In vitro and mechanism study of poly(ethylene-co-vinyl acetate)-based implant for sustained release of vitamin B12

653

Macromolecular Research, Vol. 18, No. 7, pp 653-659 (2010) www.springer.com/13233

The Polymer Society of Korea

In vitro and Mechanism Study of Poly(ethylene-co-vinyl acetate)-Based Implant

for Sustained Release of Vitamin B12

Chi Zhang*, Allan J. Easteal, Neil R. Edmonds, and Genhai Liang†

Chemistry Department, University of Auckland, New Zealand

Majid Razzak and Wayne Leech

Bomac Laboratories Ltd, New Zealand

Received December 24, 2009; Revised March 11, 2010; Accepted March 12, 2010

Abstract: A polymer-vitamin B12 implant system has been developed to overcome the disadvantages of traditional

vitamin B12 administration to sheep. The ethylene/vinyl acetate (EVAc) copolymer was pre-blended with crystalline

vitamin B12. The blends were then extruded at 100 ºC at a constant pressure and rate using a piston extruder to form

a polymer-vitamin B12 composite cord. The final implant was a cylinder 2 mm in diameter and 4 mm in length.

Determination of the rate of vitamin B12 release from the implant into a pH7.4 phosphate buffer at 37 oC showed

that the release rate was strongly dependent on the feed concentration, size and shape of vitamin B12 crystals. The

well-known Korsmeyer-Peppas exponential equation was applied to and the release mechanism was found to be a

typical anomalous transport mechanism with a strong diffusion-controlled feature between pure diffusion-controlled

and Case-II transport mechanisms because the n values obtained were between 0.45 and 0.51 for most polymer-vita-

min B12 cylindrical implant systems.

Keywords: cobalt deficiency, vitamin B12, implant, in vitro release, mechanism, ethylene/vinyl acetate (EVAc)

copolymer.

Introduction

Cyanocobolamine (vitamin B12) is an essential compound

for animal health, since it plays an important role in DNA

synthesis and neurological function. Vitamin B12 works

with folic acid to produce red blood cells. Because vitamin

B12 is water soluble and travels with bloodstream and if it

is not used by the body then it will be eliminated in urine,

which means human beings or animals need a continuous

supply of them. A deficiency of cobalt (which is provided in

easily assimilable form by vitamin B12) can lead to a broad

spectrum of hematologic and neuropsychiatric disorders.

While all animals may experience problems if subjected to a

vitamin B12 (VB12) deficiency, growing lambs are particu-

larly susceptible. Cobalt deficiency symptoms in ruminant

animals include a loss of appetite, emaciation, weakness,

anaemia, and decreased production.1 Reduced growth rates

are commonly seen and can lead to marked wasting and

even death. In New Zealand, reduced growth rates usually

appear from 3 to 4 months of age, though not all the lambs

in the flock may be affected.2

Oral and intramuscular injections are traditionally used

for administration of VB12, however the initial therapeutic

and maintenance dosages are very high3 and above the

required VB12 levels in serum.3,4 In Australia it is generally

recommended that a subcutaneous injection of 2 mg VB12

is given to lambs at the time of marking (tail docking and

castration) and a cobalt pellet is given orally at weaning for

long-term protection against cobalt deficiency.5 Judson and

co-workers5 reported that in cobalt deficient areas lambs

should be given a 2 mg dose of VB12 at least bimonthly to

reduce the risk of cobalt deficiency.

For lambs in New Zealand, VB12 concentrations of <220

pmol L-1 in serum have been associated with a 95% proba-

bility that lambs were Co deficient.6 The lambs were injected

subcutaneously with 2 mg of soluble VB12 while another

group of 36 animals served as untreated controls. Blood and

liver biopsy samples for VB12 determinations were col-

lected. The serum VB12 concentrations of the VB12 treated

lambs increased rapidly compared to the untreated lambs.

