Journal of Psychiatric Research 96 (2018) 100e107

Contents lists available at ScienceDirect

Journal of Psychiatric Research

journal homepage: www.elsevier.com/locate/psychires

Improved efficacy with targeted pharmacogenetic-guided treatment

of patients with depression and anxiety: A randomized clinical trial
demonstrating clinical utility
Paul Bradley a, h, Michael Shiekh b, Vishaal Mehra c, Keith Vrbicky d, e, f, Stacey Layle c,
Marilyn C. Olson g, Alejandra Maciel g, *, Ali Cullors g, Jorge A. Garces g,
Andrew A. Lukowiak g
a
BCG Medical Group, Internal Medicine, Mercer University School of Medicine, Savannah, GA, United States
b
Relaro Medical Trials, Dallas, TX, United States
c
Artemis Institute for Clinical Research, San Diego, CA, United States
d
Midwest Health Partners, P.C., Norfolk, NE, United States
e
Creighton University, Department of OB/GYN, United States
f
University of Nebraska Medical Center, Department of OB/GYN, United States
g
AltheaDx, San Diego, CA, United States
h
Meridian Clinical Research, Savannah, GA, United States

a r t i c l e i n f o a b s t r a c t

Article history: The objective of this study was to evaluate the effect of pharmacogenetics-guided treatment on patients
Received 24 June 2017 diagnosed with depression and/or anxiety, in a diverse set of clinical settings, as compared to the
Received in revised form standard of care. The trial design followed a prospective, randomized, subject- and rater-blinded
22 August 2017
approach enrolling 685 patients from clinical providers specializing in Psychiatry, Internal Medicine,
Accepted 22 September 2017
Obstetrics & Gynecology, and Family Medicine. The NeuroIDgenetix® test uses a genetic variant panel of
ten genes, along with concomitant medications, to make medication management recommendations
Keywords:
based on gene-drug and drug-drug interactions for over 40 medications used in the treatment of
Pharmacogenetics
Depression
depression and anxiety. Pharmacogenetic testing was performed at the initial screening visit and
Anxiety baseline patient assessments were determined using the 17-item Hamilton Rating Scale for Depression
Efficacy (HAM-D17) and the Hamilton Rating Scale for Anxiety (HAM-A). Following enrollment and randomi-
Treatment response zation, pharmacogenetic results for subjects assigned to the experimental group were provided to
Precision medicine physicians to guide treatment selection, while control subjects were treated according to the usual
standard of care. HAM-D17 and HAM-A assessments were collected at 4 weeks, 8 weeks, and 12 weeks
after baseline to assess the efficacy of therapeutic selection. In patients diagnosed with depression,
response rates (p ¼ 0.001; OR: 4.72 [1.93e11.52]) and remission rates (p ¼ 0.02; OR: 3.54 [1.27e9.88])
were significantly higher in the pharmacogenetics-guided group as compared to the control group at 12
weeks. In addition, patients in the experimental group diagnosed with anxiety showed a meaningful
improvement in HAM-A scores at both 8 and 12 weeks (p ¼ 0.02 and 0.02, respectively), along with
higher response rates (p ¼ 0.04; OR: 1.76 [1.03e2.99]). From these results, we conclude that
pharmacogenetic-guided medication selection significantly improves outcomes of patients diagnosed
with depression or anxiety, in a variety of healthcare settings.
© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Approximately 16 million Americans suffer from moderate or

* Corresponding author. AltheaDx, 10578 Science Center Drive, San Diego, CA
92121, United States.
E-mail address: AMaciel@altheadx.com (A. Maciel).
severe depression, a medical condition that affects mood, cognitive
and physical symptoms (Pratt and Brody, 2014). The costs of in-
dividuals with Major Depressive Disorder is greater than $210

