product can be marketed at a lower price than the
innovator. In the special case of generic products, at
the very end of a drug’s life cycle when patent cover-
age has expired, equivalence need only be demon-
strated pharmacokinetically (usually involving only
a small number of normal volunteers and the rele-
vant, specific types of regulatory applications). How-
ever, when the new product is challenging the
position of an older one, then equivalency trials
usually require very large numbers of patients
(often hundreds per treatment group). The overall
tactic is to show that with a well-powered study (e.g.
b¼ 0 :925) there is noclinicalor statistical differ-
ence between the two treatments. The size of the
clinical difference that is worth detecting issine qua
nondefined prospectively and forms the basis for the
power calculations, and hence study size.
Mega-trials
When it is suspected that there may only be small
differences between active treatments, and when
placebo controls are unavailable for clinical or
ethical reasons, then it is often necessary to resort
to large-scale studies (‘mega-trials’). A good,
famous example was the clinical trial known by
the acronym GUSTO, where streptokinase and
recombinant tissue plasminogen activator (t-PA)
were compared for acute coronary thrombosis
(for a commentary, see Hampton, 1996).
Unlike more orthodox studies, mega-trials do
not attempt to control for large numbers of con-
founding variables. Instead, huge numbers of
patients (tens of thousands) are randomized, ‘the
cards are allowed to fall where they may’, and faith
is placed in the notion that a largenwill automa-
tically lead to well-balanced treatment groups. This
is not always the case, and imbalance can often be
demonstrated between treatment groups of even
several thousands when enough concomitant con-
founding factors are analysed (Charlton, 1966).
Safety surveillance
The draft ICH Guidance E2E issued 11 November
2003 provides a framework for the pharmacovigi-
lance of new drug products. Each new product
should have a pharmacovigilance specification,
which basically describes the clinical hazard land-
scape for the new product, as far as it can be known
at the time of approval. The specification is essen-
tially a problem statement. Each specification
should then be accompanied by a pharmacovigi-
lanceplan. The plan might include routine adverse
event reporting and periodic safety updates to be
provided to regulators, and/or recommendations
for clarifications to product labeling. In special
cases, however, a post-marketing surveillance
study might be recommended, and this forms
another type of phase IV study.
It is typical before conducting a post-marketing
surveillance study to obtain the view of the regu-
latory authorities on its design. The study may have
been a condition of product approval, and it is both
reasonable and wise to ensure that the study design
can be expected to provide the information that is
needed both by the sponsor and the regulators.
Unblinded designs that imitate the ordinary clin-
ical situation are the norm.New indicationsAs in the early phases of drug development, the
identification of new indications for old drugs can
be both rational and serendipitous. Rarely, even
adverse events can be exploited as new indications,
and the hair-growing properties of the antihyper-
tensive drug called minoxidil is a famous example.
Finding a new indication is an obvious opportu-
nity to increase market size by enlarging the poten-
tial pool of patients that can benefit from the
product. In this case, two pivotal, well-controlled
phaseIV studies demonstrating efficacy willusually
be required, at a minimum. If there is the potential
for a new type of clinical hazard to be associated
with new disease being studied, then a safety data-
base, of a size that regulators will find acceptable,
willbe needed forthesupplemental application, too.
Clearly, whenever such a project is contemplated,
then a financial assessment is needed of the balance
between the cost of the program, the probability of
success and the size of the eventual revenue incre-
ment that may or may not justify it.122 CH10 PHASE IV DRUG DEVELOPMENT: POST-MARKETING STUDIES