Dr. Kennedy is professor in the Department of Psychiatry and Behavioral Sciences at Albert Einstein College of Medicine, and director of the Division of Geriatric Psychiatry at Montefiore Medical Center in the Bronx, New York.

Disclosure: Dr. Kennedy is a consultant to Myriad; is on the speaker’s bureaus of Forest and Pfizer; and has received grant support from Forest, Myriad, Novartis, Pfizer, and Takeda.

Please direct all correspondence to: Gary J. Kennedy, MD, Director, Department of Geriatric Psychiatry, MMC, 111 East 210th St, Klau One, Bronx, NY 10467; Tel: 718-920-4236; Fax: 718-920-6538; E-mail: gjkennedy@msn.com.


Parkinson’s disease affects as many as 1 million Americans and with advanced age is complicated by dementia in a majority of cases. However, the recognition of cognitive impairment in Parkinson’s disease is made complicated by the predominance of motor symptoms and a neuropsychiatric profile that differs from the more common dementia of the Alzheimer’s type. Differentiating the decline in personal and social activities due to cognitive impairment rather than preexisting movement disorder is difficult. Several expert bodies have addressed the use of cholinesterase inhibitors for the dementia of Parkinson’s disease, but the evidence base is far less substantial than that which exists for Alzheimer’s disease. Although most patients with Parkinson’s disease dementia should be offered a trial of anti-cholinesterase therapy, particularly those experiencing hallucinations, dramatic benefits are not common. Temporary symptomatic relief rather than disease modification is the most that can be expected. As a result, treatment should be presented as an option rather than an imperative.


Parkinson’s disease has a mean age of onset of 57 years and a prevalence of 1% to 2% among adults ≥60 years of age. There may be as many as 1 million Americans with the illness.1 It is manifested by bradykinesia, rigidity, resting tremor, postural instability, frozen gait disorder, and flexed posture. The progression and severity of Parkinson’s disease varies widely and the associated motor disability may be substantially reduced by numerous medications singly or in combination. The goal is an increase in brain dopamine through either enhanced production or reduced breakdown of the molecule.2 The disease begins as a movement disorder, but with advancing age is complicated by dementia in as many as 80% of patients. The characteristic frozen facial expression, slowed cognition (bradyphrenia), fluctuation in attention, and motor impairment compounded by depression and hallucinations make the assessment of possible dementia challenging. In addition, dopaminergic medications precipitate hallucinations with recognized frequency. Although medication to treat the dementia of Parkinson’s disease is most often modestly effective for patients in aggregate, failure to recognize dementia means that the minority of individuals who might receive substantial benefit will not be offered a therapeutic trial. As a result, practitioners need guidance to efficiently assess cognitive decline among people with Parkinson’s disease as well as realistic expectations for the benefits of dementia treatment. 

Similarities and Differences with Other Dementias

Loss of dopaminergic neurons in the substantia nigra is the hallmark of Parkinson’s disease and the basis for the use of dopamine agonists. In contrast, neuronal dropout in the entorhinal cortex and hippocampus are seen in Alzheimer’s disease. Yet, similar to Alzheimer’s disease, cholinergic deficits are common in Parkinson’s disease and parallel the decline in cognition. Hematoxilyn and eosin staining neuronal inclusions known as Lewy bodies occur in both Parkinson’s and Lewy body dementia but not in Alzheimer’s disease. However, amyloid plaques and neurofibrillary tangles thought to be the signature pathology of Alzheimer’s disease commonly occur in both Parkinson’s disease and Lewy body dementia as well, though less extensively. Differences between Alzheimer’s disease and the dementia of Parkinson’s and Lewy body disease detected by imaging studies, whether structural (magnetic resonance imaging [MRI], computerized axial tomography) or metabolic (positron emission tomography, single photon emission computed tomography, functional MRI), are too subtle for use in clinical diagnosis.   

Clinical features of Alzheimer’s, Parkinson’s, and Lewy body dementia overlap as severity progresses but significant differences are apparent in the early stages (Table 1).3-5 In Alzheimer’s disease, memory impairment is prominent with executive dysfunction, aphasia, apraxia, and anomia often present but less obvious. Apathy is more common than depression, but one or the other will be present in ~25% of affected individuals. Hallucinations occur in 10% of cases most often at the mid- to later stage of the disease.6 Severe motor impairments in gait, balance, muscle strength, and swallowing occur in the later stages. In Parkinson’s dementia, the movement disorder precedes the onset of cognitive impairment. Inattention, executive dysfunction, bradyphrenia, and visuospatial deficits may be more noticeable than impaired memory. Hallucinations are four times more frequent and depressive symptoms somewhat less so than in Alzheimer’s disease. Irritability, anger, aggression, and delusions may be more prominent in Alzheimer’s disease. More severe postural instability and gait disorder predict the onset of dementia among people with Parkinson’s disease. Incident hallucinations also predict the subsequent emergence of dementia.3


Hallucinations are also a distinguishing feature of Lewy body dementia. Marked fluctuations in attentiveness and mild impairment in memory—both of which precede the appearance of rigidity, tremor, postural instability and gait disorder—distinguish Lewy body dementia from that of Parkinson’s disease. Unanticipated sensitivity to neuroleptic-induced extrapyramidal symptoms also indicates that Lewy body dementia, rather than Alzheimer’s disease with hallucinations, is the correct diagnosis.

Efficient Screening Procedures 

Not every older adult should be screened for cognitive impairment. Screening in clinical practice is predicated on the recognition of cognitive decline interfering with personal or social activities by the patient, family, or clinician. However, Parkinson’s disease often impacts social and personal activities as a result of motor impairment. Thus, personal responsibilities may already have been abandoned before impaired cognition could have made noticeable contribution. Given the elevated frequency with which dementia emerges in Parkinson’s disease, the practitioner’s concern for impairment should be heightened. When hallucinations, apathy, or excessive daytime drowsiness appear after a period of stable treatment, dementia should be suspected.

In a 2007 review7 on the diagnosis and treatment of Parkinson’s disease dementia, the Movement Disorders Society’s taskforce recommended the Mini-Mental State Examination (MMSE) as a global measure of cognitive performance in which a score <26 indicates impairment. They also suggested a number of screening procedures to detect impairments in specific cognitive domains. These included domains of attention, visuo-constructive ability, executive function, and memory. Parkinson’s disease patients with impairments in more than one domain associated with deterioration in personal care or social activities would meet the criteria for dementia. Attention would be assessed with the serial seven subtraction task from the MMSE or by asking the patient to list months of the year in reverse order. In either test, two errors or omissions is considered evidence of impairment. For executive dysfunction, impairment is defined as failure to recite nine examples from the lexical category of words starting with the letter “S” in one minute or inability to draw a clock with the time set at 10 past 2. Visuo-constructional impairment is defined by inability to copy two intersecting pentagons from the MMSE. Impairment in memory is defined by failure to recall all three words from the MMSE’s registration and recall task. The review also provides a comprehensive listing of neuropsychological instruments that have been used to assess cognition among people with Parkinson’s disease.

However, busy practitioners may find the copyrighted MMSE and clock drawing cumbersome. As an alternative, the Memory Impairment Screen and the oral version of the Trail Making Test for executive function do not require paper and pencil, may be administered by phone, and are quite brief. Both have been validated as screening measures for use in population assessments of dementia.8 For the Memory Impairment Screen, the subject is tasked to repeat and remember four words (eg, apple, table, penny, spoon) given in sequence at 1-second intervals and then asked to recall each when prompted with a category cue (eg, fruit, furniture, money, utensil). The registration phase may be repeated up to five times before moving to the next test. Next, the subject is asked to recite the alphabet from “A to Z” and then to count from 1–25. The person is then asked to continue the sequence, which the examiner starts with “One A, Two B, Three C, Four ?” Subjects making two errors as they reach “M 13” are considered to exhibit executive impairment. Following the Trail Making test, the examiner returns to the Memory Impairment Screen by asking the subject to recall the four words that were previously rehearsed. Allowing 1 minute for free recall the examiner then provides the category cue for each word not remembered spontaneously. Words recalled freely receive a score of 2; those that required the cue for recall receive a score of 1. A total score of 4 is predictive of dementia. A free demonstration of the Trail Making Test, clock-drawing test, and other measures of executive dysfunction can be accessed on the Internet.6 Baseline assessment of cognition with simple screening procedures following the diagnosis of Parkinson’s disease will set the stage for detection of genuine impairment should warning signs of dementia emerge.

Responsiveness to Cholinesterase Inhibitors

The evidence base regarding the efficacy of pharmacotherapy for Parkinson’s disease dementia is limited9 but more extensive if one considers the illness to be part of a spectrum including dementia with Lewy bodies.10 In the largest randomized controlled trial to date, Emre and colleagues11 conducted a pivotal study of 541 people whose Parkinson’s disease was accompanied by mild-to-moderate dementia defined by MMSE score of 10–24. Patients were allocated to rivastigmine or placebo in a 2:1 ratio. Rivastigmine 1.5 mg was introduced and titrated to a maximum tolerated dose or 12 mg over 16 weeks. Exclusion criteria included a history of major depressive disorder, use of cholinesterase inhibitor or anticholinergic drug, or change in Parkinson’s disease medication within 4 weeks of enrollment. The initiation of a psychotropic medication during the study with the exception of an antipsychotic for an acute episode of psychosis was forbidden. The mean age of study participants was slightly >72 years and 66% were men. Forty percent were diagnosed with a mental disorder in addition to dementia.

Greater than 25% of participants were taking an antipsychotic at baseline, 25% were on an antidepressant, 20% were on a benzodiazepine or sedative hypnotic, and 95% were taking levodopa. The two primary efficacy measures were the Alzheimer’s Disease Assessment Scale (ADAS-cog) and the Alzheimer’s Disease Cooperative Study-Clinician’s Global Impression of Change (ADCS-CGIC). Each were separately assessed by trained raters blind to the assessment outcome. The ADAS-cog is a 70-point measure of cognitive performance. The ADCS-CGIC is a 7-point categorical scale anchored at baseline where “1” equals marked improvement, “7” equals marked worsening, and “4” denotes no change. Detectable changes that were not clinically meaningful were defined as minimal; changes associated with obvious clinical improvement were defined as moderate. Secondary efficacy measures included the ADCS measure of activities of daily living, the Neuropsychiatric Inventory, the MMSE, tests of attention and reaction time, and tests of executive function as measured by letter-category verbal fluency and clock-drawing.

At the end of dose titration, >50% of treated participants were taking rivastigmine 9–12 mg/day. Of those completing the study, 72% were in the rivastigmine group and 82% in placebo. Adverse events accounted for most of the premature withdrawals in both groups. Nausea (29%), vomiting (16.6%), and worsening tremor (10.2%) were significantly more frequent among the rivastigmine group. Of the efficacy measures comparing rivastigmine to placebo, at 24 weeks all were statistically significant. There was an 11.7% difference in cognitive performance (ADAS-cog) between drug and placebo groups. Clinically meaningful (marked to moderate) improvement was seen in 14.5% of placebo and 19.8% of the rivastigmine group. Clinically meaningful deterioration was seen in 23.1% of placebo and 13.0% of the rivastigmine group. Among the secondary measures of disability, neuropsychiatric symptoms, cognition, and executive functions all showed improvement with rivastigmine and decline with placebo with the differences being statistically significant. Of note, 34.6% of placebo and 45.4% of the rivastigmine group exhibited a ≥30% improvement in neuropsychiatric symptoms. Hallucinations disruptive enough to be considered adverse events occurred in 9.5% of placebo and 4.7% of rivastigmine participants, and the difference was significant.

Emre and colleagues11 conclude that their findings among patients with Parkinson’s disease and dementia are similar to those seen in studies of cholinesterase inhibitors for Alzheimer’s disease. Benefits are modest, representing a 6-month reprieve in the course of symptoms without genuine disease modification. However, close to one in five patients will show dramatic benefit obvious to the family and the clinician alike. Indeed, Press12 advocates the family’s impression of benefit over formal neuropsychological measures to assess the effectiveness of treatment. Given that benefits if apparent at all emerge within the first weeks of treatment, the most realistic expectation for patient and family is a 60–90-day trial of therapy rather than an open-ended course. Also noteworthy were the relatively more substantial benefits of rivastigmine for neuropsychiatric symptoms, including hallucinations, a finding similar to McKeith and colleagues’10 rivastigmine study of 120 patients with Lewy body dementia. The number of study participants receiving a neuroleptic during the Emre and colleagues11 study was considerable. Low doses of atypical antipsychotics such as olanzapine, quetiapine, and clozapine were initially used to control levodopa-induced hallucinations because they were less likely than typical antipsychotics to induce extrapyramidal symptoms.12

However, the 2005 Food and Drug Administration warning of increased mortality when prescribed to patients with dementia11 amplified by more recent reports14,15 raise the threshold at which these agents may be considered for psychosis in dementia nearly out of reach. Clearly, a trial of cholinesterase therapy should be recommended to reduce hallucinations of dementia in Parkinson’s disease before a neuroleptic is offered. The transdermal rivastigmine patch with less frequent gastrointestinal effects, not available at the time of Emre’s11 study, is an added advantage.16 It should be added that other cholinesterase inhibitors may be efficacious for the dementia of Parkinson’s disease but have not been subjected to large-scale trials. Nonetheless, the Cochrane Review and other sources17-20 find at least minimal evidence to support the use of donepezil as well (Table 2).




Dementia is such a frequent complication that every older patient with Parkinson’s disease should be screened for memory impairment and executive dysfunction, particularly if hallucinations emerge in the context of stable dopaminergic treatment. The impact of cholinesterase inhibitor therapy on cognition and activities of daily living will be obvious in only one patient in five. However, the reduction in hallucinations may be more robust. Consistent reports of elevated mortality associated with antipsychotics prescribed to people with dementia make cholinesterase inhibitor therapy preferable when hallucinations emerge. A 60–90-day trial of a cholinesterase inhibitor should be adequate to give the patient, family, and practitioner sufficient evidence on which to base a decision for ongoing treatment. In equivocal cases, the medication can be reinstituted if visible decline is observed following discontinuation.8 With adverse reactions and lack of efficacy taken into account, 40% to 60% of treated patients would withdraw from cholinesterase inhibitor therapy. However, lacking predictors of treatment responsiveness and balancing in the safety of cholinesterase inhibitor therapy, most patients with Parkinson’s disease dementia should be offered a trial of treatment. Practitioners’ zeal for treatment must be tempered by the realization that temporary symptomatic relief rather than disease modification is the most that can be expected. As a result, treatment should be presented as an option rather than an imperative. PP


1. Twelves D, Perkins KS, Counsell C. Systematic review of incidence studies of Parkinson’s disease. Mov Disord. 2002;18(1):19-31.
2. LeWitt PA. Levodopa for the treatment of Parkinson’s disease. N Engl J Med. 2008;359(23):2468-2476.
3. Emre M, Aarsland D, Brown R, et al. Clinical diagnostic criteria for dementia associated with parkinson’s disease movement disorders. Mov Disord. 2007;22(12):1689-1707.
4. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Association; 1994.
5. McKeith IG, Dickson DW, Lowe J, et al. Diagnosis and management of dementia with Lewy bodies: third report of the DLB Consortium. Neurology. 2005;65(12):1863-1872. Erratum in: Neurology. 2005 ;65(12):1992.
6. Lyketsos CG, Sheppard JM, Steinberg M, et al. Neuropychiatric disturbance in Alzheimer’s disease clusters into three groups: the Cache County study. Int J Geriat Psychiatry. 2001;16(11):1043-1053.
7. Dubois B, Burn D, Goetz C, et al. Diagnostic procedures for Parkinson’s disease dementia: recommendations from the Movement Disorder Society Task Force. Mov Disord. 2007;22(16);2314-2324.
8. Kennedy GJ, Smyth CA. Screening older adults for executive dysfunction: an essential refinement in the assessment of cognitive impairment. American Journal of Nursing. 2008;108(12):60-69.
9. Poewe W, Gauthier S, Aarsland, et al. Diagnosis and management of Parkinson’s disease dementia. Int J Clin Pract. 2008;62(10);1581-1587.
10. McKeith I, Del Ser T, Spano P, et al. Efficacy of rivastigmine in dementia with Lewy bodies: a randomized, double blind, placebo controlled international study. Lancet. 2000;356(9247):2031-2036.
11. Emre M, Aarsland, Albanese A, et al. Rivastigmine for dementia associated with Parkinson’s disease. N Eng J Med. 2004;351(24):2509-2518.
12. Press DZ. Parkinson’s disease dementia–A first step? N Engl J Med. 2004;351:(24)2547-2549.
13. Kennedy GJ. Caution vs. closure: the use of atypical antipsychotics for the treatment of behavioral disturbances in dementia. Primary Psychiatry. 2005;12(9):16-19.
14. Ballard C,  Hanney ML, Theodoulou M, et al. The dementia antipsychotic withdrawal trial (DART-AD): long-term follow-up of a randomised placebo-controlled trial. Lancet Neurol.  2009;8(2):151-157.
15. Ray WA, Chung CP, Murray KT, Hall K, Stein CM. Atypical antipsychotic drugs and the risk of sudden cardiac death. N Eng J Med. 2009;360(3):225-235.
16. Lefèvre G, Pommier F, Sedek G, et al. Pharmacokinetics and bioavailability of the novel rivastigmine transdermal patch versus rivastigmine oral solution in healthy elderly subjects. J Clin Pharmacol. 2008;48(2):246-252.
17. Maidment I, Fox C, Boustani M. Cholinesterase inhibitors for Parkinson’s disease dementia. Cochrane Database Syst Rev. 2006;(1):CD004747.
18. Ravina B, Putt M, Siderowf A, et al. Donepezil for dementia in Parkinson’s disease: a randomised, double blind, placebo controlled crossover study. J Neurol Neurosurg Psychiatry. 2005;76(7):934-939.
19. Aarsland D, Laake K, Larsen JP, Janvin C. Donepezil for cognitive impairment in Parkinson’s disease: a randomised controlled study. J Neurol Neurosurg Psychiatry. 2002;72(6):708-712.
20. Leroi I, Brandt J, Reich SG, et al. Randomized placebo-controlled trial of donepezil in cognitive impairment in Parkinson’s disease. Int J Geriatr Psychiatry. 2004;19(1):1-8.



Dr. Raby is assistant clinical professor of psychiatry at Columbia University in New York City.

Disclosure: Dr. Raby reports no affiliation with or financial interest in any organization that may pose a conflict of interest.
Off-label disclosure: This article includes discussion of the following unapproved medications for co-occurring marijuana abuse or dependence and psychotic disorders: bupropion, dronabinol, and nefazodone.

Please direct all correspondence to: Wilfrid Noël Raby, MD, PhD, Division on Substance Abuse, Unit 66, New York State Psychiatric Institute, 1051 Riverside Dr, New York, NY, 10032;
Tel: 212-923-3031; Fax: 212-568-3832; E-mail: rabywil@pi.cpmc.columbia.edu.


Focus Points

• Cannabis use before 15 years of age increases the risk of serious mental illness, especially psychotic illness later in life.
• A family history of psychiatric illness may increase the risk of cannabis-induced psychosis.
• Clinicians need to investigate not only the use of cannabis by patients, but also its effect, in order to determine vulnerability to mental illness from its ongoing use.


Marijuana abuse can lead to transient psychosis, but can it cause or worsen psychotic disorders like schizophrenia? This article reviews the evidence from key research reports, leading to the conclusion that marijuana use, especially in early adolescence, can lead to psychotic disorders in adulthood, such as schizophrenia. There is a want of treatment approaches for marijuana use in individuals with schizophrenia, or for emerging psychosis in patients dependent on marijuana. Second-generation antipsychotics, especially clozapine, appear to be the best approach to treatment for psychosis co-occurring with—and often secondary to—marijuana abuse. More research is needed to develop appropriate and effective treatments for marijuana dependence, both alone as well as in conjunction with psychosis and psychotic disorders.


Is marijuana dangerous? With an estimated 150 million people worldwide smoking or eating marijuana leaves annually,1 the question is pertinent. Marijuana is perceived as an innocuous drug in many circles due to its association with cultural and religious rituals, and with the fact that unlike alcohol, cocaine, or heroin, it rarely brings individuals to the brink of destitution. However, this perception is changing. In 1997, Tanda and colleagues2 reported that marijuana, like most drugs of abuse, increases dopamine release in the nucleus accumbens. Moreover, the increasing potency of available marijuana has led to the recognition of a withdrawal syndrome, characterized by irritability, restlessness, insomnia, anorexia, and aggressivity, which may last up to several weeks after stopping marijuana.3 Marijuana, like tobacco smoking, also increases the risk of lung cancer in young adults.4 With respect to mental health, marijuana smoking is reported to elicit psychotic disorders in individuals at risk5 as well as worsen psychotic symptoms in patients with psychotic disorders. This last point will be the focus of this article, which will review of the evidence, discuss clinical symptoms that may indicate an enhanced risk of psychosis stemming from marijuana, and present available treatment options.

Review of the Evidence Linking Marijuana to Psychosis

Marijuana use appears to be beginning at an increasingly early age. Based on the Substance Abuse and Mental Health Service Administration (SAMHSA) 2002–2003 survey, 90.8 million adults in the United States (42.9%) ≥18 years of age had used marijuana at least once in their lifetime. Among them, 2.1% had reported a first use before12 years of age, 52.7% between 12–17 years of age, and 45.2% at ≥18 years of age.6 In the same survey, 12.5% of individuals >18 years of age who reported lifetime use of marijuana were classified as having a serious mental illness in the past year. Furthermore, 21% of adults who first used marijuana before 12 years of age were classified as having a serious mental illness in the past year, as opposed to 10.5% of adults who had first used at ≥18 years of age. Strictly with respect to psychosis, results from the US National Epidemiological Catchment Area Study7 highlight that daily marijuana smokers were 2.4 times more likely to report psychotic symptoms than non-daily users, even after adjusting for psychiatric conditions and sociodemographic factors.8 Data like these have created the suspicion that marijuana may not be as innocuous as it has been previously thought.

Before inquiring about psychotic disorders, this article evaluates how prone marijuana users are to experience some form of psychosis. As early as 1972, marijuana use was stated to possibly cause acute psychosis.9 Usually, the effects of marijuana are dose related. Mild intoxication causes drowsiness, euphoria, and heightened sensory perception, while severe intoxication leads to motor incoordination, lethargy, and postural hypotension.10 Psychosis is not considered a usual manifestation of marijuana use. Cross-sectional studies have attempted to look at the types of symptoms that might be elicited by marijuana: positive (perceptual anomalies, magical or paranoid ideation), and negative (asociality, anhedonia) in non-clinical samples. While methodologic differences abound in these studies, these studies11-14 imply that marijuana users are more prone to transient positive symptoms of psychosis; one study15 found an association with negative symptoms as well. It is unclear whether these negative symptoms represent true negative symptoms or the so-called “amotivational syndrome” (loss of interests, motivation, impaired occupational performance and achievement), which is described as a subacute, reversible encephalopathy caused by chronic marijuana use.16 A review17 of randomized trials unrelated to mental health assessing the antiemetic effects of cannabis found that 6% of patients receiving cannabis experienced hallucinations and 5% paranoia, effects not seen with the other antiemetic drugs tested. Using a method called Experience Sampling Method, which is a structured daily diary method to investigate subjective experience during daily life in which subjects a prompted every three hours to complete the diary, Verdoux and colleagues18 found that in a given 3-hour period the likelihood of reporting unusual perceptions was increased if marijuana was used in the same 3-hour period, and not if used in the previous three hour period. This finding is consistent with the estimated duration of the pharmacologic effects of marijuana.19 With heavy marijuana use, symptoms of hypomania, agitation, auditory hallucinations, and thought disorder have been reported, which have tended to improve substantially after 5–7 days.20 However, one may ask if the association between marijuana use and psychotic experiences extends to psychotic disorders. National surveys support this association, such as the data from the US National Epidemiological Catchment Area study7 presented earlier. Two other national surveys also concur. The Australian National Survey of Mental Health and Well Being revealed that 12% of those diagnosed with schizophrenia also met International Classification of Diseases and Health Related Problems, Tenth Edition21 criteria for cannabis dependence. After adjusting for other disorders and sociodemographic factors, individuals with cannabis dependence were found to be nearly three times as likely to be diagnosed with schizophrenia as those not diagnosed with cannabis dependence.22 In the Netherlands, marijuana use was more prevalent among individuals with psychosis (15.3%) than those without (7.7%).23 Taken together, these findings support the association of marijuana use not only with transient psychosis, but also with the development of psychotic disorder. However, they cannot answer the question: do we need to worry that marijuana use can cause a psychotic illness?

The issue of causality is a difficult one to answer when it comes to conditions such as psychosis which can have multiple etiologies. To establish causality, three factors must be established: association (presented above), a temporal priority, and a direction of effect.24 The latter two factors can only be scrutinized in prospective studies, where a group is selected for assessment of a risk (marijuana use) and followed over time to evaluate how potent the risk is in causing a particular condition (psychotic disorders).