However, the concentration level in serum was uneven dur-

ing the daily treatment; it peaked at day 2, decreased rapidly

to day 8, and then decreased more slowly until day 24 when

there were no longer differences between the groups. Liver

DOI 10.1007/s13233-010-0711-x

*Corresponding Author. E-mail: [email protected]†Present Address: Department of Materials Engineering, Monash

University, Wellington Road, Clayton, Victoria, 3800, Australia

C. Zhang et al.

654 Macromol. Res., Vol. 18, No. 7, 2010

VB12 concentrations of the VB12 treated lambs were sig-

nificantly greater over days 8-24. A subcutaneous injection

of 2 mg of soluble VB12 was effective in increasing and

maintaining the VB12 status of lambs for about 24 days.7

In a polymer-VB12 implant system comprising VB12 dis-

persed in the polymer matrix, the implant must remain

active for the desired period and the polymer, if not biode-

gradable, may create physiological problems for the animal

and may appear in the edible carcass after slaughter.

The New Zealand existing patented technology2 has given

rise to an implant that allows release of VB12 for periods

reported to exceed 60-150 days at a daily release rate for a

single implant above 5-10 µg day-1. However, only 75% of

the VB12 in the implant is available for release, and the

technology is expensive and not conducive to large-scale

manufacture. The objective of the present work was to develop

a new polymer-VB12 blending technology by means of

crystal size and shape control together with varying ratio of

VB12/copolymer of implants which provide several months

of efficient sustained release of VB12 to give ≥220 pmol

VB12 per litre of serum. To that end we have developed a

millimeter-scale implant system based on VB12 dispersed

in ethylene-vinyl acetate (EVAc) copolymer that meets sus-

tained release requirements for lambs. A rapid, efficient

method of pre-dispersing VB12 in the matrix polymer has

been evaluated, the effect of VB12 particle size has been

assessed and used to advantage, and an upscalable implant

production process has been designed and tested.

Experimental

Materials. Dichloromethane and absolute ethanol (both

AnalaR grade) were purchased from BDH Chemicals, and

ethylene-co-vinyl acetate copolymer Elvax(R) 40W contain-

ing 40% (w) vinyl acetate (EVAc) was supplied by Dupont.

Vitamin B12 (VB12, molecular formula: C63H88CoN14O14P)

powders from two sources were utilised, designated hence-

forth as S-type (SVB12) in the form of needle-like crystals,

from Sunji, China, and H-type (HVB12) in the form of micro-

crystal aggregates from Hoechst, Germany. Both types of

VB12 satisfy pharmaceutical standard BP98. As VB12 is a

cobalt-organic ligand complex with complicated chemical

structure, VB12 crystals do not melt below 300oC8 but they

have darken at about 210-230 oC.9 VB12 is in general solu-

ble in water and alcohol,10 some groups in VB12 can be pro-

tonated or deprotonated at low or high pH values that the

solubility of VB12 depends strongly on pH.11 The water and

PBS buffer solubility of VB12 was reported as 10.7 and 10.2

mg/mL, respectively, at room temperature.12 In combination

with the ignored low water uptake of the copolymers, only a

fraction of the VB12 will be in the dissolved state in the

exposed cavities during the release.

Preliminary Experiments. Crystalline VB12 is stable at

temperatures up to 100oC for long periods, and aqueous

solutions at pH 4.0-7.0 can be autoclaved with very little

loss. However, decomposition is rapid when the vitamin is

heated at pH 9.0 or above.13 Simultaneous TGA/DTA mea-

surements have shown that VB12 loses weight, at heating

rate of 10 oC min-1 in the 20-140 oC region, due to the removal

of water.14 Using a lower heating rate (3 oC min-1), between

140 and 230 oC, VB12 showed a mass loss of about 2%

between 140-145 oC, arising from the loss of the cyano

group.14 Consequently, it is inadvisable to process VB12 at

temperature higher than 140oC and for that reason in the

present study extrusion was carried out at 100 oC.

Thermal analysis of VB12 and EVAc was carried using

differential scanning calorimetry (DSC) with a Polymer

Laboratories model 12000 instrument. The samples were

contained in aluminum pans and heated at 10oC min-1 in an

atmosphere of flowing nitrogen. The EVAc showed a low

degree of crystallinity, two sub-Tm transitions at about

-132oC and -25 oC, and melting point ~55 oC.

