Curr Psychiatry Rep (2016) 18:23
DOI 10.1007/s11920-016-0668-3
ANXIETY DISORDERS (A PELISSOLO, SECTION EDITOR)
Long-Term Pharmacological Treatments of Anxiety Disorders:
An Updated Systematic Review
Giampaolo Perna 1,2,3 & Alessandra Alciati 1 & Alice Riva 1 & Wilma Micieli 1 &
Daniela Caldirola 1
# Springer Science+Business Media New York 2016
Abstract Many aspects of long-term pharmacological treat-
ments for anxiety disorders (AnxDs) are still debated. We
undertook an updated systematic review of long-term pharma-
cological studies on panic disorder (PD), generalized anxiety
disorder (GAD), and social anxiety disorder (SAD). Relevant
studies dating from January 1, 2012 to August 31, 2015 were
identified using the PubMed database and a review of bibli-
ographies. Of 372 records identified in the search, five studies
on PD and 15 on GAD were included in the review. No studies
on SAD were found. Our review confirms the usefulness of
long-term pharmacological treatments for PD and GAD and
suggests that they can provide further improvement over that
obtained during short-term therapy. Paroxetine, escitalopram,
and clonazepam can be effective for long-term treatment of
PD. However, further studies are needed to draw conclusions
about the long-term benzodiazepine use in PD, particularly for
the possible cognitive side-effects over time. Pregabalin and
quetiapine can be effective for long-term treatment of GAD,
while preliminary suggestions emerged for agomelatine and
vortioxetine. We did not find any evidence for determining the
optimal length and/or dosage of medications to minimize the
This article is part of the Topical Collection on Anxiety Disorders
* Giampaolo Perna
pernagp@gmail.com
1
Department of Clinical Neurosciences, Hermanas Hospitalarias, Villa
San Benedetto Menni Hospital, FoRiPsi, via Roma 16, 22032,
Albese con Cassano, Como, Italy
2
Department of Psychiatry and Neuropsychology, Faculty of Health,
Medicine and Life Sciences, University of Maastricht,
Maastricht, Netherlands
3
Department of Psychiatry and Behavioral Sciences, Leonard Miller
School of Medicine, University of Miami, Miami, FL, USA
relapse risk. Few investigations have attempted to identify poten-
tial predictors of long-term treatment response. Personalized
treatments for AnxDs can be implemented using predictive tools
to explore those factors affecting treatment response/tolerability
heterogeneity, including neurobiological functions/clinical pro-
files, comorbidity, biomarkers, and genetic features, and to tailor
medications according to each patient’s unique features.
Keywords Anxiety . Longterm . Pharmacological treatment .
Panic disorder . Generalized anxiety disorder . Social anxiety
disorder
Introduction
Anxiety disorders (AnxDs) are the most frequent psychiatric
illnesses, with higher prevalence among women [1]. The most
common type of AnxD is specific phobia, with a lifetime
prevalence of 15.6 %, although individuals suffering from this
disorder rarely seek treatment. The second most common
AnxD type is social anxiety disorder (SAD), with a lifetime
prevalence of 10.7 %, followed by generalized anxiety disor-
der (GAD), with a lifetime prevalence of 4.3 %, and panic
disorder (PD) with/without agoraphobia, with a lifetime prev-
alence of 3.8 % [1]. SAD is characterized by marked fear or
anxiety, with avoidant behaviors in multiple social situations.
PD is characterized by recurrent unexpected panic attacks
(PAs) characterized by sudden intense fear/discomfort epi-
sodes, with a surge of somatic symptoms such as chest pain,
palpitations, dyspnea, and breathlessness. Patients may also
exhibit anticipatory anxiety and maladaptive changes in be-
havior related to PAs. Most subjects with PD fear or avoid
multiple situations in which PAs can occur (i.e., agoraphobia).
GAD is characterized by excessive anxiety and apprehensive
expectation about numerous events or activities [2].
23 Page 2 of 16
Survival analyses showed an overall chronic course of
AnxDs, with a relatively low recovery rate and high recur-
rence probability [3]. AnxDs negatively affect the quality of
life and daily-life functioning, with social, personal, and eco-
nomic consequences at least comparable to those of major
depression [4–6]. Comorbid psychiatric conditions are fre-
quent in subjects with AnxDs and are associated with higher
functional impairment, lower likelihood of recovery from
AnxDs, and increased likelihood of AnxD recurrence [3, 7].
Overall, these factors highlight the importance of considering
long-term treatments for AnxDs. Several pharmacological op-
tions have become available, such as tricyclic antidepressants
(TCAs), benzodiazepines (BDZs), selective serotonin reup-
take inhibitors (SSRIs), serotonin–noradrenaline reuptake in-
hibitors (SNRIs), and pregabalin. Although the efficacies and
limitations of short-term treatments for AnxDs have been
established, many aspects of long-term treatment are still de-
bated, including the following: (1) the optimal duration of
pharmacotherapy to minimize the risk for relapse after discon-
tinuation and the appropriate tapering regimen; (2) whether
long-term pharmacological treatments with lower doses are
sufficient to maintain short-term improvements; (3) the extent
of further improvements during maintenance treatment; (4)
whether, among the available pharmacological options, some
compounds display higher efficacy for long-term treatment;
(5) to what extent long-term treatments improve the functional
disability and treat the psychiatric comorbidities; (6) the bur-
den of side-effects during long-term treatments; and (7) po-
tential predictors of long-term outcome. Considering these
open issues, we undertook an updated systematic review of
long-term pharmacological studies on PD, GAD, and SAD
published over the past 3 years.
Methods
A PubMed database search from January 1, 2012 to August
31, 2015 was performed using the following search terms:
(((Bgeneralized anxiety disorder^ OR Bgeneralised anxiety
disorder^) OR Bpanic^ OR (Bsocial anxiety^ OR social
phobi*) AND (drug* OR psychopharmacolog* OR
pharmacolog* OR Bantidepressant^ OR SSRI* OR
(Bselective^ AND Bserotonin^) OR benzodiaz* OR SNRI*
OR (Bselective^ AND Bnorepinephrine^) OR TCA* OR
Btricyclic^ OR Batypical^ OR Bstabilizer^ OR Bepileptic^
OR iMAO* OR Bmirtazapine^ OR Btrazodone^ OR
Bpregabalin^ OR BHypnotics^ OR Bbupropion^ OR drug
OR psychopharmacology OR pharmacology OR antidepres-
sant OR SSRI OR (selective serotonin reuptake inhibitor) OR
benzodiazepine OR SNRI OR (selective norepinephrine reup-
take inhibitor) OR TCA OR (atypical antipsychotic) OR
(atypical antipsychotics) OR (mood stabilizer) OR (mood sta-
bilizers) OR anti-epileptic OR iMAO OR mirtazapine OR
Curr Psychiatry Rep (2016) 18:23
trazodone OR pregabalin OR hypnotics OR bupropion)
AND (Blong^ OR Byears^ OR Bmonths^ OR relapse* OR
Brecurrence^ OR Bmaintenance^ OR Bcontinuation^ OR long
OR years OR months OR relapse OR recurrence OR mainte-
nance therapy OR continuation therapy))) AND (B2012/01/
01^[Date - Publication]: B2015/08/31^[Date - Publication]).
We also used the reference lists of relevant studies and pertinent
review articles to retrieve additional literature. Of 372 records
identified in the search, five studies on PD and 15 on GAD
were included in the review. No studies on SAD were found.
Studies were included into the review if they included sub-
jects with primary PD diagnosis with or without agoraphobia,
GAD, or SAD according to DSM-III/DSM-III-R criteria [8,
9], DSM-IV/DSM-IV-TR [10, 11], ICD-9/ICD-9-CM [12,
13], or ICD-10 [14] criteria; participants were >18 years of
age; subjects underwent pharmacological treatments lasting
>12 weeks [15]; they provided separate results on subgroups
with PD, GAD, or SAD if multiple AnxDs were studied; they
were peer-reviewed; and they were written in the English lan-
guage. A summary of the selection process is shown in Fig. 1.
Results
The 20 studies reviewed are summarized in Tables 1 and 2.
Panic Disorder
Randomized Studies
Three randomized studies were included (Table 1).
An extension of a previous short-term (8 weeks) random-
ized open study [16] compared the long-term (34 months)
efficacy and safety of clonazepam and paroxetine [17]. In
the short-term, the two compounds showed similar efficacy
in reducing PA number and the global level of anxiety symp-
toms according to the Hamilton Anxiety (HAMA) and
Clinical Global Impression-Severity (CGI-S) scale scores, al-
though clonazepam had a significantly faster onset of action
and fewer adverse events (AEs) than paroxetine. Patients en-
rolled in the long-term extension continued monotherapy with
the same drug and dose, except for a subgroup of partial/non-
responders in the short-term phase who received combination
therapy with clonazepam and paroxetine. Both groups receiv-
ing long-term clonazepam or paroxetine maintained the sig-
nificant decrease in PA frequency obtained during the short-
term phase and showed further reductions in HAMA and
CGI-S scale scores. The clonazepam group reported signifi-
cantly fewer AEs than the paroxetine group. The most com-
mon AEs in the clonazepam group were drowsiness/fatigue,
memory/concentration problems, and sexual dysfunction, but
the rates were significantly lower than in the paroxetine group.
In the paroxetine group, several other AEs were significantly
Curr Psychiatry Rep (2016) 18:23
Fig. 1 Study selection
higher than in the clonazepam group, including appetite
change/weight gain, diarrhea/constipation, and dry mouth.
A study [18] compared the improvement rates in PA fre-
quency during 1 year for three treatment modalities: cognitive
behavioral therapy (CBT), SSRIs, or CBT + SSRI. In the
CBT + SSRI group, the two treatments started concurrently
and were delivered in parallel. All three groups showed sig-
nificantly reduced PA frequencies, resulting in virtually no
PAs by the end of the 1-year treatment. However, CBT was
associated with a significantly slower decline in PA frequency
during treatment and with relatively more PAs throughout the
entire year than SSRIs and CBT + SSRIs. Although PA
frequency did not differ between SSRI and CBT + SSRI
groups, improvement was significantly faster with CBT +
SSRIs than with SSRIs or CBT among patients with mod-
erate or severe AG.
A secondary analysis of a previous small, randomized con-
trolled study of paroxetine and CBT efficacy for late-life PD
with agoraphobia (PDA) [19] evaluated the predictive values
of age, age of onset, and illness duration on 14-week treatment
outcome [20]. Those patients with late-onset PDA or shorter
illness duration had more favorable outcomes with CBT than
paroxetine on the severity of catastrophic agoraphobic cogni-
tions and phobic avoidance as assessed by self-reported
questionnaires.
Non-Randomized Studies
Two non-randomized studies were included (Table 2).
A prospective, naturalistic, open-label, multicenter study
investigated the efficacy of a 24-week treatment with
escitalopram on clinical severity and functional impairment
in PD [21]. The outcome measures included panic symptom
measures and functional disability at work, in social relation-
ships, and in responsibilities at home/family. At the end of the
study, the remission rate was 73.1 %. The response rates, the
panic symptom, and functional disability measures showed a
continuous significant improvement throughout the treatment.
In the second study, approximately half of the patients
achieved remission during a 12-month naturalistic
Page 3 of 16 23
pharmacological study designed to assess whether Panic-
Agoraphobic-Spectrum (PAS) dimensions can predict treat-
ment response in PD [22]. The study found a significant as-
sociation between the Bseparation anxiety^ dimension and a
lower likelihood of remission. The sample included mixed
pharmacological regimens, such as monotherapy, combined
therapy, or switching from one drug to another. Therefore, it
did not provide conclusions about specific medications or
pharmacological approaches for the long-term PD treatment.
However, it underscored the clinical role of adult separation
anxiety in treatment outcome.
Generalized Anxiety Disorder
Randomized Studies
Seven randomized studies were included (Table 1).
Antidepressants
Agomelatine
A multicenter trial evaluated the long-term efficacy of
agomelatine, a selective melatonergic MT1/MT2 receptor ag-
onist and 5HT2c antagonist, for preventing relapse in GAD
patients [23]. Responders to the open-label phase were ran-
domized to receive agomelatine at the same dose or a placebo
for 26 weeks. Subsequently, agomelatine-treated patients were
re-randomized to receive either placebo or agomelatine (same
dose) for 1 week to assess discontinuation symptoms. At the
end of the 26 weeks, patients receiving agomelatine exhibited
a significantly lower relapse rate than those receiving placebo.
There were no discontinuation symptoms in patients switched
to placebo compared with those maintained on agomelatine.
Escitalopram
A controlled study examined whether sequential treatment
with escitalopram and CBT can prevent relapse in older pa-
tients (≥60 years) with GAD [24]. Escitalopram plus CBT
 23

