International Journal of Pharmaceutics 337 (2007) 161168

Stability of cyanocobalamin in sugar-coated tablets

Shinji Ohmori a, , Masumi Kataoka a , Hiroyoshi Koyama b
a

Health Science Laboratories, Consumer Healthcare Company, Takeda Pharmaceutical Company Limited, 17-85,
Jusohonmachi 2-chome, Yodogawa-ku, Osaka 532-8686, Japan
b Pharmaceutical Technology Research and Development Laboratories, Pharmaceutical Production Division,
Takeda Pharmaceutical Company Limited, 17-85, Jusohonmachi 2-chome, Yodogawa-ku, Osaka 532-8686, Japan
Received 31 July 2006; received in revised form 24 November 2006; accepted 28 December 2006
Available online 9 January 2007

Abstract
The purpose of this study was to clarify the stability of cyanocobalamin (VB12 -CN) in sugar-coated tablets containing fursultiamine hydrochloride
(TTFD-HCl), riboflavin (VB2 ), and pyridoxine hydrochloride (VB6 ), and to identify the factors affecting the stability of VB12 -CN in these sugarcoated tablets. The stability of VB12 -CN was investigated using high-performance liquid chromatography while decomposition was evaluated
kinetically. The decomposition of VB12 -CN in sugar-coated tablets with high equilibrium relative humidity (more than 60%) under closed conditions
showed complex kinetics and followed an AvramiErofeev equation, which expresses a random nucleation (two-dimensional growth of nuclei)
model. We showed that equilibrium relative humidity, the incorporation of VB2 and VB6 , and sugar coating, are the main factors influencing
decomposition and that these factors cause the complex decomposition kinetics.
2007 Elsevier B.V. All rights reserved.
Keywords: Cyanocobalamin; Stability; Sugar-coated tablets; Equilibrium relative humidity; Riboflavin; Pyridoxine hydrochloride

1. Introduction
Vitamin B combinations are a well-recognized treatment of
nerve pain, muscle pain, tired eyes, stiff shoulders, and lumbar pain, and usually consist of a combination of thiamine
(VB1 ), riboflavin (VB2 ), pyridoxine hydrochloride (VB6 ),
and cyanocobalamin (VB12 -CN). Fursultiamine hydrochloride
(TTFD-HCl) is a well-known thiamine derivative with a better bioavailability than thiamine in humans, and is therefore
often used in preference to thiamine in such combinations.
Several dosage forms of such vitamin B combinations are available, such as bottled nutritive drinks, capsules, and tablets. It is
often necessary to take vitamin B combinations for long periods and, as patients usually prefer to take sugar-coated tablets,
these are often used to increase compliance. Sugar-coated tablets
have several advantages over other formulations: masking of the
unpleasant odor and taste of vitamins, elegant appearance, and
ease of swallowing.
When developing sugar-coated tablets containing a vitamin
B combination, it is necessary to consider the stability of the

Corresponding author. Tel.: +81 6 6300 6896; fax: +81 6 6300 6788.
E-mail address: Ohmori [email protected] (S. Ohmori).

0378-5173/$ see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijpharm.2006.12.044

vitamins, in particular the stability of VB12 -CN. There have

been a number of studies on the stability of VB12 -CN, which is
known to be incompatible with other vitamins such as ascorbic acid, niacinamide, VB1 , and VB6 (Harada et al., 1969;
Terao et al., 1980; DeRitter, 1982; Toyama Pharmaceutical
Research Association, 1992; Ichikawa et al., 2005; The Japanese
Pharmacopoeia Fourteenth Edition, 2001). Several different
analytical methods for the measurement of VB12 -CN content
have been described in the literature. While both a colorimetric method and a bioassay method have been used (Harada
et al., 1969; Terao et al., 1980; The Japanese Pharmacopoeia
Fourteenth Edition, 2001), neither method is capable of discriminating between VB12 -CN and its decomposition products.
Therefore, a high-performance liquid chromatography (HPLC)
method which is able to discriminate between these compounds
has been developed. This method has been widely used for
measurements of VB12 -CN content (Toyama Pharmaceutical
Research Association, 1992; Ichikawa et al., 2005). From the
viewpoint of separation analysis, the HPLC method is now
widely recognized as being more accurate and reliable than other
analytical methods. There are, however, few studies on the stability of VB12 -CN or on the decomposition kinetics of VB12 -CN in
solid dosage forms which have used the HPLC method, in spite
of the fact that VB12 -CN is a well-known vitamin and widely