Two landmark prospective studies will be reviewed: the Swedish Conscript Cohort25,26 and the Dunedin study from New Zealand.27 The Swedish study examined a cohort of 50,087 conscripts and found a dose-response relationship between marijuana use at 18 years of age and a schizophrenia diagnosis. Self-described “heavy marijuana users” (>50 lifetime use) were 2.3 times more likely than non-users to have a schizophrenia diagnosis 15 year later (after controlling for pre-existing psychosis).25 When the analysis was extended to 27 years, heavy users were 6.7 times more likely than non-users to carry a schizophrenia diagnosis, after controlling for drug use other than marijuana, low intelligence quotient, and antisocial personality, among other factors.26 Restricting the analysis to a 5-year window past 18 years of age to examine whether cannabis use might be a result of prodromal psychosis did not change those risks, leading the authors to state that their results were “consistent with a causal relationship between cannabis use and schizophrenia.”26

The prodromal phase of schizophrenia is marked by gradual but profound changes in behavior, perception, and cognition, raising the question as to whether marijuana use may be a consequence of emerging psychosis rather than a cause of it. Although small in contrast to the Swedish study, the Dunedin study27 provided unique insights in this regard, studying a birth cohort of 1,037 individuals born in Dunedin, New Zealand between 1972–1973, with a 96% follow-up rate at 26 years of age. It gathered information on self-reported psychotic experiences at 11 years of age, before the onset of marijuana use, and on self-reported use of marijuana at 15 and 18 years of age. All individuals were assessed to yield Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition28 diagnoses if present at 26 years of age, allowing the investigators to note the presence of psychotic symptoms along a continuum or the presence of a formally diagnosed psychotic disorder. Psychotic symptoms stemming from alcohol or other drugs were ruled out. Cannabis use by 15 and 18 years of age, respectively, led to higher rates of psychosis at age 26 compared to non-users, even after controlling for psychotic experiences preceding marijuana use. Age of first marijuana use was a significant factor: 10.3% of individuals who had used marijuana by 15 years of age were diagnosed with schizophreniform disorder at 26 years of age, compared to 3% of controls. The risk for adult schizophreniform disorder remained elevated after controlling for psychotic experiences at 11 years of age, with an odds ratio of 3.1. Marijuana use by 15 years of age did not predict depression at 26 years of age, and other drug use did not pose a risk for schizophreniform disorder above the one posed by marijuana. Marijuana use begun between ages 15 and 18 was associated with a heightened risk for schizophreniform disorder, but only if preceded by psychotic experiences at 11 years of age. This study corroborated the notion that marijuana use in adolescence is a risk factor for schizophrenia in later life, especially if used at an early age, suggesting both a temporal priority and direction between early marijuana use and schizophrenia. The issue of age may be especially important because at  ≤15 years of age, the developing brain may be especially susceptible to suspected trophic and neurobiologic effects of marijuana exposure, for which there is accumulating evidence.29-31

Since the Dunedin study, other studies and reviews have lent support to its findings. Semple and colleagues32 conducted a meta-analysis in which odds ratio from 2–9 were found between early exposure to marijuana and psychosis, leading them to conclude that early marijuana is an independent risk factor for psychosis and psychotic disorders. Arendt and colleagues33 reported on a cohort of 535 patients who had been diagnosed with marijuana-induced psychotic disorder and found that 47% of the patients received a diagnosis of schizophrenia 1 year later. Ferdinand and colleagues34 concluded, after a 14-year follow-up study of 1,580 individuals 4–16 years of age at study entry, that there was a specific link between marijuana use and psychosis, independent of other forms of psychopathology. In a nationwide population-based sample of 2 million individuals, the authors concluded that marijuana-induced psychosis could be an early sign of schizophrenia rather than a distinct form of psychosis.35 In individuals with prodromal symptoms of schizophrenia, marijuana increased the intensity and frequency of psychotic symptoms, especially hallucinations, and did so during and shortly after marijuana use.36 This lead the authors to ponder whether marijuana could worsen prodromal symptoms and increase the likelihood of developing schizophrenia  in young adolescents at risk. Genetic predisposition may further enhance this risk. In a study of the Dunedin cohort, Caspi and colleagues37 reported that individuals with a functional polymorphism in the catechol-O-methyltransferase (COMT) gene were at increased risk of schizophreniform disorder after use of marijuana during adolescence as compared with those who did not carry this polymorphism. Similar evidence is being found for polymorphisms at the cannabinoid receptor (CB1).38 These genetic factors may influence future risk of schizophrenia by interacting with other potential risk factors. For example, accumulating evidence points to dysregulation of the endogenous cannabinoid anandamide in patients with schizophrenia, with elevation of anandamide levels in blood and cerebrospinal fluid during acute exacerbations of psychosis and resolution after treatment (Figure).10,39-41 Hence, exogenous cannabinoids may worsen preexisting states that could make some individuals more at risk to develop schizophrenia from consuming marijuana.




Faced with this mounting evidence that marijuana use, particularly at an early age, can increase the risk of schizophrenia in adults, what is a clinician to do? The next section will describe an approach that may help in advising patients on the risk inherent to marijuana use and on the risk of developing a psychotic disorder.

Clinical Approaches to Advising Patients

Given the prevalence of marijuana use, psychiatrists and clinicians will encounter patients who smoke marijuana. As a starting point, not only is it important to know which substances have been and are being used (marijuana in this instance), but it is useful to ask the patient about what they experienced when smoking marijuana. Usually, mild intoxication can be followed by drowsiness, euphoria, heightened sensory awareness, and altered time perception. Moderate intoxication may produce memory impairments, depersonalization, and mood alteration. Severe intoxication can lead to decreased motor coordination, lethargy, slurred speech, and postural hypotension. These are the usual symptoms of marijuana use. Individuals who consistently experience these symptoms may have smoked marijuana for many years, with perhaps an ensuing decline in motivation, mental acuity, and a stalling in their personal and professional achievements. These later symptoms are often those that bring these patients to seek treatment. For those other patients who may be unknowingly more at risk of psychosis from marijuana, the experience of consuming marijuana seems to be different.

Clinicians are encouraged to look for any symptoms that might differ from the usual effects of marijuana stated above. After first smoking marijuana, or after some time thereafter, some patients may experience dysphoria, restlessness, generalized anxiety, panic attacks, paranoia, and sometimes hallucinations (Table). In most cases, marijuana-induced psychiatric symptoms, such as panic attacks, agitation, or persecutory delusions, are transient.5,42 Although no literature appears to exist looking at how these early effects of marijuana may portend future risk of mental illness, they may represent a first warning. In this author’s experience, marijuana-smoking patients with these symptoms frequently have a family history of psychiatric illness as well, be it depression, bipolar disorder, anxiety disorders, or schizophrenia. How these familial risks enhance the probability of acquiring a psychotic disorder from marijuana is not yet elucidated. Nonetheless, this author has witnessed patients with these anomalous effects of marijuana go on to develop autonomous psychotic disorders from not stopping their marijuana use in time. In the face of current evidence, the most conservative stance would dictate that patients be told that symptoms contrary to the usual effects of marijuana may signal that continued use of marijuana may possibly and seriously jeopardize their future mental health, although no definite proof of this exists for now. Presently, the state of reimbursement for clinical care forbids ancillary testing that might substantiate this risk, such as genetic testing for polymorphisms in the COMT or CB1 receptor genes.



Treatment for Marijuana-related Psychosis

Compared to psychosis unrelated to marijuana, marijuana-associated psychosis is emerging as more challenging to treat. In established schizophrenia, marijuana or other drug abuse leads to decreased adherence to treatment43 as well as increases recurrence of symptoms,44 episodes of violence,45 victimization (such as being used as drug “mules” to carry drugs),46 hospitalizations,47 and suicide.48 This underlies the seriousness of the problem, and the importance of developing effective treatments. Before moving on to potential medication treatments, the context of treatment deserves special mention. Programs that integrate counseling for substance abuse, psychosocial support for mental illness, and medication treatment provide the continuity and comprehensiveness that is more likely to make such treatment a success with the severely mentally ill. The inclusion of cognitive-behavioral and motivational interviewing approaches enhances treatment success.44,49-51 Features such as contingency management, where abstinence is rewarded with small prizes, can further increase success.52 For the most recalcitrant patients, long-term residential programs must be considered.53 However, many clinics are not equipped to provide such comprehensive services, and much remains to be overcome to disseminate such services throughout the current mental health network and to a wider population.

There are few guidelines concerning the pharmacologic treatment of co-occurring marijuana abuse or dependence and psychotic disorders. With respect to marijuana dependence itself, the cannabinoid receptor antagonist rimonabant is showing promise in primate trials to alter marijuana-seeking behavior.54 Low dose naltrexone (12 mg) has been reported to reduce the effects of marijuana, an approach that may hold promise in schizophrenia patients.55 Nefazodone, buspirone, and dronabinol show some promise as well in attenuating the manifestations of marijuana withdrawal.56 However, this research is in the preliminary stages, and it is yet to be made clear how these various approaches can be implemented in schizophrenia patients with marijuana dependence. As psychosis must be addressed in any approach to treatment for these patients, antipsychotics have featured prominently in the attempts to treat psychosis and marijuana dependence.

The first-generation antipsychotics appear to have little role in the treatment of other cannabis use disorders, and indeed, there are reports that they may worsen substance abuse.57 Older antipsychotics, especially high-potency dopamine antagonists, may further disrupt an already dysregulated mesocorticolimbic dopamine pathway, a feature common to both schizophrenia58 and drug dependence.59 Marijuana or other drug use may very well transiently relieve core deficits in schizophrenia patients, even though it may worsen psychotic symptoms.60,61 Buckley and colleagues62 reported on a 6-month study with clozapine, showing equal response in individuals for schizophrenia who did and did not use recreational drugs. Outcomes from dual diagnosis programs are of interest as well; for the 36 out of 151 schizophrenia patients on clozapine, remission rates from marijuana and alcohol were reported at 67% to 79%, compared to 34% for the remaining patients on first-generation antipsychotics.63 A 10-year follow-up study of this group showed that schizophrenia patients on clozapine and in remission had an 8% relapse risk in the following year compared to 40% on typical antipsychotics.64 Results have been more equivocal for the other second-generation antipsychotics.65 These antipsychotics, most prominently clozapine, seem to offer the best approach to the treatment of marijuana-associated psychosis, based on the literature available and in the personal experience of the author who has treated many patients with emerging psychosis due to marijuana use. Few treatments for individuals with emerging psychosis from marijuana use can be sifted from the existing literature. This author has found clozapine, olanzapine, and aripiprazole to be most useful in treating such patients that do not yet meet criteria for schizophrenia or other psychotic disorders.


Despite the major public health problems posed by marijuana abuse, the weight of disability imposed by schizophrenia, and the emerging consensus that marijuana use—especially at an early age—can lead to psychotic disorders in adults, treatment approaches to schizophrenia patients with marijuana dependence or for emerging marijuana-related psychosis are still sorely lacking.66 Any medication approach will likely not deliver its promise without the proper supportive and psychotherapeutic environment. Although medications like naltrexone or rimonabant may be applicable to the treatment of marijuana dependence in patients with schizophrenia as these might be less likely to exacerbate psychosis, they remain to be tested. Clozapine offers the best promise thus far among antipsychotics to mitigate both psychosis and marijuana misuse, both in individuals with schizophrenia and in patients with incipient psychosis due to marijuana. The difficulties of using clozapine have reduced its acceptability to patients and still pose a major hurdle to its more widespread use. Alternatives to clozapine that preserve its benefits and shed its severe liabilities are being actively sought after. Much more work is necessary to address the issue of marijuana-related psychosis, especially in light of the risk of serious mental illness posed by marijuana use in adolescents. Intervening early to stop marijuana use, as with overall drug use in the US, must remain a public health priority and may represent a unique and significant preventative measure to preserve good mental health in individuals at risks. PP


1.    Global Illicit Drug Trends 2002. New York, NY: United Nations Office for Drug Control and Crime Prevention; 2002.
2.    Tanda G, Pontieri FE, Di Chiara G. Cannabinoid and heroin activation of mesolimbic dopamine transmission by a common mu-1 opioid receptor mechanism. Science. 1997;276(5321):2048-2050.
3.    Budney AJ, Moore BA, Vandrey RG, Hughes JR. The time course and significance of cannabis withdrawal. J Abnorm Psychol. 2003;112(3):393-402.
4.    Aldington S, Harwood M, Cox B, et al. Cannabis use and risk of lung cancer: a case-controlled study. Eur Respir J. 2008;31(2):280-286.
5.    Arseneault L, Cannon M, Witton J, Murray RM. Causal association between cannabis and psychosis: examination of the evidence. Br J Psychiatry. 2004;84:110-117.
6.    SAMHSA. Office of Applied Studies. The NSDUH Report:  Age at First Use of Marijuana and Past Year Serious Mental Illness. Available at: http://oas.samhsa.gov/2k5/MJageSMI/MJageSMI.cfm. Accessed February 26, 2009.
7.    Robins LN, Regier DA. Psychiatric Disorders in America: The Epidemiologic Catchment Area Study. New York, NY: The Free Press; 1991.
8.    Tien AY, Anthony JC. Epidemiological analysis of alcohol and drug use as risk factors for psychotic experiences. J Nerv Ment Dis. 1990;178(8):473-480.
9.    Halikas JA, Goodwin DW, Guze SB. Marijuana use and psychiatric illness. Arch Gen Psychiatry. 1972;27(2):162-165.
10.    Emrich HM, Leweke FM, Schneider U. Towards a cannabinoid hypothesis of schizophrenia: cognitive impairment due to dysregulation of the endogenous cannabinoid system. Pharmacol Biochem Behav. 1997;56(4):803-807.
11.    Williams JH, Wellman NA, Rawlins JN. Cannabis use correlates with schizotypy in healthy people. Addiction. 1996;91(6):869-877.
12.    Kwapil TR. A longitudinal study of drug and alcohol use by psychosis-prone and impulsive-nonconforming individuals. J Abnorm Psychol. 1996;105(1):114-123.
13.    Skosnick PD, Spatz-Glen L, Parks S. Cannabis use is associated with schizotypy and attentional disinhibition. Schizophr Res. 2001;48(1):83-92.
14.    Nunn J, Rizza F, Peters ER. The incidence of schizotypy among cannabis and alcohol users. J Nerv Ment Dis. 2001;189(11):741-748.
15.    Verdoux H, Sorbara F, Gindre C, Swenden D, van Os J. Cannabis and dimensions of psychosis in a non-clinical population of female students. Schizophr Res. 2003;59(1):77-84.
16.    Johns A. Psychiatric effects of cannabis. Br J Psychiatry. 2001;178:116-122.
17.    Tramèr MR, Carroll D, Campbell FA, Reynolds DJ, Moore RA, McQuay HJ. Cannabinoids for control of chemotherapy induced nausea and vomiting: quantitative systematic review. BMJ. 2001;323(7303):16-21.
18.    Verdoux H, Gindre C, Sorbara F, Tournier M, Swendsen J. Effects of cannabis use and psychosis vulnerability in daily life: an experience sampling test study. Psychol Med. 2003;33(1):23-32.
19.    Ashton C. Pharmacology and effects of cannabis: a brief review. Br J Psychiatry. 2001;178:101-106.
20.    Hall W, Degenhardt L. Is there a specific “cannabis psychosis”? In: Murray R, Castle D, eds. Marijuana and Madness. New York, NY: Cambrige University Press; 2004:89-100.
21.    International Statistical Classification of Diseases and Health Related Problems. 10th rev. 2nd ed. Geneva, Switzerland: World Health Organization; 2004.
22.    Hall W, Degenhardt L. Cannabis use and psychosis: a review of clinical and epidemiological evidence. Aust N Z J Psychiatry. 2000;34(1):26-34.
23.    Van Os J, Bak M, Bijl RV, De Graaf R, Verdoux H. Cannabis use and psychosis: a longitudinal population-based study. Am J Epidemiol. 2002;156(4):319-327.
24.    Susser M. What is a cause and how do we know one? A grammar for pragmatic epidemiology. Am J Epidemiol. 1991;133(7):635-648.
25.    Andreasson S, Allebeck P, Engstrom A, Rydberg U. Cannabis and schizophrenia: a longitudinal study of Swedish conscripts. Lancet. 1987;2(8574):1483-1485.
26.    Zammit S, Allebeck P, Andreasson S, Lundberg I, Lewis G. Self-reported cannabis use as a risk factor for schizophrenia: further analysis of the 1969 Swedish conscript cohort. BMJ. 2002;325(7374):1199-1201.
27.    Arseneault L, Cannon M, Poulton R, Murray R, Caspi A, Moffitt TE. Cannabis use in adolescence and risk for adult psychosis: longitudinal prospective study. BMJ. 2002;325(7374):1212-1213.
28.    Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Association; 1994.
29.    Pistis M, Perra S, Pillolla G, Melis M, Muntoni AL, Gessa GL. Adolescent exposure to cannabinoids induces long-lasting changes in the response to drugs of abuse in rat midbrain dopamine neurons. Biol Psychiatry. 2004;56(2):86-94.
30.    Schneider, M. Puberty as a highly vulnerable developmental period for the consequences of cannabis exposure. Addict Biol. 2008;13(2):253-263.
31.    Psychoyos D, Hungund B, Cooper T, Finnell RH. A cannabinoid analogue of delta(9)-tetrahydrocannabinol disrupts neural development in chick. Birth Defects Res B Dev Reprod Toxicol. 2008;83(5):477-488.
32.    Semple DM, McIntosh AM, Lawrie SM. Cannabis as a risk factor for psychosis: systematic review. J Psychopharmacol. 2005;19(2):187-194.
33.    Arendt M, Rosenberg R, Foldager L, Perto G, Munk-Jorgensen P. Cannabis-induced psychosis and subsequent schizophrenia spectrum disorders: follow-up study of 535 incident cases. Br J Psychiatry. 2005;187:510-515.
34.    Ferdinand RF, van der Ende J, Bongers I, Selten JP, Huizink A, Verhulst FC. Cannabis-psychosis pathway independent of other types of psychopathology. Schizophr Res. 2005;79(2-3):289-295.
35.    Arendt M, Mortensen PB, Rosenberg R, Pedersen CB, Waltoft BL. Familial predisposition for psychiatric disorder: comparison of subjects treated for cannabis-induced psychosis and schizophrenia. Arch Gen Psychiatry. 2008;65(11):1269-1274.
36.    Corcoran CM, Kimhy D, Stanford A, et al. Temporal association of cannabis use with symptoms in individuals at clinical high risk for psychosis. Schizophr Res. 2008;106(2-3):286-293.
37.    Caspi A, Moffitt TE, Cannon M, et al. Moderation of the effect of adolescent-onset cannabis use on adult psychosis by a functional polymorphism in the catechol-O-methyltransferase gene: longitudinal evidence of a gene x environment interaction. Biol Psychiatry. 2005;57(10):1117-1127.
38.    Agrawal A, Wetherill L, Dick DM, et al. Evidence for association between polymorphisms in the cannabinoid receptor 1 (CNR1) gene and cannabis dependence. Am J Med Genet B Neuropsychiatr Genet. 2008 Nov 14 [epub ahead of print].
39.    De Marchi N, De Petrocellis L, Orlando P, Daniele F, Fezza F, Di Marzo V. Endocannabinoid signalling in the blood of patients with schizophrenia. Lipids Health Dis. 2003;2:5.
40.    Leweke FM, Giuffrida A, Wurster U, Emrich HM, Piomelli D. Elevated endogenous cannabinoid in schizophrenia. Neuroreport. 1999;10(8):1665-1669.
41.    Giuffrida A, Leweke FM, Gerth CW, et al. Cerebrospinal anandamide levels are elevated in acute schizophrenia and are inversely correlated with psychotic symptoms. Neuropsychopharmacology. 2004;29(11):2108-2114.
42.    Zvolensky MJ, Lewinsohn, P, Bernstein A, et al. Prospective association between cannabis use, abuse, and dependence and panic attacks and disorder. J Psychiatr Res. 2008;42(12):1017-1023.
43.    Owen RR, Fischer EP, Booth BM, Cuffel BJ. Medication noncompliance and substance abuse among patients with schizophrenia. Psychiatr Serv. 1996;47(8):853-858.
44.    Swartz MS, Wagner HR, Swanson JW, et al. Substance use in persons with schizophrenia: baseline prevalence and correlates from the NIMH CATIE study. J Nerv Mental Dis. 2006;194(3):164-172.
45.    Abram KM, Teplin LA. Co-occurring disorders among mentally ill jail detainees. Implications for public policy. Am Psychol. 1991;46(10):1036-1045.
46.    Goodman LA, Salyers MP, Mueser KT, et al. Recent victimization in women and men with severe mental illness: prevalence and correlates. J Trauma Stress. 2001;14(4):615-632.
47.    Drake RE, Mueser KT, Brunette MF, McHugo GJ. A review of treatments for people with severe mental illness and co-occurring substance use disorders. Psychiatr Rehabil J. 2004;27(4):360-374.
48.    Potvin S, Stip E, Roy JY. Clozapine, quetiapine, and olanzepine among addicted schizophrenic patients: towards testable hypotheses. Int Clin Psychopharmacol. 2003;18(3):121-132.
49.    Mueser KT, Noordsy D, Drake RE, Fox M. Integrated Treatment for Dual Disorders: A Guide to Effective Practice. New York, NY: Guilford Press; 2003.
50.    Bellack AS, Bennett ME, Gearon JS, Brown CH, Yang Y. A randomized clinical trial of a new behavioral treatment for drug abuse with people with severe and persistent mental illness. Arch Gen Psychiatry. 2006;63(4):426-432.
51.    Weiss RD, Griffin ML, Kolodziej ME, et al. A randomized trial of integrated group therapy versus group drug counseling for patients with bipolar disorder and substance dependence. Am J Psychiatry. 2007;164(1):100-1007.
52.    Drebing CE, Van Ormer EA, Krebs C, et al. The impact of enhanced incentives on vocational rehabilitation outcomes in dually diagnosed veterans. J Appl Behav Anal. 2005;38(3):359-372.
53.    Brunette MF, Mueser KT, Drake RE. A review of research on residential programs for people with severe mental illness and co-occurring substance use disorders. Drug Alcohol Rev. 2004;23(4):471-481.
54.    Justinova Z, Munzar P, Panlilio LV, et al. Blockade of THC-seeking behavior and relapse in monkeys by the cannabin oid CB(1)-receptor antagonist rimonabant. Neuropsychopharmacology. 2008;33(12):2870-2877.
55.    Haney M. Opioid antagonism of cannabinoid effects: differences between marijuana smokers and non-marijuana smokers. Neuropsychopharmacology. 2007;32(6):1391-1403.
56.    Elkashef A, Vocci F, Huestis M, et al. Marijuana neurobiology and treatment. Subst Abus. 2008;29(3):17-29.
57.    Brady K, Anton R, Ballenger JC, Lydiard RB, Adinoff B, Selander J. Cocaine abuse among schizophrenic patients. Am J Psychiatry. 1990;147(9):1164-1167.
58.    Svensson TH, Mathé JM, Andersson JL, Nomikos GG, Hildebrand BE, Marcus M. Mode of action of atypical neuroleptic in relation to the phencyclidine model of schizophrenia: role of 5-HT2 receptor and alpha-1 adrenoreceptor antagonism. J Clin Psychopharmacol. 2005;15(1 suppl 1):11S-18S. Erratum in: J Clin Psychopharmacol. 1995;15(2):154.
59.    Raby WN, Levin, FR, Nunes EV. Pharmacological treatment of substance abuse disorders. In: Tasman A, Kay J, Lieberman JA, First MB, Maj M, eds. Psychiatry. 3rd ed. West Sussex, UK: Wiley-Blackwell; 2008:2390-2412.
60.    Green AI, Zimmet SV, Strous RD, Schildkraut JJ. Clozapine for co morbid substance use disorder and schizophrenia: Do patients with schizophrenia have a reward-deficiency syndrome that can be ameliorated by clozapine? Harvard Rev Psychiatry. 1999;6(6):287-296.
61.    Roth RM, Brunette MF, Green AI. Treatment of substance use disorders in schizophrenia: a unifying neurobiological mechanism? Curr Psychiatry Rep. 2005;7(4):283-291.
62.    Buckley P, Thompson P, Way L, Meltzer HY. Substance abuse among patients with treatment-resistant schizophrenia: characteristics and implications for clozapine therapy. Am J Psychiatry. 1994;151(3):385-389.
63.    Drake RE, Xie H, McHugo GJ, Green AI. The effects of clozapine on alcohol and drug use disorders among patients with schizophrenia. Schizophr Bull. 2000;26(2):441-449.
64.    Brunette M, Drake RE, Xie H, McHugo GJ, Green AI. Clozapine use and relapses of substance use disorder among patients with co-occurring schizophrenia and substance use disorders. Schizophr Bull. 2006;32(4):637-643.
65.    Green AI, Noordsy DL, Brunette MF, O’Keefe C. Substance abuse and schizophrenia: pharmacotherapeutic interventions. J Subst Abuse Treat. 2008;34(1):61-71.
66.    Bodkin LL, Singh A, Corcoran C. Cannabis as a risk factor for psychosis in vulnerable teens: implications for treatment. Primary Psychiatry. 2008;15(1):51-57.



Dr. Compton is assistant professor in the Department of Psychiatry and Behavioral Sciences and Ms. Ramsay is research coordinator of the Atlanta Cohort on the Early Course of Schizophrenia Project, both at Emory University School of Medicine in Atlanta, Georgia.

Disclosures: Dr. Compton receives grant support from the Emory University Research Committee and the National Institute of Mental Health. Ms. Ramsay reports no affiliation with or financial interest in any organization that may pose a conflict of interest.

Please direct all correspondence to: Michael T. Compton, MD, MPH, Assistant Professor, Department of Psychiatry and Behavioral Sciences, Emory University School of Medicine, 49 Jesse Hill Jr. Drive, S.E., Room #333, Atlanta, GA 30303; Tel: 404-778-1486; Fax:  404-616-3241; E-mail: mcompto@emory.edu.


Focus Points

• Cannabis, or marijuana, is a drug that is commonly abused by adolescents and young adults; it is the most frequently abused illicit drug among people with schizophrenia or other psychotic disorders.
• Among people with comorbid substance abuse and schizophrenia or other psychotic disorders, substance use and abuse are typically initiated prior to the overt onset of the psychotic disorder.
• Some research suggests that cannabis use prior to onset may be associated with an earlier age at onset of psychosis, although it is difficult to establish whether this association is causal.
• Athough additional research is needed, preliminary research suggests that cannabis use prior to any psychiatric symptoms may be associated with an earlier age at onset of the prodromal symptoms that commonly precede the onset of schizophrenia.



