1. Choosing an Initial Treatment
CBT and SRIs are recommended on the basis of clinical trial
results as safe and effective first-line treatments for OCD. SRIs
include clomipramine and all of the SSRIs. Whether to recommend
a form of CBT, an SRI, or combined treatment will depend on a number
of factors. These include the nature and severity of the patient's
symptoms, the nature of any co-occurring psychiatric and medical
conditions and their treatments, the availability of CBT, and the
patient's past treatment history, current medications,
and preferences. Because most treatment studies have been of 3–4 months' duration,
only limited data are available to guide long-term treatment decisions
The evidence base for the form of CBT that relies primarily
on behavioral techniques, such as ERP (57), is the
strongest (58–60). Data also support the use
of CBT that focuses on cognitive techniques aimed at identifying, challenging,
and modifying dysfunctional beliefs (61–64)
if these techniques are combined with behavioral experiments. However,
some data suggest, and many clinical experts believe, that the most
effective form of CBT for OCD integrates exposure, response prevention
(behavior that results in not performing rituals), discussion of
feared consequences and dysfunctional beliefs, and relapse prevention.
There are few data from controlled trials to support cognitive therapy
without ERP or behavioral experiments.
CBT alone, consisting of ERP, is recommended as initial treatment
for a patient who is not too depressed, anxious, or severely ill
to cooperate with this treatment modality, or who prefers not to take
medications. The patient must be willing to do the work that CBT
requires (e.g., regular behavioral homework).
An SRI alone is recommended for a patient who has previously
responded well to a given drug or prefers treatment with an SRI
alone. Starting with an SRI alone may enhance cooperation with treatment
by diminishing symptom severity. Thus, an SRI alone may also be
considered in patients who have severe OCD or are not otherwise
able to cooperate with the demands of CBT. An SRI alone may also
be necessary if CBT is not accessible or available.
The available data suggest that combining an SRI and CBT consisting
of ERP is more effective than monotherapy in some patients but is
not necessary for all (65). Combined treatment should
be considered for patients with an unsatisfactory response to monotherapy,
for those with co-occurring psychiatric conditions for which SRIs
are effective, and for those who wish to limit the duration of treatment
with medication. In the latter instance, uncontrolled follow-up
studies suggest that CBT consisting of ERP may delay or mitigate
relapse when SRI treatment is discontinued (66–68).
Combined treatment may also be considered in patients with severe
OCD, since the medication may diminish symptom severity sufficiently
to allow the patient to engage in CBT.
2. Choosing a Specific Pharmacological
Clomipramine, fluoxetine, fluvoxamine, paroxetine, and sertraline,
which are approved by the FDA for treatment of OCD, are recommended
pharmacological agents. Although meta-analyses (59, 69, 70)
of placebo-controlled trials suggest greater efficacy for clomipramine
than for fluoxetine, fluvoxamine, and sertraline, the results of
head-to-head trials comparing clomipramine and SSRIs directly do
not support this impression (Section V.A.1). Because the SSRIs have
a less troublesome side-effect profile than clomipramine (see Section II.B.2.b), an SSRI is preferred for a first medication trial. Although
all SSRIs (including citalopram and escitalopram) appear to be equally effective,
individual patients may respond well to one and not to another.
The reasons for this patient-specific response are unknown, and
no demographic or clinical variables are sufficiently accurate predictors
of treatment outcome to permit their use in selecting medications
In choosing among the SSRIs, the psychiatrist should consider
the safety and acceptability of particular side effects for the
patient, including any applicable FDA warnings, potential drug interactions,
past treatment response, and the presence of co-occurring general
medical conditions. For example, paroxetine, the SSRI most associated
with weight gain (72) and the most anticholinergic
SSRI, would not be the first choice for patients with obesity, diabetes
mellitus, constipation, or urinary hesitancy.
