Dose-response relationships

A central question of drug therapy is: “What is the proper dose of a drug
that will produce the desired therapeutic effect without producing
(excessive) harmful side effects”.

Relationships between
doses and responses are
discussed in terms of log
(dose) - responses (or LDR)
curves. There are threshold
doses for action, as well as
maximally-effective doses.
 What about homeopathy? How does it work?
Simple answer – it doesn’t! The principles of homeopathy go completely
counter to the idea of dose-response relationships as defined by
pharmacologists (previous slide).

From Merriam-Webster: “Homeopathy is a system of medical practice that

treats a disease especially by the administration of minute doses of a remedy
that would in larger amounts produce in healthy persons symptoms similar to
those of the disease.” For example, if I want to treat chest pain in a patient I
would give him a tiny dose of a drug that would actually produce chest pain in
a healthy individual.

The more dilute the drug, the more potent it is. In some cases (eg. 30C
dilutions), homeopathic ‘medicines’ have been diluted so much that not a
single molecule of the original substance remains. This is illogical and
completely unsupported by double-blind clinical studies.

Here’s a link to a hilarious take on homeopathy (pretty accurate depiction

too). My favourite part is when the doctor wants to treat the car accident
victim with a dilute solution of scrapings of the car that hit him.
 Properties of LDR Curves

1. These curves describe the dose-

response relationship over a
wide range of doses.
2. LDR curves are typically
sigmoidal (“S” shaped). The
middle section is approximately
straight, facilitating statistical
analyses.
3. The same effect is produced by
different drugs that act by
identical or at least similar
mechanisms.
Dose-response relationships
allow one to quantitatively
compare drugs.

Linear

Question: How do
the potencies of
drugs A and B
compare?
Log
 Interpreting LDR curves: receptor occupancy

Drugs must attach to their sites of action (receptors) in the body in

order to exert their actions. Thus the intensity of drug action is
proportional to the occupancy of receptors (i.e., to the concentration
of drug-receptor [DR] complexes).

Response = α [DR] where α is a proportionality constant referred

to as intrinsic activity.
k+1
 DR
D+R 
kd = k-1/k+1 and is referred to
as the dissociation
k-1
constant of the receptor.

α [D] [R]
Response =
K + [D]

This K refers to potency NOT binding affinity, so don’t confuse k d with K

 α [D] [R]
Response =
K + [D]
can be converted into the straight line expression

1 1 K 1
= + x
Resp Respmax Respmax [D]
where Resp and Respmax refer to the
response observed and the maximally
attainable response.

The abscissa (y=0) gives you -1/K, the

ordinate (x=0) gives you 1/Respmax and the
slope of the line gives you K / Respmax.

This is an example of a graded response. Double reciprocal plot

 All or none effects

Normal distribution
Graded responses are those in which a
slight increase in the dose is expected to
slightly increase the responses to drug. In
contrast, for quantal or all-or-none
responses, either the drug does something
or it doesn’t. Examples include convulsions,
anaesthesia and death.
In discussing quantal effects, either the
percentage or actual number of patients

Cumulative distribution
responding to a given dose of a drug is
recorded.
Graded effects (eg. heart rate) can be
converted to quantal effects. Example:
drug effect on heart rate measure is
graded, but % subjects showing 50%
increase in heart rate in response to drug
is quantal.
 Median effective and median lethal doses

It is possible to evaluate doses to which 20%, 70% or 84% or any

other percentage of patients will respond. Such effective doses
are described as ED20, ED70 or ED84. The mean effective dose is
the ED50.

When drugs have parallel LDR curves, their potencies can be

compared and the more potent drug has a lower ED50.

LD50 is the dose of a drug that kills 50% of animals tested.

 Evaluation of drug safety

Toxicity as well as efficacy can be represented on LDR curves. Drugs

are considered safe if their toxicity LDR curves are considerably to the
right of their efficacy LDR curves. One way of describing the safety of
drugs is in terms of the therapeutic index (TI) where,
TI = TD50 / ED50 TI should be as high as possible.

However, TI does not adequately

characterize the relative safeties of
different drugs. Despite have equal TIs,
drug A is much safer than drug B
because the maximal therapeutic dose of
drug A, but not B, causes no toxicity.
 Certain safety factor (CSF) = TD1 / ED99

ED50 50.00 T.I. 40.00

ED99 300.00 C.S.F. 0.67
TD1 200.00
TD50 2000.00

CSF will be low for an unsafe drug (like drug B) and higher for a safer
drug (like drug A).

The LDR curves for the various responses to a drug, whether

therapeutic, toxic, or lethal, need not be parallel.

Evaluations of TI and CSF do not require assumptions of LDR curve

parallelism.