Schizophrenia is currently conceptualized as an illness that is caused by both genetic predispositions and exposure to stressors or environmental factors, particularly during early childhood and adolescence. This article focuses on one such environmental factor, cannabis use, especially use occurring prior to the onset of clinically evident psychiatric symptoms. Cannabis is commonly abused by adolescents and is the most abused illicit drug in the context of schizophrenia. Several first-episode studies document that the initiation of substance use and abuse typically precedes the onset of psychosis. This article highlights eight studies that characterize the impact of cannabis use on the age at onset of psychosis and three studies that provide early information on the impact of cannabis use on the age at onset of even earlier prodromal symptoms. Future research is needed to better characterize the impact of cannabis use on the onset of psychotic disorders and to determine if cannabis use increases the risk of developing a psychotic disorder, as several other studies suggest. Based on emerging evidence, preventing or reducing cannabis use among adolescents, particularly those at elevated risk of developing psychosis, may delay the onset of psychosis in some.


Schizophrenia is currently conceptualized from the perspectives of the neurodevelopmental and diathesis-stress models.1-3 The neurodevelopmental model integrates altered pre- or perinatal brain development, adolescent developmental abnormalities, and potentially progressive processes that occur after illness onset.3 The diathesis-stress model suggests that symptomatic manifestations of the biologic vulnerability for schizophrenia are influenced by exposure to stressors or environmental factors.4 Based on these conceptualizations, the following points are fairly well accepted among schizophrenia researchers: First, the etiology of the disorder is most likely related to a number of genetic and early environmental risk factors. Second, later risk factors (during adolescence and young adulthood) and neurohormonal changes likely impact the manifestation of underlying vulnerability. Third, genes and environmental risk factors may interact to affect risk. Fourth, the sequential onset of symptoms usually occurs in a gradual  fashion from the premorbid phase to the prodrome to the onset of full psychosis. Fifth, symptom onset, phenomenology, and course are highly heterogeneous. Last, both genetic and environmental factors contribute to this heterogeneity.

This article focuses on one such environmental factor, cannabis use, especially that which occurs prior to the onset of clinically evident psychiatric symptoms. Although adolescent-onset cannabis use has been shown by epidemiologic research to be a risk factor (presumably a causal risk factor, or component cause) for schizophrenia,5-10 the present article examines this environmental factor in terms of its potential to adversely affect two key features of disease onset—age at onset of psychosis and age at onset of even earlier prodromal symptoms. This qualitative summary of the literature is not meant as a systematic review, but as a synthesis of select studies in this area.

The Crossroads of Schizophrenia and Cannabis Use

Cannabis is the most commonly used illicit drug in the United States. According to the 2006 National Survey on Drug Use and Health, 45.4% of Americans ≥12 years of age have tried cannabis at least once.11 Among those ≥18 years of age with lifetime cannabis use, >50% report first using it between 12 and 17 years of age.12,13 Earlier onset of drug use has consistently been associated with greater risk of developing abuse and dependence.12,14-17 Cannabis use disorders occur in approximately 4% of the general US population, with a peak in the 18–29-year age range.15,18,19 Some 56% of those seen in treatment for cannabis abuse/dependence began using by 14 years of age, and 92% began by 18 years of age.13,20 Cannabis use is now considered a substantial public health problem by many, due to several reasons, as noted previously.21 First, US adolescents and young adults have very high rates of cannabis use. Second, cannabis dependence in youths predicts increased risks of using other illicit drugs and underperforming in school.22 Third, the cannabinoid content of smoked cannabis has increased substantially during recent decades,23 potentially resulting in a larger “dose” of psychoactive cannabinoids during drug use.

Unsurprisingly, given the aforementioned high population prevalence of cannabis use, cannabis is the most abused illicit drug in the context of schizophrenia.24 The Epidemiologic Catchment Area study found the lifetime prevalence of a cannabis use disorder in people with schizophrenia to be 19.7%.25 Many studies confirm high rates (20% to 70%) of cannabis use in patients with schizophrenia.26-31 Data from 53 studies of schizophrenia revealed that 12-month prevalence estimates of use and misuse of cannabis were 29% and 19%, and lifetime use and misuse estimates were 42% and 23%, respectively.32 Researchers report rates of cannabis misuse ranging from 15% to 65% in first-episode samples.33-39

Several first-episode studies document that the initiation of substance use and abuse typically precedes the onset of psychosis, often by several years.31,38,40-42 When prodromal symptoms are taken into account, one German study37 of 232 patients with a first episode of psychosis found that 29.5% of those using drugs had a drug problem >1 year before the earliest sign of an emerging psychotic disorder. In an additional 34.6%, drug abuse emerged at the same time as the first symptoms. In a first-episode sample (n=133) from the Netherlands, among those patients who had used cannabis by the time of their first treatment contact, 64.3% reported initiating cannabis use before the onset of social and/or occupational dysfunction and 85.7% before the onset of psychosis.43 In a sample of 109 hospitalized patients with first-episode non-affective psychosis in the US, 79.8% had used cannabis at least once in the years prior to hospitalization (Compton MT, unpublished data, March 2009). While mean ages at the onset of prodromal symptoms and psychotic symptoms in that sample were 19.4±5.3 and 21.8±4.7 years, respectively, the mean age at first use of cannabis among the 87 who had used it was 15.8±4.0. These and other studies indicate a high prevalence of cannabis use occurring prior to the onset of psychiatric symptoms in people who develop a psychotic disorder.

Although numerous studies show that the initiation of substance use commonly precedes the onset of psychosis, this does not necessarily imply a directional or causal association.  It is not surprising that substance use often precedes psychosis given that initiation of substance use usually occurs during adolescence.  However, research establishing that early-course patients typically begin substance use prior to onset confirms temporality (ie, that exposure precedes outcome in a plausible way), which is one criterion in ultimately establishing a causal relationship. This article, which largely focuses on the possibility that pre-onset cannabis use may hasten onset of psychotic disorders, notes as important the substantial research showing that, among patients with comorbidity, substance use often precedes the manifestation of symptoms.

The biologic pathways linking cannabis use and psychosis are being actively studied. Numerous research findings, six of which are briefly described here, may demonstrate the biologic plausibility of pre-onset cannabis use impacting not only vulnerability for developing schizophrenia, but also the age at onset among those who do develop the disorder. First, exogenous cannabinoids (eg, marijuana) are extremely lipid soluble, accumulating in fatty tissues from which they are slowly released back into body compartments, including the brain,23 suggesting that even occasional cannabis use leads to long-term exposure of central receptors to cannabinoids. Second, exogenous and endogenous (eg, anandamide) cannabinoids exert their effects (such as modulating the release of neurotransmitters including glutamate, norepinephrine, and dopamine) by interactions with specific cannabinoid (CB1) receptors44,45 that are distributed in brain regions implicated in the pathophysiology of schizophrenia (including the cerebral cortex, limbic areas, basal ganglia, and thalamus).23 Third, cannabis increases mesolimbic dopaminergic transmission and inhibits glutamatergic release.46 Fourth, several studies have shown an increased CB1 receptor density in brain regions of interest in schizophrenia, including the dorsolateral prefrontal cortex and the anterior cingulate cortex,47-49 and elevated levels of endogenous cannabinoids in the blood and cerebrospinal fluid of patients with schizophrenia.50-52 Fifth, gene variants of the CB1 receptor may be associated with schizophrenia and risk of substance abuse in individuals with schizophrenia.48,53,54 However, other studies have not found an association with risk for schizophrenia,55 and a recent meta-analysis did not implicate these gene variants among 24 showing significant effects.56 Sixth, acute administration of cannabis causes both patients and controls to experience transient increases in cognitive impairments and schizophrenia-like positive and negative symptoms.57 It could be argued that these six points provide only a weak argument for a causal effect of cannabis on hastening onset. For example, the findings of increased CB1 receptor density in regions implicated in schizophrenia are not surprising given that CB1 receptors are relatively widely dispersed.  However, when taken together, these findings do suggest biologic plausibility, which, like temporality, is one criterion for eventually demonstrating causality. 

Having provided some evidence supporting potential biologic plausibility, the remainder of this article focuses on two themes—the potential impact of early-course cannabis use on both the age at onset of psychotic symptoms and the age at onset of even earlier prodromal symptoms. Age at onset of the prodrome and psychosis are critical variables to understand because they are important prognostic factors. An earlier age at onset is associated with a higher degree of cognitive impairment, increased severity of psychosocial and functional disability, more severe symptoms and behavioral deterioration, less responsiveness to antipsychotics, decreased ability to tolerate discontinuation of medication, and greater likelihood of rehospitalization.58-69 Given extensive literature connecting earlier onset with poorer course and outcomes, discovering potentially modifiable determinants of age at onset is crucial. Could pre-onset cannabis use in adolescence be one such determinant?

The Impact of Pre-psychotic Cannabis Use on the Age at Onset of Psychosis

At least eight studies, generally collecting cross-sectional or retrospective information from individuals with a recent onset of psychosis, have examined the potential impact of cannabis use on the age at onset of psychosis. These studies, discussed briefly here, are also summarized in the Table (Compton MT, unpublished data, March 2009).28,34,38,42,43,70,71 Although numerous older studies explored the relationship between substance abuse and psychosis, Hambrecht and Häfner28 were perhaps the first to study the exact timing of the onset of substance use and symptoms in first-episode psychosis. In the Age, Beginning, and Course (ABC) schizophrenia study,28 they found that the mean age at onset of the first negative symptom, first positive symptom, and first admission were lower in the 32% who had abused drugs prior to admission than in those who had not. The mean ages at first positive and negative symptoms were each 5.7 years younger for those with a history of drug abuse than for those with no substance abuse history (21.1 compared to 26.8, and 24.3 compared to 30.0, respectively). Among those who reported a history of drug abuse, the mean age at onset of drug abuse was 18.6 years, or 1.5 and 5.7 years before the mean age at onset of the first negative and first positive symptoms. However, the analysis of age at onset was not restricted to those who had initiated drug abuse specifically before, rather than during or after, the onset of illness. In addition, the independent effects of particular substances of abuse were not considered (although 90% of those abusing drugs in the sample had abused cannabis, 63% had used other drugs as well). Finally, the study did not examine the impact of drug use before it reached the threshold of drug abuse.
Among a sample of patients in New York State with a first episode of either affective or non-affective psychosis (n=541), symptom onset among men and women was examined by Rabinowitz and colleagues38 in three groups: those with no lifetime substance use disorder diagnosis, those in remission or with mild substance use, and those with current moderate-to-severe substance abuse at the time of admission. Females with current moderate-to-severe substance abuse were 6 years younger at the onset of first psychotic symptoms than their counterparts with no lifetime substance use. No significant impact on the age at onset was observed for males. However, the inclusion of patients with affective as well as non-affective psychoses introduces heterogeneity in the expected age at onset, disease mechanisms, gender distribution, and rates of comorbid substance abuse. Nonetheless, this study is notable for accounting for the severity of substance use.


In the Calgary Early Psychosis Program, which assesses and treats patients with a recent onset of psychosis, 44% of 357 consecutively admitted patients had substance abuse or dependence in the previous year.70 In this sample, Van Mastrigt and colleagues70 found that patients who had misused cannabis (or cannabis and alcohol) were younger and had an earlier age at onset of positive psychotic symptoms than non-users or those who misused alcohol only or alcohol and other drugs. These findings suggest that a link may exist between cannabis use and age at onset of psychosis, and that this effect may be related to the cannabis, per se, as opposed to personality traits or other vulnerabilities that lead to a substance use disorder, given that misuse of other substances did not carry the same association. However, without data on the age at first cannabis use or abuse, or on the onset of prodromal or negative symptoms, it is not possible to establish the directionality of the association between cannabis use and the age at onset of symptoms.

Veen and colleagues43 used an incidence cohort of patients in the Netherlands to examine the independent influences of gender and cannabis use on early course features. The sample (n=133) included natives of the Netherlands, first- and second-generation immigrants from Surinam and Morocco, and individuals from other racial/ethnic groups. Patients who used cannabis (n=70) had an earlier median age at onset compared to the 63 patients not using cannabis. In a multiple regression analysis, male cannabis users (n=55) were found to have had their first psychotic episode a mean of 6.9 years earlier than 37 male nonusers. Cannabis use was a stronger predictor of age at first psychotic episode than gender. However, the study did not control for the effects of family history or the use of other substances (eg, alcohol, cocaine). Furthermore, the study treated cannabis use as categoric/dichotomous variable (which is true of most studies conducted to date) and therefore could not examine potential dose-effect relationships.

Mauri and colleagues42 retrospectively studied 285 first-episode patients in Italy and found that patients abusing cannabis had an earlier age at onset compared with those who did not abuse cannabis, though it is unclear how onset was operationalized. Further, this comparison failed to control for the influence of gender, and only 18% of females had used substances compared to 44% of males. Additionally, much of the data were obtained by retrospectively reviewing medical records (which are presumably less thorough and accurate than formal research assessments), 56% of patients having used drugs were multi-drug abusers (and this apparently was not controlled for), and the amount and duration of substance use was not considered.

In London, Barnes and colleagues34 assessed 152 first-episode patients and found that those reporting past substance use were significantly younger at the onset of psychotic symptoms compared with those who had not used substances. In a linear regression, use of any substances other than cannabis was not significantly related to age at onset, though gender and cannabis use were. The age at onset of psychosis was on average 4.2 years older for women and 5.0 years younger for participants using cannabis, adjusting for the other variables. In this study, like most others, cannabis was the most prevalently used illicit substance, and, therefore, detecting an effect of cannabis may be easier than finding effects of other drugs, given issues of statistical power. Unfortunately, inquiries about past substance use did not include detailed assessment of the frequency and quantity of drugs taken, and it is unclear whether the initiation of cannabis use in fact preceded the onset of symptoms in the patients included in the analysis.

González-Pinto and colleagues71 found, among 131 first-episode patients in Spain, a significant, gradual reduction in age at onset as the level of use of cannabis increased—a decrement of 7, 8.5, and 12 years for patients with cannabis use, abuse, and dependence, respectively. The effect was not explained by the use of other drugs or gender. However, the study included patients with affective psychosis, who would be expected to have a later age at onset, and likely a lower prevalence of comorbid cannabis use. In addition, the study did not take into account the duration of cannabis use and it is not clear exactly how age at onset was operationalized.

Recently, Compton and colleagues (Compton MT, unpublished data, March 2009) examined the impact of prior cannabis use on the age at onset of psychosis in 109 patients hospitalized for a first episode of psychosis. This group found that both daily cannabis and daily tobacco use occurring before onset of psychosis had a significant effect on the risk of onset of psychosis (hazard ratios of 2.0 and 1.8, respectively, P<.05), when the level of frequency of use was treated as a time-dependent covariate in Cox regressions (ie, progression to daily use was associated with a higher risk of onset). Of note, cannabis and tobacco use were highly correlated (eg, having ever used nicotine was highly associated with having ever used cannabis, χ2=25.5, P<.001 [Compton MT, unpublished data, March 2009]) and therefore may not represent two independent risk factors. A gender by progression to daily cannabis use interaction was observed—progression to daily use was related to a much larger increased relative risk for onset of psychosis in females (hazard ratio=5.1) than in males (hazard ratio=3.4). Although this study took into account the frequency of cannabis use (ie, never, ever but not weekly use, weekly but not daily use, or daily use), it did not assess quantity of use and did not gather detailed information on patterns of use.

Impact of Pre-prodromal Cannabis Use on the Age at Onset of the Prodrome

While the previously reviewed studies suggest, through various analytic designs, that cannabis use may be associated with a younger age at onset of psychotic symptoms, only a few groups have attempted to determine if cannabis use is associated with a younger age at onset of prodromal symptoms. The prodrome is the syndromal period commonly comprised of non-specific psychiatric symptoms, emerging attenuated positive symptoms, negative symptoms, and psychosocial decline—commonly lasting several months to a few years—that precedes the emergence of frank psychosis in most patients. One critique of the literature is that the possible influence of cannabis use on prodromal symptoms has not been adequately explored.34 Doing so could shed light on the competing hypotheses that substance use precipitates or hastens onset of the illness versus that very early, subtle symptoms of the illness make patients vulnerable to substance use.

Hambrecht and Häfner28 conducted one of very few studies that included an analysis of prodromal symptoms in relation to substance use. In 232 first-episode patients from the ABC schizophrenia study, they found that the mean age at onset of the first sign was lower in the 32% who had abused drugs prior to admission than in those who had not. First signs included the first negative, positive, or non-specific psychiatric symptom if it occurred continuously until the onset of psychosis. In this way, the “first sign” represented the onset of the prodrome, if a prodrome had occurred, or psychosis, if there had been no prodrome. The age at first-sign onset was 7.2 years younger among those who had abused drugs than among those who had no history of drug or alcohol abuse (18.5 compared to 25.7 years).

In the report by Veen and colleagues,43 the relationship between prior cannabis use and the age at onset of the first sign of social or occupational dysfunction, which could be considered a proxy for the age at onset of the prodrome, was examined. In this cohort, the median age at onset was 18.1 years among patients using cannabis, as compared to 27.7 years in those not using cannabis. However, in a linear regression, gender was a more important predictor of age at onset, and after controlling for this variable, cannabis use was not a significant predictor. This study did not use a precise indicator of the onset of the prodrome, but it is noteworthy for its analysis of the effect of prior cannabis use on onset. Like their analysis of age at onset of psychosis, Veen and colleagues43 did not control for family history or other substance abuse, and they treated cannabis use as a categorical variable.

Similar to their findings pertaining to age at onset of psychosis, Compton and colleagues (Compton MT, unpublished data, March 2009) examined the impact of prior cannabis use on the age at onset of illness/prodromal symptoms in 109 hospitalized first-episode patients. When considering the level of frequency of use as a time-dependent covariate in Cox regressions, both daily cannabis and tobacco use had a significant effect on the risk of onset of the symptoms (which represented the onset of the prodrome in 70% of the sample), hastening onset (hazard ratios=2.1 and 1.8, respectively). As noted above, although this analysis accounted for the frequency of cannabis use, other methodologic limitations make the results preliminary, requiring further research.

Discussion and Unanswered Questions

In summary, several studies suggest that cannabis use among first-episode patients prior to onset may be associated with an earlier age at onset of positive psychotic symptoms. Much less is known about potential associations between pre-prodromal cannabis use and the onset of prodromal symptoms. The Figure depicts hypothesized symptom development and the accumulation of functional impairment in the early course of schizophrenia in patients with a history of cannabis use compared to those without a history of cannabis use. Further research is needed to show whether pre-onset cannabis use is in fact an independent risk factor for developing a psychotic disorder or for an earlier emergence of symptoms among those who do develop a disorder. Support for the psychotogenic properties of cannabis during a prodromal period comes from the finding that perceptual disturbances fluctuate over time with cannabis use in a clinical high-risk cohort.72


As stated above, an earlier age at onset of psychosis is a poor prognostic indicator. If further research proves a link between adolescent, pre-onset cannabis use and age at onset, then decreasing cannabis use among adolescents may delay the onset of psychosis among those destined to develop a psychotic disorder. Some have argued that there now exists sufficient evidence to inform the public that using cannabis could increase the risk of developing a psychotic illness73; this may be especially relevant for at-risk groups. Programs to decrease cannabis use may be particularly beneficial in adolescents identified as being at very high risk for psychosis, by virtue of a positive family history or by being identified as potentially prodromal or at “ultra-high risk” based on emerging psychiatric symptoms and functional decline. Just as decreasing cannabis use has been suggested as a potential preventive intervention to reduce the incidence of schizophrenia,74,75 reducing cannabis use also could delay onset among those who do, nonetheless, develop the disorder.76

Numerous unanswered questions should be the focus of ongoing research. First, given the high comorbidity between cannabis abuse and the abuse of other addictive substances, especially nicotine and alcohol,77,78 the independent effects of pre-onset cannabis use, as well as pre-onset nicotine, alcohol, and other drug use, should be examined. Although the neurobiologic feasibility of the cannabis/psychosis link was pointed out above, it must be recognized that alcohol and other drugs impact upon dopaminergic, glutamatergic, and other neurotransmitter systems affected by schizophrenia. Regarding potential effects of alcohol and cannabis, for example, on psychosis risk, the same neural structures are indeed affected by both substances, at least at a global level (mesolimbic dopamine pathways and the central reward system in general), though alcohol and cannabis exert their effects partly through different receptors (ie, g-aminobutyric acid-ergic/glutamatergic and cannabinoid receptors, respectively). Some evidence suggests that alcohol may exert modulatory actions in the endogenous cannabinoid system.79 In addition to cannabis and alcohol, nicotine use is also a critical variable to examine for a couple of reasons: Cigarette smoking is highly prevalent in individuals with schizophrenia80,81 (even in those with first-episode psychosis),82 and there is increasing interest in the literature in biologic links between the central nicotinic system and schizophrenia.83-85 However, though such elucidation of single-drug effects would be beneficial, it must be emphasized, as noted above, that the independent effects of each substance will be difficult to parse given the high degree of comorbidity across substances, especially cannabis, alcohol, and nicotine. Relatively large sample sizes likely will be necessary to examine independent effects.

A second direction for future research pertains to the issue of causality versus association. Like any observational study, the studies described here cannot rule out the possibility of reverse causality, in which the disease processes associated with the later development of psychosis render an individual more susceptible to the initiation of substance use and abuse earlier in life. Even if further research indicates that pre-onset cannabis use is associated with an earlier age at onset, sorting out whether the cannabis use causes an earlier onset or is a marker of a disease process or subtype associated with earlier onset will be challenging. Similar questions pertain to the link between cannabis use and risk of developing schizophrenia—causality remains difficult to prove, and a shared diathesis for both psychotic illnesses and substance abuse may be at play.

Third, regarding a potential impact of cannabis use on prodromal symptomatology, it is possible that cannabis use causes prodrome-like (but not definitively prodromal) symptoms in patients who later develop a psychotic disorder. That is, cannabis use in adolescence among patients who later develop a psychotic disorder may lead to apathy, academic problems, and other prodromal-appearing difficulties, though such problems may not necessarily be inherent to the schizophrenia process or they may not alter the course of the developing psychotic disorder in a way that conveys long-term course implications. A number of studies suggest that cannabis use is associated with schizotypal features in people who may or may not develop schizophrenia.21 Furthermore, in a recent study86 involving 6,330 adolescents (15–16 years of age) in a Finnish prospective birth cohort, the 356 (5.6%) participants who had used cannabis endorsed a higher mean number of “prodromal” symptoms, and a dose-response relationship was evident. However, actual prodromal symptoms can only be confirmed retrospectively; it remains to be determined through further longitudinal research whether or not the symptoms assessed in that study actually represented a prodrome.

A fourth important issue requiring further, more methodologically rigorous research, relates to the measurement of substance use in relation to ages at onset of prodromal and psychotic symptoms. As noted above, not all prior studies commented on (in fact, most did not) or restricted the analyses to those whose drug use/abuse preceded the onset of psychotic and/or prodromal symptoms,28,34,70,71 which is critical to the research question. To advance the field in this area, future research should carefully examine the timing of the initiation of cannabis use and the development of psychotic symptoms as well as earlier prodromal symptoms using well-defined operationalizations of onset. Such retrospective measurement is admittedly a difficult task. Comprehensive assessments of cannabis and other substance use with reliable and valid retrospective measures that incorporate calendars, timelines, and significant life events should be used to gather data on amount, duration, frequency, and patterns of use, thus allowing for the examination of temporal relationships and potential dose effects. Additionally, because the limited research in this area to date has generally consisted of cross-sectional and retrospective studies, other research designs, including case-control and longitudinal studies, would be beneficial.

Fifth, relevant covariates, including gender and family history, must be examined when studying age at onset. The fact that gender is a predictor of age at onset is one of the most consistent findings in schizophrenia research.87,88 For example, results from the ABC schizophrenia study indicate that women are 3–4 years older than men at illness onset, as defined by the onset of positive symptoms, negative symptoms, or psychosocial impairment.89,90 Family history of schizophrenia is also associated with an earlier age at onset,91-94 and should be assessed as a covariate. It should be noted, however, that even if future studies are more rigorous, it may still be difficult to establish with certainty that cannabis use hastens onset and that all relevant covariates have been taken into account. Confounding (the distortion of an apparent effect of cannabis use on risk brought about by an association with other significant risk factors), must be seriously considered.

These five issues, among others, suggest a need for further research to substantiate the early reports that pre-onset cannabis use, typically occurring in adolescence, may be associated with (and perhaps even causative of) an earlier age at onset. This line of research—in addition to ongoing research on the neurobiologic interface between cannabinoid systems and the neurocircuitry involved in schizophrenia, cannabis use as a potential component cause of schizophrenia, and the influence of cannabis use on symptom and neurocognitive profiles—may advance the field in terms of both further elucidating psychotic disorders and informing future preventive interventions.


