Multivariate analysis could not resolve if the holo-TC II
abnormalities in renal failure and diabetes were inde-
pendent of each other.
Not surprisingly, both holo-TC I and holo-TC II levels
correlated with cobalamin levels (P ˆ 0.0001 for each).
Among the other laboratory tests, only serum creatinine
levels (P ˆ 0.007) correlated with holo-TC II. Holo-TC I
correlated with creatinine (r ˆ 0.35; P ˆ 0.0001) and
alkaline phosphatase levels (r ˆ 0.34; P ˆ 0.0001)
and inversely with albumin levels (r ˆ )0.40; P ˆ
0.0001) and blood haemoglobin levels (r ˆ )0.22;
P ˆ 0.02). Multivariate logistic regression analysis showed
that albumin (P ˆ 0.0001), creatinine (P ˆ 0.0002) and
alkaline phosphatase (P ˆ 0.03) were independent predic-
tors of holo-TC I levels.
The clinical data from the seven patients whose holo-
TC II levels were so high that they actually exceeded the
holo-TC I levels were reviewed individually, but no
Cancer: Yes 9/17 (52.9%) 8/17 (47.1%)
No 47/103 (45.6%) 51/103 (49.5%)
Hypertension Yes 29/66 (43.9%) 35/66 (53.0%)
No 27/55 (49.1%) 25/55 (45.5%)
Altered Yes 10/30 (33.3%) 13/30 (43.3%)
mental status No 46/91 (50.5%) 47/91 (51.6%)
None of the comparisons between presence and absence of
disrder showed signi®cant differences in prevalence of holo-
transcobalamin abnormalities except those so marked (in boldface
type). *, P ˆ 0.001;  , P ˆ 0.02; à, P ˆ 0.01.
de®nite clinical associations were discerned in these
extreme cases.
Apotranscobalamin levels
The plasma unsaturated cobalamin-binding capacity,
which includes apo-TC I, apo-TC II and minor binding pro-
teins (Carmel et al., 1977), was elevated (> 1420 pmol/l) in
all but six of the patients with high cobalamin levels. Its near
universality, thus, provided no further insight into the high
cobalamin levels or the holotranscobalamin abnormalities.
Fractionation showed seven extremely elevated apo-TC I
levels of > 2000 pmol/l, which resembled the charac-
teristic pattern described in leukaemic, leukaemoid and
neoplastic disorders. However, only one of these accom-
panied a leukaemoid reaction (WBC 75,600/ll in a patient
Ó 2001 Blackwell Science Ltd., Clin. Lab. Haem., 23, 365±371

who died with sepsis), and only one occurred in a patient
with cancer (resection of the colon cancer showed regional
metastases, but no distant metastases were identi®ed). None
of the chromatographic elution patterns indicated the
presence of autoantibody to TC I or TC II.
Discussion
Abnormal increases of serum cobalamin levels have been
described in various conditions (Table 4). The most
prosaic of these is a recent injection of cobalamin and
the most dramatic and best studied ones include certain
leukaemias and myeloproliferative disorders in which TC I
elevation is paramount.
Our data show that high cobalamin levels occur almost
twice as often as low ones (14% vs. 7.6%, P ˆ 0.002).
However, the best known causes such as those just
mentioned appear to be uncommon, even rare, in the
usual clinical setting where cobalamin assay is requested
by physicians. Instead, the most common cause of
abnormally elevated cobalamin levels appears to be renal
Table 4. Known causes and presumed mechanisms of elevated
serum cobalamin levels
a Recent parenteral injection of cobalamin
b Abnormalities of transcobalamin (TC)
1. Elevated TC I levels
Leukaemias and myeloproliferative disorders
Chronic myelogenous leukaemia, acute
promyelocytic leukaemia, eosinophilic leukaemia
Some cases of subacute and acute leukaemia
Some cases of polycythaemia vera and myelo®brosis
Hypereosinophilic syndrome
Leukaemoid reactions
Cancer
Some cases of metastatic cancer, especially
involving liver
Some cases of hepatocellular carcinoma
2. Circulating anti-TC antibody
Antibody to TC II
Some cases with antibody to TC I
c Unknown mechanisms (increased release of cellular
cobalamin?, decreased clearance of cobalamin from
plasma?, etc.)
1. Renal failure
2. Liver disease
Acute hepatitis
Some cases of cirrhosis
Others?
3. Miscellaneous
Kwashiorkor (? due to liver disease)
Cystic ®brosis (? due to liver disease)
Ó 2001 Blackwell Science Ltd., Clin. Lab. Haem., 23, 365±371
R. Carmel et al. 369
failure. Although the association between renal failure
and high cobalamin levels is well established in the
literature (Matthews & Beckett, 1962; Nyberg et al., 1964;
Sevitt & Hoffbrand, 1969; Milman, 1980), it is not a
generally well known one. Few internists, or even nephro-
logists and haematologists, are aware of it (R. Carmel,
pers. observ.). The high frequency of renal failure in our
survey could not be attributed easily to selection factors.