Kinetic studies of thermal degradation of EVAc at varying

temperatures revealed that at 175oC the oxidative induction

times for EVAc with 28% and 33% VA were 720 and 240 min,

respectively, and at 200oC the oxidation induction times

were 60 and 30 min, respectively.15 It follows that the EVAc

should not have undergone any thermally induced degrada-

tion at the processing temperature that was used in the present

work. According to Dupont, ethylene-co-vinyl acetate copoly-

mer Elvax(R) 40W melts at 47 oC with maximum processing

temperature at 230oC and this copolymer complies with US

Food and Drug Administration Regulation.16 Due to the low

melting point, softness of the EVAc and the crystalline mor-

phologies of VB12s, it is possible to use extrusion way to make

the EVAc-VB12 implant without degrading the copolymer

and readily for processing. Further, extrusion makes the

needle-like crystals of VB12 orientated along the direction

of the extrusion stress and results in the inter-connections of

channels ready for VB12 release and solution transport.

Biodegradable polymers such as polycaprolactone (PCL,

melting point ~59-64oC, depending on its crystallinity17) with

different crystallinity, molecular weights or their blends are

also candidates for replacing non-biodegradable EVAc copoly-

mer in our study, however, in general the cost of PCL is two

to three-fold higher than that of EVAc copolymers.18 Fur-

ther, biodegradable implant matrix makes the mechanism of

release of VB12 more complicated due to the erosion/degra-

dation of the PCLs. The choice of the EVAc copolymer as

materials for the implant study and application was then

determined in present work. Our next generation material

for VB12-implant has been designed and tried using PCL

but the research work will not be presented in this paper.

Blending of Copolymer with VB12 and Implant Man-

ufacture. In a typical preparation 6 g EVAc was dissolved

in 36 mL of dichloromethane in a sealed flask at room tem-

perature. The mixture was continuously stirred for 1-2 h

until the polymer was completely dissolved, whereupon a

EVAc Implant for Sustained Release of Vitamin B12

Macromol. Res., Vol. 18, No. 7, 2010 655

quantity of VB12 powder sufficient to give 35%, 37.5%,

40% or 45% (w/w) VB12, based on VB12 and EVAc, was

added with stirring. The flask was wrapped with aluminium

foil to avoid exposure to light and the contents of the flask

were stirred for 2 h. At that point the dispersion of VB12 in

polymer solution was rapidly poured in a 24 cm diameter

glass petri dish, whereupon a solid sheet of VB12/EVAc

composite was rapidly formed by evaporation of the solvent.

After standing at ambient temperature overnight, residual

solvent was removed by vacuum (10-20 mmHg) drying at

40oC to constant weight.

After removal of residual solvent the VB12/EVAc com-

posite sheet was cut into pieces a few mm in size. The mate-

rial was added to a purpose-built, manually operated piston

extruder capable of temperature control to ±1 oC at 100 oC.

The VB12/EVAc composite was pre-heated at 100oC for

2 h, then extruded with an applied pressure of 3 to 20 kg cm-2

through a 2.0 mm i.d. cylindrical die. The resulting cord

was cut into 0.5-1 cm segments and re-extruded under the

same conditions. A chopper with rotating blade was used to

cut the final cord into pellets 4 mm long and 2 mm in diameter.

Vitamin B12 Assay. Assay of VB12 was carried out spec-

trophotometrically using a Philips PU 8620 UV/VIS/NIR

spectrophotometer. Absorbance measurements at 361 nm

showed conformity to the Beer-Lambert law for VB12 solu-

tions in water containing up to 50 mg L-1. The relationship

between VB12 concentration (C/µg L-1) and UV absorption

(A) was found to be:

A = 0.0186×C+0.0014 (1)

with correlation coefficient R2=0.9999. The concentration

was calculated assuming VB12 purity of 96.1%.

