Professional Documents
Culture Documents
What is a Drug?
A drug is any substance other than food, that when inhaled, ingested, injected or absorbed after
administration in appropriate doses by other means causes a change in body functions (or in a pathogenic
agent) which can be exploited for clinical benefit prophylactically, therapeutically, or diagnostically. Drugs
may also act on physiological situations to achieve a wanted effect.
Using an excess dose causes toxicity, using too small doses doesnt give you the wanted effect.
NB!: in the case of rational design, as it is clear from the example, the discovery of the drug is not the
first step of the process! First of all, there is a phase of target selection and validation, with the
following aims:
o Define the unmet medical need (disease)
o Understand the molecular mechanism of the disease
o Identify a therapeutic target in that pathway (e.g gene, key enzyme, receptor, ion-channel,
nuclear receptor)
o Demonstrate that target is relevant to disease mechanism using genetics, animal models, lead
compounds, antibodies, RNAi, etc.
But determining a target is not enough! A compound must be found and proved to be able to interfere
with the normal target activity. So, the following evaluations must be made (these are valid
independently from which was the original way to achieve the discovery, eg: also if you produce a drug
by structural modifications This is my guess, these info were taken by the slides and not mentioned in class):
Develop an assay to evaluate activity of compounds on the target
- in vitro (e.g. enzyme assay)
- in vivo (animal model or pharmacodynamic assay)
Identify a lead compound
- screen collection of compounds (compound library)
- compound from published literature
- screen Natural Products
- structure-based design (rational drug design)
Optimize to give a proof-of-concept moleculeone that shows efficacy in an animal disease
model
Optimize to give drug-like propertiespharmacokinetics, metabolism, off-target activities
When this phase is completed, the compound such produced is the Preclinical Candidate, ready to go
through safety assessment
- Structural modification of a drug which is already known to cause an effect, with the aim to improve
the beneficial function.
Eg: Cimetidine was very effective and safe, but had side effects on the endocrine system, and needed
to be given 3 times a day because it was eliminated rapidly. Therefore, other companies tried to
improve these aspects, by modifying the original molecule, its conformation, to make it more specific
for H2, thus avoiding interactions with other targets, which is what causes side effects.
Curiosity: Nowadays new drugs that act directly on the proton pump to inhibit it are used, so
Cimetidine became obsolete
This process often creates a class of drugs, for example penicillins. Originally penicillin was only
effective on gram + bacteria, was modified by some enzymes in the body and couldnt be given orally
because was destroyed by gastric enzymes. Therefore, modifications of the original compound were
created and now there are penicillins with a wider spectrum of targets and some versions can be
given orally.
This approach is cheaper than producing drugs from rational design, since you can take advantage of
the knowledge collected previously about the target and the mechanism of the disease pathogenesis,
plus you already have a molecule to work with. The disadvantage is that the pharmaceutical company
following this approach wont be the first to put something new on the market, it wont own the
market, and it will have a lot of competition because other companies will be trying to do the same.
If it wasnt clear enough already, profit and money are obviously a leading theme in pharmacology.
Every time a new drug is developed through structural modification, the company will of course claim
that it is better than the original version. Is it true? In many cases it isnt. It might just be the same
and more expensive, or may even be worse, or simply there is not enough information to say it yet,
because not enough time has passed to observe possible side effects or to see effects on populations
not included in the clinical study.
How is this possible? Imagine the following example: a company wants to market a drug to treat
pain. Many drugs like these are present already. The company collects all the data, all the
experiments, and brings the dossier to the agency to be approved. Approval will simply require that
the drug is better than nothing, it does not have to be better than those already present, it will be
the market itself to determine its success or its failure.
Curiosity box!
A drug is released with 2 information: the SPC (summary of product characteristics), for
the physician, and the patient information leaflet, much simpler, not comparative. It tells you the
use, when to take it, the dose, the possible side effects,.
