Stability of vitamin B12 in the presence of

ascorbic acid1
Harold L. Newmark, M.S., Jacob Schemer, M.S., Martin Marcus, M.S.,
and Mukund Prabhudesai, M.D.

ABSTRACT Experiments were performed in two independent laboratories, each using their
own meal preparations which were exactly similar in composition to the meals described by Herbert
and Jacob (J. Am. Med. Assoc. 230: 241, 1974), in order to check their report that incubating meals
(portions of daily food intake by man) of “modest” or “high” vitamin B3 content with increasing
levels of added L-ascorbic acid (vitamin C) produced increasing destruction of vitamin B12. The
present studies were performed with standardized and official methods. Vitamin B12 was determined
microbiologically and by radioassay method. The results showed that 1) the vitamin B12 contents of
these meals were in general agreement with values calculated from the literature for the foods
involved, 2) the values obtained were manyfold higher than those reported by Herbert and Jacob,
and 3) there was no deleterious effect of added ascorbic acid on the vitamin B12 content of meals,
contrary to their published results. Am. J. C/in. Nutr. 29: 645-649, 1976.

Recently, Herbert and Jacob (1) reported Experimental methods

that increasing levels of ascorbic acid, added Microbiological assays
to homogenized meals prior to incubation at
Four sets each of the “modest” (Table I ) and “high”
37 C for 30 mm as a laboratory mimic of the (Table 2) vitamin B12 meals of the composition and
gastric environment in man, produced in- preparation previously described (1) were prepared in the
creasing destruction of vitamin B12. This was Product Development Laboratories, H offm an n- La
Roche Inc., and incubated at 37 C for 30 mm with 0, 0.1,
determined by radioassay against a series of
0.25, or 0.5 g L-ascorbic acid (vitamin C). Immediately
intrinsic factor concentrate standards. From
after this incubation the meals were frozen and kept
these results they inferred that “high doses of frozen until assayed. The diets were assayed either I)
ascorbic acid, popularly used as a home directly, 2) after extractions by the Association of
remedy against the common cold, destroy Official Analytical Chemists (AOAC) procedure (6), or
3) by the procedure published by the British Analytical
substantial amounts of vitamin B12 when
Methods Committee (7). The following quantities of
ingested with food.” If this is true, the timing AOAC extractant were added to 20 g of the homogenized
of supplementary ascorbic acid ingestion meals: “modest B12”, 25 ml; “high B12 meal”, 300 ml.
would have to be adjusted to a minimum The meals were then extracted for 10 mm at 15 psi steam
pressure (121 C), brought to appropriate volume, cen-
exposure to vitamin B12 in the digestive tract.
trifuged, the pH of aliquot was adjusted to 6.0 and di-
Review of the Herbert and Jacob paper
luted to assay range. For the British method, 4 ml of a I
revealed that the reported vitamin B12 assay mg/mI KCN solution and 16 ml water were added to 20 g
values of the meals that were used were of homogenized meal; the pH was adjusted to 4.6 to 5.0.
considerably lower for the foods involved After standing for 30 mm with occasional stirring, the
pH was readjusted to 4.6 to 5.0 as necessary. The
than were calculated values taken from previ-
containers were placed in a boiling water bath and were
ously published food composition tables left in the bath for 30 mm after the internal temperature
(2-5), which in turn led to concern about the reached 90 C. The meal samples were cooled and brought
analytical methodology employed in the to a volume of 100 ml. A portion of the samples was
centrifuged and an aliquot was adjusted to pH 6.0 and
investigation. In view of these observations
diluted to assay range. All assays for vitamin B13 were
which could have influenced the results of
run by the AOAC microbiological procedure (6), using
Herbert and Jacob, it was decided to reexam- Lactobacillus leichmannii.
ine the effect of supplementary ascorbic acid
“From the Product Development Department, Hoff-
on vitamin B12 in meals with standardized mann-La Roche Inc., Nutley, New Jersey 07110, and the
and official extraction procedures, using both Department of Pathology, Fordham Hospital, Miseri-
microbiological and radioassay methods. cordia-Fordham Affiliation, Bronx, New York 10458.