In vitro Drug Release. The rate of release of VB12 into

pH 7.4 phosphate buffer at 37 oC was determined spectro-

photometrically. Due to the porosity of the pellets their den-

sity was typically about 10% smaller than the density of the

buffer solution (~1 g cm-3). To ensure that during the release

experiments the implant pellets were completely immersed

in buffer solution, each pellet was enclosed in a nickel mesh

cage (~1 cm long × 0.5 cm in diameter) which prevented

the pellet from floating on the surface of the buffer solution.

Each pellet was immersed in 20 mL buffer solution main-

tained at 37 oC, and five replicates were used for each com-

position. At appropriate time intervals, the solutions with

released VB12 were separated from the pellets then trans-

ferred to brown glass vials for UV absorbance measurement.

Where necessary, solutions were diluted to ensure that the

VB12 concentrations were within the range of the calibra-

tion curve. A compressed air flow was used to remove the

residual solution from the nickel cages and pellets, and each

pellet was transferred to a vial containing 20 mL of fresh

buffer solution at 37 oC and the release of VB12 continued.

Figure 1. SEM micrograph for (a) HVB12 powder; (b) 40 wt% HVB12 aggregates in EVAc (× 500).

Figure 2. SEM micrograph for (a) SVB12 powder; (b) 40 wt% SVB12 aggregates in EVAc (× 500).

C. Zhang et al.


Microscopic Examination. The VB12 and VB12-loaded

pellets were examined using a Philips XL30 scanning elec-

tron microscope after gold coating the samples.

Results and Discussion

Morphology Study of VB12 Powders and Implants. The

rate of release of VB12 from the implant pellets was found

to depend markedly on the size of the VB12 crystals and

their morphology. Figures 1(a) and 2(a) show that HVB12

was made up of 50-100 µm aggregates of very small crys-

tals, whereas the SVB12 material was in the form of needle-

like single crystals with size 100-200 µm. Figures 1(b) and

2(b) show that the initial blended states of VB12 aggregates

in EVAc, respectively, for HVB12 and SVB12.

Figure 3 shows the smooth cross-section of a typical EVAc

pellet without VB12. The change in the surface morphology

after loading with the microcrystalline VB12 but before

immersion in pH 7.4 buffer is apparent in Figures 4(a) and

(b), and the subsequent changes after immersion in buffer

solution (for 14 days) are revealed in Figures 4(c) and (d).

Figures 4(a) and (b) show that aggregates of HVB12 are

trapped in the EVAc matrix and that the HVB12 aggregates

create cavities in the pellet and on implant surface. How-

ever, after 14 days immersion in buffer at 37oC (shown in

Figure 4(c)), the exposed end-face of HVB12-EVAc pellet

shows many more cavities distributed throughout the sur-

face. By contrast the outer surface of the pellet is changed to

a much lesser extent by immersion in buffer solution (see

Figure 4(b) and (d)), implying that VB12 release has occurred

predominantly through the end faces of the pellet.

Figures 5(a-d) show SEM micrographs of SVB12-EVAc

pellets, corresponding to the micrographs in Figures 4(a-d).

Figure 3. SEM micrograph of EVAc pellet cross-section without

VB12 (× 125).

Figure 4. SEM micrographs of HVB12-EVAc implant. (a) cross-section before in vitro release (× 125); (b) exterior before in vitro

release (× 250); (c) cross-section after in vitro release (× 125); (d) exterior after in vitro release (× 250).



The needle-like SVB12 crystals similarly create cavities

and channels in the end face of the pellet before immersion

in buffer, and the number of cavities increases after immer-

sion. In this case there is clear evidence (Figure 5(d)) of loss

of VB12 through the exterior surface as well as via the end

faces of the pellet. The morphology changes noted above,

resulting from immersion in buffer solution, suggest that the

larger crystal size of SVB12 creates more cavities and chan-

nels in the pellet than does HVB12, allowing more efficient

access of fluid to the VB12 particles and hence more rapid

leaching of VB12 from the pellet.