https://doi.org/10.1016/j.jpsychires.2017.09.024
0022-3956/© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 P. Bradley et al. / Journal of Psychiatric Research 96 (2018) 100e107
billion annually including direct costs such as medical and phar-
maceutical services in addition to workplace costs (Greenberg et al.,
2015). One in eleven US adults reports the use of an antidepressant
medication in the past 30 days, making antidepressants the third
most commonly prescribed drug type (US Dept. Health and Human
Services, Centers for Disease Control and Prevention, 2013). Despite
their broad use in treating depression, initial success in medication
selection can be low. Findings from the Sequenced Treatment Al-
ternatives to Relieve Depression (STAR*D) study demonstrated that
only one-third of the patients achieved remission during the first
treatment level (Trivedi et al., 2006). Furthermore, patients with
difficult-to-treat depression may achieve remission after trying
several treatment strategies, but the odds diminish with every
additional treatment strategy demonstrating the limitations of
trial-and-error medication management (Trivedi et al., 2006;
Huynh and McIntyre, 2008).
While often overshadowed by depression, anxiety disorders are
the most prevalent mental disease in the United States with an
estimated 40 million affected adults age 18 or older and account for
more than $42 billion in annual healthcare costs due to frequent
medical evaluations and treatment of physical manifestations of
the disorder (Bystritsky, 2006; Anxiety and Depression Association
of America, 2016). Approximately one-third of those suffering from
an anxiety disorder receive treatment with only about 60% of those
patients responding to treatments (Bystritsky, 2006). Further
complicating the treatment of anxiety is that it is frequently part of
a dual diagnosis with depression (comorbidity estimated at 50%),
decreasing the odds of remission and increasing the response time
to pharmacotherapy (Anxiety and Depression Association of
America, 2016; Wittchen et al., 2002; Kessler et al., 2010; Fava
et al., 2008).
Therapeutic selection today relies on an empirical approach
using physician experience in combination with clinical indicators
to determine treatment options, resulting in a “trial-and-error”
approach to medication management (Mrazek, 2010; Rush et al.,
2006). In contrast, pharmacogenetic testing can help reduce the
uncertainty inherent in psychiatric pharmacotherapy by detecting
genetic factors that predict clinical response and side effects, such
as genetic variations that impact drug-metabolizing enzymes, drug
transporters or drug targets (Evans and Relling, 1999; Evans and
Johnson, 2001; Altar et al., 2013; Porcelli et al., 2011). These asso-
ciations are well supported in the scientific literature and are
becoming prevalent in professional guidelines and drug labeling
(US Food and Drug Administration, 2016; Caudle et al., 2014; Hicks
et al., 2015). In fact, the American Psychiatric Association guidelines
for the treatment of major depressive disorder acknowledge indi-
vidual pharmacokinetic and pharmacodynamic differences among
individuals that may necessitate doses higher than those approved
by the FDA to achieve therapeutic benefit (Gelenberg et al., 2010).
Additionally, FDA guidance for new drug approvals further sup-
ports the importance of pharmacogenetic testing to reduce adverse
drug events (ADEs) and personalize dosing (US Food and Drug
Administration, 2013).
Previous studies have explored the clinical impact of pharma-
cogenetic testing in the management of patients diagnosed with
depression. In a review of 465 published records conducted by
Helfand et al., in 2016, two randomized clinical trials were identi-
fied that compared the effectiveness of pharmacogenetic-testing to
standard of care (Helfand et al., 2016). Winner et al., found
improved remission rates in patients with previous failed treat-
ments but the results did not reach statistical significance (Gene-
sight: 20% vs 8%, RR 2.40, N ¼ 51) (Winner et al., 2013). Singh
reported improved remission rates in a pharmacogenetic-guided
group compared to standard of care among Australian patients
with baseline HAM-D scores of 25 (CNSDose: 72% vs. 28%; RR 2.52,
101
95% CI 1.71e3.73, N ¼ 148) (Singh, 2015). While informative, the
narrow patient population and the low sample number of these
studies limits the applicability of the results to the general popu-
lation. For example, both Winner and Singh exclude patients with
comorbid psychiatric disorders and enroll patients only seen at
psychiatric clinics. Given the high number of non-psychiatric cli-
nicians prescribing anti-depressants and the high rate of comor-
bidity for depression and anxiety, a clinical study examining the
clinical utility of pharmacogenetics in depression patients should
aim to reflect the real-world clinical setting.
The present study utilizes the NeuroIDgenetix® pharmacoge-
netic test to detect and interpret the impact of an individual pa-
tient's genetic variants associated with pharmacokinetic and
pharmacodynamic response, as well as metabolic interactions
related to concurrent medications, over the counter drugs, herbal
supplements, and diet. The NeuroIDgenetix® Test uses a genetic
variant panel of ten genes, along with concomitant medications, to
make medication management recommendations based on gene-
drug and drug-drug interactions for over 40 medications used in
the treatment of depression and anxiety (see Supplemental Table 1
and Supplemental Table 2). This study was designed to determine
whether incorporation of the pharmacogenetic test report can
improve outcomes in the treatments of anxiety and depression and
address shortcomings of current prescribing practices. Specifically,
this study compares the therapeutic outcomes of pharmacogenetic-
guided treatment in patients diagnosed with depression or anxiety
with standard of care medication management.
2. Methods
2.1. Study design
This multicenter study (registered at www.clinicaltrials.gov
under identifier number NCT02878928) was designed to deter-
mine if pharmacogenetics can help guide therapeutic decisions for
patients with depression and/or anxiety disorders and improve
patient outcomes by maximizing drug efficacy. The trial was con-
ducted across 20 independent clinical sites within the US special-
izing in Psychiatry, Internal Medicine, Obstetrics and Gynecology,
and Family Medicine. Study participants were randomized by dis-
ease and severity and allocated in a 1:1 ratio to the experimental