162

S. Ohmori et al. / International Journal of Pharmaceutics 337 (2007) 161168

used in solid dosage forms in the pharmaceutical field. In particular, there are few stability studies of VB12 -CN in sugar-coated
tablets using the HPLC method. In the present study, we investigated the stability and decomposition kinetics of VB12 -CN in
sugar-coated tablets containing TTFD-HCl, VB2 , and VB6 using
the HPLC method, and compared it with the stability of VB12 CN in powders, granules, and plain tablets. In order to clarify
the factors affecting the stability of VB12 -CN in sugar-coated
tablets, we evaluated the stability of VB12 -CN in sugar-coated
tablets under various conditions of equilibrium relative humidity
(ERH), in sugar-coated tablets lacking one of the other vitamins,
and in halved sugar-coated tablets.
2. Materials and methods
2.1. Materials
The following vitamins were used in the study: cyanocobalamin (VB12 -CN, Aventis Pharma S.A., France), fursultiamine
hydrochloride (TTFD-HCl, Takeda Pharmaceutical Company
Limited, Japan), riboflavin (VB2 , Takeda Pharmaceutical Company Limited, Japan), and pyridoxine hydrochloride (VB6 ,
Takeda Pharmaceutical Company Limited, Japan). Lactose
(Meggle GmbH, Germany) was used as a diluent. Corn starch
(Nihon Cornstarch Co., Japan) was used as both a diluent
and a disintegrant. Hydroxypropylcellulose (HPC-L, Nippon
Soda Co., Ltd., Japan) was used as a binder. Magnesium
stearate (Sakai Chemical Industrial Co., Ltd., Japan) was used
as a lubricant. Sucrose (Ensuiko Sugar Refining Co., Japan),
talc (Matsumura Sangyo Co., Japan), titanium dioxide (Ishihara Sangyo Kaisha Ltd., Japan), and powdered acacia (San-ei
Yakuhin Boueki Co., Japan) were used for sugar coating. The
above vitamins and excipients were in compliance with The
Japanese Pharmacopoeia Fourteenth Edition (2001). All other
chemicals were of reagent grade.
2.2. Manufacture of granules and plain tablets
TTFD-HCl, VB2 , VB6 , lactose, and corn starch were granulated with 0.12% VB12 -CN aqueous solution and 6% HPC-L
aqueous solution using a fluidized-bed granulator (FD-3S,
Powrex Co., Japan). Loaded weight in the fluidized-bed granulator was 1092 g, inlet air temperature 70 C, spray air pressure
0.1 MPa, and spray feed rate 810 g/min. After granulation,
the granules were milled using a screen mill (Power Mill,
Showa Kagaku Kikai Kosakusho Co., Ltd., Japan) with 1.5 mm
screen size. The milled granules were mixed manually with
corn starch and magnesium stearate in a plastic bag. Lubricated
granules were compressed using a rotary compression machine
(Correct 19 K, Kikusui Seisakusho Ltd., Japan) with 8-mm
diameter convex punches (radius 6.5 mm). Tablet weight was
170 mg, compression pressure 68 kN/punch, and rotation speed
20 rpm.
Each plain tablet contained 0.01 mg VB12 -CN, 54.58 mg
TTFD-HCl, 4.95 mg VB2 , 5.0 mg VB6 , 68.46 mg lactose,
31.5 mg corn starch, 5.0 mg HPC-L and 0.5 mg magnesium
stearate. When tablets lacking one of the vitamins were man-

ufactured, the quantity of lactose was adjusted to maintain a

constant tablet weight of 170 mg as a plain tablet.
2.3. Sugar coating
Sugar coating was performed manually on 5500 plain tablets
in a 12-in. onion pan (Kikusui Seisakusyo Ltd., Japan) using a
dusting method (Ohmori et al., 2004a,b). The dusting method
can be divided into four steps: (1) subcoating, (2) smoothing,
(3) colouring, and (4) polishing.
(1) The subcoating is applied to round the edges and build up the
tablet size. The subcoating step consists of alternately applying a sugar-coating suspension to the tablets followed by
dusting with powders and drying. Firstly, the sugar-coating
suspension is added to the core tablets. Secondly, the tablets
are stirred by hand to distribute the suspension. Thirdly, the
dusting powder is applied until all the tablets are coated
and tumble freely. Finally, the tablets are dried by hot air
at 60 C and the excess dusting powder is removed through
the exhaust. The sugar-coating suspension formulation is
45.0% (w/w) sucrose, 25.4% (w/w) talc, 2.8% (w/w) titanium dioxide, 4.3% (w/w) powdered acacia, and 22.5%
(w/w) purified water. The dusting powder formulation is
98% (w/w) talc and 2% (w/w) powdered acacia. The weight
of a final subcoated tablet was 252.3 mg.
(2) The smoothing step is to smoothe out the tablet surface prior
to application of the colouring, and consists of alternately
applying the sugar-coating suspension to the tablets and drying. The drying temperature was 60 C. The weight of a final
smoothed tablet was 294.5 mg.
(3) In the colouring step, the desired colour is imparted to
the tablets by alternately applying a colouring syrup and
drying. The colouring syrup formulation is 66.55% (w/w)
sucrose, 0.12% (w/w) VB2 , and 33.33% (w/w) purified
water. The drying temperature is initially 55 C, and is gradually reduced to 25 C. The weight of a final coloured tablet
was 320.0 mg.
(4) The polishing step is to achieve a final gloss, and involves
application of a mixture of waxes (beeswax and carnauba
wax) to the tablets in a polishing pan.
2.4. Moisture content
2.4.1. LOD of powders and granules
Moisture contents of VB12 -CN powders and granules containing VB12 -CN, TTFD-HCl, VB2 , and VB6 were measured by
loss on drying (LOD). The samples (1 g) were dried at 60 C for
5 h under the vacuum condition (0.53 0.13 kPa) using equipment (TD-40G, Toyama Sangyo Co., Ltd., Japan).
2.4.2. ERH of tablets
The ERH is water activity expressed as a percentage
(Heidemann and Jarosz, 1991). This was used as a measure of
the moisture content of the tablets and was measured using a
water activity analyzer (Hygroskop DT, Rotronic, Switzerland).
About 1 g of crushed tablets was used for measurement.