In vitro VB12 Release. Release profiles for a range of

loadings of SVB12, HVB12 and a mixture of SVB12 and

HVB12 in equal proportions by weight are shown in Figure 6,

which indicates that VB12 release profiles depend strongly

on the size and morphology of the VB12 particles, as well

as the drug loading. For the pellets loaded with SVB12

(curves 1 and 2), the cumulative release increases with drug

loading, but for the smaller HVB12 particles the cumulative

release is almost independent of drug loading (curves 4, 5,

and 6, respectively).

It is clear from Figure 6 that, at least within the timescale

of the experiment, the proportion of VB12 that is released

from the pellet is very much higher for the SVB12 material

Figure 5. SEM micrograph of SVB12-EVAc implant. (a) cross-section before in vitro release (× 125); (b) exterior before in vitro release

(× 250); (c) cross-section after in vitro release (× 125); (d) exterior after in vitro release (× 250).

Figure 6. Comparison of in vitro release behaviour of SVB12

and HVB12 in VB12/EVAc pellets in pH 7.4 phosphate buffer

(PSB) at 37o

C. 1 and 2 denote pellets containing 45% and 40%

weight percent of SVB12, 3 denotes 40% mixed VB12s (mixing

ratio of SVB12/HVB12=1:1); 4, 5 and 6 denote pellets contain-

ing 37.5%, 35% and 40% weight percent of HVB12, respec-

tively. The data shown are averages from at least 5 replicate

determinations. The vertical lines are error bars from standard

deviations.

C. Zhang et al.


than for HVB12. It appears that a significant proportion of

the small HVB12 crystals become completely coated with

EVAc during blending and are not leachable from the pellet,

with consequent reduction in the proportion of active mate-

rial released.

For the intended application of these implants, SVB12 is

released too rapidly and the proportion of HVB12 released

is unacceptably small. A possible solution to that problem

that we have investigated is to utilize a mixture of SVB12

and HVB12 as the active material. In Figure 6, curve 3

shows release profiles for pellets with a 40% loading of

SVB12, HVB12 and a mixture of SVB12 and HVB12 in

equal proportions by weight. The rate of release for the pel-

let with the mixture of SVB12 and HVB12 is nearly optimal

for the intended application, and the proportion of active

material remaining as a residue in the pellet is of the order

20% to 30%. In vivo trials have confirmed that the perfor-

mance of these implants in live sheep meets the require-

ments of effective concentration level for treatment of cobalt

deficiency up to 6 months.

The release data were fitted to the exponential equation

(Korsmeyer-Peppas equation19), which is often used to

describe the drug release behavior from polymeric systems

where the mechanism is not well-known or more than one

type of release phenomena is involved:

M(t)/M(f ) = ktn (2)

where k is a kinetic constant relevant to the structural and

geometric characteristics of the matrix pellets, n is the expo-

nent characterizing release mechanism20,21 and M(t)/M( f )

represents the drug dissolved fraction or cumulative release

percentage at time t.21 M( f ) is the total amount of VB12

loaded in pellets and M(t) is the cumulative released amount

of VB12 at time t. This equation is normally acceptable as a

valid limit of M(t)/M( f )≤0.6.19,20

The validation can be seen in Figure 7 when the linearity

is kept between log (Qt%) vs log(t) at low cumulative% (Qt%).

To clarify the release exponent for different sets of pellets,

the log value of the percentage of cumulative VB12 released

was plotted against log time for each pellet set according to

eq. (3).

Log[M(t)/M( f )] = log(k) + nlog(t) (3)

The estimation values for n and log(k) were obtained and

list in Table I.

In case of slab geometry, n=0.5 indicates a pure diffusion-

controlled drug release and n=1.0 means that the release is

swelling-controlled.21 In case of cylinder geometry, n=0.45

is a Fickian diffusion or pure diffusion-controlled mecha-

nism while n=0.89 is swelling-controlled release or Case-II

transport mechanism.20,21 These k values in Table I show that

the higher feed wt% (45% and 40%, pellet set ID: 1, 2, 3

and 6) have higher kinetic constants, indicating that initial

release rate of VB12s depends strongly on feed wt% of

VB12s. In vivo trials have confirmed that the performance

of these implants in live sheep meets the requirements of

effective concentration level for treatment of cobalt defi-

ciency up to 6 months with k=12.59 and n=0.47. It is under-

stood that high k results in a fast release within a certain

period of time and low k has the opposite effects.