Several first-episode studies document that the initiation of substance use and abuse typically precedes the onset of psychosis, often by several years. Studies reviewed also suggest that cannabis use among first-episode patients prior to onset may be associated with an earlier age at onset of positive psychotic symptoms. Much less is known about potential associations between pre-prodromal cannabis use and the onset of prodromal symptoms, though preliminary evidence suggests that an association may be present.  Future research should examine the effects of cannabis, independent of other substances used; establish the causal direction of these associations; clarify whether prodrome-like symptoms observed during concurrent cannabis abuse are, indeed, related to the subsequent psychosis; include more rigorous research design; and control for all significant covariates such as gender and family history. PP


1.    Keshavan MS, Gilbert AR, Diwadkar VA. Neurodevelopmental theories. In: Lieberman JA, Stroup TS, Perkins DO, eds. Textbook of Schizophrenia. Washington, DC: American Psychiatric Publishing, Inc.; 2006:69-83.
2.    Mueser KT, McGurk SR. Schizophrenia. Lancet. 2004;363(9426):2063-2072.
3.    Walker EF. Developmentally moderated expressions of the neuropathology underlying schizophrenia. Schizophr Bull. 1994;20(3):453-480.
4.    Walker EF, Diforio D. Schizophrenia: a neural diathesis-stress model. Psychol Rev. 1997;104(4):667-685.
5.    Andréasson S, Allebeck P, Rydberg U. Schizophrenia in users and nonusers of cannabis. A longitudinal study in Stockholm County. Acta Psychiatr Scand. 1989;79(5):505-510.
6.    Arsenault L, Cannon M, Witton J, Murray R. Cannabis as a potential causal factor in schizophrenia. In: Castle DJ, Murray R, eds. Marijuana and Madness. Cambridge, MA: Cambridge University Press; 2004:101-118.
7.    Semple DM, McIntosh AM, Lawrie SM. Cannabis as a risk factor for psychosis: systematic review. J Psychopharmacol. 2005;19(2):187-194.
8.    Smit F, Boiler L, Cuijpers P. Cannabis use and the risk of later schizophrenia: a review. Addiction. 2004;99(4):425-430.
9.    Weiser M, Noy S. Interpreting the association between cannabis use and increased risk for schizophrenia. Dialogues Clin Neurosci. 2005;7(1):81-85.
10.    Zammit S, Allebeck P, Andréasson S, Lundberg I, Lewis G. Self reported cannabis use as a risk factor for schizophrenia in Swedish conscripts of 1969: historical cohort study. BMJ. 2002;325(7374):1199-1203.
11.    Substance Abuse and Mental Health Services Administration (SAMHSA). Office of Applied Studies. 2006 National Survey on Drug Use and Health: Detailed Tables. Available at: http://oas.samhsa.gov/NSDUH/2k6NSDUH/tabs/Sect1peTabs1to46.htm#Tab1.19B. Accessed February 20, 2009.
12.    Substance Abuse and Mental Health Services Administration (SAMHSA). Results from the 2004 National Survey on Drug Use and Health: National Findings. NSDUH Series H-25. DHHS Pub. No. 04-3964. Rockville, MD: SAMHSA; 2005.
13.    National Institute on Drug Abuse (NIDA). InfoFacts. Marijuana. Available at: www.drugabuse.gov/PDF/InfoFacts/Marijuana06.pdf. Accessed February 20, 2009.
14.    Anthony JC, Petronis KR. Early-onset drug use and risk of later drug problems. Drug Alcohol Depend. 1995;40(1):9-15.
15.    Compton WT, Grant BF, Colliver JD, Glantz MD, Stinson FS. Prevalence of marijuana use disorders in the United States: 1991-1992 and 2001-2002. JAMA. 2004;291(17):2114-2121.
16.    Grant BF, Dawson DA. Age at onset of drug use and its association with DSM-IV drug abuse and dependence: results from the National Longitudinal Alcohol Epidemiologic Survey. J Subst Abuse. 1998;10(2):163-173.
17.    Lynskey MT, Heath AC, Bucholz KK, et al. Escalation of drug use in early-onset cannabis users vs co-twin controls. JAMA. 2003;289(4):427-433.
18.    Robins LN, Regier DA. Psychiatric Disorders in America. New York, NY: Free Press; 1991.
19.    Kessler RC, McGonagle KA, Zhao S, et al. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Arch Gen Psychiatry. 1994;51(1):8-19.
20.    Substance Abuse and Mental Health Services Administration (SAMHSA). Treatment Episode Data Set (TEDS). Highlights – 2003. National Admissions to Substance Abuse Treatment Services. DASIS Series: S-27, DHHS Publication No. (SMA 05-4043). Rockville, MD: SAMHSA; 2005.
21.    Compton MT, Goulding SM, Walker EF. Cannabis use, first-episode psychosis, and schizotypy: a summary and synthesis of recent literature. Cur Psychiatry Rev. 2007;3(3):161-171.
22.    Hall WD. Cannabis use and the mental health of young people. Austr N Z J Psychiatry. 2006;40(2):105-113.
23.    Ashton CH. Pharmacology and effects of cannabis: a brief review. Br J Psychiatry. 2001;178(2):101-106.
24.    Mueser KT, Yarnold PR, Levinson DF, et al. Prevalence of substance abuse in schizophrenia: demographic and clinical correlates. Schizophr Bull. 1990;16(1):31-56.
25.    Regier DA, Farmer ME, Rae DS, et al. Comorbidity of mental disorders with alcohol and other drug abuse: results from the Epidemiological Catchment Area (ECA) Study. JAMA. 1990;264(19):2511-2518.
26.    Bersani G, Orlandi V, Kotzalidis GD, Pancheri P. Cannabis and schizophrenia: impact on onset, course, psychopathology and outcomes. Eur Arch Psychiatry Clin Neurosci. 2002;252(2):86-92.
27.    Dubertret C, Bidard I, Adès J, Gorwood P. Lifetime positive symptoms in patients with schizophrenia and cannabis abuse are partially explained by co-morbid addiction. Schizophr Res. 2006;86(1-3):284-290.
28.    Hambrecht M, Häfner H. Substance abuse and the onset of schizophrenia. Biol Psychiatry. 1996;40(11):1155-1163.
29.    Hides L, Dawe S, Kavanagh DJ, Young RM. Psychotic symptom and cannabis relapse in recent-onset psychosis. Br J Psychiatry. 2006;189(2):137-143.
30.    Linszen DH, Dingemans PM, Lenior ME. Cannabis abuse and the course of recent onset schizophrenic disorders. Arch Gen Psychiatry. 1994;51(4):273-279.
31.    Sevy S, Robinson DG, Holloway S, et al. Correlates of substance abuse in patients with first-episode schizophrenia and schizoaffective disorder. Acta Psychiatr Scand. 2001;104(5):367-374.
32.    Green B, Young R, Kavanagh D. Cannabis use and misuse prevalence among people with psychosis. Br J Psychiatry. 2005;187(4):306-313.
33.    Addington J, Addington D. Impact of an early psychosis program on substance use. Psychiatr Rehab J. 2001;25(1):60-68.
34.    Barnes TR, Mutsatsa SH, Hutton SB, Watt HC, Joyce EM. Comorbid substance use and age at onset of schizophrenia. Br J Psychiatry. 2006;188(3):237-242.
35.    Compton MT, Furman AC, Kaslow NJ. Lower negative symptom scores among cannabis-dependent patients with schizophrenia-spectrum disorders: preliminary evidence from an African American first-episode sample. Schizophr Res. 2004;71(1):61-64.
36.    Cantwell R, Brewin J, Glazebrook C, et al. Prevalence of substance misuse in first-episode psychosis. Br J Psychiatry. 1999;174(2):150-153.
37.    Hambrecht M, Häfner, H. Cannabis, vulnerability, and the onset of schizophrenia: an epidemiological perspective. Austr N Z J Psychiatry. 2000;34(3):468-475.
38.    Rabinowitz J, Bromet EJ, Lavelle J, Carlson G, Kovasznay B, Schwartz JE. Prevalence and severity of substance use disorder and onset of psychosis in first-admission psychotic patients. Psychol Med. 1998;28(6):1411-1419.
39.    Wade D, Harrigan S, McGorry PD, Burgess PM, Whelan G. Impact of severity of substance use disorder on symptomatic and functional outcome in young individuals with first-episode psychosis. J Clin Psychiatry. 2007;68(5):767-774.
40.    Silver H, Abboud E. Drug abuse in schizophrenia: comparison of patients who began drug abuse before their first admission with those who began abusing drugs after their first admission. Schizophr Res. 1994;13(1):57-63.
41.    Bühler B, Hambrecht M, Löffler W, an der Heiden W, Häfner H. Precipitation and determination of the onset and course of schizophrenia by substance abuse—a retrospective and prospective study of 232 population-based first illness episodes. Schizophr Res. 2002;54(3):243-251.
42.    Mauri M, Volonteri L, De Gaspari I, Colasanti A, Brambilla M, Cerruti L. Substance abuse in first-episode schizophrenic patients: a retrospective study. Clin Pract Epidemiol Ment Health. 2006;2:4-12.
43.    Veen ND, Selten JP, van der Tweel I, Feller WG, Hoek HW, Kahn RS. Cannabis use and age at onset of schizophrenia. Am J Psychiatry. 2004;161(3):501-506.
44.    Devane WA, Dysarz FA 3rd, Johnson MR, Melvin LS, Howlett AC. Determination and characterization of a cannabinoid receptor in rat brain. Mol Pharmacol. 1988;34(5):605-613.
45.    Iversen L. Cannabis and the brain. Brain. 2003;126(6):1252-1270.
46.    Zammit S, Lewis G. Exploring the relationship between cannabis use and psychosis. Addiction. 2004;98(11):1353-1355.
47.    Dean B, Sundram S, Bradbury R, Scarr E, Copolov D. Studies on [3H]CP-55940 binding in the human central nervous system: regional specific changes in density of cannabinoid-1 receptors associated with schizophrenia and cannabis use. Neurosci. 2001;103(1): 9-15.
48.    Sundram S, Dean B, Copolov D. The endogenous cannabinoid system in schizophrenia. In: Castle DJ, Murray R, eds. Marijuana and Madness. Cambridge, MA: Cambridge University Press; 2004:127-141.
49.    Zavitsanou K, Garrick T, Huang XF. Selective antagonist [3H]SR141716A binding to cannabinoid CB1 receptors is increased in the anterior cingulate cortex in schizophrenia. Progress Neuro-Psychopharmacol Biol Psychiatry. 2004;28(2):355-360.
50.    De Marchi N, De Petrocellis L, Orlando P, Daniele F, Fezza F, Di Marzo V. Endocannabinoid signaling in the blood of patients with schizophrenia. Lipids Health Dis. 2003;2(5):5-14.
51.    Giuffrida A, Leweke FM, Gerth CW, et al. Cerebrospinal anandamide levels are elevated in acute schizophrenia and are inversely correlated with psychotic symptoms. Neuropsychopharmacology. 2004;29(11):2108-2114.
52.    Leweke FM, Giuffrida A, Wurster U, Emrich HM, Piomelli D. Elevated endogenous cannabinoids in schizophrenia. Neuroreport. 1999;10(8):1665-1669.
53.    Leroy S, Griffon N, Bourdel MC, Olié JP, Poirier MF, Krebs MO. Schizophrenia and the cannabinoid receptor type 1 (CB1): association study using a single-base polymorphism in coding exon 1. Am J Med Genet (Neuropsychiatr Genet). 2001;105(8):749-752.
54.    Ujike H, Takaki M, Nakata K, et al. CNR1, central cannabinoid receptor gene, associated with susceptibility to hebephrenic schizophrenia. Mol Psychiatry. 2002;7(5):515-518.
55.    Zammit S, Spurlock G, Williams H, Norton N, Williams N, O’Donovan M, Owen MJ. Genotype effects of CHRNA7, CNR1 and COMT in schizophrenia: interactions with tobacco and cannabis use. Br J Psychiatry. 2007;191(5):402-407.
56.    Allen NC, Bagade S, McQueen MB, et al. Systematic meta-analyses and field synopsis of genetic association studies in schizophrenia: the SzGene database. Nat Genet. 2008;40(7):827-834.
57.    D’Souza DC, Abi-Saab WM, Madonick S, et al. Delta-9-tetrahydrocannabinol effects in schizophrenia: implications for cognition, psychosis, and addiction. Biol Psychiatry. 2005;57(6):594-608.
58.    Johnstone EC, Owens DG, Bydder GM, Colter N, Crow TJ, Frith CD. The spectrum of structural brain changes in schizophrenia: age at onset as a predictor of cognitive and clinical impairments and their cerebral correlates. Psycholog Med. 1989;19(1):91-103.
59.    Mayer C, Kelterborn G, Naber D. Age at onset in schizophrenia: relations to psychopathology and gender. Br J Psychiatry. 1993;162(5):665-671.
60.    Häfner H, Nowotny B. Epidemiology of early-onset schizophrenia. Eur Arch Psychiatry Clin Neurosci. 1995;245(2):80-92.
61.    Sharma RP, Dowd SM, Davis JM, Janicak PG. Age of illness onset and schizophrenic symptomatology during an inpatient washout period. Schizophr Res. 1996;20(3):295-300.
62.    Jeste DV, McAdams LA, Palmer BW, et al. Relationship of neuropsychological and MRI measures to age at onset of schizophrenia. Acta Psychiatr Scand. 1998;98(2):156-164.
63.    Banaschewski T, Schulz E, Martin M, Remschmidt H. Cognitive functions and psychopathological symptoms in early-onset schizophrenia. Eur Child Adolesc Psychiatry. 2000;9(1):11-20.
64.    Lay B, Blanz B, Hartmann M, Schmidt MH. The psychosocial outcome of adolescent-onset schizophrenia: a 12-year followup. Schizophr Bull. 2000;26(4):801-816.
65.    Tuulio-Henriksson A, Partonen T, Suvisaari J, Haukka J, Lonnqvist J. Age at onset and cognitive functioning in schizophrenia. Br J Psychiatry. 2004;185(3):215-219.
66.    Rhinewine JP, Lencz T, Thaden EP, et al. Neurocognitive profile in adolescents with early-onset schizophrenia: clinical correlates. Biol Psychiatry. 2005;58(9):705-712.
67.    Ropcke B, Eggers C. Early-onset schizophrenia: a 15-year follow-up. Eur Child Adolesc Psychiatry. 2005;14(6):341-350.
68.    Amminger GP, Leicester S, Yung AR, et al. Early-onset of symptoms predicts conversion to non-affective psychosis in ultra-high risk individuals. Schizophr Res. 2006;84(1):67-76.
69.    Malla A, Norman R, Schmitz N, et al. Predictors of rate and time to remission in first-episode psychosis: a two-year outcome study. Psychol Med. 2006;36(5):649-658.
70.    Van Mastrigt S, Addington J, Addington D. Substance misuse at presentation to an early psychosis program. Soc Psychiatry Psychiatr Epidemiol. 2004;39(1):69-72.
71.    González-Pinto A, Vega P, Ibáñez B, et al. Impact of cannabis and other drugs on age at onset of psychosis. J Clin Psychiatry. 2008;69(8):1210-1216.
72.    Corcoran CM, Kimhy D, Stanford A, et al. Temporal association of cannabis use with symptoms in individuals at clinical high risk for psychosis. Schizophr Res. 2008;106(2-3):286-293.
73.    Moore TH, Zammit S, Lingford-Hughes A, et al. Cannabis use and risk of psychotic or affective mental health outcomes. Lancet. 2007;370(9584):319-328.
74.    Ferdinand RF, Sondeijker F, van der Ende J, Selten JP, Huizink A, Verhulst FC. Cannabis use predicts future psychotic symptoms, and vice versa. Addiction. 2005;100(5):612-618.
75.    Verdoux H, Tournier M, Cougnard A. Impact of substance use on the onset and course of early psychosis. Schizophr Res. 2005;79(1):69-75.
76.    Bodkin L, Singh A, Corcoran C. Cannabis as a risk factor for psychosis in vulnerable teens: implications for treatment. Primary Psychiatry. 2008;15(6):51-57.
77.    Degenhardt L, Hall W, Lynskey M. The relationship between cannabis use and other substance use in the general population. Drug Alcohol Depend. 2001;64(3):319-327.
78.    Grant BF, Hansan DS, Chou SP, Stinson FS, Dawson DA. Nicotine dependence and psychiatric disorders in the United States. Arch Gen Psychiatry. 2004;61(11):1107-1115.
79.    Hunglund BL, Basavarajappa BS. Are anandamide and cannabinoid receptors involved in ethanol tolerance? A review of the evidence. Alcohol Alcohol. 2000;35(2):126-133.
80.    Poirier MF, Canceil O, Bayle F, et al. Prevalence of smoking in psychiatric patients. Progress Neuropsychopharmacol Biol Psychiatry. 2002;26(3):529-537.
81.    Üçok A, Polat A, Bozkurt O, Meteris H. Cigarette smoking among patients with schizophrenia and bipolar disorders. Psychiatry Clin Neurosci. 2004;58(4):434-437.
82.    Weiser M, Reichenberg A, Grotto I, et al. Higher rates of cigarette smoking in male adolescents before the onset of schizophrenia: a historical-prospective cohort study. Am J Psychiatry. 2004;161(7):1219-1223.
83.    Sacco KA, Termine A, Seyal A, et al. Effects of cigarette smoking on spatial working memory and attentional deficits in schizophrenia: involvement of nicotinic receptor mechanisms. Arch Gen Psychiatry. 2005;62(6):649-659.
84.    Watkins SS, Koob GF, Markou A. Neural mechanisms underlying nicotine addiction: acute positive reinforcement and withdrawal. Nicotine Tob Res. 2000;2(1):19-37.
85.    Ziedonis DM, George TP. Schizophrenia and nicotine use: Report of pilot smoking cessation programs and review of neurobiological and clinical issues. Schizophr Bull. 1997;23(2):247-254.
86.    Miettunen J, Törmänen S, Murray GK, et al. Association of cannabis use with prodromal symptoms of psychosis in adolescence. Br J Psychiatry. 2006;192(6):140-141.
87.    Angermeyer MC, Kühn L. Gender differences in age at onset of schizophrenia: an overview. Eur Arch Psychiatry Neurol Sci. 1988;237(6):351-364.
88.    DeLisi LE. The significance of age at onset for schizophrenia. Schizophr Bull. 1992;18(2):209-215.
89.    Leung A, Chue P. Sex differences in schizophrenia, a review of the literature. Acta Psychiatr Scand Suppl. 2000;401:3-38.
90.    Häfner H, Maurer K, Löffler W, Reicher-Rossler A. The influence of age and sex on the onset and early course of schizophrenia. Br J Psychiatry. 1993;162(1):80-86.
91.    Häfner H, Maurer K, Löffler W, et al. The ABC Schizophrenia Study: a preliminary overview of the results. Soc Psychiatry Psychiatr Epidemiol. 1998;33(8):380-386.
92.    Shimizu A, Kurachi M, Yamaguchi N, Torii H, Isaki K. Does family history of schizophrenia influence age at onset of schizophrenia? Acta Psychiatr Scand. 1988;78(6):716-719.
93.    Gorwood P, Leboyer M, Jay M, Payan C, Feingold J. Gender and age at onset in schizophrenia: impact of family history. Am J Psychiatry. 1995;152(1):208-212.
94.    Stompe T, Ortwein-Swoboda G, Strobl R, Friedmann A. The age at onset of schizophrenia and the theory of anticipation. Psychiatry Res. 2000;93(2):125-134.



Dr. Sussman is editor of Primary Psychiatry as well as Associate Dean for Post-Graduate Programs and professor of psychiatry at the New York University School of Medicine in New York City.

Dr. Sussman reports no affiliation with or financial interest in any organization that may pose a conflict of interest.

Email questions or comments to ns@mblcommunications.com

While reading the Wall Street Journal on the morning of March 8, 2010, I came across the following headline under the New Medical Findings section:“Say What? New Risk in Pain-Reliever Use.”1 The brief article, based on a study published in the American Journal of Medicine,2 went on to report the following:

     “Regular use of pain-relief medicine appears to increase men’s risk of hearing loss, especially among middle-aged men, according to an American Journal of Medicine study. Researchers surveyed nearly 27,000 men every two years from 1986 to 2004; about one-fourth of the men said they had been diagnosed with hearing loss. Men who used pain relievers at least twice a week were more likely than non-users to be diagnosed. Aspirin users were 12% more likely, those on ibuprofen-like drugs were 21% more likely and users of acetaminophen, 22% more likely. Men from 45 to 50 years old at the start of the study faced the greatest risk—a 33% increase for aspirin, 61% for ibuprofen and 99% for acetaminophen. Previous nonhuman research has found some substances in pain-relievers can decrease blood flow to the cochlea, the part of the inner ear that converts waves sound into brain signals.”2

This was of particular personal interest to me because in recent years there has been decline in my hearing, along with an increase in tinnitus. It got suddenly worse ~1 year ago. The physicians at my institution gave me the full work up—magnetic resonance imaging, computerized axial tomography scan, and advanced auditory testing—trying to determine why, other than age, there had been such a sudden change. Nothing turned up. However, reading this article made me question whether years of using ibuprofen for various ailments had caused or contributed to my hearing problems. It also caused me to realize how often it is that we only find out about some serious drug adverse effect many years or even decades after it has come into clinical use. This is certainly the case with drugs in all therapeutic areas, including those used to treat mental disorders.

For example, there is growing evidence that use of antidepressants may be associated with an increase in risk for developing diabetes. Two studies3,4 have been published in recent months that strongly suggest that patients treated with these agents be monitored closely for evidence of glucose dysregulation.

One study3 found that use of selective serotonin reuptake inhibitors (SSRIs) was only associated with a significantly reduced risk of developing type 2 diabetes compared to using tricyclic antidepressants (TCAs) alone (37.5% versus 44.2%), but that using TCAs/SSRIs concurrently was associated with an increased risk of type 2 diabetes compared to using a TCA alone (59.8% versus 44.2%). After adjusting for sex, age, number of physician visits, and use of augmentation therapy, only the use of TCAs and SSRIs concurrently was associated with an increased risk of type 2 diabetes compared to using TCAs only. Using multiple antidepressants or SSRI monotherapy was not associated with an increased risk of diabetes compared to using TCAs alone. SSRI use in the study was associated with weight gain during the study. The investigators found that elevated depression inventory scores, which were present in 10.3% of patients on study entry, did not predict whether patients would develop diabetes during the study, but baseline antidepressant use did. When other factors associated with the risk of developing diabetes were controlled, elevated depression scores at baseline or during the study were not associated with diabetes risk in any arm.

As part of the study, some patients were put on metformin, an antidiabetic drug, as prophylactic treatment. The study showed that treatment with metformin did have a protective effect—those participants on antidepressants with metformin did not develop diabetes. Other findings were that antidepressant treatment for shorter periods or with lower daily doses were not associated with an increased risk. Recent use of other antidepressants was associated with an 80% increase in risk of diabetes; however, a dose or duration effect could not be detected, probably because of the rather low number of exposed case and comparison subjects.

Another study found results that are consistent with the data from the randomized Diabetes Prevention Program trial cited above. This article4 reported on a large observational study, which included >160,000 patients with depressive disorder treated with antidepressants for up to 2 years. It was found that recent long-term use of antidepressants in moderate to high daily doses was associated with an 84% increase in risk of diabetes. This association was present for both TCAs and SSRIs. The authors reported that antidepressant treatment for shorter periods or with lower daily doses was not associated with increased risk. Recent use of other antidepressants was associated with an 80% increase in risk of diabetes. However, a dose or duration effect could not be detected, probably because of the low number of exposed case and comparison subjects. There was a four-fold increase in risk of diabetes associated with long-term therapy with paroxetine in daily doses >20 mg/day but not with long-term use of fluoxetine, citalopram, or sertraline. Paroxetine, venlafaxine, fluvoxamine, and amitriptyline were associated with the highest risk.

These two studies3,4 are consistent with a third study, a Canadian review5 of the medical history of 2,400 people who were diagnosed with depression and were taking antidepressants. The investigators set out to determine whether there was a clear correlation between that disease and type 2 diabetes. They found that people with a history of depression had a 30% increased risk of type 2 diabetes. They divided the group into four categories: those taking older antidepressants, using newer treatments, using a combination of both old and new treatments, and who were switching medications. The risk of diabetes almost doubled for the patients who were using two types of therapies at the same time, TCAs and SSRIs. Collectively, these findings suggest a need for regular screening for diabetes in depression, particularly those taking more than one antidepressant.

An obvious question that arises as a result of these findings is whether the growing practice of augmenting antidepressants with second-generation antipsychotic, which have a well-established risk profile for causing cardiometabolic disturbances, will result in an even higher incidence of type 2 diabetes among patients on combination therapy. This issue surely should be on our radar.

I want to thank Brendan T. Carroll, MD, and Francisco Appiani, MD, for serving as guest editors for this issue of Primary Psychiatry. Their guest editorial provides an overview on the articles that focus on this month’s theme—catatonia. I would also urge you to read the case report by Michael M. Messer, MD, and Irina V. Haller, PhD, MS, which describes exciting findings about the effectiveness of ketamine as an antidepressant. In recent years, it has become clear that this anesthetic/recreational drug may pave the way for a new generation of effective antidepressants.  PP



1.    Singer-Vine J. Say what? New risk in pain-reliever use. Wall Streeet Journal. March 8, 2010. Pg D4.
2.    Curhan SG, Eavey R, Shargorodsky J, Curhan GC. Analgesic use and the risk of hearing loss in men. Am J Med. 2010;123(3):231-237.
3.    Rubin RR, Ma Y, Marrero DG, et al. Elevated depression symptoms, antidepressant medicine use, and risk of developing diabetes during the Diabetes Prevention Program. Diabetes Care. 2008;31(3):420-426.
4.    Andersohn F, Schade R, Suissa S, Garbe E. Long-term use of antidepressants for depressive disorders and the risk of diabetes mellitus. Am J Psychiatry. 2009;166(5):591-598.
5.    Brown LC, Majumdar SR, Johnson JA. Type of antidepressant therapy and risk of type 2 diabetes in people with depression. Diabetes Res Clin Pract. 2008:79(1);61-67.



This interview took place on January 15, 2010 and was conducted by Norman Sussman, MD.

Disclosure: Dr. Chang is consultant to Bristol-Myers Squibb, Eli Lilly, and GlaxoSmithKline; is on the speaker’s bureaus of Bristol-Myers Squibb, Eli Lilly, and Merck; and receives research support from GlaxoSmithKline, the National Alliance for Research on Schizophrenia and Depression, and the National Institute of Mental Health.


Dr. Chang is associate professor of Psychiatry and Behavioral Sciences in the Division of Child Psychiatry at the Stanford University School of Medicine in California. He is director of the Pediatric Bipolar Disorders Clinic and Research Program, where he specializes in pediatric psychopharmacology and treatment of depression and bipolar disorder in children and adolescents. Dr. Chang’s research includes brain imaging, genetics, and medication and psychotherapy trials.


How is bipolar disorder diagnosed in children and adolescents?

