Pharmacogenetic tests: the need for a level playing field
a level playing fieldMunir Pirmohamed and Dyfrig A. Hughes
The delivery of more personalized medicine could be accelerated by addressing the substantial differences in the level of evidence required for the inclusion of pharmacogenetic tests in treatment guidelines, drug labelling and reimbursement schemes compared with that needed for non-genetic diagnostic tests.
There is an increasing drive, within both drug develop to those for patients with normal renal function. Indeed,
ment and clinical practice, to stratify or personalize such a stipulation would be impractical and costly, and
medicines to improve clinical outcomes for patients. would never be done during the drug development
Pharmacogenetics is perhaps the most welldeveloped process, potentially disadvantaging vulnerable patient
of the various technologies that can be used to personal populations.
ize medicines, but it has been criticized for not deliver
Another major determinant of variability in drug ex
ing on this promise. There are several sciencerelated po sure is genetic polymorphisms in drugmetabolizing
reasons why many pharmacogenetic studies do not enzymes and drug transporters. For example, cyto
progress beyond the stage of biomarker discovery, such chrome P450 2D6 (CYP2D6) is responsible for the
as inadequate sample sizes, poor clinical phenotype or metabolism of ~25% of drugs, but ~7% of the Caucasian
poor study design. However, we would also strongly population (termed poor metabolizers (PMs)) lack this
argue that the slow progress in the implementation of enzyme, whereas another 2% of Northern Europeans,
pharmacogenetic (and indeed other genetic) tests can 10% of Southern Europeans and up to 30% of some
partly be explained by the fact that different criteria are African populations carry more than two copies of the
applied when considering genetic testing compared gene, and are termed ultrarapid metabolizers (URMs).
with nongenetic diagnostic tests. Three specific areas Atomoxetine, a drug widely used for attention deficit
hyperactivity disorder, is metabolized in the liver by
First, drug response (efficacy or toxicity) is related CYP2D6. The states that the
to the dose administered, but more importantly to dose should be reduced by 50% in patients with hepatic
the systemic and cellular exposure to the drug and its impairment (ChildPugh class B), as drug exposure goes
metabolites. As most drugs and/or metabolites are elimi up by twofold. It is also known that drug exposure is
nated via the kidney and/or the liver, there are numer increased by a similar amount in CYP2D6 PMs; how
ous regulatory documents that provide guidance on ever, although the SmPC for atomoxetine mentions the
performing pharmacokinetic and pharmacodynamic effect of CYP2D6 polymorphisms, it does not mandate
assessments on patients with impaired liver or kidney testing for their presence. It has been stated that atomo
function and using this data to develop dosing recom xetine is equally efficacious and well tolerated among
mendations. For example, thpatients with the different CYP2D6 genotypes1, but
Personalised Medicine, University of Liverpool,
tates: “after an given the occurrence of numerous adverse effects in
initial usual dose, the dosage of aztreonam should be children, including arrhythmias and hepatotoxicity, can
halved in patients with estimated creatinine clearances we be sure? The role of such polymorphisms in predis
between 10 and 30 mL/min/1.73 m2”. There are many posing to toxicity from increased drug and/or metabo
other drugs that contain similar recommendations lite exposure is only now becoming clear, even with
within their labelling. Of course, this is logical in order old drugs such as codeine; there is increasing concern
to prevent variability in exposure and thus in the efficacy about the risk of respiratory depression in children who
and/or toxicity of a drug. There is no regulatory require are CYP2D6 URMs because of increased conversion
ment to undertake clinical trials to show that the dosing of codeine to morphine2. Notably, the new
recommendations for patients with, for example, renal
impairment are equivalent in terms of clinical outcomes has recommendations on
NATURE REVIEWS | DRUG DISCOVERY
VOLUME 12 | JANUARY 2013 | 3
2013 Macmillan Publishers Limited. All rights reserved
dosing evaluation in patients with polymorphisms in treated with carbamazepine — do not enjoy these ben
known metabolic pathways, an important pioneering efits. Subsidizing treatments to manage, but not tests
development that will begin to ensure harmonization of to prevent, serious conditions appears inconsistent.
dosing strategies as part of drug development.
