Effects and Cost-Effectiveness of Pharmacogenetic Screening among Elderly Starters with Antidepressants: A Pragmatic Randomized Controlled Trial

Depression is common among elderly with an estimated prevalence of 5% and due
to ageing the national burden will double in the coming decade. Antidepressants
as TCAs and SSRIs are effective in reducing symptoms, particularly if the
symptoms are more serious as in psychiatric settings (Arroll B, 2009, Nelson et
al, 2008, Tedeschini et al, 2011)). In the Netherlands in 2009, already more
than 12,000 and 20,000 elderly patients aged 65 years or older were using
nortriptyline and venlafaxine, respectively, the current first choice drugs
according to the Dutch Guideline Depression (Addendum Elderly). Also the
American Geriatrics Society mentions these drugs in their Geriatrics At Your
Fingertips (Reuben et al, 2011). However, dose-finding in elderly is still too
slow. Although there are indications for applicability of CYP2D6 genotyping in
dose finding for nortriptyline and venlafaxine (Kootstra-Ros et al, 2006,
Vandel et al, 2007, Whyte et al, 2006), the evidence base for adequate tailor
made dosing and tolerability of antidepressant drugs is lacking (Pollock et al,
2009). Especially in this more seriously ill population, in initiating
treatment with antidepressants it is important to increase the dose as fast as
possible to the desired effective dose. This will increase adherence and
enhance drug efficacy. A too fast dose escalation may lead to strong adverse
effects and discontinuation of therapy.

In advanced age, the enzyme activity of hepatic enzymes involved in drug
metabolism is decreased (Bebia et al, 2004; Jin et al, 2010) as well as
delivery of medication from the blood to the hepatocytes as well as liver size
(Lotrich and Pollock, 2005). Additionally, CYP2D6 and CYP2C19 show
pharmacogenetic differences relevant to different drugs resulting in poor (PM),
intermediate(IM), extensive(EM) and ultrarapid metabolisers(UM). In
antidepressant therapy, genotyping for CYP2D6 only, allows the prediction of a
major proportion of the intra-individual variation of plasma levels of most
antidepressants (Kirchheiner 2004; Maier et al.2008) and genotyping for other
potential polymorphisms has not yet demonstrated an important role. The
Pharmacogenetics Working Group of the Royal Dutch Pharmacists Association
developed guidelines for dose-adaptation for drugs predominantly metabolised by
CYP2D6 and CYP2C19 based on the genotype of these enzymes (Swen et al, 2008 and
2011). In this guideline dose-interventions based on all deviating
genotypes(PM, IM, UM) for CYP2D6 are advised for nortriptyline and venlafaxine
accompanied by Therapeutic Drug Monitoring. However, data to support the
guideline for elderly are lacking. In general most TCAs are relatively
contraindicated in older persons (Beers criteria(American Geriatrics Society
2012) and STOPP/START criteria(Gallagher et al, 2008)). Although in the
guideline nortriptyline is a first choice drug within the tricyclic
antidepressants (TCA), serious side effects because of anticholinergic and
anti-*-adrenergic effects should be considered. Venlafaxine is first choice if
a TCA is contraindicated according to the Dutch Guideline Depression-addendum
Elderly. Additionally in this guideline application of Therapeutic Drug
Monotoring is emphasized, which is also corroborated for these drugs by
Mitchell (2004), Sjöqvist and Eliasson (2007), Wille et al (2008), Hiemke
(2008) and Preskorn (2010). Intermediate metabolisers are sensitive to minor
decrease in enzyme activity induced by enzyme inhibition by interacting
co-medication. Although speculative, our hypothesis is that the inhibitory
effect of older age on enzyme activity has a comparable effect on intermediate
metabolisers as interacting co-medication in the younger population.

As mentioned above depression is common among elderly and increasing in the
population due to aging. There is a lack of data on the impact of
pharmacogenetics on tailormade therapy with the guideline based antidepressants
nortriptyline and venlafaxine with respect to efficacy and tolerability in
elderly. It is however well documented that in the general population both
nortriptyline and venlafaxine are sensitive to polymorphism of CYP2D6 resulting
in dose changes up to 60% (Kirchheiner 2004; Maier et al. 2008; Swen et al.
2008 and 2009 and references mentioned therein). However, since a longer
duration of depression enhances the duration of an increased suicide-risk and
often results in a worse prognosis increasing the dose as fast as possible in
the more seriously ill population as adressed in this study is of major
clinical importance.
Also the role of polymorphisms in the multi-drug resistance gene ABCB1 which
encodes P-glycoprotein, a critical component of the Blood-Brain Barrier seems
to be implicated via the drug availability in the central nervous system in the
central actions of nortriptyline and venlafaxine (Roberts et al. 2002; Perroud
et al. 2011), but there are no data available in elderly.
This clinical trial will advance scientific knowledge on the added value of
pharmacogenetics to treatment of the elderly with nortriptyline or venlafaxine
with serum drug levels as a proxy for efficacy as often required by drug
registration authorities for special populations and emphasized in the Dutch
guidelines. From a patient and societal perspective we will also record adverse
events and non-adherence, quality of life and cost-effectiveness from a
societal perspective.