S. Ohmori et al. / International Journal of Pharmaceutics 337 (2007) 161168

2.5. Stability test of VB12 -CN

2.5.1. Storage conditions
2.5.1.1. Open conditions. The powders or granules were put
into uncapped glass bottles which were stored in desiccators
containing various saturated solutions at 40 C. Humidity conditions depended on the saturated solution used. In other words,
under open conditions, the powders or granules in the glass bottle
were exposed to both moisture and heat.
2.5.1.2. Closed conditions. The tablets were put into a glass
bottle which was then sealed with a metal cap. The closed bottle
was stored in the oven at 40 or 60 C. In other words, under
closed conditions, the tablets in the glass bottle were exposed to
only heat.
In the storage stability experiments, VB12 -CN powders, and
granules (after milling) containing VB12 -CN, TTFD-HCl, VB2 ,
and VB6 were stored in open glass bottles placed within desiccators at 40 C/6%RH, 40 C/57%RH, and 40 C/75%RH
(open conditions). Plain tablets containing VB12 -CN, TTFDHCl, VB2 , and VB6 were moisturized in desiccators using
various saturated solutions at 25 C (open conditions) and then
stored in glass bottles at 40 C under closed conditions. Sugarcoated tablets, whose ERH was controlled during the process
of sugar coating, were stored in glass bottles at 40 C under
closed conditions. In the experiment investigating the influence of sugar coating, plain tablets, sugar-coated tablets, and
halved sugar-coated tablets were stored at 60 C under closed
conditions.

tablet in 10 mL purified water) was measured using a pH meter

(F-16, Horiba, Japan).
3. Results and discussion
3.1. Stability of VB12 -CN in powders and granules
It is well known that vitamins are susceptible to moisture.
Therefore, we investigated the stability of VB12 -CN in powders,
and in granules containing TTFD-HCl, VB2 , and VB6 under
various conditions of humidity. Fig. 1(a) shows the stability of
VB12 -CN in powders. VB12 -CN was stable in powders stored at
40 C/6%RH and 40 C/57%RH but unstable in powders stored
at 40 C/75%RH. The stability of VB12 -CN in powders is therefore susceptible to moisture. Fig. 2(a) shows the stability of
VB12 -CN in granules. VB12 -CN was stable in granules stored at
40 C/6%RH, but unstable in granules stored at 40 C/57%RH
or 40 C/75%RH. The stability of VB12 -CN in granules was
more susceptible to moisture than in powders.
In powders and granules stored under conditions of high
humidity, decomposition followed neither zero-order nor firstorder kinetics. A zero-order equation can be expressed as shown
in the following equation:
(1 X) = K0 t + 1

(1)

where X is decomposition rate, t is time, and K0 is a zero-order

rate constant. A first-order equation can be expressed as shown
in the following equation:
ln(1 X) = K1 t

2.5.2. Measurements of VB12 -CN content

When measuring the VB12 -CN content of powders and granules, a sample of powders or granules corresponding to the
weight of five tablets was accurately weighed. The sample was
then dissolved in 25 mL diluent methanol (10%). The sample
suspension was shaken for 30 min in a shaker (Recipro shaker
SR-2w, Taitec Co., Ltd., Japan) and centrifuged for 10 min
at 10,000 rpm (O5P-21 centrifuge, Hitachi, Japan). The clear
supernatant liquid was used for the measurement of VB12 CN content. When measuring the VB12 -CN content of tablets,
10 tablets were crushed using a mill (8000 SPEX CertiPrep
Mixer/Mill, SPEX CertiPrep Inc., USA). The same procedure
as described above was then used to prepare the samples.
Measurements were performed using an HPLC method with
a system equipped with a controller (600, Waters, USA) and an
ultraviolet-visible light detector (SPD-6AV, Shimadzu, Japan).
HPLC conditions were as follows: column, YMC-Pack ODSA-312 (YMC Co., Ltd., Japan); column temperature, 25 C;
detection wavelength, 550 nm; mobile phase, 0.05 mol/L ammonium dihydrogenphosphate aqueous solution (pH 3.5):methanol
(3:1); flow rate, 1.0 mL/min; injection volume, 100 L.