Table 3. Dosing of Serotonin Reuptake Inhibitors (SRIs) in Obsessive-Compulsive
| Add to My POL
serotonin uptake inhibitorsin obsessive-compulsive disorderTable 3. Dosing of Serotonin Reuptake Inhibitors (SRIs) in Obsessive-Compulsive
|SRI||Starting Dose and Incremental Dose (mg/day)a||Usual Target Dose (mg/day)||Usual Maximum Dose (mg/day)||Occasionally Prescribed Maximum Dose (mg/day)b|
patients may need to start at half this dose or less to minimize
undesired side effects such as nausea or to accommodate anxiety
about taking medications.
bThese doses are sometimes used
for rapid metabolizers or for patients with no or mild side effects
and inadequate therapeutic response after 8 weeks or more at the
usual maximum dose.
cCombined plasma levels of clomipramine
plus desmethylclomipramine 12 hours after the dose should be kept
below 500 ng/mL to minimize risk of seizures and cardiac
dSertraline, alone among the SSRIs,
is better absorbed with food.
Another factor in choosing among medications
is the degree to which they alter metabolism through the hepatic
cytochrome P450 enzyme system or uridine 5'-diphosphate
glucuronosyltransferases (UGTs), act at the P-glycoprotein transporter,
or displace drugs tightly bound to plasma proteins (e.g., see references
73–76). Many interactions, however, reflect only in vitro
data, and their clinical importance is not established. Citalopram,
escitalopram, and sertraline (along with venlafaxine and mirtazapine)
have few or no important known drug interactions. Web sites providing
data on potential drug interactions include http://medicine.iupui.edu/flockhart
and http://mhc.com/Cytochromes. For up-to-date
clinical reports of interactions between specific SRIs and other
medications, psychiatrists can consult the federal National Library
of Medicine database at http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?DB=pubmed, which
is also accessible by entering the term "pubmed" in
a search engine. Since there are very few serious risks associated
with treatment with SSRIs (e.g., the risk of serotonin syndrome
from adding an SSRI to an MAOI, tramadol, meperidine, or dextromethorphan [77,
78]), the psychiatrist will much more often have to consider
the relative potential for specific SSRIs to interact with the patient's
other medications, particularly given the higher doses of SSRIs
that are often used in treating OCD.
Although no definitive data are available, the response of
first-degree relatives to particular medications may be predictive
of the patient's response because of genetic similarity.
This is a subject, however, for future research.
a. Implementing Pharmacotherapy
The need to educate the patient about any medication recommended
has been emphasized earlier. Table 3 displays suggested starting
doses, known effective doses, maximum recommended doses, and maximum
doses occasionally prescribed for each SRI. For most patients, the
starting dose is that recommended by the manufacturer.
For patients who are worried about side effects, the medication
can be started at half the listed dose or less, since many SSRIs
(e.g., citalopram, escitalopram, fluoxetine, paroxetine, and sertraline)
are available in liquid form or in pills that can be split. Lower
initial medication starting doses and more gradual dose titration
may also be needed in patients with co-occurring anxiety disorders,
who may experience a transient worsening of anxiety symptoms. Experience
with pharmacotherapy in the elderly indicates that lower starting
doses of medication and a more gradual approach to dose increase
are often advisable. Most patients will not experience substantial improvement
until 4–6 weeks after starting medication, and some who
will ultimately respond will experience little improvement for as
many as 10–12 weeks. Available trial data suggest that higher
SSRI doses produce a somewhat higher response rate and somewhat
greater magnitude of symptom relief (79–82).
Some patients, such as those who have had little response to previous
treatments and are tolerating the medication well, may benefit from
even higher doses than those shown in the last column of Table 3.
In these instances, the patient should be closely monitored for
side effects, including the serotonin syndrome. Moreover, patients
who have not responded to a known effective dose after 10–12
weeks may respond at higher doses (83, 84).
For this reason, some clinicians prefer to titrate doses more rapidly
(in weekly increments to the maximum recommended dose if this is
comfortably tolerated), rather than waiting for 1–2 months before
each dose increment to evaluate results.
