Solid state stability of digoxin as a function of temperature and humidity

MISSING IMAGE

Material Information

Title:
Solid state stability of digoxin as a function of temperature and humidity
Physical Description:
vii, 63 leaves : ill. ; 29 cm.
Language:
English
Creator:
Kibbe, Arthur H., 1943-
Publication Date:

Subjects

Subjects / Keywords:
Digitalis   ( mesh )
Chemistry, Pharmaceutical   ( mesh )
Cardiac Glycosides   ( mesh )
Pharmacy thesis Ph.D   ( mesh )
Dissertations, Academic -- Pharmacy -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1973.
Bibliography:
Bibliography: leaves 61-62.
Statement of Responsibility:
by Arthur H. Kibbe.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000901173
oclc - 20058215
notis - AEK9998
System ID:
AA00009131:00001


This item is only available as the following downloads:


Full Text




















SOLID STATE STABILITY OF DIGOXIN AS A FUNCTION
OF TEMPERATURE AND IUll DITY






By



Arthur H. Kibbe


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY








UNIVERSITY OF FLORIDA
1973















ACKNOWLEDGEMENTS


Sincere appreciation is expressed to Dr. Oscar E. Araujo, Chairman

of the Supervisory Committee, for his understanding and guidance.

Appreciation is also extended to the members of the Supervisory

Committee, Drs. Charles H. Becker, Richard H. Hammer, Stephen G.

Schulman, and especially to John I. Thornby for his help with the

statistics.

Gratitude is also expressed to Ginny Currin for typing the manu-

script and to Tony Capomacchia for preparing the illustrations.

















TABLE OF c'::frrxE S

Page

ACKNOWLEDGEMi NTS . . .... ii

LIST OF TABLES . . . iv

LIST OF FIGURES . . . vi

ABSTRACT . . . vii

INTRODUCTION . . . 1

EXPERIMENTAL . .. ........... 12

RESULTS AND DISCUSSION . . 18

APPENDIX . . . 28

A. Tables . . 28
B. Figures . . 54

REFERENCES . . . 61

BIOGRAPHICAL SKETCH . . 63











LIST OF TABLES


Table


I. RELATIVE HUMIDITIES EMPLOYED . 29

II. TLC PLATE TO PLATE VARIATION (USING 0.799 MICROGRAP-'
OF DIGOXIN) . . 30

III. THIN LAYER CHROMATOGRAPHY Rf VALUES FOR VARIOUS SOLVENT
SYSTEMS . . . 31

IV. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
230 AND 18.8% RELATIVE HUMIDITY. . ... 32

V. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
230 AND 47.2% RELATIVE HUMIDITY. . ... 33

VI. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
230 AND 80.5% RELATIVE HUMIDITY. . ... 34

VII. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
40 AND 18.8% RELATIVE HUMIDITY. . ... 35

VIII. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
40 AND 47.2% RELATIVE HUMIDITY . ... 36

IX. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
400 AND 80.5% RELATIVE HUMIDITY. . ... 37

X. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
50 AND 18.8% RELATIVE HUMIDITY. . ... 38

XI. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
50 AND 47.2% RELATIVE HUMIDITY . 39

XII. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
50 AND 80.5% RELATIVE HUMIDITY. . ... 40

XIII. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY BURROUGHS
WELLCOME 230 AND 18.8% RELATIVE HUMIDITY . 41

XIV. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY BURROUGHS
WELLCOME 230 AND 47.2% RELATIVE HUMIDITY . 42

XV. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY BURROUGHS
WELLCOME 230 AND 80.5% RELATIVE HUMIDITY . 43


Page










LIST OF TABLES (Continued)


Table Page

XVI. THE DETlL.DAfTION OF DIGOXIN TABLETS SUPPLIED BY BURROUGHS
WELLCOME 40 AND 18.87 RELATIVE !:U1rDITY . 44

XVII. THE P-Gi'.D.-J'ION OF DIOXIN TABLETS SUPPLIED BY BURROUGHS
WELLCOME 400 AND 47.2% }IL.ATT.L IL' DITIY . 45

XVIII. THE DEl,RAD,.\ON OF DIOXIN TABLETS SUPPLIED BY BURROUGHS
v:'ELLCOM. 400 A~.D 80.5% RELATIVE HUMIDITY . 46

XIX. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY BURROUGHS
WELLCOME 50 AND 18.8% RELATIVE HUMIDITY . 47

XX. THE DECilDATION OF DIGOXIN TABLETS SUPPLIED BY BURROUGHS
WELLCOME 500 AND 47.2% RELATIVE HUMIDITY . 48

XXI. THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY BURROUGHS
WELLCO;-M'L 50 AND 80.5% RELATIVE HUMIDITY . 49

XXII. A COMPARISON OF RESIDUAL MEAN SQUARES FOR LINEAR AND
EXPONENTIAL FITS . . 50

XXIII. 3X3X2 FACTORIAL DESIGN OF COEFFICIENTS OF REGRESSION 51

XXIV. ANALYSIS OF VARIANCE OF COEFFICIENTS OF REGRESSION 52

XXV. A COMPARISON OF WEIGHT VARIATION, HARDNESS, AND
DISINTEGRATION RATE OF TABLETS . 53
















LIST OF FIGURES


Figure Page

I. STANDARD CUR'.T FOR U.S.P. ASSAY . 56

II. STANDARD CURVE FOR THIN LAYER CiROMATOGP.."PHi 58

III. ARRHENIUS PLOT OF REGr:. :sION COEFFICIENTS . 60









Abstract of Dissertation Presented to the
Graduate Council of the University of Florida in Partial
Fulfillment of the Requirements for the Degree of Doctor of Philosophy


SOLID STATE STABILITY OF DIGOXIN AS A FUNCTION
OF TL';'ERATURE AND HUMIDITY


By

Arthur H. Kibbe

March, 1973


Chairman: Oscar E. Araujo, Ph.D.

Major Department: Pharmacy



Digoxin in tablet form was subjected to various conditions of tem-

perature and humidity. The results were statistically analyzed using

a 3x3x2 factorial design. A thin layer quantitative assay was developed

for single tablet determinations after a gas chromatographic approach

failed. The results indicated a strong correlation between temperature

and rate of decay as well as humidity and rate of degradation. A high

degree of tablet to tablet variation was noted in addition to a dif-

ference between manufacturers. Tablet hardness, weight, and disinte-

gration time for both manufacturers was checked. A comparison of the

results with solution kinetic theory was made and an explanation for

the inconsistencies attempted.


vii














I ; R.., ACTIONN


Digoxin is a naturally occurring cardiac glycoside employed in medi-

cine for the treatment of congestive heart failure. Cardiac glycosides,

in one form or another, have been used since 1785 when William Withering

published "An Account of the Foxglove and Some of its Medical Uses".

Digoxin was isolated from Digitalis lanata in 1930 by Smith. Its

plant precursor is lanatoside-C or digilamide-C and the digoxin molecule

is formed by loss of a glucose and acetic acid molecule through hydro-

lysis.2

Digoxin is a colorless to white crystalline powder. It is insoluble

in water, chloroform and ether. It is soluble in pyridine and dilute

alcohol.3

The structure of digoxin consists of three components: (1) an a,

t unsaturated lactone, (2) a steroid ring system and (3) three digitoxoses.

OH
CH3






H CHo

1 OH
(Digitoxose)3 3 (










The U.S.P. XVIII monograph states: "Digoxin is a cardiotonic

glycoside obtained from the leaves of Digitalis lanata Ll.rh. (Fam.

Scrophulariaceae). It contains not less than 96% C41H64014 calculated

.on a dried basis. CAUTION- Handle digoxin with exceptional care, since

it is highly potent." The preceding information is about all that is

learned from the official compendium on the chemistry, stability and

pharmacology of one of the most commonly used drugs among people over

fifty years of age.

Digoxin is particularly important because of its nonmetabolic path-

way of elimination, slow excretion and good reabsorption from the ali-

mentary tract. Most individuals under digoxin therapy use tablets manu-

factured and marketed by one of several drug companies. The patients

must maintain the necessary concentration of the drug in the cardiac

tissue or relapse into congestive heart failure. It therefore becomes

obvious that after two to three weeks of therapy with tablets below

labeled potency, a dangerously low level of digoxin in the blood and

heart can occur. These facts make the importance of an accurate daily

dose quite apparent.