The American Journal of Clinical Nutrition 29: JUNE 1976, pp. 645-649. Printed in U.S.A. 645

Downloaded from https://academic.oup.com/ajcn/article-abstract/29/6/645/4648834

by guest
on 07 February 2018
 646 NEWMARK ET AL.

Radioassays For the “modest B12” content meal, with

For the radioassay, separate preparations similar to and without added ascorbic acid, microbio-
those used for the microbiological assays of both the logical assays indicate that the AOAC extrac-
“modest B12” content meal and a “high B12” content tion procedure gives somewhat higher results
meal were prepared in a Fordham Hospital laboratory.
than the British procedure, while the results
Proportional amounts of freshly prepared ascorbic acid
solution were added to 20 g portions of these homoge- without extraction are about ‘/2 to ‘/ of the
nized meals equivalent to 0. 1, 0.25, and 0.5 g per meal. total B12 (P < 0.01; Table 4). It is noteworthy
These preparations, as well as control meal aliquots that the various levels of added ascorbic acid
without ascorbic acid, were incubated for 30 mm at 37 C.
have significant effect neither on the free
After incubation, KCN was added to one set and the pH
vitamin B12 (no extraction after homogeniza-
was adjusted to 4.8. This set was placed in a boiling water
bath and timed for 30 mm after the temperature reached tion) content of the meal nor on total vitamin
90 C. The pH ofthe other set was adjusted to 4.8 and the B12 after extraction.
samples were kept at room temperature.
Both sets were subdiluted with deionized water so that TABLE 2
expected values would lie within the linear portion of the Composition of high vitamin B12 meal
standard curve. Four ml of a solution containing 0.6 g
of glutamic acid and 2 mg of KCN per 100 ml adjusted Component Amount
to pH 3.3 were added to 1.0-mI sample aliquots. Prelimi- Chicken noodle soup 177 ml
nary assays revealed the presence of high nonspecific
Crackers 6 g
binding which was eliminated by heating the sample to
Grilled beef liver 90 g
90 C for 2 1 hr. Cyanocobalamin standards that were simi-
Catsup 15 ml
larly treated were unaffected by this procedure. Aliquots
Whole kernel corn 90 g
of the final dilution were used for the saturation analysis Egg salad 45 g
which followed. With the exception of extended heating
Apple l75g
time the method was essentially that of Wide and Kil- Bread 25g
lander (8). Butter 5 g
Milk 237 ml
Results
The major sources of vitamin B12 in the TABLE 3
Calculated and assay values for modest vitamin
“modest” meal (Table 1) are cottage cheese B12 and high vitamin B12 meals
and milk, each of which supply about 1 g of
vitamin B12 to the meal. In the “high B12” Microbiological
assay extraction #{149} Herbert
meal (Table 2), the major source of the Meal
Cal- procedure Radio- and
culated assay Jacob
vitamin is liver, with milk adding a small
AOAC British (1)
fraction of the total content. A comparison of
calculated values and the various assay values sg/meal

of the two meals is given in Table 3. Clearly Modest B12 ca 2.0 2.7 2.0 1.6 0.37
High B12 ca73” 96 78 46 8.9
the results of the microbiological assays and
radioassays by standardized extraction proce- “Based on values in Home Economics Research
dures employed in this study are in rough Report 36, United States Department of Agriculture,
agreement with the values calculated from the 1969, Washington, D.C. “Assuming 90 g grilled liver
had B12 content similar to 90 g fresh liver.
literature. On the other hand, the results of
Herbert and Jacob (1) are only a small frac- TABLE 4
tion of the expected values. Effect of extraction procedure on apparent
vitamin B12 content of modest B12 meal