No evidence suggests that OCD treatment outcome is related
to plasma levels of clomipramine (85), fluoxetine (86),
fluvoxamine (87), or sertraline (82).
However, these studies were not designed to identify whether rapid
or ultra-rapid metabolizers of these medications were more likely
to be nonresponders.
b. Managing Medication Side Effects
Unlike the SSRIs, clomipramine also blocks norepinephrine
reuptake, muscarinic cholinergic receptors, H1 histamine
receptors, and 1-adrenergic
receptors, as well as sodium channels in the heart and brain. Thus,
clomipramine is more likely to induce anticholinergic effects such
as tachycardia, dry mouth, constipation, and blurred vision, although
these typically diminish over time. Clomipramine is also more likely
to induce delayed urination or, uncommonly, urinary retention. Histaminic
blockade is associated with weight gain and sedation. Adrenergic
blockade may lead to orthostatic hypotension and postural dizziness.
Sodium channel blockade can induce cardiac arrhythmias or seizures
(estimated to occur in 0.7% of patients treated with clomipramine
at a dose of up to 300 mg/day for as many as 6 years ).
In view of clomipramine's less favorable side-effect profile,
expert opinion favors one or more SSRI trials before trying clomipramine
(89). Starting clomipramine at a dose of 25 mg/day
or less will increase its early tolerability (90).
The most common side effects of the SSRIs
include gastrointestinal distress (especially in the first weeks
of treatment), agitation, insomnia or somnolence, increased tendency
to sweat, and sexual side effects, including diminished libido and
difficulty with erection and orgasm. A first step in managing any
side effect is to consider whether lowering the drug dose may alleviate
the side effect without loss of therapeutic effect. When this is
not possible, specific interventions may be considered. Gastrointestinal
distress can be minimized by starting with low doses; if mild queasiness
or nausea occurs, it will usually disappear within 1–2
weeks at a constant dose. Insomnia may respond to the patient's
taking the medication in the morning or to standard sleep hygiene measures,
or may require addition of a sleep-promoting agent. Fatigue or sleepiness
may respond to the addition of modest doses of modafinil (91).
Cases of successful treatment of sweating have been reported with
low doses of anticholinergic agents such as benztropine (92, 93),
and with clonidine (94), cyproheptadine (95),
and mirtazapine (96).
Few controlled trials have been published regarding the management
of sexual side effects, which may affect one-third or more of patients
(97). Management approaches include reducing the dose
to the minimal effective dose; waiting for the symptom to remit
(which clinical impression suggests may occur within 2 months in
about 10% of patients); trying a once-weekly, one-day "drug
holiday" before engaging in sexual activity; switching
to another SSRI (which may relieve the sexual dysfunction but not
control the patient's OCD); or adding a counteracting pharmacologic
agent. The drug holiday approach may alleviate difficulties with
erection or orgasm but not with libido. It is not effective for
fluoxetine because of its long half-life (98) and may
induce withdrawal symptoms if attempted with paroxetine or venlafaxine.
Taking drug holidays more than once weekly risks a return of OCD
symptoms. Case series and primarily uncontrolled studies report
modest success in restoring libido, erection, and orgasm by addition
of amantadine, bupropion, buspirone, yohimbine, Ginkgo biloba extract
(99), ropinirole (100), or stimulants
such as methylphenidate or dextroamphetamine. Adding bupropion has
the best evidence base (101), but even this literature
is mixed (102). Case series and uncontrolled studies
report modest success in restoring erection or orgasm, but not libido,
with cyproheptadine or mirtazapine (99). Controlled
trials support the use of sildenafil, tadalafil, and vardenafil
(103, 104) to restore erection and, in
men (105) and women (106), to restore
Primarily on the basis of data in children and adolescents
(Section III.B.4), concerns have been raised about the potential
for increases in self-harming or suicidal behaviors in individuals
treated with antidepressant medications, including SRIs. A meta-analysis
in adults treated with SSRIs did not demonstrate increases in rates
of suicide or suicidal thoughts, although there was weak evidence
for an increase in self-harming behavior (107). A second
meta-analysis noted an increase in the odds ratio for suicide attempts
but did not report on the risk of other suicidal behaviors (108).