Most patients use a single tablet of digoxin daily as a means of

obtaining their medication. The use of tablets as a dosage form can be

traced back to the 16th century. Tablets can be made either by molding

or compression. The United States Pharmacopeia defines a tablet as "a

solid dosage form containing medicinal substances with or without

suitable diluents." On a small scale, they may be prepared by various

molding techniques or by fusion; however, large scale production requires

the use of a tablet compressing machine. To efficiently and uniformly










operate such equipment a ho; )ger.eous mixture of solid particles possessing

suitable physical characteristics is required.5

A well-made compressed tablet is able to withstand the stress

involved in production, packaging, shipment and dispensing. When the

patient receives the tablet, it should be free from obvious defects such

as cracks, chipped edges, discoloration, speaking and contamination.

More importantly, however, it should be reasonably stable to chemical

and physical change in the active ingredient, while maintaining its

ability to release the medicament in a reproducible and predictable man-

ner. All of this should lead to a consistent bioavailability from tab-

let to tablet.

Machines built to compress tablets are of two general types: the

single punch and the multistation rotary press. Each performs the same

task of converting the mixture (called a granulation) of drug, diluents,

binders, lubricants and disintegrants into a tablet. The mixture must

possess the essential characteristics of fluidity and compressibility.

Fluidity is necessary because of the nature of the tabletting machines

which requires the sized granules to flow from a storage hopper through

a feed.frame into the dye that shapes the tablet. If the granulation

does not flow smoothly, vibrations may introduce a serious problem of

stratification or separation of different size particles. This may cause

changes in total tablet weight or uniformity of content or both. The

ideal shape of the granulation particles, therefore, would be spheres

since they would flow smoothly. Granulation is the pharmaceutical process

that attempts to connect powdered materials into aggregates called

granules that are approximate spheres.










Compressibility is the property of forming a stable, compact mass

when pressure is applied. The bonding of particles in a tablet is

probably due to a number of mechanisms. During compression, surfaces

are brought into close proximity by plastic deformation. Mechanical

interlocking, as well as electostatic and Van der Waals forces, combine

to add to the strength of the compressed tablet.

Since granulation increases both fluidity and compressibility,

almost all compressed tablets are formed from granulation produced by

either the dry or wet method.

Diluents are employed as a bulking agent in tablets. Lactose, which

is frequently used, is relatively inexpensive and is available in a coarse

or regular granular size. Starch is also used, as well as the less com-

mon diluents mannitol, microcrystalline cellulose and sucrose.

Binders are substances that cause powders to adhere and form granules.

They can be added dry and then activated with water or added as a slurry

to the other powders. Acacia is a common binder used as a 10 to 20 per

cent solution. Other binders that are used in various per cent solu-

tions of water or water-alcohol are tragacanth, gelatin, sucrose, methyl-

cellulose and polyvinylpyrrolidone.

Once the granulation is made, an additional additive, a lubricant,

is used to improve flow properties and to prevent the tablet from sticking

to the dyes and punches employed in the tabletting process. This is

employed in a fine particle size and can be either a metallic stearate,

high melting point wax or talc.

The last additive, the disintegrant, is one that will cause the

tablet to break apart when placed in an aqueous environment. Most

compounds used for this purpose swell when placed in water. Starch does










not, however, work in this manner since it does not swell at normal

temperatures of gastric fluid. Curlin6 suggests that it acts as a

wick bringing aqueous fluid into the tablet where it dissolves the

binding agent and causes other additives to swell. In addition to

starch, which is the most common disintegrant, gums, cellulose deriva-

tives and alginates may be used.

After granulating and tabletting, the final product is evaluated

through several standard tests. These tests are the basis for guaranteeing

quality control on a batch to batch basis. Some of these tests are

hardness, disintegration time, weight and dissolution rate.

Traditionally, to determine tablet strength or hardness the tablet

was broken between the thumb and first and second finger. If there was

a sharp snap the tablet was satisfactory. Several companies have since

developed hardness testers, all of which actually measure resistance to

crushing. There is, however, no correlation between the scales of each

tester.

A commonly used apparatus for determining tablet hardness is a

Monsanto tester which consists of a barrel containing a compressible

spring held between two plungers. As the spring is compressed by turning

the threaded bolt, a pointer rides along a gauge in the barrel and indi-

cates the pressure at which the tablet fractures.

The U.S.P. tablet disintegration test is specific for the type of

tablet being tested. It consists of an apparatus that moves the tablet

up and down through a test fluid maintained at a temperature between 350

and 39 at a rate of 28 to 32 cycles per minute with a stroke distance

of from 5 to 6 cm. per cycle. The in vitro procedure does not actually

simulate physiologic conditions. It is used, rather, as a means of









quality control to insure product uniformity. The test does not indi-

cate complete solution of the tablet or evn dissolution of its active

constituents.

Wagner Wurster and Taylor and Pernarowski, Woo and Searl ll

have worked on apparatuses for dissolution rate studies. The U.S.P.

has since adopted an official device for this purpose. The tablet is

placed in a test medium which is sampled at various time intervals and

thus the rate at which the active ingredient goes into solution is

determined. The test medium approximates the gastrointestinal fluid

and the dissolution medium is maintained at 370 so as to closely parallel

the in vivo process.

The last several revisions of the U.S.P. have included a require-

ment for weight variation tolerance. These requirements are generous

in that they allow a 10 per cent difference in two tablets out of 20,

for tablets weighing 130 mg. or less. Until recently, the assay of drug

content of tablets involved the grinding of a large sample followed by

analysis of an aliquot. Results obtained were then expressed on an

individual tablet basis. Efforts to overcome this problem brought the

content uniformity test described in the U.S.P.:

Where a requirement of Content uniformity is specified
in the individual monograph, select a representative
sample of 30 tablets. Assay 10 of these individually
as directed in the Assay in the monograph. If the
amount of drug in a single tablet is less than that
required in the assay procedure, the degree of dilu-
tion of the solutions and/or the volume of aliquots may
be adjusted so that the concentration of the drug in
the final solution will be of the same order as that
obtained in the Assay provided in the monograph. The
requirements of this test are met if all 10 results
fall within the limits of 85 per cent and 115 per cent
of the average of the tolerances specified in the
respective monograph. If 1, but not more than 1.











result falls outside these limits, assay the
remaining 20 tablets individually. The require-
ments are met if not more than I of the 30 results
is outside the limits of 85 per cent and 115 per
cent.

From the preceding information, one can see that the formulation

and manufacture of tablets is a complex process and thus is liable for

some serious errors. This is especially true for tablets that contain

a low percentage of active ingredient, such as digoxin tablets. A small

change in the overall process might affect the properties of the active

ingredient or its distribution between the tablets to an inordinate

degree.

Various colorimetric assays for digoxin have been reported. Alka-

line picrate was used by Baljet xanthydrol by Pesez and 3,5 dini-
12
trobenzoic acid by Tattje.

An automated assay of single tablets of digoxin which depend on

chemical reactions such as the oxidation of the terminal sugar of the
13
glycoside into a malonic dialdehyde was proposed by both Khoury and
14
Myrick. A fluorometric micromethod for simultaneous determination of

digitoxin and digoxin was proposed by Jakovljevic while an automated

fluorometric procedure was proposed by Cullen, Packman and Papariello.16

These methods involve reactions, which can also occur with other

glycosides and mono- or di-digitoxose aglycons and therefore render

the assays non-specific.

Thin layer chromatography combined with colorimetric reactions,

as well as gas chromatographic assay procedures, have been reported.

They are all based on the theories of chromatography on thin layers of

adsorbent which were conceived as early as 1938 and developed largely










during the early 1950's. Chromatography was developed because of a

specific need for a rapid r-thod of separating minute amounts of com-

pounds. It was not limited to colored compounds because of the possi-

bility of carrying out certain reactions which would make otherwise

colorless compounds visible in either normal or ultraviolet light.

Chromatography can be considered from three viewpoints. One is

qualitative, the second is preparative and the third is quantitative.

All of the techniques are based upon a single simple principle involving

a moving system of some type of gas or liquid in equilibrium with a

stationary phase. When a solid is used as the stationary phase and the

substances being separated are adsorbed onto it, the method is called

adsorption chromatography. If, however, the stationary phase is a

liquid or gas held on some type of support, it is partition chromatography.

Gas chromatography can be of either type. The advantages of gas

chromatography are its speed of operation, its high degree of resolution

and the fact that it can yield quantitative results. Its major disadvan-

tage, other than high cost, is that the substances to be separated must

have at least some vapor pressure at a workable temperature.