Drug companies are not legally allowed to advertise a drug for a use they are not approved
for, but there are ways around this: a pharmaceutical company may present to a doctor a drug
that is approved for something, telling them that they also have positive evidences that it could
have a good effect on something else, such as gastritis. A doctor may be intrigued and carry on a
research on his own patients, monitoring the effect of said drug on a disease the drug was not
designed for. Getting better in 90% of the cases is thanks to a placebo effect, so patients may
actually get better independently from the actual effect of the drug on a biochemical level. The
doctor thanks to these positive results can then publish the research, go to conferences to present
it, and other doctors may decide to try that alternative use for their patients too.
Doing something like this is actually illegal, because the drug is being prescribed with an use for
which it was not approved, so it might have unknown consequences.
But this is a blurry grey area, in fact in other situations you may be legally justified to deviate from
the approved use (off label use). This is extremely common, in many fields, because there might
be no drugs approved for a given condition, so you try anyways to do something that seemed to
be useful in similar patients.
- Random observation, luck: the discovery happens thanks to chance + ability to pick up signals, its
not simply random luck. The majority of drugs are discovered in this way.
Example: beta blockers, that have function of blocking beta receptors for catecholamines.
Catecholamines increase the heart rate and might favour arrhythmias, so pharmaceutical companies
developed beta blockers originally with the idea of fighting arrhythmias. The results of the
administration of these drugs were mixed, in some cases arrhythmias increased, in some cases
decreased. But then it was observed that when given to people that had high pressure too, pressure
was lowered. Nowadays this is their main use. The understanding of the mechanism underlying this
effect took much more time.
- Random screening: its a method based on a mixture of rational design and random finding:
companies use automatized technological methods to synthesise automatically, at random,
hundreds of different molecules. These molecules are then used to test effects in biological systems,
to see to which receptor they bind to with high specificity or affinity. Afterwards they reason about
what that binding could be exploited for, what effect it will have in the human system, so for which
disease it might be useful. Paradoxically, the reasoning process has been reversed compared to the
process followed when starting from a rational design.
PRE-CLINICAL DEVELOPMENT
Chemistry and pharmaceutics: how is the drug preserved, in which environment it is stable; info
about how the drug has to be administered (encapsulated, dissolved in liquid, injected)
Pharmacology:
- Pharmacodynamic data: they describe what the drug does to the biological system you apply it
to, the effect of the drug on different systems, and the different effects resulting from a certain
drug concentration.
- Pharmacokinetic data: describe what the body does to the drug, how long it survives in the body,
the way it is excreted (feces, urine, sweat, breath), the route of elimination, the possible
modification of the drugs by the body (metabolized by body enzymes), the chance for it to be
metabolized into something toxic. In brief, it describes the drug concentration-time courses in
body fluids resulting from administration of a certain drug dose
This is important because if the body gets rid fast of a drug, you need to administer it many times
a day, while others may survive in the body for months and need less administrations.
The rationale for PK/PD-modelling is to link pharmacokinetics and pharmacodynamics in order to
establish and evaluate dose-concentration-response relationships and subsequently describe and
predict the effect-time courses resulting from a drug dose
Toxicology: is based mainly on experiments done in animals. To keep annoying animalists at bay,
regulations to do a study in animals are extremely strict!
Animal Studies:
- Not completely replaceable by in vitro studies
- Indispensable to ensure safety for human use
- Predictive of therapeutic and adverse effects in humans (as well as pharmacokinetics)
- Privilege utilization of species which are low in the phylogenic scale (rats and mice)
- Conducted under strict, ethically-driven regulations
- We try to use as much as possible in vitro systems to limit use of animals for costs and
ethical reasons
-
All these studies are performed according to special quality requirements, codified by international
regulations, to have rigorous standards in terms of ethics and science:
- GLP: good laboratory practice for preclinical studies
- GCL: good clinical practice for clinical studies
Its a very complicated process, to ensure that everything is documented. For example, a scientist applying
any small change to an experiment must write it down with the precise date when it happened and sign it.
