The "New" Newborn Screening and the Tyranny of the Randomised Controlled Trial
Bridgit Wilcken
Newborn screening tests detect babies who need treatment for rare diseases. Because the diseases ARE very rare, it is very difficult to evaluate the effects of early diagnosis and treatment by the currently accepted standard of the randomised controlled trial and lesser evidence often needs to be gathered. It is hard to get this evidence accepted by the scientific community.
Newborn screening tests
The aim of performing biochemical
tests in newborns is to detect treatable diseases before the adverse effects
occur. Newborn screening was started in the 1960’s with a simple robust test
for phenylketonuria (PKU), a metabolic disorder which causes severe mental
retardation unless diet treatment is started in the first few weeks of life.
The test was developed by an American microbiologist, Dr Bob Guthrie, and was
conducted then using a bacterial inhibition assay, on a blood sample dried on
filter paper. The use of a dried blood sample made mass testing possible,
cheaply, at a central laboratory. This was a dramatic step in preventive
medicine. For the first time, a disease that caused severe mental retardation
could be completely controlled. Our PKU patients can now lead normal lives by
following a low-phenylalanine diet. The next important step in newborn
screening was in the 1970’s with the introduction of a test for hypothyroidism
(cretinism), a hormone deficiency that also leads to mental retardation if it
is diagnosed too late. Several other disorders where there is a benefit from
early treatment can be detected by newborn screening, but each needs a
separate test, and this is time consuming and expensive. In New South Wales
before 1998 we only tested for 4 disorders, PKU, hypothyroidism, cystic
fibrosis, and galactosaemia. Testing for the even rarer disorders couldn’t
easily be justified. That is, until the advent of electrospray tandem mass
spectrometry (MS/MS), which was suitable for the fast through-put needed to
test hundreds of babies per day.
The new development, tandem mass spectrometry
MS/MS, by scanning groups of
structurally related compounds simultaneously and sorting by mass, offers the
possibility of broad-spectrum screening which will detect a large number of
disorders by a single test. So testing can be offered for disorders
that, because of their extreme rarity, would not easily qualify for inclusion
if a separate test were needed. At present we measure selected amino acids and
acylcarnitines.
An example of finding a very rare disorder before harm
is done
In 1991 a 16-month-old boy was seen
for the first time with heart failure. He was thoroughly investigated, and
several tests were done to find the cause. Before a crucial test result came
back he died suddenly and unexpectedly. He was found to have had a very rare
disorder, carnitine transporter defect, (only then described in a handful of
cases world-wide) and was the first patient in Australia identified with this
disorder. Carnitine is an important component of the normal diet, and is also
produced endogenously. It is necessary for the transport of long-chain fats
into the mitochondrion, to be used in energy production. Babies and children
with this disorder have defective carnitine absorption in the intestine, and
also lose a lot in the urine. As a result they have very low blood and tissue
levels of carnitine, and this interferes with how they can use fat for energy.
The disorder most commonly causes a cardiomyopathy (poor heart muscle
function), and leads to heart failure and, if not treated, death.
In 1999, Baby Douglas had a newborn screening blood test, just as all other babies have done for over 30 years. But Douglas’s test was performed in the new way by tandem mass spectrometry. He was found to have a very low level of carnitine in the blood, and he too has the carnitine transporter defect. But things will be very different for Douglas. From the age of three weeks he was treated with extra carnitine which he takes as a pleasant tasting liquid. This will successfully treat the disorder and keep him healthy. He is one of a number of babies to have benefited already from the exciting new technology.
Newborn screening by tandem mass
spectrometry.
For all of New South Wales (NSW) and
the Australian Capital Territory (ACT), newborn screening is conducted at one
central laboratory at the Children’s Hospital at Westmead, Sydney. Similar
centralised laboratory services in Brisbane, Melbourne, Adelaide and Perth
cover the rest of Australia. In late 1997 at the Children’s Hospital Westmead,
we were able to purchase a tandem mass spectrometer for the New South Wales
Newborn Screening Programme. Tandem mass spectrometry is a “great leap
forward” for newborn screening. It enormously expands the number of disorders
that can be detected. At present we analyse a range of amino acids and
acylcarnitines in one dried blood spot, containing 3 microlitres of blood
(much less than a single drop), the test taking a little over 2 minutes per
sample. Apart from PKU, (which is now tested for with the new technology) we
can detect about 20 other disorders of amino acid and organic acid metabolism,
and 10 or more disorders of fatty acid oxidation. As our screening programme
tests all the babies born in NSW and ACT, over 90,000 babies each year. This
means that almost 400 babies are tested every working day. In fact, the
testing process runs overnight and takes over 14 hours to complete. In the
first 4 years of MS/MS testing in NSW we have detected over 50 babies with
rare treatable disorders, (in addition to the approximately 280 detected with
the less rare disorders previously tested for). South Australia started this
screening in 1999, and Victoria in 2002.
How can we evaluate newborn
screening tests?
A randomised controlled trial is a
powerful tool for investigating the merits of an intervention. Subjects are
randomly assigned to receive the intervention under investigation (in this
case it would be a completed newborn screening test) or to receive standard
medical care (in this case, no completed test). When drug treatments are
compared, usually the trial is “blind” , that is, neither the investigator nor
the patient knows who has been assigned to which group. The end-points need to
be defined: cure of an existing condition, reduction in deaths or defined
complications, or some other predetermined outcome measure. Importantly,
there need to be enough subjects for the differences in outcome to reach
statistical significance – and therein lies the problem for assessing the
effects of newborn screening tests. Because the disorders are rare, literally
millions of babies may need to be tested for there to be enough affected
babies to be statistically useful in assessing outcomes. The disorders sought
before MS/MS came were indeed rare. PKU occurs in NSW in 1:10,000 babies. The
commonest disorder sought is cystic fibrosis, occurring in about 1:2,500
babies. These disorders can be sought by mass screening because the tests are
relatively cheap. Of all newborn screening throughout the world, only for
cystic fibrosis has a randomised trial been carried out, in Wisconsin and in
the UK, and although the trial protocols were approved by ethics committees,
in the US a family is suing because of delayed diagnosis for their child, who
was randomised not to receive a completed screening test.
Bridget Wilcken is
Associate Professor and Director, NSW Biochemical Genetics and Newborn
Screening Services, New Children's Hospital, Westmead.