Biological standardization or bioassay of drugs is defined as

the assessment of the activity of a preparation by

measuring
its effect on living animals or tissues. It is essential that the
amount of the active principle in a dose be uniform and of
known potency. This enables the clinician to prescribe a dose
that wil always, as nearly as possible, induce the same
magnitude of action. Thus, bioassay is a procedure for
determining the quantitative relationship between the dose (or
concentration) of a drug and the magnitude of response it
evokes. Such procedures utilize intact animals, isolated
living tissues, animal preparations or micro-organisms.
Selection of the Method
The quantitative estimation of a drug may be done by
physical, physico-chemical or biological methods. In
physical methods, a physical property of the drug like colour,
fluorescence or irradiation is used; in physico-chemical
methods, a combination of a physical property and aa
chemical reaction is used, eg., changing the colour by a
chemical reaction and then measuring the intensity of the
colour; and in biological methods some biological property is
used to estimate activity. Whatever method is chosen, it
should fulfil the criteria for reliability, i.e., it should be
sensitive, specific and accurate; and practicability, i.e., it should
be simple, quick and cheap. Based on these criteria the
physical and physico-chemical methods are usually
superior to biological methods. However, the biological
methods for assay of drugs are still used to measure the
potency of substances for which physical or
hysico-chemical methods are not available. If for a certain
drug both physico-chemical and bioassay methods are
available, the former are preferred as they are more
accurate, less time consuming and convenient to follow.

Bioassay methods are usually requird for substances

derived from plant or animal sources. Synthetic drugs do
not require bioassay for standardization, as if they fulfil the
purity criteria, mere weighing or measuring is a reliable
gauge of their activity.
 The foremost principle of bioassay is to compare
quantitatively the effect of the preparation under test
(sample) with that of standard preparation. The accuracy and
reliability of these assays depends on : () specially trained
staff; (i) details of the procedure as described in the official
compendia (pharmacopoeias); and (iii) proper selection of
the animals and the test procedure.
The details of bioassay methods to be followed are given
in the pharmacopoeias (IP, BP etc.). For many drugs there
are more than one bioassay methods. The
assay parameters
recorded for some drugs are given below
Adrenaline. Blood pressure raising effect in the cat or dog.
Noradrenaline. Rise of blood pressure in the spinal cat.
Histamine. Contraction of the isolated guinea pig ileum.
Insulin. Hypoglycaemic convulsions in mice.
Acetylcholine. Contraction of the frog's rectus adbominis
muscle.
d-Tubocurarine. Rabbit head drop due to paralysis of neck
muscles.
Androgen. Growth of the capon's comb.
Oestrogen. Cornification of the vagina in rats.
Digitalis. Death due to cardiac arrest in systole in the
guinea pig.
Application of Bioassay Methods
The various applications of bioassay methods are:
i) Standardization of drugs of natural origin.
(ii) Estimation of biologically active substances like
acetylcholine, adrenaline, noradrenaline and serotonin
in body fluids or tissue extracts.
(ii) Screening of new compounds for biological activity. Even
synthetic products are subjected to these methods, as
chemical structure does not truly predict pharmaco
logical activity.

49
 (iv) Bioassay is employed if the drug is composed of a
complex mixture of substances of varying structure
and activity, e.g., digitalis, posterior pituitary.
() Diagnosis and research. The concentration of of
gonadotrophins in the blood or urine may be
estimated by injecting these fluids in animals
(chemical methods may be employed, if available).
(vi) Estimation of the dose of a drug required to produce a
therapeutic or toxic response, e.g, EDso (effective dose
50) or LD5o (lethal dose 50).
When results of biological and chemical methods for the
assay of an active substance disagree widely, the bioassay
result is upheld.