Components of GCP and GLP (skipped in class)
- Appropriateness of equipment and logistics
- Trained, specialized personnel
- Chain of responsibilities defined and tracked
- Detailed protocols
- Standard operating procedures (SOPs)
- Registration of procedures and protocols
- Standardized data recording and reporting
- Internal and external quality control
- Ethics and safety paramount
Key Items Addressed in Pharmacology Studies
- Mechanism(s) of action: from a molecular point of view, making sure that the drug binds to what it
is expected to bind
- Activity profile in models predictive of therapeutic efficacy in the target indication, ie: use animals
that reproduce the disease in question. For example, Alloxan is a drug with toxic effects on beta cells
of pancreas, so you can use it to cause diabetes in rats, then you can test your potential drug for
diabetes on such produced experimental animals. Nowadays you can also produce genetically
modified animals that reproduce the genetical modification that causes a given disease to have an
even more accurate model
- Activity profile in other organs and systems: check if the drug affects other systems of the body, to
try to avoid side effects, to produce a safe drug, to advise not to include a certain category of patients
(eg: pregnant women), or also to find additional useful indications
- PK, including ADME: ADME is an abbreviation in pharmacokinetics for "absorption, distribution,
metabolism, and excretion," and describes the disposition of a pharmaceutical compound within an
organism. The four criteria all influence the drug levels and kinetics of drug exposure to the tissues
and hence influence the performance and pharmacological activity of the compound as a drug.
- Drug interactions: people typically may take many drugs at the same time, so you want to find out if
your drug interacts positively with commonly administered drugs. These experiments are usually
performed in vitro
We have progressed a lot in understanding drugs interactions. Many drugs introduced in the body are bio-
transformed into metabolites. Now we know exactly for the majority of drugs what specific enzymes are
responsible for this. So we can test in vitro what a certain molecule does to the enzymes involved in this
process.
For example, statins: all of them are metabolized by the same cytochromes, enzymes in the liver (CYP3A3
and CYP3A4). You might want to develop a drug for any indication, and during the preclinical tests you find
out that your drug blocks the activity of these cytochrome enzymes. In this case you can know immediately
that if it is administered to a patient that is taking statins, the metabolism of statins will be stopped because
the enzyme responsible for it stops working. This will most likely stop the company from keeping developing
the drug, because it would have too much important interactions with drugs already on the market.
In the past the interactions we are speaking about were discovered during development or post-marketing,
by accident, nowadays screens are always made at the first stage in development of a drug to understand
interactions with the main drugs taken among the population. If interactions are found, development
stops, or the molecule is kept being modified until you get rid of the undesired interaction.
Examples: the following drugs were taken off the market, or stopped late in clinical development because
of adverse properties that could have been picked up very early in in vitro systems.
Key concept: we can identify very early adverse effects of drugs including drug interactions, without the
need to test it in vivo
Toxicity testing
Fundamental! Aimed at identifying potential for adverse effects, target organs for toxicity, and their
dependency on dose and duration of exposure. The drug must be safe!
There are standardized tests that need to be done, classified based on duration of treatment and specific
objectives. They need to be performed in animals.
Acute: observe the patient toxicity after a single dose
Subacute: the animal is exposed to the drug for 4 weeks
Subchronic: 12 weeks
Chronic: at least 6 months
Special toxicology tests:
- mutagenesis, cancerogenesis, teratogenesis : possibility that the drug causes cancer or a birth defect
when given to pregnant women
- second generation effects: drugs that do not cause a congenital malformation in the offspring of an
exposed animal, but they may cause some other effect later in the development of the offspring. For
example, Valproic acid is a drug for epilepsy that can cause a number of malformations in a baby if
taken during pregnancy, but it was also discovered that it can casue impaired intellectual
development that only becomes evident when the child is 5/6 y.o., so the child may be perfectly
healthy at birth, but later on develop mental retardation. Another example, there was a drug used
to cause abortion and to treat endocrinology problems. In the first case there was no problem, but
in the second case, if afterward a woman gave birth to a female child, she would develop cervical
cancer when she reached puberty.
