Return to December
2001 contents page
More combination vaccines needed
Faced with a bigger than expected
shortfall in supplies of some vaccines, countries must make choices.
Phyllida Brown reports
WHEN governments began to set out their
immunization plans in their proposals to GAVI last year, most said
they wanted to use new combination vaccines those that immunize
a child against four or five diseases in one shot rather
than have two or three separate vaccines. The benefits are obvious
fewer needles, fewer procedures, fewer demands on staff,
children and parents. The challenge at the end of 2001 is that the
supply of these vaccines is still far short of the demand, and that
it will be perhaps three more years before all countries that want
them will receive them.
is frustrating for immunization programme managers and all those
they serve and for the Alliance as a whole. GAVI policy clearly
states a preference for combination vaccines where possible(1).
At issue are two key combinations: diphtheria, tetanus and pertussis
plus hepatitis B (DTPHepB) and a five-antigen (pentavalent) combination
that also includes Haemophilus influenzae Type B (DTPHepBHib).
Last year, the Alliance had been told that there would be some 30
million doses in total of these two vaccines by 2002. Now the manufacturer,
Glaxo SmithKline, has told GAVI that the true figure will be closer
to 20 million. "This has created a big challenge for us," says Tore
Godal, Executive Secretary of GAVI.
Already, last year, the Alliance had
advised countries that there were too few doses of these relatively
new vaccines for everyone who wanted them(see 2),
and the GAVI Board had agreed a policy to allocate the available
vaccines. Countries were allocated vaccine on the basis of need,
giving priority to countries with the most fragile immunization
systems and the lowest routine coverage. It was argued that the
more fragile the immunization system, the greater the difficulties
in introducing separate additional antigens. However, in addition,
the Board ruled out Pakistan and Bangladesh because the large populations
of these two countries alone could have used up the entire supply.
Even though the manufacturer has almost
doubled its production in a year, its output has not grown as fast
as expected. As a result, the 12 countries that have already received
some supplies of these combination vaccines (see Table) will continue
to receive them, but no new countries will join the list in 2002
or 2003. GSK says it will be up and running with significantly increased
production by early 2004, but it will probably be 2005 before every
country gets what it wants.
Table 1: The countries that
already get the combination vaccines
Cambodia, Cote d'Ivoire, Eritrea,
Laos PDR, Madagascar, Mozambique, Tanzania
Countries receiving DTP+HepB+Hib:
Ghana, Kenya, Malawi, Rwanda,
So what happened? First, it appears
that the manufacturer was wary of putting its full investment into
scaling up the production of the vaccines until it had firm evidence
of the amounts of vaccine that GAVI would buy. It can take three
to five years to scale up production of a vaccine, including adapting
the manufacturing plant. But scale-up will only happen when the
manufacturer is sure that the buyers are there. The combination
vaccines had first been released on the market in 1996, but, says
Walter Vandersmissen of GSK, there had been little demand for them
until 1999. "The response at first was zero," he says. Surprised,
the company sat tight, wondering if there would ever be any customers.
Then, in 1999, the Revolving Fund of the Pan American Health Organization
(PAHO) started to buy the pentavalent DTPHepBHib combination. "Until
then, obviously, we had not stepped up production," says Vandersmissen.
In 2000 and 2001, GSK has been making
as much of the two combination vaccines as it can: "We are at the
limit of our capacity," says Vandersmissen. The company's estimate
that it would have 30 million doses available for GAVI by 2002 was
thus vulnerable to any quirk, however small, in production.
Countries needing more time
Vandersmissen points out
that GAVI itself has perhaps inevitably been
slower in approving countries for receiving the vaccines than
the ambitious original timetable it originally set itself.
Although the GAVI process has been faster than traditional
funding mechanisms, its independent review committee has had
to request further information from some countries before
approving their proposals, and some countries found that they
needed more time to gather the information. As a result, says
Steve Jarrett of UNICEF, which buys the vaccines on behalf
of countries with approvals from GAVI and funding from the
Vaccine Fund, "The amount [of vaccine] that we have bought
this year is relatively low compared with what we thought
we would buy... as there has been a slow uptake by countries."
This appears to have made GSK wary of committing itself to
immediate full-speed scale-up.
Originally, says Jarrett, UNICEF
planned to buy 24 million doses of the two combinations in
2001, but deliveries to most of the approved countries did
not start until the autumn. This was due, he says, to combination
vaccine not being readily available until the second half
of the year, as well as some countries' desire to start deliveries
late in the year to allow prior preparation.