Table 1 Randomized trials in panic disorder and generalized anxiety disorder

Authors Study design Duration Subjects Number Treatments (number Daily doses (mg) Main outcome measures Other outcome measures or
[reference] of subjects) other results

Panic disorder
Page 4 of 16

Nardi et al., 2012 Randomized, open 34 weeks, following Outpatients with 105 Clonazepam (n = 47), Mean doses (SD): Assessment of PAs: patients’ CGI-S, HAMA: similar
[17] design; naturalistic an acute 8-week PD with or paroxetine (n = 37), clonazepam = 1.9 diaries + clinical improvement in
setting; treatment phase without AG combination (n = 21) (0.2), interviews. Patients free clonazepam/paroxetine
monocenter paroxetine = 37.7 from PAs: group. CGI-I in clonazepam
(3.8) clonazepam = 87 %, group, 1.06 (0.16), > CGI-I
paroxetine = 85 %, in paroxetine group, 1.11
combination = 61 % (0.14), (p = 0.04). AEs in
clonazepam group
(28.9 %) < AEs in paroxe-
tine group (70.6 %)
(p < 0.001)
Van Apeldoorn Randomized, open 52 weeks (including a Outpatients with 150 SSRIs (fluoxetine or Not specified. Assessment of PAs: patients In patients with moderate/
et al., 2013 design; medication taper PD with or paroxetine or Clinicians were free colored a box in a panic severe AG: CBT + SSRIs
[18] multicenter period: weeks without AG sertraline or to adjust dosages plot each time a PA provided faster improve-
40–52) citalopram or according to occurred (event-contingent ment than SSRIs or CBT
fluvoxamine) (n = 53; clinical response recording). In all the 3 alone. Gains were main-
completers, n = 24), and tolerability. groups: significant tained over the medication
CBT (n = 55; improvement on frequency taper period.
completers, n = 32), of PAs.
CBT + SSRI (n = 52, SSRIs = CBT + SSRIs;
completers, n = 27) SSRIs/
CBT + SSRIs > CBT. The
analyses were performed
on the completers.
Hendriks et al., Randomized, open 14 weeks Outpatients, 49 Paroxetine (n = 17), CBT Week 1: 10; from – In patients with late-onset PDA
2012 [20] design; 60 years and (n = 20), waiting-list week 2 to week 4: or shorter duration of ill-
monocenter older, with control condition daily dose ness, CBT condition had
PDA (n = 12) increased with better outcome than paroxe-
10 mg per week; tine on self-reported ACQ
from week 5: 40 scores and MIA total scores.
Generalized anxiety disorder
Stein, 2012 [23] Randomized, double- 26 (+1) weeks of Outpatients with 227 (responders to Agomelatine (n = 113), Agomelatine = 25–50 Relapse = HAMA score ≥15 Significantly higher rates of
blind, placebo- maintenance GAD period 1) placebo (n = 114) or clinical judgment of a patients presented at least 1
controlled relapse treatment following lack of efficacy. Significant emergent AE in the
prevention study; a previous 16-week lower rates of relapse in the agomelatine group than in
multicenter open-label treat- agomelatine group (17.7 % the placebo group (40.7 and
ment (period 1) in the whole group; 28.8 % 27.2 %, respectively). The
in the severely ill patients) majority of the AEs were
than in the placebo group mild to moderate. 17
(30.7 % in the whole patients (3.6 %) had at least
group; 42.9 % in the 1 emergent PCSA liver
Curr Psychiatry Rep (2016) 18:23

severely ill patients). enzyme value (13 taking

agomelatine, 4 placebo); all
values of patients taking
agomelatine returned to
baseline levels.
Table 1 (continued)

Authors Study design Duration Subjects Number Treatments (number Daily doses (mg) Main outcome measures Other outcome measures or
[reference] of subjects) other results

Wetherell et al., Randomized, 16 weeks Patients with 73 (responders to Responders to the open- Escitalopram = 10–20 Response = HAMA score –
2013 [24] placebo-controlled (augmentation GAD ≥ 60 yea- period 1) label period were ran- ≤10, PSWQ = decrease
study; multicenter phase) + 28 weeks rs old domly assigned to one ≥8.5. Relapse = two
(maintenance of 4 conditions: consecutive assessments of
phase) following a 16 weeks of 1) HAMA ≥ 14 and 2)
previous 12-week escitalopram + CBT meeting DSM-IV criteria
open-label treat- (augmentation phase), for GAD. Augmentation
Curr Psychiatry Rep (2016) 18:23

ment (period 1) followed by 28 weeks phase: response HAMA:

of escitalopram main- escitalopram
tenance or 28 weeks alone = escitalopram + CB-
of placebo; 16 weeks T; response PSWQ:
of escitalopram alone, escitalopram + CBT
followed by 28 weeks significantly > than
of escitalopram main- escitalopram alone.
tenance or 28 weeks Maintenance phase: relapse
of placebo with escitalopram (2.7 %)
significantly < than placebo
(46.1 %). Participants tak-
ing placebo who received
CBT had significantly
lower rates of relapse
(25.0 %) than those who
did not (66.4 %).
Baldwin et al., Randomized, double- 24–56 weeks Outpatients with 687 in the open-label Vortioxetine (Lu Vortioxetine = 5–10 Relapse during the double- No significant difference in the
2012 [26] blind, placebo- following a GAD phase, 459 in the AA21004) (n = 229), (flexible dose in the blind phase = HAMA score incidence of AEs between
controlled relapse previous 20-week double-blind placebo (n = 230) 8 weeks-open-label ≥15 or clinical judgment of the vortioxetine group
prevention study; open-label treat- phase (responders phase, fixed dose in insufficient therapeutic re- (55.5 %) and the placebo
multicenter ment to the open-label the following sponse. Risk of relapse group (53.9 %).
phase) weeks) with placebo: 3 times > -
than with vortioxetine. A
significantly higher rate of
patients with placebo re-
lapsed compared to those
with vortioxetine (34 and
15 %, respectively).
Kasper et al., Randomized, double- 12 week lorazepam- Patients with GAD 615 entering period Period 2: pregabalin low Pregabalin low Following the taper of period Improvements in HAMA
2014 [27] blind, placebo- controlled period 1, 463 responders dose (n = 112), dose = 150–300, 2, DESS events did not scores and in both CGI-I
and lorazepam- (period 1) + 12- entered period 2 pregabalin high dose pregabalin high differ between patients on and CGI-S scores obtained
controlled trial; week lorazepam (n = 121), lorazepam dose = 450–600, active medication and during period 1 were main-
multicenter and placebo- (n = 114), placebo lorazepam = 3–4 patients on placebo (range tained in period 2 in patients
controlled period (n = 116) (flexible dose in the 22.3–31.2 % and 13.3– with both active treatment
(period 2), each first 6 weeks, fixed 31.0 % respectively). and placebo.
followed by a dose in the Small increase in PWC
1-week taper and 2 following weeks) scores in both active
post- treatment and placebo
discontinuation as- groups was found. Rates of
sessments rebound anxiety (HAMA
Page 5 of 16 23

scores) were low at both 12

and 24 weeks (0–6 %).
 23

Table 1 (continued)