An ideal method of chromatography should lend itself to quantitative

interpretation and should be technically simple in design. Both of these

conditions are fulfilled by thin layer chromatography. The development

time is shorter than that for paper chromatography and a number of pro-

cedures have been published for quantitative assay.

Quantitative evaluation of thin layer chromatograms fall into two

general categories. In one case, a mixture is separated on the thin

layer and then eluted for measurement by a spectrophotometric or colorimetric

method. The second and simpler method involves the measurement of spot










area and these values are then related to substance amount in some

manner.

The spot area analysis is based on a mathematical relationship

between the spot area and the weight of a given substance. The method

is widely applicable and involves only a small number of mechanical

operations. 17-23 Seher obtained results with only 5 per cent error

using calibration curves obtained by plotting spot area against weight.24

Brenner and Niederwieser obtained a linear relationship using a plot
25
of the logarithm of the weight of sample against its spot area. The

most complete study involving the technique was made by Purdy and Truter,

who found a linear relationship using the square root of the spot area

vs. the logarithm of its weight. They developed a simple algebraic

relationship which for 540 observations yielded a standard deviation

of 2.7 per cent.26'27

Eric Watson and S. Kalman assayed plasma levels of digoxin by gas

chromatography. Their assay procedure took five hours per assay and
28
involved both thin layer and gas chromatography steps. All previous

work on gas chromatography of digoxin had only been of a qualitative

nature. Watson used electron capture detection while Wilson et al.

used silyl ethers of cardenolides which required high temperature and
29
high flow rates of the carrier gas. A recent study by Tan questioned

the formation of the silyl ethers reported by Wilson and others.30

Separation of digoxin from the other glycosides was achieved by

Svendsen and Jensen.31 Thin layer chromatography on digoxin was also

employed by Stahl, who used methylene chloride-methanol-formamide as

the solvent system.32 Heusser proposed a method based on thin layer









chromatography separation of the glycoside and subsequent colorimetric

determination using xanthydrol.

Michalska, Zurkowaka and Ozarowski studied the stability of digi-

toxin at room temperature, 40, and A7'. They reported, however, that

their determinations of total glycosides do not correlate with the actual

changes in therapeutic value of the drug. They further stated that the

total glycoside content stayed constant over a period of seven months

at elevated temperatures but that there was a decrease in digitoxin con-

tent. However, they omitted any discussion concerning the interconver-

sions which would account for the decrease in digitoxin.34,35

Grollman states that all digitalis preparations are indicated for

any of the forms of congestive heart failure irrespective of type of

rhythm.36 The action of the glycoside in treating atrial fibrillation

is due to the specific effect of the strength of contraction and on

conduction of the cardiac muscle. Digoxin has an onset of action of

two hours if taken orally and of five minutes if given intravenously,

with an average time of maximal effect at eight hours after oral and

five hours after intravenous administration respectively. The duration

of effect is from four to seven hours. The myocardium displays no special

affinity for the cardiac glycosides which are present in highest concen-

tration in the kidney. The glycosides are bound reversibly to the plasma
37
albumin. A normally digitalized man excretes about 32 to 44 mcg. daily.

Doherty, Hall, Murphy and Beard presented a critical review of

digitalis metabolism. They reported that digoxin is 80 to 90 per cent

absorbed, recycled to a small extent, only slightly protein-bound, poorly

metabolized and excreted largely unchanged in the urine. Consequently,

its half-life of 1.5 days is shorter than digitoxin and its excretion

may be directly related to creatinine clearance.










Cattel et al. state that digitoxin is 100 per cent absorbed.39

The variation of this degree of absorption from the 80-90 per cent

attributed to digoxin by Doh rty can best be explained by the slight

difference in structure of the two compounds, resulting in a change in

polarity.40 This difference was noted to have a large effect on the
41
half-life since digitoxin is recycled and digoxin is not.

Doherty and Perkins surgically induced biliary fistula in human

subjects to prevent recycling. When digoxin was given orally to these

subjects as well as to another group of normal humans, there appeared

to be no difference in the serum half-life.42

Doherty recommended the following dosage regimEn for digoxin:

One to 2.5 mg. as a single dose orally in a 24-hour period with a

maintenance dose of from 0.125 to 0.75 daily based on renal function,

and organ response, absorption and excretion.43

The present study proposed to examine the effects of humidity

and temperature on possible solid state degradation of different commer-

cial tablets of digoxin, as well as to investigate tablet to tablet

variation for digoxin. Where the U.S.P. assay calls for the use of 20

tablets to give an average content, individual tablet assay will be

carried out in this study. In this manner an indication of the com-

bined effect of these factors on tablet degradation can be ascertained.

















Digoxin tablets manufactured by two companies were used through-

out this study: 1000 tablets (0.25 mg.) from Purepac Pharmaceutical

Company, Division of Elizabeth Labs, Elizabeth, New Jersey, Lot #0192F3

and 1000 tablets (0.25 mg.) from Burroughs Wellcome Company, Research

Triangle Park, North Carolina, Lot #196E. The tablets from each company

were divided at random into groups of 100 and either used as zero time

samples or placed in the containers for exposure to the various tempera-

tures and relative humidities used in the degradation studies.

Desiccators containing sulfuric acid-water solutions provided the

atmospheres for the three relative humidities employed, as shown in

Table I. 44

The desiccators for each relative humidity were stored at three

different temperatures, 230, 40 and 500. The ovens used were Thelco,

Model 6M, manufactured by Precision Scientific. The temperature was

controlled to within one degree.


Assay Procedures

Gas-Liquid Chromatography

An attempt was made to quantify the work done by Wilson et al.

using a Varian 2100 gas chromatograph equipped with a Hydrogen Flame

Detector. The support was supplied by Applied Science complete with a

precoating of 2.5 per cent OV-1 on Chromsorb W 80-100 mesh. The sil-

anizing reagent was prepared just prior to use by mixing 1 ml. of trimethyl


r 7t 'lI:T1;TAL










chlorosilane and 10 ml. of hexamnethyldisilazane with 10 ml. of dry

pyridine as a solvent. The pyridine was stored over potassium hydro-

xide to insure dryness. The he:amethyldisilazane and trimethyl chloro-

silane were stored under nitrogen and extracted from sealed vials with

a syringe. The volume was then replaced with dry nitrogen gas. The

sample to be silanized was made to react with 1 ml. of the above reagent

at room temperature before being injected into the gas chromatograph.

Initially glass columns 6 feet in length with 2 mm. ID were used.

Due to the high temperature of operation suggested by Wilson, difficulty

was experienced in maintaining a firm seal between the glass column and

the metal connections. The Teflon ferrel, which normally would be used

for this purpose, began to melt at 3000 to 350* temperature at which

the injector and detector were operated. Red rubber "0" rings supplied

by Applied Science were used to replace the Teflon ferrels. They, how-

ever, deteriorate rapidly at the high temperature.

To alleviate this problem the use of copper columns having a 4 mm.

ID but only a 1 1/2 foot length was employed. This enabled the use of

swage lock fittings which prevented leaking and slipping. Nitrogen

was employed as a carrier gas in all the experiments and the rate of

flow was varied from between 50 ml/min. to 100 ml/min.

The genins, as well as digoxin itself, were silanated separately

and in mixture. Different concentrations of each were injected in vol-

umes of between 2 to 5 microliters. After repeated injection over a

period of days, a high noise level developed which required the column

packing to be changed. This was very time consuming and introduced varia-

tions in the results since no two columns are exactly alike.










U.S.P. Assay

The U.S.P. assay was performed on groups of 20 tablets at the

beginning and at the end of the stability studies.

Alkaline dinitrobenzene solution---Mix 1 ml. of
tetramethylammonium hydroxide T.S. with 140 ml. of
dehydrated alcohol, titrate a portion of the solu-
tion with 0.01 N hydrochloric acid, and adjust the
remaining solution to a concentration of 0.008 N
by dilution with dehydrated alcohol. Immediately
before use, mix 40 ml. of this solution with 60 ml.
of a 1 in 20 solution on m-dinitrobenzene in
benzene.
Standard preparation---Dissolve 25.0 mg. of U.S.P.
Digoxin Reference Standard, previously dried in
vacuum at 105 for 1 hour, in 50 ml. of hot alco-
hol, cool, add alcohol to make 100.0 ml., and mix.
Dilute 10.0 ml. of this solution with alcohol to
100.0 ml.
Assay preparation---Weigh accurately about 25 mg.
of Digoxin, and prepare solutions as directed under
Standard .preparation.
Procedure--- Pipet 5 ml. each of the Assay prepara-
tion and the Standard preparation into separate,
small conical flasks, and treat each solution as
follows: Evaporate on a steam bath, with the aid
of a current of air, to dryness, and cool the resi-
due in a vacuum desiccator for 15 minutes. Add
5.0 ml. of Alkaline dinitrobenzene solution, and
allow to stand, with frequent swirling, for 5 min-
utes at a temperature not exceeding 30. Determine
the absorbance of the solution, relative to a reagent
blank, in a 1-cm. cell at 620 mu with a suitable
spectrophotometer, repeating the measurement at 1-
minute intervals until maximum absorbance is
obtained.
Calculate the quantity, in mg., of C41H64014 in
the portion of Digoxin taken by the formula 25(Au/As),
in which Au and As are the maximum absorbances of
the solutions from the Assay preparation and Stan-
dard preparation, respectively.