Join the line:
vaccine production must be planned months or years ahead
GSK's own business
decisions may also have played a role. In deciding how to allocate
antigens between the manufacture of each of the two combination
vaccines, the company appears to have chosen to make relatively
more of the DTPHepBHib (pentavalent) vaccine, for which initial
demand from countries was more modest, than the DTPHepB (tetravalent)
vaccine, for which demand was much greater. Each dose of pentavalent
vaccine is priced at above $3, compared with around $1 for the tetravalent
vaccine. "It is difficult to understand why there was this huge
offer of pentavalent vaccine when both the UNICEF tender and the
[GAVI] forecasting group projections of demand were much lower,"
says Julie Milstien in the WHO Department of Vaccines and Biologicals.
Vandersmissen at GSK admits that the
higher-priced pentavalent vaccine is more attractive for the company
to make. "There is a difference in price and clearly that makes
it more interesting to us to have the newer product added to our
output," he says. But he stresses that the prices of the two products
are not directly comparable because the tetravalent vaccine is packed
in 10-dose vials whereas the pentavalent vaccine is in two-dose
Nor is profit the sole consideration,
he says. First, the company had been told that demand for the pentavalent
vaccine was likely to rise further in future, he says. Second, the
Hib component of the pentavalent vaccine must be freeze-dried, and
competition for freeze-drying capacity in the plant creates a "severe
bottle neck," says Vandersmissen. Having committed a significant
part of that capacity to Hib, the company was not about to waste
it. "If you have a scarce resource you must put it to the best use,"
While the shortfall continues, there
are alternatives. There are plentiful supplies of hepatitis B vaccine
available in monovalent form, and, given the delay until the combined
vaccine will be fully available, most countries may choose to use
it alongside DTP. So far, countries have been "very pragmatic" about
the shortfall between demand and supply, says Godal, and he is optimistic
that the impact of the delays will be small.
In the medium term, an alternative combination
vaccine could become available from 2002. A liquid-form tetravalent
vaccine combining DTP and Hib, made by Chiron Vaccines, is expected
to be prequalified by WHO shortly and a freeze-dried combination
of the same four antigens has been put forward to WHO for approval.
Some countries might choose to use one of these together with monovalent
hepatitis B vaccine, although the available quantities are not yet
Despite the new money available for
vaccine purchase, the overall system is still in need of improvement,
says Klaus Friederich, head of government and international institutional
policy for Chiron Vaccines in Marburg. Industry needs earlier decisions
from the public sector on the number of doses of each vaccine it
needs, and firmer commitment to buy, he says. He has been told to
wait for a final decision on the amount of the liquid DTPHib that
UNICEF will buy from Chiron for delivery from September 2002 onwards.
"But the lead time to produce the product is more than 50 weeks,"
he says. "What do I tell the production guys?"
But Milstien at WHO points out that
the vaccine is still not prequalified, with WHO awaiting "minor"
information from Chiron. "It's a bit of a catch-22." Everyone, it
seems, is waiting for someone else. "We are asking countries to
consider what substitute they want," says Godal. "We will be seeking
to make decisions in early February."
All players can suggest ways to improve
the process. Friederich is puzzled that the public sector purchasers
cannot commit themselves earlier to buy vaccines whose shelf life
is more than a year. Public-sector representatives point out, in
turn, that since the formation of GAVI there have been forecasting
meetings at which the industry has always been present, giving them
knowledge of further demand.
Ultimately, the small number of players
in the vaccine industry creates a seller's market for certain products,
which some observers feel is inappropriate. For example, says Friederich,
PAHO is shortly to tell Chiron how much of the DTPHib vaccine it
wants, effectively coming forward as a customer before UNICEF. With
a limited amount of vaccine available, Friederich does not believe
it should be left to a company to decide which public-sector buyers
should get the vaccine.
"Everyone is trying very hard,
and this is a mutual challenge," he says. "But the system really
has to be fine-tuned."
(1) Immunize Every
Child, GAVI policy document, February 2000, p 10 http://www.VaccineAlliance.org/download/immunize_every_child.doc
(2) GAVI Board Policy
on Vaccines of Limited Supply. http://www.vaccinealliance.org/reference/teleconf/october00.html#criteria
Return to December 2001 contents