Authors Study design Duration Subjects Number Treatments (number Daily doses (mg) Main outcome measures Other outcome measures or
[reference] of subjects) other results

Hadley et al., Randomized, double- 12-week randomized, Outpatients with 138 entering the Alprazolam (n = 106, of Alprazolam = 1–4 At the endpoint: rate of In both groups, the HAMD and
2012 [28] blind, placebo double-blind treat- lifetime GAD stabilization which n = 56 (taper: 25 % per patients who remained the PHQ-9 scores showed
Page 6 of 16

controlled; ment with either who had been phase, 106 randomized to week), alprazolam-free did not minimal endpoint changes.
multicenter pregabalin or place- treated with entering the pregabalin and n = 50 pregabalin = 300– differ in pregabalin group The DSST (assessing cog-
bo (after benzodiaze- double-blind to placebo for the 12- 600 (51.4 %) and placebo nitive performance and psy-
2–4 weeks of stabi- pines for 8– phase week double-blind group (37.0 %). Pregabalin chomotor speed) showed a
lization with alpraz- 52 weeks phase) was associated with signif- similar endpoint increase in
olam). icantly greater reduction in both the pregabalin and pla-
Simultaneously, HAMA, PWC, CGI-S, cebo groups. Use of high-
6 weeks of gradual CGI-I, and PGI-I scores. potency benzodiazepine be-
alprazolam taper, fore the beginning of the
followed by study significantly reduced
6 weeks of the ability to achieve an
alprazolam-free alprazolam-free status.
phase
Sheaan et al., Randomized- Open-label run-in Outpatients with 1224 entering the Quetiapine XR (n = 216), Quetiapine XR = 50– Quetiapine XR significantly Quetiapine XR was
2013 [30] withdrawal, phase: 4–8 weeks; GAD open-label phase, placebo (n = 216) 300 reduced the risk of an significantly more effective
double-blind, open-label stabiliza- 432 entering the anxiety event by 81 % than placebo at maintaining
parallel group, tion phase: 12– double-blind compared with placebo. SDS total score, SDS
placebo-controlled 18 weeks; random- phase Bfamily life/home
relapse prevention ized, double-blind responsibilities^ and work-
study; multicenter phase: 52 weeks. related Bdays lost^ domain
scores, and sleep improve-
ment (PSQI score).
Endicott et al., Randomized- Open-label run-in Outpatients with 1248 entering the Quetiapine XR (n = 216), Quetiapine XR = 50– Q-LES-Q-SF percent Quetiapine XR maintained a
2013 [31] withdrawal, phase: 4–8 weeks; GAD open-label phase, placebo (n = 216) 300 maximum total score, item significantly better SDS and
double-blind, open-label stabiliza- 432 entering the 15 (Bsatisfaction with PSQI total scores.
parallel group, tion phase: 12– double-blind medication^), and item 16
placebo-controlled 18 weeks; randomi- phase (Boverall life satisfaction^)
relapse prevention zation phase: significantly improved in
study; multicenter 52 weeks. the quetiapine XR group
and were significantly
better maintained
compared with placebo
group.

ACQ Agoraphobic Cognitions Questionnaire, AE(s) adverse event(s), AG agoraphobia, CBT cognitive behavioral therapy, CGI-I Clinical Global Impression-Improvement scale, CGI-S Clinical Global
Impression-Severity scale, DESS Discontinuation-Emergent Signs and Symptoms, DSM-IV Diagnostic and Statistical Manual of mental disorders-IV Edition, DSST Digit-Symbol Substitution Test, GAD
generalized anxiety disorder, HAMA Hamilton Anxiety rating scale, HAMD Hamilton Depression rating scale, MIA Mobility Inventory Avoidance scale, PA(s) panic attack(s), PCSA potentially clinically
significant abnormal, PD panic disorder, PDA panic disorder with agoraphobia, PGI-I Patient Global Impression of change-Improvement score, PHQ-9 9-item Patient Health Questionnaire, PSQI
Pittsburgh Sleep Quality Index, PSWQ Penn State Worry Questionnaire, PWC Physician Withdrawal Checklist, Q-LES-Q-SF Quality of Life and Satisfaction Questionnaire-short form, SD standard
deviation, SDS Sheehan Disability Scale, SSRIs selective serotonin reuptake inhibitors, XR extended release
Curr Psychiatry Rep (2016) 18:23
Table 2 Non-randomized trials in panic disorder and generalized anxiety disorder

Authors Study design Duration Subjects Number Treatments (number of Daily doses (mg) Main outcome measures Other outcome measures or
[reference] subjects) other results

Panic disorder
Choi et al., 2012 Prospective, 24 weeks Outpatients with PD 119 Escitalopram Initial dose: 5 or 10; Remission = absence of full At the end of the study:
[21] naturalistic, open- with or without it was increased PAs + global score PDSS significant improvement in
label; multicenter AG up to 20 mg per ≤7. Response = > 50 % all the three SDS domains
day, according to reduction of the PDSS. At (disability at work, social
the clinicians’ the end of the study: relationship,
Curr Psychiatry Rep (2016) 18:23

judgment. Mean remission rate = 73.1 %, responsibilities at home/

dose (SD): 11.65 response rate = 80.7 %. family). Continuous
(3.83) An intention-to-treat anal- significant improvement in
ysis was performed. response rates (45.4, 69.7,
80.7 after 4, 12, 24 weeks,
respectively). PDSS and
SDS scores were found
over the entire treatment.
Miniati et al., Prospective, 12 months Outpatients with PD 57 SSRIs (fluoxetine, – Remission = PDSS scores <5. The last month PAS-SR Badult
2012 [22] naturalistic, open- with or without paroxetine, sertraline, At the end of the study: separation anxiety^ factor
label; monocenter. AG, selected citalopram, remission rate = 48.1 %. was associated with lower
At each time point, among individuals fluvoxamine), SNRI likelihood of remission.
PDSS was adminis- referred by (venlafaxine), TCAs
tered by raters not primary care (imipramine,
involved in the treat- physicians trimipramine,
ment clomipramine), BDZ
(delorazepam,
alprazolam)
Generalized anxiety disorder
Rickels et al., Open-label treatment 6 months Outpatients with 268 Venlafaxine XR (n = 268) Venlafaxine = 75– Remission = HAMA score Baseline low Eysenck
2013 [33] trial GAD 225 (flexible ≤7. Among completers neuroticism score
dose). Mean (n = 159), 59.8 % had significantly predicted
doses (SD) = 144 remission after 3 months remission. CGI-I score of 1
(74) in non remit- and 79.9 % after or 2 after 8 weeks of treat-
ters, 158 (61) in 6 months. ment predicted later remis-
remitters (non sion (78 % accuracy) and
significant differ- non-remission (91 % accu-
ences) racy)
Narasimhan et Open-label treatment 6 months Outpatients with 112 (European- Venlafaxine XR (n = 112) Venlafaxine = 75– Response = HAMA Response = CGI-I score = 1 or
al., 2012 [34] trial; monocenter GAD American 225 (flexible reduction of ≥50 %. 2. Remission = CGI-I score
population) dose) Remission = HAMA score = 1. A nominal statistically
≤7. No significant significant association was
association between found between the A-allele
HAMA reduction and the (Met) and global CGI-I re-
COMT functional variant sponse.
rs4680 (Val158Met).
Cooper et al., Open-label treatment 6 months Outpatients with 108 (European- Venlafaxine XR (n = 108) Venlafaxine = 75– Response = HAMA Response = CGI-I score = 1 or
2013 [35] trial; monocenter GAD American 225 (flexible reduction of >50 %. 2. Remission = CGI-I score
Page 7 of 16 23

population) dose) Remission = HAMA score = 1. No significant

<7. No significant association at genotypic or
 23

Table 2 (continued)

Authors Study design Duration Subjects Number Treatments (number of Daily doses (mg) Main outcome measures Other outcome measures or
[reference] subjects) other results

association at genotypic or allelic level between the

Page 8 of 16

allelic level between the OPRM1 A118G

OPRM1 A118G polymorphism and either
polymorphism and either CGI-I response or remis-
HAMA response or sion.
remission.
Cooper et al., Open-label treatment 6 months Outpatients with 109 (European- Venlafaxine (n = 109) Venlafaxine = 75– Response = HAMA Response = CGI-I score =1 or
2013 [38] trial; monocenter GAD American 225 (flexible reduction of ≥50 %. 2. Remission = CGI-I score
population) dose) Remission = HAMA score =1. No significant
≤7. The PACAP gene associations with the
polymorphism rs2856966 PACAP gene
(Asp54Gly) showed a sig- polymorphism rs2856966
nificant association with (Asp54Gly).
remission.
Lohoff et al., Open-label treatment 6 months Outpatients with 112 (European- Venlafaxine XR (n = 112) Venlafaxine = 75– Response = HAMA Response = CGI-I score =1 or
2013 [36••] trial; monocenter GAD American 225 (flexible reduction of ≥50 %. 2. Remission = CGI-I score
population) dose) Remission = HAMA score =1. One or two copies of
≤7. One or two copies of the serotonin receptor 2A
the serotonin receptor 2A (HTR2A) gene rs7997012
(HTR2A) gene rs7997012 SNP G allele were signifi-
SNP G allele were signifi- cantly associated with re-
cantly associated with re- sponse.
sponse. This effect was
increasingly statistically
significant as time
progressed, up to a mean
HAMA difference of 4.8
between the A/A group
and the G/A + G/G group
at 6 months.
Lohoff et al., Open-label treatment 6 months Outpatients with 112 (European- Venlafaxine XR (n = 112) Venlafaxine = 75– Serotonin transporter gene –
2013 [37] trial; monocenter GAD American 225 (flexible (SLC6A4) 5-HTTLPR/
population) dose) rs25531 haplotype (La/La,
La/S or S/S) and the sero-
tonin 2A receptor gene
(HTR2A) SNP rs7997012
(G or A) interact in the
prediction of outcome.
Genotypes La/La + G/G or
La/La + G/A had signifi-
cantly lower HAMA
Curr Psychiatry Rep (2016) 18:23

scores than genotypes La/

S + A/A or S/S + A/A.
Montgomery et Open-label safety trial, 51 weeks Subgroup of patients 330 Pregabalin (n = 330) Pregabalin = 150– Most common AEs: Reduction of illness severity
al., 2013 [39] multicenter with GAD 600 (flexible dizziness (10.3 %), (CGI-S score) at week 27.
dose). Mean dose somnolence (7.0 %), The improvement was
weight gain (4.2 %), sustained through week 51.
Curr Psychiatry Rep (2016) 18:23 Page 9 of 16 23