TABLE I
Ascorbic acid
Composition of modest vitamin B12 meal Microbiological assay
added to meal Herbert
extraction procedure
prior to and
Component Amount incubation Jacob (I)
None AOAC” British”
at 37 C
Cream of potato soup 177 ml
g g/meaI
Cottage cheese 100 g
Lettuce lOg 0 0.98 I 2.7 I 2.0 0.37
Canned peaches 360g 0.1 0.90 3.1 2.1 0.21”
Crackers 6g 0.25 0.92 3.0 2.2 O.07b
Bread 25g 0.5 0.99 3.5 2.0 0.02”
Butter 5g
Milk 237 ml “Average of duplicate assays. “Calculated from
data in paper.

Downloaded from https://academic.oup.com/ajcn/article-abstract/29/6/645/4648834

by guest
on 07 February 2018
 STABILITY OF VITAMIN B12 IN THE PRESENCE OF ASCORBIC ACID 647

For the “high B12” content meal with and Discussion

without added ascorbic acid, the results ob- A review of the literature indicated that the
tamed by microbiological assays follow a
results obtained with L. leichmannii for the
similar pattern (Table 5). The data indicate “high B12” and “modest B12” meals were in
once again that neither the free vitamin B12 agreement with assessments of the vitamin
levels nor the total B12 levels by AOAC B12 content as assayed by a variety of
extraction show significant changes after in- methods. The vitamin B12 of the “high B12”
cubation at 37 C for 30 mm with graded levels meal came almost entirely from liver. In this
of added ascorbic acid. The results obtained context it had been shown by Kamikubo and
by the British extraction and microbiological Oguni (3) and by Lichtenstein et al. (4) that
assay indicate that the results for all three for beef liver the results of L. leichmannii
levels of ascorbic acid addition do not differ assays were essentially similar to the results
significantly, considering the normal variabil- of Ochromonas malhamensis assays, the lat-
ity of this method of assay.
ter organism also being more specific for
As a further check on the validity of the
cyanocobalamin and hydroxycobalamin than
assay results, extracts of both the “high B12” for “pseudo” cobalamins (1 1). The vitamin
and “modest B12” diets prepared by the
B12 of the “modest B12” diet came from milk
AOAC procedure were hydrolyzed at pH and cottage cheese. In this case Gregory (5)
12 for 30 mm at 121 C. This procedure is
and Lichtenstein et al. (4) had shown that
known to destroy vitamin B12 (9, 10) and any
similar results were obtained for milk and
microbiological activity remaining after such cheese, respectively, by both of the above
treatment could be considered as microbio-
assay organisms.
logical growth factors for the assay organism When the vitamin B12 content of the meals
that are not cobalamins. Assays of these as determined by Herbert and Jacob (1) was
treated extracts indicated that at best only a
compared to the amount calculated from
small fraction (e.g., less than 20%) of the literature values of the meal components (2),
vitamin B12 activity as measured by L. leich- it was evident that the values reported by
mannii in the diets might be due to nonspe- these investigators were considerably lower.
cific stimulation of the test organism. There- Thus, the “modest” meal was one-fifth and
fore, the microbiological activity reported is the “high” meal one-eighth of the calculated
essentially vitamin , activity. values. These discrepancies in vitamin B12
The radioassays indicate that lower results content suggested the possibility of inade-
are obtained in the absence of initial heat and quate methodology.
cyanide treatment than when they are used (P It is now accepted that a common error in
< 0.2; Table 6). Hence by another analytical the early years of vitamin B12 research into
approach, neither set of results for the “mod-
est B12” meal or the “high B12” meal indi- TABLE 6
Radioassays on modest B12 and high B12
cates any loss of vitamin B12 activity in the
meals after incubation at 37 C
presence of the added levels of ascorbic acid.
Herbert
Ascorbic acid No added Added
and
TABLE 5 added cyanide cyanide
Jacob (I)
Effect of extraction procedure on apparent
vitamin B12 content of high B1, meal g pg/meal