group (guided by the NeuroIDgenetix® test) or control group
(standard of care). For patients within the experimental group, the
NeuroIDgenetix® test report was released to participating clinicians
to consider the recommendations from the report when making
medication decisions. For the control group, patients were sub-
jected to sample collection procedures identical to the experi-
mental group, however NeuroIDgenetix® results were withheld
and clinicians were asked to manage control subjects according to
the standard of care. Test results were provided before the first
study visit at 2 weeks (±1 week), to guide medication changes in
the experimental group. Medication changes for both groups
occurred by the 2 week visit. Patients were monitored and assessed
for depression and anxiety symptoms using HAM-A and HAM-D17
interviews by trained raters at study initiation (baseline) and at the
4, 8, and 12 week follow-up visits. All patients and raters were
blinded to the patient's assigned study group until the end of the
trial. Medication changes in both groups (new medications,
discontinuation, and dose adjustments) and prescription use were
collected at all study visits, as were the occurrence of ADEs, hos-
pitalizations, medical visits, disability, and missed days of work.
2.2. Study Participants
A total of 764 patients were assessed for eligibility, 79 subjects
102 P. Bradley et al. / Journal of Psychiatric Research 96 (2018) 100e107
were excluded resulting in the enrollment and randomization of
685 patients for the study (Fig. 1). Study participants included pa-
tients between the ages of 19 and 87 with a diagnosis of depression
and/or anxiety using the DSM-V criteria or standard of care site
procedures and verified by the MINI Psychiatric Interview (Sheehan
et al., 1998). Participants were either ‘New to Treatment’ (newly
diagnosed or taking medications for less than 6 weeks) or ‘Inade-
quately Controlled’ with medications as defined by lack of efficacy
or treatment discontinuation due to adverse events or intolera-
bility. Exclusion criteria consisted of a concurrent diagnosis of bi-
polar disorder, schizophrenia, personality disorder, traumatic
physical injury (i.e., traumatic brain injury), significant risk for
suicide and hospitalization, patients with a history of chronic renal
dysfunction or Chronic Kidney Disease (Stage 4 or 5), malabsorp-
tion (short gut syndrome), pregnancy, or abnormal hepatic function
(INR >1.2 not attributable to anticoagulant medications, AST/
aspartate aminotransferase or ALT/alanine aminotransferase >1.5
normal, or suspected cirrhosis). This study was carried out in
accordance with the Declaration of Helsinki, approved by an
institutional review board (Aspire IRB) and all participants signed
an approved consent form prior to initiating study activities. All
subjects were monitored over a 12 week period.
2.3. Pharmacogenetic testing
Buccal swabs were collected from all patients prior to randomi-
zation and shipped to AltheaDx (San Diego, CA) for processing.
Samples were processed in accordance with the Centers for Medi-
care and Medicaid Services (CMS) regulations in a College of Amer-
ican Pathologists (CAP)-accredited, Clinical Laboratory Improvement
Amendments (CLIA)- certified laboratory (AltheaDx). In summary,
genomic DNA was extracted from buccal samples using the Qiagen
DSP DNA Midi Kit (Qiagen, Valencia, CA, USA). Pharmacokinetic and
Pharmacodynamic variants associated with 10 genes were tested,
including CYP1A2, CYP2C9, CYP2C19, CYP2D6, CYP3A4, CYP3A5,
SLC6A4 [NM_001045.5:c.-1760C > T], SLC6A4 [5-HTTLPR], COMT
[NM_000754.3:c.472G > A], HTR2A [NM_000621.4:c.-998G > A],
Fig. 1. Study participants.
HTR2A [NM_000621.4:c.614e2211T > C], MTHFR
[NM_005957.4:c.665C > T], MTHFR [NM_005057.3:c.1286A > C] (see
Supplemental Table 1). The variants tested represent an ethnically-
diverse set of genetic polymorphisms, appropriate for the US pop-
ulation (see Supplemental Table 1). Molecular methods used for
detection included end-point PCR, real-time PCR for CYP2D6 copy
number variation determination, and capillary electrophoresis for
the genotyping of SLC6A4 5-HTTLPR (Pratt et al., 2010; Lesch et al.,
1996). In the case of CYP genes, genotypes are translated into dip-
lotypes and then diplotypes are assigned a metabolic phenotype for
clinical management. As an example, a subject homozygous for the
CYP2D6 1846G > A variant, would be assigned a CYP2D6*4/*4 dip-
lotype and a corresponding “Poor Metabolizer” phenotype for the
purpose of medication selection. Complementing this pharmacoge-
netic analysis, the IDgenetix® algorithm also screens for potential
metabolic interactions between concomitant medications as well as
a variety of lifestyle factors; including the use of alcohol, tobacco, and
herbal supplements (see Supplemental Fig. 1 for an example of the
NeuroIDgenetix® Report).
NeuroIDgenetix® reports were sent to the treating physicians,
who were blind to the patients' Hamilton Scores, for patients in the
experimental group within approximately 3 days after sample
receipt by the laboratory. The NeuroIDgenetix® report classifies
therapeutically related medications as one of two options; “Use as
Directed” (UAD) or “Use with Caution and/or Increased Monitoring”
(UWC). The medications classified as “Use as Directed” may be
administered in accordance to the manufacturer's standard pre-
scribing information since among the genes analyzed and current
drug regimen, no genetic variants or metabolic interactions were
identified that would suggest a need for increased caution or dose
adjustment. Medications classified as “Use with Caution and/or
Increased Monitoring” have been identified by the IDgenetix® al-
gorithm as having one or more reasons for avoidance, including
possible drug-gene and drug-drug interactions. These warnings are
accompanied by brief descriptions of the reasons for caution (e.g.,
lack of efficacy or toxicity) along with recommendations for
appropriate clinical action.
 P. Bradley et al. / Journal of Psychiatric Research 96 (2018) 100e107
2.4. Patient assessments
Patients diagnosed with depression were graded using the 17
item Hamilton Depression Rating Scale (HAM-D17) to determine
depression severity at baseline and at the 4, 8 and 12 week follow-
up visits (Hamilton, 1960; Cusin et al., 2010). The HAM-D17 inter-
view was administered by trained clinicians blinded to treatment
decisions and patient status throughout the study. Total scores
were categorized as of 0e7 normal, 8e17 mild depression, 18e24
moderate depression, and >24 severe depression. Patients diag-