Unfortunately, we are still using the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition1 criteria because that is all we were given. We still do not have a good biological marker to help with early detection and diagnosis. That is the focus of my research, but in the meantime we are left with phenomenology and the DSM-IV. The biggest thing to remember in children is that they are not small adults and that, naturally in their developing brains, they have rapid mood shifts and rapid emotional changes; at each age, what is considered normal in development changes. A 3-year-old child is very normally going to have rapid mood shifts, and an adolescent is going to have fairly rapid mood shifts as well. Clinicians must remember that context when making a diagnosis in youths. In light of these rapid mood shifts, clinicians must realize that much of it is normal. Even if it is an abnormal mood, whether it is depression or mania, the child is not usually going to sit in the mood for the whole 2 weeks for depression or the full 1 week for mania. The crux of the whole diagnostic question always comes down to the episode: Can a child reach a full episode? Youths typically do not meet full episode criteria because their brains are not geared that way.

Since their brains are still developing, the criteria for bipolar disorder do not apply correctly to children. However, clinicians still try to make the diagnosis fit the criteria as much as possible, looking at the overarching episode as a distinct period of change in mood and not necessarily requiring that the child is in that mood state 24 hours a day.


How is that diagnostic process different than for an adult?

It is similar in that adults in mania are not necessarily manic 24 hours a day. Their moods do change throughout the day; it is just a predominant mood. However, a child’s mood fluctuation is definitely more extreme than in adults. When that normal developmental context is clear, a clinician can make a more appropriate diagnosis based on what is normal for a child.

The individual symptoms can be different as well, because the symptoms were originally developed for adults. Grandiosity in a 32-year-old does not look the same as grandiosity in a 7-year-old. Again, one has to be aware of what is normal versus abnormal grandiosity in a 7-year-old.


Do you think that bipolar disorder has become a default diagnosis in children and adolescents?

Yes, I think it has become one of the diagnoses, along with pervasive developmental disorders in children. A study2 examining the last 10 years in an Health Maintenance Organization sample found a 4,000-fold increase in community diagnoses of bipolar disorders in children. That does not mean that they really have bipolar disorder, rather that there has been a greater diagnosis in the community. I think it is a reflection of the fact that people do not know what it looks like. When behavioral problems occur, a diagnosis of bipolar disorder has become one of the first thoughts that people have. It is good in that at least we are not missing diagnosing children who have real bipolar disorders. However, clearly there are children also being incorrectly diagnosed and overdiagnosed. Thus, it is really important to get the accurate diagnosis.


Some people would argue that the increase in diagnosis is because there are now drugs approved for that diagnosis and that such drugs are being promoted.

Without just flat out saying that is complete lunacy, there is some concern about pharmaceuticals’ involvement to the extent where now pharma is reluctant to even market such drugs towards children, regardless of the indications for children, because of the fear of the backlash. If anything, I think that kind of backlash will be negative and that we are going to start missing diagnoses and missing treatment opportunities because people are so fearful that they may be looked upon as agents of the pharmaceutical industry.

Before the last 2 years, nothing was indicated for bipolar disorder in children other than lithium, and that was only down to age 12. There are still children who need treatment. I think it is a little bit of an incorrect kind of accusation.


Upon getting a history, even without seeing a child’s or adolescent’s symptoms, what would a clinician look for that would put a youth at high risk for bipolar disorder?

The first thing we always look at is family history. The more loaded the family history for mood disorder, particularly bipolar disorder, and the closer to the child (eg, a first-degree relative, like a sibling or a parent), the greater the risk.

Then, after that, we take a really close look at two main kinds of syndromes. First, unipolar depression: a child with depression who has a strong family history is at very high risk to go on to develop bipolar disorder, ie, possibly 30% to 50%. Second, attention-deficit/hyperactivity disorder (ADHD) plus mood dysregulation: if they have symptoms of ADHD at an early age and then go on to develop more mood problems and they have a family history of bipolar, that also puts them at higher risk.

Those are the two main pathways that my colleagues and I have been investigating in our research. We are pretty sure there are other pathways as well, including anxiety disorders, and even children who present with what we call bipolar not otherwise specified (NOS); the latter opens up a whole other can of worms, but is really important because probably most of these children in the community are being diagnosed with bipolar disorder NOS without really good, firm criteria to back up those diagnoses.


Does family history of alcoholism help predict whether there is a higher risk of bipolar disorder?

We have not seen alcoholism per se do that, but whenever we get a family history of alcoholism we examine more closely to see if there is a mood component. With family histories, we will go very closely and try to diagnose by proxy, asking the parents questions about their direct relatives (eg, “Did your brother have this? Did he have a manic episode? Did he have a depressive episode?”) Oftentimes, if the patient had a history of drug abuse or alcoholism, we will find either depression or sometimes even bipolar disorder. However, in the absence of that, we have not really seen a big increase in risk.


Do you think bipolar disorder is caught most of the time now, or is there still a ways to go in terms of recognition?

I think there is still a ways to go, depending on where you are. In the United States, certainly in major urban areas, there is probably the tendency towards overdiagnosis, but in other areas it is still being underdiagnosed. Clearly, the increase in incidence is due to better recognition and over-recognition, but it may also be due to an actual increase in incidence due to genetic and environmental reasons. The age of onset appears to be getting lower throughout the generations, so that would indicate that there are more children and adolescents with bipolar disorder now than there were in previous generations.

I think we have a long way to go in terms of catching bipolar disorder. The majority of children diagnosed in the community are diagnosed with NOS, but there are no firm criteria. We have been trying to develop criteria, along with other research institutions across the country so that we can speak the same research language, and that can be translated to better clinical recognition of bipolar NOS.

Currently, the way we are using NOS is that patients have to have the criteria A symptoms for the manic episode—which are elation, euphoria, or irritability—but they need one less criteria B symptom than they normally would, or their episodic duration needs only to be 4 hours rather than 4–7 days. However, during the 4 hours, the symptom has to be something that repeats itself on different occasions and does cause functional impairment.

It is a little controversial, but when this type of criteria is applied to children, many children are found to meet this diagnosis. This has been well-studied in the Course and Outcome of Bipolar Youth study in Pittsburgh.3 Those researchers have found that ~38% of these children with bipolar disorder NOS go on to develop full bipolar I or II disorder over 4 years.


Why are early recognition and intervention important in dealing with children and adolescents?

Detecting the manic episode early on is important to not only inform better treatment, but also to prevent future episodes from reoccurring. My colleagues and I are big believers in the kindling hypothesis, which proposes that with each subsequent mood episode, it becomes easier for the brain to flip into that mood episode. Preventing mood episodes is actually preventing recurrence of worse and worse mood episodes that become more difficult to treat.

Thus, it seems prudent to figure out which children are at the highest risk for developing bipolar disorder and intervene even before the first manic episode. Granted, there is much controversy with treating children who have not developed full bipolar disorder; however, as I mentioned, a lot of these states are already problematic, whether it is bipolar disorder NOS, depression, or ADHD plus mood symptoms. They are already usually being treated anyway in the community. If we can tailor that treatment and guide it more towards bipolar disorder prevention, we will be able to hopefully even stave off the first manic episode and prevent a whole life filled with morbidity.


Is there any new evidence that confirms or contradicts reports about there being specific polymorphisms associated with bipolar disorder?

What is new is that we are realizing the complexity involved with these genetic polymorphisms. For example, with brain derived neurotrophic factor or with the serotonin transporter gene, it does seem like certain polymorphisms confer perhaps a very small increased risk for developing bipolar disorder. Yet, the complexity we are realizing is that these polymorphisms then affect development of certain brain regions and brain activation patterns that then lead to problems with mood regulation. The future of understanding how these candidate genes help increase risk for major mood disorders, including bipolar disorder, is going to look at their effects on brain structure and function that then lead to these mood episodes. That is an area of research that my colleagues and I are actively engaged in.


In treating a child as opposed to a middle-aged adult, are there different considerations when choosing which medication to start with?

Absolutely. We used to base our treatment plans on adult data because there were no childhood data. In the last 4–5 years there has been an explosion of data, so that now we have many positive, placebo-controlled, large-scale trials for acute mania in children down to 10 years of age.

Unfortunately, some of those trials have been negative, as well—for example, with oxcarbazepine and, surprisingly, divalproex. This has somewhat changed the landscape of treatment choices in children in the community and probably also in academia, where now we are gearing a little bit more towards atypical antipsychotics because of their relatively positive data compared to anticonvulsants.


You seem to be referring to acute treatment. Does the answer change in regards to maintenance or long-term treatment?

Definitely, again because of our concerns with the side effects from the atypical antipsychotics, whether it be weight gain or metabolic problems. I refer to children who have even more weight gain than adults exposed to these medications. From a long-term standpoint, we are very careful when continuing medication doses at the doses that were used to get them stable.

It becomes a different issue. Unfortunately, there are very little maintenance data in children. There are a couple studies being conducted now in that area, but only with a couple agents, so we need a lot more data before being able to understand what is the best for efficacy. However, for long-term side effects profile, we are concerned about continuing atypical antipsychotics. Thus, we will be very active in treating the acute symptoms. Once we get things under control, we are also very active in trying to pare down the medication regimen so that the children do not have these side effects, because very little is known about the adverse effects on the developing brain and body of these children.


Are risks of side effects greater in children, as opposed to adults, taking medications to treat bipolar disorder?

Yes, the risks become much greater when dealing with children. Many are psychological risks because clinicians are afraid to impact children more so than they are afraid to impact adults, as children are in a more vulnerable position. They cannot consent for themselves, but can only give assent. It is a little bit trickier.

Still, when it comes to a major mood disorder like bipolar disorder that carries great morbidity and a great suicide rate, it becomes clear to these families that some of these unknown side effects can be handled down the road. My colleagues and I believe endocrine effects, such as polycystic ovarian syndrome, are somewhat reversible. Maybe not the kidney effects—almost nothing is known about that in children in long term.

The idea is to try to stabilize a child and improve his or her functioning and prevent the suicidality now because that is the most important thing. Then we will worry about the other kinds of side effects down the road after the child has been stabilized.


Is there any evidence that the early use of antidepressants or stimulants to treat either depression or ADHD has led to an increase or a worsening of bipolar disorder when used in a vulnerable child?

Clinicians are much more aware of this possibility now than they were 10 years ago. However, time and time again, we still see patients who are exposed to these medications that have manic episodes, and we wonder if it is completely due to these medications or not. We used to be very concerned about stimulants and that they may promote kindling in these vulnerable children, but more and more data have come out that suggest that, overall, stimulants are probably not a problem. In fact, they may be beneficial in some children.

Tillman and Geller4 followed children with ADHD. Approximately 25% of them developed bipolar disorder, eventually experiencing a manic episode. Those children treated with stimulants were less likely to develop mania than children who were not treated with stimulants, almost suggesting that there may be some protective effects of improving a child’s psychosocial and school functioning as well as against developing a mood disorder. Even with stimulants currently, once in a while a child still may have a manic reaction. Of course, that child should probably not be on this class of medications. Still, overall, it appears stimulants may actually be beneficial in these children and not speed up the process of bipolar disorder.

Antidepressants are a different story. A lot of retrospective evidence suggests that antidepressants are much more problematic in causing manic episodes in children who otherwise may not have developed mania. My colleagues and I are examining that area very closely. The conversion from unipolar depression to mania often coincides with giving a child an antidepressant.

Again, we look very carefully at family history and other early signs of hypomania that might indicate that this child really has underlying bipolar brain chemistry, and then in those children we are very careful before starting antidepressants. We still feel that some of them will be fine, but there is a certain proportion—probably significant somewhere on the order of 30% to 40% of children who have strong family histories of bipolar disorder and depression—who may respond adversely to antidepressants. I am very careful before using them, and if we do use them, we start very low and go very slowly, and ask the parents to really closely monitor for any manic symptoms.

The alternatives, though, are the mood stabilizers lithium and lamotrigine, which have not been well studied yet in this population. Some parents will opt to go that direction rather than risk having a manic episode. There are more side effects associated with those medications, so we are very careful to explain all the risks and benefits to the family members.


People tend to ask whether there is some kind of imaging or other diagnostic test that will help confirm the diagnosis. What are your thoughts on this?

I tell parents, family members, and patients that right now the general consensus among all researchers in the field is that bipolar disorder or most psychiatric disorders cannot be diagnosed using any kind of brain imaging. In fact, at this point brain imaging does not inform diagnosis or treatment selection, particularly for bipolar disorder.

I do tell them there are private enterprises that use brain imaging more as a clinical gestalt tool because most of their data have not been published and have not been rigorously analyzed to figure out what are signs associated with specific diagnoses or specific treatment response. A clinician involved in such enterprises might see that the last five people treated with a certain medication who had a certain symptom complex did really well on a specific drug. The clinician will then continue that drug. It is almost as if they are using this tool in the same manner, saying, “The last five people we saw with this brain pattern responded well to this medication profile, so we will use this.” Some have received positive results but some have not. I just caution them to be completely aware of the possible pitfalls of that approach.

That is all a long-winded way to say that we really have no idea or evidence that these things work at all. I do not recommend obtaining brain imaging for diagnostic purposes at this time, but it is an excellent research tool, and we are getting close to making it useful clinically.  PP



1.    Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Association; 1994.
2.    Moreno C, Laje G, Blanco C, Jiang H, Schmidt AB, Olfson M. National trends in the outpatient diagnosis and treatment of bipolar disorder in youth. Arch Gen Psychiatry. 2007;64(9):1032-1039.
3.    Birmaher B, Axelson D, Goldstein B, et al. Four-year longitudinal course of children and adolescents with bipolar spectrum disorders: the Course and Outcome of Bipolar Youth (COBY) study. Am J Psychiatry. 2009;166(7):795-804.
4.    Tillman R, Geller B. Controlled study of switching from attention-deficit/hyperactivity disorder to a prepubertal and early adolescent bipolar I disorder phenotype during 6-year prospective follow-up: rate, risk, and predictors. Dev Psychopathol. 2006;18(4):1037-1053.



Dr. Luo is associate clinical professor in the Department of Psychiatry and Biobehavioral Sciences at the University of California in Los Angeles; past president of the American Association for Technology in Psychiatry (AATP) in New York City; and Gores Informatics Advocacy chair at the AATP.

Disclosure: Dr. Luo reports no affiliation with or financial interest in any organization that may pose a conflict of interest.


Technology has enabled mental health practitioners to have greater reach to their patients with telemedicine, and electronic prescribing systems have made transcribing errors a faint memory. However, the day-to-day practice of medicine still requires many other elements in the office for communication and documentation such as e-mail, scheduling, transcription, and voicemail. Some practices can afford to hire staff full time, whereas others consider part-time and even off-site assistants for cost savings. With today’s more technologically savvy patients who look up their physicians online and review medical information, it is not a far stretch that these same tech-savvy patients want that same convenience with online scheduling and one stop for messages. This month’s column reviews these virtual assistants which can enhance the clinician’s busy practice and save time.

Online Scheduling

It is highly unlikely that a private practice would use an online calendar such as Google Calendar1 to schedule appointments for the office. Although it is free and has options whereby Gmail users can share their calendars, it clearly lacks privacy features and the ability to set appointments. Many electronic health record systems such as Practice Fusion2 and Valant Electronic Medical Record3 offer appointment scheduling entered by providers and administrative staff in realtime on these Web-based systems, typically at the end of an appointment. However, a true time saver would be if patients could independently reschedule or schedule their appointments directly online without staff intervention. Companies such as SCI Solutions’ Schedule Maximizer,4 Appointment Quest,5 NetAppointment,6 and Appointment-Plus7 offer these services. Patients are able to see available times only and do not see appointment times taken by other patients. They can log in at any time and set up how they want to be reminded, such as via e-mail or phone. In addition, these systems can be set up to collect co-pays and other information in advance of the appointment. Many of these systems can export the schedule into Microsoft Outlook or a spreadsheet for offline backup. Finally, many of these systems also provide analysis, which may help providers determine which patients cancel the most often or who change their appointments frequently.


Secure Communication

Many healthcare providers utilize e-mail for communication with patients, usually initiated by the patient who finds it convenient for asking for refills or a change an appointment. The 1998 American Medical Informatics Association Guidelines for the Clinical Use of Electronic Mail with Patients recognized that encryption was not widely available, user friendly, or practical, and therefore did not demand its use.8 Although encryption availability in standard e-mail client software still has not become user friendly or practical, there are alternative solutions to ensure that electronic communication between patients and providers is secure enough to maintain privacy. LuxSci9 and 4SecureMail10 are companies that provides Health Insurance Portability and Accountability Act-compliant e-mail solutions. One method is using a a secured ‘escrow’ account where patients and providers go online via an SSL connection to the company e-mail server providing a secured portal as all messages are left there. Alternatively, both companies also offer secured SMTP (simple mail transfer protocol) relaying services so that providers can use an e-mail client such as Microsoft Outlook or Eudora to send encrypted e-mail. RelayHealth provides a secure communication portal where patients can leave specific messages for providers, such as request/cancel appointment, request a lab/test result, request medication refills, send a note to the doctor/office staff, or request a referral.11 It structures the patient queries via specific messaging templates that have fields such as pharmacy and medication request, so that the prescription can be sent via fax directly in the system on the SureScript network. RelayHealth also offers webVisits, where patients can enter specific queries or consultations on non-urgent topics, which are responded to by providers using a template-based reply system to save time. For offices with a single provider and on a limited budget, TeleHealth Connect provides a free secure communication system for providers with their patients using the basic account.12 TeleHealth Connect has partnered with Microsoft’s HealthVault, a personal health record system, so that all secure messages and attachments are stored in HealthVault. These messages are controlled by the patient, who can delete and discontinue them at any time. Patients always use the TeleHealth Connect system at no cost, whether the healthcare provider is in solo practice or part of an enterprise system.

Although there may be some reluctance on the part of patients to entrust their medical information with a large company such as Microsoft, HealthVault was launched in 2007 and was designed to balance privacy features with management of health information. In particular, the privacy protections in HealthVault reflect the privacy principles of the Coalition for Patient Privacy,13 a non-profit health privacy watchdog organization founded in 2004 by Debora C. Peel, MD. Encouraging patients to set up a free account with HealthVault has benefits for healthcare providers as many Internet-based healthcare applications and medical devices interface with HealthVault and store information there. For example, Quest Diagnostics,14 a large national laboratory testing company, can export patient test results directly from their doctor into their HealthVault account upon authorization. In this way, patients can have a copy of their blood tests after reviewing them in the office with their physician. If the physician is using the National ERx initiative software from Allscripts ePrescribe,15 patients can request to have an electronic record of medications, conditions, and allergy information be placed into their HealthVault account. Similarly, patients can have their pharmacy records from CVS16 and Walgreens17 linked to their HealthVault account for management. Personal health records were primarily touted as a way for patients to keep track of their own health information; however, a secondary benefit is that more accurate health information is organized and now available to any physician, which will save valuable time spent sending for records or contacting pharmacies for medication names, dosages, and frequencies.


Central Communication

Many physicians carry a pager, check messages on an office voicemail system, receive fax requests for medication refills, and use e-mail for communication. All theses systems are enough to make anyone feel like they need several heads like the hyrda beast to keep up with these systems. A variety of products help tame the information overload in different ways. Google Voice18 provides one phone number to manage all the clinician’s phones. This number is linked to the clinician’s account, not a physical location or a particular device. Voicemail can be heard by dialing in or on the computer while online. Voicemail can even be transcribed and sent either via e-mail or text messages. The Google Voice number can be used to route phone calls to any phone number such as a mobile phone or office phone. One very handy feature is the ability to use a Google Voice number as the caller ID, which preserves the privacy of the mobile device. Innoport19 offers hosted unified communication solutions, which provides similar features as Google Voice, but it includes a hosted PBX that provides directory assistance, customizable greetings, and Internet fax to e-mail as well as e-mail to fax. YouMail20 provides access to voice mail either on phone, online, or via e-mail attachment. It also provides voicemail filtering to avoid telemarketing as well as smart greetings to distinguish between family, friends, and patients. Jott21 offers similar services for voicemail, but also allows voice dictation to send notes and to-dos, set reminders and appointments, send e-mail and text messages, and post to blogs, all with a phone call. These time-saving solutions all work by keeping it simple to respond to different mechanisms of received messages.



Although the technologies described above at first seem obvious and mundane, the minutes of time saving and decreased hassle all add up. Unfortunately, one system does not provide everything needed for the office practice, akin to how finding the perfect electronic health record system today is very much an exercise in frustration. There are times when a jack-of-all-trades or unified solution is better than numerous expert or customized system solutions; however, many of these virtual assistants do add up to save clinicians time and frustration. Why not employ one of them now?  PP



1.    Google Calendar. Available at: http://calendar.google.com. Accessed March 4, 2010.
2.    Practice Fusion. Available at: www.practicefusion.com. Accessed March 4, 2010.
3.    Valant EMR. Available at: www.valentmed.com. Accessed March 4, 2010.
4.    SCI Solutions. Available at: www.scisolutions.com/solutions/schedule-maximizer.asp. Accessed March 4, 2010.
5.    Appointment Quest. Available at: www.appointmentquest.com. Accessed March 4, 2010.
6.    NetAppointment. Available at: www.netappointment.com. Accessed March 4, 2010.
7.    Appointment-Plus. Available at: www.appointment-plus.com. Accessed March 4, 2010.
8.     Kane B, Sands DZ. Guidelines for the clinical use of electronic mail with patients. The AMIA Internet Working Group, Task Force on Guidelines for the Use of Clinic-Patient Electronic Mail. J Am Med Inform Assoc. 1998;5(1):104-111.
9.    LuxSci. Available at: www.luxsci.com. Accessed March 5, 2010.
10.    4SecureMail. Available at: www.4securemail.com. Accessed March 5, 2010.
11.    RelayHealth. Available at: www.relayhealth.com. Accessed March 5, 2010.
12.    TeleHealth Connect. Available at: www.telehealthconnect.com. Accessed March 7, 2010.
13.    Patient Privacy Rights Foundation. Available at: http://patientprivacyrights.org. Accessed March 7, 2010.
14.     Quest Diagnostics Partnership. Available at: www.healthvault.com/websites/QuestDiagnostics-MyCare360.html. Accessed March 7, 2010.
15.     Allscripts ePrescribe Partnership. Available at: www.healthvault.com/websites/allscripts-AllscriptsePrescribe.html. Accessed March 7, 2010.
16.     CVS Partnership. Available at: www.healthvault.com/websites/CVSPharmacy-CVSPharmacy.html. Accessed March 7, 2010.
17.     Walgreen Partnership. Available at: www.healthvault.com/websites/Walgreens-Walgreens.html. Accessed March 7, 2010.
18.    Google Voice. Available at: voice.google.com. Accessed March 9, 2010.
19.    Innoport. Available at: www.innoport.com. Accessed March 9, 2010.
20.    YouMail. Available at: www.youmail.com. Accessed March 9, 2010.
21.    Jott. Available at: www.jott.com. Accessed March 9, 2010.


Parents Report Inadequate Support in Schools for Behavioral and Emotional Issues

Children with behavioral, emotional, or family problems typically first seek support from school psychologists, counselors, and social workers. Behavioral problems, including attention-deficit/hyperactivity disorder (ADHD), along with emotional and family problems, bullying, and homelessness, can negatively impact children’s academic success.

Approximately 1,100 parents across the United States were surveyed by the C.S. Mott Children’s Hospital National Poll on Children’s Health regarding how well their children’s public schools supported children with behavioral, emotional, or family problems. The poll revealed that 37% of parents surveyed gave an A to primary schools’ support for children with ADHD and other behavioral problems, and 34% gave an A for support for children with emotional or family problems. For secondary schools, 22% of parents surveyed gave an A for support for children with behavioral, emotional, or family problems. For overall education, 52% of parents surveyed gave primary schools an A, whereas 38% gave secondary schools an A.

“Most surprising is the lower parental confidence in secondary schools versus primary schools in public systems, in terms of support for children with emotional, family, or behavioral problems,” said Matthew Davis, MD, MAPP, director of the poll and associate professor of Pediatrics and Internal Medicine at the University of Michigan Medical School in Ann Arbor. “Generally, adolescents are at a greater risk than younger children, in terms of harming themselves and harming others. Therefore, we as a society need to offer more, rather than less, support for them and their emotional struggles.”

Some stakeholders believe that school funding should be limited to educational services due to the economy. If students need special support services, drastic cuts to these areas may interfere with students’ ability to learn and work to their full academic potential.

Dr. Davis said that in the most optimal circumstances, primary care physicians (PCPs) and schools should act together to support children and their families. However, there are many potential obstacles to good teamwork between PCPs and school staff.

“Our results indicate that one major problem is that many parents do not feel that their children’s schools are providing adequate support,” Dr. Davis said. “Based on our findings, I would encourage PCPs to ask parents openly about the amount of support they and their children are receiving from schools. If parents seem discouraged, PCPs should try to help them find other types of support within the community rather than insisting they keep trying to find support within the school setting.”

Funding for this research was provided by the Department of Pediatrics and Communicable Diseases and part of the CHEAR Unit at the University of Michigan Health System. To view the poll, please visit the University of Michigan Health System Website. (www.med.umich.edu/mott/npch/). – JV


fMRI Reveals Functioning Behind Stunted Emotional Processing in Generalized Anxiety Disorder

In a recent imaging study, patients with generalized anxiety disorder (GAD) demonstrated inhibited emotional processing. This, according to researchers, was explained by the inability of GAD patients’ brains to regulate the amygdala by engaging the pregenual anterior cingulate.

The investigation, headed by Amit Etkin, MD, PhD, at Stanford University, and colleagues, focused on implicit emotional regulation in GAD.

Etkin and colleagues scanned 17 GAD patients (mean 31.5 years of age, 65% female) and 24 healthy comparison subjects (mean 36.5 years of age, 75% female) with functional magnetic resonance imaging (fMRI) while displaying happy or fearful facial expressions. Each image was overlaid with a “happy” or “fear” caption. Some facial expressions had mismatching captions.