These effects can only be found thanks to studies that keep following over time the offspring of
animals exposed to the drug during pregnancy
- developmental toxicology
- All the information on dosage are critical to determine if the drug will be viable or not to be given to
humans, ie: the first part of the preclinical tests are aimed at finding out which concentration is
needed to produce a therapeutic effect, but if in this phase the dose needed for therapeutic effect is
found to have high toxicity, development stops
- Select species with a metabolic activity comparable to that in humans. How can you know this? There
is overlap between clinical and preclinical development, especially in chronic testing: before starting
the long term experiments, that require exposure of animals for months and cost huge amounts of
money, healthy volunteers will take a single dose and the fate of the drug in humans will be
evaluated, then you can start the long term experiment with animals and make sure that the drug
has comparable fate
Here he started speaking of stuff not mentioned on the slides, triggered by questions
How is the price of a drug determined?
It varies depending on which part of the world you live in. In the US, no regulations on prices are present, the
drug just needs to be approved according to the standards of safety and efficacy, but it is the company that
chooses the price. There might be some interference when a company needs to negotiate with the insurance
system or health care provider for a supply of medications to hospitals.
In Europe its more complicated, the EMA does the same work of the FDA, but after approval the company
has legally the right to market the drug in all the countries of UE, and legally allowed to set the price they
want (same as in the US). But there is a big difference: health care is largely provided by the national health
service in Europe (especially in the Western countries), so if someone needs to buy a drug, the price will be
reimbursed by his national health care system. In this situation if a company markets a drug at the price it
wants, Italys national health system for example is forced to take it in the market, but it can decide which
will be reimbursed for the patients and which will be not. So, a company is free to sell it, but people wont
receive it for free from the state, so they most likely will not buy it if they have another similar option that
instead is. Therefore, a company always negotiates the price for a drug with the national health care systems
of the different countries, to make sure that it will be reimbursed, otherwise people wouldnt pay for it.
NB! This system leads to the consequence that the same drug might have different prices in different
countries of the EU. But the EU is a free market, so a distributor might refurnish for a drug in a country where
the price is lower, such as Italy, repackage the packet to give information in German, sell it at the German
price and pocket the difference in price between the Italian and the German price. This parallel market causes
biased statistics of consumption of a certain drug, that in Italy results extremely high and in Germany
extremely low.
But! What does the same means? You test this on healthy volunteers, usually around 20-30 people, each is
given in a random way either the generic or the brand.
NB!By doing this you soon realize that there is variability: even if you give the brand twice to the same person,
with a time interval of a month, the level checked in blood will be different (even if the subject and the drug
are the same!). You then group the measurements and make a mean between these profiles, usually the
difference of any profile from the mean is in the interval of 80%-125%, with a confidence limit of 95% (ie:
95% of the profiles surely are comprised in the limit 0.8-1.25, but 5% may not be (?)).
You do the means with the measurements of both people who took the bran and people who took the generic
and end up having two curves. On average the two curves must be superimposable, which does not mean
that they need to be identical, but that to determine that the two curves
are equivalent the 95% confidence limit of certain parameters must fall
between 80% and 125% of the mean, ie: the ratio between all the profiles
with the brand and all the profiles with the generic must have confidence
limits that fall between 0.8 and 1.25 of the mean. If the generic has an
average that is the same as the brand, but the confidence limits are broader,
ie: the variability is bigger, it will not be approved.
The point is, the generic shouldnt be too far from the brand but since it is normal for variability to be present,
even if you take the same drug, they must not be identical, but simply not having a greater variability than
the intrinsic variability of the brand itself.
There are some exceptions: the drug levels may be too variable even for the brand, or drug levels may be
undetectable (eg: drugs for asthma, that are inhaled and act locally, reach the site of action directly without
entering the concentration). In this case the only way to mark the generic is to test it on patients with the
disease, but in this case you need at least 400 subjects. In this case you must prove that safety and efficacy
is the same as the brand. This is a longer and more expensive process.
Companies that sell brands of course do not like the coming of generics into the market. Therefore, there is
a huge investment to discredit generics. This is done by doing poor quality studies, cleverly designed, biased,
with the objective to show that generics are worse. These studies are then advertised to doctors and even
more to patients, so we should always be careful to filter the information received.
The evaluation of generics is the only way for a drug to get to the market without going through the EMA,
but simply through the national agencies, to not overload EMA. It does not require a lot of expertise to check
safety and efficiency of generics. If a generic is approved in a nation, then there is a process to enlarge the
approval also in the other European countries, so a company even if it starts the approval by going
peripherally, it can then spread generally anyways.