nosed with anxiety were assessed for symptoms and severity of
anxiety at baseline and at the 4, 8 and 12 week follow-up visits
using the Hamilton Rating for Anxiety (HAM-A) (Hamilton, 1959;
Goodman, 2004). Similar to the HAM-D17 assessments, the HAM-
A interviews were conducted by trained and blinded clinicians
throughout the study. Total scores of 8e17 indicated mild anxiety,
18e24 moderate anxiety and >24 severe anxiety.
2.5. Adverse drug events
The Adverse Drug Event form used in this clinical study was
developed and reviewed by external clinical psychiatric consultants
and used uniformly by all participating sites. The form is divided
into two sections. The first section is completed during a face to face
interview by the study coordinator. The type of ADE (e.g. nausea,
dry mouth, weight gain), suspected drug and start date as well as
the event onset and end date (if applicable) are collected. The
second section is completed by a blinded clinician and involves
grading the seriousness of the ADE on a scale of mild, moderate,
severe, life-threatening or death, and documenting the manage-
ment and outcome of the ADE. Lastly, the clinician evaluates the
likelihood that the suspected drug was the actual cause of the
adverse event. The assessment of causality is rated as one of the
following: unrelated, unlikely, possibly, probably or definitely
related to the suspected drug. All ADE forms were completed by
trained and blinded clinicians and completed at study enrollment
and at each follow-up visit.
2.6. Statistical analysis
This study was designed to determine the clinical utility of
pharmacogenetics-guided treatment (experimental group) as
compared to standard of care treatment (control group) by
measuring the changes in the severity of depression and anxiety
symptoms. The efficacy endpoints included HAM-D17 changes for
patients with depression and HAM-A changes for patients with
anxiety measured at baseline and the 4, 8, and 12 week follow-up
visits. Per the study protocol, only those patients with moderate
or severe Hamilton scores (18) at baseline were included in the
efficacy analysis. HAM-D17 and HAM-A scores from patients with
both depression and anxiety were included in both the HAM-D17
and HAM-A analyses. The impact of pharmacogenetics guided
treatment on response rates (50% reduction in HAM scores) and
remission rates (HAM scores 7) were evaluated using Fisher's
exact test and logistic regression analyses were used to compare
response and remission rates between the experimental group and
control group. Response and remission rates were defined and
calculated as a subject achieving response or remission at any time
point after baseline. Changes in the HAM-D17 and HAM-A scores
were compared between the experimental and control groups by t-
test and also in a regression analysis accounting for baseline scores
and comorbidity of depression and anxiety. In addition, a mixed
model repeated measures (MMRM) analysis was performed using
all available data from baseline to the 12 week visit.
103
3. Results
3.1. Participant characteristics
A total of 764 patients diagnosed with depression and/or anxi-
ety were eligible to participate in this study of which, 79 were
excluded due to not meeting inclusion/exclusion criteria or
declining participation (Fig. 1). The remaining 685 patients were
randomized and allocated to either the experimental group
(n ¼ 352) or the control group (n ¼ 333) and followed for period of
up to 12 weeks. At baseline and each follow-up visit (4, 8 and 12
weeks) patients were assessed for depression and/or anxiety
symptoms. Patients were allowed to miss up to one visit and
remain in the study. A total of 579 patients (297 experimental, 282
control) completed the study.
Study participants had a mean age of 48 years, with approxi-
mately 73% female and 27% male, approximating a gender bias
distribution consistent with other reports (Table 1). Patients were
categorized into three different diagnosis categories: depression
(n ¼ 246), anxiety (n ¼ 235) and both depression and anxiety
(n ¼ 204). HAM-D17 assessments were conducted for patients
diagnosed with depression and depression and anxiety. HAM-A
assessments were conducted for patients with anxiety and pa-
tients with both depression and anxiety. Baseline HAM-D17 and
HAM-A scores were similar between the experimental and control
groups as shown in Table 1. The distribution of metabolizer phe-
notypes, based on Clinical Pharmacogenetics Implementation
Consortium (CPIC) guidelines (Whirl-Carrillo et al., 2012), and other
genotypes were statistically indistinguishable between the exper-
imental and control groups (Supplemental Table 3).
3.2. Depression outcomes
To determine whether NeuroIDgenetix®-guided treatment can
improve outcomes for patients diagnosed with moderate or severe
depression at baseline, we compared HAM-D17 scores between the
experimental and control groups at 4, 8 and 12 weeks of follow-up.
Remission rates at the 8 week follow-up visit for experimental
patients with severe depression were 25%, whereas the control
group remission rates were 9% (Fig. 2a). At the 12 week visit,
remission rates for experimental patients with severe depression
were 35%, 95% CI [21%, 52%], whereas the control group remission
rates were 13%, 95% CI [5%, 25%], p ¼ 0.02, OR: 3.54 (1.27e9.88),
NNT ¼ 4.6 (Fig. 2a).
Response rates were significantly higher for patients in the
experimental group as well. At the 8 week follow-up visit, the
response rate was higher in the experimental group compared to
the control group (55% versus 28% respectively) for patients diag-
nosed with severe depression (Fig. 2b). At the 12 week follow-up
visit, response rates trended even higher in the experimental
group (73%, 95% CI [56%, 85%]) as compared to the control group
(36%, 95% CI [23%, 50%], p ¼ 0.001, OR: 4.72 (1.93e11.52),
NNT ¼ 2.7). When both moderate and severe patients were
included in the analysis (HAM-D17  20), 8 week response rates
were 49% for the experimental group and 41% for the control group.
The 12 week follow-up visit response rates were significantly
higher for patients in the experimental group compared to the
control group, 64%, 95% CI [55%, 72%], vs. 46%, 95% CI [37%, 56%],
p ¼ 0.01, OR:2.03 (1.23e3.33), NNT ¼ 5.8 (Fig. 2c).
Our study did not find significant improvement in patients with
mild depression. This finding aligns with research suggesting that
patients with mild depression may not benefit from anti-
depressant medication (Barbui et al., 2011).
104 P. Bradley et al. / Journal of Psychiatric Research 96 (2018) 100e107