The comparison subjects non-intentionally regulated the emotional conflicts presented by mismatched captions. The GAD group, however, had impaired emotional adaptability and delayed reaction times. fMRI and performance results were so significantly correlated with symptoms that patients could be divided accurately by that sole criterion.

This study was funded by grants from the National Institute of Health and the residencey program of the Veterans Affairs–Palo Alto Health Care System. (Am J Psychiatry. Epub February 1, 2010). – LS


Database of Brain Tissue to Advance Research of Major Psychiatric Diseases

Biomarkers, which are biologic footprints left behind by illness, can be found in brain tissue, the most reliable vehicle that contains direct clues to psychiatric and neurologic disorders. Biomarkers can divulge clues regarding the origin of a disorder and the chain of events that cause full-blown disease.

Robert Yolken, MD, a neurovirologist at Johns Hopkins Children’s Center in Baltimore, Maryland, and colleagues from the Stanley Medical Research Institute in Baltimore, have developed a large repository of brain and tissue samples to advance the understanding and treatment of major psychiatric diseases such as bipolar disorder, major depressive disorder (MDD), and schizophrenia. 

“The medications currently used to treat mental illness can have significant side effects and often do not alleviate symptoms,” Dr. Yolken said. “We anticipate that the database will facilitate the identification of molecular pathways that can be targeted for the development of new and improved medications that primary care physicians can use more successfully in their practices.”

During the last 12 years, information from 45 human brains was obtained postmortem from people who were diagnosed with a psychiatric disorder. In total, the database contains 60 human brains, including 15 brains from people diagnosed with schizophrenia, 15 from those with bipolar disorder, 15 from those with MDD, and 15 from unaffected brains. The database also utilizes software that provides powerful analyses of 1,749 neuropathology datasets and several gene expression datasets.

Also available is information regarding other patient-specific factors, including age, sex, duration of illness, brain pH, alcohol, drug use, and smoking. The database also offers blood, lymphocytes, cerebrospinal fluid, and liver and spleen tissue from the same patients. These samples are meant to provide clues to disease activity. Yolken’s laboratory has also conducted research on the role that viral and bacterial infections, such as the human herpes virus, Epstein-Barr virus, toxoplasmosis, and influenza, play in the development of schizophrenia and bipolar disorder.

“We have found that antibody levels against Epstein-Barr virus and human herpes virus type 6 were significantly increased in the serum of patients with bipolar disorder as compared to control subjects,” Dr. Yolken explained. “Moreover, the antibody levels were significantly correlated with the dopamine receptor mRNA levels on the hippocampus of the same subjects. This suggests that exposure to certain viruses may contribute to the risk for developing psychiatric disorders by interacting with or regulating neurotransmitters.”

Funding for this study was provided by the Stanley Medical Research Institute.  (www.stanleyresearch.org/dnn/BrainResearchLaboratorybrBrainCollection/tabid/83/Default.aspx). – JV

Psychiatric dispatches is written by Lonnie Stoltzfoos and Jennifer Verlangieri.



Print Friendly 

Genetics and Genomics in Schizophrenia

Ming T. Tsuang, MD, PhD, DSc, Stephen J. Glatt, PhD,
and Stephen V. Faraone, PhD


Primary Psychiatry. 2003;10(3):37-40,50

Dr. DeVane is professor of psychiatry and behavioral sciences at the Medical University of South Carolina in Charleston. Dr. Nemeroff is Reunette W. Harris Professor and chairman of the Department of Psychiatry and Behavioral Sciences at Emory University School of Medicine in Atlanta, Ga.

Disclaimer: Although every effort has been made to ensure that drug doses and other information are presented accurately in this article, the ultimate responsibility rests with the prescribing physician. Neither the publishers nor the authors can be held responsible for errors or for any consequences arising from the use of information contained herein. Readers are strongly urged to consult any relevant primary literature. No claims or endorsements are made for any drug or compound currently under clinical investigation.

Acknowledgments: The authors report no financial, academic, or other support of this work.



The present “2002 Guide to Psychotropic Drug Interactions” is an update of the past 2000 edition. Since the appearance of the 2000 Guide, new psychotropic drugs have been introduced which have specific data related to their potential drug interactions. Documentation of drug interactions with commonly used psychotropics continues to appear in the literature at a steady pace.

As this guide is intended to serve an educational role for both the psychiatrist-in-training and the nonpsychiatric physician less familiar with the interactions of psychoactive drugs, the bulk of the background discussion on drug metabolism and mechanisms of drug interactions remains unchanged. For the repeat reader, we have summarized in Table 1 important new findings on drug interactions appearing since the last update. The interactions of three new psychoactive drugs introduced recently to the market (oxcarbazepine, modafinil, and ziprasidone) are covered in Tables 17, 22, and 32. Other additions in the tables reflect new case reports and further documentation of drug interactions.

New knowledge related to the benefits of psychiatric drug treatment results in earlier initiation of drug therapy for some psychiatric disorders, and maintenance therapy is more and more commonplace during asymptomatic periods. In fact, maintenance therapy for affective anxiety and psychotic disorders, often continuing for years or decades, is now the accepted standard of care, especially for patients with a history of recurrent episodes of illness. Long-term pharmacotherapy requires awareness and management of drug interactions.

As the population ages, more drugs are prescribed on a chronic basis for maintenance of health without treatment of overt symptoms. Increasing numbers of patients take one of the serum lipid-lowering compounds from the class of 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors. These drugs can be taken for primary prevention, regardless of whether or not the patient has previously experienced a vascular event such as myocardial infarction or stroke. With the exception of pravastatin, the drugs in this class are highly metabolized by cytochrome P450 (CYP) 3A4, a hepatic enzyme whose action can be inhibited by several antidepressants. As will be explained later, some knowledge of how the major antidepressants interact with specific liver enzymes allows the choice of an antidepressant that avoids such potential drug-drug interactions.

New drugs to treat psychiatric illness have been introduced to clinical practice in recent years. Additional antidepressants and antipsychotics are expected over the next few years. The recent introduction of ziprasidone reflects the high level of activity in drug development for treatment of psychotic conditions. Additional new drugs in this category are currently being tested in clinical trials. Each of these compounds possesses a certain potential to interact with other drugs. This is especially true since psychoactive drugs are generally highly metabolized compounds. Laboratory methodologies developed in recent years can identify the specific enzymes mediating various metabolic pathways. This information can be used to predict how a new drug will interact pharmacokinetically with a variety of other drugs already marketed. Some background knowledge of major drug-metabolizing enzymes is helpful in understanding how these predictions are made. Of course, in vitro predictions must be confirmed with in vivo studies, but supporting clinical data may not be available for months or years.

This guide summarizes psychotropic drug interactions from several viewpoints. First, examples of pharmacokinetics will be discussed to aid the reader in understanding how drugs may interact during the course of their absorption and elimination from the body. Secondly, because many interactions with psychotropic drugs occur via specific interactions with the CYP system, this hepatic enzyme system will be described and the most important enzymes involved in the metabolism or interactions of psychoactive drugs will be discussed. Some principles of drug interactions operating through competitive inhibition of hepatic enzymes will be explained, so that the reader may make informed judgments about the possibility of an interaction.

The bulk of the guide will be concerned with drug interactions that have been described with specific psychoactive drug classes. The degree of documentation varies for many interactions, from theoretical conjecture, to clinical experience with patients, to well-established research outcomes. The sources of interaction data will be noted to help identify the appropriate level of confidence in the predicted consequences of combining drugs in therapy. When possible, specific management guidelines are provided to avoid or minimize some potentially negative interactions. The major psychoactive drugs, classified according to their primary therapeutic indication, are listed in Table 2. Subsequent tables will list important drug interactions for each of these classes.


Classification of Drug Interactions

Drug interactions are commonly classified as occurring by either pharmacodynamic or pharmacokinetic mechanisms. A third category, pharmaceutical interactions, occurs from physical incompatibility of drugs. Examples of this last class include the precipitation of drugs following their addition to intravenous fluids of inappropriate pH, or the physical absorption of drugs to intravenous tubing. The intravenous dose of diazepam delivered can be far less than expected if it is injected into intravenous tubing distal to the point of venipuncture, due to drug absorption to plasticizer in the tubing. However, these types of interactions are rarely of concern, because the vast majority of psychoactive drugs are prescribed for oral administration.

Pharmacodynamic Drug Interactions

A pharmacodynamic drug interaction occurs when the pharmacologic response of one drug is modified by another drug without the effects being the result of a change in drug concentration. These interactions occur at the sites of drug action. Such sites can include receptors, ion channels, cell membranes, and enzymes. We lack a thorough understanding of these drug interactions, as they are generally more difficult to detect and study than pharmacokinetic interactions. The latter are more easily documented and quantified through measurement of plasma drug concentrations. The pharmacologic effects of psychoactive drugs can be difficult to measure, especially changes in behavior or mental status. Some examples are illustrative of pharmacodynamic interactions.

Drugs that produce sedation by different mechanisms often produce additive sedation when administered together. The combination of traditional antihistamines with benzodiazepines or alcohol provides an example. Another well-known pharmacodynamic interaction is the combination of a nonselective monoamine oxidase inhibitor (MAOI) with an over-the-counter (OTC) sympathomimetic nasal decongestant or foods rich in tyramine. Because of differing mechanisms of action, overstimulation of the sympathetic nervous system can result in pressor effects that produce hypertension. This interaction is becoming less of a clinical concern due to the diminishing use of the MAOIs. A more relevant example for current clinical practice is provided by serotonin syndrome. This is a potentially fatal disorder, which can result from combining highly serotonergic drugs. It was first recognized in laboratory animals given MAOIs and L-tryptophan, but has been documented with the newer antidepressants and other agents that have prominent serotonergic actions. It occurs in the absence of pharmacokinetic changes in drug disposition.

Some drug interactions at sites of action are specifically exploited for their therapeutic benefits. The pharmacodynamic interactions of competitive antagonists at receptor sites are the basis for development of several therapeutically useful drugs. Naloxone, propranolol, and flumazenil reverse the effects of opiates, catecholamines, and benzodiazepines at their respective receptor sites when given in close temporal proximity to their agonists. When adjunctive agents are combined with antidepressants (eg, lithium), or thyroid hormone with tricyclic antidepressants (TCAs), or pindolol with selective serotonin reuptake inhibitors (SSRIs), it is hoped that a pharmacodynamic interaction will result in an improvement in patient response.

Pharmacokinetic Drug Interactions

A pharmacokinetic interaction occurs when one drug alters the disposition of another drug, thereby resulting in a change in plasma or tissue drug concentration. The change in concentration may or may not result in clinically significant consequences. Any of the major components of drug disposition illustrated in Figure 1 can theoretically be affected.


For the psychoactive drugs, the drug dose is usually administered orally. Absorption occurs most often in the small intestine, where a favorable pH promotes transit across the gastrointestinal (GI) membranes. Some portion of the absorbed dose undergoes glomerular filtration and passes out through the urine in an unchanged form. The proportion varies both among individuals and between drugs. Generally, the psychoactive drugs are excreted unchanged only to a minor degree. Exceptions are lithium and gabapentin, which are excreted unchanged. Most drugs are biotransformed to either active or inactive metabolites. Either the administered parent drug and/or active metabolites can produce pharmacologic effects at various sites of action. In turn, metabolites are to some degree excreted in the urine, or they can be further metabolized. Eventually, the biotransformation process results in a metabolite that is sufficiently water-soluble to be renally excreted. Drug interactions may involve any of these various steps in the drug disposition process.

Interactions Involving Absorption

Absorption of orally administered drugs is a multistep process. Once a solid form (tablets, capsules) of a drug dosage is dissolved into solution in the GI tract, it transverses the gut lumen and wall in transit to the liver. A portion of the drug dose may never be absorbed, due to inadequate dissolution or drug interactions that promote further passage beyond the small intestine and elimination in the feces. The possible sites of drug elimination during absorption are shown in Figure 2. Drugs such as cholestyramine can physically bind to drugs in the GI tract and produce this effect. The nonabsorbable fat substitutes may also reduce the absorption of other drugs. Cimetidine, by altering GI pH, may reduce the rate or extent of absorption of many psychoactive drugs. Similarly, anticholinergic drugs can decrease the motility of the gut and alter drug absorption.

Drugs are subject to elimination during their absorption through the gut wall by the action of carrier proteins and metabolizing enzymes. P-glycoprotein (PGP) and CYP 3A4 act in concert to limit the absorption of a number of drugs. PGP is a carrier protein that exports drug molecules back into the GI tract. This creates a continual recycling of a portion of the unabsorbed drug dose and has the effect of increasing the exposure to CYP 3A4 and first-pass elimination (Figure 2). PGP transport is a saturable process, which partially explains why increasing absorption may occur with an increased dose.

The gut wall is the site of interaction of PGP or CYP 3A4 inhibitors that can increase the bioavailability of some drugs. Some natural chemicals in grapefruit juice down-regulate, or decrease, protein expression of CYP 3A4 in the gut wall, which allows greater amounts of drugs that are prominent 3A4 substrates to be absorbed. For cyclosporine, this interaction with grapefruit juice can increase drug bioavailability and result in decreased dosage requirements for immunosuppression and economic cost savings for patients.

The role of PGP in drug interactions is being increasingly recognized. The cardiac glycoside digoxin is not metabolized, but renally excreted, and St. John’s wort (SJW) decreases its plasma concentration. The likely mechanism is induction of intestinal PGP to limit digoxin’s oral absorption. A similar mechanism or CYP 3A4 induction may explain the lowering by SJW of indinavir, alprazolam, and cyclosporine plasma concentration. PGP also serves a protective function to limit access of drugs to the brain, due to its presence in capillary endothelial cells which comprise the blood-brain barrier. Tolerance to the analgesic effects of morphine in rats was recently shown to result from induction of PGP synthesis. Induction or inhibition of PGP is a drug interaction mechanism likely to be documented in future reports altering the actions of many psychoactive drugs.

Interactions Involving Distribution and Protein Binding

Almost all drugs circulate in blood, bound by some degree to specific plasma proteins, most often albumin and lipoproteins. This process presents an opportunity for drug-drug interactions to occur by one highly bound drug displacing another from its protein-binding sites. The potential consequences of this interaction can be seen in Figure 3. Normally, drug bound to protein in plasma is in equilibrium with unbound drug. It is an accepted principle of pharmacology that only unbound drug is free to diffuse to sites of action, usually in tissues, and produce pharmacologic effects. When the amount of unbound drug in plasma is increased due to displacement from proteins by another drug, more unbound drug is available to distribute to tissues, where it can produce increased pharmacologic effects.

Although several drug interactions can be shown to occur through protein-binding displacement, this type of pharmacokinetic interaction may not be significant unless the binding displacement actually modifies a drug’s dose-effect relationship. A classic example of this type of interaction is the displacement of warfarin from serum albumin-binding sites by phenylbutazone or salicylate analgesics. An increase in the plasma concentration of warfarin occurs accompanied by an increase in its pharmacologic effects, a prolongation of prothrombin time. However, as a result of more free (unbound) drug being in the systemic circulation not bound to plasma protein, more drug becomes available for hepatic metabolism. Eventually, the total concentration of warfarin in plasma returns to the pre- interaction level. This is a time-limited interaction in which homeostatic changes play a role in buffering the consequences of the increased free warfarin concentration.

Protein-binding interactions have been hypothesized to occur with most of the members of the SSRI class of antidepressants due to their high degree of plasma protein binding (>95% for some drugs); however, such interactions have not been shown to be a prevalent clinical problem. For example, sertraline produced a small increase in the free fraction of warfarin and a modest increase in prothrombin time in a study involving healthy male volunteers, but neither effect was considered to be clinically significant. The plasma binding of antidepressants and antipsychotics is generally greater to lipoproteins than to albumin, and, hence, warfarin-binding displacement interactions from albumin have been of more theoretical than practical significance. Nevertheless, these drugs may have a hypoprothrombinemic effect related to perturbations in platelet serotonin apart from any protein-binding interactions with anticoagulants. Alternatively, fluvoxamine may modify the enzymatic metabolism of warfarin, directly leading to enhanced pharmacologic effects.

Among the interactions of psychoactive drugs, the anticonvulsant mood stabilizers are most often involved in altering plasma protein binding. Valproate is highly bound to plasma proteins (>90%) and can displace the binding of diazepam, phenytoin, tolbutamide, and warfarin from their plasma albumin-binding sites. Valproate is also a weak inhibitor of several hepatic enzymes and may increase the pharmacologic effects of coadministered drugs. Overall, interactions involving protein binding occur with psychoactive drugs, but the examples are limited despite many psychoactive drugs being highly plasma protein-bound.

Interactions Involving Metabolism and/or Elimination

The liver is the primary site of elimination of most psychoactive drugs. It contains numerous Phase I and Phase II enzymes that oxidize or conjugate drugs, respectively. The most important of these enzymes in terms of understanding pharmacokinetic drug interactions is the Phase I CYP system. The majority of drug interactions of concern during the course of psychopharmacological treatment involve alterations of drug metabolism. Drug metabolism can occur in several tissues in the body, but hepatic metabolism is generally recognized as the most important, because proportionally the liver contains the highest enzyme content compared with other organs and is therefore most responsible for drug biotransformation.

Potential drug interactions involving Phase II metabolism are increasingly being recognized. The most important Phase II enzymes involved in drug metabolism are the glucuronosyltransferases. These enzymes perform conjugations by combing drug molecules with glucuronic acid, mostly in the liver. Three benzodiazepines (lorazepam, oxazepam, and temazepam) undergo Phase II reactions exclusively before being excreted into the urine. Both inducers and inhibitors of glucuronosyltransferases are known and have the potential to affect the plasma concentration and actions of important psychotropic drugs.

Drug interactions involving metabolism arise from enzyme induction or inhibition. Cigarette smoking and some specific drugs are recognized as inducers of hepatic oxidizing enzymes. The administration of these drugs can stimulate the synthesis of additional enzymes. Eventually, the increased enzyme activity results in an enhanced clearance of drugs that are substrates for the induced enzyme. Plasma drug concentration may fall, leading to diminished pharmacologic effects. An example is the treatment of a patient with carbamazepine who is taking an oral contraceptive. Carbamazepine can induce the activity of CYP 3A4, leading to increased steroid metabolism and a loss of contraceptive effect.

An interaction involving enzyme inhibition results in impaired drug clearance and a rise in plasma drug concentration. While several types of enzyme inhibition can occur, the most common is known as competitive enzyme inhibition. This occurs when two drugs have such a strong affinity for the same enzyme that one is preferentially metabolized at the expense of the other. The concentration of the drug whose elimination has been inhibited will rise with continued dosing, due to decreased clearance. The magnitude of inhibition depends upon several factors, including the affinity of the drugs for the enzyme, the drug concentration in the plasma, the degree of partitioning into hepatocytes, and others.

Interactions involving hepatic enzyme induction or inhibition are characterized by dose and time dependence. The greater the dose of an inhibitor that is administered within the range of clinically useful doses, the greater the extent of the inhibition that should occur. For example, fluoxetine is a competitive inhibitor of CYP 2D6 and should produce a greater inhibitory effect at a dose of 40 mg or 60 mg than at 20 mg/day. Eventually, increasing doses of an inhibitor will result in a maximum inhibition with no further effect from increasing doses.

Interactions involving competitive enzyme inhibition occur with the first dose of inhibitor, as it is the presence of the two competing drugs at the enzymatic site in the liver or GI tract that results in an interaction. In contrast, interactions occurring as a result of enzyme induction require several days to become apparent, as the inducing agent must stimulate the synthesis of additional metabolizing enzymes.

Drug interactions involving changes in renal elimination of drugs are infrequent with psychoactive drugs. An exception is lithium, which is totally renally cleared. Drugs and physiologic conditions that alter renal function affect lithium clearance. Foremost among the drugs that inhibit lithium clearance and increase its plasma concentration are most non-steroidal anti-inflammatory drugs (NSAIDs) and the thiazide diuretics. Drug interactions involving changes in renal elimination are unlikely to occur with the antidepressants, antipsychotics, and anxiolytics because these are highly metabolized drugs with typically less than 5% of an administered dose excreted in the urine in an unchanged form.

Prediction of Metabolic Drug Interactions

Based on an abundance of theoretical and experimental data, drug interactions as a result of competitive inhibition for the same metabolizing enzyme can be predicted. Prediction rests upon knowledge of substrate specificity for particular enzymes, the degree of affinity of a competing drug for the same enzyme, and the concentrations of the substrate and inhibitor. Mathematical equations can predict the degree of change in clearance of one drug by another under these circumstances in in vitro laboratory experiments using liver slices, intact hepatocyte preparations, or microsomes. Rarely is such complete information available for patients under clinical circumstances. In practical terms, by knowing the metabolic pathways of a drug (ie, which enzymes are involved in its metabolism) and whether a drug to be combined in therapy has inhibitory effects on that enzyme, an interaction can be predicted. The degree of interaction and whether the consequences will be clinically meaningful will depend upon multiple factors. Some of these include the specific drugs involved, drug dosage and length of therapy, and the clinical state of the patient.

While many enzymes in the liver are capable of biotransformation reactions, emphasis has focused recently on the CYP enzymes because it is estimated that collectively they participate in the metabolism of greater than 80% of all available drugs used in humans. CYP enzymes play additional roles in the metabolism of some endogenous substrates, including prostaglandins and steroids. At least 30 related enzymes are divided into different families according to their amino acid homology. Some enzymes exist in a polymorphic form, meaning that a small percentage of the population possesses mutant genes that alter the activity of the enzyme, usually by diminishing or abolishing activity. A genetic polymorphism has been well characterized with the CYP 2C19 and CYP 2D6 genes. Recently discovered but poorly categorized are polymorphisms of CYP 3A4. Table 3 lists the most important CYP enzymes, along with some of their substrates. Remarkably, for many drugs in clinical use for years, the enzymes involved in their metabolism have not been identified. This is an active research area, and information is continually being updated.

In the current approach to new drug development, candidate compounds are screened for their affinity for various P450 enzymes. A high affinity for one or more enzymes suggests a likelihood of interactions with other drugs metabolized by the same enzyme. These predictions can then be confirmed with targeted drug interaction studies in human volunteers or patients. The degree to which an interaction will occur also depends upon the concentration of the substrate and inhibitor at the enzyme site, which in turn depends upon the size of administered doses. The significance of blocking or inducing a particular cytochrome enzyme for a drug interaction will depend upon the importance of the enzyme in the overall elimination of the drug. Most drugs are eliminated through more than one pathway, and some degree of renal clearance also contributes to the elimination of many drugs. The existence of parallel pathways of elimination moderates the effects of inhibiting a single enzymatic pathway.

A qualitative approach to the prediction of drug interactions can be used by clinicians to identify the combinations of drugs that should be used cautiously or avoided, especially when preexisting information about their potential interaction is unavailable. Psychoactive drugs that inhibit or induce the enzymes listed in Table 2 would be expected to interact with the substrates of those particular enzymes. This approach provides a rough screen to predict the potential for pharmacokinetic interactions. It should be remembered that concentration changes do not necessarily translate into clinically meaningful interactions. Most drugs have acceptable therapeutic indices so that minor alterations in clearance, steady-state plasma concentration, or half-life, although statistically significant, may be clinically unimportant. Also, pharmacodynamic interactions are not predicted by this approximation and may occur in addition to or apart from pharmacokinetic interactions.

CYP Enzymes

CYP enzymes exist in a variety of body tissues, including the brain. Clearly, their presence in the GI tract (especially CYP 3A4) and in the liver is important for the elimination of administered drugs. The molecular and pharmacologic characterization of CYP enzymes and the corresponding genes that determine their synthesis is an active research area. The most prominent enzymes are discussed below, due to their importance for drug metabolism and participation in drug interactions.


The CYP 1A subfamily includes CYP 1A1 and CYP 1A2, with both genes located on human chromosome 15. CYP 1A2 is an important enzyme in the metabolism of several widely used drugs (Table 3). It comprises about 13% of the total P450 content of the human liver and is highly inducible.

Nonpsychiatric drugs metabolized by CYP 1A2 include theophylline,  aminophylline, caffeine, and the antiarrhythmic propafenone. The ß-blocker propranolol is believed to have a minor component of its biotransformation mediated by CYP 1A2. The tertiary amine tricyclic antidepressants undergo demethylation to their secondary amine active metabolites by this enzyme. The traditional antipsychotic drug haloperidol and the newer atypical antipsychotics clozapine and olanzapine are partially metabolized by CYP 1A2. Tetrahydroacridinamine (tacrine) is hydroxylated by CYP 1A2.

CYP 1A2 is induced by cigarette smoke, charcoal-broiled foods, and some cruciferous vegetables (eg, Brussels sprouts). The effect of cigarette smoking can be prominent, and patients who stop or substantially reduce smoking can be expected over the subsequent few weeks to have a return to baseline of their CYP 1A2 activity. This situation has resulted in the appearance of seizures in a patient taking clozapine who quit smoking during therapy.

Fluvoxamine and ciprofloxacin are potent inhibitors of CYP 1A2, and interactions have been described with theophylline and clozapine. One of the most notable interactions of fluvoxamine is its ability to inhibit theophylline metabolism. Because the elevation of serum theophylline could double or more, it is recommended that when this antidepressant is prescribed for a patient receiving this bronchodilator, the patient’s theophylline dose be reduced by one third of the prior dosage. Fluvoxamine is unique among the newer antidepressants in the ability to inhibit CYP 1A2. While the choice of another antidepressant in these circumstances could avoid this potential interaction, these drugs may be used safely together when dosed appropriately and cautiously. Appropriate clinical care would include monitoring of theophylline plasma concentration and vigilance to the appearance of side effects. Although other psychoactive drugs, including haloperidol, some tertiary amine tricyclic antidepressants, and olanzapine, are partially metabolized by CYP 1A2, their participation in competitive enzyme interactions appears to be a result of a stronger affinity for enzymes other than CYP 1A2.