Primary Objective:
* To determine the effects of pharmacogenetic screening for CYP2D6 on the time
to reach adequate serum drug levels according to
the guidelines.

Secondary Objective(s):
* To determine the effects of pharmacogenetic screening for CYP2D6 on the
adverse events by using an abbreviated self-reported
Antidepressant Side Effect Checklist (ASEC),
self-reported functional health status and general and disease-specific quality
of life
by means of EQ5D (including VAS).
* To determine the incremental cost-effectiveness ratio for intervention versus
control strategy from a societal perspective including
uncertainty analysis (bootstrapping, Fieller*s
estimates), discounting, scenario analysis, (probabilistic) sensitivity
analysis and
presentation in cost-effectiveness acceptability curves.
* To determine the effects of the genotype for ABCB1 alone or in combination
with CYP2D6 on the serum drug levels.
* To determine the effects of the genotype for ABCB1 alone or in combination
with CYP2D6 on the adverse events
by using an abbreviated self-reported Antidepressant Side
Effect Checklist (ASEC), self-reported functional health status
and general and disease-specific quality of life by means
of EQ5D (including VAS).

Design: To determine the value of pharmacogenetic screening when starting
anti-depressants among the elderly in a real world setting, we propose the
conduct of a pragmatic randomized controlled trial. The trial is designed
according to the CONSORT guidelines (Begg et al, 1996).

Setting and study population: All patients aged 60 years and older with a
depression in whom drug treatment with either nortriptyline or venlafaxine is
started in any of the participating psychiatric clinics in the Netherlands are
eligible for first screening to be included into the trial. Use of clinically
relevant CYP2D6 inhibitors (e.g. fluoxetine, paroxetine, quinidine, ritonavir,
efavirenz) or inducers (e.g. dexamethasone, rifampicine) or other drugs that
affect plasma levels as co-medication is an exclusion criterion, since dosing
schedules need to be adapted differently.

First pharmacogenetic screening: After informed consent has been given in
writing for pharmacogenetic screening, blood will be drawn from eligible
participants by use of a minimal invasive, but clinically validated Dry Blood
Spot Method (drop of blood from finger on filtration paper; De Boer et al,
2011, Edelbroek et al, 2009, Wijnen et al, 2008). Based on their genetic
information on CYP2D6, patients will then be classified as poor, intermediate,
extensive (normal) or ultrarapid metaboliser. Then, after randomization (see
below) informed consent will be asked in writing for participation in the part
of the trial in which an intervention based on genotype is applied.

Randomization and intervention: Patients with a deviant polymorphism for which
the guideline advices to adapt the standard dosing schedule (nortriptyline: PM,
IM, UM; venlafaxine: PM, IM, UM), will be randomly allocated by computer to an
intervention or control strategy according to daily practice. Based on
extensive literature (Tamminga et al, 1999 and 2000, Raimundo et al, 2000,
Sachse et al 1997, Bradford 2002, Zanger et al, 2004, Van Schaik et al, 2006)
we expect that about 20 to 30% within this group of starters will be eligible
for randomization. In the intervention group, an advice for dose adaptation
based on blood level of the drug and/or the genotype will be communicated by a
dedicated research team to the treating physician. In the control group, an
advice for dose adaptation based on blood level of the drug only will be given
to the physician according to current daily practice. Therefore the additional
burden for the patients caused by participating in this trial will be
restricted to the fact of the randomization, but will be alleviated in
comparison to standard care by taking blood by a fingerprick instead of
intravenously.

External control study group: From the first screened study population we will
randomly sample 75 extensive metabolisers for CYP2D6 and follow them up
similarly as those within the trial. This group will provide reference data on
outcomes for the deviant patient groups.

Clinical follow-up: At the second scheduled visit a blood spot will be taken
for assessment of the serum drug level. Thereafter, blood spots will be taken
at the next visits by the treating physician for assessment of the serum drug
level according to daily practice. Such sampling will start at least one week
after dose adaptation and will be repeated until an adequate drug serum level
is reached. Normally it takes approximately two weeks to reach such levels. It
is expected that in these deviant patients the time doubles to 4 weeks. Our
hypothesis is that by giving dose adaptation based on genotypes, the time to
reach adequate drug levels will be halved to be comparable with controls
without deviant genotypes.

Duration: 2- 4 weeks for individual patients

The risk of participation in the first part of the study is limited to a finger
prick to obtain blood for pharmacogenetic screening.
The patients who will participate in the second/main part of the study will be
alleviated in comparison to standard care by taking blood for the assessment of
the serum drug level according to daily practice by a finger prick instead of
intravenously. Furthermore the following data are collected extra for the
study: self-reported functional health status and general and disease-specific
quality of life by means of EQ5D (including VAS), data on health care
associated resource use (e.g. visits to the specific specialists including
diagnoses; drug use including exact dosing and durations of prescriptions;
hospitalizations, inclusive exact intensities of care; and lab values if
relevant).