Table 1
Demographic and clinical characteristics of study participants.

Experimental Group (N ¼ 352) Control Group (N ¼ 333)

Age
Years 47.8 (±14.5) 47.3 (±15.2)
Gender
Female 257 [73%] 241 [72%]
Male 95 [27%] 92 [28%]
Ethnicity
African-American 63 [18%] 59 [18%]
Asian 5 [1%] 4 [1%]
Caucasian 221 [63%] 209 [63%]
Hispanic 57 [16%] 58 [17%]
Other 6 [2%] 3 [1%]
Total Population
Depression (n ¼ 450) 237 213
Baseline HAM-D17 Mean (S.D) 20(±5.8) 20(±5.6)
Baseline HAM-D17 Median 20 20
Anxiety (n ¼ 439) 219 220
Baseline HAM-A Mean (S.D) 25(±7.4) 25(±7.1)
Baseline HAM-A Median 25 24
Moderate and Severe Population(s)
Depression HAM-D17  25 (n ¼ 93) 40 53
Depression HAM-D17  20 (n ¼ 261) 140 121
Anxiety HAM-A 18 (n ¼ 396) 200 196
Anxiety Only HAM-A  18 (n ¼ 224) 109 115

Out of 352 subjects in the experimental group, 133 were classified as depression only, 115 were classified as anxiety only, 104 were classified as depression
and anxiety, leading to a total of 237 (133 þ 104) patients in the experimental arm of depression and 219 (115 þ 104) in the experimental arm of anxiety. Out
of 333 subjects in the control group, 113 were classified as depression only, 120 were classified as anxiety only, 100 were classified as depression and anxiety,
leading to a total of 213 (113 þ 100) patients in the control arm of depression and 220 (120 þ 100) in the control arm of anxiety.