If you market a generic, the regulatory agency must be provided with researches and evidences explained
above. The generics are produces in eastern or developing countries (Bulgaria, Romania, Kazakhstan, South
Africa, India), but still they are strictly controlled and regularly checked by inspectors.
About brands, the drug sold by the same name may not actually be always equal, because a company may
decide for economical convenience to change excipients, ie: if a new excipient comes on the market and is
cheaper than the old one, they change the manufacturing process. They are basically doing the same process
to create a generic, and they need to bring the same proofs to the regulatory agency, but they dont tell you,
and you think you are buying the same compound when you are not. Not only, even the doctor does not
know that he is switching a patient from a certain product to one slightly different (while it is obvious when
switching to a generic), so if side effects were to be caused by these modifications he couldnt even think to
attribute it to the change in said excipients.
Typically, when generics are approved, the company brings all the documentation to get it approved, but
they also may sell that documentation to many other companies, allowing them to use the same
documentation to get the right to produce and market the product: the same exact generic in this way is
produced and sold by many different companies. So there isnt so much variability among all the generics on
the market in the end.
Also, it might happen that the company producing the brand may do the same, ie: give the right to a company
to sell a drug identical to the brand, but with a generic name. in this case the generic is actually exactly
identical to the brand.
The system if not full proof, there is potential for fraud, but it is not so common. There are ways to keep the
situation controlled, for example physicians that switch a patient to a generic are required by law to send
reports to AIFA if they find that suddenly there is a loss of therapeutical efficacy. If many reports about the
same pro++duct come up, its an alarm and the agency does the experiments that the company is also
required to do, to check if the results of the studies might have been frauded. Most of the times these are
false alarms, but in few cases fraud can be identified.
Example: Lamotrigine is a drug for epilepsy, one of the most common, very profitable. There were large
numbers of reports of problems switching to the generic. These were supported by scientific papers, mostly
financed by the company protecting its exclusivity for its own interest, that stated
that if you have 50 generics on the market, all fitting in the confidence limits 0.8-
1.25, they will all be approved. This approval is based on the comparison to the
brand, ie: they all have an equal or smaller variability compared to the brand.
But different generics may have very different confidence limits among
themselves, so you cant guarantee that two generics are equivalent, you can just
guarantee that the generic is equal to the brand. Ie: on average two different
generics can be very different, one having higher levels in blood, one having lower
levels. They might not respect the confidence limits one in respect to the other.
These papers caused a lot of fears especially in consumers associations in countries with a system like the US
(in Italy it is different because if you go to a pharmacy and ask for a generic of a specific company, even if
they dont have it they will be able to get it in the next 24 h). In the US if you ask for a specific companys
product in a pharmacy, they wont do that, they will just give you the first generic they have already there,
so you will keep switch among one generic and another (which is also bad because different generics have
different colours so it might create confusion).
The US government financed the FDA to perform studies to understand if this was actually a problem:
experiments were made, they selected the most extreme versions among the 50 generics of Lamotrigine and
it was proven that these extremes that were supposed to have different effects actually had the same effects.
They also took a large population of people with epilepsy and kept rotating between the two extreme
generics and no harmful effect was recorded. Then another study was performed specifically on those
patients that had complained about a loss of efficacy, in a double blinded way, so that they wouldnt know
about what they were receiving, if the brand or the generic. The result is that even if they were switched
form the brand to the brand, so no change was really applied, they got worse. The conclusion is that it was
just a reversed placebo effect caused by the fear of change.
NB! Not all the countries have so strict controls, for example in Brazil there have been many cases of herbal
medicaments claiming to be natural actually containing chemical compounds such as ibuprofen, or drugs
claiming to contain a certain active principle that were found to not even have a trace of it.
Curiosity: in some countries it is legal to market a generic with a brand name, which means not the same
name of the original one, but not necessarily the name of the chemical compound, a simply random name.
These are called Branded generics: they are generics but have a specific name, different from the other
generics, so that they cant be confounded.