However, interpretation of these results is difficult because of
the confounds involved in calculating risks of suicide and other
suicidal behaviors from meta-analyses (109). This is
particularly true with regard to the use of antidepressants to treat
OCD, because the majority of SSRI clinical trials involve depressed
subjects, not subjects with OCD. In addition, studies using other
methodologies, including a nested case–control study (110),
a retrospective analysis of a large sample of computerized health
plan records (111), and a large prospective effectiveness
study in adult subjects with bipolar disorder (112),
showed no increases in the likelihood of suicide or suicide
attempts with antidepressant treatment. An additional nested case–control
study also showed no increase in the risk of suicide but did note
a small increase in the likelihood of self-harm (113).
Although antidepressant treatment in adults does not appear to be
associated with an increase in suicide, it is conceivable that side
effects (e.g., anxiety, agitation, insomnia, irritability) may increase
the chance of self-harming behaviors in some individuals (109).
Thus, careful monitoring for such side effects, as well as for evidence
of self-harming or suicidal thoughts or behaviors, is important,
particularly in the early phases of treatment and after increases
in antidepressant dose. Against these small risks, the benefits
of antidepressant treatment must always be considered (114, 115).
SSRIs may be associated with increased intra-operative blood
loss in patients also taking nonsteroidal anti-inflammatory drugs
(116) and, along with clomipramine, may interact with
anesthetics and opiate pain relievers. Patients should inform their
surgeon and anesthesiologist if they are taking an SRI.
A drug discontinuation syndrome consisting most often of dizziness,
nausea/vomiting, headache, and lethargy, but also including
agitation, insomnia, myoclonic jerks, and paresthesias (117, 118),
may occur if medication is suddenly stopped. The syndrome may occur
after stopping any SRI but is most often seen after rapid discontinuation
of paroxetine or the serotonin-norepinephrine reuptake inhibitor
(SNRI) venlafaxine (117). Particularly for these latter
medications, a slow taper over several weeks or more will minimize the
likelihood of discontinuation symptoms. If symptoms do occur, raising
the medication dose and slowing the taper may bring relief.
3. Choosing a Specific Form of Psychotherapy
CBT is the only form of psychotherapy for
OCD whose effectiveness is supported by controlled trials. There
are no controlled studies that demonstrate effectiveness of dynamic
psychotherapy or psychoanalysis in dealing with the core symptoms
of OCD. Psychodynamic psychotherapy may still be useful in helping
patients overcome their resistance to accepting a recommended treatment by
illuminating their reasons for wanting to stay as they are (e.g.,
best adaptation, secondary gains). It may also be useful in addressing
the interpersonal consequences of the OCD symptoms (119).
Motivational interviewing may also help overcome resistance to treatment.
As noted in Section II.B.1, the CBT variant that relies primarily
on behavioral techniques such as ERP has the strongest evidence
base (59, 60). Some data suggest that
ERP is more effective if it integrates habituation with discussions
of feared consequences and dysfunctional beliefs (120, 121)
and with relapse prevention (122–125). For
OCD patients without co-occurring depression, data from one large
(N = 122) randomized controlled
trial suggest that intensive CBT consisting of ERP may
be superior to clomipramine monotherapy (123, 126).
Data also support CBT that primarily utilizes cognitive techniques
such as identifying, challenging, and modifying dysfunctional beliefs
when these techniques are combined with behavioral experiments (64, 121, 127–129),
which can resemble ERP depending on how they are done. In direct
comparisons, CBT utilizing cognitive techniques and behavioral experiments
had efficacy similar to CBT consisting of ERP that focused only
on habituation. Whether incorporating cognitive techniques with
ERP is more effective than either treatment alone is under
investigation (122). Only case reports support attempting
to treat OCD with cognitive therapy alone, without exposure or behavioral
experiments (129, 130). No controlled
trials have evaluated other psychosocial methods for treating OCD
that have been used in clinical practice (e.g., the "brain-lock
technique" , other
mindfulness approaches, acceptance and commitment therapy).
