Are genetic tests for depression treatment worthwhile? Or is this an expensive technology that is not ready for routine use?
Peter Roy-Byrne, writing in NEJM Journal Watch seems to say no.
Although some clinicians may argue that such testing “can’t hurt and might help,” current psychopharmacological practice is complex, usually including combinations of multiple medications, and patients can have multiple comorbidities, both psychiatric and medical. Hence the meager available evidence, although promising, does not support routine use of these kits (although testing in some treatment-resistant patients might be reasonable). Even in medical specialties (e.g., oncology) that are much farther along in the understanding of genetic factors in illness pathophysiology, genetic testing to guide treatment is still in its infancy.
Since we’ve been using these tests for a couple of years in patients who have failed to or more previous treatments or who have demonstrated significant and unusual adverse effects to treatment we wanted to get our hands on the article and read it carefully to see if we agreed with Dr. Roy-Byrne’s summary.
The review in Lancet begins by distinguishing between genetic tests and clinical support tools. A genetic test, like the Amplichip from Roche Pharmaceuticals, determines a person’s genotype, but does not provide any guidance about what the results mean.
Given the complexity of the literature on genetics, most clinicians do not have the ability to make clinical decisions based on genetic test results.
A pharmacogenetic decision-support tool takes the results from a panel of genetic tests and provides a clinical interpretation of the results that can be used to guide treatment selection.
The first question in assessing the value of a decision-support tool, is the selection of genetic tests included in the panel.
In the review, the author identified a total of 46 genes included in one or more of the 21 decision support tools available worldwide. Of the 21 tools, 15 were available in the United States.
The complete list of genes includes 18 pharmacokinetic genes (genes that influence how drugs are metabolized) which are colored blue below and 28 pharmacodynamic genes (genes that influence how the body responds to medications) which are colored in red. There was much greater consistency in the choice of pharmacokinetic genes (9 out of 18 were included in more than 20% of the assays – these are bolded below) than in the pharmacodynamic genes (5 out of 28 were in 20% or more of the assays, also bolded below).
ABCB1=ATP-binding cassette, sub-family B, member 1.
ABCC1=ATP-binding cassette, sub-family C, member 1.
ADRA2A=adrenoceptor α 2A.
AKT1=v-akt murine thymoma viral oncogene homologue 1.
BDNF=brain-derived neurotrophic factor.
CACNA1C=calcium channel, voltage-dependent, L type, α 1C subunit.
CACNG2=calcium channel, voltage-dependent, gamma subunit 2.
CYP1A2=cytochrome P450, family 1, subfamily A, polypeptide 2.
CYP2B6=cytochrome P450, family 2, subfamily B, polypeptide 6.
* CYP2C9=cytochrome P450, family 2, subfamily C, polypeptide 9.
** CYP2C19=cytochrome P450, family 2, subfamily C, polypeptide 19.
** CYP2D6=cytochrome P450, family 2, subfamily D, polypeptide 6.
CYP3A4=cytochrome P450, family 3, subfamily A, polypeptide 4.
CYP3A5=cytochrome P450, family 3, subfamily A, polypeptide 5.
DDIT4=DNA-damage-inducible transcript 4.
DRD2=dopamine receptor D2.
DRD3=dopamine receptor D3.
* DPYD=dihydropyrimidine dehydrogenase.
EPHX1=epoxide hydrolase 1.
F2=coagulation factor II.
F5=coagulation factor V.
FHSD1=facio-scapulo-humeral dystrophy 1.
GRIA3=glutamate receptor, ionotropic, AMPA 3.
* G6PD=glucose-6-phosphate dehydrogenase.
* GRIK4=glutamate receptor, ionotropic, kainate 4.
HLA-A=major histocompatibility complex, class I, A.
** HLA-B=major histocompatibility complex, class I, B.
HTR2A=serotonin receptor 2A.
HTR2C=serotonin receptor 2C.
* IFNL3=interferon, lambda 3.
OPRM1=opioid receptor, mu 1.
RPTOR=regulatory associated protein of MTOR, complex 1.
* SLCO1B1=solute carrier organic anion transporter family, member 1B1.
SULT4A1=sulfotransferase family 4A, member 1.