The genes for the expression of the CYP 2A subfamily are localized on the long arm of chromosome 19. Three genes for CYP 2A6, CYP 2A7, and CYP 2A13 have been identified and sequenced. A variant allele for CYP 2A6 has been associated with individuals who are deficient in their ability to metabolize warfarin. In in vitro studies, orphenadrine decreased the activity of CYP 2A6, but the clinical significance of this effect, if any, is unknown. CYP 2A6 comprises about 4% of the P450 content of the human liver, and its contribution to the metabolism of therapeutically used drugs is probably small.


The cytochrome 2B subfamily consists of the closely related P450s 2B1, 2B2, and 2B6. CYP 2B1 has been the focus of study as it oxidizes toluene, aniline, benzene, and other solvents to reactive metabolites thought to be important in promoting carcinogenesis. It can be induced by acetone, phenobarbital, and carbamazepine. It plays a minor role in the metabolism of a few drugs used in humans, including caffeine, theophylline, coumarin, and lidocaine. In animal studies, clonazepam has been found to be a potent inhibitor of catalytic activities mediated by CYP 2B in microsomes derived from phenobarbital-pretreated rats. The MAOIs selegiline and clorgyline have been found to inactivate the activity of CYP 2B in vitro. The clinical significance of these effects is unknown.

CYP 2B6 is thought to be a minor component of P450 content in the liver, normally constituting less than 0.5% of total P450, although substantial interindividual variability has been observed. CYP 2B6 plays a role in the metabolism of the anticancer drug cyclophosphamide and is the major enzyme responsible for converting bupropion to its primary active metabolite, hydroxybupropion. Orphenadrine is a CYP 2B6 inhibitor in vitro. In a human pharmacokinetic study, carbamazepine and valproate both increased hydroxybupropion concentration, but their function as CYP 2B6 inhibitors has yet to be established.

CYP 2C9/19

The CYP 2C subfamily consists of several closely related enzymes: 2C9, 2C10, 2C19, and others. CYP 2C comprises about 18% of the total P450 content of the human liver. A genetic polymorphism exists with CYP 2C19, with approximately 18% of Japanese and African Americans reported as poor metabolizers of CYP 2C19 substrates. Only about 3% to 5% of whites inherit this deficiency. Affected individuals are identifiable by phenotyping with mephenytoin administration. Poor metabolizers have higher than normal plasma concentrations of the CYP 2C19 substrates from usual doses (Table 3). Rare polymorphisms of CYP 2C9 have been discovered.

Nonpsychiatric drugs metabolized by the CYP 2C subfamily include S-mephenytoin (2C19), phenytoin (2C19), tolbutamide (2C9), S-warfarin (2C9), ibuprofen (2C9), diclofenac (2C9), and piroxicam (2C9). Other substrates of CYP 2C9 and CYP 2C19 include diazepam, clomipramine, amitriptyline, and imipramine (Table 2). Several of the NSAIDs are substrates of CYP 2C, but clinically significant metabolic interactions with negative consequences have not been described involving psychoactive drugs combined with NSAIDs.

Several antidepressants with affinity for CYP 2C (sertraline, fluoxetine, fluvoxamine) appear to have a moderate although measurable affinity for the CYP 2C isoenzymes. The nature of the dose response curves for the NSAIDs may minimize or preclude important interactions unless substantial rises in plasma drug concentration occur. In general, drug interactions are likely to be of significance when a small increase in the concentration of an inhibited drug results in substantially increased pharmacologic effects. This situation characterizes phenytoin, and significant interactions involving this anticonvulsant with fluoxetine have been reported.


This is the best characterized of the CYP enzymes. The CYP 2D6 gene locus is on chromosome 22. A genetic polymorphism exists, with 7% to 10% of whites inheriting an autosomal recessively transmitted defective allele. Four genotypes can be distinguished: homozygous and heterozygous efficient metabolizers, homozygous poor metabolizers, and ultrarapid metabolizers carrying a duplicated or multiduplicated CYP 2D6 gene. In African Americans, the percentage of poor metabolizers is less, generally between 1% and 4%. Poor metabolizers among Asians are rare. These ethnic differences may explain different dosage requirements of some drugs in different populations.

Poor metabolizers lack sufficient functional enzyme to metabolize the CYP 2D6 substrates listed in Table 3. They can therefore be expected to have higher plasma drug concentrations and prolonged elimination half-lives of these drugs when given in usual doses. The significance of this metabolic defect is that an exaggerated pharmacologic response is possible following standard doses of drugs that are CYP 2D6 substrates.

CYP 2D6 comprises a small percentage of the total P450 content of the liver, about 1.5%, but many useful drugs are specific substrates. Nonpsychiatric drugs metabolized by CYP 2D6 include propranolol (also 1A2 and possibly 2C19), metoprolol, timolol, mexiletine, propafenone (also 1A2 and 3A4), codeine, and dextromethorphan (also 3A4). Several of the newer antidepressants are partially metabolized by CYP 2D6. They include paroxetine, venlafaxine, and fluoxetine. The tertiary amine tricyclic antidepressants are hydroxylated by CYP 2D6.

No inducers of CYP 2D6 have been identified. While CYP 2D6 substrates have shown decreased plasma concentration under conditions of cigarette smoking and barbiturate administration, this is not a laboratory-reproducible phenomenon. Alternative explanations include effects on other enzymes that mediate parallel pathways of elimination, or increases in hepatic blood flow that increase drug clearance.

Several antidepressants, discussed below, are inhibitors of CYP 2D6, but they vary widely in their potency. For example, adding fluoxetine or paroxetine to a drug regimen including desipramine will increase the plasma TCA concentration by interference with the hydroxylation pathway. Fluvoxamine, citalopram, and sertraline in low doses are less likely to exert a similar effect.


This subfamily of enzymes, with genes localized on chromosome 10, is important in the bioactivation of several carcinogens and the metabolism of organic solvents. Cytochrome 2E1 is the focus of current research for its role in alcohol metabolism. It comprises about 7% of the total P450 content of the human liver. Substrates of CYP 2E1 include chlorzoxazone, acetaminophen, halothane, enflurane, and methoxyflurane. In in vitro studies, significant inhibition of CYP 2E1 occurred with TCAs, phenothiazines, and flurazepam. Although these psychoactive drugs are not substrates for CYP 2E1, they have the potential to modulate the toxicity of nondrug xenobiotics metabolized by this isoenzyme. CYP 2E1 is induced by alcohol, which may be an important factor in its toxicity. CYP 2E1 is an active area of investigation, with limited current relevance, however, for the practice of clinical psychopharmacology.


This enzyme metabolizes the largest number of drugs used therapeutically. It constitutes approximately 30% of the P450 present in the liver and 70% of the cytochrome enzymes in the gut wall. There is little evidence for a genetic polymorphism. Everyone possesses CYP 3A4 hepatic enzyme, although there is broad variability in expressed activity among subjects. A study of the metabolism of carbamazepine suggested that CYP 3A4 activity may peak in children and show a gradual decline to adult levels of activity. This would partly account for why older children and adolescents require larger doses of some drugs than adults. The elderly, especially individuals aged 70 years and above, show a reduction in overall drug metabolism related to a decrease in CYP content, although comparative rates of decline in specific CYP enzymes are not well characterized.

Nonpsychiatric drugs metabolized by CYP 3A4 include diltiazem, verapamil (also 1A2), nifedipine, alfentanil, tamoxifen, testosterone, cortisol, progesterone, ethinyl estradiol, cisapride, cyclosporine, terfenadine, astemizole, quinidine, and the protease inhibitors (Table 3). Psychoactive drugs that are metabolized by CYP 3A4 include alprazolam, diazepam (also 2C19), triazolam, carbamazepine, nefazodone, and sertraline.

Marked enzyme induction of CYP 3A4 occurs after long-term administration of rifampin and rifabutin. Other inducers include carbamazepine, dexamethasone, and phenobarbital. Significant inhibition of CYP 3A4 substrates occurs after administration of nefazodone and fluvoxamine. The most potent inhibitors of CYP 3A4 are the azole antifungal drugs (eg, ketoconazole) and the macrolide antibiotics. A recent report of the sudden death of a child receiving pimozide who was treated with clarithromycin is a case of suspected CYP 3A4 inhibition by this antibiotic. Inhibition of terfenadine metabolism by ketoconazole, itraconazole, erythromycin, or clarithromycin poses a risk of cardiotoxicity. The noncardioactive metabolite of terfenadine, carboxyterfenadine, was recently marketed as a nonsedating antihistamine, and either this agent or loratadine is strongly preferred if a psychoactive drug must be prescribed together with an antihistamine. Among the SSRIs, paroxetine, fluoxetine, and sertraline have been specifically combined with terfenadine in in vivo pharmacokinetic studies and found not to produce a significant interaction. Fluvoxamine and nefazodone, among the newer antidepressants, are contraindicated in combination with terfenadine due to their potent CYP 3A4 isoenzyme inhibition.


The recent focus on psychotropic drug interactions has primarily emphasized the Phase I CYP system. The metabolism of drugs by Phase II reactions is accomplished by a variety of enzymes, but the emerging role of the glucuronosyltransferases as important in clinical psychopharmacology is being increasingly recognized. The uridine diphosphate-glucuronosyltransferases exist as multiple families of enzymes and have been defined with a nomenclature similar to that used to define the P450 system. The symbol UGT has been chosen to represent the superfamily of enzymes. Different UGT families are defined as having <45% amino acid sequence homology, while in subfamilies there is approximately 60% homology. As many as 33 families have been defined, with three families identified in humans. The most important of the enzymes for psychopharmacology are discussed below and listed with prominent substrates in Table 4.


The UGT 1A subfamily includes enzymes which can glucuronidate bilirubin, phenol derivatives, and estrogens. UGT 1A1 has been implicated in the metabolism of several opiate analgesics, including buprenorphine, nalorphine, and morphine. Phenobarbital and rifampin have been shown to induce UGT 1A1. Rifampin is also a PGP inducer.

UGT 1A3/1A4

Several tricyclic antidepressants undergo conjugation mediated by UGT 1A3 and UGT 1A4. In addition, chlorpromazine, lamotrigine, cyproheptadine, and zidovudine are substrates. Probenecid and valproate are inhibitors while several anticonvulsants/mood stabilizers are inducers. Olanzapine circulates in plasma, to a large extent, as a glucuronide conjugate, but the precise UGT enzymes have not been identified.


The benzodiazepines metabolized exclusively or primarily by conjugation (oxazepam, tenazepam, lorazepam) are glucuronidated by UGT 2B7, along with some opiate analgesics. A number of NSAIDS are competitive inhibitors. Phenobarbital, rifampin, and oral contraceptives appear to act as inducers of UGT 2B7.

Specific Drug Interactions

In this section, specific drug interactions are discussed for some of the major psychoactive agents in widespread clinical use. For each drug class, tables are presented that list the medications with which the drugs in the class may interact, how the drugs may interact, and the type of data that support the relevance of the interaction. Guidelines for management are also presented.

While these tables summarize the current state of our knowledge regarding interactions of psychoactive drugs, new agents are being introduced to the market at a rapid pace, and new or suspected interactions are increasingly being described in the biomedical literature each month. Suspected drug interactions generally appear first in the form of clinical case reports. This is frequently the first indication to the physician that two drugs may interact in a previously undescribed manner. The publication of several case reports of a similar nature frequently stimulates further investigation in the form of formal pharmacokinetic studies. Often, the period of time between the publication of a previously undescribed drug interaction and subsequent prospective investigation is considerable. Given the importance of case reports to the clinician, who must decide whether a particular case represents a sufficiently significant finding to merit a change in prescribing behavior, questions are posed in Table 5 as guidelines for interpretation of reports of suspected drug interactions. Consideration of these issues may be helpful in determining the potential risks or benefits of combining similar drugs.


Remarkably, TCAs are still extensively prescribed in some communities. Their generic status, allowing for relatively low cost, is a major factor in their continued prescription. Some significant interactions have been documented, which are summarized in Table 6.

The TCAs are metabolized by several P450 enzymes. CYP 1A2, 2C, and 3A4 are thought to be involved in the demethylation of the TCAs that are administered as tertiary amines (clomipramine, amitriptyline, imipramine). CYP 2D6 is involved in the hydroxylation of the secondary amine TCAs (desipramine, nortriptyline). They are further glucuronidated before being excreted in the urine. While not all TCAs have been carefully scrutinized, it can be expected that, for example, the metabolism of doxepin and trimipramine proceeds in a similar fashion.

Coadministration of the TCAs with MAOIs is contraindicated. Hyperpyretic crises or severe seizures may occur in patients receiving such combinations. At least 2 weeks should elapse between the discontinuation of an MAOI and the initiation of a TCA.

Cimetidine is a broad CYP enzyme inhibitor and has been documented to increase the plasma concentration of several TCAs. Increased side effects, including anticholinergic-induced delirium, are a possible consequence of cimetidine and other inhibitor-induced concentration elevations. All of the SSRIs have been noted in case reports to increase TCA plasma concentrations. Their relative potency in this regard is discussed in the section below. Whenever an SSRI is prescribed to a patient already receiving a TCA, caution should be exercised and the dose of the TCA reduced, if necessary.

Enzyme inducers, including cigarette smoking, carbamazepine, phenobarbital, and phenytoin, can increase the clearance of TCAs and lower their plasma concentration. Thus, in smokers, average TCA doses may be higher than in nonsmokers. Because plasma concentration measurements of the TCAs are widely available, this resource can be used to monitor the effect of adding or eliminating other drugs in a TCA-treated patient.



Drug interactions with the SSRIs have been the subject of intensive study. Five drugs are available for prescribing that vary considerably in their specificity and potency to inhibit various P450 enzymes. It was noted at an early point in the development of the SSRIs that inhibition of CYP enzymes, particularly CYP 2D6 in vitro, was a property of the majority of these drugs. Since their initial clinical use, numerous studies and reports have clarified some differences among these drugs. A summary of the inhibitory potential of the SSRIs and other newer antidepressants is provided in Table 7. The estimated potencies are based on a consideration of in vitro evidence, case reports, and formal pharmacokinetic studies. The significance of a predicted interaction in an individual patient may vary widely. A summary of the interactions with the SSRIs is provided in Table 8.

The first SSRI marketed in the United States, fluoxetine, is a potent in vitro and in vivo inhibitor of CYP 2D6. It produces an active metabolite with similar potency. The extended elimination half-life of fluoxetine and norfluoxetine means that when CYP 2D6 substrates are combined in treatment (Table 3), their metabolic elimination mediated by this enzyme can be compromised. This effect can lead to higher drug concentrations, an extended elimination half-life, and potentially increased pharmacologic effects. Interactions have been most often documented with TCAs. Fluoxetine also has some inhibitory effects on CYP 2C19, though it is not as potent an inhibitor on this enzyme as it is on CYP 2D6. Its effect on the former enzyme is sufficient to interact with diazepam and phenytoin. These drugs, therefore, should be used cautiously with fluoxetine. Fluoxetine has no recognized inhibitory potential for CYP 1A2 substrates, but its effects on CYP 3A4 are complex. A drug interaction has been noted in a pharmacokinetic study with carbamazepine, a well-documented CYP 3A4 substrate, but fluoxetine appears not to alter the metabolism of terfenadine. Fluoxetine has a potential to interact with CYP 3A4 substrates, especially as its metabolite possesses CYP 3A4 inhibition, but few reports of interactions when combined with such substrates are available.

Paroxetine is also a potent in vivo and in vitro inhibitor of CYP 2D6, and lower doses of drugs that are substrates for the isoenzyme should be used if paroxetine is combined in treatment. Paroxetine has no clinically meaningful effects on other CYP enzymes.

Sertraline is a relatively weak inhibitor of CYP 2D6, CYP 2C19, and CYP 3A4, but when used in the upper range of clinically recommended doses, it may inhibit CYP 2D6 substrates to a significant extent. This effect is inconsistent across patients but should be recognized as a possible interaction when sertraline is prescribed. The drug’s effects on tolbutamide, a CYP 2C19 substrate, were documented in a pharmacokinetic study, but clinically significant case reports involving patients are lacking.

Fluvoxamine is the only SSRI that has potent inhibitory effects on the CYP 1A2 enzyme. Interactions are documented with several substrates, including clozapine, TCAs, and theophylline. This last combination requires substantial dosage decreases of the bronchodilator to avoid potential toxicity. Fluvoxamine also inhibits CYP 2C19 and CYP 3A4 to a significant extent, and dosage modifications are recommended for some substrates, such as alprazolam.

Citalopram has been shown in a pharmacokinetic study to raise plasma concentrations of desipramine, a CYP 2D6 substrate. However, its potency as an inhibitor is quite weak, and this SSRI has the least potential to interact with P450 substrates compared to the other drugs in its class. Recently, a case was reported of citalopram combined with clomipramine in which the suspected mechanism of increased tricyclic plasma concentration was glucuronosyltransferase inhibition.

Several drugs can potentially elevate concentrations of the SSRIs. This has not been shown to be a major concern in clinical practice because patients tolerate a broad range of SSRI plasma concentrations. However, when using cimetidine or another known inhibitor in combination with an SSRI, caution should be exercised.


Other Newer Antidepressants

SJW is one of the most commonly utilized herbal agents. Available data from clinical studies and case reports suggests that SJW is unlikely to inhibit CYP 3A4 or 2D6, but it is likely an inducer of CYP 3A4 and possibly PGP. The accumulating evidence of significant drug interactions with SJW (Table 13) should serve as an example for clinicians to be aware of the potential for herbal products to participate in important herb-drug interactions. Concomitant use of herbal agents and conventional medications should be discouraged until further information is available.

Bupropion is thought to produce its antidepressant effects primarily through enhancement of noradrenergic and perhaps dopaminergic neurotransmission without any appreciable serotonergic effects. These properties should theoretically confer a low propensity to interact pharmacodynamically with other drugs to produce a serotonin syndrome. Bupropion’s proconvulsant effects in a small number of patients suggest that it should be combined cautiously with other drugs that may increase the seizure threshold, though the sustained-release form of the drug has reduced this risk. Bupropion is metabolized by multiple pathways and enzymes. Theoretically, CYP 2B1, CYP 2D6, or CYP 3A4 inhibitors could increase its clinical effects, but specific documentation is lacking. Although bupropion and its major metabolite, hydroxybupropion, are not CYP 2D6 substrates, in a healthy volunteer study one or both are potent inhibitors of this enzyme as indicated by a 2- to 5-fold rise in desipramine plasma concentration. The pharmacokinetic consequences of coadministration of bupropion with other CYP 2D6 substrates have not been published, but caution is advised for this potential interaction. Selected drug-drug interactions related to bupropion are summarized in Table 9.

Nefazodone possesses serotonergic activity as a 5-HT2 antagonist and a serotonin reuptake inhibitor. The usual precautions involving combinations of drugs resulting in excessive serotonergic activity are warranted for nefazodone. The drug is a very potent CYP 3A4 inhibitor and will theoretically inhibit the metabolism of the relevant substrates listed in Table 3. Specific interactions have been documented with alprazolam and triazolam. Nefazodone increased the plasma concentration of alprazolam 2-fold and that of triazolam 4-fold. Thus, doses of these benzodiazepines should be reduced whenever nefazodone is coadministered or when initiating anxiolytic therapy in the presence of nefazodone. One favorable report used the combination of nefazodone and alprazolam to advantage to lengthen the interdosing interval of the antipanic medication. Nefazodone’s drug interactions are summarized in Table 10.

Mirtazapine has multiple effects on serotonergic neurotransmission, acting as a 5-HT2, 5-HT3, and presynaptic α2-receptor antagonist. While mirtazapine is highly metabolized, it apparently possesses insufficient affinity for any of the specific CYP enzymes to be a meaningful metabolic inhibitor. Thus, specific interactions of this type have not been reported. Mirtazapine possesses significant sedative effects, so that in combination with other drugs producing sedation or psychomotor impairment, additive or synergistic effects are possible. Mirtazapine’s drug interactions are summarized in Table 11.

Venlafaxine is a structurally novel antidepressant that inhibits norepinephrine and serotonin reuptake, with the latter action being the more potent of the two, and predominant at lower doses. It has a low propensity for drug-drug interactions. While its active metabolite has a measurable CYP 2D6 inhibitory effect, reports of clinically significant metabolic interactions with CYP 2D6 substrates are lacking. It does, however, have the potential to interact pharmacodynamically with potent serotonergic agents, and toxicity has been reported when combined with MAOIs. Venlafaxine’s drug interactions are summarized in Table 12.


Interactions of the MAOIs are summarized in Table 14. Some unusual interactions have been reported, including their combination with meperidine or fentanyl to produce an apparent serotonin syndrome. The interactions of MAOIs with the TCAs have already been discussed. The extensive list of medications that these drugs have been reported to interact with has limited their popularity, despite their efficacy for major depression, atypical depression, panic disorder, and other anxiety syndromes.

The most feared interaction of the MAOIs has been the possible hypertensive crisis from combination with tyramine-rich foods or various OTC or prescription sympathomimetic amines. This possibility requires the counseling of patients receiving these drugs regarding the potential for diet constituents and OTC medications to interact with MAOIs.


Lithium has a very narrow therapeutic range of serum concentration associated with therapeutic effects, above which serious toxicity can occur. Lithium is renally cleared, and drugs and physiologic conditions that influence its renal elimination pose a potential risk to increase serum lithium concentration. Among the commonly used drugs that pose such a risk are thiazide diuretics, NSAIDs, and angiotensin-converting enzyme (ACE) inhibitors. They all increase plasma lithium levels.

Concomitant use of diuretics has long been associated with the development of lithium toxicity, but the risk varies with the type of diuretic. Lithium is completely filtered and then reabsorbed along the proximal renal tubule in parallel with sodium. The thiazide diuretics act distally and produce a natriuresis that leads to an increase in the reabsorption of sodium and lithium. Diuretics that act on the proximal tubule, such as furosemide, have less effect on lithium reabsorption. The degree of these interactions is variable, but a decrease in lithium dosage is almost always necessary, especially in patients receiving a thiazide diuretic.

The osmotic diuretics enhance lithium excretion and have been used in the treatment of lithium toxicity. Potassium-sparing diuretics (triamterene, amiloride, spironolactone) have exerted variable effects on lithium clearance, sometimes increasing its clearance. Theophylline and caffeine decrease lithium concentrations to a significant degree, and dosage adjustments are likely when used together.

When the NSAIDs are used with lithium, plasma concentrations can rise to a toxic level. Because some of these drugs are now available OTC, there is controversy as to whether the lower recommended OTC doses produce as dramatic a change in lithium clearance as prescribed doses. When an NSAID must be used in combination with lithium, aspirin and sulindac are recommended because they exert the least increase, if any, on lithium concentration.

Lithium toxicity has been reported with the concomitant use of ACE inhibitors and valsartan. Case series and formal pharmacokinetic evaluations document the interaction, but the precise mechanism is uncertain. Frequent monitoring of lithium concentration is recommended when these therapies are used together. The calcium channel antagonists diltiazem and verapamil have been associated with lithium toxicity through an unknown mechanism but likely involve changes in lithium’s renal clearance. These combinations require close monitoring. The continued development of anticonvulsant mood stabilizers for treatment of bipolar disorder means that some patients will receive these drugs in combination with lithium. Topiramate transiently decreased lithium concentrations when added to a lithium regimen in healthy volunteers. A similar effect in patients has not yet been reported, but closer monitoring of lithium serum concentration appears warranted when these drugs are used together. Drug-drug interactions of lithium are summarized in Table 15 .

Other Mood Stabilizers

Carbamazepine is both a substrate of CYP 3A4 and an inducer. These characteristics account for the autoinduction and decrease in its plasma concentration observed several weeks following initiation of dosing. As a CYP 3A4 substrate, carbamazepine’s clearance and plasma concentration are subject to change in the presence of inhibitors, including valproate, nefazodone, cimetidine, and others. Erythromycin can significantly increase carbamazepine concentration and produce signs of toxicity. These commonly include confusion, sedation, and ataxia. Should these appear, dosage should be decreased and plasma drug concentration should be assessed for subsequent monitoring. Valproate is often combined with carbamazepine and it may slightly impair carbamazepine clearance; carbamazepine may decrease valproate concentration. This situation requires plasma concentration monitoring of both drugs to avoid excessive concentration changes, and therefore guides dosing. Carbamazepine added to a regimen of lamotrigine decreased the latter’s plasma concentration by 40%, but lamotrigine had no effect on carbamazepine concentration. The concentration of carbamazepine epoxide was increased in one study, so plasma concentration monitoring is recommended if these drugs are used concurrently. Carbamazepine and gabapentin do not affect each other’s disposition.

Carbamazepine has been reported to decrease the concentration of other CYP 3A4 substrates as a result of its enzyme-inducing effects. Some dosage adjustments may be necessary. A significant interaction is the well-described effect of diminishing the concentration of oral contraceptives. These interactions are summarized in Table 16.

Oxcarbazepine, structurally related to carbamazepine, appears to be as effective as carbamazepine in the treatment of epilepsy and slightly better tolerated. Thus, it may find utility as a mood stabilizer alternative to carbamazepine. It appears to possess dose-dependent enzyme induction, like carbamazepine, and may participate in a variety of similar drug interactions (Table 17).

Valproate’s interactions (see Table 18) result from mild enzyme inhibition and the additional capacity to displace other drugs from their plasma protein-binding sites. Caution is warranted when combining valproate with aspirin, because the free fraction of valproate may increase dramatically (see Figure 3). This may not be reflected by an increased measurement of total drug concentration in plasma. In turn, valproate may increase the anticoagulant effects of aspirin.

The precise interactions between valproate and specific CYP isoenzymes are unclear. It inhibits glucuronosyltransferase, as evidenced by an effect on zidovudine and lorazepam, as well as producing apparent inhibitory effects on substrates of CYP 2C9 and CYP 2C19 (phenytoin and diazepam). Its interactions with other mood stabilizers are complex. An interaction with phenytoin may result from both a metabolic inhibition and an increased concentration of unbound phenytoin, but without an apparent increase in total drug concentration. When lamotrigine was added to existing valproate therapy, valproate concentrations decreased by 25%. When valproate was added to lamotrigine therapy, lamotrigine concentrations increased 2-fold. These changes suggest that close monitoring of combined mood stabilizer therapy is necessary to optimize treatment and avoid adverse effects. Gabapentin pharmacokinetic parameters are unaffected by valproate.