Methyltransferase, GAD generalized anxiety disorder, HAMA Hamilton Anxiety rating scale, MOS-S Medical Outcomes Study-Sleep scale, OPRM1 μ-opioid receptor gene, PACAP pituitary adenylate

deviation, SDS Sheehan Disability Scale, SNP single nucleotide polymorphism, SNRI serotonin noradrenaline reuptake inhibitors, SSRIs selective serotonin reuptake inhibitors, TCAs tricyclic antidepres-
AE(s) adverse event(s), AG agoraphobia, BDZ benzodiazepines, CGI-I Clinical Global Impression-Improvement scale, CGI-S Clinical Global Impression-Severity scale, COMT Catechol-O-

cyclase-activating peptide, PA(s) panic attack(s), PAS-SR self-report form of the panic-agoraphobic-spectrum clinical interview, PD panic disorder, PDSS Panic Disorder Severity Scale, SD standard
Other outcome measures or

mediated effect of anxiety

significantly increased response scores on the Penn State

remained (direct effect)

positive effect on sleep
Worry Questionnaire, but not HAMA scores, compared with

after discounting the

70 % of the pregabalin
escitalopram alone. Escitalopram maintenance with or with-

improvement.
out CBT was associated with significantly lower relapse rates
other results

than placebo. Patients receiving placebo during the mainte-

nance phase following CBT augmentation had significantly
lower rates of relapse than patients without CBT treatment.
Relapse rate did not differ significantly between patients who
insomnia (4.2 %). The rate

symptoms (HAMA score)

(MOS-S score) than usual

received maintenance escitalopram without previous CBT
Main outcome measures

Pregabalin group reported

of severe AEs was low

and sleep disturbance

reduction in anxiety

(5.3 %) and in those who received previous CBT but no main-

significantly higher

care group. tenance medication (i.e., placebo) (25.0 %).

(3.6 %).

Vortioxetine

Vortioxetine (Lu AA21004) is a novel antidepressant with

multimodal serotonergic activity [25]. A 24–56-week multi-
Treatments (number of Daily doses (mg)

at the endpoint:

center, double-blind, randomized placebo-controlled relapse

prevention study investigated the long-term efficacy of
205.3.

vortioxetine for GAD in responders to a previous 20-week,

open-label vortioxetine treatment [26]. The time to relapse
–

(primary efficacy endpoint) as well as the time to relapse in

switched to) (n = 984),

stable responders (i.e., patients responding to treatment for at

Pregabalin (included in

usual care (n = 562)

least 12 weeks, the secondary efficacy endpoint) were signif-

the treatment or

icantly longer in patients treated with vortioxetine than with

placebo. The incidence of AEs, including sexual dysfunctions,
subjects)

and the withdrawal rate from the study due to AEs did not
differ significantly between vortioxetine and placebo groups.
Moreover, there were no significant symptoms following
abrupt discontinuation of vortioxetine.

Anticonvulsants
Number

1546

Pregabalin
sponse to previous
anxiolytic therapy

A placebo- and lorazepam-controlled, randomized, double-

showed a poor/
insufficient re-
Outpatients with

blind, multicenter trial [27] assessed discontinuation symp-

GAD who

toms following taper from short (12 weeks) or long-term treat-

Subjects

ment (24 weeks) with pregabalin. The study consisted of two

12-week treatment periods (periods 1 and 2), each followed by
a 1-week taper and two post-discontinuation assessments, im-
Duration

Ruiz et al., 2015 Post hoc analysis of data 6 months

mediately following the taper and 1 week post-taper. In both

high- and low-dose pregabalin responders who continued with
treatment to 12 or 24 weeks, there was a low incidence of
from a prospective
non-interventional

discontinuation symptoms during the 2-week discontinuation

trial; multicenter

evaluation at the end of each treatment period.

Study design

The efficacy of pregabalin in facilitating taper-off from

sants, XR extended release

long-term BDZ therapy was evaluated in patients with lifetime

GAD [28]. No difference was found between proportions of
Table 2 (continued)

patients remaining benzodiazepine-free among those receiv-

ing pregabalin compared with those receiving placebo.
[reference]

Compared with placebo, treatment with pregabalin was asso-

Authors

[40]