Nephrologists at our hospital do not routinely request
cobalamin measurements in patients with renal failure or
in dialysis programs, nor is cobalamin supplementation
given to such patients routinely (K. Neelakantappa, pers.
comm.). Even if the prominence of renal failure as a cause
of high cobalamin levels in our study had been in¯uenced
by a higher rate of test ordering, it nevertheless re¯ects the
relative frequencies seen in clinical practice, which was
the purpose of our survey.
Diabetes mellitus was also signi®cantly associated with
high cobalamin levels in our study, but multivariate
analysis suggests that the association re¯ects the renal
failure common in diabetes. It is uncertain whether renal
failure with diabetes elevates cobalamin levels more often
than renal failure due to other causes or does so in a
different fashion.
Studies in animals have suggested that the kidney plays
a large role in regulating cobalamin ¯ux (Scott et al.,
1984; Lindemans et al., 1986; Seetharam, 1999),
although studies in humans are limited and it is not clear
how well animal data correspond to the situation in
humans. The kidneys are major organ repositories of
tissue cobalamin. A study of 11 tissues and ¯uids in
humans found the kidney to have the third highest
cobalamin concentration (Matthews, 1979). In the dog,
the kidney may also be a source of TC II (Rappazzo & Hall,
1972). The kidney is also one of the richest sources of
cellular receptors for TC II (Bose et al., 1995a,b). In
addition to having the TC II receptors common to all
tissues, the renal tubule is rich in megalin (Moestrup et al.,
1996), whose physiologic role is uncertain (Seetharam,
1999), and in the intrinsic factor receptor, cubilin
(Kozyraki et al., 1998).
The mechanisms responsible for high cobalamin levels
in renal failure are still unknown. The new demonstration
here of an abnormal distribution of holotranscobalamins
offers partial support for a failure of renal uptake of TC
II-bound cobalamin. Although free cobalamin is cleared by
glomerular ®ltration, much like inulin, and is apparently
neither secreted nor reabsorbed in the tubule (Nelp et al.,
1964), cobalamin rarely exists in the free state in the
circulation. Indeed, prevention of renal loss is presumed to
be an important function of the transcobalamins.
370 High cobalamin levels
In view of the major role of TC II-mediated uptake of
cobalamin in the kidney, it was of interest that holo-TC II
was proportionately even more elevated than holo-TC I in
renal failure in our study. Normally, little or no holo-TC II
is present in the circulation because holo-TC II delivers its
cobalamin to cells and is rapidly cleared (Hall, 1975,
1977; England et al., 1976; MacDonald et al., 1977;
Nexo & Andersen, 1977; Carmel, 1985). Although we
cannot exclude the possibility that TC II is functionally
altered by renal failure, failure of the diseased kidney to
clear holo-TC II may play a major role in the high
circulating cobalamin levels. However, holo-TC I levels are
also elevated in renal failure, which suggests a complex
interplay of events and mechanisms.
Despite being by far the most common association, renal
failure explained only about one-half of the high cobal-
amin levels in the study. The operative causes in most of
the other cases are unknown. The signi®cant association
with diminished albumin levels may be an important clue
but awaits clari®cation.
Just as noteworthy as the frequent association with
renal failure was the surprising infrequency of the better
known causes of high cobalamin levels such as hyper-
leucocytic disorders. Cobalamin levels did not correlate
with leucocyte counts. Although apotranscobalamin lev-
els were elevated in 90% of the patients with high
cobalamin levels, very few cases displayed the strikingly
high apo-TC I levels seen in some leucocytic and
myeloproliferative disorders and only one of those patients
had a very high leucocyte count. Similarly noteworthy
was the lack of an association with liver disease, whose
frequency was no higher than in control patients with
normal cobalamin levels. This may be explained by the
variability of cobalamin patterns in different liver diseases
(Jones et al., 1957; Cowling & Mackay, 1959; Stevenson &
Beard, 1959; Nelson & Doctor, 1960; Holdsworth et al.,
1964; Rachmilewitz & Eliakim, 1968; Adams et al.,
1971). Consistently high cobalamin levels tend to occur
only in some hepatic disorders, such as those involving
hepatocellular necrosis, and cirrhosis tends to have mild
or no effects (Nelson & Doctor, 1960; Holdsworth et al.,
1964). Finally, although black people have higher cobal-
amin levels than white people (Carmel, 1999), the
frequency of abnormally high cobalamin levels showed
no ethnic/racial predilections in our study.