The absorbance was measured on a Fischer colorimeter. One possible

mechanism which could lead to the development of the color in the U.S.P.

assay is an electron transfer from the digoxin to the m-dinitrobenzene,

which lowers its absorbance energy from 300 mu to 620 mp. In order for










digoxin to transfer this electron, a ch-: in structure nust be caused

by the tetramethyammonium hydroxide. The probable steps in this reaction

are shown in the following scheme.
O


R -- R R




The reason for choosing this scheme over one that would involve

attachment of the m-dinitrobenzene to the opened lactone ring is that

the added conjugation would not lower the energy of the system enough

to make it visible. An electron transfer, however, followed by a com-

plexation step would lower the energy enough to explain the large change

in absorbance wave length.

A standard curve was produced using five different concentrations

of digoxin in alcoholic solution. They were treated according to the

U.S.P. method and the results appear in Figure I.


Thin Layer Chromatography

Thin layer chromatography was performed on Adsorbosil-5-Prekotes

supplied by Applied Science, College Park, Pennsylvania. The solvents

were supplied by Matheson, Coleman and Bell, Atlanta, Georgia. Various

grades of solvents were used. Pyridine, cyclohexane, acetone and acetic

anhydride were Spectra grade, while the ethanol was U.S.P. 95 per cent.

All other solvents were Technical grade. Pure powdered digoxin was pur-

chased from K & K Labs, Plainview, N.Y., Lot #15614, as were the pure

powdered samples of the aglycons. All powdered samples were tested for

purity using thin layer chromatography.










The samples were assayed quantitatively by the following procedure:

A digoxin tablet was removed from the desiccator and placed in a 5 mm.

test tube. Fifty microliters of distilled water were then added to the

test tube and allowed to stand for 30 minutes. At this point 0.5 ml.

of pyridine was added, and the mixture agitated for 30 minutes at 5

minute intervals. A 2 microliter aliquot was removed from the clear

solution and spotted on a plate. An additional 0.5 ml. aliquot of

pyridine was then added, mixed, and a second 2 microliter portion was

spotted on the same plate. A 2 microliter aliquot from a known stan-

dard solution was also spotted. Its concentration was estimated to be

between that of the two unknown aliquots. Step by step the assay pro-

cedure is as follows:

Tablet

1) Add 5 microliters of water

2) stand for 30 minutes

3) add 0.5 ml. pyridine

4) Mix for 30 min. 5) take
2 microliter sample> 6) spot on TLC plate

7) add second 0.5 ml. pyridine

8) Mix 9) take
second 2 microliter sample 10) spot on TLC plate

The plate was then developed in glass unlined tanks using a mix-

ture of cyclohexane, acetone and acetic acid (49-49-2). The plate was

removed after the solvent front had migrated the full distance of the

plate and dried before visualization, employing the spray reagent.

A spray consisting of a mixture of a 3 per cent aqueous solution

of chloramine T and a 25 per cent alcoholic solution of trichloroacetic

acid was prepared. The two solutions, in a ratio of 1:4, were mixed










just prior to use. After spraying, the plate was heated in an oven at

1100 for seven minutes and a blue spot was observed under UV light

(385 my). The chemical reaction which could possibly explain the genera-

tion of the fluorescence is the oxidation of the lactone ring to a 4i-

ketone which when protonated will fluoresce.

The color development for both the U.S.P. assay and the thin

layer visualization are non-specific reactions for any compound which

has a lactone ring. The specificity of the thin layer procedure arises

from the separation step preceding the quantification.

During the course of the stability studies, single tablets were

taken at random from a group of tablets being exposed to each of the

nine sets of conditions. The assay described in the discussion of

thin layer chromatography was then performed and the results tabulated

for analysis. The data was analyzed for linear and exponential fit,

using two programs currently used by the University of Florida Com-

puter Center. Further statistical calculations were carried out on

a standard desk calculator supplied by Monroe. An analysis of variance,

according to a 2x3x3 factorial design, was performed on the linear

regression coefficients. This analysis demonstrates the effect of tem-

perature, humidity and method of manufacture of the digoxin tablets.














RESULTS AND DISCUSSION


An exhaustive attempt to reproduce the gas chromatographic assay

proposed by Wilson was made. It was possible only to obtain good

results for the pure genin. The glycosides showed no detectable peak

under any of the conditions used. This may have been due to the low

temperatures at which digoxin has been reported to degrade, as well as

the high vapor pressure produced by the high molecular weight compound.

At low oven temperatures, the compound moves so slowly that it just

gives rise to base line drift and noise, while at high temperatures,

it fragments and comes off as small pieces which has the same effect

as base line drift.

It was possible, however, to identify the pure genin (aglycone).

To quantify the amount of each of the glycosides (digoxin, monodigitoxose,

digoxigenin and bis-digitoxose digoxigenin) would require a hydrolysis

step following either thin layer or column chromatography. The pure

genin could then have been assayed and the amount of parent compound

calculated. This method would require many time consuming manipulative

steps and probably introduce unnecessary errors. As a result, it was

decided to rely on quantitative thin layer chromatography as the main

method of assay.

The twenty tablets, assayed by the U.S.P. method before the start

of the stability study, showed an average content of 0.25 mg. of digoxin

per tablet. When the assay was repeated at the end of the study on 20

tablets taken from each of the nine conditions to which they were










subjected they again showed an average content of 0.25 mg. per tablet.

The fact that no change in total glycoside content was evident is a

result of the type of assay used. Since the assay depends on a reaction

that occurs at the lactone ring, the presence of one, two or three

sugars on the other end of the steroid has no effect. It is possible,

therefore, to assay a group of tablets which have been completely hydro-

lyzed to the pure genin or partly degraded through removal of one or

two digitoxose sugars and still obtain results in terms of the parent

compound. For this reason the U.S.P. assay could not be used as a means

for determination of the amount of degradation which occurred in the tablet.

It is true that the U.S.P. assay could be adjusted to evaluate a single

tablet, but only in terms of total glycoside content, not in terms of

the four possible compounds of interest in this study.


Thin Layer Chromatography

Figure 2 shows a plot of the logarithm of the weight of digoxin spotted

versus the square root of the area of the developed spot for a series of

solutions. Eleven different solutions were prepared by dilution of a stock

solution and randomly spotted in duplicate (two microliter spots on six dif-

ferent plates). No attempt was made to prepare even numbers of replicates

as the graph was not intended to be used as a standard curve but only as a

test for linearity. The scatter about the line can be partially attributed

to a plate to plate variation. Each silica gel plate used in the thin layer

chromatography assay cannot be prepared in exactly the same manner, which af-

fects the area covered by a given weight of digoxin. Table II shows a com-

parison of six plates chosen at random. An analysis of variation was done

using the plate as the treatment. The large F shows that there is a










statistically significant effect of the plate on the size of a spot

from a given amount of drug. Consequently, all weights of unknown sam-

ples were calculated mathematically by an area comparison between the

unknown and a standard solution of digoxin spotted on the same plate.

This procedure eliminates any plate to plate variation which would

have otherwise occurred.

The weight of the unknown was calculated using the following for-

mula: Log W = Log Ws + [(w 7~s)/(7 1d)](log d)

Where W = weight of unknown sample

Ws = weight of standard sample

A = spot area of the unknown

A = spot area of.the standard
s
Ad = spot area of the diluted unknown sample

d = dilution factor

A straight line relationship between log W and fX has been demon-

strated to be true using a series of standard solutions (Figure 2).

Assuming coordinate C(r, log W), (Vs log Ws), and (d', log W/d)

to be three points on the line, the slope can be calculated as

(log W Log Ws)/f '7s) or (Log W Log W/d)/(TA 7d).