CLINICAL DEVELOPMENT:
Objectives of Clinical Drug Development: obtain all the necessary information for approval:
Confirm pharmacological properties identified preclinically, including PK-> confirm they are also
present in humans
Determine therapeutic potential, safety and tolerability-> show that the drug does something and
quantify efficiency.
Safety and tolerability are similar concepts, but safety has more to do with serious side effects,
requiring stopping the drug, while tolerability is about side effects that you can tolerate, but are still
adverse.
Define:
- indications (conditions in which you are legally allowed to prescribe a drug)
- contraindications (conditions in which you are legally not allowed to administer a drug to a certain
category of patients, because there is a risk for them)
- mode of use (doses, route of administration, dosing schedules, precautions eg: monitoring the level
in blood,)
Obtain marketing approval (requires demonstration of efficacy and safety)
Continue postmarketing evaluation of efficacy, safety and additional indications / contraindications:
pharmacosurveillance, companies are legally bound to do this and to file a report every year (or
even more often) concerning what is happening mostly in terms of safety to the drug, plus also
information about how much the drug is being sold and used, where, I which conditions,
3. Phase III: hundreds or thousands of health-impaired patients take the drug. The aim is to provide
evidence that the drug is safe and works under broader conditions of use
Information learned
1. Benefit/risk relationship of drug
2. Less common and longer term side effects
3. Labelling information
PHASE 1:
Peculiar use of drugs, borderline in terms of ethics. Normally you can apply an intervention only if
there is a reasonable chance of positive effect. If you give a drug to a healthy subject, you only do
that to look for negative effects, so the only thing that could happen is a worsening of the volunteers
health. This is justified because it is done with the aim of protecting patients from negative possible
effects. Volunteers are supposed to have a greater freedom than patients in deciding whether to
participate or not (this is also the reason why they must not be in economical need to be chosen and
they are not paid more than they would get paid in any other job for the same amount of time
Experiments in phase 1 are relatively cheap compared to the total investment for development: it
costs around 2-3 million dollars.
Objectives:
- Obtain preliminary evaluation of tolerability (adverse effects, maximum tolerated dose)
- Determine PK properties in humans (mostly done through blood samples and urine samples during
administration of single doses and multiple doses afterwards)
- Confirm (whenever possible) the pharmacological properties identified preclinically
PHASE 2
It is a critical phase, its aim is to reach an answer as reliable as possible, to the question: is this going
to be a valuable medication?
Performed on patients with the disease in question, but on a limited number
Objectives:
- Demonstrate with rigorous methodology the expected pharmacological / therapeutic properties
(proof of concept studies): the extreme scientific rigour is needed to get as much information as
possible about safety and therapeutic potential with a limited number of subjects.
Determine the proof of concept means that you have an hypothesis, the aim now is to understand
if it is correct and the drug will actually do what is meant to do.
- Obtain further data on adverse effect potential
- Characterize dose-response relationships and PK-PD relationships
The dosage is critical with respect to the risk to benefit ratio. When a drug is released physicians need
to have information about the relationship between dose and therapeutic effect, or a side effect.
Individuals are different among themselves and may answer differently.
- Assess potential drug interactions in the target population
Often, for ethical reasons, if you are developing a drug for which there is another treatment, it is
unethical to test it in a sick person, because you dont know if it works, plus someone must end up
receiving a placebo, so if it isnt a trivial condition people would be prevented from getting the best
treatment available. Therefore, usually phase II experiments are performed in people resistant to a
certain already existent treatment, the new drug is simply added on to the established treatment
they are already receiving. So already at this point it is important to check for potential interactions.
Curiosity: For this reason, it is common that initially drug are released only with the indication of people
not responsive to available treatment, then, when more information are acquired, the drug may start
to be used alone, not in addition to the other treatment.
This is also because if you only tested the compound on people taking also another treatment, you can
only get approval to do that. To get an approval to administer the treatment without concomitant
treatments, you have to conduct an experiment where that is done and checked.
The ideal experiment compares the effect of placebo, of different doses (at least 3, selected on the
basis of the experiments in phase I). Also, it includes the standard treatment used for that condition
(if there is one).