Cobalamin therapy or supplementation also was found
infrequently. The ®nding that only three high cobalamin
levels could be explained by recent cobalamin injections
may have underestimated the frequency of cobalamin
injections slightly as a cause of high cobalamin levels. The
study design, in order to assure that all elevations in
cobalamin levels were real, excluded several patients
whose high cobalamin levels could not be con®rmed in
follow-up testing and the cobalamin level rise after
injection is transient.
The ®ndings presented here and a review of the
literature suggest several recommendations to clinicians
faced with a high cobalamin level. A serum creatinine level
and a history of recent cobalamin treatment are simple
things to examine and may resolve half of the cases. If
these are unrevealing, a more complex search will be
needed, when clinically indicated, to identify rarer causes,
such as cancer; but an answer may not always be found.
Clinicians should take one further consideration into
account, namely the original reason they tested the
patient's cobalamin level. Most often, the cobalamin assay
has been requested in order to rule out cobalamin
de®ciency. The latter can still exist occasionally despite a
high cobalamin level. Competing in¯uences on serum
cobalamin levels, such as the TC I changes in chronic
myelogenous leukaemia, may prevent cobalamin levels
from becoming low even in the face of coexisting
pernicious anaemia (Britt & Rose, 1966; Corcino et al.,
1971). A similar coexistence of cobalamin de®ciency
conceivably could be masked in a patient with renal
failure or other cause of high cobalamin levels. Metabolic
testing may be necessary in such cases, but may be
complicated because homocysteine and methylmalonic
acid levels are often falsely elevated in renal insuf®ciency.
Acknowledgement
This study was supported in part by grant DK32640 from
the National Institutes of Health. We are indebted to
Dajun Qian of the University of Southern California School
of Medicine for help with statistical analysis, Dr Bela
Bhuskute for help in collecting information on some of
the patients, and Sarika Singh and Ming Deng-Yun for
technical assistance.
References
Adams J.F., Rowan T., Kennedy E.H. & McEwan F.C. (1971)
Clinical value of serum vitamin B12 estimation in jaundice.
Digestion 4, 289±294.
Bose S., Seetharam S., Hammond T.G. & Seetharam B. (1995a)
Regulation of expression of transcobalamin II-receptor in the
rat. Biochemistry Journal 310, 923±929.
Bose S., Seetharam S. & Seetharam B. (1995b) Membrane
expression and interactions of human transcobalamin II
receptor. Journal of Biological Chemistry 270, 8152±8157.
Ó 2001 Blackwell Science Ltd., Clin. Lab. Haem., 23, 365±371

Britt R.P. & Rose D.P. (1966) Pernicious anemia with a normal
serum vitamin B12 level in a case of chronic granulocytic
leukemia. Archives of Internal Medicine 117, 32±33.
Carmel R. (1981) Cobalamin-binding proteins in man. In: Con-
temporary Hematology-Oncology (eds R. Silber, A. S. Gordon,
J. LoBue & F. M. Muggia), Vol. 2. pp. 79±129. Plenum, New
York.
Carmel R. (1985) The distribution of endogenous cobalamin
among cobalamin-binding proteins in the blood in normal and
abnormal states. American Journal of Clinical Nutrition 41,
713±719.
Carmel R. (1999) Ethnic and racial factors in cobalamin meta-
bolism and its disorders. Seminars in Hematology 36, 88±100.
Carmel R. (2000) Current concepts in cobalamin de®ciency.
Annual Review of Medicine 51, 357±375.
Carmel R., Tatsis B. & Baril L. (1977) Circulating antibody to
transcobalamin II causing retention of vitamin B12 in the
blood. Blood 49, 987±1000.
Corcino J.J., Zalusky R., Greenberg M. & Herbert V. (1971)
Coexistence of pernicious anaemia and chronic myeloid leuk-
aemia: an experiment of nature involving vitamin B12 meta-
bolism. British Journal of Haematology 20, 511±520.
Cowling D.C. & Mackay I.R. (1959) Serum vitamin B12 levels in
liver disease. Medical Journal of Australia 2, 558±562.
England J.M., Down M.C., Wise I.J. & Linnell J.C. (1976) The
transport of endogenous vitamin B12 in normal human serum.
Clinical Science and Molecular Medicine 51, 47±52.
Hall C.A. (1975) Transcobalamins I and II as natural transport
proteins of vitamin B12. Journal of Clinical Investigation 56,
1125±1131.
Hall C.A. (1977) The carriers of native vitamin B12 in nor-
mal human serum. Clinical Science and Molecular Medicine 53,
453±457.
Holdsworth C.D., Atkinson M., Dossett J.A. & Hall R. (1964) An
assessment of the diagnostic and prognostic value of serum
vitamin B12 levels in liver disease. Gut 5, 601±606.