163

(2)

where X is decomposition rate, t is time, and K1 is a first-order

rate constant.

2.6. Measurement of pH in the suspensions

The pH of suspensions of TTFD-HCl (10 g in 10 mL purified
water), VB6 (10 g in 10 mL purified water), granules (1 g in
10 mL purified water), and crushed sugar-coated tablets (one

Fig. 1. Stability of VB12 -CN in powders. (X) Decomposition rate. (a) Zero-order
plots; (b) Jander plots. () 40 C/6%RH; () 40 C/57%RH; () 40 C/75%RH.

164

S. Ohmori et al. / International Journal of Pharmaceutics 337 (2007) 161168

Table 1
Rate constants and R2 values in powders and granules

Fig. 2. Stability of VB12 -CN in granules containing TTFD-HCl, VB2 , and VB6 .
(X) Decomposition rate. (a) Zero-order plots; (b) Jander plots. () 40 C/6%RH;
() 40 C/57%RH; () 40 C/75%RH.

As decomposition followed neither zero-order nor first-order

equations, a Jander equation was applied to the decomposition
kinetics (Horioka et al., 1974; Okazaki et al., 1989, 1990).
Fig. 1(b) and Fig. 2(b) show Jander plots of the stability of VB12 CN in the powders and granules. The decomposition in powders
stored at 40 C/75%RH and in granules stored at 40 C/57%RH
and 40 C/75%RH was shown to follow the Jander equation,
which can be expressed as shown in the following equation:
(1 X)1/3 = KJ t 1/2 + 1

(3)

where X is decomposition rate, t is time, and KJ is a Jander rate

constant. In this case, the powder and the granules were kept
under dry conditions before the experiment. During the experiment, the surrounding moisture reacted with the interfaces of the
powders and granules, and VB12 -CN decomposed. The decomposition fitted neither the zero-order equation, which expresses
a drug-concentration-independent constant reaction model, nor
the first-order equation, which expresses a drug-concentrationdependent constant reaction model. The decomposition did fit
the Jander equation, which expresses a three-dimensional diffusion model. Rate constants and R2 values in powders and granules are listed in Table 1. The rate constant for the granules was
larger than that for the powders. Moisture contents of the powders stored at 40 C/6%RH, 40 C/57%RH, and 40 C/75%RH
were 3.9%, 12.8%, and 15.5%, respectively. Moisture contents of the granules stored at 40 C/6%RH, 40 C/57%RH,
and 40 C/75%RH were 2.2%, 3.5%, and 4.1%, respectively.
Although the differences in the decomposition rate between in
the granules stored at different humidity conditions were due to
the differences in moisture contents, the differences in decom-

Rate constant

R2

Powders at 40 C/75%RH
Zero-order
First-order
Jander

0.03
0.04
0.02

0.7968
0.8092
0.9553

Granules at 40 C/75%RH
Zero-order
First-order
Jander

0.07
0.03
0.05

0.7813
0.8243
0.9886

Granules at 40 C/57%RH
Zero-order
First-order
Jander

0.05
0.02
0.03

0.7194
0.7450
0.9780

position rate between in the powders and in the granules stored

at the same humidity conditions were not due to the differences
in moisture contents. The stability of VB12 -CN in the granules
could be affected by the other vitamins such as TTFD-HCl,
VB2 , and VB6 . Furthermore, during the granulation process,
VB12 -CN aqueous solution was adhered onto the surface of the
ingredients, resulting in the dramatic increase of the surface area
in drying process. The increase of the surface area of VB12 -CN
in the granules could contribute to the larger rate constant.
3.2. Stability of VB12 -CN in plain tablets
Moisture in solid dosage forms is generally classified as
mobile or immobile water. The measurement of mobile water is
expressed by ERH. Much research has been carried out on the
relationship between stability and ERH (Heidemann and Jarosz,
1991; Ohmori et al., 2004c), and it is now well recognised that
ERH is correlated with the stability of moisture-sensitive drugs.
Fig. 3 shows the stability of VB12 -CN in plain tablets containing TTFD-HCl, VB2 , and VB6 with different ERH. VB12 -CN
in plain tablets with a high ERH (65%) was less stable than
in tablets with a low ERH (44%). The stability of VB12 -CN in
plain tablets was also susceptible to moisture. Although decomposition in powders and granules followed the Jander equation,
decomposition in the plain tablets followed the zero-order equation. The difference in decomposition kinetics may be due to

Fig. 3. Stability of VB12 -CN in plain tablets containing TTFD-HCl, VB2 , and
VB6 at 40 C. (X) Decomposition rate. () ERH 44%; () ERH 65%.