* TPMT=thiopurine S-methyltransferase.
UGT1A1=UDP glucuronosyltransferase 1 family, polypeptide A1.
UGT1A4=UDP glucuronosyltransferase 1 family, polypeptide A4.
UGT2B15=UDP glucuronosyltransferase 2 family, polypeptide B15.
UGT2B7=UDP glucuronosyltransferase 2 family, polypeptide B7.
* VKORC1=vitamin K epoxide reductase complex, subunit 1.
Blue color – pharmacokinetic genes. Bolded meant included in more than 20% of tools.
Red color – pharmacodynamic genes. Bolded if included in more than 20% of tools.
* Single star if the evidence supporting use met the highest quality.
** Double star if the evidence was of the highest quality and relevant to many psychotropics routinely used.
Having noted the significant heterogeneity in the selection of genes in the panel, the authors then reviewed the evidence that panels can improve outcomes.
For most of the panels there was limited evidence and not much evidence of an interest in collecting this evidence. However three of the tools had evidence supporting them, and this is summarized below.
Three tools (GeneSight, Genecept, and CNSDose), currently available only in the USA or Australia, are building a relatively strong evidence base for pre-emptive (ie, before first prescription) use in psychiatry via clinical trials or cost-effectiveness studies…
The GeneSight tool was examined in two prospective cohorts, one double-blind, randomised clinical trial (RCT), two cost-savings studies, and one cost-effectiveness study.
The two prospective cohort studies recruited patients (n=4416 and n=16515) with major depressive disorder (MDD) and showed between two and four times improvement in depressive symptoms among patients receiving treatment guided by GeneSight compared with patients receiving unguided treatment.
In an independent double-blind RCT of 49 MDD patients, five (20%) of 26 patients who received treatment guided by GeneSight achieved symptom remission (ie, Hamilton Depression Rating Scale-17 score less than or equal to 7) compared with 8% of patients receiving treatment as usual.10 Although this equates to a 2·5 times difference in remission rates, nine patients would need to be genotyped for one patient to remit. This number needed to genotype might be acceptable given findings from cost-saving and cost-effectiveness studies of the GeneSight decision-support tool.
People who received GeneSight-guided treatment saved an estimated $1035 in total medication costs and $5188 in direct and indirect health-care costs compared with those receiving treatment as usual.
The Genecept tool was evaluated in a large (n=685) 3-month, naturalistic, unblinded, prospective cohort as well as in a retrospective adherence and health-care costs
study. The naturalistic cohort study showed significant decreases in depression and anxiety symptoms as well as decreases in side-effects. However, a treatment as usual comparator was not included, inhibiting any firm conclusions about Genecept’s effectiveness. The retrospective study examined health claims data in two independent cohorts and found those receiving Genecept guided treatment had a 6% greater increase in medication adherence compared with individuals receiving treatment as usual and demonstrated a relative saving of 9·5% ($562) in outpatient costs over a 4-month period.
The CNSDose tool was examined in a 12-week, double blind RCT in 148 white adults with MDD. Individuals who received CNSDose-guided prescribing remitted 2·5 times more than those receiving unguided prescribing. The number needed to genotype was only three, suggesting the CNSDose tool might be more cost-effective than GeneSight. However, to our knowledge no head-to-head comparisons of these tools or any others have been conducted.
We have had experience with GeneSight and Genecept and have found that using these tools provides information that is useful in about a third of the patients we test – patients who have failed two or more antidepressants (or psychotropic agents) or have had extreme and unusual side effects to medications.
Data from the STAR*D NIMH sponsored trial on the treatment of depression suggests that around 15% of patients who have failed two trials will respond to a third agent (or augmentation), and it is this group of patients who were shown to preferentially respond to TMS as opposed to another medication trial (and TMS costs around 30,000$ for a course of treatment) so it seems reasonable to use these tools (which cost between 1500 and 3500$) to guide selection of a medication if the tools can significantly increase the response rate.
Bousman CA, Hopwood M. Commercial pharmacogenetic-based decision-support tools in psychiatry. Lancet Psychiatry. 2016 Jun;3(6):585-90. doi: 10.1016/S2215-0366(16)00017-1. Epub 2016 Apr 25. Review. PubMed PMID: 27133546.