Principles of Bioassay of Drugs

Burn and Dale have enunciated certain principles for the
conduction of bioassay procedures. These must be followed
for obtaining valid results.
Principle 1. All bioassays (laboratory studies, toxicity studies,
clinical trials) must be comparative against a standard drug or
preparation. This is one way of minimizing error in bioassays
due to biological variation (see later), as the new sample is
simultaneously compared with the standard preparation.
Standard preparations (Reference Standards). A standard
preparation is a selected representative of the substance
which serves as a basis for the comparative measurement of
activity. These standard preparations are distributed to
experts, pharmaceutical firms and research laboratories to
standardize new batches of their
preparations.
In India these Reference Standards are maintained and
distributed by the Central Drug Research
and the Central Research Institute, Kasauli.Laboratory,
Calcutta,
In USA they are
maintained by the US Pharmacopeial Authorities, and in Great
Britain by the National Institute
The Expert Committee on
of Medical Research, London.
Biological Standardization of the WHO
also maintains and distributes these
laboratories. preparations to various
 In India these Reference
Standards maintained and
are
distributed by the Central Drug Research
and the Central Research Institute, Kasauli.Laboratory, Calcutta,
In USA they are
maintained by the US Pharmacopeial Authorities, and in Great
Britain by the National Institute of Medical
The Expert Committee on Research, London.
Biological Standardization of the WHO
also maintains and distributes these
laboratories. preparations to various
A Unit is defined as a definite weight of one of these
Standard Preparations which
terms of the
produces a certain effect in
preparation under test.
Principle 2. The Standard and the new drug should be, as far
as possible identical to each other.
If it is so, their
curves will have the same
dose-response
slope and would be parallel, i.e.,
the potency ratio would be constant all
levels. However, if the curves are not
along the response
is invalid.
parallel, this
statement

Principle 3. The method for comparing the unknown and the

standard drug should preferably (but not
essentially) test the
therapeutic property of the drug. Ideally an analgesic should be
tested for analgesic activity, and an anticonvulsant for
anticonvulsant activity. However, it is not always possible to
do so, e.g., for insulin lowering of blood
glucose would be
the ideal potency index, but the commonly used method is
the mouse convulsion method, in which the percentage of mice
developing convulsions with the standard and new insulin
is compared.
Principle 4. The method should estimate, as far as possible,
and allouw an estimate of the error due to biological variation in
different animals/persons at any one time, and in the the same
animal/person at different times. The precautions to be taken to
minimize error due to biological variation are: (i) the
experimental conditions must be kept constant; (i) the
biological response (indicator) should be sensitive to h
drug: (ii) the indicator should be insensitive to other druoe.e
(iu) the indicator should be capable of giving constant and
reproducible results; (v)a preparation of known
concentration (standard) must be available to compare the
response; (vi) the number of experiments should be
be
sufficiently large, and the assay designed to minimize
minimize
biological variation; and (vi) the animals should be of the
same species and strain (preferably litter mates), of similar
ilar age,
age,
weight and sex, kept on similar diet and housed under similar
conditions. In certain bioassay designs cross-over of animals
rceiving the sample and the standard is possible.
In short, the sources of error in bioassays are mainly of
two types. Firstly, errors due to biological variation; and
secondly errors due to faulty methodology.

Biological Variation
Individuals and animals show marked variability in their
response to a drug. The response of any drug even in a
fixed dose in a most carefully selected group of animals is
variable, due to the law of biological variation. If such an
experiment is conducted on a large number of animals, and
the
the frequency of the response (vertical axis) is plotted
normal
against the dose (horizontal axis), one obtains the
distribution curve (Fig. 1.9.1). The features of
the
frequency
normal distribution curve are that it is bell shaped,
with two tails which never actually touch the
symmetrical, it,
horizontal axis, although they Continuously approach
and the mean lies at the peak of the curve i.e., the highest
frequency
 SD -3 -2 -1 +1 +2 +3 SD
68%-
95%
99.7%

Fig. 1.9.1: Normal frequency distribution curve (Gaussian Curve). X = mean;

SD = standard deviation.

The normal distribution curve has a symmetrical

shape,
but the exact location and scale of distribution depends on
two constants, namely the mean and the standard deviation
(see later). The normal distribution curve indicates that a
majority of the animals show a particular degree of response,
but there are animals which are hyporeactors (left half of the
curve), and others are hyperreactors (right half of the curve).
The mean indicates the average response in that group or
animals.
To assess the extent of biological variation, the extent of
spread of responses (scatter) is to be indicated by calculating
the Standard Deviation (SD). The SD may be geometrically
described as the distance from the mean to the steepest part
of the normal distribution curve vertically (dotted line), and
this is an excellent measure of variability. An analysis of the
normal distribution curve (Fig. 1.9.1) shows that mean t 1
SD includes about 68% of animals showing responses within
this range. Mean t 2 SD includes about 95% of the animals,
and Mean t 3SD includes about 99.7% of the animal
responses. The wider the base of this curve, the greater the
variability. The maximum frequency is midway between the
highest (+) and the lowest ( point on the horizontal axis
i.e., at the mean value.
The method of expressing the error due to biological
variation, in the form of standard deviation or standard error
of the mean value, or the significance of the difference
between two means is described in the section on
"biostatistics" in this Chapter. In expressing the limits of
error of biological assays the term "limits of error (P=
0.99)"
is used. This statement is based on the convention that for
practical purposes, a probability of 0.99 is equivalent to
certainty, i.e., the result of the assay will be within the stated
limits 99 times out of 100 times that the assay is made.