In this way you can:
- Check what placebo does: your drug must do more
- Find lowest dose producing an effect
- Find highest tolerated dose
- Find the best dose in between these two limits (maybe studying 3 doses wont be enough for this),
but this will require lots of subjects and it will cost a lot
- Find how your drug compares with the already available treatment, possibly the gold standard. This
is not required for approval, you just need to prove it is better than placebo, but it can be a plus when
the drug gets to the market. Plus, this can be a way to gain more investors to finance phase III (that
is much more expensive).
NB: one company will probably not bias the study at this point, because it is useful also for them to
gain reliable information (if they want to go for fraud they will probably do it later). The problem is,
if the result of the study is not what they wished for, i.e: the competitor drug results to work better,
the study will never be published, since there is no legal obligation, and physicians will never know
that such a study was done and it could help them to choose the best treatment for their patients.
Every time it is possible in development of a drug we try to measure what eventually is clinically relevant, but
if the evidence from using a surrogate endpoint is a reliable predictor of benefit, measurements can be based
on it.
Nb: of course there is some risk: it needs to be accepted that surrogate endpoint will translate into a clinical
benefit.
Eg: drugs for diabetes: for Metformin there are long term studies conducted for may many years, that have
shown exactly the mechanisms by which this drug reduces risks linked to diabetes: cardiovascular mortality,
diabetic retinopathy, diabetic neuropathy. Other new drugs instead for sure reduce sugar blood level, but it
cant be known if this will translate into reduced clinical risks.
PHASE III
Similar to phase II, but in a much broader population, closer to the real world, without so strict
inclusion criteria (but still not completely equal to the normal population)
Collect all the information necessary to get the permission to put the drug on the market
Cost of phase 3 is huge! 10s or 100s of millions of dollars. If a drug fails in phase III it is an economically
disaster, so a lot of efforts are placed to avoid this.
Objectives
- Confirm and characterize in greater detail efficacy and adverse effects
- Improve understanding of dose-response relationships and long-term effects
- Evaluate influence of variables (confounders) typically excluded from phase II studies, such as age,
concomitant diseases and concomitant treatments.
NB! Drugs that will eventually be used around 50% in elderly people in the real world, often at the
end of phase III will only be evaluated on a ridiculous proportion elderly, i.e. around 10-15% of the
total population that has been studied.
- Compare investigational drug with alternative treatments (gold standards or other), if not done
before.
NB! At this stage, companies may start to think about the possible benefit that a study could bring if it
gives good results, and studies unfavourable may be hidden
Methodology
- Larger populations than Phase II, and broadened eligibility criteria (n = 500-5000)
- Special groups (children, the elderly, patients with co-morbidities)
o Elderly
o Children: Children normally eliminate the drugs faster than adults, so companies should perform
studies to give to the physicians information about dosage. Sadly, usually companies dont, and
rely on off-label use to gather information after the drug is put on the market.
o Patients with co-morbidities should also be studied to determine dosages. E.g. dosage in a
patient with kidney disease, or with cirrhosis, or in a patient undergoing dialysis, or so on.
- Randomized double-blind controlled trials vs placebo or active controls, open-label follow-up
studies, specific safety studies
Remember the main problem about the difference with the true population with the disease and the
population included in the studies.
For example, a metanalysis compared patients with heart failure in general clinical practice with those
included in clinical trials:
Regulatory (clinical) trials: mean age = 62; 81% males.
In general practice (average target for the drug): mean age = 75; 47% males.
This can lead to unforeseen problems if no studies on females or elderly were done
PHASE IV
Post-marketing
Objectives:
- Improve knowledge on efficacy and adverse effects
- Asssess long-term outcomes
- Evaluate effects in population potentially at special risk
- Identify rare or delayed adverse effects
- Obtain data on drug utilization and pharmaco-economics. Pharmaco-economic evaluation is usually
done post marketing by national health service if the health is a public service (cost benefit ratio in
comparison to other drugs).
NB! Studies investigating new indications, new dosages or new models of use are classified as phase III.