Jacob E. & Herbert V. (1975) Measurement of unsaturated
`granulocyte-related' (TC I and TC III) and `liver-related' (TC II)
B12 binders by instant batch separation using a micro®ne
precipitate of silica (QUSO G32). Journal of Laboratory and
Clinical Medicine 86, 505±512.
Jones P.N., Mills E.H. & Capps R.B. (1957) The effect of liver
disease on serum vitamin B12 concentrations. Journal of
Laboratory and Clinical Medicine 49, 910±922.
Kozyraki R., Kristiansen M., Silahtaroglu A., Hansen C., Jacobsen
C., Tommerup N., Verroust P.J. & Moestrup S.K. (1998) The
human intrinsic factor-vitamin B12 receptor, cubilin: molecular
characterization and chromosomal mapping of the gene to 10p
within the autosomal recessive megaloblastic anemia (MGA1)
region. Blood 91, 3593±3600.
Lindemans J., van Kapel J. & Abels J. (1986) Uptake of transco-
balamin II-bound cobalamin by isolated rat kidney tubule cells.
Scandinavian Journal of Clinical and Laboratory Investigations 46,
223±232.
Ó 2001 Blackwell Science Ltd., Clin. Lab. Haem., 23, 365±371
R. Carmel et al. 371
MacDonald C.M.L.A., Farquharson J., Bessent R.G. & Adams J.F.
(1977) The forms of vitamin B12 on the transcobalamins.
Clinical Science Molecular Medicine 52, 215±218.
Matthews D.M. (1979) Distribution of cobalamins in the animal
body. In: Vitamin B12. Proceedings of the Third European
Symposium on Vitamin B12 and Intrinsic Factor (eds B. Zagalak
& W. Friedrich), pp. 681±694. Walter de Gruyter, Berlin.
Matthews D.M. & Beckett A.G. (1962) Serum vitamin B12 in
renal failure. Journal of Clinical Pathology 15, 456±458.
Milman N. (1980) Serum vitamin B12 and erythrocyte folate in
chronic uraemia and after renal transplantation. Scandinavian
Journal of Haematology 25, 151±157.
Moestrup S.K., Birn H., Fischer P.B., Petersen C.M., Verroust
P.J., Sim R.B., Christensen E.I. & Nexo E. (1996) Megalin-
mediated endocytosis of transcobalamin-vitamin B12 com-
plexes suggests a role of the receptor in vitamin B12 home-
ostasis. Proceedings of the National Academy of Sciences of the
USA 93, 8612±8617.
Nelp W.N., Wagner H.N. Jr & Reba R.C. (1964) Renal excretion
of vitamin B12 and its use in measurement of glomerular
®ltration rate in man. Journal of Laboratory Clinical Medicine 63,
480±491.
Nelson R.S. & Doctor V.M. (1960) Hepatic and serum vita-
min B12 content in liver disease. Gastroenterology 38,
188±193.
Nexo E. & Andersen J. (1977) Unsaturated and cobalamin sat-
urated transcobalamin I and II in normal human plasma.
Scandinavian Journal of Clinical and Laboratory Investigation 37,
723±728.
Nyberg W., KuhlbaÈck B., Pasternack A. & Tallgren A.G. (1964)
Renal failure and serum vitamin B12 concentration. Acta
Medical Scandinavica 412 (Suppl.), 73±77.
Rachmilewitz M. & Eliakim M. (1968) Serum vitamin B12 ± a
diagnostic test in liver disease. Israel Journal of Medical Sciences
4, 47±54.
Rappazzo M.E. & Hall C.A. (1972) Transport function of
transcobalamin II. Journal of Clinical and Laboratory Investi-
gation 51, 1915±1918.
Scott J.S.D., Treston A.M., Bowman E.P.W., Owens J.A. &
Cooksley W.G.E. (1984) The regulatory roles of liver and
kidney in cobalamin (vitamin B12) metabolism in the rat: the
uptake and intracellular binding of cobalamin and the activity
of the cobalamin-dependent enzymes in response to varying
cobalamin supply. Clinical Science 67, 299±306.
Seetharam B. (1999) Receptor-mediated endocytosis of cobal-
amin (vitamin B12). Annual Review of Nutrition 19, 173±195.
Sevitt L.H. & Hoffbrand A.V. (1969) Serum folate and vitamin
B12 levels in acute and chronic renal disease. Effect of perito-
neal dialysis. British Medical Journal 2, 18±21.
Stabler S.P., Allen R.H., Savage D. & Lindenbaum J. (1991)
Response (letter). Blood 77, 1853±1854.
Stevenson T.D. & Beard M.F. (1959) Serum vitamin B12 content
in liver disease. New England Journal of Medicine 260,
206±210.