Methods of Bioassay
Four methods are usually enmployed tor interpreting results
of biological assay.
Method 1. Threshold dose measured on each animnal (Direct
end-point assays). In this type of assay the threshold dose is
measured for obtaining a specific biologic end-point in each
animal. In these assays the drug (standard or test sample) is
administered slowly to the animal until some sharply
observable effect (end-point) is produced. The amount of the
for end-point is then noted.
drug or
the time reaching the
For example, the assay of digitalis in cats is carried out by
infusing the extract containing digitalis into a vein at the rate
of 1 m per minute till the heart stops and the blood pressure
falls to zero. The volume of the fluid passed into the vein
(threshold dose) is noted in each animal. This is the minimal
lethal dose of digitalis in cats. Two series of experiments are
done with the standard preparation and the unknown
preparation, and the potency ratio is calculated. Direct
end-point assays are applicable when an end-point can be
reached in each animal; the drug effect appears rapidly; and
the drug effect is directly proportional to the dose.
 Method 1. Threshold dose measured on each animal (Direct
end-point assays). In this type of assay the threshold dose is
measured for obtaining a specific biologic end-point in each
animal. In these assays the drug (standard or test sample) is
administered slowly to the animal until some sharply
observable effect (end-point) is produced. The amount of the
drug or the time for reaching the end-point is then noted.
For example, the assay of digitalis in cats is carried out by
infusing the extract containing digitalis into a vein at the rate
of 1 ml per minute till the heart stops and the blood pressure
falls to zero. The volume of the fluid passed into the vein
(threshold dose) is noted in each animal. This is the minimal
lethal dose of digitalis in cats. Two series of experiments are
done with the standard preparation and the unknown
preparation, and the potency ratio is calculated. Direct
end-point assays are applicable when an end-point can be
reached in each animal; the drug effect appears rapidly; and
the drug effect is directly proportional to the dose.
Method 2. Responses recorded or measured (Graded response
assays). At times the effect of the drug can be observed
repeatedly on the same tissue. The standard preparation and
the test sample are administered alternately, and the doses
adjusted until they give equal effects. Precisely, a fixed dose
of the standard is matched against varying doses of the test
sample, the object being to find out equivalent doses of the
two. The concentration of the standard being known, the
concentr on of the unknown is easily found out. This
technique is known as matching assay. The assay of
histamine on the isolated guinea pig ileum is an example of
matching assay methodology. Very small quantities of
histamine can only be assayed biologically. Similar
methodology can be used for the bioassay of the posterior
pituitary extract, adrenaline and acetylcholine.
Many of the vitamins and hormones cann be assayed by
obtaining their graded responses. One randomly selected
group of animals receives doses of the standard preparation,
and the other similar group receives doses of the unknown
preparation to be tested. The experimentis repeated
measuring the response (eg., liver glycogen contehih MBRI
for glucocorticoids; percent fall in blood sugar for insulin in
rabbits) at two or three dose levels, separatelA the
standard and the unknown preparation and thPNeaAn th earAn-
response at the various dose levels is obtainedTheT
relationship between the dose and the response is plotted on
a graph paper to give the dose-response curve which is not
a straight line as on the ordinary arithmetic scale the doses
are not equally spaced. But if the response (y-axis) is plotted
against the log dose (x-axis) a straight line is obtained, as the
logarithm of the dose is linearly related to the response. This
is termed as the log dose-response curve (See Chap. 1.6). Thus
two or more fixed doses of both standard and unknouwn are
used, and the responses to each measured in an adequate
number of animals. Depending on the number of doses (two
or three) of the standard and unknown used, the assay is
known as a 2 x 2 or 3 x 3 assay.
From the results, log dose-response curves are
constructed (Fig. 1.9.2), and the potency of the unknown
sample relative to the reference standard is estimated by
fitting parallel lines to the data statistically and by calculating
the horizontal distance between the two lines. This gives the
logarithm of the relative potency. Such assays are only
applicable and valid if the two log dose-response lines are
parallel, i.e., when the potency ratio is constant with both
low and high doses. For this reason a statistical test for
parallelism is included in the routine analysis of such assays.
Non-parallelism indicates that the potency ratio is not
constant, and the active commponents are not present in the
same proportion in the standard and the unknown
preparation, or may be the unknown preparation has an
interfering compound which makes its mechanism of action
different from that of the standard preparation. The statistical
calculations involved are complex and have been excluded.
 Unknown
Sa
Standard
A