Equating the two:

Log W Log W Log W Log W/d


Combining the terms:

Log W Logog W, = Log d

s d
Rearranging:

Log W Log Ws = [( As )/(Qf- 9 )](log d)
S d










Finally:

Log W = Log Ws + [( ?s/A 7Ad)](log d)

The solvent system was chosen on the basis of the difference in Rf

values between digoxin and its genins. This would allow easy identifi-

cation of simple hydrolysis products since these would show up higher

on the plate than digoxin. The pure sugar digitoxose was found to appear

well below the digoxin spot. Other components of the tablet such as talc,

starch and stearic acid were found to remain behind at the original

point of spotting and were not considered as interfering with the assay.

Table III shows the Rf values for the various solvent systems. Based

on these findings, a combination of cyclohexane-acetone-acetic acid

(49-49-2) was chosen as the solvent system with a one-half hour drying

time between two successive runs. This procedure gave maximum separation.

The degradation studies were carried out by assaying a tablet

taken from each desiccator at seven-day intervals. The results obtained

along with their analysis of variance and test for significance are

shown in Tables IV-XXI. The degradation data was analyzed by the use

of two programs already in the computer. These programs treated the

data first as a linear model and then as an exponential one. A compari-

son of the residual mean squares of the two fits shows that the linear

model values are all smaller than those of the exponential one indi-

cating that the latter is not better and probably worse than the linear

fit. On this basis the simpler model was chosen. The comparison of

the mean squares is shown in Table XXII.

It is not to be inferred from the above discussion that in some

instances a non-linear component does not exist. It is not, however,










exponential in nature. The choice of the linear model was based, there-

fore, on the fact that it would provide a reasonably close approximation

of the degradation kinetics at play. The coefficient of regression was

chosen as the statistical parameter for the evaluation of the temperature,

humidity, and manufacturer effects.

The various coefficients of regression were then treated using

3x3x2 factorial design shown in Table XXIII. An analysis of variance

was performed and the results appear in Table XXIV. The data clearly

indicates a significant effect due to temperature and humidity and a

slight effect due to manufacturing differences between tablets.

The results of three control routine tests performed on the manu-

factured tablets are shown in Table XXV. They indicate the hardness,

weight and tablet to tablet weight variation and the disintegration

times of a sample of each of the groups of tablets used in the stability

studies. Note that the Purepac brand is harder and disintegrates

slower than the Burroughs Wellcome brand, as well as the weight, on

the average, being ten milligrams more. There is also more variation

in tablet weight and hardness in the Purepac brand than with the

Burroughs Wellcome brand as indicated by the large standard deviation

for both of these parameters. The results may be explained in a number

of ways. It is most probable that the Purepac Company intended to pro-

duce a harder tablet so that fewer tablets would be lost due to breakage

during manufacture, packaging and storage. This made it necessary to

add a stronger binder or more binder in the base formulation and sub-

jected the tablet to a possibly greater pressure during the compression

cycle of the tabletting operation. Purepac has also succeeded in obtaining










a good disintegration rate which is probably due to a disintegrant

in the formulation. These two changes possibly explain the added ten

milligrams of weight of the average tablet. Variation in granulation

particle size and improper classification of the granulation before

compression would lead to a larger weight variation from tablet to tab-

let. These slight changes in total formulation would show up later in

the stability studies as a significant effect on overall resistance to

hydrolysis.

As can be seen in Table IV and Table XIII, the F ratio is low.

This signifies that the coefficient of regression for these two sets

of conditions cannot be said, with any confidence, to vary from zero.

On the other hand, a comparison of the F ratio for more strenuous con-

ditions of temperature and humidity, as seen in Tables XII and XXI,

shows that a definite slope is apparent. The first fact is a result

of the high tablet to tablet variation as well as the very slight rate

of degradation of digoxin. When all of the regression coefficients

are considered together, ignoring their relative F ratios, a trend is

apparent that justifies the overall analysis of variance that was per-

formed.

In Table XXIV the effect of humidity showed the largest F ratio

which would indicate that it is the single most important effect. From

this fact, it could be deduced that in the absence of moisture, no

reaction would occur and that the primary reaction being considered

here must involve water in the transition state. The large positive F

for temperature indicates an increase in reaction rate with an increase

in temperature as would be expected in any reaction which must overcome

an energy peak as postulated by Arrhenius. However, when an attempt










was made to fit the data for the coefficients of regression on a stan-

dard Arrhenius plot, it appears to hold only at the lowest humidity.

As the humidity is increased, the linearity of the plot is lost and a

curve results. The high F ratio for hie cross term, humidity x tem-

perature, gives an indication that this would occur. This might be

accounted for by the effect of humidity on the apparent energy of

activation. The data indicates that the higher the humidity, the

higher the apparent energy of activation (Figure 3). This fact might

at first glance seem opposite to what one would expect, since the in-

crease in humidity should in all probability increase the rate but not

change the energy of activation. If indeed the same energy state or

transition state is in operation in all cases, the increase in humidity

while affecting the rate should not in any way affect the slope of the

Arrhenius plot. As indicated, however, the results are not in accord

with this line of reasoning. The most obvious explanation appears to

be that the parameters actually being measured are two different energy

pathways and that the rate determining step is shifting from the one

dependent on the transfer of water through the tablet to the one depen-

dent on the reaction time of digoxin with water. One could look at the

overall reaction as involving two steps. The first is the transfer of

water from the atmosphere to the site of hydrolysis and the second

being the actual hydrolysis itself due to moisture present. This is to

say that at low humidity the adsorption and diffusion effects are regulating

factors, while at higher humidities most sites for reaction are saturated

with water, and the above factors no longer predominate. This can be

visualized in the following scheme:









Water k> water & digoxin k2 genin & digitoxose

in which kI is a constant of diffusion and k2 is the rate constant for

hydrolysis. Accordingly, if the slope of the Arrhenius plot is examined

in Figure 3 for the 18.8 per cent relative humidity conditions, an

energy of activation is obtained of approximately 4 kilocalories/mole

while the shift in slope for the 80.5 per cent relative humidity resulted

in a calculated energy of activation of 13 kilocalories/mole.

Any tablet, regardless of how well formulated and manufactured,

cannot be as uniform in solid form concentration as in solution. Tab-

lets are made by compressing granules of a drug together with inactive

ingredients, while solutions allow for dissolution of the drug on the

molecular level. Crystal size of an active drug will affect the amount

of surface exposed to the moisture and thereby its chances for hydro-

lysis. The various mixing and other processes involved in granulation

and tablet preparation are bound to affect the uniformity of drug con-

centration across the tablet. A granulation, whether it is prepared

by the slugging process or by the wet process, will contain fine and

coarse particles which are liable to stratify. Furthermore, migration

in a tablet granulation can occur through routine handling and if the

active ingredient happens to be concentrated in one particular area,

then, upon compression, irregularities in the amount 'of active ingre-

dient will occur. These factors may be largely responsible for tablet

to tablet variation which causes each tablet to contain a different

initial amount of drug. It is, therefore, difficult or impossible to

establish a zero time level. Each tablet acts as its own reaction

vessel and the nature of that vessel changes from tablet to tablet.

Porosity, hardness and total weight thus contribute to creating a









different environment for reaction. Also, the change in the amount of

various additives, such as starch, can change the rate at which the

humidity in the atmosphere can migrate through the tablet to the site of

hydrolysis.

Most manufacturers are conscious of the problems involved with

possible concentration variation ariong tablets. The fact remains that

even with the greatest of manufacturing care, no two tablets can be

exactly alike. This fact is reflected in the large scatter about the

regression in each of the sets of data for the various conditions. It

also shows up in Table XXIV as a significant F ratio for manufacturer

effect.

To analyze the results in terms of solution kinetics would over-

simplify the situation. In solution kinetics, factors such as pH, con-

centration, temperature and time can all be measured with a great

degree of accuracy. In a tablet there is a given amount of the drug

per total weight of tablet but this does not really qualify as an

absolute concentration. The active ingredient is not dispersed uni-

formly throughout the tablet as it would be in a solution. It is in

pockets of pure crystal and reacts as a pure crystal. If the hydrolysis

is assumed to require dissolution on the drug into a multimolecular

layer of water adsorbed on the surface of the crystal, then the solu-

bility of the drug will affect the overall rate and the scheme must be

expanded to include solution rate.