Figures 4 and 5 showed that VB12-EVAc pellets were

shown as cylinder type, no matter what size and shape of

VB12 particles embedded in the polymer matrix. The com-

parison with the above two figures and Table I (set 2 for

SVB12 and 6 for HVB12, respectively) also shows that at

40% VB12 feed, after 14 days release, ~59% of VB12 was

dissolved from the deep channels of SVB12-EVAc pellets

while only about 33% of VB12 was released out of HVB12-

EVAc pellets. It is obvious that needle-like VB12 particles

provide more opportunities for connections between the

VB12 particles in the polymer matrix. The n values for pel-

let sets 1 to 5 (45% and 40% of SVB12, 40% mixed VB12s

(1:1), 37.5% and 35.0% HVB12, respectively) are all not far

from 0.45 but between 0.45 and 0.89 (0.45<n<0.89). Taking

EVAc as a non-swellable polymer matrix (slightly swelling

in water due to the polar groups of acetate) into account, for

these n values, their release behavior can be regarded as a

combined mechanisms of diffusion-controlled (dominant)

with swelling-controlled releases (minor) or say “anomalous

transport mechanism” (ATM mechanism).21

It was known that for hydrophilic-hydrophobic poly(eth-

Figure 7. Plot of log(t) against log(Qt%) for estimation of n and k

values.

Table I. Estimates for n, Log(k) and k (Data Derived from Fig.

7 and eq.(3) and the Denotes 1 to 6 as the Same in Fig. 6)

Pellet Set ID 1 2 3 4 5 6

Log(k) 1.337 1.253 1.100 1.004 0.9318 1.115

k 21.73 17.91 12.59 10.09 8.547 13.03

n 0.5062 0.4548 0.4700 0.4691 0.4900 0.3478



ylene glycol)/poly(butylene terephthalate) multiblock (PEG/

PBT) copolymers the release rate can be effectively affected

by PEG-segment lengths, and, the swelling and the crystal-

linity of the matrix determine the diffusion coefficient of

VB12 through the copolymers. In general, the swelling can

mainly be varied by changing the PEG-segment length. At a

constant PEG-segment length, the crystallinity increases at

increasing wt% PBT.22 In our study using EVAc Elvax(R)

40W, due to the general hydrophobic nature of ethylene and

vinyl acetate with slightly swelling feature of the acetate

groups in water, the combined mechanisms could be under-

stood. In addition, there are so many groups such as

NH2CO-, -OH, -CH2CH2- in the VB12 structure, the inter-

actions between groups in EVAc copolymer and VB12 may

exist and lead to VB12 slightly soluble/miscible in EVAc

copolymer as a composite. Consequently, “ATM” mecha-

nism was seen in EVAc-VB12 implant release behaviour.

Recently, it was known that in the case of release from the

two-dimensional percolation fractal, the pellet set 6 (40%

HVB12) with n about 0.35 is probably with such a fractal

dimension where the values of n ranged from 0.35 to 0.39.23

The lower value of n in the percolation cluster reflects the

slowing down of the diffusion process in the disordered

medium or the release device is not homogeneous.23 The

greater errors of cumulative release% in HVB12-pellets

(Table I, sets 4, 5 and 6) also indicate the heterogeneousness

of blending of EVAc polymer and HVB12 particles. For

commercial reasons, the details of polymer-drug blending

procedures and in vivo experimental data will not be present

in this paper. However, it is very likely that the unusual low

n value and heterogeneousness for 40% HVB12 in the form

of microcrystal aggregates may be caused by extraordinary

high viscosity of molten polymer-VB12 blend from 40%

HVB12 loading.