Ascorbic acid Modest B12

Microbiological assay
added to meal Herbe 0 0.84 1.6 0.37
extraction procedure
prior to and
0.1 1.1 2.0 0.21#{176}
incubation Jacob (I)
None AOAC” British’ 0.25 1.4 1.7 0.07#{176}
at 37 C
0.5 0.9 1.9 0.02#{176}
pg/meal
g I
0 37 96 78 8.9 High B12
0.1 35 89 I 68 ca8.5b 0 32 46 8.9
0.25 34 I 83 I 61 6.7” 0.1 36 49 ca8.5#{176}
0.5 37 96 60 ca5.l” 0.25 31 62 6.7#{176}
0.5 31 47 ca5.I#{176}
Average
#{176} of duplicate assays. “Calculated from
data in paper. Calculated
#{176} from data in paper.

Downloaded from https://academic.oup.com/ajcn/article-abstract/29/6/645/4648834

by guest
on 07 February 2018
 648 NEWMARK ET AL.

the cobalamin content of food was inadequate that the iron content was the stabilizer of
extraction from the tightly-bound cobalamin vitamin B12 in liver extracts. Commercial
forms. Several extensive extraction proce- pharmaceutical liquid formulations have used
dures have become universally accepted for low levels of iron salts or complexes to
food assays for cobalamin content over the stabilize vitamin B12 against degradation by
last two decades. However, Herbert and high concentrations of ascorbic acid for long
Jacob (1) used a method which was developed periods of months or years. Levels of iron as
for assays of vitamin B12 in serum, where low as 17 ppm in the solution (14) or 12 ppm
mild extraction appeared to be adequate by (15) have been used successfully in such
the references cited. On the other hand, systems. The United States Food and Drug
consulting the now extensive literature reveals Administration has recognized this use of
that a more extended extraction procedure is iron as a vitamin B12 stabilizer in aqueous
required to release the more tightly-bound multivitamin solutions in a Food Additive
cobalamins present in many foods (7, 9, 12). Regulation (16). Many foods have iron levels
In addition, according to Skeggs (9), many in this range.
discrepancies in assay values determined by The vitamin B12 in foods, however, unlike
radioactive methods are “the result of failure that in pharmaceutical preparations, is usu-
to extract the total cobalamin present in the ally tightly bound. For example, the major
sample and convert it to cyanocobalamin part of the cobalamin present in liver is in the
against which it is to be compared.” The form of a coenzyme bound to a liver protein.
short extraction time apparently used by This binding serves to enhance further the
Herbert and Jacob (1) was considerably less stability of vitamin B12 activity, presumably
than that found necessary by Rosenblum (12) by reducing accessibility of vitamin B12 bind-
for total cobalamin in liver, the recommended ing sites to prolonged chemical attack.
extraction time being boiling for 1 hr at pH
4.0 in the presence of sodium nitrite and
potassium cyanide. Summary
In the present study the vitamin B12 values
Herbert and Jacob have reported vitamin
obtained by standardized extraction and
B12 values for two specific test meals that are
assay methods of both meals were in rough
only a small portion (one-fifth to one-eighth)
agreement with those calculated from litera-
of the levels calculated from literature values
ture values and manyfold higher than those
of the meal components. Their “low” results
reported by Herbert and Jacob. No signifi-
are apparently due to inadequate extraction
cant deleterious effect of supplementary as-
of the vitamin B12 in the samples during
corbic acid on the vitamin B12 content of the
preparation for assay.
meals was found, contrary to that indicated
The vitamin B12 values obtained by offi-
by Herbert and Jacob.
cially recognized methods of extraction and
It is interesting to note that the microbio-
assayed by microbiological or radioassay
logical assay medium for vitamin B,2 used by
were manyfold higher than those of Herbert
both the AOAC (6) and the United States
and Jacob and were in rough agreement with
Pharmacopeia (13) contains ascorbic acid
those calculated from literature values.
added at the level of 4 g/liter. This is about
When performed with standardized and
106 times the level of vitamin B12 at the lowest
official extraction methods and assayed either
level of standard used in the microbiological
microbiologically or by radioassay, there is
assay. The vitamin B12 and vitamin C are
no significant deleterious effect of added
autoclaved together in this media for 5 mm at
ascorbic acid on vitamin B12 stability in foods
121 C as part of the normal test procedure
when tested in meals and under conditions
without destruction
It has long been known
of vitamin B12.
that low levels of
reported by Herbert and Jacob. a
iron in solution in various ionic or complexed
The authors wish to thank Miss Ann Dowell and Mr.
forms will markedly enhance vitamin B12
Sami Wassef for helping in performing these assays and
stability in the presence of ascorbic acid. Drs. Myron Brin and Jack Bauernfeind for their helpful
Thus, Shenoy and Ramasarma (10) found suggestions.