S2
log
-potency-
ratio
U S

LOG DOSE
(x-axis)
Fig. 1.9.2: A typical 3 x 3 graded response assay (diagrammatic). The log
aose-response curves are parallel. The unknown preparation is more active
than the reference standard by a factor which is given by the antilogarithm of
the horizontal distance between the two lines. Confidence limits can be
calculated from the scatter of the experimental points around the fitted parallel
lines.

Cross-over test. When the effect of the drug can be

observed and measured more than once in the same animal,
the accuracy is greatly increased by arranging the experiment
according to the cross-over design. On the first day one
group of animals receives the standard drug and another
similar group receives the unknown preparation and the
effects are measured; on another day the standard and
unknown are crossed over, so that the first group receives
the unknown preparation and the second group receives the
standard. The average effect of each preparation, for both
stages of the experiment is then calculated, and the results
interpreted by the method of drawing log dose-response curves
as already discussed. The advantage
of this arrangement is
that it eliminates errors between one animal and
another,
since both the standard and the unknown are tested on the
same animal.
This experimental design is used for the assays of insulin
on rabbits, designated as the twin cross-over test. It
combines the advantages of a cross-over test with those of
2 x 2 assay, according to the
following design :
Group First injection Second injection
Standard dilution 2 Sample dilution 1
2 Standard dilution 1 Sample dilution 2
3 Sample dilution 2 Standard dilution 1
4 Sample dilution 1 Standard dilution 2

Rabbits of a specific weight are maintained on a

prescribed feeding schedule, and the second injection is
made at least 24 hours after the first. Obviously, the
cross-over design cannot be employed if an assay results in
the death of the animal, or an organ is removed from the
body in the test procedure.
Method 3. Percentage of positive effects measured (Quantal
assays or All-or-none assays). A quantal response is obtained
and the percentage of positive effects at each dose is
computed. The unknown is then compared with a standard
with respect to potency is causing the quantal effects.
However, the most familiar example of quantal assays is in
toxicity testing eg., LD5o determinations.
 The median lethal dose
(LDs0) is that dose of the
drug/poison which kills 50 per cent of a test population of
animals. For the determination of LDs0 the drug/poison in
different doses is injected into groups of animals, and the
percentage mortality determined. This percent mortality is
plotted against the dose, and a curve is drawn through the
plotted points. This curve is sigmoid or S-shaped, as in low
doses all >imals live, and in higher doses all animals die
(Fig. 1.9.3).
00

75

1 50

25

LD 50

DOSE
Fig. 1.9.3: Determination of LDs0 of a drug/poison (diagrammatic). The per
cent mortality is plotted against the different doses injected into animals
(rats/mice); and the dose estimated which killed 50 per cent of the animals
(dotted line).
Such quantal assay methods can also be employed for
the bioassay of insulin, where the presence or absence of
hypoglycaemic convulsions (in mice) at the doses tried is
taken as an indicator. The curves for the unknown insulin
and standard insulin are compared with respect to potency.

Method 4. Microbiological assays. These methods are used

for assaying antibiotics. The test preparation is compared
with the standard preparation for its inhibitory effect on
particular organisms, sensitive to the agent.

Biostatistics
Statistics is defined as the science which deals with
collection, presentation and analysis of data, and the
draw an
interpretation of results obtained, with an aim to
inference. According to R.A. Fisher. statistics essentially is
a branch of applied mathematics, and may
be regarded as
mathematics applied to observational data. Biostatistics is
statistics applied to biological sciences and to medicine.
Whatever the definition, statistics is the science of
information presentation in a numerical form which ernables
us to maximize our understanding of such
information.
Biostatistics has two sub-divisions.

(a) Descriptive Statistics. This deals with methods of

describing and presenting a large mass of data, with
the help
help of tables, graphs, histograms, pictograms,
piecharts, bar charts, statistical maps, multiple bar
charts and others.
(b) Analytical Statistics. This deals with methods that
enable us to draw a conclusion from the data
obtained, with the help of significance tests,
correlation, regression and other methods.
in
already discussed
Biological responses are variable, as
the section on biological variation in this Chapter. Thus it is
k did in a laroe