3.3. Anxiety outcomes
The impact of NeuroIDgenetix® testing on guiding therapy for
patients diagnosed with anxiety was also evaluated in this study.
Patients included in this analysis were classified as having moder-
ate to severe anxiety based on HAM-A scores of 18. Anxiety or
both depression/anxiety subjects with medication therapy guided
by the NeuroIDgenetix® test showed a reduction in anxiety
symptoms compared to the control group (Table 2).
The percent reduction in HAM-A scores was even more signif-
icant when patients with anxiety only were analyzed. The percent
reduction in HAM-A scores at the 4, 8 and 12 week follow-up visits
within this cohort were 39%, 47% and 54% for the experimental
group versus 26%, 35% and 42% for the control group (p ¼ 0.004,
0.02 and 0.02) (Table 2). Furthermore, at the 12 week follow-up
visit, the response rate was significantly higher in the experi-
mental group (63%, 95% CI [54%, 72%]) compared to the control
group (50%, 95% CI [40%, 59%], p ¼ 0.04, OR: 1.76 (1.03e2.99),
NNT ¼ 7.3) for patients diagnosed with anxiety only (Fig. 3).
3.4. Medication changes
similar for subjects between groups (p ¼ 0.70). Medication changes
in the experimental group were aligned with the report recom-
mendations 70% of the time, as opposed to random selection in the
control group, which aligned with report recommendations
(withheld from the treating physician) only 29% of the time. Ulti-
mately, 83% of the subjects in the experimental group had at least
one medication concordant with the report. Overall treatment
changes from baseline were comparable between the experimental
and control groups, supporting our conclusion that report recom-
mendations were highly informative as the experimental group
experienced significantly better efficacy over the control.
3.5. Adverse drug events
We did not find any statistical difference in ADEs between the
control and the experimental group (p-value ¼ 0.21). Ninety-four
percent of total ADEs reported by both groups were not severe,
and included ADEs such as drowsiness, nausea and dry mouth. Of
the total number of ADEs, 6% were severe and were equally
distributed across the control and experimental groups. The fre-
quency of severe ADEs was too low for statistical comparison.

Physicians implemented or modified therapy two weeks after 4. Discussion

patient enrollment (±1 week). All depression and anxiety medi-
cation changes were recorded along with a list of prescription
medication use. Medication changes were categorized as one of the
following: a new medication added, dose adjustment, or current
medication discontinued. The number of medication changes
occurring in the experimental group was significantly higher than
the control group during the 2e12 week follow-up visits
(p < 0.0001, Table 3). Most medication changes occurred at the 2
week follow up visit; 81% of the subjects in the experimental group
versus 64% of subjects in the control group along with additional
changes throughout the study. Within the experimental group,
there were more subjects with new medications added and med-
ications discontinued, as compared to the control group (p < 0.0001
and p < 0.0001, respectively), however dose adjustments were
Diagnoses of anxiety and depression have been increasing
rapidly in the US over the last several decades, with the lifetime
prevalence of a major depressive disorder near 17% and anxiety
disorders affecting 29% of the population (Kessler et al., 2005). The
U.S. Preventive Services Task Force recommends screening for
depression in the general adult population and includes effective
treatment as part of the implementation plan (Siu and US
Preventative Services Task Force, 2016). Consequently, the re-
sponsibility of treating neuropsychiatric illness is no longer falling
on psychiatrists alone but a wide spectrum of health care providers.
This paradigm shift is already well under way as psychiatrists are
only responsible for ~20% of the antidepressants prescribed today,
with the remaining ~80% of prescriptions written by primary-care
 P. Bradley et al. / Journal of Psychiatric Research 96 (2018) 100e107 105

Table 2
a Patients Achieving Remission Percent Change in HAM-A scores for Patients with Anxiety.
Severe Depression
40%
Patients with Depression (%)

35% Diagnosis Experimental Group Control Group

P
N % Change (SD) N % Change (SD)
30%
25%
Patients with Anxiety or Both Depression/Anxiety
4 week visit 177 34% (32) 175 27% (31) 0.05
20%
8 week visit 165 45% (33) 169 37% (33) 0.02
13%
12 week visit 159 51% (33) 162 44% (33) 0.06
9%
10% Patients with Anxiety Only
4 week visit 95 39% (32) 103 26% (31) 0.004
8 week visit 89 47% (32) 99 35% (35) 0.02
0% 12 week visit 82 54% (35) 95 42% (34) 0.02
8 weeks 12 weeks
Control NeuroIDgenetix Percent change in HAM-A scores is represented as mean change (±SD %). P values
were calculated using two sided t-test.

Patients Achieving Response Fig. 3. Response Rates for Patients with Anxiety. Response rates (50% reduction in
c Moderate/Severe Depression
HAM-A scores) in the experimental group (red bars, n ¼ 109) compared to the control
70% group (blue bars, n ¼ 115) for moderate and severe anxiety only patients. The response
Patients with Depression (%)