Downloaded from https://academic.oup.com/ajcn/article-abstract/29/6/645/4648834

by guest
on 07 February 2018
 STABILITY OF VITAMIN B1, IN THE PRESENCE OF ASCORBIC ACID 649

References 8. WIDE, L. AND A. KILLANDER. A radiosorbent tech-

nique for the assay of serum vitamin B12 Scand. J.
I. HERBERT, V., AND E. JACOB. Destruction of vitamin Clin. Lab. Invest. 27: 151, 1971.
B1, by ascorbic acid. J. Am. Med. Assoc. 230: 241, 9. SKEGGS, H. R. “The Vitamins” (2nd ed.) New York:
1974. Academic Press, 1967, vol. 7, pp. 277-293.
2. “Pantothenic Acid, Vitamin B, and Vitamin B1, in 10. SHENOY, K. G. AND G. B. RAMASARMA. Iron as a
Foods” Home Economics Research Report No. 36, stabilizer of vitamin B1, activity in liver extracts and
Agricultural Research Service, U.S. Dept. of Agri- the nature of so-called alkali stable factor. Arch.
culture, Washington, D.C., 1969. Biochem. Biophys. 55: 293, 1955.
3. KAMIKUBO, T., AND Y. OGUNI. Microbiological 11. BRIGGS, G. M. AND F. S. DAFT. Water soluble
determination of vitamin B1, with Ochromonas vitamins, part I (Vitamin B13, folic Acid, choline and
malhamensis. J. Vitaminol. 5: 51, 1959. para-aminobenzoic acid). Ann. Rev. Biochem. 24:
4. LICHTENSTEIN, H., A. BELOIAN AND H. REYNOLDS. 339, 1955.
Comparative vitamin B1, assay of foods of animal 12. ROSENBLUM, C. Analytical applications of radioac-
origin by Lactobadillus leichmannii and Ochromonas tive vitamin B12. Talanta II: 255, 1964.
maihamensis. J. Agr. Food Chem. 7: 771, 1959. 13. The United States Pharmacopeia, 19th Rev. United
5. GREGORY, M. E. The microbiological assay of States Pharmacopeial Convention, Inc., Rockville,
‘vitamin B12’ in the milk of different animal species. Md. 1975, p. 613.
Brit. J. Nutr. 8: 340, 1954. 14. NEWMARK, H. L. U.S. Patent 2,823, 167-February
6. Official Methods of Analysis (12th ed). Washing- II, 1958, Stable vitamin B12-containing solutions.
ton, D.C.: Association of Official Analytical 15. FREEDMAN, L., M. BLITZ, D. B. SABINE AND E.
Chemists, 1975, p. 842. EIGEN. US. Patent 2,939,82 I-June 7, 1960, Stable
7. Analytical Methods Committee: The estimation of vitamin B1, solutions.
vitamin B12. Analyst 81: 132, 1956. 16. Food Additives. Federal Register 27: 883, 1962.

Downloaded from https://academic.oup.com/ajcn/article-abstract/29/6/645/4648834

by guest
on 07 February 2018