S. Ohmori et al. / International Journal of Pharmaceutics 337 (2007) 161168

the different storage conditions. While the storage conditions

for the powders and granules were open, storage conditions for
the plain tablets were closed. Furthermore, the storage of plain
tablets was started after controlling their ERH in desiccators
using saturated solutions at 25 C. In other words, the moisture
conditions affecting the decomposition were different. Under
open conditions, surrounding moisture is able to attack VB12 CN and the decomposition reaction sites would therefore be
these interfaces. Under closed conditions, moisture in the plain
tablets, which can attack VB12 -CN, would be more uniform
inside the tablets compared with that in the granules because
of the process of controlling ERH. The decomposition reaction
sites would therefore be unlimited. Therefore, the decomposition
of VB12 -CN in plain tablets under closed conditions followed
the zero-order equation.
3.3. Stability of VB12 -CN in sugar-coated tablets
Fig. 4(a) shows the stability of VB12 -CN in sugar-coated
tablets containing TTFD-HCl, VB2 , and VB6 . The decomposi-

165

tion of VB12 -CN in sugar-coated tablets with low ERH (39%

and 41%) followed the zero-order equation. The decomposition
kinetics were the same as those for plain tablets. In general, the
ERH of sugar-coated tablets is more than 60% because of the
dusting method used, which is known as the intermittent spray
method. It is more difficult to dry the tablets using this method
than using the film-coating method (a continuous spray mist
method) as the sugar-coating makes it difficult to remove moisture from inside the tablets during the drying process (Ohmori
et al., 2004b, c). The stability of VB12 -CN in the sugar-coated
tablets with high ERH (>60%) was inferior to that in the sugarcoated tablets with low ERH. The decomposition kinetics of
VB12 -CN in the sugar-coated tablets with high ERH was evaluated kinetically and found not to follow any of the equations
mentioned (i.e., neither zero-order, first-order, nor Jander equations). Instead, the decomposition showed complex kinetics,
which could be due to a complex decomposition mechanism
such as occurs when the decomposition products themselves
accelerate the decomposition. In this case, a ProutTompkins
equation is used, which can be expressed as shown in the following equation:


X
ln
(4)
= Kpt t + C
1X
where X is decomposition rate, t is time, Kpt is a ProutTompkins
rate constant, and C is a constant. For example, Carstensen and
Attarchi (1988) reported that the decomposition of aspirin in
the solid state, in the presence of limited amounts of moisture,
followed the ProutTompkins equation. However, in the present
study, the decomposition of VB12 -CN in the sugar-coated tablets
did not follow the ProutTompkins equation. Yoshioka (1995),
when considering that some drugs in solid dosage forms show
complex decomposition kinetics, which are not described by any
theoretical kinetic equations, suggested that, when it is impossible to apply theoretical decomposition kinetics equations to the
decomposition of the drug, the Weibull equation is empirically
applicable to the decomposition. Both Okusa (1975) and Seki et
al. (1980) have applied the Weibull equation to the decomposition of drugs. Accordingly, we empirically applied the Weibull
equation to the decomposition curves. As shown in Fig. 4 (b), the
decomposition of VB12 -CN in the sugar-coated tablets with high
ERH followed the Weibull equation, which can be expressed as
shown in the following equation:


1
= ln k + m ln t
(5)
ln ln
1X

Fig. 4. Stability of VB12 -CN in sugar-coated tablets containing TTFD-HCl,

VB2 , and VB6 at 40 C. (X) decomposition rate. (a) Zero-order plots; (b) Weibull
plots; (c) AvramiErofeev plots. () ERH 39%; () ERH 41%; () ERH 65%;
() ERH 68%.

where X is decomposition rate, t is time, k and m are parameters.

The R2 values were 0.9575 (ERH 65%) and 0.9415 (ERH 68%).
Although the Weibull equation is empirical, kinetic equations
can be predicted by using the shape parameter, m, in Eq. (5). The
parameter, m, on the decomposition in the sugar-coated tablets
were 1.95 (ERH 65%) and 2.17 (ERH 68%). According to the
method of HancockSharp (Table 2), an AvramiErofeev equation would be applicable (Hancock and Sharp, 1972; Umeda et
al., 1985). As shown in Fig. 4(c), the decomposition of VB12 CN in the sugar-coated tablets with high ERH followed the

166

S. Ohmori et al. / International Journal of Pharmaceutics 337 (2007) 161168

Table 2
Values of m for solid-state decompositions (compiled from Hancock and Sharp, 1972; Umeda et al., 1985)
m