64%
60% rate by 12 weeks was higher in the experimental group (p ¼ 0.04). P value was
calculated using two-sided Fisher's exact test.
49% 46%
50%
41%
40%
30%
Table 3
20% Proportion of patients experiencing at least one medication change.
10% Medication Change Control Group Experimental P
0% (n ¼ 333) Group (n ¼ 352)
8 weeks 12 weeks
N % Changes N % Changes
Control NeuroIDgenetix Any Medication Change
2 weeks 212 64% 284 81%
4 weeks 105 32% 131 37%
Fig. 2. Remission and Response Rates for Patients with Depression. (a) Remission rates
8 weeks 75 23% 91 26%
(HAM-D17 scores 7) and (b) response rates (50% reduction in HAM-D17 scores) for
12 weeks 30 9% 46 13% <0.0001
the experimental group (red bars, n ¼ 40) compared to the control group (blue bars,
New Medication
n ¼ 53) for patients with severe depression. Percentage of patients achieving remission
2 weeks 151 45% 240 68%
and response at 12 weeks was higher in the experimental group (p ¼ 0.02, p ¼ 0.001
4 weeks 50 15% 62 18%
respectively). (c) Response rates for patients with moderate and severe depression
8 weeks 30 9% 31 9%
(HAM-D17 scores 20) were higher (p ¼ 0.01) in the experimental group (red bars,
12 weeks 4 1% 13 4% <0.0001
n ¼ 140) versus control (blue bars, n ¼ 121). P values were calculated using two-sided
Medication Discontinued
Fisher's exact test.
2 weeks 27 8% 91 26%
4 weeks 33 10% 44 13%
8 weeks 23 7% 28 8%
12 weeks 7 2% 22 6% <0.0001
physicians (Mark et al., 2009). Considering the majority of anxiety Dose Adjustment
and depression patients are being treated by primary care physi- 2 weeks 79 24% 70 20%
cians, pharmacogenetic tests which demonstrate utility in psychi- 4 weeks 56 17% 73 21%
8 weeks 46 14% 62 18%
atric and non-psychiatric practices are likely to enhance the overall 12 weeks 22 7% 25 7% 0.70
quality of psychiatric treatment. While previous studies have found
P values were calculated using two-sided Fisher's exact test.
clinical benefits of PGx-guided medication management in patients
with depression, the results are limited in their applicability to real-
world settings; excluding patients with comorbid psychiatric dis- pharmacogenetic testing in a manner that most accurately reflects
orders and enrolling patients only seen at psychiatric clinics. The clinical practice today. In order to achieve this, our study enrolled
aim of this study was to demonstrate the clinical utility of patients from 20 independent clinical sites throughout the US, with
106 P. Bradley et al. / Journal of Psychiatric Research 96 (2018) 100e107
treating physicians specializing in Psychiatry, Internal Medicine,
Obstetrics and Gynecology, and Family Medicine. Patients enrolled
included those new to therapy or those currently failing treatment.
Finally, our study population included patients with a broad range
of severities for depression or anxiety, as well as comorbid patients
displaying both depression and anxiety.
We observed statistically significant improvements for depres-
sion patients in the experimental group, in both the time to
remission and response, as measured by HAM-D17. These results
are well aligned with the growing body of scientific evidence
demonstrating that pharmacogenetic information can improve
antidepressant efficacy (Hall-Flavin et al., 2012, 2013; Winner et al.,
2013). In order to compare our findings with previously published
studies, we focused a post-hoc analysis on the ‘Inadequately
Controlled’ cohort, as published literature has focused exclusively
on the effects of pharmacogenetic-guided treatment in patients
with treatment-resistant depression. An improvement in remission
(42% vs. 27%, p-value ¼ 0.03) rates was evident in the experimental
group compared to control. Similarly, the experimental group also
showed greater improvement (62% vs. 44%, p-value ¼ 0.01) in
response rates than the control arm. We believe this post-hoc
analysis demonstrates strong alignment with what has already
been reported in the current literature. Furthermore, the design of
the study expands the utility of pharmacogenetic testing to include
‘New to Treatment’ patients as well as administration of the test in
both psychiatric and non-psychiatric settings.
In addition, we believe this to be the first report of using phar-
macogenetics to improve outcomes for patients diagnosed with
anxiety. While statistical significance was found in the percent
reduction of HAM-A scores of the experimental group comprised of
‘anxiety or both depression/anxiety’ subjects, ‘anxiety only’ sub-
jects appeared to benefit even more from pharmacogenetic-
guidance. While additional work may be required to explain this
difference, this outcome may reflect a unique set of pharmaco-
therapy requirements for treating anxiety in comorbid patients or it
may represent a clinical prioritization of managing depression
symptoms ahead of anxiety.
The blind nature of this clinical trial refers to blinding of both
subjects and rating clinicians providing efficacy assessments.
Specimen collection occurred for all subjects regardless of study
arm to maintain blinding, however, patients were not asked to
guess their treatment assignment at the end of the study. While the
treating physicians could not be blinded to patient arm assignment,
they were blind to the HAM assessment scores, thus ensuring that
any expectancy bias did not affect the clinical outcomes measured.
Similarly, clinicians conducting the HAM assessments were blind to
the therapeutic decisions made by the prescribing physician and
blind to patient treatment arm assignment. Additionally, all
participating site staff (including physicians, study coordinators
and HAM-raters) were thoroughly trained on the protocol and
followed procedures designed to maintain blinding.
A critical attribute of any diagnostic test, and perhaps the most
significant demonstration of clinical utility, is the extent to which a
test is implemented by a health care provider. In this study, phy-
sicians made at least one medication change in 81% of subjects in
the experimental group as compared to 64% of subjects in the
control group (p < 0.0001) at the two week time-point. These re-
sults strongly suggest that the NeuroIDgenetix® test results pro-
vided physicians with clear concise information to justify
additional, meaningful medication changes. While these additional,
pharmacogenetically-guided changes are the likely cause of
improved efficacy in treatment, no significant difference could be
found between the experimental and control groups in regards to
ADEs. While these results were initially surprising (as the avoid-
ance of ADEs is an integral part of the messaging in the
NeuroIDgenetix® report) we believe this represents a clinical
“trade-off” between implementing a new treatment with an
increased probability of effectiveness and the patient's awareness
of new side effects associated with titrating medications. Given that
medication changes were higher in the experimental group, and
ADE occurrence rates were statistically indistinguishable between
the experimental and control groups, we believe the ADE
messaging in the IDgenetix® algorithm effectively provides a crit-
ical balance in improving efficacy, without a subsequent increase in
the incidence or severity of ADEs.
This study demonstrates the clinical validity and utility of
pharmacogenetic-guided treatment for depression and anxiety.
While medication selection based on pharmacogenetic testing
continues to become closer to routine clinical practice, test imple-
mentation will result from pharmacogenetic algorithms that pro-
vide clinically validated information in an easily actionable format.
In alignment with these requirements, this study demonstrates
that the NeuroIDgenetix® test improves outcomes for patients with
both depression and anxiety, in a diverse set of relevant clinical
settings.
Conflict of interest
P. Bradley, M. Shiekh, V. Mehra, K. Vrbicky and S. Layle have no
conflict of interest to disclose.
M. Olson, A. Maciel, A. Cullors, J. Garces, and A. Lukowiak are
employed by AltheaDx.
Acknowledgments
We thank the other investigators, clinical site personnel and the
participants of this study for their valuable contributions. We thank
Dr. Richard Weisler, Dr. Mustafa Hussain, Dr. Robert Detweiler and
Dr. Jeffrey Ross for their input on study design. We also thank Dr.
Ping-Yu Liu (Liu Associates) for statistical analysis and Innovis, LLC
for their assistance with clinical trial management. Finally, we
thank Joel Centeno for his contributions throughout the entirety of
the project.