In solution kinetics, there can be an optimum pH for stability of

the drug. In solid state systems, the pH or hydrogen ion concentration

at the site of reaction is fixed by the solubility of the drug and the










additives adjacent to it. 'inerefore, in solid state kinetics, the pH

at the site of reaction can vary from tablet to tablet and even within

one tablet.

The humidity can be considered to be directly proportional to

the amount of water at the site of reaction, only as long as diffusion

through the tablet is a rate limiting step. Thus at low humidities,

the less porous the tablet, the slower the rate of reaction. The type

of "inert" ingredient may either slow or speed the passage of water

through the tablet. If it is hydrophobic in nature, it can retard

degradation, whereas if it is hydrophilic, it might increase the rate

of degradation.

As a result of this study, there is a need for continuous monitoring

on a tablet to tablet basis of lots from all the manufacturers of potent

pharmaceuticals. It is not possible to make the normal prediction of

stability based on the Arrhenius relationship which only applies to

solution kinetics. The need now exists for the development of theory

and methodology that can adequately describe the type of reaction which

occurs in most tablets on the market, so as to make valid long range

stability predictions based on current data.


































APPENDIX

A. Tables














TABLE I

RELATIVE IJ:1DITIES EMPLOYED


Relative Humidity

80.5%

47.2%

18.8%


Density of Solution

1.20

1.35

1.50


W/W % Sulfuric Acid

28%

45%

60%











TABLE II

TLC PLATE TO PLATE VARIATION
(USING 0.799 MICROGRAMS OF DIGOXIN)


2

0.1581

0.1732

0.1868

0.1732

0.1837

0.1658


1.0408

0.18112

0.1805

0.1735


3

0.1541

0.1541

0.1616

0.1581

0.1581

0.1498


6

0.9319

0.14479

0.1447

0.1553


4

0.1658

0.1936

0.1694

0.1541

0.1803

0.1904


6

1.0536

0.18617

0.1850

0.1750


5

0.1904

0.2208

0.1935

0.2000

0.2150

0.2398


6

1.2595

0.26617

0.2644

0.2099


Source

Among plates

Within plates

Total


Analysis of Variance

Sum of Squares Mean Square

0.01684 0.00336

0.00486 0.00016

0.02170


F Ratio

21.00***


*** Significant at the 0.1% level (P<.001)


1

0.1457

0.1412

0.1500

0.1500

0.1541

0.1411


0.8821

0.12982

0.1297

0.1470


r


Sum of X

Sum of X2

(Sum of X)2
6
X


Total















36

6.3213

1.13166


6

0.2179

0.1969

0.1837

0.1968

0.1801

0.1801


6

1.1555

0.22359

0.2225

0.1926









TABLE III

THIX; LAYER CHROMAI,'GRAPHY Rf VALUES
FOR VARIOUS SOLVENT SYSTEMS


Rf Values

D D1 D2


Ethyl Acetate

Ethyl Acetate-
Pyridine-Water
(50-10-40)

Cyclohaxane-Acetone-
Acetic Acid
(49-49-2)

Same as above but run
twice with drying between

Ethyl Acetate-
Methanol-Water
(90-5-5)

Benzene-Ether
(70-30)

Benzene-Ether-Ethanol
(70-30-20)

Benzene-Ethanol
(82-18)

Benzene-Ethanol
(70-30)

Benzene-Ethanol
(82-23)

Ethyl Acetate-
Methanol-Water
(16-1-1)

Ethyl Acetate-Methanol
(80-10)


Key: D = digoxigenin; D1
bis-digitoxose.


0.3708 0.3708 0.1685

0.8108 0.8784 0.8243


0.3537



0.7381


0.7561



0


0.041


0.2812


0.9354


0.6571


0.6969



0.8286


0.3049



0.6548


0.7560



0


0.041


0.2187


0.8871


0.6428


0.6666



0.8714


0.2805



0.5595


0.6585



0


0.0137


0.2031


0.7903


0.6143


0.5909



0.8570


Digoxin

0.1348

0.8108



0.2561



0.5000


0.6219



0


0.0137


0.1875


0.7742


0.6000


0.5606



0.8429


= digoxigenin mono-digitoxose; D = digoxigenin
2


Solvent System
















TABLE IV

THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
230 AND 18.8% RELATIVE HUMIDITY


Time in Days

39

50

56

70

77

84


Micrograms of Digoxin

0.995

1.240

1.110

1.950

1.160

0.996


Time in Days

91

98

105

114

120


Micrograms of Digoxin

1.580

1.240

1.240

0.951

0.651


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.076 0.076

1.092 0.121


Variables in Equation

Coefficient -0.00328

Standard Error 0.00415

F to Remove 0.6271


F Ratio

0.627














TABLE V

THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
230 AND 47.2% RELATIVE HUMIDITY


Time in Days

38

50

56

63

70

77


Micrograms of Digoxin

1.020

0.992

1.310

1.200

1.540

1.260


Time in Days

84

91

98

105

114

120


Micrograms of Digoxin

1.150

0.961

0.618

1.020

0.595

0.612


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.379 0.379

0.595 0.059


Variables in Equation

Coefficient -0.00712


Standard Error

F to Remove


0.00282

6.3753


F Ratio

6.375














TABLE VI

THE DEGRADATION OF DIGOXED TABLETS SUPPLIED BY PUREPAC
230 AND 80.5 U'i.:'li'IVE HUMIDITY


Time in Days

39

50

56

63

70

77


Micrograms of Digoxin

1.080

1.550

1.190

1.600

1.590

0.976


Time in Days

84

91

98

105

114

120


Micrograms of Digoxin

1.110

0.802

1.040

0.803

1.090

0.581


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.501 0.501

0.662 0.066


Variables in Equation

Coefficient -0.00823

Standard Error 0.00299

F to Remove 7.5584


F Ratio

7.558













TABLE VII

THE DEGRADATI::i OF DIC..L ': T;A T L SUPPLIED BY PUREPAC
400 AND 18.8% RELATIVE HUMIDITY


Time in Days

38

45

50

56

63

70

77


Micrograms of Digoxin

0.745

1.190

0.804

1.310

1.100

1.030

0.894


Time in Days

84

91

98

105

114

120


Micrograms of Digoxin

1.110

0.966

0.854

0.899

0.502

0.434


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.250 0.250

0.518 0.047


Variables in Equation

Coefficient -0.00537

Standard Error 0.00233

F to Remove 5.3040


F Ratio

5.304













TABLE VIII

THE DEGRADATION OF DIGOXIN TABF.LFfS SUPPLIED BY PUREPAC
400 AND 47.2% RELATIVE HUMIDITY


Time in Days

38

45

50

56

63

70

77


Micrograms of Digoxin

1.350

1.024

1.590

1.240

1.410

1.110

.0.990


Time in Days

84

91

98

105

114

120


Micrograms of Digoxin

1.060

1.028

1.600

0.317

0.634

0.725


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.630 0.630

1.030 0.094


Variables in Equation

Coefficient -0.00854

Standard Error 0.00329

F to Remove 6.7352


F Ratio

6.735














TABLE IX

THE DEGIPAD.TION OF DITO.1:.. TABLETS SUPPLIED BY PUREPAC
40 AND 80.52 RELATIVE HUMIDITY


Time in Days

38

45

50

56

63

70

77


Micrograms of Digoxin

1.508

0.988

1.150

1.360

1.180

1.050

0.900


Time in Days

84

91

98

105

114

120


Micrograms of Digoxin

1.400

0.877

0.854

0.888

0.498

0.513


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.716 0.716

0.435 0.040


Variables in Equation

Coefficient -0.00911

Standard Error 0.00214

F to Remove 18.1205


F Ratio

18.121















TABLE X

THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
500 AND 18.8% RELATIVE HUMIDITY