Conclusions

The in vitro release behavior of VB12 depends strongly

on the size and morphology of VB12 crystals and the con-

centration of VB12 in VB12-EVAc composites. The VB12-

EVAc implant can be considered as a device with both ends

open. Relatively large needle-like VB12 crystals in EVAc

copolymer matrix create deeper and larger channels which

allow more VB12 to be released with higher cumulative

release and higher release rate, compared to aggregates of

very small VB12 crystals. In general, the current study

reveals that the release mechanisms for VB12-EVAc cylin-

drical pellets releasing VB12 in pH 7.4 PSB are mainly dif-

fusion-controlled anomalous transport mechanism and the

Korsmeyer-Peppas exponential equation was successful to

apply for the study. The exponent n values depend on the

percentage of VB12 feed, the form of crystals and on geom-

etry of polymer carrier.

Acknowledgement. Authors are grateful to Technology

New Zealand for granting the TIF scholarships and the col-

laborations from Bomac Laboratory Ltd, New Zealand.

References

(1) R. Alexandrova, R. Tudose, E. Arnaudova, O. Costisor, and

L. Patron, Experimental Pathology and Parasitology, 7, 3

(2004).

(2) D. G. McLaren, S. Z. Hashmi, I. G. Tucker, M. H. Abdul-Raz-

zak, and M. L. Xiahong, GB99-17757 2340395, 19990729

(2000).

(3) R. Oh and L. B. David, American Family Physician, 67, 979

(2003).

(4) H. W. Baik and R. M. Russell, Annual Review of Nutrition,

19, 357 (1999).

(5) G. J. Judson and P. J. Babidge, Australian Veterinary Journal,

80, 777 (2002).

(6) N. D. Grace, S. O. Knowles, G. R. Sinclair, and J. Lee, New

Zealand Veterinary Journal, 51, 89 (2003).

(7) N. D. Grace, D. M. West, and N. D. Sargison, New Zealand

Veterinary Journal, 46, 194 (1998).

(8) K. Folkers, Abstracts Communs. 1st Intern. Congr. Biochem.,

86 (1949).

(9) E. L. Rickes and T. R. Wood, Vitamin B12. 2563794 (1951).

(10) R. R. Williams and R. E. Waterman, J. Biol. Chem., 68, 499

(1926).

(11) C. W. Tsai and M. D. Morris, Anal. Chim. Acta, 76, 193

(1975).

(12) X. Wang, L. Wei, and L. P. Kotra, Bioorg. Med. Chem., 15,

1780 (2007).

(13) D. S. C. Lee and B. W. Griffiths, Clin. Biochem., 18, 261

(1985).

(14) S. Goldstein and A. Duca, Thermochim. Acta, 59, 211 (1982).

(15) M. Giurginca, L. Popa, and T. Zaharescu, Polym. Degrad.

Stab., 82, 463 (2003).

(16) Dupont Elvax resins Product Data Sheet. http://www2.

dupont. com/Elvax/en_US/assets/downloads/.

(17) V. R. Sinha, K. Bansal, R. Kaushik, R. Kumria, and A. Tre-

han, Int. J. Pharm., 278, 1 (2004).

(18) G. Bogoeva-Gaceva, M. Avella, M. Malinconico, A. Buzarovska,

A. Grozdanov, G. Gentile, and M. E. Errico, Polym. Comp., 28,

98 (2007).

(19) R. W. Korsmeyer, R. Gurny, E. Doelker, P. Buri, and N. A. Pep-

pas, J. Pham. Sci., 72, 1189 (1983).

(20) V. Papadopoulou, K. Kosmidis, M. Vlachou, and P. Mach-

eras, Int. J. Pharm., 309, 44 (2006).

(21) J. Siepmann and N. A. Peppas, Adv. Drug Deliver. Rev., 48,

139 (2001).

(22) R. van Dijkhuizen-Radersma, F. L. A. M. A. Peters, N. A. Stien-

stra, D. W. Grijpma, J. Feijen, K. de Groot, and J. M. Beze-

mer, Biomaterials, 23, 1527 (2002).

(23) K. Kosmidis, P. Argyrakis, and P. Macheras, J. Chem. Phys.,

119, 6373 (2003).

Documents

In vitro and mechanism study of poly(ethylene-co-vinyl acetate)-based implant for sustained release of vitamin B12