Moreover, many treatments for rare diseases are very
The second example relates to the use of diagnostic expensive and exceed conventional thresholds of cost
tests. In general, the evidence required for implementa effectiveness. However, these are still available in most
tion of nongenetic tests in clinical practice seems to be markets, implying that more lenient criteria are applied
relatively haphazard and varies between genetic and non for reimbursement. Pharmacogenetic tests typical y cost
genetic tests. For example, renal impairment is widely less than their valuebased price (that is, the price that
regarded as a risk factor for hypersensitivity to the gout results in a costeffectiveness ratio of £30,000 per qual
drug al opurinol, with drug package inserts recommend ityadjusted lifeyear in the United Kingdom), and much
ing the use of lower doses in patients with renal impair less than the premium prices of many orphan drugs. As
ment. The origin of this association is from case reports with other diagnostics, however, their value is appropri
and case series in the 1970s and 1980s, and it has become ated to the treatments they accompany. Whereas the
widely accepted despite the fact that other studies have annual cost of abacavir is ~£6,100, for instance, the cost
shown no differences in adverse events according to cre of HLA-B*57:01
testing is £50 (less than 1% of the cost
atinine clearance, and there have been concerns about of abacavir). However, the value of testing is not lim
underdosing to control urate levels in clinical practice. ited to the reduction in the incidence of hypersensitivity
By contrast, there is now extensive evidence showing reactions: it also increased HIV clinicians’ confidence
that the presence of the HLA-B*58:01
allele predisposes in the drug and therefore led to more widespread use
to serious cutaneous adverse drug reactions to allopu of abacavir. Different strategies for setting the price of
rinol3, but this has not been implemented in guidelines pharmacogenetic tests may therefore be warranted, such
or as yet in drug package inserts. Prospective studies are that their true value is reflected in their price.
quoted as being necessary to demonstrate evidence of
Our arguments here are not meant to suggest that
clinical utility. Indeed, this has been done for two other special consideration should be given to genetic tests,
drugs that can cause severe hypersensitivity reactions: but more that the same level of evidence needs to be
the antiHIV drug abacavir4 (for which the presence of applied to genetic and nongenetic tests for adoption
al ele has been associated with hyper into guidelines and drug package inserts. Incentives for
sensitivity) and the antiepileptic drug carbamazepine5 the development of genetic tests, their pricing and reim
(for which the presence of the HLA-B*15:02
allele has bursement arrangements must also be aligned with other
been associated with a lifethreatening hypersensitivity health technologies.
disorder known as Stevens–Johnson syndrome). Part of 1. Sauer, J. M., Ring, B. J. & Witcher, J. W. Clinical pharmacokinetics
the reason for differences in evidential standards may
of atomoxetine. Clin. Pharmacokinet. 44
, 571–590 (2005).
be the familiarity and availability of nongenetic tests, 2. Ciszkowski, C., Madadi, P., Phillips, M. S., Lauwers, A. E. & Koren, G.
Codeine, ultrarapid-metabolism genotype, and postoperative death.
and the perceived low costs of these tests. However, it is
N. Engl. J. Med. 361
, 827–828 (2009).
important to note that genetic tests only need to be done 3. Zineh, I. et al.
Allopurinol pharmacogenetics: assessment of
potential clinical usefulness. Pharmacogenomics 12
once, whereas protein or metabolitebased tests, such as
renal function tests, need to be performed repeatedly. So, 4. Mallal, S. et al.
HLA-B*5701 screening for hypersensitivity
to abacavir. N. Engl. J. Med. 358
, 568–579 (2008).
over a lifetime, the costs of the nongenetic tests may be 5. Chen, P. et al.
Carbamazepine-induced toxic effects and
equivalent and indeed may become greater as the costs
HLA-B*1502 screening in Taiwan. N. Engl. J. Med. 364
A final difference between genetic and other health Disclaimer
The views expressed in this article are those of the authors, and not of any
technologies is apparent with respect to legislation on institutions that they represent.
orphan drugs in various regions. Such legislation pro Competing financial interests
vides incentives for manufacturers to develop treatments The authors declare no competing financial interests.
for lifethreatening or chronically debilitating diseases
that are rare (in the European Union, diseases that affect
European Medicines Agency publishes guideline on use of
no more than 5 in 10,000 people). The incentives range
pharmacogenetics in evaluating pharmacokinetics of medicines:
from reductions in regulatory fees to 10 years of mar
ket exclusivity once authorized (7 years in the United
States). However, manufacturers of genetic tests that
prevent equally severe and rare conditions — such as
SmPC for aztreonam: http://www.medicines.org.uk/EMC/medicine/549/SPC/
testing for the presence of HLA-B*15:02
to avoid the
development of Stevens–Johnson syndrome in patients
ALL LINKS ARE ACTIVE IN THE ONLINE PDF
| JANUARY 2013 | VOLUME 12
2013 Macmillan Publishers Limited. All rights reserved
Joseph Schifini, M.D. DATE ____/____/____ PLEASE FILL OUT THIS FORM COMPLETELY SO WE CAN HELP YOU WITH YOUR PAINFUL CONDITION . NAME :_____________________________ AGE :_____ HEIGHT ___ ' ___ " WEIGHT ______ lbs OCCUPATION :__________________________________________ DATE LAST WORKED ____/____/____ CC: WHY WERE YOU REFERRED TO OUR CLINIC?_____________
Information for Patients General information: The site where an implant has been placed will normally be sutured to reduce the amount of bleeding and begin the healing process. However, a clot will still need to develop under the gum and trouble-free healing is dependent on this initially fragile clot rejoining undisturbed as much as possible. Most implant sites heal with minimal trouble. For