Time in Days

43

50

56

63

70

77


Micrograms of Digoxin

0.707

0.799

1.260

1.690

1.590

1.050


Time in Days

84

98

105

114

120


Micrograms of Digoxin

1.050

1.400

0.274

0.583

0.632


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.314 0.314

1.705 0.189


Variables in Equation

Coefficient -0.00672

Standard Error 0.00522

F to Remove 1.6601


F Ratio

1.660














TABLE XI

THE DEfRC.KUAi'ION OF DIGOXIN TABLETS SUPPLIED BY PUREPAC
500 AX.D 47.2% RELATIVE HUMIDITY


Time in Days

43

50

56

63

70

77


Microgratas of Digoxin

0.876

1.120

0.679

1.910

1.970

1.220


Time in Days

84

91

98

105

114

120


Micrograms of Digoxin

1.070

1.040

0.863

0.294

0.683

0.672


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.522 0.522

2.146 0.215


Variables in Equation

Coefficient -0.00859

Standard Error 0.00551

F to Remove 2.4339


F Ratio

2.434















TABLE XII

THE DEGRADATION OF DIGOXC.: T.'- F-S SUPPLIED BY PUREPAC
500 AND 80.5% RELATIVE HUMIDITY


Time in Days

43

50

56

63

70

77


Micrograms of Digoxin

1.360

1.320

1.810

1.800

1.590

1.430


Time in Days

84

91

98

105

114

120


Micrograms of Digoxin

1.420

0.515

0.208

0.114

0.961

0.667


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

2.000 2.000

1.878 0.188


Variables in Equation

.ent -0.01681

Error 0.00515

iove 10.6472


F Ratio

10.647


Coeffici

Standard

F to Reu














TABLE XIII

THE DEGRADATION OF DIGOXIN TABULT5 SUPPLIED BY BURROUGHS WELLCOME
230 A);D 18.8% RELATIVE HUMIDITY


Time in Days

38

45

50

56

63

77


Micrograms of Digoxin

1.110

0.749

0.736

1.590

1.470

1.420


Time in Days

84

91

98

105

114

120


Micrograms of Digoxin

1.580

0.887

1.860

0.992

0.698

0.227


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.135 0.135

2.387 0.239


Variables in Equation

Coefficient -0.00397

Standard Error 0.00528

F to Remove 0.5669


F Ratio

0.567















TABLE XIV

THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY BURROUGHS WELLCOME
230 AND 47.2% RELATIVE HUMIDITY


Time in Days

38

45

50

56

63

77


Micrograms of Digoxin

1.024

0.905

0.834

1.330

1.050

1.270


Time in Days

84

91

98

105

114

120


Micrograms of Digoxin

1.200

0.627

1.500

0.056

0.745

0.789


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.172 0.172

1.442 0.144


Variables in Equation

Coefficient -0.00447

Standard Error 0.00410

F to Remove 1.1896


F Ratio

1.190





43








TABLE XV

THE DEGRADATION OF DICOC::" TABLETS SUIPPIlED BY BURROUGHS WELLCOME
230 AND 80.5% RELATIVE HLbiIDITY


Time in Days

38

45

50

56

63

77


Micrograms of Digoxin

1.094

0.798

0.438

0.854

0.799 ,

1.140


Time in Days

84

91

98

105

114

120


Micrograms of Digoxin

1.200

0.574

1.390

0.039

0.799

0.124


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.193 0.193

1.728 0.173


Variables in Equation

Coefficient -0.00474

Standard Error 0.00449

F to Remove 1.1142


F Ratio

1.114















TABLE XVI

THE DEGRA'AVION OF DICOXIT' TABLETS SUPPLIED BY BURROUGHS WELLCOME
40 AND 18.8%7 RELATIVE HUMIDITY


Time in Days

43

50

56

63

70

77


Micrograms of Digoxin

1.260

1.470

1.230

0.488

1.390

0.518


Time in Days

84

91

98

114

120


Micrograms of Digoxin

1.210

0.711

1.410

0.999

0.784


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.105 0.105

1.205 0.134


Variables in Equation

Coefficient -0.00404

Standard Error 0.00456

F to Remove 0.7851


F Ratio

0.785














TABLE X'.'I I

THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY BURROUGHS WELLCOME
400 AND 47.2% RELATIVE HUMIDITY


Time in Days

43

50

56

63

70


Micrograms of Digoxin

1.140

1.230

1.380

1.770

1.420


Time in Days

84

91

98

114

120


Micrograms of Digoxin

1.280

1.170

1.830

0.799

0.481


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.320 0.320

1.147 0.143


Variables in Equation

Coefficient -0.00705

Standard Error 0.00472

F to Remove 2.2331


F Ratio

2.233














TABLE XVIII

THE DLGRAD..TION OF DIGOXIN T...i.h SUPPLIED BY BURROUGHS WELLCOME
40 AND 80.57 i:L.. U'F: HUMIDITY


Time in Days

43

50

56

63

70


Micrograms of Digoxin

1.020

1.330

1.600

1.250

1.140


Time in Days

84

91

98

114

120


Micrograms of Digoxin

1.100

0.849

0.950

0.819

0.677


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.406 0.406

0.264 0.033


Variables in Equation

Coefficient -0.00794

Standard Error 0.00226

F to Remove 12.2984


F Ratio

12.298














TABLE XIX

THE DEGRADATION OF DIGOXIN TABLETS SUPPLIED BY BURROUGHS WELLCOME
500 AND 18.8% RELATIVE HUMIDITY


Time in Days

50

56

63

70

84


Micrograms of Digoxin

1.180

0.027

1.130

1.650

1.090


Time in Days

91

98

105

114

120


Micrograms of Digoxin

0.914

1.200

0.809

0.799

0.196


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.128 0.128

1.968 0.246


Variables in Equation

Coefficient -0.00484

Standard Error 0.00672

F to Remove 0.5195


F Ratio

0.520














TABLE XX

THE DEGRADATION OF DIGOXIN TALLr.TS SUPPLIED BY EiLRP.OUGHS WELLCOME
50 and 47.2% RPEL- V\'I HUMIDITY


Time in Days

50

56

63

70

77

84


Micrograms of Digoxin

1.380

0.799

1.710

1.460

0.925

0.885


Time in Days

91

98

105

114

120


Micrograms of Digoxin

0.639

1.640

1.190

0.509

0.696


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

0.346 0.346

1.419 0.158


Variables in Equation

Coefficient -0.00793

Standard Error 0.00535

F to Remove 2.1960


F Ratio

2.196














TABLE YXXI

THE DEGRADATION OF DIGC,O:I' TAIjLLT'- SUPPLIED BY HURROUGICS WELLCO11E
50 AND 80.5% I.ILATIVE .4L",'IDITY


Time in Days

50

56

63

70

84


Micrograms of Digoxin

1.270

1.130

1.570

1.600

1.110


Time in Days

91

98

105

114

120


Micrograms of Digoxin

0.847

1.040

0.545

0.607

0.035


Source

Regression

Residual


Analysis of Variance

Sum of Squares Mean Square

1.471 1.471

0.627 0.078


Variables in Equation

Coefficient -0.01643

Standard Error 0.00379

F to Remove 18.7600


F Ratio

18.760













TABLE XXII

A COMPARISON OF i.. DUAL MEAN SQL'AUES
FOR LINEAR AND F ''K' LK:fIAL FITS


Linear Exp.

1 0.121 0.122

2 0.059 0.062

3 0.066 0.070

4 0.047 0.050

5 0.094 0.097

6 0.040 0.043

7 0.189 0.194

8 0.215 0.223

9 0.188 0.193

10 0.239 0.241

11 0.144 0.146

12 0.173 0.175

13 0.134 0.133

14 0.143 0.149

15 0.033 0.036

16 0.246 0.248

17 0.158 0.160

18 0.078 0.114













TABLE ;

3X3X2 FACiOC?.IJ. DESIGN OF


M

M2

Total

M1

M
2
Total

M1l

M2

Total


Key: MI

H1

H3

T


H1

3.28

3.97

7.25

5.37

4.04

9.41

6.72

4.84

28.22


CE .LFFICIENTS OF REGRFSSION


H2

7.12

4.47

11.59

8.54

7.05

15.59

8.59

7.93

43.70


H3

8.23

4.74

12.93

9.11

7.94

17.05

16.81

16.43

63.26


Total

18.63

13.18

31.81

23.02

19.03

42.05

32.12

29.20

135.18


= Purepac; M2 = Burroughs Wellcome

= 18.8% relative humidity; H2 = 47.2% relative humidity;

= 80.5% relative humidity

= 23; T2 = 40; T3 = 500

For ease of analysis all values were multiplied
by -100.