ciated with greater reduction of anxiety symptoms and lower

withdrawal severity during both the taper and the
23 Page 10 of 16
benzodiazepine-free phase. Pregabalin did not show negative
effects on cognitive performance/psychomotor speed.
Atypical Antipsychotics
Quetiapine Extended Release
A secondary analysis of data from a previous long-term main-
tenance study [29] showed that the risk of an anxiety event
was significantly lower in the quetiapine extended release
(XR) group than in the placebo group and that quetiapine
XR was more effective than placebo for maintaining patients’
overall daily-life functioning, functioning in family life/home
responsibilities, and work-related Bdays lost^ domains. In ad-
dition, the quetiapine XR group showed improvement across
multiple sleep domains (sleep quality, latency, duration, and
daytime dysfunction) [30].
Another secondary analysis of the same previous long-term
maintenance study [29] showed that once-daily quetiapine XR
better maintained improvements in health-related global qual-
ity of life, Boverall life satisfaction,^ and Bsatisfaction with
medication^ than placebo. The study also confirmed that
quetiapine XR maintained significantly better overall sleep
quality and daily-life functioning than placebo [31].
Non-Randomized Studies
Eight non-randomized studies were included (Table 2).
Venlafaxine Extended Release In the following studies, the
GAD remission rate as well as clinical and genetic predictors
of remission were investigated after 6 months of open-label
treatment with venlafaxine XR. These studies were secondary
analyses of the lead-in phase of a previous 18-month-long
relapse study [32]. A study [33] found an increasing remission
rate from 3 to 6 months of treatment, suggesting that a
prolonged treatment period optimizes the outcome. Low base-
line neuroticism was the only significant predictor of remis-
sion at the end of the study. The CGI-I score at week 8 was a
significant early predictor of endpoint remission/non-remis-
sion. Five studies explored the predictive value of genetic
variants of catechol-O-methyltransferase (COMT) gene [34],
μ-opioid receptor gene (OPRM1) [35], and serotonin receptor
2A (HtR2A) gene [36••] and its interaction with serotonin
transporter (SLC6A4) gene [37], pituitary adenylate cyclase-
activating peptide (PACAP), and its receptor (PAC1) genes
[38] (results are summarized in Table 2).
Pregabalin A 12-month, open-label safety trial was conduct-
ed in patients with GAD, SAD, or PD who had previously
completed one of four randomized, double-blind, placebo-
controlled short-term studies of pregabalin [39]. In the GAD
subgroup, long-term treatment with pregabalin was generally
Curr Psychiatry Rep (2016) 18:23
well tolerated. Dizziness was the only treatment-related ad-
verse event (AE) occurring in ≥10 % of subjects. The other
most common AEs included somnolence, weight gain, and
insomnia. Majority of weight gain occurred during the first
27 weeks of treatment but then leveled off. Pregabalin treat-
ment was associated with global clinical severity reduction,
and the improvement was sustained up to the endpoint.
A post hoc analysis [40] of data from a previous 6-month,
longitudinal, non-interventional study [41] explored the ef-
fects of pregabalin on self-reported changes in sleep distur-
bances. Switching to pregabalin or including pregabalin in the
usual care (UC) treatment resulted in greater improvement of
sleep disturbances and anxiety symptoms than UC without
pregabalin. Mediation analysis suggested that most of the
sleep improvement resulted from a direct effect of pregabalin,
independent of its effect on anxiety symptoms.
Social Anxiety Disorder
There are no studies relating to social anxiety disorder.
Discussion
Panic Disorder
Numerous short-term clinical trials have shown that the
SSRIs, the SNRI venlafaxine, the TCAs imipramine and clo-
mipramine, and the BDZs alprazolam and clonazepam are
effective for PD. All these compounds, except the SSRI
fluvoxamine, are approved by the European Medicines
Agency (EMA) or the United States Food and Drug
Administration (FDA) for PD treatment. TCAs are considered
second-choice medications because of their side-effects as
well as BDZs because of risk of dependence and cognitive
impairment [42–44].
Unfortunately, our literature search found few long-term
pharmacological studies on PD conducted over the past
3 years. This is disappointing because PD has a long-lasting
course and pharmacotherapy discontinuation often results in
relapse. Similar to findings in naturalistic long-term studies,
only 18 % of patients showed complete recovery 15 years
after alprazolam/imipramine treatment, 13 % recovered but
were still on medication, 51 % experienced recurrent PAs,
and 18 % still met diagnostic criteria for PD [45]. Relapse or
continued panic-phobic symptoms has been found in 15–50 %
of PD patients treated with TCAs, SSRIs, or venlafaxine with-
in 6–12 months after discontinuation [43, 44, 46]. Current
practice guidelines and meta-analyses suggest that continuing
pharmacotherapy for 6–12 months after the acute response
can lead to further symptom reduction and decreased recur-
rence risk, although the optimum length of treatment follow-
ing response is still uncertain. During the discontinuation
Curr Psychiatry Rep (2016) 18:23
phase, tapering the medication gradually over weeks to
months may reduce the likelihood of discontinuation symp-
toms and reveal early relapse signs [15, 47, 48•, 49]. Recent
recommendations suggest that pharmacotherapy should be
continued even longer than 12 months if drug discontinuation
leads to recurrent symptoms, if the disorder is particularly
severe, or if the patient’s clinical history suggests that extend-
ed treatment is needed [48•]. Our review supports the indica-
tion that long-term treatments may provide further symptom
improvement over that obtained in the short-term but unfortu-
nately does not provide advice concerning the optimal length
of pharmacotherapy to minimize the relapse risk after discon-
tinuation. Long-term treatment with clonazepam or paroxetine
further reduced global severity of the disorder and anxiety
symptoms [17], while a 24-week treatment with escitalopram
yielded a continuous increase in the response rate as well as
improvement in both panic symptoms and functional impair-
ment [21]. Only one reviewed study included a tapering peri-
od (12 weeks) after 40 weeks of SSRI treatment.
Discontinuation was well tolerated, and gains were main-
tained throughout medication taper. This suggests that a cau-
tious and slow discontinuation of SSRIs does not lead to re-
lapse during tapering. However, the PA frequency was not
evaluated beyond the full year of treatment [18]. Thus, this
crucial issue of optimal pharmacotherapy duration remains
open. Previous results are mixed, and much more research is
needed. For example, in patients treated with imipramine, a
prospective study showed that a second year of treatment was
more efficacious than 1 year for preventing relapse within
6 months after discontinuation [50], whereas a subsequent
study did not find a difference in relapse after discontinuation
from 6 months or 12–30 months of treatment [51]. Similarly,
the extension of paroxetine maintenance treatment from 12 to
24 months did not provide additional benefits for decreasing
the relapse risk after discontinuation [46].
Recent recommendations suggest that long-term treatment
with SSRIs or SNRIs should be continued at the same dose
that was successful for short-term treatment [48•]. Our review
does not provide data about the possibility that maintenance
treatment with lower dosage is sufficient to maintain short-
term improvement and/or to decrease the relapse risk; there-
fore, this issue remains unclear. Preliminary results suggested
that continuing imipramine at half the dosage is sufficient to
maintain the improvement and that the daily BDZ dosage can
be reduced while remaining effective [44], but no reliable
conclusions can be drawn to date.
Except for the one study reporting similar efficacies of
long-term paroxetine and clonazepam treatments [17], the
reviewed studies did not compare the long-term efficacy of
different anti-panic compounds. Preliminary data suggest
equivalent efficacies of long-term paroxetine and clomipra-
mine [52] as well as imipramine and alprazolam [53] for main-
taining the improvement obtained with short-term treatment
Page 11 of 16 23
and that all are superior to placebo. Similarly, a naturalistic
open-label 12-month comparison study did not show differ-
ences in panic symptom improvement among patients treated
with paroxetine, citalopram, fluoxetine, or fluvoxamine [54].
However, additional studies are needed to draw conclusions
because preliminary short-term studies have suggested differ-
ential effectiveness among anti-panic compounds. One-month
fluvoxamine treatment had the weakest anti-panic properties
in patients with respiratory PD compared with other SSRIs
and clomipramine [55], whereas 2-month paroxetine treat-
ment showed a trend toward higher efficacy in reducing un-
expected panic attacks than citalopram [56]. These discrepan-
cies may be related to differences in drug pharmacodynamics,
patient features, or both, and underscore the need for addition-
al long-term comparative studies in well distinguished patient
subgroups.
Although the efficacy of CBT for PD treatment is beyond
the scope of this review, a study [18] suggested that long-term
treatment with SSRIs alone or SSRIs combined with CBT
have similar effects that are superior to CBT alone.
Moreover, the results suggest that combined treatment with
SSRIs and CBT is recommended for patients with moderate/
severe AG. These findings are in contrast to other studies
showing that combined therapy was as effective as CBT alone
[57], more effective than antidepressants alone [57, 58], and
that no different treatments were indicated for patients with
PD with or without and AG [57].
Current practice guidelines suggest caution in the long-
term use of BDZs due to potential abuse/misuse and physical
dependence, possibly resulting in withdrawal and rebound
symptoms when therapy is discontinued, as well as trouble-
some side-effects (e.g., sedation, fatigue, memory impairment,
risk of motor vehicle accidents) [15, 47]. Recent data [17]
showed no significant increase of memory/concentration im-
pairment in patients receiving long-term treatment with clo-
nazepam. Conversely, at the end of the study, patients receiv-
ing clonazepam experienced a lower incidence of memory
impairment than those receiving paroxetine. This is in line
with previous studies showing only limited evidence for more
pronounced cognitive impairment in chronically BDZ-
medicated PD patients [59]. Although others have reported
links between long-term BDZ use and cognitive impairment
[60, 61], the reviewed study [17] found no increased tolerance
risk or dose escalation during long-term use of BDZs because