4. Implementing Cognitive-Behavioral
Cognitive-behavioral therapies have been delivered in individual,
group (132–134), and family therapy sessions,
with session length varying from less than 1 hour to 2 hours (135, 136)
(for a summary of group and family therapy studies, see references 136, 137).
One group has explored the delivery of behavior therapy by means
of a self-instructional workbook that allows the patient to design
and implement an individualized treatment plan. The patient couples
the plan with a voice-activated telephone-response system accessible
24 hours a day to monitor and report progress (138, 139).
The literature and expert opinion suggest that CBT sessions should
be scheduled at least once weekly (63, 140).
One study suggests that five sessions per week of CBT consisting
of ERP may be more effective than once-weekly ERP sessions, but
are not necessarily more effective than twice-weekly ERP sessions
(141). The number of treatment sessions, their length, and the duration
of an adequate trial have not been established, but expert consensus
recommends 13–20 weekly sessions for most patients (140).
More severely ill patients may require longer treatment and/or
more frequent sessions. On the basis of clinical experience, clinicians
should also consider booster sessions for more severely ill patients,
patients who have relapsed in the past, and patients who show signs
of early relapse. Finally, because the majority of treatment trials
have been only 8–16 weeks long, the long-term persistence
of treatment effects and the utility of periodic "booster
sessions" require further study.
The psychiatrist may elect to conduct CBT or to refer the
patient for this or another adjunctive psychotherapy. Psychiatrists
wishing to utilize various forms of CBT are encouraged to pursue training
through workshops, conferences, and other training programs. In
addition, they can consult a number of treatment manuals (128, 142–146)
or other publications (33, 147–149)
or obtain consultation from a clinician specializing in the use
of CBT. The psychiatrist initiating CBT should explain to the patient
the nature of the treatment, including its here-and-now focus, the
rationale underlying treatment procedures, and what the patient
will be required to do. When resources for CBT are not available,
the psychiatrist can suggest and supervise the use of self-help treatment
guides (Appendix) and recommend support groups such as those accessible
through the Obsessive Compulsive Foundation (Appendix), although
these interventions have not been subjected to controlled study.
At the start of therapy, the psychiatrist can use the Y-BOCS
Symptom Checklist (10) to help the patient create a
list of target symptoms, including obsessions, compulsions, and
items or situations that are avoided because of OCD concerns. The
patient ranks the listed items from least to most anxiety-provoking.
In CBT consisting of ERP, patients are taught
to confront feared situations and objects (i.e., exposure) and to
refrain from performing rituals (i.e., response prevention). Exposures
may include in vivo confrontations (e.g., touching objects in public
bathrooms) and imaginal confrontations of feared consequences (e.g.,
imagining becoming "dirty" from "contamination").
Exposures that provoke moderate anxiety are prescribed first, followed
as quickly as tolerable by exposures of increasing difficulty. Moving
at too slow a pace can diminish faith in the treatment and motivation to
continue. Patients must face their fears for a prolonged period
without ritualizing, allowing the anxiety or discomfort to dissipate
on its own ("habituation"). The goal is to weaken
the connections between feared stimuli and distress and between
carrying out rituals and relief from distress.
As noted earlier, ERP can be combined with formal cognitive
techniques aimed at dysfunctional beliefs (122). Dysfunctional
beliefs in OCD include magical thinking (e.g., contamination by
proximity), an inflated sense of responsibility for unwanted events,
overestimations of the probability of feared events, the assumption
that thoughts are morally equivalent to actions or inevitably lead
to action ("thought-action fusion"), perfectionism,
the belief that anxiety/ discomfort will persist forever, and
the need for control. Whether ERP with cognitive therapy is superior
to ERP alone is currently under investigation. However, many experts
believe that integrating exposures (or behavioral experiments) with
discussions of dysfunctional beliefs and feared consequences is
likely to be the most effective treatment (120). Modifications
of ERP that include formal cognitive techniques as well as other
interventions have been suggested for OCD patients with certain
symptoms (e.g., hoarding ) and
those without overt rituals (e.g., see references 59, 149, 151).
5. Monitoring the Patient's
The frequency of follow-up visits after a new pharmacotherapy
is initiated may vary from a few days to 2 weeks. The indicated
frequency of visits will depend on the severity of the patient's symptoms,
the complexities introduced by co-occurring conditions, whether
suicidal ideation is present, and the likelihood of troubling side
effects. Patients should be seen as often as necessary for psychiatric
management. They should be encouraged to telephone between visits
if medication questions arise. If telephone calls become reassurance
rituals, the physician should work with the patient and the patient's
family to limit call frequency, using exposure (e.g., to the anxiety
that prompts the urge to call) and response prevention (e.g., limiting
calls), just as in treating any other ritual.
As noted earlier, symptom rating scales can sometimes be helpful
for monitoring the response of OCD, co-occurring depression, or
co-occurring anxiety disorders, or for keeping an objective record
over time in those patients whose symptoms do not respond to initial
treatments. Although use of scales is not expected in routine practice,
keeping an objective record over time for patients who do not respond
to initial treatments may be helpful.
6. Determining When and Whether to
The physician's goals are always to reduce suffering
and disability while minimizing the adverse effects of treatment.
First treatments rarely produce freedom from all OCD symptoms. In
clinical trials, OCD "responders" are variously
defined as those whose Y-BOCS scores decrease by at least 25%–35% from
baseline or who are rated much improved or very much improved on
the Clinical Global Impressions–Improvement scale (CGI-I)
(152). But even these degrees of improvement leave
much room for additional gains. Thus, the psychiatrist must decide
with the patient when, whether, and how to alter the treatment approach.
Table 4. Factors to Consider at Each Treatment Step
| Add to My POL
obsessive-compulsive disordertreatmentTable 4. Factors to Consider at Each Treatment Step
Patient's motivation and ability to comply
Nature and severity of OCD symptoms
Co-occurring psychiatric disorders and their treatment
Presence/absence of suicide risk
Co-occurring medical disorders and their treatment
Likely drug side effects
If patient is a woman of childbearing age or is elderly
Patient's past treatment history
Family or caregivers' involvement in symptoms
Patient's preferences regarding
treatment modality, acceptable risks, and expected benefits
Deciding whether to alter the treatment approach
will depend on the degree of suffering and disability the patient
wishes to accept. Because illness can bring secondary gains (familial
attention and caring and freedom from responsibilities), and because
depressed mood can diminish hopefulness, the psychiatrist may have
to address these issues when the patient is not well motivated to pursue
further treatments despite limited improvement in his or her OCD.
In the opinion of CBT experts, 13–20 sessions of weekly
outpatient CBT with daily homework or weekday daily CBT for 3 weeks
(about 50 hours, half therapist-guided, half homework) is an adequate
dose after which next steps can be considered (140).
With regard to SRIs, expert opinion supports changing medication
strategy (switching or augmenting) after a trial of 8–12
weeks, with at least 4–6 weeks at the highest comfortably
tolerated dose (140). However, some patients may respond
simply to a longer period of continued treatment with the same medication
(84). There is no apparent relationship between OCD
treatment outcome and plasma levels of SRIs (82, 85–87).
When the outcome of initial treatment has been unsatisfactory,
the psychiatrist should first consider the possible contribution
of several factors: problems in the therapeutic alliance; interference
by co-occurring conditions such as panic disorder, major depression,
alcohol or substance use disorders, or severe personality disorder;
inadequate patient adherence to treatment; the presence of psychosocial
stressors; the level of family members' accommodation to
the obsessive-compulsive symptoms (153); and an inability
to tolerate an adequate trial of psychotherapy or the maximum recommended drug
doses. The psychiatrist should next consider extending or intensifying
the psychotherapeutic or pharmacotherapeutic intervention. Figure
1 displays a treatment algorithm outlining potential next steps;
Table 4 lists factors to be considered at each treatment step.
Figure 1 Algorithm for the Treatment of Obsessive-Compulsive
response" means clinically significant but inadequate response. *Treatment
with little supporting evidence (e.g., one or few small trials or
case reports or uncontrolled case series). CBT = cognitive-behavioral
therapy; ERP = exposure and response prevention; MAOI = monoamine
oxidase inhibitor; SRI = serotonin reuptake inhibitor;
SSRI = selective serotonin reuptake inhibitor.
7. Pursuing Sequential Treatment
When the patient has an inadequate response to the initial
treatment and no interfering factor can be identified, the psychiatrist
and patient must decide on next treatment steps without the benefit of
data from controlled trials comparing all the possibilities. The
sequence of treatment trials shown in Figure 1 is based on expert
opinion (e.g., reference 140 and contributors to this guideline).
Given the absence of definitive trial data, augmentation strategies
might be preferred to switching strategies in patients who have
had a partial response to the initial treatment. Modest evidence supports
augmentation of SRIs with antipsychotic medications, including haloperidol
(154), risperidone (155–157),
quetiapine (158), or olanzapine (159).
These trials report response rates in the range of 40% to
55%. Patients who do not respond to one antipsychotic augmenting
agent may respond to another. A chart review study found that discontinuing
successful augmentation after 1–12 months resulted in relapse
for more than 80% (15/18) of patients, most within
2 months of discontinuation (160). Despite these promising
data regarding antipsychotic augmentation in OCD, many questions about
this treatment strategy remain unanswered, including the optimal
dose for each drug, long-term tolerability, when and how to discontinue
treatment, the drugs' relative augmentation efficacy, and
the reasons that only some patients benefit. Further data are also
needed on subsets of patients who may respond preferentially to
specific augmentation strategies. For example, one study (154)
suggests that augmentation with haloperidol helps only those patients
with co-occurring tic disorders.
Modest evidence supports augmentation of SRIs with CBT (specifically,
ERP) (161–163) and augmentation
of CBT with SRIs (164, 165). Some studies
have demonstrated no added benefit from combining SRIs and CBT (61, 123),
but these findings are limited by high refusal and dropout rates
and uncertainty about levels of treatment resistance (166, 167).
Some evidence suggests that adding cognitive therapy to ERP may
enhance the results, but this remains to be established. In the
absence of definitive data, combination treatment is provided in clinical
situations that include efforts to treat a co-occurring disorder
that is SRI responsive, to augment a partial response to monotherapy
(163), and to reduce the chance of relapse when medication
is discontinued (67).
For patients who do not respond to their first SRI, expert
opinion and clinical trial data support switching to a different
SRI (85, 87, 168–171).
However, the evidence does not allow one to predict the patient's
chance of response to the new SRI. Clinical experience suggests
that response rates to a second SRI trial are close to 50% but may
diminish as the number of failed adequate trials increases. A switch
to venlafaxine at doses ranging from 225 mg/day to 350
mg/day is supported by active comparator trials and open-label studies
suggesting its effectiveness in treating OCD (171–173).
A switch to mirtazapine is supported by one open pilot study and
a double-blind discontinuation trial (174).
If the strategies described above are not effective, augmentation
with other pharmacotherapies may also be considered. Expert opinion
(140, 175) and three open-label studies
(176–178) support clomipramine
augmentation of SSRIs. If clomipramine is added, plasma
levels of clomipramine and desmethylclomipramine should be assayed
2–3 weeks after reaching a dose of 50 mg/day,
and the total plasma concentration should be kept below 500 ng/mL
to avoid cardiac and central nervous system toxicity. Fluvoxamine
most increases plasma clomipramine levels (178), but
substantial increases may occur with fluoxetine and paroxetine.
A screening electrocardiogram may be advisable in patients suspected
of having heart disease or in patients over the age of 40. Pulse
rate and blood pressure should be monitored as the dose of clomipramine
Positive case reports exist for lithium augmentation, and
positive case series have been reported for buspirone augmentation.
However, small controlled but methodologically limited trials of
lithium and buspirone augmentation have been negative. Adding pindolol
2.5 mg three times daily was effective in one small, double-blind,
placebo-controlled trial (179) but not in another (180).
A small, 12-week, open-label study reported that augmentation of
SRIs with riluzole 50 mg two times daily was often helpful, but
methodological limitations prevent confidence that the benefit was
due to riluzole itself (501). Small controlled augmentation
trials with l-triiodothyronine and
desipramine have produced generally negative results, and a double-blind, placebo-controlled
trial found St. John's wort to be no better than placebo
Adding once-weekly oral morphine sulfate 30–45 mg
to various SSRIs with or without other augmenting agents was superior
to placebo in a double-blind crossover study (182).
However, morphine sulfate should be avoided in patients with contraindications
to opiate administration, including a history of substance or prescription
medication abuse, psychosis, mania, antisocial personality disorder,
chronic obstructive pulmonary disease, or cardiovascular compromise.
In addition, the psychiatrist should consider what precautions and
documentation may be neededfor example, those described
by the American Academy of Pain Medicine (www.painmed.org). In addition,
positive case reports exist, along with a positive case series,
for monotherapy with the weak narcotic agonist tramadol (183).
And two small, double-blind, placebo-controlled, single-dose studies
reported positive results for d-amphetamine
30 mg in unmedicated OCD subjects (184, 185).
However, these drugs should be avoided in some patients (e.g., those
with a history of alcohol or other substance abuse or dependence).
Other treatment strategies that are supported only by case
series, case reports, or small open trialsliteratures
that are less likely to include negative experiencesinclude
anticonvulsants, MAOIs, ondansetron, l-tryptophan,
nicotine delivered via transdermal patch or chewing gum (186),
and Kundalini yoga. For patients with severe treatment-resistant
OCD, partial hospitalization (49, 187)
and intensive residential treatment (48, 188,189)
have been used.
Other somatic therapies should be considered only after first-
and second-line treatments and well-supported augmentation strategies
have been exhausted. With treatments such as these for which efficacy
is uncertain, it is especially important to weigh the potential
benefits against the side effects and other risks of therapy. For
example, evidence for the use of electroconvulsive therapy (ECT)
in OCD is limited to a single case series using a nonstandard form
of ECT administration (190). In addition, ECT carries
the risks of anesthesia and has side effects such as memory impairment.
As a result, ECT cannot be recommended for the treatment of OCD
but may be considered for treating co-occurring conditions for which
it may be indicated (e.g., major depression, uncontrollable mania,
and schizophrenia) (191–194).
Transcranial magnetic stimulation (TMS) is associated with less
potential for side effects, but evidence for its efficacy is limited.
Deep brain stimulation (DBS), a reversible and adjustable neurosurgical
intervention, has been reported to show efficacy in a few case series
of individuals with severe, highly treatment-resistant OCD (195)
but also has potential side effects.
The efficacy of ablative neurosurgery (anterior capsulotomy,
limbic leucotomy, cingulotomy, and gamma-knife radiosurgery) in
patients with severe, treatment-refractory, or intractable OCD has been
evaluated in case reports and unblinded studies. Improvement rates
have ranged from 35% to 50% (196–198).
Although some studies report relatively high rates of improvement,
the unblinded nature of these studies and the ongoing treatment
of many patients limit interpretation of these results. In addition,
potential adverse events range from personality changes, seizures,
and hydrocephalus to transient mania and mild transient side effects
such as urinary dysfunction. The recent development of less invasive
(DBS) and non-invasive (TMS) procedures makes it harder to consider
ablative neurosurgery as an alternative for highly treatment-resistant
or intractable OCD. For the time being, DBS and ablative neurosurgical
treatment for OCD should be performed only at sites with expertise
in both OCD and these treatment approaches.