Equation

0.54
0.57
0.57
0.62
1.00
1.07
1.11
1.24
2.00
3.00

1 (1 X)1/3

Mechanism
= kt1/2

(Jander equation)
(1 X)ln(1 X) + X = kt
(1 2X/3) (1 X)2/3 = kt (GinstlingBrounshtein equation)
X2 = kt
ln(1/(1 X)) = kt
1 (1 X)1/3 = kt
1 (1 X)1/2 = kt
X = kt
ln(1/(1 X))1/2 = kt (AvramiErofeev equation)
ln(1/(1 X))1/3 = kt (AvramiErofeev equation)

AvramiErofeev equation, which can be expressed as shown

in the following equation:

1/2
1
ln
= Kae t
(6)
1X
where X is decomposition rate, t is time, Kae is an
AvramiErofeev rate constant. The decomposition mechanism
of VB12 -CN in the sugar-coated tablets with high ERH may be
a random nucleation (two-dimensional growth of nuclei). Rate
constants and R2 values in plain and sugar-coated tablets are
listed in Table 3.
In order to clarify the factors affecting the stability of VB12 CN in sugar-coated tablets, we investigated the influences of
TTFD-HCl, VB2 , and VB6 on the stability of VB12 -CN in
sugar-coated tablets (Fig. 5). The stability of VB12 -CN in sugarTable 3
Rate constants and R2 values in plain and sugar-coated tablets
Rate constant
Plain tablets (ERH 65%) at 40 C
Zero-order
0.07
Plain tablets (ERH 44%) at
Zero-order

R2
0.9996

40 C
0.04

0.9887

Sugar-coated tablets (ERH 68%) at 40 C

Zero-order
0.15
First-order
0.32
Jander
0.17
ProutTompkins
0.83
AvramiErofeev
0.25

0.9238
0.9195
0.7779
0.8921
0.9606

Sugar-coated tablets (ERH 65%) at 40 C

Zero-order
0.15
First-order
0.28
Jander
0.15
ProutTompkins
0.74
AvramiErofeev
0.24

0.9456
0.9368
0.7995
0.9144
0.9693

Sugar-coated tablets (ERH 41%) at 40 C

Zero-order
0.02

0.9915

Three-dimensional diffusion
Two-dimensional diffusion
Three-dimensional diffusion
One-dimensional diffusion
Random nucleation (one nucleus on each particle)
Phase boundary reaction (spherical symmetry)
Phase boundary reaction (cylindrical symmetry)
Zero-order mechanism
Random nucleation (two-dimensional growth of nuclei)
Random nucleation (three-dimensional growth of nuclei)

coated tablets without VB2 or VB6 , whose ERH is high (67%

or 68%), was superior to that in sugar-coated tablets containing
VB2 and VB6 . The decomposition of VB12 -CN in sugar-coated
tablets without VB2 or VB6 followed the zero-order equation.
We demonstrated that both VB2 and VB6 reduced the stability of
VB12 -CN and changed the decomposition kinetics of VB12 -CN
in sugar-coated tablets, with VB6 having the greater effect. On
the other hand, the stability of VB12 -CN in sugar-coated tablets
without TTFD-HCl was inferior to that in sugar-coated tablets
containing TTFD-HCl. TTFD-HCl therefore improved the stability of VB12 -CN in sugar-coated tablets. Although VB12 -CN
is known to be incompatible with thiamine, the precise stabilization mechanism of TTFD-HCl has not been determined in
this study. The difference between thiamine and TTFD-HCl is
tetrahydrofurfuryl group. Tetrahydrofurfuryl group could contribute to the stabilization of VB12 -CN in a measure.
pH is one of the most common factors influencing stability. In aqueous solutions VB12 -CN is stable between pH 4.0
and pH 7.0, while it is unstable in acidic solutions. Therefore,
we measured the pH of suspensions of TTFD-HCl, VB6 , the
granules, and crushed sugar-coated tablets and found them to
be 2.3, 2.2, 3.1, and 4.1, respectively. Although TTFD-HCl stabilized VB12 -CN and VB6 destabilized VB12 -CN, there was
no appreciable difference between the pHs of the suspensions
of TTFD-HCl and VB6 . In addition, although VB12 -CN in the
sugar-coated tablets was unstable, the pH of the suspension of

40 C

Sugar-coated tablets (ERH 39%) at

Zero-order
0.02
0.9710
Sugar-coated tablets (without VB2 ) (ERH 67%) at 40 C
Zero-order
0.05
0.9799
Sugar-coated tablets (without VB6 ) (ERH 68%) at 40 C
Zero-order
0.03
0.9975

Fig. 5. Influences of TTFD-HCl, VB2 , and VB6 on stability of VB12 -CN in

sugar-coated tablets at 40 C. (X) Decomposition rate. () Sugar-coated tablets
without VB6 (ERH 68%); () sugar-coated tablets without VB2 (ERH 67%);
() sugar-coated tablets containing TTFD-HCl, VB2 , and VB6 (ERH 65%); ()
sugar-coated tablets without TTFD-HCl (ERH 70%).