T.E 1. c XXI V

ANALYSIS OF VARIANCE OF COEFFICIENTS
OF REGRESSION


Source

Humidity

Temperature

Manufacturer

Humidity X
Temperature

Humidity X
Manufacturer

Temperature X
Manufacturer

Error

Total


** Significant

* Significant

# Significant


at the

at the

at the


Sum of Squares

102.7792

74.8350

8.4872


51.6099


0.6448


0.5377

4.9058

243.7996


1% level (P<.01)

5% level (P<.05)

10% level (P<.10)


Mean Square

51.3896

37.4175

8.4872


12.9025


0.3224


0.2688

1.2264

---


F Ratio

41.90**

30.51**

6.86#


10.52*


0.26


0.22












TABLE XXV

A COMPARISON OF WEIGHT VARIATION, HARDNESS,
AND DISINTEGRATION RATE OF TABLETS


Burroughs Wellcome


Tablet weight in milligrams


135.8
136.8
134.5
135.0
138.3
131.3
132.6
131.0
134.4
136.4

Sum of
Sum of

s


140.2
137.0
134.8
134.7
136.0
137.3
134.1
136.6
135.7
137.3


2709.8
367,244.2
135.49
2.217


Hardness in Kilograms


4.0
3.5
5.0
Sum of X
Sum of X2
X
s


3.0
3.0


18.5
71.25
3.7
0.836


Tablet weight in milligrams

122.7 124.4
126.8 125.5
124.8 129.8
128.4 124.6
127.5 126.5
125.8 123.5
125.7 125.5
126.2 124.9
125.6 126.7
126.6 124.7

2516.3
316,638.73
125.82
1.629


Hardness in Kilograms


1.5
1.5
2.0
8
13.0
1.6
0.224


1.5
1.5


Disintegration rate
in seconds


21
23
22
Sum of X
Sum of X2

S


Disintegration rate
in seconds


22
23

111
2467
22.2
0.836


16
17
17
84
1414
16.8
0.836


Purepac




































B. Figures


































FIGURE I

STANDARD CURVE FOR U.S.P. ASSAY

A = Absorbance measured on the Fisher Colorimeter

C = Concentration of Digoxin in mg/ml.
















1.0



0-8



06


0.4



0*2




0-0


C (mgn /m)


0-02 0-04 0-06 0"08 0.10

































FIGURE II

STANDARD CURVE FOR THIN LAYER CHROMATOGRAPHY

The square root of the area of the spot is plotted versus
the log of the weight of Digoxin in pg. spotted.



















r\)
o 0
. *


0
.0

.0
t

s.0

0



5


. .*. .


84*



























FIGURE III

ARRHENIUS PLOT OF REGRESSION COEFFICIENTS

The linear regression coefficients are plotted versus
the inverse of the temperature in degrees Kelvin.

Key: M-i = Purepac

M-2 = Burroughs Wellcome

H-I = 18.8% Relative Humidity

H-2 = 47.2% Relative Humidity

H-3 = 80.5% Relative Humidity












x

x





0
0~


H-I 0
H-2 o
H-3 X
M-1 -
M-2 -


5-2
(1/ T)X IO3


-0


X
-.0


3-0


3-4


0


(














RFFt r.;iJCrSF

1. Smith, S., J. Ch-ri. Soc., 508 (1930).

2. Stool, A., J. Am. Pharm. Sci. Ed., 27, 761 (1938).

3. Wilson and Grisvold, Textbook of Organic Medicinal and Pharmaceutical
Chemistry, Lippincott Pub. Co., 4th Ed. (1962).

4. The United States Pharmacopia, Mack Publishing Company, XVIII Ed.,
198-199 (1970).

5. Ibid.

6. Curlin, L.C., J. A. Ph. A. Sci. Ed., 44, 16 (1955).

7. Wagner, J., J. Pharm. Sci., 50, 359 (1961).

8. Wurster, D.E. and Taylor, R.S., J. Pharm. Sci., 57, 1419 (1968).

9. Pernarowski, M., Woo, W., and Searl, R.O., J. Pharm. Sci., 57,
1419 (1968).

10. Baljet, H., Schweiz. Apoth. Ztg., 56, 71 and 84 (1918).

11. Pesez, M., Ann. Pharm. Franc., 10, 104 (1952).

12. Tattje, D.H.E., J. Pharm. Pharmacol., 9, 29 (1957).

13. Khoury, A.J., Automation in Analytical Chemistry Technicon Symposium
166, Vol. 1, 192-195, Mediad, Inc., White Plains, N.Y. (1967).

14. Myrick, J.W., J. Pharm. Sci., 58, No. 8, 1019 (August 1969).

15. Jakovljevic, I.M., Analytical Chemistry, 35, No. 10, 1513-1516,
(September 1963).

16. Cullen, L.F., Packman, D.L., Papariello, G.J., J. Pharm. Sci.
59, No. 5, 697 (May 1970).

17. Mangold, H.K. and Kammereck, R., J. Am. Oil Chemists Soc., 1032 (1961).

18. Pastuska, G. and Petrowitz, A., J. Chemiker Ztg., 86, 311 (1962).

19. Purdy, S.J. and Truter, E.V., Chem. & Ind. (London), 506 (1962).

20. Purdy, S.J. and Truter, E.V., Analyst, 87, 802 (1962).

21. Reitsema, R.H., Cramer, F.J. and Fass, W.E., J. Agr. Food Chem.,
5, 779 (1957).










22. Seher, A., Nahrung, 4, 466 (1960).

23. Seher, A., Mikrochin Acta, 308 (1961).

24. Seher, A., Nahrung, 4, 466 (1960).

25. Brenner, M. and Niederwieser, A., Experientia, 16, 378 (1960).

26. Purdy, S.J. and Truter, E.V., Chem. & Ind. (London), 506 (1962).

27. Purdy, S.J. and Truter, E.V., Analyst, 87, 802 (1962).

28. Watson, E. and Kalman, S.M., J. Chromatogr., 56, 209-218 (1971).

29. Wilson, W.E., Johnson, S.A., Perkins, W.H. and Ripley, J.E.,
Analytical Chemistry, 39, 40 (1967).

30. Tan, L., J. Chromatog., 45, 68-75 (1969).

31. Svendsen, A.B. and Jensen, K.B., Pharm. Acta Helv., 25, 241 (1950).

32. Stahl, E., J. Chromatog., 5, 458 (1961).

33. Heusser, V.W., Planta. Med., 12, 237 (1964).

34. Michalska, B., Zurkowska, J. and Ozarowski, A., Acta Poloniae
Pharmaceutica XXIII, No. 1, 64-67 (1966).

35. Ibid., No. 2, 268-273 (1966).

36. Grollman, A., Pharmacology and Therapeutics, Lea & Febiger, 6th Ed.,
483 (1965).

37. Ibid., 484-513.

38. Koherty, J.D., Hall, W.H., Murphy, C.L. and Beard, O.W., Chest,
59, No. 4 (April 1971).

39. Cattell, McKeen, Gold, H., Modell, W., Kwit, N.T., Kramer, M.L.
and Zahm, W., J. Pharmacol. Exp. Ther., 82, 187-195 (1944).

40. Doherty et al.

41. Meyers, J., J. Pharmacol. Ther., 109, 45-57 (1953).

42. Doherty, J.E., Perkins, W.H., Amer. Heart J., 63, 528-536 (1962).

43. Doherty, Hall, Murphy and Beard.

44. Handbook of Chemistry and Physics, 53rd Ed., E40 and F7, Chemical
Rubber Company (1971).














BIOGRAPHICAL S1.' L11


Arthur H. Kibbe was born February 10, 1943 at Staten Island, New

York. He enrolled in the College of Pharmacy, Columbia University,

where he received the degree of Bachelor of Science in Pharmacy in

June 1966.

In January of 1967 he entered the Graduate School at the Univer-

sity of Florida and received his Master of Science in Pharmacy degree

in December of 1969.

He served in the U.S. Army from January 1970 to December 1971 at

which time he reentered the University of Florida to continue his studies

towards the degree of Doctor of Philosophy.

He is married to the former Gerda I. Poppel of Titusville, Florida.

His major is Physical Pharmacy with a minor in Pharmaceutical

Chemistry.











N











I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.



Oscar E. Araujo, Chairin
Associate Professor of Pharmacy

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.



Charles H. Becke/
Professor of Pharmacy

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.



Richard H. Hammer
Associate Professor of Pharmaceutical
Chemistry

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.



Stephen Schulan
Assistant Professor of Pharmaceutical
Chemistry

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.



John I. Thorn-by -C
Associate Professor of Statistics










This dissertation was submitted to the Dean of the College of Pharmacy
and to the Graduate Council, and was accepted as partial fulfillment of
the requirements for the degree of Doctor of Philosophy.

March, 1973




Dean, College of Pharmacy




Dean, Graduace School





































UNIVERSITY OF FLORIDA
l111111111111111111 1111 111111111111111111 l ll 1
3 1262 08554 3360




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E6G9U2JFD_DBUXXJ INGEST_TIME 2012-02-29T16:30:04Z PACKAGE AA00009131_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES