Chemical abuse and dependency aie common in patients presenting with complaints of anxiety. According to the National Comorbidity Survey, among patients who meet DSM-UI-R criteria for any anxiety disorder, approximately 15% also have a substance use disorder during the previous 12 months.1 The use of pharmacotherapy in such patients presents some special problems that have not been addressed systematically in previously published algorithms for the treatment of anxiety disorders.2"4 The authors decided to pool their many years of clinical experience working with dualdiagnosis populations, survey the literature and expert opinion, and prepare a series of algorithms for these patients. These algorithms were presented at the American Psychiatric Association 1998 Annual Meeting in Toronto, Ontario, Canada. After further review and improvement, they will be converted to an Internet "virtual psychopharmacology consultant" format and will become part of the portfolio of applications in the Harvard Psychopharmacology Algorithm Project described in this issue of Psychiatric Annals.
In crafting this algorithm, the goal was to propose a plausible, supportable, possibly cost-effective sequence that could be a "default" method for the optimum psychopharmacologic care for an average case of these often complicated and treatment-resistant patients. We tried to address some of the nuances that may affect the prescribing decision of the psychiatrist, nuances that may differ depending on the setting in which the patient is being seen (eg, outpatient or institutional care).
There is a relative paucity of research directly relevant to this effort. Much of this research was summarized in the March 1998 issue of Psychiatric Annals, which was devoted to treating the anxious person with an addiction. Those articles strongly advocated minimizing the use of benzodiazepines (BZs) in this patient population,5 but the editorial questioned whether passion on this issue had obscured the quest for balance.6 The controversy over the role of BZs was a central issue that had to be addressed in the development of the present algorithms, and a balanced set of recommendations was considered essential if the final product was to be useful to physicians facing real-world exigencies.
We want to stress that in proposing this educated guess of what constitutes reasonable practice, it was not our intention to discourage innovation. Creative new thinking often leads to significant advances. Innovators need only explain why the new approach might be better, and then try it. Ultimately, those innovations that pass the test of time and replication can replace these initially proposed benchmarks. A mechanism for collection of patient outcome data over the Internet was described in the introduction to this issue.
INITIAL SCREENING FOR PATIENT APPROPRIATENESS FOR THIS ALGORITHM
The entry into the algorithm begins with having a patient with anxiety symptoms who meets DSM-IV7 criteria for substance or alcohol abuse or dependency. The start of the flowchart (Fig. 1) requires an evaluation of whether the patient is currently using drugs or alcohol. If so, the patient should have completed withdrawal from his or her drug of dependence and from any pharmacotherapy used for withdrawal, and be abstinent for at least one additional week. This appears to be the minimum time after which drug-placebo differences have been demonstrated in the treatment of anxiety and depressive disorders that persist following alcohol withdrawal.8,9 Symptoms occurring after abstinence of less than a week may be more likely to be attributed to the residual effects of the substance. If symptoms appear to be diminishing during the first week of sobriety, and there is no history of these symptoms prior to onset of the substance abuse or dependency or during previous periods of extended sobriety, it is reasonable to wait at least another week before initiating pharmacotherapy.
Withdrawal from some other substances, such as methadone hydrochloride, can be prolonged and the residual effects of their presence can affect subsequent medications that may be given. For example, methadone hydrochloride has a half-life of approximately 2 days and its effects as an inhibitor of cytochrome P-450 2D6 could persist for more than a week.
This algorithm does not address the pharmacotherapy of anxiety in patients who are currently intoxicated with alcohol or substances. There are occasions when it may be appropriate and reasonably safe to use medication in such patients. Another group not specifically addressed is individuals who use prescription or illicit drugs "recreationally" but who do not meet DSM-TV criteria for abuse or dependence: should the sequence of antianxiety treatments be any different for them? These areas are under-researched and deserve future algorithm development efforts.
After the patient achieves abstinence, the physician makes a thorough review of the differential diagnosis of the anxiety symptoms, ruling out or treating any medical causes, including the ongoing use of anxiety-producing substances such as caffeine, and the ongoing presence of other addictions such as nicotine. This process should be repeated, throughout the algorithm, after each sequential treatment trial if the response is still unsatisfactory. At each decision point, the physician should also always consider the possibility of an undetected substance abuse relapse. The hidden relapse could be manifested as an aggravation of anxiety symptoms.
If the patient was alcohol dependent, and his or her anxiety is accompanied by and appears to be due to preoccupation with or craving for alcohol, consideration should be given for a 1- to 2week trial of naltrexone hydrochloride, as noted in Figure 1. If the patient experiences immediate relief of urges to drink, this may obviate the need for further pharmacotherapy for the anxiety.10 The best outcomes with naltrexone hydrochloride seem to occur in patients who experience rapid elimination of craving11 and who receive intensive psychosocial treatment that helps maintain compliance with the medication.12 If the patient was also opiate dependent, there should be greater caution in using naltrexone hydrochloride because of the risk of opiate relapse and subsequent opiate overdose. Death has occurred as a complication of these events, according to the product monograph. However, if there is a pattern of alcohol relapse occurring first, followed by opiate use, and, in addition, if this is a patient with severe, treatment-refractory alcohol addiction, then with appropriate informed consent and medical review (eg, of the status of the patient's liver) the patient may still be considered for naltrexone hydrochloride.13
If the patient's anxiety does not appear to be related to craving, or if the latter has been treated by naltrexone hydrochloride and the anxiety persists, a more specific diagnostic evaluation of the anxiety disorder is now conducted. Separate algorithms will be presented for panic disorder, social anxiety disorder, posttraumatic stress disorder (PTSD), obsessive-compulsive disorder, and generalized anxiety disorder. All should be diagnosed by DSM-IV criteria for maximum validity of the recommendations that will follow.
Many patients may meet the criteria for more than one anxiety disorder,14 or the diagnosis may be uncertain. In that case, the physician should follow the algorithm for each one to the possible diagnoses, and compare the recommendations. There may be overlapping recommendations that cover the diagnostic ambiguity.
WHICH ANXIETY DISORDER IS PRESENT?
The physician should be sure that this condition is being diagnosed accurately, as errors appear to be frequent. In particular, note that in DSM-IV panic disorder the panic attacks are "unexpected" (ie, not obviously precipitated by any known stress or trigger). The panic attacks of many patients occur under specific circumstances, such as in a social situation or at night in a dark room when reexperiendng of a previous trauma may occur. The treatment sequence for panic disorder may not be appropriate for these patients. A second commonly overlooked diagnostic requirement is that the anxiety attacks in panic disorder come on suddenly and reach their peak within 10 minutes. Actually, 10 minutes is an outer limit: most often the attacks become maximum in seconds to minutes. Patients with other anxiety disorders have symptoms that wax and wane irregularly during the course of hours. If this was the pattern from the onset, it is unlikely to be panic disorder. Note, however, that many patients with panic disorder develop adaptations to the attacks (with or without the help oí formal psychotherapy), and, in time, they become able to blunt the intensity, duration, and rapid progression of symptoms.
The recommendation for first-line treatment of panic disorder in patients with a history of substance abuse who have not previously received pharmacotherapy for their anxiety disorder is a combination of a selective serotonin reuptake inhibitor (SSRI) or nefazodone hydrochloride, plus psychotherapy.15 The most research support is for cognitive-behavioral psychotherapy, but psychodynamic therapy could produce similar results16 (controlled studies are in progress). Figure 2 is the flowchart for the treatment of panic disorder.
Comments on the Use of BZs. BZs are also a first-line treatment for panic disorder, and they have numerous advantages, including rapid action, modest side effects, and long-term efficacy.17 There are some risks of abuse and dependency developing in patients with a history of chemical dependence, and this has led many experts to take an "avoidant" stance with respect to their use in this population. For example, the "Practice Guidelines for the Treatment of Patients with Panic Disorder" of the American Psychiatric Association (1998) state that ". . . benzodiazepines should be avoided whenever possible. . ." in these patients.18 The issue is made more problematic by the rigid positions that Alcoholics Anonymous and Narcotics Anonymous have taken against BZ use. However, there are important research data contradicting this position.19 Epidemiologic studies, for example, show that substance abusers who use BZs are more likely to employ them to prevent or treat withdrawal rather than as primary recreational agents. These conclusions are disputed by others,20 but it is clear that much more research is needed before it may be conclusively stated that prescription of BZs significantly prolongs the course or increases the morbidity associated with substance abuse disorders.
We propose the following criteria for consideration of a BZ in the initial treatment of substanceabusing or dependent patients with panic disorder. The same considerations could lead to selection of a BZ at a subsequent point in the algorithm it the standard recommendation appears unlikely to be acceptable to the patient or the clinician. These principles should apply, as well, to BZ use in other anxiety disorders (to be described) in which BZs have been found useful.
1. If the patient has a history of response to a BZ without significant abuse or misuse.
2. If there is a family history of panic disorder responsive to a BZ without evidence of abuse or misuse.
3. If the patient (especially an outpatient with few social supports) seems highly likely to experience a relapse of substance abuse or dependence if required to wait for a slower-acting antianxiety treatment to take effect, the physician could consider offering a single prescription of a BZ for a brief period (eg, a 1-week supply) to test the patient's ability to comply with appropriate use.
4. High-risk suicidal patients, especially inpatients but even some outpatients, who have panic attacks comorbid with severe dysphoria with agitation may be saved from a disastrous outcome by the short-term use of a rapidly acting BZ to relieve their distress.6
5. If the patient is a high-functioning, wellmotivated individual who has had a sustained period of abstinence, or a relatively mild level of dependency.
In all these cases, given the risk management issues occasioned by the litigious and politicized atmosphere that surrounds this issue, the clinician should discuss with the patient and document in the record the risks and benefits and the structure of the plan for the proposed BZ use. It may be helpful to obtain the patient's permission to discuss the matter with the spouse or other significant stakeholders in the patient's well-being.21
Alprazolam and diazepam seem to be the most abuse-prone BZs, in alcoholics,22 although the considerable street value of clonazepam and lorazepam corifirms that addicts find them useful as well.5 It may be argued that these should be reserved for patients who have failed to respond to first-line trials with others such as oxazepam or chlordiazepoxide hydrochloride. However, sometimes the patient's past experience is that the higher abuse-prone agents work better for the anxiety, and the physician may calculate that a brief trial may be needed to avert relapse for that patient. Or, the physician may "inherit" a patient who was started on one of these agents by another prescriben The patient may strenuously resist any suggestion to switch to another medication. The physician may (perhaps supported by a consultation with a colleague) elect to continue it if the benefits appear to justify the ongoing risks.
The algorithm's recommendations for non-BZ treatments are likely to work best in a structured setting where patients can be assisted with medication compliance and where the benefits of medication can be maximized by concomitant group and individual psychosocial treatment focused on relapse prevention. In the unstructured outpatient setting, there is much less chance that the physician will be able to methodically go through a sequence of slow-acting treatments before the patient has a relapse. BZs may sometimes be necessary to engage the patient in treatment. Whenever possible, however, such unreliable patients should be treated by more than one clinician, or with the help of other responsible adults in the patient's life.
Back to the Algorithm. Returning to a discussion of the first-line recommendation of SSRIs or nefazodone hydrochloride, it may be noted that all of the SSRIs have been found efficacious in panic disorder, but fluoxetine hydrochloride23 may be more likely to be activating (producing "jitteriness," agitation, or anxiety) than fluvoxamine maléate, nefazodone hydrochloride, paroxetine hydrochloride, and sertraline hydrochloride (if the physician begins with 25 mg/d24). Citalopram hydrobromide is probably not activating, despite one report.23 The sedating effects of nefazodone hydrochloride may be particularly desirable in the outpatient setting where the immediate impact of the sedation may help motivate the patient to complete the therapeutic trial. If fluoxetine hydrochloride is used, it may have to be started at doses of 2.5 to 5 mg and gradually increased.25 Somewhat less established options that may be considered are venlaiaxine hydrochloride26 and (anecdotally) mirtazapine. The sedating effects of the mirtazapine might be particularly desirable, although the weight gain would not be. Venlaiaxine hydrochloride is associated with the slightly inconvenient requirement to check baseline and subsequent blood pressures, although at the doses reported (50 to 75 mg/d) blood pressure should rarely be a problem.27
The anti-panic effect of these medications usually takes up to several weeks to become significant, and during the initial period of treatment the physician is faced with the dilemma of what to prescribe that can be of temporary benefit. Trazodone hydrochloride, which may have antianxiety effects equivalent to those of BZs and even some anti-panic effects,28 may be tried as an adjunctive sedating agent, on an as-needed basis, 25 to 50 mg up to several times a day.
Adjunctive BZs are routinely given for temporary relief when non-substance-abusing patients are treated with antidepressants. The main problem with this use of BZs (in all patients) is that it is difficult to get patients off them even when the intention is to use them for only a few weeks.29 Nevertheless, the considerations listed earlier may mitigate in favor of adjunctive BZ use in some patients with substance abuse or dependence.
Psychotherapy for the patient with panic disorder who has a history of substance abuse should focus on the stress-related and other environmental factors fostering substance relapse, and on the development of new coping skills for the anxiety. These skills may supplement the effects of medication and, ultimately, facilitate withdrawal from medication and maintenance in a medication-free state, if this becomes desirable in the future.
The recommendation for second-line pharmacotherapy treatment, if the first-line medication produces an unsatisfactory response after 2 to 6 weeks, or the medication is not tolerated, is to try a second SSRI (or nefazodone hydrochloride if that was not used at first). Again, venlaiaxine hydrochloride and mirtazapine would be reasonable choices. Tricyclic antidepressants may be considered here as well, although their side effect burden is somewhat more problematic compared with the others. They should be avoided if the patient has a high suicide or seizure risk. Again, BZs may be considered for supplemental or monotherapy use at this second-line level, if they were not used before, with the same considerations as noted above being applied to justify this use. In general, as more non-BZ treatments have been sequentially tried without success, and the likelihood of success with them therefore becomes more questionable, the potential benefits to risk ratio of trying a BZ rises and it becomes slightly more reasonable to consider one.
The third-line option is a tricyclic, if the patient is not a high suicide or seizure risk. The fourth is valproic acid,30 if the patient does not have a persistent liver problem such as hepatitis B or C. At the third- and fourth-line levels, the option of adding the additional treatments rather than employing sequential monotherapies may be considered, in the event of a partial response to the previous regimen.17 In patients with comorbid chemical dependency it seems reasonable to be more cautious about employing combination treatments due to the risk of toxicity or misuse of the medication if the patient resumes substance abuse. More benign augmentation options might be buspirone hydrochloride and (based on very preliminary evidence31) gabapentin.
Monoamine oxidase inhibitors (MAOIs) are highly effective treatments for panic disorder, but in the patient with substance abuse or dependence their use would involve major safety risks. They could be considered as a last resort for a very reliable patient.
Social Anxiety Disorder (Social Phobia)
Social phobia can be a severe, disabling anxiety disorder associated with marked reduction in quality of life. It is the third most common psychiatric disorder in the United States, with a 13% lifetime prevalence rate, according to the National Comorbidity Survey.32 Major depression was first with 17% and alcohol dependency second at 14%. Despite its frequency, social phobia is still under-recognized and untreated, especially in primary care. Furthermore, it is typically stigmatized and trivialized. The patients may consider themselves to be "shy" and believe this is normal.33
Among the anxiety disorders, social phobia is the only one where there is evidence of higher comorbid alcohol abuse or dependency compared with the general population (20% to 30%).34 This evidence deserves further replication, but it suggests that self-medication may be more of a factor in alcohol use by patients with social phobia compared with other anxiety disorders discussed in this article.
In the "specific type" of social phobia, dysfunction is relatively limited. There may be anxiety in particular situations, such as in a group or while urinating in a public bathroom. In Figure 3, the first-line treatment recommendations for specific social phobia are listed as cognitive-behavioral therapy or beta-blockers taken 1 to 3 hours before major performances.35 The usual dose of propranolol hydrochloride is 10 to 40 mg. A test dose at some other prior time is advisable to determine how the medication will be tolerated.
The "generalized type" of social anxiety is a more pervasive problem, similar (and sometimes identical) to avoidant personality disorder in DSM-TV.33 Its onset is in childhood or adolescence. The patient blushes or trembles in many performance situations, develops anxiety that can rise to panic, levels, and experiences negative cognitions that others are thinking critical thoughts about him. or her. In the more severely disabled patients, there is often comorbid chemical dependency, depression, and other anxiety disorders.
If there is comorbid depression (Fig. 3), the firstline recommendation is for a combination of an SSRI or nefazodone hydrochloride and cognitive-behavioral therapy. These patients are more resistant to treatment and will probably require the potential additive effects of the two kinds of therapy. Some of these desperately uncomfortable, precariously stabilized, and /or suicidal individuals may require consideration of a BZ, as described in the BZ criteria of panic disorder.
All SSRIs and nefazodone hydrochloride have evidence of effectiveness, even the newest entry, Citalopram.36'38 Unlike the situation with panic disorder, there is no evidence of initial agitation with antidepressants, so it is usually not necessary to start an adjunctive sedative.39 However, benefits take 2 to 4 weeks to begin, and the full effect may not occur for months as patients test their ability to handle different social demands. Patients with social anxiety disorder may be particularly vulnerable to sexual side effects, as they seem to have a high rate of preexisting dysfunction in this area.40 This creates a possible advantage for nefazodone hydrochloride. Also, the more immediate sedative properties of nefazodone hydrochloride can be another advantage, as noted in our discussion of panic disorder.
Support exists at the case report level for bupropion hydrochloride in social anxiety disorder.15 We are awaiting information about mirtazapine, whose sedative properties could be an advantage.
Tricyclics and beta-blockers have not been effective.15 Nonselective MAOIs are very effective, but generally have an unacceptable safety profile in the substance-dependent population. The MAOIs selective for MAO-A such as moclobemide (not available in the United States) are safer, but appear to be only weakly effective in social phobia, according to the more recent studies.41
If the social anxiety improves but the depression persists, or if neither improves, the depression should be the target for further pharmacotherapy. The physician should consider obtaining a "virtual consultation" from our Algorithm for the Pharmacotherapy of Depression on the Internet at www.mhc.com/expert.html and published elsewhere.42
If there is no comorbid depression (Fig. 3), the firstline recommendation is for a trial of cognitive-behavioral therapy, which is well established as effective.43 Many patients will prefer this, but others may elect to begin with pharmacotherapy, which may facilitate their ability to use group treatment in particular. If or when medication is started, SSRIs or nefazodone hydrochloride are again recommended. Buspirone hydrochloride could be considered, but the most recent placebo-controlled study showed no efficacy.44 Clonazepam is one BZ that has a strong evidence basis for efficacy in social phobia in non-substance-abusing patients.45 The controversies that have surrounded its use in patients with social phobia and substance abuse46 may, again, be addressed by considering the criteria for BZ use proposed in the section on panic disorder.
Should the patient fail to respond to this SSRI or tolerate it poorly, the second-line recommendation is to try a different one. The third-line recommendation could be another monotherapy, or an augmentation. The addition of buspirone hydrochloride to an SSRI was reported to result in a 70% response rate, at 45 mg/d. Gabapentin (600 to 3600 mg/d)47 and venlafaxine hydrochloride (50 to 300 mg/d)15 have studies to support them as alternatives for monotherapy. MAOIs could possibly be considered, for ideal candidates. The addition of cognitive-behavioral therapy, if not already ongoing by this time, could augment the improvement from medication and offer the chance of sustained improvement if medication is discontinued at some point in the future.48
Posttraumatic Stress Disorder
Figure 4 displays the flowchart of the algorithm for PTSD and chemical dependency. The first recommendation is for the provision of some psychotherapeutic intervention. The nature of the therapy depends on the setting and the needs of the patient, In short- to intermediate-term structured residential treatment where support is available to ensure abstinence, trauma-focused therapy techniques may be appropriate and helpful. Eye movement desensitization and reprocessing therapy may be applicable, as well as other specific behavioral techniques of "grounding" during dissociative and flashback experiences.49 In the outpatient setting where there are fewer supports, lack of control over triggers of traumatic memories, and perhaps even ongoing perpetration of further trauma, the psychotherapy probably should be more purely supportive and focused on substance relapse prevention, marital or family counseling, and practical assistance with life management.
The first-line choice of medication depends on whether there is comorbid DSM-G?* major depression or dysthymia. If the patient is not clinically depressed, we would focus first on whether there is any sleep disturbance, which some consider to be a hallmark of the psychopathology of FTSD.50 If the patient has significant distressing insomnia, successful treatment of this might have a considerable restorative effect on the patient's daytime functioning and level of distress. Relief of chronic sleep deprivation may obviate the need for an antidepressant or other pharmacotherapy targeting PTSD symptoms. Medical causes of the insomnia, such as obstructive sleep apnea or restless leg syndrome, should first be diagnosed and treated.51
The first-line recommendation for a hypnotic would be for a trial of trazodone hydrochloride, 25 to 300 mg at bedtime as required. This generally benign and inexpensive agent has come into widespread use, in low doses, for insomnia from a variety of causes; however, there have been no controlled studies and the indication is not approved by the Food and Drug Administration.52,53 In FTSD, it can improve sleep and reduce traumatic nightmares, possibly through its suppressant action on rapid eye movement.54,55 We generally start with 50 mg and give instructions that if the patient experiences significant "hangover" in the morning, the dose should be lowered to 25 mg the next night. If 50 mg is ineffective after an hour, or early awakening occurs, a second 50-mg tablet may be taken. In some cases, a patient may require 150 to 300 mg. A second-choice recommendation for a hypnotic, if the first-line recommendation fails, would be a low dose (25 to 75 mg) of a sedating, tertiary amine tricyclic antidepressant such as doxepin hydrochloride or amitriptyline hydrochloride, if the patient is not at high risk for suicide or seizures and the atropinic and cardiac risks are acceptable.52 Another option used by some clinicians is a sedating antihistamine such as diphenhydramine hydrochloride, but tolerance tends to develop quickly to the sedating effect. The serotoninblocking agent cyproheptadine hydrochloride seems to be locally popular, but a recent report found it ineffective for nightmares in PTSD patients.56
A BZ hypnotic might be considered under the circumstances described in the algorithm for panic disorder: high-functioning, cooperative patients with relatively mild substance dependency, and perhaps some marginally compensated patients in the private practice setting at very high risk for relapse if immediate, reliable relief of their insomnia is not provided. All are to be given small amounts at first. Keep in mind, however, that there is evidence that PTSD patients (regardless of whether they are dual diagnosed with chemical dependency) are at very high risk for abusing prescribed medication.57 Also, BZs seem to have little effect on the core symptoms of PTSD.54
If the patient does not have prominent insomnia, or if use of the recommended hypnotics does not result in improvement of the patient's daytime PTSD symptoms such as hyperarousal, reexperiencing, or "numbing" (avoidant behavior), the first-line recommendation for PTSD pharmacotherapy will be an SSRI or nefazodone hydrochloride.58,59 The evidence for nefazodone hydrochloride is weaker,60,61 but there were three new studies presented in posters at recent meetings, so it looks promising. These antidepressants are also the first-line recommendation if the patient has comorbid depression. It would appear unlikely that a hypnotic alone would be sufficient in these patients, hence the algorithm branches away from the hypnotics questions if this comorbidity is present. There is some evidence that SSRIs will also be of benefit to reduce alcohol craving and consumption.62 Benefits in patients with PTSD, if they occur, are usually seen in 4 to 8 weeks.62a Hypnotics (eg, trazodone hydrochloride) may be used for SSRI-induced insomnia or nightmares.63
If the patient has not responded satisfactorily to an antidepressant, we ask whether he or she has a PTSD-related psychosis characterized by trauma-based hallucinations. Although there has been little research concerning this problem, it appears to be fairly common in patients with severe PTSD.64*65 One clinical report found antipsychotics helpful,66 and this supports what appears to be a common practice.54 The availability of the three newer-generation antipsychotics with apparently reduced risk of tardive dyskinesia makes it reasonable to consider using them in PTSD patients earlier than one would have used neuroleptics in the past. If the psychotic symptoms do not respond to the trial with an antidepressant, as above, it seems reasonable to try to add an antipsychotic.
If the patient does not have psychotic symptoms, or if he or she has been treated and PTSD symptoms persist, the second-line recommendation is to try a second SSRI (or nefazodone hydrochloride, if this was not tried before). The third-line recommendation would be to try one more antidepressant, this time considering a tertiary amine tricyclic (because there is evidence that they have a larger effect size than secondary amines such as desiprarnine hydrochloride67). The patient should not be at high risk for suicide or seizures. An alternative might be one of the newer generation of antidepressants that may be more tricyclic-like and sometimes more powerful than SSRIs: venlaiaxine hydrochloride or mirtazapine.68"71 Three of six treatment-resistant PTSD patients responded to rnirtazapine in one report.72 Bupropion hydrochloride was not effective for PTSD in a recent report, although it was helpful for comorbid depression.73 MAOIs, although supported by some studies, would usually be too risky in this population.
Fourth-line recommendations for this now clearly treatment-resistant case of PTSD might follow a reassessment of the patient's PTSD symptoms. If hyperarousal symptoms are prominent, an antiadrenergic agent such as Clonidine hydrochloride (or the long-acting guanfacine hydrochloride) might be particularly effective. Although the literature is small, some experts are enthusiastic.54 If impulsivity and aggression are prominent, a mood stabilizer such as lithium or the anticonvulsants valproate sodium74 and carbamazepine should be considered, if persistent liver disease is not a contraindication.
Beyond this, combinations of some of the above may be considered, individualized to the patient's particular symptom requirements. Again, however, our general preference in the substance-abusing patient population is to keep medication regimens as uncomplicated as possible through at least the first several "lines" of recommendations.
Obsessive-compulsive disorder (OCD) is still sometimes confused with obsessive-compulsive personality disorder (OCPD), a member of cluster C in DSM-IV. The distinction is important, as OCPD is not associated with response to any particular pharmacotherapy. Individuals with OCPD have a preoccupation with orderliness, rigid rules, inflexible ethics, and stinginess with money. Unlike patients with OCD, they are not tormented by unwanted obsessions; instead, they cause distress for others by insisting that everyone think the way they do.
After reviewing a variety of algorithms, reviews, and guidelines regarding the treatment of OCD, it was concluded that differences in the pharmacotherapy approach to dual diagnosed compared with uncomplicated OCD seem relatively minor.2,4,75-77
Most experts call for a series of two or three monotherapy trials with SSRIs, moving to high doses if moderate doses do not produce the desired response in the first 4 to 10 weeks. All SSRIs appear effective, including the newest, Citalopram.78 Clomipramine hydrochloride is generally the next recommendation, based on meta-analysis suggesting that it is more effective than the SSRIs.79 In the dual-diagnosis patient there would be the usual questions regarding suicide, and especially seizure risk, before considering clomipramine hydrochloride.
After these monotherapies, a variety of augmentation strategies are generally proposed.77,80 For our patients, buspirone hydrochloride would be the most benign, but it also has minimal evidence of effectiveness. Antipsychotics may be helpful, especially if there are tics, schizotypal personality, or obsessions that are fixed at a delusional level. Clomipramine hydrochloride is a potentially more powerful augmentation for an SSRI, but predictable cytochrome P450 drug interactions must be addressed80 and, again, this must be avoided if suicide or seizures are at issue. Another frequent augmentation recommendation in the literature, although it is weakly based on evidence, is for a BZ such as clonazepam. BZ use in the OCD patient with substance abuse may be considered, with the caveats described elsewhere in this article.
The role of adjunctive cognitive-behavior therapy has been difficult to establish in the studies conducted to date, but most experts consider it an important potential source of augmented benefit to all patients.
Generalized Anxiety Disorder
No diagnostic category has changed as much as generalized anxiety disorder (GAD) in the past 20 years. The present criteria reflect, in part, the results of psychopharmacologic studies designed to improve the specificity of treatments for this disorder. The current DSM-IV criteria describe patients with a core problem of chronic excessive worrying, focused in a number of areas, that is difficult to control and causes impairment. If the worry is confined to the domains of the other anxiety disorders (eg, social anxiety, fear of panic attacks, unwanted obsessions, or recurrent traumatic experiences), or other Axis I disorders, then the other disorder will be the primary diagnosis. GAD is no longer predominantly a disorder of autonomic, motor, or other somatic manifestations of anxiety, as it was in DSM-UI-R.81 Many former GAD patients are now classified as having somatoform disorders in DSM-IV. Supporting these changes is a study by Rickels et al. that showed that the psychic (worry) symptoms of the old DSM-EQ-R criteria responded well to an antidepressant (Imipramine hydrochloride 145 mg), but not did not respond to a BZ (diazepam 26 mg).82 By contrast, the diazepam was better than the Imipramine hydrochloride for the somatic and arousal symptoms of the old criteria, at least during the first 2 weeks. By weeks 3 to 8, the antidepressant was just as effective as diazepam, and trazodone hydrochloride 225 mg (which was also an arm in the study) was almost as effective. If this study can predict future experience, DSM-IV GAD should be more responsive to antidepressants than to BZs.
To our knowledge, no other direct comparisons of antidepressants and BZs have since been completed. There are two large completed studies of venlafaxine hydrochloride in DSM-TV GAD, one a comparison with placebo and the other a comparison with buspirone hydrochloride and placebo. Venlafaxine hydrochloride was better than placebo in both studies and better than buspirone hydrochloride on some measures.83,84 SSRI antidepressants deserve further investigation in DSM-IV GAD: one open study showed that paroxetine hydrochloride produced significant improvement in the character-temperamental traits of harm avoidance and self-directedness of GAD patients.85 Nefazodone hydrochloride and mirtazapine might also be helpful, given their sedative and anxiolytic effects in depressed patients.
Seventy-five percent to 90% of patients with DSM-HI-R GAD have current or lifetime comorbid Axis I disorders.86,87 At the entry point to this algorithm (Fig. 5), we ask whether the patient meets criteria for one of the other anxiety disorders for which we have algorithms. If so, the algorithm for that disorder or disorders should be explored first, as the treatment options are more delineated and successful treatment of the comorbid anxiety disorder may result in improvement in the GAD.
If there are no comorbid anxiety disorders, the next question is whether there is comorbid major depression or dysthymia. If not, our first-line recommendation for treatment is a trial of psychotherapy. This may be introduced now or at any of the subsequent pharmacotherapy steps. Options include cognitive-behavioral, relaxation, or brief psychodynamic approaches combined with a chemical dependence relapse-prevention paradigm.88 Cognitive therapy would seem particularly applicable given the new definition of GAD as a disorder of preoccupation with exaggerated perception of danger. If psychotherapy has been tried and despite several weeks of effort the results are not promising, or if the patient has a preference for early pharmacotherapy, then the first-line recommendation for pharmacotherapy would be buspirone hydrochloride. This suggestion is based on positive evidence of efficacy in placebo-controlled studies in recently abstinent alcoholic patients, who had DSM-III-R GAD.8,89 The average dose in these patients was 50 mg/d. Also, it is based on extrapolation from evidence of buspirone hydrochloride's possible efficacy in depression90: the assumption is being made that the psychic, or worrying, component of GAD may be related to a mood disturbance. The relatively mild side effect profile of buspirone hydrochloride is also a consideration, including its lack of impairment of automobile driving performance.91 It even partly blunts the adverse effect of alcohol on driving ability.92
Studies using DSM-III in the 1980s showed that patients, with heterogeneous diagnoses, who are given buspirone hydrochloride after previous use of BZs do not respond well to it. This is an important issue, as many patients with chemical dependence have used BZs. A more recent study involving GAD patients only, meeting DSM-III-R criteria, showed that the patients did well with buspirone hydrochloride (and better than placebo) after previous lorazepam use for 8 to 14 weeks.93 The lorazepam was tapered slowly during 2 weeks. However, another study, a metaanalysis of 9 studies, found that buspirone hydrochloride worked best in GAD patients who had terminated their previous BZ use more than 4 weeks before starting the drug.94 These studies provide reasonable support for our proposal of buspirone hydrochloride for first-line use in the patient population under discussion. BZs, even if perhaps not the ideal drugs for GAD based on DSM-IV criteria, could be justified as an alternative first-line or supplemental treatment because of the major advantage of their rapid action. BZ caveats mentioned in the discussion of panic disorder earlier in this article would apply.
If this buspirone hydrochloride trial does not produce a satisfactory result after 2 to 4 weeks, the second-line pharmacotherapy recommendation is for an antidepressant: an SSRI or nefazodone hydrochloride. As indicated, the evidence basis for this derives from an educated guess only, but these have a slight safety advantage over venlaiaxine hydrochloride (recently approved by the Food and Drug Administration for GAD) or mirtazapine. The third-line recommendation would be to try another SSRI (or nefazodone hydrochloride if that was not tried before), venlafaxine hydrochloride, or mirtazapine. The fourth-line choice might be yet another antidepressant. A BZ might be reconsidered at any point if the risks can be justified for a particular patient.
What if the patient with GAD does have comorbid depression? Technically, the physician would not diagnose GAD in that case because GAD is hierarchical to mood disorders in the DSM-W (ie, the physician would not make a separate diagnosis of GAD). The first-line recommendation then is for an antidepressant: an SSRI or nefazodone hydrochloride. Further treatment might follow the sequence of steps in our Algorithm for the Pharmacotherapy of Depression.42
ALGORITHMS FOR THE PHARMACOTHERAPY OP AHXIETY DISORDERS IH PATIEHTS WITH CHEMICAL ARUSE AHD DEPEHDEHCE: OVERVIEW AHD CONCLUSIONS
Although the evidence basis is limited, the treatment strategies for anxiety disorders in dualdiagnosed patients seem to have a lot in common. Psychotherapeutic interventions and the newer generation of antidepressants have significant roles in most cases. BZs and tricyclics have a more restricted role than they do in the general population. Specific situations in which to consider buspirone hydrochloride, naltrexone hydrochloride, trazodone hydrochloride, valproate sodium, and the newer antipsychotics were described for some of the disorders.
These guidelines may be particularly useful and applicable in institutional and other settings where there are restrictions on the availability of certain medications: if a first choice is not available, suggestions for alternatives are described. We welcome suggestions for how the algorithms could be improved or amplified to account for particular patient subgroups not addressed in the current version. Readers may write to the corresponding author. Also, watch our web site, www.mhc.com, for announcements about when this algorithm is ready for interactive use on the Internet.
1. Kessler RC, Nelson CB, McGonagle KA, Edlund MG, Frank RG, Leaf PJ. The epidemiology of co-occurring mental disorders and substance use disorders in me National Comorbidity Survey: implications far prevention and service utilization. Am J Orthopsychiatry. 1996;66:17-31.
2. Sussman N. Interactive treatment algorithms for anxiety disorders. Primary Psychiatry. 1997;4:16-80.
3. Stein DJ, Jobson KO. Pharmacotherapy algorithms for anxiety disorders. Psychiatric Annals. 1996;26:190-232.
4. Jobson KO, Potter WZ. International psychopharmacology algorithm project report. Psychopharmacol Bull. 1995;31:483-507.
5. Longo LP. Non-benzodiazepine pharmacotherapy of anxiety and panic in substance abusing patients. Psychiatric Annals. 1998;28:142-153.
6. Fawcett J. Science, balance, and passion. Psychiatric Annals. 1998;28:116-117.
7. American Psychological Association. Diagnostic and Statistical Manual of Mental Disorders, 4th ed. Washington, DC: American Psychiatric Association; 1994.
8. Kranzler HR, Burleson JA, Del Boca FK, et al. Buspirone treatment of anxious alcoholics: a placebo controlled trial. Arch Gen Psychiatry. 1994;51:720-731.
9. Mason BJ, Kocsis JH, Ritvo EC, Cutler RB. A double-blind, placebo-controlled trial of desipramine for primary alcohol dependence, stratified on the presence or absence of major depression JAMA. 1996;275:761-767.
10. Osser DN. How is naltrexone used in the treatment of alcoholism? The Harvard Mental Health Letter. 1998;15:8.
11. Perez-Cruet JF, Iricanin T. Immediate naltrexone decreases in alcohol intake urges as predictor of a good outcome (NR 387). American Psychiatric Association Annual Meeting; San Diego, California; 1997.
12. Volpicelli JR, Rhines KC, Thines JS, Volpiceli! LA, Alterarian AI, O'Brien CP. Naltrexone and alcohol dependence: role of subject compliance. Arch Gen Psychiatry. 1997;54:737-742.
13. O'Mailey SS. Integration of opioid antagonists and psychosocial therapy in the treatment of narcotic and alcohol dependence. J Clin Psychiatry. 1995;56:30-38.
14. Goldenberg IM, White K, Yonkers K, et al. The infrequency of "pure culture" diagnoses among the anxiety disorders. J Clin Psychiatry. 1996;57:528-533.
15. Brawman-Mintzer O, Lydiard RB. Psychopharmacology of anxiety disorders: treatment resistance. Psychiatr Clin North Am. 1996;3:66-67.
16. Milrod B, Busch F, Cooper AM, Shapiro T, Manual of Panic-Focused Psychodynamic Psychotherapy. Washington, DC: American Psychiatric Press; 1997:128.
17. Rosenbaum JF, Pollack MH, Fredman SJ. The pharmacotherapy of panic disorder. In: Rosenbaum JF, Pollack MH, eds. Panic Disorder and Its Treatment. New York: Marcel Dekker; 1998:153-180.
18. American Psychiatric Association. Practice guidelines for the treatment of patients with panic disorder. Am J Psychiatry. 1998;155:23.
19. Woods JH, Katz JL, Winger G. Benzodiazepines: use, abuse, and consequences. Pharmacol Rev. 1992;44:151-347.
20. Fraser AD. Use and abuse of benzodiazepines. Ther Drug Monit. 1998;20:481-489.
21. DuPont RL. Anxiety arid addiction: a clinical perspective on comorbidity. Bull Menninger Clin. 1995;59:A53-A72.
22. Ckaulo DA, Bamhill JG, Ciraulo AM, et al. Alterations in pharmacodynamics of anxiolytics in abstinent alcoholic men: subjective responses, abuse liability, and electroencephalographic effects of alprazolam, diazepam, and buspirone. J CHn Pharmacol. 1997;37:64-73.
23. Humble M, Wistedt B. Serotonin, panic disorder, and agoraphobia: short-term and long-term efficacy of Citalopram in panic disorders. Int Clin Psychopharmacol. 1992;6:21-39.
24. Pollack MH, Otto MW, Worthington JJ- Manfro GG, Wolkow R. Sertraline in the treatment of panic disorder: a flexible-dose multicenter trial. Arch Gen Psychiatry. 1998;55:1010-1016.
25. Schneier FR, Liebowitz MR, Davies SO, et al. Fluoxetine in panic disorder. J Gin Psychopharmacol. 1990;10:119-121.
26. Pollack MH, Worthington JJ/ Otto MW, et al. Venlafaxine for panic disorder: results from a double-blind, placebocontrolled study. PsychopharmacQl Bull. 1996;32:667-670.
27. Thase ME. Effects of venlafaxine on blood pressure: a meta-analysis of original data from 3744 depressed patients. J Clin Psychiatry. 1998;59:502-508.
28. Mavissakalian M, Perei J, Bowler K, Dealy R. Trazodone in the treatment of panic disorder and agoraphobia with panic attacks. Am J Psychiatry. 1987;144:785-787.
29. Woods SW, Nagy LM, Koleszar AS, Krystal JH, Heninger GR, Charney DS. Controlled trial of alprazolam supplementation during imipramine treatment of panic disorder. J CKm Psychoptiarmacol. 1992;12:32-38.
30. Woodman CL, Noyes R Jr. Panic disorder: treatment with valproate. J Clin Psychiatry. 1994;55:134-136.
31. Crockatt JO, Greiner M, Out LL, Pande AC. Treatment of panic disorder with gabapentin (Poster 154). New Clinical Drug Evaluation Unit Program; Boca Raton, Florida; National Institute of Mental Health; 1998.
32. Kessler RC, McGonagle KA, Zhao S, et al. Lifetime and 12-month prevalence of DSM-HI-R psychiatric disorders in the United States: results from the National Comorbidity Survey. Arch Gen Psychiatry, 1994;51:8-19.
33. Fones CSL, Manfro GG, Pollack MH. Social phobia: an update. Harv Rev Psychiatry. 1998;5:247-259.
34. Schuckit MA, Hellelbrock V. Alcohol dependence and anxiety disorders: what is the relationship? Am J Psychiatry. 1994;151:1723-1734.
35. Lockwood AH. Medical problems of musicians. N Engl J Med. 1989;320:221-227.
36. Bouwer C, Stein DJ. Use of the selective serotonin reuptake inhibitor Citalopram in the treatment of generalized social phobia. J Affect Disord. 1998;49:79-82.
37. Van Ameringeh MA, Mancini CL, Oakman JM. Nefazodone in social phobia. J Clin Psychiatry. 1999;60: 96-100.
38. Worthington JJ III, Zucker BG, Fones CSL, Otto MW, Pollack MH. Nefazodone for social phobia: a clinical case series. Depress Anxiety. 1998;8:131-133.
39. Davidson JRT Pharmacotherapy of social anxiety disorder. J Clin Psychiatry. 1998;59(suppl 17):52.
40. Ware MW, Emmanuel NP, Johnson MR, BrawmanMintzer O, Kapp R, Lydiard RB. Self-reported sexual dysfunctions in anxiety disorder patients (NR676). American Psychiatric Association Annual Meeting; New York, New York; 1996.
41. Schneier FR, Goetz D, Campeas R, Fallon B, Marshall R, Liebowitz MR. Placebo-controlled trial of moclobemide in social phobia. Br J Psychiatry. 1998;172:70-77.
42. Osser DN, Patterson RD. Algorithms for the pharmacotherapy of depression: part one and part two. Directions in Psychiatry. 1998;18:303-336.
43. Heimberg RG, Liebowitz MR, Hope DA, et al. Cognitive behavioral group therapy vs phenelzine therapy for social phobia. Arch Gen Psychiatry. 1998;55:1133-1141.
44. van Vliet IM, den Boer JA, Westenberg HGM, Ho Pian KL. Clinical effects of buspirone in social phobia: a double-blind placebo-controlled study. J Clin Psychiatry. 1997;58:164-168.
45. Conner KM, Davidson JRT, Potts NLS, et al. Discontinuation of clonazepam in the treatment of social phobia. J Clin Psychopharmacol. 1998;18:373-378.
46. Marshall JR. The diagnosis and treatment of social phobia and alcohol abuse. Bull Menninger Clin. 1994;58:A58-A66.
47. Pande AC, Davidson JRT, Greist MH, Jefferson JW, Weisler R, Sutherland S. A placebo-controlled study of gabapentin in social phobia. American College of Neuropsychopharmacology Annual Meeting; San Juan, Puerto Rico; 1997.
48. Shear MK, Beidel DC Psychotherapy in the overall management strategy for social anxiety disorder. J Clin Psychiatry. 1998;59(suppl 17)3944.
49. DeBeIl C, Jones RD. As good as it seems? a review of EMDR experimental research. Professional Psychology: Research and Practice. 1997;28:153-163.
50. Ross RJ, Ball WA, Sullivan KA, Caroff SN. Sleep disturbance as the hallmark of posttraumatic stress disorder. Am J Psychiatry. 1989;146:697-707.
51. Kushida CA, Guilleminault C, Dement WC, Simon RD Jr, Ball EM. Diagnostic and treatment strategies for the obstructive sleep apnea syndrome. Journal of Clinical Outcomes Management 1998;5:49-65.
52. Kupfer DJ, Reynolds CF HI. Management of insomnia. N Engl J Med. 1997;336:341-345.
53. Rascati K. Drug utilization review of concomitant use of specific serotonin reuptake inhibitors or clomipramine with antianxiety /sleep medications. Clin Ther. 1995;17:786-790.
54. Friedman MJ. Current and future drug treatment for posttraumatic stress disorder patients. Psychiatric Annals. 1998;28:461468.
55. Hertzberg MA, Feldman MD, Beckham JC, Davidson JRT. Trial of trazodone for posttraumatic stress disorder using a multiple baseline group design. J Clin Psychopharmacol. 1996;16:294-298.
56. Clark R, Canive J, Brugger R, Vosburgh T, Tuason V. Cyproheptadine in the treatment of nightmares with posttraumatic stress disorder (Poster 158). New Clinical Drug Evaluation Unit Program, 38th Annual Meeting; Boca Raton, Florida; National Institutes of Mental Health; 1998.
57. Chilcoat HD, Breslau N. Posttraumatic stress disorder and drug disorders. Arch Gen Psychiatry. 1998;55:913-917.
58. Marshall RD, Schneier FR, Fallon BA, et al. An open trial of paroxetine in patients with posttraumatic stress disorder. Int Clin Psychopharmacol. 1998;18:10-18.
59. Connor KM, Davidson JRT. The role of serotonin in posttraumatic stress disorder: neurobiology and pharmacotherapy. CNS Spectrums. 1998;3:43-51.
60. Davidson JRT, Weisler RH, Malik ML, Connor KM. Treatment of posttraumatic stress disorder with nefazodone. fat Clin Psychopharmacol. 1998;13:111-113.
61. Hertzberg MA, Feldman ME, Beckham JC, Moore SD, Davidson JRT. Open trial of nefazodone for combat-related posttraumatic stress disorder. J Clin Psychiatry. 1998;59:460-464.
62. Cornelius JR, Salloum IM, Ehler JG, et al. Fluoxetine in depressed alcoholics: a double-blind, placebo-controlled trial. Arch Gen Psychiatry. 1997;54:700-705.
62a. Brady KX Sonne SC, Roberts JM. Sertraline treatment of comorbid posttraumatic stress disorder and alcohol dependence. J Clin Psychiatry. 1995;56:502-505.
63. Nierenberg AA, Adler LA, Peselow E, Zomberg G, Rosenthal M. Trazodone for antidepressant-associated insomnia. Am J Psychiatry. 1994;151:1069-1072.
64. David D, Kutcher GS, Jackson EI, Mellman TA. Psychotic symptoms in combat-related posttraumatic stress disorder. J Clin Psychiatry. 1999;60:29-32.
65. Hammer MB, Ulmer H, Home DF, Frueh C, Twomey TJ, Chobot K. Psychotic features and symptom severity in PTSD (NR 611). American Psychiatric Association Annual Meeting; Toronto, Ontario, Canada; 1998.
66. Hammer MB, Ulmer H, Huber MG, Measom MO. Risperidone treatment of psychotic features in PTSD (NR 610). American Psychiatric Association Annual Meeting; Toronto, Ontario, Canada; 1998.
67. Penava SJ, Otto MW, Pollack MH, Rosenbaum JF. Current status of pharmacotherapy for PTSD: an effect size analysis of controlled studies. Depress Anxiety. 1997;4:240-242.
68. Fawcett J, Barkin RL. A meta-analysis of eight randomized, double-blind, controlled clinical trials of mirtazapine for the treatment of patients with major depression and symptoms of anxiety. J Clin Psychiatry. 1998;59: 123-127.
69. Nierenberg AA, Feighner JP, Rudolph R, Cole JO, Sullivan J. Venlaiaxine for treatment-resistant unipolar depression. J Clin Psychopharmacol. 1994;14:419-423.
70. Wheatley DP, van Moffaert M, Timmerman L, Kremer CME, MFS G. Mirtazapine: efficacy and tolerability in comparison with fluoxetine in patients with moderate to severe major depressive disorder. J Clin Psychiatry. 1998;59:306-312.
71. Clerc GE, Ruimy P, Verdeau-Pailles J, VFIS G. A doubleblind comparison of venlaiaxine and fluoxetine in patients hospitalized for major depression and melancholia. Int Clin Psychopharmacol. 1994;9:139-143.
72. Connor KM, Davidson JRT, Weisler RH. A pilot study of mirtazapine in posttraumatic stress disorder (Poster III98). 37th Annual Meeting of the American College of Neuropsychopharmacology; Las Croabas, Puerto Rico; 1998.
73. Canive JM, Clark RD, Calais LA, Quails C, Tuason VB. Bupropion treatment in veterans with posttraumatic stress disorder: an open study. J Clin Psychopharmacol. 1998;18:379-383.
74. Fesler FA. Valproate in combat-related posttraumatic stress disorder. J Clin Psychiatry. 1991;52:361-364.
75. Goodman WK, Ward HE, Murphy TK. Biologic approaches to treatment-refractory obsessive-compulsive disorder. Psychiatric Annals. 1998;28:641-649.
76. Greist JH, Jefferson JW. Pharmacotherapy for obsessivecompulsive disorder. Br J Psychiatry. 1998;173:64-70.
77. March JS, Francis A, Carpenter D, Kahn DA. The expert consensus guidelines series: treatment of obsessive-compulsive disorder. J Clin Psychiatry. 1997;58(suppl 4):l-72.
78. Montgomery S. Citalopram treatment of obsessive-compulsive disorder: results from a double-blind, placebocontrolled trial (Poster LTI-102). 37th Annual Meeting of the American College of Neuropsychopharmacology; Las Croabas, Puerto Rico; 1998.
79. Abramowitz JS. Effectiveness of psychological and pharmacological treatment for obsessive-compulsive disorder. journal of Clinical and Consulting Psychology. 1997;65:44-52.
80. Oesterheld JR, Osser DN. Drug interactions in augmentation strategies for pharmacotherapy of obsessive-compulsive disorder. Journal of Practical Psychiatry and Behavioral Health. 1999. In press.
81. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 3rd ed. revised. Washington, DC: American Psychiatric Association; 1987.
82. Rickels K, Downing R, Schweizer E, Hassman H. Antidepressants for the treatment of generalized anxiety disorder: a placebo-controlled comparison of Imipramine, trazodone, and diazepam. Arch Gen Psychiatry. 1993;50:884-895.
83. Entsush R, Derivan AT, Haskins T, Rudolph RL, Aguiar L. Double-blind, placebo-controlled study of once- daily venlafaxine extended release and buspirone in outpatients with GAD (NR 644). American Psychiatric Association Annual Meeting; Toronto, Ontario, Canada; 1998.
84. Aguiar AM, Haskins T, Rudolf RL, Pallay A, Derivan AT. Double-blind, placebo-controlled study of once-daily venlafaxine extended release in outpatients with GAD (NP 643). American Psychiatric Association Annual Meeting; Toronto, Ontario, Canada; 1998.
85. Allgulander C, Cloninger CR, Przybeck TR, Brandt L. Changes on the temperament and character inventory after paroxetine treatment in volunteers with generalized anxiety disorder. Psychopharmacol Bull. 1998;34:165-166.
86. Wittchen HU, Zhao S, Kessler RC, Eaton WW. DSM-UI-R generalized anxiety disorder in the national comorbidity survey. Arch Gen Psychiatry. 1994;51:355-364.
87. Brawman-Mintzer O, Lydiard RB, Emmanuel N, et al. Psychiatric comorbidity in patients with generalized anxiety disorder. Am J Psychiatry. 1993;150:1216-1218.
88. Thompson PM. Generalized anxiety disorder treatment algorithm. Psychiatric Annals. 1996;26:227-232.
89. Tollefson GD, Montague-Clouse J, Tollefson SL. Treatment of comorbid generalized anxiety in a recently detoxified alcoholic population with a selective serotonergic drug (buspirone), J Clin Psychopharmacol. 1992;12: 19-26.
90. Apter JT, Allen LA. Buspirone: future directions. J Clin Psychopharmacol. 1999;19:86-93.
91. van Laar MW, Volkerts ER, van Willigenburg APR Therapeutic effects and effects on actual driving performance of chronically administered buspirone and diazepam in anxious outpatients. J Clin Psychopharmacol. 1992;12:86-95.
92. Smiley A, Moskowitz H. Effects of long term administration of buspirone and diazepam on driver steering control. Am J Med. 1986;80:22-29.
93. Delle Chiaie R, Pancheri P, Casacchia M, Stratta P, Kotzalidis GD, Zibellini M. Assessment of the efficacy of buspirone in patients affected by generalized anxiety disorder, shifting to buspirone from prior treatment with lorazepam: a placebo-controlled, double-blind study. J Clin Psychopharmacol. 1995;15:12-19.
94. DeMartinis NA, Rickels K, Mandos L, Rynn M, Schweizer E. Prior benzodiazepine use and buspirone response in the treatment of generalized anxiety disorder (Poster ?76). 37th Annual Meeting of the American College of Neuropsychopharmacology; Las Croabas, Puerto Rico; 1998.