S. Ohmori et al. / International Journal of Pharmaceutics 337 (2007) 161168

Fig. 6. Influence of sugar coating on stability of VB12 -CN in tablets containing

TTFD-HCl, VB2 , and VB6 at 60 C. (X) Decomposition rate. () Plain tablets
(ERH 65%); () sugar-coated tablets (ERH 65%); () halved sugar-coated
tablets (ERH 65%).

167

other vitamins, decompose VB12 -CN. We therefore propose the

following decomposition mechanism: the combination of sufficient quantities of moisture, VB2 , and VB6 , generates gases
in the sugar-coated tablets during storage. Due to the gas-tight
coating layer, these gases cannot escape and are free to attack
the VB12 -CN in the tablets, causing decomposition. This proposed mechanism would account for the complex decomposition
kinetics. Acetic acid, formic acid, furfural, and furfur alcohol
can all be detected in sugar-coated tablets after storage at 40 C
for 8 months. Unfortunately, the gaseous components related
to decomposition were not fully identified in this study. As a
result of these findings, we propose that separating VB12 -CN
from VB6 and reducing the ERH in the tablets would stabilize
VB12 -CN in sugar-coated tablets.
4. Conclusions

the crushed sugar-coated tablets was in the stable region (pH

4.07.0). This showed that the different influences of TTFDHCl, VB2 , and VB6 on the stability of VB12 -CN were not due
to pH.
Fig. 6 shows the influence of sugar coating on the stability
of VB12 -CN in tablets. Although the ERH levels of the tablets
were same, the stability of VB12 -CN in the sugar-coated tablets
was inferior to that in the plain tablets. Furthermore, although
the ingredients and ERH levels in the tablets were the same,
the stability of VB12 -CN in the halved sugar-coated tablets was
superior to that in whole sugar-coated tablets. This suggests that
while the ingredients of the sugar-coating layer did not significantly reduce the stability of VB12 -CN in the tablets, the
sugar-coating layer itself did. Budavari et al. (2001) reported
that film-coated tablets containing folic acid, VB6 , and VB12 -CN
showed improved stability of VB12 -CN compared with uncoated
tablets. In the present study, sugar coating did not improve the
stability of VB12 -CN. The difference between sugar-coating and
film-coating is the structure of the coating layer. Sugar-coating
is a denser coating layer than film-coating, and shows a superior masking ability for unpleasant odors (Ohmori et al., 2004c,
2005), being a gas-tight layer.
Jacob et al. (1968) reported that the content of VB12 -CN
in film-coated multivitamin tablets was reduced after exposure to methanol vapor for 1 month at room temperature. They
postulated that methanol vapor causes decomposition of either
ascorbic acid or VB1 , or both, the decomposition products of
which in turn influence the stability of VB12 -CN. Harada et
al. (1969) reported that formic acid, which arises from VB1
decomposition, decomposed VB12 -CN, while Terao et al. (1980)
reported that sodium bisulfite decomposed VB12 -CN. DeRitter
(1982) wrote that VB12 -CN is decomposed by the decomposition products of ascorbic acid, the combination of ascorbic
acid-Cu2+ or dehydroascorbic acid-Cu2+ , the 4-methyl-5-(hydroxyethyl) thiazole formed by cleavage of thiamine, or
the combination of thiamine and niacinamide. Several compounds, in particular the decomposition products of vitamins,
have been shown to decompose VB12 -CN, but unfortunately
they have not been fully identified. Our results suggest that
the gases generated inside the sugar-coated tablets containing

The decomposition kinetics of VB12 -CN in sugar-coated

tablets containing TTFD-HCl, VB2 , and VB6 with high ERH
(more than 60%) under closed conditions was different from
that in powders kept under open conditions, granules kept
under open conditions, plain tablets kept under closed conditions, sugar-coated tablets with low ERH (ca. 40%) kept under
closed conditions, and sugar-coated tablets without VB2 or
VB6 kept under closed conditions. The decomposition kinetics of VB12 -CN in powders and granules kept under open
conditions followed the Jander equation, which expresses a
three-dimensional diffusion model. The decomposition kinetics of VB12 -CN in plain tablets, in sugar-coated tablets with
low ERH (ca. 40%), and in sugar-coated tablets without VB2
or VB6 kept under closed conditions, followed the zero-order
equation, which expresses a drug-concentration-independent
constant reaction model. The decomposition kinetics of VB12 CN in sugar-coated tablets with high ERH (more than 60%)
kept under closed conditions showed complex kinetics and
followed the AvramiErofeev equation, which expresses a random nucleation (two-dimensional growth of nuclei) model. We
demonstrated that ERH, the presence of other vitamins (VB2
and VB6 ), and sugar-coating, reduced the stability of VB12 -CN
and changed the decomposition kinetics of VB12 -CN in sugarcoated tablets from a zero-order equation to a complex kinetics
equation, the AvramiErofeev equation.
Acknowledgment
We thank Mr. Masaharu Imamura for developing the HPLC
method for measurement of cyanocobalamin content.
References
Budavari, Z., Zelko, R., Antal, I., Marton, S., Racz, I., 2001. Comparison of the
stability of different tablet formulations containing folic acid, vitamin B6
and B12 . Pharmazie 56, 668.
Carstensen, J.T., Attarchi, F., 1988. Decomposition of aspirin in the solid state
in the presence of limited amounts of moisture. III. Effect of temperature
and a possible mechanism. J. Pharm. Sci. 77, 318321.
DeRitter, E., 1982. Vitamins in pharmaceutical formulations. J. Pharm. Sci. 71,
10731096.