patients maintained stable doses of clonazepam with no recur-
rence of symptoms, consistent with previous studies of clo-
nazepam, alprazolam, and other BDZs [53, 62, 63]. Finally, a
previous study of the same group showed that clonazepam can
be successfully discontinued after 3 years of treatment by
gradual tapering without any major withdrawal symptoms
while maintaining clinical improvement at the 8-month fol-
low-up [63]. Overall, these data suggest that the role of BDZ
in PD treatment need to be carefully reappraised.
23 Page 12 of 16
The issue of the side-effects during long-term treatment
was investigated only by Nardi et al. [17]. This study showed
a greater burden of side-effects with paroxetine than with clo-
nazepam, including sexual dysfunction and weight gain.
Similarly, a previous study indicated that 24-month paroxetine
treatment was associated with weight gain, mainly in the first
year of treatment, and sexual side-effects [46]. Although com-
parative studies are limited, different side-effect profiles were
found among SSRIs in long-term treatment, with lower rates
of sexual dysfunction with fluoxetine or fluvoxamine than
paroxetine or citalopram and a lower weight gain rate with
fluvoxamine than with other SSRIs [54]. Because side-
effects may negatively affect both therapy compliance and
patient quality of life, clinicians should carefully consider
the balance between efficacy and side-effects when choosing
long-term treatment for each patient.
In conclusion, the recent data confirms the usefulness of
long-term pharmacological treatments for PD. They indicate
that long-term therapy with SSRIs or clonazepam provide
further improvement over that obtained during short-term
therapy and that gradual taper of SSRIs does not cause relapse
during tapering. Clonazepam was as effective as paroxetine
but better tolerated and was not associated with tolerance/dose
escalation. However, further studies are needed to draw reli-
able conclusions about the efficacy and safety of long-term
BDZ use in PD, particularly the possible cognitive side-effects
over time. We did not find any evidence for determining the
optimal length and/or dosage of medications to minimize the
relapse risk. Few investigations have attempted to identify
potential predictors of long-term treatment response.
Associations were found between adult separation anxiety
and a lower likelihood of PD remission and between late-
onset PDA/shorter duration of illness and a better efficacy of
CBT than paroxetine on AG in older patients. However, the
latter study provided results only about AG because it did not
include PAs in the outcome measures. Because a previous
study on the same sample but with different aims found that
the panic-free rate increased significantly among patients re-
ceiving paroxetine but not in those receiving CBT [19], pre-
dictors of outcome may differ depending on whether PA or
AG is considered.
No studies assessed the role of psychiatric comorbidity in
long-term treatment outcome.
Generalized Anxiety Disorder
Previous studies and meta-analysis [48•, 49, 64] showed that
long-term treatment with SSRIs/SNRIs (first-line therapy) in
GAD patients significantly reduced relapse rates over 6–
18 months compared with placebo (10–19 % versus 39–
56 %). The reviewed studies support the usefulness of long-
term therapy with escitalopram or venlafaxine XR for relapse
prevention. However, although escitalopram is highly
Curr Psychiatry Rep (2016) 18:23
effective in preventing relapse in older patients, many may
obtain sustained remission by CBT without long-term phar-
macological treatment [24]. Our review suggests that other
antidepressants with different mechanisms of action than
SSRIs/SNRIs, such as agomelatine and vortioxetine or other
compounds with anxiolytic properties such as pregabalin and
quetiapine XR, also have a high efficacy for the long-term
GAD treatment.
The relapse prevention trials with agomelatine or
vortioxetine (neither of which is FDA/EMA approved for
GAD treatment) reported relapse rates for active treatment
similar to those found with SSRIs/SNRIs. Although these re-
cent data support the long-term efficacy of vortioxetine for
GAD and showed a favorable tolerability profile (lack of dis-
continuation symptoms and AE rates similar to placebo) [26],
the conflicting short-term studies and limited clinical experi-
ence recommend its use only as second-line therapy.
Agomelatine was found to be effective in short-term therapy,
and the reviewed study [23] demonstrated its ability to reduce
the relapse risk over 6 months. Despite incidences of most
AEs similar to those observed with placebo, agomelatine can-
not be widely recommended for the long-term GAD treatment
due to concerns regarding its potential to induce liver injury. A
recent review concluded that agomelatine was associated with
higher rates of liver injury than both placebo and SSRIs [65•].
Previous investigations support the efficacy of pregabalin
both for the acute phase and for long-term relapse prevention
of GAD [66]. One reviewed open-label study showed that
pregabalin produces a sustained improvement in illness sever-
ity over 1 year of treatment [39], while another [27] showed
that the improvements obtained during the initial 12 weeks of
treatment were maintained over the course of the second 12-
week treatment period, both in patients remaining on
pregabalin and those switched to placebo. This result appears
to contradict previous findings of the superiority of pregabalin
over placebo. However, a possible explanation is the shorter
duration of GAD in this subject sample than those of previous
investigations, suggesting that patients with a more recent
onset of GAD may maintain gains achieved in the short-
term phase of treatment without the need for continued treat-
ment. Finally, both short-term and long-term treatments with
pregabalin were not associated with clinically meaningful dis-
continuation symptoms or rebound anxiety following a 1-
week taper [27].
Among unlabelled compounds, quetiapine has the most
robust evidence for short- and long-term GAD treatment.
The reviewed studies confirmed the effectiveness of long-
term quetiapine in reducing the occurrence of anxiety events
and improving daily-life function and quality of life [30, 31].
Despite this proven efficacy, some authors do not recommend
quetiapine as first-line therapy for GAD owing to long-term
metabolic side-effects and the associated cardiovascular risk
[67]. Neither between-drug comparisons on relapse
Curr Psychiatry Rep (2016) 18:23
prevention nor studies on optimal doses to minimize the re-
lapse risk have been recently conducted. However, in a
reviewed study [27], the response maintenance at 24 weeks
was similar for low-dose pregabalin, high-dose pregabalin,
and lorazepam.
Recent reviewed studies provided indications on suit-
able compounds to counteract specific GAD symptoms.
Treatment with pregabalin was associated with significant
improvements in GAD-related sleep disturbance [40] and
facilitated taper-off from chronic benzodiazepines, signif-
icantly reducing anxiety symptoms and withdrawal sever-
ity [28].
To date, few studies have investigated factors predictive of
which GAD patients will benefit from long-term treatment.
Some reviewed studies investigated predictors of remission
after 6 months of open-label treatment with venlafaxine XR.
They suggested low neuroticism as a clinical predictor [33], in
accord with previous reports of different compounds [68]. In
addition, several genetic polymorphisms have been suggested
as biological predictors, mainly in genes involved in the sero-
tonin system and pituitary adenylate cyclase-activating pep-
tide [36••, 37, 38]. Interestingly, patients with one or two
copies of the serotonin receptor 2A (HTR2A) gene
rs7997012 SNP G allele showed an increasingly significant
anxiety symptom improvement as time progressed compared
with those with an A/A genotype, suggesting that GAD pa-
tients with at least one G allele should continue antidepressant
treatment beyond 12 weeks to maximize their potential re-
sponse to long-term treatment.
In conclusion, recent studies suggest that agomelatine,
vortioxetine, pregabalin, and quetiapine are effective for
long-term GAD treatment. No additional data was provided
about the optimal duration of long-term treatment indicated by
current clinical guidelines (i.e., to continue treatment for 6–
12 months after improvement is obtained by short-term ther-
apy) [48•, 64]. Similarly, no useful data was found about op-
timal compounds or dosages to minimize the relapse risk. No
recent studies have investigated to what extent long-term ther-
apies can treat psychiatric comorbidities.
Social Anxiety Disorder
No studies were found. For recent reviews, see [69•, 70•].
Conclusions
Much more work is evidently required to clarify the open
issues about long-term treatments for AnxDs. Several limita-
tions of the available recent studies should be considered and
addressed by future research. One is the paucity of random-
ized, controlled, comparative studies (for SAD in particular,
recent studies were completely lacking), while another is the
Page 13 of 16 23
lack of standardized methodological criteria to define treat-
ment outcome. The frequent use of unsuitable outcome mea-
sures, such as clinical global improvement (CGI scale), and, in
PD, the total panic-phobic symptoms score or the self-
reported number of PAs without clinical interview, is also
problematic. This approach does not disentangle medication
effects on different clinical features of the disorder and may
affect the validity of the results. A fourth limitation is the lack
of biomarker or endophenotype approaches to decrease the
variability within nosographic diagnoses. For example, in
PD, hypersensitivity to hypercapnia is considered an
endophenotype of panic, is strongly related to genetic factors,
and is associated with respiratory symptoms, higher PA fre-
quency, and familiarity for PD, probably typifying a
respiratory-panic subtype. Abnormalities in respiratory/auto-
nomic/balance system functions were found in PD patients
that may result in different clinical symptoms and outcomes
[43, 71••]. Among older GAD patients, some had higher cor-
tisol levels associated with poorer cognitive performance that
may benefit from specific interventions [72]. Incorporating
endophenotypes or patterns of neurobiological functions/
clinical profiles in pharmacological studies may help to select
truly homogeneous patients to identify more appropriate out-
come measures and test the efficacies of compounds on spe-
cific symptoms and functions. Another limitation is the pau-
city of investigations about predictors of response/relapse to
identify patients that may benefit most from long-term treat-
ments. Predictors can include clinical profiles/neurobiological
functions, biomarkers, neuroimaging, and allelic variations
modulating treatment responses [73, 74]. Some preliminary