This research was funded by AltheaDx. The statistical analyses
were conducted by an independent company, Liu Associates, and
were paid for by AltheaDx.
Appendix A. Supplementary data
Supplementary data related to this article can be found at
https://doi.org/10.1016/j.jpsychires.2017.09.024.
References
Altar, C.A., Hornberger, J., Shewade, A., Cruz, V., Garrison, J., Mrazek, D., 2013.
Clinical validity of cytochrome P450 metabolism and serotonin gene variants in
psychiatric pharmacotherapy. Int. Rev. Psychiatry 25, 509e533.
Anxiety and Depression Association of America, Updated August, 2016. Facts and
Statistics. https://www.adaa.org (Accessed on 8 February, 2017).
Barbui, C., Cipriani, A., Patel, V., Ayuso-Mateos, J.L., van Ommeren, M., 2011. Efficacy
of antidepressants and benzodiazepines in minor depression: systematic re-
view and meta-analysis. Br. J. Psychiatry 198 (1), 11e16.
Bystritsky, A., 2006. Treatment-resistant anxiety disorders. Mol. Psychiatry 11,
805e814.
Caudle, K.E., Klein, T.E., Hoffman, J.M., Muller, D.J., Whirl-Carillo, M., Gong, L., et al.,
2014. Incorporation of pharmacogenomics into routine clinical practice: the
Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline
development process. Curr. Drug Metab. 15, 209e217.
Cusin, C., Yang, H., Yeung, A., Fava, M., 2010. Rating scales for depression. In: Baer, L.,
Blais, M.A. (Eds.), Handbook of Clinical Rating Scales and Assessment in Psy-
chiatry and Mental Health. Humana Press a Part of Springer Science, New York,
NY, pp. 7e35.
Evans, W.E., Johnson, J.A., 2001. Pharmacogenomics: the inherited basis for inter-
individual differences in drug response. Annu. Rev. Genomics Hum. Genet. 2,
9e39.
 P. Bradley et al. / Journal of Psychiatric Research 96 (2018) 100e107
Evans, W.E., Relling, M.V., 1999. Pharmacogenomics: translating functional geno-
mics into rational therapeutics. Science 286, 487e491.
Fava, M., Rush, A.J., Alpert, J.E., Balasubramani, G.K., Wisniewski, S.R., Carmin, C.N.,
et al., 2008. Difference in treatment outcome in outpatients with anxious versus
nonanxious depression: a STAR*D report. Am. J. Psychiatry 165, 342e351.
Gelenberg, A.J., Freeman, M.P., Markowitz, J.C., Rosenbaum, J.F., Thase, M.E.,
Trivedi, M.H., Oct. 2010. Practice Guideline for the Treatment of Patients with
Major Depressive Disorder. American Psychiatric Association. Available at:
https://psychiatryonline.org/pb/assets/raw/sitewide/practice_guidelines/
guidelines/mdd.pdf (Accessed on 10 February 2017).
Goodman, W.K., 2004. Selecting pharmacotherapy for generalized anxiety disorder.
J. Clin. Psychiatry 65 (13 Suppl. l), 8e13.
Greenberg, P.E., Fournier, A.A., Sisitsky, T., Pike, C.T., Kessler, R.C., 2015. The eco-
nomic burden of adults with major depressive disorder in the United States
(2005 and 2010). J. Clin. Psychiatry 76, 155e162.
Hall-Flavin, D.K., Winner, J.G., Allen, J.D., Jordan, J.J., Nesheim, R.S., Snyder, K.A.,
et al., 2012. Using a pharmacogenomic algorithm to guide the treatment of
depression. Transl. Psychiatry 2, e172. https://doi.org/10.1038/tp.2012.99.
Hall-Flavin, D.K., Winner, J.G., Allen, J.D., Carhart, J.M., Proctor, B., Snyder, K.A., et al.,
2013. Utility of integrated pharmacogenomic testing to support the treatment
of major depressive disorder in a psychiatric outpatient setting. Pharmacogenet
Genomics 23, 535e548.
Hamilton, M., 1959. The assessment of anxiety states by rating. Br. J. Med. Psychol.
32, 50e55.
Hamilton, M., 1960. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry
23, 56e62.
Helfand, M., Peterson, K., Dieperink, E., Ferguson, L., Anderson, J., May 2016. Evi-
dence Brief: the Comparative Effectiveness, Harms and Cost-effectiveness of
Pharmacogenomics-guided Antidepressant Treatment versus Usual Care for
Major Depressive Disorder. Evidence-based Synthesis Program Coordinating
Center, Portland CA Health Care System.
Hicks, J.K., Bishop, J.R., Sangkuhi, K., Muller, D.J., Ji, Y., Leckband, S.G., et al., 2015.
Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline for
CYP2D6 and CYP2C19 genotypes and dosing of selective serotonin reuptake
inhibitors. Clin. Pharmacol. Ther. 98, 127e134.
Huynh, N.N., McIntyre, R.S., 2008. What are the implications of the STAR*D Trial for
primary care? A review and synthesis. Prim. Care Companion J. Clin. Psychiatry
10, 91e96.
Kessler, R.C., Berglund, P., Demler, O., et al., 2005. Lifetime prevalence and age-of-
onset distributions of DSM-IV disorders in the National Comorbidity Survey
Replication. Arch. Gen. Psychiatry 62, 593e602.
Kessler, R.C., Birnbaum, H., Shahly, V., Bromet, E., Hwang, I., McLaughlin, K.A., 2010.
Age differences in the prevalence and comorbidity of DSM-IV major depressive
episodes: results from the WHO World Mental Health Survey Initiative. April
Depress Anxiety 27 (4), 351e364. https://doi.org/10.1002/da.20634.
Lesch, K.-P., Bengel, D., Heils, A., Sabol, S.Z., Greenberg, B.D., Petri, S., et al., 1996.
Association of anxiety-related traits with a polymorphism in the serotonin
transporter gene regulatory region. Science 274, 1527e1531.
Mark, T.L., Levit, K.R., Buck, J.A., 2009. Psychotropic drug prescriptions by medical
specialty. Psychiatr. ServSept 60 (9), 1667. https://doi.org/10.1176/ps_
2009.60.9.1167.
107
Mrazek, D.A., 2010. Psychiatric pharmacogenomic testing in clinical practice. Di-
alogues Clin. Neurosci. 12, 69e76.
Porcelli, S., Drago, A., Fabbri, C., Gibiino, S., Calati, R., Serretti, A., 2011. Pharmaco-
genetics of antidepressant response. J. Psychiatry Neurosci. 36, 87e113. https://
doi.org/10.1503/jpn.100059.
Pratt, L.A., Brody, D.J., 2014 Dec. Depression in the U.S. household population, 2009-
2012. NCHS Data Brief. 172, 1e8.
Pratt, V.M., Zehnbauer, B., Wilson, J.A., Baak, R., Babic, N., Bettinotti, M., et al., 2010.
Characterization of 107 genomic DNA reference materials for CYP2D6, CYP2C19,
CYP2C9, VKORC1, and UGT1A1: a GeT-RM and Association for Molecular Pa-
thology collaborative project. J. Mol. Diagn 12, 835e846.
Rush, A.J., Trivedi, M.H., Wisniewski, S.R., Nierenberg, A.A., Stewart, J.W., Warden, D.,
et al., 2006. Acute and longer-term outcomes in depressed outpatients
requiring one or several treatment steps: a STAR*D report. Nov Am. J. Psychiatry
163, 1905e1917.
Sheehan, D.V., Lecrubier, Y., Sheehan, K.H., Amorim, P., Janavs, J., Weiller, E., et al.,
1998. The Mini-International Neuropsychiatric Interview (M.I.N.I.): the devel-
opment and validation of a structured diagnostic psychiatric interview for
DSM-IV and ICD-10. J. Clin. Psychiatry 59 (20 Suppl. l), 22e33.
Singh, A.B., 2015. Improved antidepressant remission in major depression via a
pharmacokinetic pathway polygene pharmacogenetic report. Clin. Psycho-
pharmacol. Neurosci. 13 (2), 150e156 [PMC free article] [PubMed] [Reference
list].
Siu, A.L., U.S. Preventative services Task Force, 2016. Screening for depression in
adults US preventative services Task Force recommendation statement. JAMA
315, 380e387.
Trivedi, M.H., Rush, A.J., Wisniewski, S.R., Nierenberg, A.A., Warden, D., Ritz, L., et al.,
2006. Evaluation of outcomes with citalopram for depression using
measurement-based care in STAR*D implications for clinical practice. Am. J.
Psychiatry 163, 28e40.
U.S Department of Health and Human Services, Centers for Disease Control and
Prevention, National Center for Health Statistics. Health, United States, 2013.
With Special Feature on Prescription Drugs. Available at: https://www.cdc.gov/
nchs/data/hus/hus13.pdf (Accessed on 10, February 2017).
U.S Food and Drug Administration. Table of pharmacogenomics biomarkers in drug
labeling. www.fda.gov. Updated 7 November, 2016. (Accessed on 10 February,
2017).
U.S. Food and Drug Administration, CDER, CBER, CDRH, January 2013. Guidance for
Industry, Clinical Pharmacogenomics: Premarket Evaluation in Early-phase
Clinical Studies and Recommendations for Labeling.
Whirl-Carrillo, M., McDonagh, E.M., Hebert, J.M., Gong, L., Sangkuhl, K., Thorn, C.F.,
Altman, R.B., Klein, T.E., 2012. Pharmacogenomics knowledge for personalized
medicine. Clin. Pharmacol. Ther. 92 (4), 414e417. https://www.pharmgkb.org/
page/pgxGeneRef.
Winner, J.G., Carhart, J.M., Altar, C.A., Allen, J.D., Dechairo, B.M., et al., 2013.
A Prospective, randomized, double-blind study assessing the clinical impact of
integrated pharmacogenomic testing for major depressive disorder. Discov.
Med. 16, 219e227.
Wittchen, H.U., Kessler, R.C., Beesdo, K., Krause, P., Hofler, M., Hoyer, J., 2002.
Generalized anxiety and depression in primary care: prevalence, recognition,
and management. J. Clin. Psychiatry 63, 24e34.