Lamotrigine is metabolized predominantly by conjugation with glucuronic acid, a Phase II metabolic process by 1A4, with little or no involvement of CYP enzymes. The drug has not been reported to affect CYP enzymes. Its interactions have only been systematically studied with the common anticonvulsants. With the exception of valproate, the addition of lamotrigine to other mood stabilizers does not affect their steady-state plasma concentration. No significant effect was noted after the addition of lamotrigine to regimens of phenytoin or carbamazepine. As noted above, lamotrigine decreased valproate concentration. Phenytoin and carbamazepine decrease and valproate increases concentrations of lamotrigine. Lamotrigine is approximately 55% bound to human plasma proteins, so drug interactions secondary to binding displacement are not expected. No clinical value has yet been shown from monitoring plasma concentrations of lamotrigine. Its potential interactions with other drugs should be monitored by close clinical observation. The drug interactions of lamotrigine are summarized in Table 19.

Topiramate is an anticonvulsant with possible mood-stabilizing effects. When combined with other anticonvulsants, such as carbamazepine, phenobarbital, or primodone, topiramate has no effect on their concentrations. Nor does it have clinically relevant effects on plasma levels of classical neuroleptics, TCAs, theophylline, and warfarin. However, concomitant use of this compound with central nervous system (CNS) depressants can cause excessive sedation. When combined with acetazolamide or other carbonic anhydrase inhibitors, it can increase the risk of renal stones. Also, topiramate can interfere with the efficacy of contraceptive medication by decreasing levels of ethinyl estradiol by one third (Table 20).

Gabapentin has been reported to have mood stabilizing effects and to be effective for social phobia. Gabapentin is not metabolized by the liver and has no significant pharmacokinetic interactions. Its elimination is reduced in patients with impaired renal function. Gabapentin does not interact with hepatic enzymes, causing neither inhibition nor induction.


The psychostimulants methyl-phenidate, dextroamphetamine, and pemoline are among the most common medications used in child and adolescent psychiatry, and are often used in combination with other medications. A variety of case reports describe suspected metabolic drug interactions, but sparse data from systematic study are available. Methylphenidate appears to be involved primarily in pharmacokinetic interactions suggestive of CYP inhibition, while dextroamphetamine and pemoline are more often involved in apparent pharmacodynamic interactions. Selected interactions are summarized in Table 21.

Methylphenidate is highly metabolized but the specific enzymes involved have not been characterized. A pharmacokinetic interaction study observing methylphenidate concentration with and without quinidine found no evidence for the involvement of CYP 2D6 in its metabolism. Methylphenidate plasma concentration monitoring is seldom practiced clinically. The drug’s reported interactions all involve the effect of methylphenidate on the disposition of other drugs. No reports have been published that document alterations in methylphenidate concentration. Potential drug interactions should be monitored by careful patient observation of signs and symptoms suggestive of enhanced or diminished effects.

Modafinil is a recently introduced psychostimulant labeled for the treatment of narcolepsy. It may find use as a treatment for attention-deficit/hyperactivity disorder and other conditions. In vitro examination of its enzyme inductive/inhibitory effects has found little evidence for potential drug interactions (Table 22).


The drugs used as anxiolytics are primarily the benzodiazepines and buspirone. The benzodiazepines zolpidem and zaleplon are used as hypnotics. Their interactions are summarized in Tables 23, 24, and 25. The benzodiazepines increase the sedative and CNS-depressive effects of other drugs. Some metabolic interactions have been documented (eg, alprazolam and diazepam concentrations increased when coadministered with CYP 3A4 inhibitor/antidepressants—nefazodone, fluoxetine, and fluvoxamine). Dosage adjustments are necessary to avoid excessive effects. These interactions usually present clinically as an exaggeration of the expected pharmacologic effects (Table 26).

Zaleplon is a hypnotic agent indicated for the short-term management of insomnia. It is metabolized by CYP 3A4 with a short half-life of 1 hour. It has been shown to lack any pharamokinetic interaction with digoxin, ibuprofen, or thioridazine; however, it had an additive pharacodynamic effect with thioridazine on psychomotor testing. The short half-life of zaleplon should preclude most clinically significant interactions with CYP 3A4 inhibitors. Considerations that apply to zolpidem influence by CYP 3A4 inducers and inhibitors shoud also apply to zaleplon. In combination with alcohol or other CNS depressants, enhanced residual effects should be kept in mind.


Antipsychotic Agents

Drug interactions involving the conventional and atypical antipsychotics are summarized in Tables 25–31. These are all highly metabolized drugs producing multiple metabolites. The specific oxidizing enzymes for the metabolism of haloperidol and the atypical drugs have been reported, but fewer data are available for the older conventional drugs from which to predict drug-drug interactions. Hence, the interactions of the phenothiazines are grouped together while haloperidol and the newer drugs are considered separately.

Numerous drug interactions have been reported with the conventional antipsychotics. Antacids and anticholinergics may reduce their absorption. Formal pharmacokinetic studies have revealed mutual metabolic interactions with the TCAs, but dosage adjustments as a result are rarely considered in clinical practice. As these drugs are likely metabolized by several P450 enzymes, broad enzyme inducers, such as barbiturates, and inhibitors, such as cimetidine, predictably lead to altered plasma concentrations in the expected direction.

The metabolism of haloperidol has been studied for more than 30 years. One metabolite, reduced haloperidol, possesses 10% to 20% of the pharmacologic activity of haloperidol. The interconversion of haloperidol with its metabolite was initially hypothesized to involve CYP 2D6, based on evidence that haloperidol is apparently a CYP 2D6 inhibitor. Subsequent studies with poor and extensive CYP 2D6 metabolizers have failed to confirm evidence for CYP 2D6 involvement. There is more substantial evidence of CYP 3A4 and CYP 1A2 involvement in the metabolism of haloperidol. Rifampin, a potent CYP 3A4 inducer, decreases the concentration of haloperidol, as does carbamazepine. Nefazodone increases its concentration, as do fluoxetine and fluvoxamine, agents with CYP 3A4 inhibitory effects. Reduced haloperidol has recently been shown to be a potent CYP 2D6 inhibitor, which suggests a basis for interactions of haloperidol and CYP 2D6 substrates. Although long known to cause dose-related QTc interval prolongation, the package insert of Mellaril (thioridazine) was recently changed to reflect warnings that the CYP 2D6-mediated metabolism of thioridazine results in elevated drug plasma concentrations in patients with CYP 2D6 deficiency or in patients receiving drugs that potently inhibit CYP 2D6. Thioridazine is now contraindicated by its manufacturer with certain other drugs, including fluvoxamine, propranolol, pindolol, and any drug that inhibits CYP 2D6 (paroxetine, fluoxetine, quiaidine).

Clozapine was the first atypical antipsychotic marketed in the US. It undergoes extensive hepatic metabolism to over 10 metabolites in humans. Multiple CYP enzymes are involved in its metabolism; however, two prominent enzymes are CYP 1A2 and CYP 3A4. There is less evidence for involvement of CYP 2D6. Clozapine disposition was found to co-vary with CYP 1A2 activity, and fluvoxamine has caused robust increases in clozapine and desmethylclozapine plasma concentrations. Sertraline, paroxetine, and fluoxetine have been reported to increase plasma concentrations of clozapine. Reports are available in which coadministration of erythromycin, a relatively specific inhibitor of CYP 3A4, resulted in significant increase in clozapine concentration. Additionally, coadministration of clozapine with carbamazepine and rifampin has been shown to diminish clozapine concentration. Because the plasma concentration of clozapine has been related to its antipsychotic effect in more than six controlled studies, concomitant use with inducers or inhibitors should be accompanied by plasma concentration and clinical monitoring.

Risperidone produces a pharmacologically active metabolite, 9-hydroxy-risperidone, mediated by the actions of CYP 2D6. Its formation is highly correlated with the patient’s phenotype. Combining risperidone and its metabolite in poor or extensive CYP 2D6 metabolizers did not affect the overall pharmacologic effects. These findings suggest that CYP 2D6 inhibitors will interact to alter the plasma concentration of risperidone, but its effects may be unchanged. No routine dosage adjustments are recommended for coadministration of risperidone with CYP 2D6 inhibitors. An interaction with carbamazepine has been reported by the manufacturer, but confirmatory reports of patient complications are lacking. Risperidone’s metabolism is mediated to a minor degree by CYP 3A4. Drugs that induce/inhibit CYP 3A4 may alter risperidone plasma concentrations, but the clinical significance of such interactions appears to be minimal. Multiple studies and case reports document a lack of significant problems when combining risperidone with SSRIs. Overall, risperidone appears to have a relatively benign drug interaction profile.

Olanzapine undergoes extensive hepatic metabolism, with at least 10 metabolites identified. Principal enzymatic pathways involve CYP 1A2 and glucuronidation. Although plasma concentration monitoring of olanzapine is not a routine clinical procedure, preliminary data suggest that plasma concentrations may predict clinical response. Theoretical drug interactions with olanzapine can be proposed, but few actual reports are available.

In vitro studies indicate that CYP 3A4 is the primary enzyme involved in the metabolism of quetiapine. A lesser role has been found for CYP 2D6. Coadministration of the CYP 3A4 inducer phenytoin resulted in a 5-fold increase in the clearance of quetiapine; however, coadministration of cimetidine did not significantly affect its steady-state concentration. Unexpectedly, thioridazine, which is regarded as a CYP 2D6 inhibitor, decreased the concentration of quetiapine. Other interactions are theoretical involving CYP 3A4 inducers or inhibitors. Because the plasma concentration of quetiapine has not been reported to be correlated with clinical responses, monitoring cannot be recommended at the present time.

Ziprasidone has been introduced for oral administration as an antipsychotic. Its major routes of elimination include metabolism by a non-P450 enzyme, aldehyde oxidase, and CYP 3A4 and CYP 1A2 oxidation. Ziprasidone had little in vitro inhibitory effects on the major P450 enzymes and would be expected to participate in few pharmacokinetic interactions (Table 32).

Cholinesterase Inhibitors

There are currently three cholinesterase inhibitors available for the treatment of Alzheimer’s disease (AD): donepezil, tacrine, and rivastrigmine. These drugs work by enhancing cholinergic function, and are based on theories that some AD symptoms are due to a deficiency in cholinergic neurotransmission. Due to this mechanism of action, these drugs will interfere with and be counteracted by the activity of any anticholinergic medications, and this combination should therefore be avoided. Similarly, a synergistic effect may be expected when cholinesterase inhibitors are given concurrently with succinylcholine, similar neuromuscular blocking agents, or cholinergic agonists such as bethanechol. They are, therefore, likely to exaggerate succinylcholine-type muscle relaxation during anesthesia, and a clinically appropriate washout period is recommended. No in vivo clinical trials have investigated the effect of donepezil on the clearance of cisapride, terfenadine (CYP 3A3/4), or CYP 2D6 substrates. However, in vitro studies show a low rate of binding to these enzymes, which indicates little likelihood of interference. Ketoconazole and quinidine, inhibitors of CYP 450, 3A4, and 2D6, respectively, inhibit donepezil metabolism in vitro. Whether there is a clinical effect is unknown. Inducers of CYP 2D6 and CYP 3A4 (eg, phenytoin, carbamazepine, dexamethasone, rifampin, and phenobarbital) could increase the rate of elimination of donepezil.

Coadministration of tacrine with theophylline increases theophylline plasma concentrations via competition with CYP 1A2. Theophylline concentration levels should therefore be monitored upon coadministration, and the dose of theophylline should be reduced as necessary. Formal interaction studies suggest that donepezil does not have a significant interaction with digoxin, warfarin, theophylline, and cimetidine. Rivastigmine is minimally metabolized by CYP enzymes, has low protein binding, a short plasma half-life, and a relatively short duration of action. Combination with a variety of drugs has not revealed any significant pattern of pharmacodynamic drug interactions. Rivastigmine is not thought to have CYP drug interactions. No pharmacokinetic interactions were apparent with diazepam, digoxin, fluoxetine, or warfarin. Selected drug interactions related to the cholinesterase inhibitors used in the treatment of AD are summarized in Table 33.


Anorectic/Anti-Obesity Agents

The anorectic agents should not be administered with MAOIs. It is advised to wait 14 days following the administration of an MAOI before taking these drugs.

Phentermine may decrease the hypotensive effect of adrenergic neuron-blocking drugs such as guanethidine. Combination with phentermine may result in overstimulation, restlessness, dizziness, insomnia, or tremors at some doses. Phentermine may alter insulin requirements for patients with diabetes mellitus. Related drug interactions are highlighted in Table 34.

Sibutramine, a newer anorectic agent that works as a sympathomimetic amine, is expected to have side effects similar to other anorectic agents. It has potential for causing hypertension, should not be combined with MAOIs, and may cause serotonin syndrome when combined with SSRIs.

Orlistat, a new selective inhibitor of GI lipases, reduces dietary fat absorption and could potentially interfere with the absorption of coadministered drugs. It has been shown not to affect the absorption of oral contraceptives, nifedipine, atenolol, furosemide, captopril, phenytoin, warfarin, and vitamin A. It did significantly reduce the absorption of vitamin E, which is taken by some patients for treatment of movement disorders. The influence, if any, on absorption of other drugs taken for psychotropic effects has not been reported.


Methadone is a synthetic opiate agonist that is used in psychiatry primarily in the detoxification and maintenance treatment of opiate addiction, as well as in chronic pain management programs. Despite the therapeutic use of methadone for nearly 50 years, details of its pharmacokinetics are incomplete. Consequently, regimens for methadone are often empirical, titrating dosage against clinical response. Methadone appears to be metabolized extensively by CYP 3A4 and secondarily by CYP 2D6. Methadone is a mild in vitro inhibitor of CYP 2D6, which explains its ability to increase desipramine plasma concentration. It has also blocked nifedipine oxidation, a CYP 3A4 pathway in vitro, but case reports of methadone inhibiting CYP 3A4 substrates are lacking. Fluvoxamine, more potently than fluoxetine, increased methadone plasma concentration when added to chronic therapy. Thus, any CYP 3A4 inhibitors should be used with caution in patients treated with methadone. Table 35 lists selected methadone interactions.


Clinicians need to be alert for possible interactions in patients using multiple drugs. Many drug interactions probably cause subtle effects that are not recognized clinically. Most drug interactions are not life-threatening. Nevertheless, some interactions cause side effects that interfere with compliance or cause a decrease in drug efficacy. Whatever their consequences, drug-drug interactions represent a major public health concern. Preventable drug therapy problems increase medical costs by nearly $100 billion annually, and about 20% of that additional cost is attributed to drug-drug interactions.

Much of the emphasis on drug interactions focuses on the CYP system. The importance of other factors as determinants of plasma drug concentrations is underscored by findings involving serum protein binding and extrahepatic drug disposition. For example, serum α-1-acid glycoprotein, a serum protein to which drugs bind, fluctuates in various disorders. It is elevated in depression, arthritis, and autoimmune disorders. These elevated levels alter the disposition and actions of highly bound drugs, such as the TCAs and the SSRIs. The lungs have also been found to function as a reservoir for drugs, with high affinity for the serotonin transporter. Another agent may displace an antidepressant that has accumulated in the lungs, with a resultant increase in plasma concentrations and possible toxicity.

No discussion of potential psychotropic drug interactions can be all-inclusive. Current understanding of the variables that contribute to drug pharmacokinetics and pharmacogenetics is incomplete, and no interactions can be predicted or ruled out with absolute certainty. Drugs known to be potent enzyme inhibitors may fail to produce a predicted interaction, while a supposedly “clean” drug can cause a fatal interaction. New information emerges daily. Readers are encouraged to supplement this article with other sources and to be familiar with drug interactions listed in the product information sheets included in the packaging of each drug they prescribe.   PP


1. Akula SK, Rege AB, Dreisbach AW, Dejace PM, Lertora JJ. Valproic acid increases cerebrospinal fluid zidovudine levels in a patient with AIDS. Am J Med Sci. 1997;313:244-246.
2. Apseloff G, Wilner KD, Gerber N, Tremaine LM. Effect of sertraline on protein binding of warfarin. Clin Pharmacokinet. 1997;32(suppl 1):37-42.
3. Barbhaya RH, Shukla UA, Kroboth PD, Greene DS. Co-administration of nefazodone and benzodiazepines: a pharmacokinetic interaction study with triazolam. J Clin Psychopharmacol. 1995;15:320-326.
4. Bergstorm RF, Goldberg MJ, Cerimele BJ, Hatcher BL. Assessment of the potential for pharmacokinetic interaction between fluoxetine and terfenadine. Clin Pharmacol Ther. 1997;62:643-651.
5. Bergstrom RF, Peyton AL, Lemberger L, et al. Quantification and mechanism of the fluoxetine and tricyclic antidepressant interaction. Clin Pharmacol Ther. 1992;51:239-248.
6. Bertz RJ, Granneman GR. Use of in vitro and in vivo data to estimate the likelihood of metabolic pharmacokinetic interactions. Clin Pharmacokinet. 1997;32:210-258.
7. Brosen K. Recent developments in hepatic drug oxidation: implications for clinical pharmacokinetics. Clin Pharmacokinet. 1990;18:220-239.
8. Curry SH, DeVane CL, Wolfe MM. Cimetidine interactions with amitriptyline. Eur J Clin Pharmacol. 1985;29:429-433.
9. DeVane CL. Principles of pharmacokinetics and pharmacodynamics. In: Schatzberg AF, Nemeroff CB, eds. Textbook of Psychopharmacology. 2nd ed. New York, NY: APA Press. 1998:155-169.
10. DeVane CL. Clinical implications of dose-dependent cytochrome P450 drug-drug interactions with antidepressants. Hum Psychopharmacol. 1998;13:329-336.
11. Edge SC, Markowitz JS, DeVane CL. Clozapine drug-drug interactions: a review of the literature. Hum Psychopharmacol. 1997;12:5-20.
12. Ereshefsky L, Riesenman C, Lam YW. Antidepressant drug interactions and the cytochrome P450 system: the role of cytochrome P450 2D6. Clin Pharmacokinet. 1995;29(suppl 1):10-19.
13. Finley PR, Warner MD, Peabody CA. Clinical relevance of drug interactions with lithium. Clin Pharmacokinet. 1995;29:172-191.
14. Flockhart DA, Richard E, Woosley RL, Pearle PL, Drici MD. A metabolic interaction between clarithromycin and pimozide may result in cardiac toxicity [abstract]. Clin Pharmacol Ther. 1996;59:189.
15. Greene DS, Salazar DE, Dockers RC, Kroboth P, Barbhaya RH. Co-administration of nefazodone and benzodiazepines: III. A pharmacokinetic interaction study with alprazam. J Clin Psychopharmacol. 1995;15:399-408.
16. Ketter TA, Callahan AM, Post R. Nefazodone relief of alprazolam interdose dysphoria: a potential therapeutic benefit of 3A3/4 inhibition [letter]. J Clin Psychiatry. 1996;57:307.
17. Ketter TA, Flockhart DA, Post RM, et al. The emerging role of cytochrome P450 3A in psychopharmacology. J Clin Psychopharmacol. 1995;15:387-398.
18. Korinthenberg R, Haug C, Hannak D. The metabolism of carbamazepine to CBZ-10,11 epoxide in children from the newborn age to adolescence. Neuropediatrics. 1994;25:214-216.
19. Martin DE, Zussman BD, Everitt DE, Benincosa LJ, Etheredge RC, Jorkasky DK. Paroxetine does not affect the cardiac safety and pharmacokinetics of trefenadine in healthy adult men. J Clin Psychopharmacol. 1997;17:451-459.
20. Maynard GL, Soni P. Thioridazine interferences with imipramine metabolism and measurement. Ther Drug Monitor. 1996;18:729-731.
21. McCarthy R. Seizure following smoking cessation in a clozapine responder. Pharmacopsychiatry. 1994;27:210-211.
22. Nemeroff CB, DeVane CL, Pollock BG. Newer antidepressants and the cytochrome P450 system. Am J Psychiatry. 1996;153:311-320.
23. Owen JR, Nemeroff CB. New antidepressants and the cytochrome P450 system: focus on venlafaxine, nefazodone, and mirtazapine. Depress Anxiety. 1998;7(suppl 1):24-32.
24. Preskorn SH, Alderman J, Chung M, et al. Pharmacokinetics of desipramine co-administered with sertraline or fluoxetine. J Clin Psychopharmacol. 1994;14:90-98.
25. Rau SE, Bend JR, Arnold MO, Tran LT, Spence JD, Bailey DG. Grapefruit juice—terfenadine single-dose interaction: magnitude, mechanism, and relevance. Clin Pharmacol Ther. 1997;61:401-409.
26. Spina E, Pisani F, Perucca E. Clinically significant pharmacokinetic drug interactions with carbamazepine: an update. Clin Pharmacokinet. 1996;31:198-214.
27. Sporer KA.?The serotonin syndrome: implicated drugs, pathophysiology and management. Drug Safety. 1995;13:94-104.
28. Suhara T, Sudo Y, Yoshida K, et al. Lung as reservoir for antidepressants in pharmacokinetic drug interactions. Lancet. 1998;351:332-335.
29. von Moltke LL, Greenblatt DJ, Cotreau-Bibbo MM, et al. Inhibition of desipramine hydroxylation in vitro by serotonin-reuptake inhibitor antidepressants, and by quinidine and ketoconazole: a model system to predict drug interactions in vivo. J Pharmacol Exp Ther. 1994;268:1278-1283.
30. Wilkinson GR. Plasma and tissue binding considerations in drug disposition. Drug Metab Rev. 1983;14:427.
31. Wong SL, Cavanaugh J, Shi H, Awni WM, Granneman GR. Effects of divalproex sodium on amitriptyline and nortriptyline pharmacokinetics. Clin Pharmacol Ther. 1996;60:48-53.
32. Wrighton SA, Stevens JC. The human hepatic cytochromes P450 involved in drug metabolism. Crit Rev Toxicol. 1992;22:1-21.
33. Chang TKH, Weber GF, Crespi CL, Waxman DJ. Differential activation of cyclophoshamide and ifosphamide by cytochromes P450 2B and 3A in human liver microsomes. Can Res. 1993;53:5629-5637.
34. Liston HL, Markowitz JS, DeVane CL. Drug glucuronidation in clinical psychopharmacology. J Clin Psychopharmacol. 2001;21:500-515.
35. Miners JO, Mackenzie PI. Drug glucuronidation in humans. Pharmacol Ther. 1991;51:347-369.
36. Green MD, Tephley TR. Glucuronidation of amine substrates by purified and expressed UDP-glucuronosyltransferase proteins. Drug Metab Disp. 1998;26:860-867.
37. Greenblatt DJ, von Moltke LL, Harmatz JS, et al. Kinetic and dynamic interaction study of zolpidem with ketoconazole, itraconazole, and fluconazole. Clin Pharmacol Ther. 1998;64:661-671.
38. Hartman D, Guzelhan C, Zuiderwijk PB, Odink J. Lack of interaction between orlistat and oral contraceptives. Eur J Clin Pharmacol. 1996;50:421-424.
39. Weber C, Tam YK, Schmidke-Schrezenmeier G, Jonkmann JH, van Brummelen P. Effect of the lipase inhibitor orlistat on the pharmacokinetics of four different antihypertensive drugs in healthy volunteers. J Clin Pharmacol. 1996;51:87-90.
40. Melia AT, Mulligan TE, Zhi J. Lack of effect of orlistat on the bioavailability of a single dose of nifedipine extended-release tablets in healthy volunteers. J Clin Pharmacol. 1996;36:352-355.
41. Zhi J, Melia AT, Guerciolini R, et al. The effect of orlistat on the pharmacokinetics and pharmacodynamics of warfarin in healthy volunteers. J Clin Pharmacol. 1996;36:659-666.
42. Melia AT, Mulligan TE, Zhi J. The effect of orlistat on the pharmacokinetics of phenytoin in healthy volunteers. J Clin Pharmacol. 1996;36:654-658.
43. Melia AT, Koss-Twardy SG, Zhi J. The effect of orlistat, an inhibitor of dietary fat absorption, on the absorption of vitamin A and E in healthy volunteers. J Clin Pharmacol. 1996;647-653.
44. Haffen E, Vandel P, Bgonin B, Vandel S. Citalopram pharmacokinetic interaction with clomipramine. UDP-glucuronosyltransferase inhibition??A case report. Therapie. 1999:54:767-770.
45. Johne A, Brockmoller J, Bauer S, Maurer A, Langheinrich M, Roots I. Pharmacokinetic interaction of digoxin with an herbal extract from St. John’s wort (hypericum perforatum). Clin Pharmacol Ther. 1999;66:338-345.
46. Piscitelli SC, Burstein AH, Chaitt D, Alfaro RM, Falloon J. Indinavir concentrations and St John’s wort. Lancet. 2000;355(9203):547-548.
47. Nebel A, Schneider BJ, Baker RK, Kroll DJ. Potential metabolic interaction between St. John’s wort and theophyline. Ann Pharmacotherapy. 1999;33:502.
48. Ruschitzka F, Meier PJ, Turina M, Luscher TF, Noll G. Acute heart transplant rejection due to St. John’s wort. Lancet. 2000;355(9203):548-549.
49. Markowitz JS, DeVane CL, Boulton DW, Carson SW, Nahas Z, Risch SC. Effect of St. John’s wort (hypericum perforatum) on cytochrome P-450 2D6 and 3A4 activity in healthy volunteers. Life Sci. 2000;66:PL133-139.



Print Friendly 

Antidepressant Withdrawal Syndrome in Two Patients Taking Fluoxetine

John Norton, MD


Primary Psychiatry. 2003;10(3):51-52