168

S. Ohmori et al. / International Journal of Pharmaceutics 337 (2007) 161168

Hancock, J.D., Sharp, J.H., 1972. Method of comparing solid-state kinetic data
and its application to the decomposition of kaolinite, brucite, and BaCO3 . J.
Am. Ceram. Soc. 55, 7477.
Harada, K., Saito, I., Miura, H., Suzuki, K., Utsumi, I., 1969. Studies on vitamin
B12 . I. Thiamine as a degradation factor of cobalamins. Yakugaku Zasshi
89, 464468.
Heidemann, D.R., Jarosz, P.J., 1991. Preformulation studies involving moisture
uptake in solid dosage forms. Pharm. Res. 8, 292297.
Horioka, M., Aoyama, T., Takata, K., Maeda, T., Shirahama, K., 1974. Degradation of propantheline bromide and dried aluminum hydroxide gel in
powdered preparations. Yakuzaigaku 34, 1621.
Ichikawa, M., Ide, N., Shiraishi, S., Ono, K., 2005. Effect of various halide
salts on the incompatibility of cyanocobalamin and ascorbic acid in aqueous
solution. Chem. Pharm. Bull. 53, 688690.
Jacob, J.T., Nessel, R.J., Blodinger, J., 1968. Stability of cyanocobalamin in
film-coated multivitamin tablets. J. Pharm. Sci. 57, 18541857.
Ohmori, S., Ohno, Y., Makino, T., Kashihara, T., 2004a. Improvement of impact
toughness of sugar-coated tablets manufactured by the dusting method.
Chem. Pharm. Bull. 52, 322328.
Ohmori, S., Ohno, Y., Makino, T., Kashihara, T., 2004b. Effect of moisture
on impact toughness of sugar-coated tablets manufactured by the dusting
method. Chem. Pharm. Bull. 52, 329334.
Ohmori, S., Ohno, Y., Makino, T., Kashihara, T., 2004c. Development and evaluation of the tablets coated with the novel formulation termed thin-layer
sugarless coated tablets. Int. J. Pharm. 278, 459469.
Ohmori, S., Ohno, Y., Makino, T., Kashihara, T., 2005. Application of an electronic nose system for evaluation of unpleasant odor in coated tablets. Eur.
J. Pharm. Biopharm. 59, 289297.

Okazaki, K., Nishigaki, R., Hanano, M., 1989. Equation and accelerating factor
of degradation in cocarboxylase free ester lyophilizate. Yakuzaigaku 49,
141147.
Okazaki, K., Nishigaki, R., Hanano, M., 1990. Relationship between degradation rate and relative humidity, temperature, or physical pretreatments of
carboxylase free ester lyophilizate. Yakuzaigaku 50, 141148.
Okusa, N., 1975. Prediction of stability of drugs. III. Application of Weibull
probability paper to prediction of stability. Chem. Pharm. Bull. 23, 794
802.
Seki, H., Hayashi, T., Okusa, N., 1980. An application of weighted least-squares
analysis to Weibull probability paper for prediction of stability of drug
products. Yakuzaigaku 40, 201207.
Terao, M., Marui, E., Tanaka, K., Nakao, Y., 1980. Studies on the formulation and admixture of parental preparations. I. Degradation of ascorbic acid
and cyanocobalamin by sodium bisulfite added to ascorbic acid injection.
Yakugaku Zasshi 100, 8187.
The Japanese Pharmacopoeia Fourteenth Edition, 2001. Hirokawa Publishing
Co., Japan.
Toyama Pharmaceutical Research Association, 1992. Studies on the Stability
of Vitamin B Complex-Combined-Preparations, Stability of Vitamin B1 ,
B6 , B12 preparations, following proprietary standard. Kateiyakukenkyu (in
Japanese) 11, 3340.
Umeda, T., Ohnishi, N., Yokoyama, T., Kuroda, T., Kita, Y., Kuroda, K., Tatsumi, E., Matsuda, Y., 1985. A kinetic study on the isothermal transition of
polymorphic forms of tolbutamide and mefenamic acid in the solid state at
high temperatures. Chem. Pharm. Bull. 33, 20732078.
Yoshioka, S., 1995. Stability of Drugs and Dosage Forms. Nankodo Co. Ltd,
Japan (in Japanese).