studies have investigated potential predictors of short-term
outcome in AnxDs. In GAD adults, a lower pre-treatment
amygdala response to facial expressions and greater activation
of the anterior cingulate cortex have been associated with
better responses to venlafaxine [75], while a stronger pre-
treatment loudness dependence of the auditory evoked poten-
tial predicted response to escitalopram [76]. In older GAD
patients treated with escitalopram, treatment-related cortisol
reduction predicted improvements in neuropsychological
function [77]. Although results are scant and mixed, genetic
biomarkers of short-term treatment outcome have been inves-
tigated in PD, particularly serotonergic candidate genes [74].
Similarly, serotonin transporter promoter haplotype has been
associated with variability in short-term SSRI efficacy for
GAD [78], while variants in corticotropin-releasing hormone
receptor 1, dopamine receptor D3, nuclear receptor subfamily
group C member 1, and phosphodiesterase 1A have been as-
sociated with duloxetine response [79]. Finally, our group
found that in PD patients, the decrease in behavioral reactivity
to hypercapnia after the first week of treatment with TCAs or
SSRIs was a significant predictor of good clinical outcome
after 1 month [55]. Unfortunately, there are few studies on
outcome predictors of long-term pharmacological treatment
23 Page 14 of 16
(see previous sections), indicating that this research area needs
substantial expansion.
The ultimate goal of a personalized approach to pharmaco-
therapy is to tailor medications according to each patient’s
unique features. Personalized treatments for AnxDs can be
implemented using predictive tools [80] to explore those fac-
tors affecting treatment response/tolerability heterogeneity, in-
cluding gender, familiarity, clinical profiles, medical/
psychiatric comorbidity, biomarkers, and genetic features,
and to identify the most appropriate medications for each in-
dividual in terms of efficacy, tolerability, and treatment
duration.
Acknowledgments The authors thank Massimiliano Grassi, MSc, and
Raffaele Balletta, MD, Department of Clinical Neurosciences, Villa San
Benedetto Menni, Hermanas Hospitalarias, FoRiPsi, Albese con
Cassano, Como, Italy, for their contribution to literature search. Mr.
Grassi and Mr. Balletta have no conflicts of interest to declare.
Compliance with Ethical Standards
Conflict of Interest The authors declare that they have no competing
interests.
Human and Animal Rights and Informed Consent This article does
not contain any studies with human or animal subjects performed by any
of the authors.
References
Papers of particular interest, published recently, have been
highlighted as:
• Of importance
•• Of major importance
1. Kessler RC, Petukhova M, Sampson NA, Zaslavsky AM,
Wittchen HU. Twelve-month and lifetime prevalence and life-
time morbid risk of anxiety and mood disorders in the United
States. Int J Methods Psychiatr Res. 2012;21(3):169–84. doi:
10.1002/mpr.1359.
2. American Psychiatric Association. Diagnostic and statistical man-
ual of mental disorders (fifth ed.). Arlington: American Psychiatric
Association; 2013.
3. Bruce SE, Yonkers KA, Otto MW, Eisen JL, Weisberg RB, Pagano
M, et al. Influence of psychiatric comorbidity on recovery and
recurrence in generalized anxiety disorder, social phobia, and panic
disorder: a 12-year prospective study. Am J Psychiatry.
2005;162(6):1179–87. doi:10.1176/appi.ajp.162.6.1179.
4. Berger A, Edelsberg J, Bollu V, Alvir JM, Dugar A, Joshi AV, et
al. Healthcare utilization and costs in patients beginning phar-
macotherapy for generalized anxiety disorder: a retrospective
cohort study. BMC Psychiatry. 2011;11:193. doi:10.1186/
1471-244X-11-193.
5. Batelaan N, Smit F, de Graaf R, van Balkom A, Vollebergh W,
Beekman A. Economic costs of full-blown and subthreshold panic
disorder. J Affect Disord. 2007;104(1-3):127–36. doi:10.1016/j.jad.
2007.03.013.
Curr Psychiatry Rep (2016) 18:23
6. Wittchen HU, Fuetsch M, Sonntag H, Muller N, Liebowitz M.
Disability and quality of life in pure and comorbid social pho-
bia. Findings from a controlled study. Eur Psychiatry.
2000;15(1):46–58.
7. Tyrer P, Seivewright H, Johnson T. The Nottingham Study of
Neurotic Disorder: predictors of 12-year outcome of dysthymic,
panic and generalized anxiety disorder. Psychol Med. 2004;34(8):
1385–94.
8. American Psychiatric Association. Diagnostic and statistical man-
ual of mental disorders DSM-III. Washington, DC: Amer
Psychiatric Pub; 1980.
9. American Psychiatric Association. Diagnostic and statistical man-
ual of mental disorders DSM-III-R. Washington, DC: Amer
Psychiatric PuB; 1987.
10. American Psychiatric Association. Diagnostic and statistical man-
ual of mental disorders DSM-IV. Washington, DC: Amer
Psychiatric Pub; 1994.
11. American Psychiatric Association. Diagnostic and statistical man-
ual of mental disorders DSM-IV-TR fourth edition (text revision).
4th ed. Washington, DC: Amer Psychiatric Pub; 2000.
12. Centers for Disease Control and Prevention. International
Classification of Diseases, Ninth Revision (ICD-9). http://www.
cdc.gov/nchs/icd/icd9.htm.
13. Centers for Disease Control and Prevention. International
Classification of Diseases, Ninth Revision, Clinical Modification
(ICD-9-CM). http://www.cdc.gov/nchs/icd/icd9cm.htm.
14. Centers for Disease Control and Prevention. International
Classification of Diseases, Tenth Revision (ICD-10). http://www.
cdc.gov/nchs/icd/icd10.htm.
15. American Psychiatric Association. Practice Guidelines for the treat-
ment of patients with Panic Disorder. Second ed. American
Psychiatric Association; 2009. http://www.psychiatryonline.com/
pracGuide/pracGuideTopic_9.aspx
16. Nardi AE, Valenca AM, Freire RC, Amrein R, Sardinha A, Levitan
MN, et al. Randomized, open naturalistic, acute treatment of panic
disorder with clonazepam or paroxetine. J Clin Psychopharmacol.
2011;31(2):259–61. doi:10.1097/JCP.0b013e318210b4ee.
17. Nardi AE, Freire RC, Mochcovitch MD, Amrein R, Levitan MN,
King AL, et al. A randomized, naturalistic, parallel-group study for
the long-term treatment of panic disorder with clonazepam or par-
oxetine. J Clin Psychopharmacol. 2012;32(1):120–6. doi:10.1097/
JCP.0b013e31823fe4bd.
18. Van Apeldoorn FJ, Van Hout WJ, Timmerman ME, Mersch PP, den
Boer JA. Rate of improvement during and across three treatments for
panic disorder with or without agoraphobia: cognitive behavioral
therapy, selective serotonin reuptake inhibitor or both combined. J
Affect Disord. 2013;150(2):313–9. doi:10.1016/j.jad.2013.04.012.
19. Hendriks GJ, Keijsers GP, Kampman M, Oude Voshaar RC,
Verbraak MJ, Broekman TG, et al. A randomized controlled study
of paroxetine and cognitive-behavioural therapy for late-life panic
disorder. Acta Psychiatr Scand. 2010;122(1):11–9. doi:10.1111/j.
1600-0447.2009.01517.x.
20. Hendriks GJ, Keijsers GP, Kampman M, Hoogduin CA, Oude
Voshaar RC. Predictors of outcome of pharmacological and psy-
chological treatment of late-life panic disorder with agoraphobia.
Int J Geriatr Psychiatry. 2012;27(2):146–50. doi:10.1002/gps.2700.
21. Choi KW, Woo JM, Kim YR, Lee SH, Lee SY, Kim EJ, et al. Long-
term escitalopram treatment in Korean patients with panic disorder:
a prospective, naturalistic, open-label, multicenter trial. Clin
Psychopharmacol Neurosci. 2012;10(1):44–8. doi:10.9758/cpn.
2012.10.1.44.
22. Miniati M, Calugi S, Rucci P, Shear MK, Benvenuti A, Santoro D,
et al. Predictors of response among patients with panic disorder
treated with medications in a naturalistic follow-up: the role of adult
separation anxiety. J Affect Disord. 2012;136(3):675–9. doi:10.
1016/j.jad.2011.10.008.
Curr Psychiatry Rep (2016) 18:23
23. Stein DJ, Ahokas A, Albarran C, Olivier V, Allgulander C.
Agomelatine prevents relapse in generalized anxiety disorder: a 6-
month randomized, double-blind, placebo-controlled discontinua-
tion study. J Clin Psychiatry. 2012;73(7):1002–8. doi:10.4088/JCP.
11m07493.
24. Wetherell JL, Petkus AJ, White KS, Nguyen H, Kornblith S,
Andreescu C, et al. Antidepressant medication augmented with
cognitive-behavioral therapy for generalized anxiety disorder in
older adults. Am J Psychiatry. 2013;170(7):782–9. doi:10.1176/
appi.ajp.2013.12081104.
25. Mork A, Pehrson A, Brennum LT, Nielsen SM, Zhong H, Lassen
AB, et al. Pharmacological effects of Lu AA21004: a novel multi-
modal compound for the treatment of major depressive disorder. J
Pharmacol Exp Ther. 2012;340(3):666–75. doi:10.1124/jpet.111.
189068.
26. Baldwin DS, Loft H, Florea I. Lu AA21004, a multimodal psycho-
tropic agent, in the prevention of relapse in adult patients with
generalized anxiety disorder. Int Clin Psychopharmacol.
2012;27(4):197–207. doi:10.1097/YIC.0b013e3283530ad7.
27. Kasper S, Iglesias-Garcia C, Schweizer E, Wilson J, DuBrava S,
Prieto R, et al. Pregabalin long-term treatment and assessment of
discontinuation in patients with generalized anxiety disorder. Int J
Neuropsychopharmacol. 2014;17(5):685–95. doi:10.1017/
S1461145713001557.
28. Hadley SJ, Mandel FS, Schweizer E. Switching from long-term
benzodiazepine therapy to pregabalin in patients with generalized
anxiety disorder: a double-blind, placebo-controlled trial. J
P s y c h o p h a r m a c o l . 2 0 1 2 ; 2 6 ( 4 ) : 4 6 1 – 7 0 . d o i :1 0 . 11 7 7 /
0269881111405360.
29. Katzman MA, Brawman-Mintzer O, Reyes EB, Olausson B, Liu S,
Eriksson H. Extended release quetiapine fumarate (quetiapine XR)
monotherapy as maintenance treatment for generalized anxiety dis-
order: a long-term, randomized, placebo-controlled trial. Int Clin
Psychopharmacol. 2011;26(1):11–24. doi:10.1097/YIC.
0b013e32833e34d9.
30. Sheehan DV, Svedsater H, Locklear JC, Eriksson H. Effects of
extended-release quetiapine fumarate on long-term functioning
and sleep quality in patients with Generalized Anxiety Disorder
(GAD): data from a randomized-withdrawal, placebo-controlled
maintenance study. J Affect Disord. 2013;151(3):906–13. doi:10.
1016/j.jad.2013.07.037.
31. Endicott J, Svedsater H, Locklear JC. Effects of once-daily extend-
ed release quetiapine fumarate on patient-reported outcomes in pa-
tients with generalized anxiety disorder. Neuropsychiatr Dis Treat.
2012;8:301–11. doi:10.2147/NDT.S32320.
32. Rickels K, Etemad B, Khalid-Khan S, Lohoff FW, Rynn MA,
Gallop RJ. Time to relapse after 6 and 12 months’ treatment of
generalized anxiety disorder with venlafaxine extended release.
Arch Gen Psychiatry. 2010;67(12):1274–81. doi:10.1001/
archgenpsychiatry.2010.170.
33. Rickels K, Etemad B, Rynn MA, Lohoff FW, Mandos LA, Gallop
R. Remission of generalized anxiety disorder after 6 months of
open-label treatment with venlafaxine XR. Psychother
Psychosom. 2013;82(6):363–71. doi:10.1159/000351410.
34. Narasimhan S, Aquino TD, Multani PK, Rickels K, Lohoff FW.
Variation in the catechol-O-methyltransferase (COMT) gene and
treatment response to venlafaxine XR in generalized anxiety disor-
der. Psychiatry Res. 2012;198(1):112–5. doi:10.1016/j.psychres.
2011.12.034.
35. Cooper AJ, Rickels K, Lohoff FW. Association analysis between
the A118G polymorphism in the OPRM1 gene and treatment re-
sponse to venlafaxine XR in generalized anxiety disorder. Hum
Psychopharmacol. 2013;28(3):258–62. doi:10.1002/hup.2317.
36.•• Lohoff FW, Aquino TD, Narasimhan S, Multani PK, Etemad B,
Rickels K. Serotonin receptor 2A (HTR2A) gene polymorphism
predicts treatment response to venlafaxine XR in generalized
Page 15 of 16 23
anxiety disorder. Pharmacogenomics J. 2013;13(1):21–6. doi:10.
1038/tpj.2011.47. The first study showing a role of serotonergic
system polymorphisms in long-term treatment response in gen-
eralized anxiety disorder.
37. Lohoff FW, Narasimhan S, Rickels K. Interaction between poly-
morphisms in serotonin transporter (SLC6A4) and serotonin recep-
tor 2A (HTR2A) genes predict treatment response to venlafaxine
XR in generalized anxiety disorder. Pharmacogenomics J.
2013;13(5):464–9. doi:10.1038/tpj.2012.33.
38. Cooper AJ, Narasimhan S, Rickels K, Lohoff FW. Genetic poly-
morphisms in the PACAP and PAC1 receptor genes and treatment
response to venlafaxine XR in generalized anxiety disorder.
Psychiatry Res. 2013;210(3):1299–300. doi:10.1016/j.psychres.
2013.07.038.
39. Montgomery S, Emir B, Haswell H, Prieto R. Long-term treatment
of anxiety disorders with pregabalin: a 1 year open-label study of
safety and tolerability. Curr Med Res Opin. 2013;29(10):1223–30.
doi:10.1185/03007995.2013.820694.
40. Ruiz MA, Alvarez E, Carrasco JL, Olivares JM, Perez M, Rejas J.
Modeling the longitudinal latent effect of pregabalin on self-
reported changes in sleep disturbances in outpatients with general-
ized anxiety disorder managed in routine clinical practice. Drug Des
Devel Ther. 2015;9:4329–40. doi:10.2147/DDDT.S88238.
41. Alvarez E, Carrasco JL, Olivares JM, Lopez-Gomez V, Vilardaga I,
Perez M. Broadening of generalized anxiety disorders definition
does not affect the response to psychiatric care: findings from the
observational ADAN study. Clin Pract Epidemiol Ment Health.
2012;8:158–68. doi:10.2174/1745017901208010158.
42. Freire RC, Hallak JE, Crippa JA, Nardi AE. New treatment options
for panic disorder: clinical trials from 2000 to 2010. Expert Opin
Pharmacother. 2011;12(9):1419–28. doi:10.1517/14656566.2011.
562200.
43. Perna G, Guerriero G, Caldirola D. Emerging drugs for panic dis-
order. Expert Opin Emerg Drugs. 2011;16(4):631–45. doi:10.1517/
14728214.2011.628313.
44. Batelaan NM, Van Balkom AJ, Stein DJ. Evidence-based pharma-
cotherapy of panic disorder: an update. Int J
Neuropsychopharmacol. 2012;15(3):403–15. doi:10.1017/
S1461145711000800.
45. Andersch S, Hetta J. A 15-year follow-up study of patients with
panic disorder. Eur Psychiatry. 2003;18(8):401–8.
46. Dannon PN, Iancu I, Cohen A, Lowengrub K, Grunhaus L, Kotler
M. Three year naturalistic outcome study of panic disorder patients
treated with paroxetine. BMC Psychiatry. 2004;4:16. doi:10.1186/
1471-244X-4-16.
47. National Collaborating Centre for Mental Health. Generalised anx-
iety disorder in adults: management in primary, secondary and com-
munity care. National clinical guideline number 113 ed. Leicester:
The British Psychological Society and The Royal College of
Psychiatrists; 2011.
48.• Bandelow B, Lichte T, Rudolf S, Wiltink J, Beutel ME. The diag-
nosis of and treatment recommendations for anxiety disorders.
Dtsch Arztebl Int. 2014;111(27-28):473–80. doi:10.3238/arztebl.
2014.0473. Recent recommendations for pharmacological and
non-pharmacological treatments in Anxiety Disorders.
49. Donovan MR, Glue P, Kolluri S, Emir B. Comparative efficacy of
antidepressants in preventing relapse in anxiety disorders—a meta-
analysis. J Affect Disord. 2010;123(1-3):9–16. doi:10.1016/j.jad.
2009.06.021.
50. Mavissakalian MR, Perel JM. 2nd year maintenance and discontin-
uation of imipramine in panic disorder with agoraphobia. Ann Clin
Psychiatry. 2001;13(2):63–7.
51. Mavissakalian MR, Perel JM. Duration of imipramine therapy and
relapse in panic disorder with agoraphobia. J Clin
Psychopharmacol. 2002;22(3):294–9.
23 Page 16 of 16
52. Lecrubier Y, Judge R. Long-term evaluation of paroxetine, clo-
mipramine and placebo in panic disorder. Collaborative
Paroxetine Panic Study Investigators. Acta Psychiatr Scand.
1997;95(2):153–60.
53. Schweizer E, Rickels K, Weiss S, Zavodnick S. Maintenance drug
treatment of panic disorder. I. Results of a prospective, placebo-
controlled comparison of alprazolam and imipramine. Arch Gen
Psychiatry. 1993;50(1):51–60.
54. Dannon PN, Iancu I, Lowengrub K, Gonopolsky Y, Musin E,
Grunhaus L, et al. A naturalistic long-term comparison study of
selective serotonin reuptake inhibitors in the treatment of panic
disorder. Clin Neuropharmacol. 2007;30(6):326–34. doi:10.1097/
WNF.0b013e318064579f.
55. Perna G, Bertani A, Caldirola D, Gabriele A, Cocchi S, Bellodi
L. Antipanic drug modulation of 35% CO2 hyperreactivity and
short-term treatment outcome. J Clin Psychopharmacol.
2002;22(3):300–8.
56. Perna G, Bertani A, Caldirola D, Smeraldi E, Bellodi L. A compar-
ison of citalopram and paroxetine in the treatment of panic disorder:
a randomized, single-blind study. Pharmacopsychiatry. 2001;34(3):
85–90. doi:10.1055/s-2001-14283.
57. Furukawa TA, Watanabe N, Churchill R. Psychotherapy plus anti-
depressant for panic disorder with or without agoraphobia: system-
atic review. Br J Psychiatry. 2006;188:305–12. doi:10.1192/bjp.
188.4.305.
58. Cuijpers P, Sijbrandij M, Koole SL, Andersson G, Beekman AT,
Reynolds 3rd CF. Adding psychotherapy to antidepressant medica-
tion in depression and anxiety disorders: a meta-analysis. World
Psychiatry. 2014;13(1):56–67. doi:10.1002/wps.20089.
59. Deckersbach T, Moshier SJ, Tuschen-Caffier B, Otto MW. Memory
dysfunction in panic disorder: an investigation of the role of chronic
benzodiazepine use. Depress Anxiety. 2011;28(11):999–1007. doi:
10.1002/da.20891.
60. Barker MJ, Greenwood KM, Jackson M, Crowe SF. Persistence of
cognitive effects after withdrawal from long-term benzodiazepine
use: a meta-analysis. Arch Clin Neuropsychol. 2004;19(3):437–54.
doi:10.1016/S0887-6177(03)00096-9.
61. Verdoux H, Lagnaoui R, Begaud B. Is benzodiazepine use a risk
factor for cognitive decline and dementia? A literature review of
epidemiological studies. Psychol Med. 2005;35(3):307–15.
62. Soumerai SB, Simoni-Wastila L, Singer C, Mah C, Gao X, Salzman
C, et al. Lack of relationship between long-term use of benzodiaz-
epines and escalation to high dosages. Psychiatr Serv. 2003;54(7):
1006–11.
63. Nardi AE, Freire RC, Valenca AM, Amrein R, de Cerqueira AC,
Lopes FL, et al. Tapering clonazepam in patients with panic disor-
der after at least 3 years of treatment. J Clin Psychopharmacol.
2010;30(3):290–3. doi:10.1097/JCP.0b013e3181dcb2f3.
64. Bandelow B, Boerner JR, Kasper S, Linden M, Wittchen HU,
Moller HJ. The diagnosis and treatment of generalized anxiety dis-
order. Dtsch Arztebl Int. 2013;110(17):300–9. doi:10.3238/arztebl.
2013.0300. quiz 10.
65.• Freiesleben SD, Furczyk K. A systematic review of agomelatine-
induced liver injury. J Mol Psychiatry. 2015;3(1):4. doi:10.1186/
s40303-015-0011-7. An updated review, based on clinical
trials, non-interventional studies and pharmacovigilance data-
bases, on potential agomelatine-related risk of liver injury.
66. Feltner D, Wittchen HU, Kavoussi R, Brock J, Baldinetti F, Pande
AC. Long-term efficacy of pregabalin in generalized anxiety disor-
der. Int Clin Psychopharmacol. 2008;23(1):18–28. doi:10.1097/
YIC.0b013e3282f0f0d7.
Curr Psychiatry Rep (2016) 18:23
67. Newcomer JW. Comparing the safety and efficacy of atypical
antipsychotics in psychiatric patients with comorbid medical
illnesses. J Clin Psychiatry. 2009;70 Suppl 3:30–6. doi:10.
4088/JCP.7075su1c.05.
68. Massion AO, Dyck IR, Shea MT, Phillips KA, Warshaw MG,
Keller MB. Personality disorders and time to remission in general-
ized anxiety disorder, social phobia, and panic disorder. Arch Gen
Psychiatry. 2002;59(5):434–40.
69.• Blanco C, Bragdon LB, Schneier FR, Liebowitz MR. The evidence-
based pharmacotherapy of social anxiety disorder. Int J
Neuropsychopharmacol. 2013;16(1):235–49. doi:10.1017/
S1461145712000119. A recent review on pharmacological
treatments for Social Anxiety Disorder based on placebo-
controlled studies and meta-analyses.
70.• Canton J, Scott KM, Glue P. Optimal treatment of social phobia:
systematic review and meta-analysis. Neuropsychiatr Dis Treat.
2012;8:203–15. doi:10.2147/NDT.S23317. A systematic
literature review and meta-analysis on the evidence-based phar-
macological and non-pharmacological treatments for social
phobia.
71.•• Perna G, Schruers K, Alciati A, Caldirola D. Novel investigational
therapeutics for panic disorder. Expert Opin Investig Drugs.
2015;24(4):491–505. doi:10.1517/13543784.2014.996286. A
recent review on novel mechanism-based anti-panic drugs un-
der current investigation in animal studies up to phase- II stud-
ies, with a focus on the translational validity of animal models.
72. Rosnick CB, Rawson KS, Butters MA, Lenze EJ. Association of
cortisol with neuropsychological assessment in older adults with
generalized anxiety disorder. Aging Ment Health. 2013;17(4):
432–40. doi:10.1080/13607863.2012.761673.
73. Maron E, Nutt D. Biological predictors of pharmacological therapy
in anxiety disorders. Dialogues Clin Neurosci. 2015;17(3):305–17.
74. Caldirola D, Perna G. Is there a role for pharmacogenetics in the
treatment of panic disorder? Pharmacogenomics. 2015;16(8):771–
4. doi:10.2217/pgs.15.66.
75. Shin LM, Davis FC, Vanelzakker MB, Dahlgren MK, Dubois SJ.
Neuroimaging predictors of treatment response in anxiety disor-
ders. Biol Mood Anxiety Disord. 2013;3(1):15. doi:10.1186/
2045-5380-3-15.
76. Park YM, Kim DW, Kim S, Im CH, Lee SH. The loudness depen-
dence of the auditory evoked potential (LDAEP) as a predictor of
the response to escitalopram in patients with generalized anxiety
disorder. Psychopharmacology (Berl). 2011;213(2-3):625–32. doi:
10.1007/s00213-010-2061-y.
77. Lenze EJ, Dixon D, Mantella RC, Dore PM, Andreescu C,
Reynolds 3rd CF, et al. Treatment-related alteration of cortisol pre-
dicts change in neuropsychological function during acute treatment
of late-life anxiety disorder. Int J Geriatr Psychiatry. 2012;27(5):
454–62. doi:10.1002/gps.2732.
78. Lenze EJ, Goate AM, Nowotny P, Dixon D, Shi P, Bies RR, et al.
Relation of serotonin transporter genetic variation to efficacy of
escitalopram for generalized anxiety disorder in older adults. J
Clin Psychopharmacol. 2010;30(6):672–7.
79. Perlis RH, Fijal B, Dharia S, Houston JP. Pharmacogenetic investi-
gation of response to duloxetine treatment in generalized anxiety
disorder. Pharmacogenomics J. 2013;13(3):280–5. doi:10.1038/tpj.
2011.62.
80. Vidyasagar M. Identifying predictive features in drug response
using machine learning: opportunities and challenges. Annu Rev
Pharmacol Toxicol. 2015;55:15–34. doi:10.1146/annurev-
pharmtox-010814-124502.