Steven J. Kingsbury, MD, PhD

A great deal of data exists about the dangers of polypharmacy. Persons with psychiatric disorders experience increased risk for adverse drug interactions because of the great frequency with which multiple medications are used.

 Using multiple antipsychotics concomitantly has been associated with increased mortality in patients with schizophrenia.

 Reports of adverse psychiatric polypharmacy effects are abundant, including increased duration of hospital stay.

In this article we define polypharmacy as the use of 2 or more medications of the same chemical class or with the same or similar pharmacological actions to treat the same condition or separate conditions. For example, prescribing thyroid medication to augment antidepressant response would be polypharmacy, while using thyroid medication to treat hypothyroidism, with concomitant antidepressant therapy for depression would not be considered polypharmacy (

Although the use of multiple medications to treat different conditions is beyond the scope of this paper, it should be realized that there is a potential for problems any time multiple medications are used. As an example, cases of carbamazepine toxicity have been reported when this mood stabilizer is used in combination with protease inhibitors for HIV because of the well-known CYP3A4 inhibition of this antiretro- viral class.

A medication that treats the adverse effects of another agent would not be considered polypharmacy by this definition. For instance, using anticholinergic agents to treat extrapyramidal side effects of first-generation antipsychotics or the use of amiloride to treat lithium-induced polyuria may be empirically based. However, the need for 2 medications might be avoided with the use of second-generation antipsychotics in the first example or valproic acid in the second example.

The term polypharmacy does not in itself suggest whether its use is warranted. Ideally, only one medication would be used to optimally treat a symptom or disorder. However, even a cursory review of the literature suggests that many, if not most, patients show suboptimal response to any one given agent. New medications more specifically target disease state symptoms and have improved side-effect profiles, making combination treatment strategies more attractive and possibly less daunting. The legitimate increase in pressures to help improve patient status has changed the standard for successful treatment in psychiatry; the bar has been raised.

Perhaps the best example of this phenomenon is in the treatment of depression. Response is no longer considered an acceptable goal-remission is now the clear end point for therapy. Multiple lines of evidence show that patients who only respond have a poorer quality of life and are more likely to relapse.

There are certainly some well-researched uses of medication combinations. Even casual reflection shows that the number of possible medication combinations to treat any psychiatric condition is enormous. Surveys have found that the use of 4 or 5 medications simultaneously is not uncommon in the treatment of bipolar disorder

 Given the overwhelming number of possible combinations, it is unlikely that more than a few of these will be researched.

Although double-blind, placebo-controlled trials are considered the standard, relatively few such studies of polypharmacy exist. Therefore, clinicians must venture beyond this evidentiary base. Less firm evidence may be gathered from clinical series and expert opinions, which may not be completely current nor deal with the specific clinical dilemma being faced. Therefore, alternative ways of approaching this problem are needed (


Two polypharmacy strategies may be considered good. The first involves using a combination of 2 medications based on research showing that the combination is more effective than the use of either one alone. Since such studies are in short supply, the second strategy, rational polypharmacy, would be to extrapolate the combined pharmacodynamics of concomitant medications from existing research results, expert opinions, or hypotheses.

Perhaps the largest group of research studies supporting polypharmacy exists for manic episodes in bipolar disorder. The combination use of either lithium or valproate with the second-generation antipsychotics olanzapine, quetiapine, or risperidone has proved to be effective in the treatment of manic episodes.

 Therefore, since research supports using 3 of the second-generation antipsychotics in combination with mood stabilizers for mania, it would seem reasonable to generalize these findings as applicable to the other 2 second-generation antipsychotics, aripiprazole and ziprasidone. Moreover, they too are approved for the treatment of manic episodes in bipolar disorder, so combining them with a mood stabilizer would be rational polypharmacy.

Various algorithms exist reflecting both research and clinical experience from authorities and panels of experts. These guidelines frequently suggest rational approaches to the concomitant use of more than 2 drugs. Post and colleagues

 report on the complexity and sophistication necessary for such combinations, particularly in the treatment of bipolar disorder.

Knowledge about the biology of a disorder and the pharmacodynamic properties of medications can be foundations on which rational polypharmacy regimens can be built. For example, the addition of bupropion, which possesses norepinephrine and dopamine reuptake inhibition, to an SSRI would appear rational in the treatment of depression, despite the lack of firm double-blind, placebo-controlled evidence that this combination is effective.

It should be realized that regardless of the rationale or quality of the research, the combination chosen might not be effective for a given patient. Therefore, it is incumbent upon the clinician to carefully monitor the effects of such combinations. Although the use of detailed research scales would be ideal, time constraints in most practices mitigate against this. However, the use of selected items from such scales or the use of a clinical global impression scale would be helpful in avoiding long-term use of additional medications that provide little or no benefit.

Schizophrenia is one area where such monitoring is necessary because polypharmacy research is either lacking or contradictory. For example, although there are case studies suggesting that the addition of risperidone may augment response to clozapine, the only double-blind studies of this combination report contradictory results.

Since some positive results exist, using risperidone and clozapine together may be beneficial, but the combination requires monitoring for symptom improvement to ensure the adequacy of such treatment. We have often been consulted about cases of schizophrenia in which patients are taking several antipsychotics simultaneously and the clinician is seeking to add another. When asked whether they had considered a trial of clozapine alone for the patient, the answer is invariably no.


For the purposes of this article, bad polypharmacy will refer to the combination of 2 or more medications without attention to pharmacodynamic properties. Inattention to pharmacokinetic properties will be discussed in the next section.

Several reasons for bad polypharmacy have been described.

 One example is clinicians' fear of "rocking the boat," so they refrain from discontinuing any current medications that are partially effective. An additional medication is then added to the existing regimen in an attempt to achieve full therapeutic effects. In these cases, the additional medication may be sufficient as monotherapy, which would make a medication switch more appropriate than a combination approach.

Another cause of bad polypharmacy deserves mention: following fads. For example, we have frequently seen the addition of low-dose quetiapine for sleep to suboptimal doses of antipsychotics or mood stabilizers. Although clearly more expensive, it is unclear whether the use of low-dose quetiapine provides any greater benefit than a pure antihistamine agent. Another recent example we observed was the addition of gabapentin to a large variety of medications in the hope of augmenting response in schizophrenia, bipolar disorder, depression, and other conditions despite an absence of sound evidence for its efficacy.

We have unfortunately observed plentiful inattention to the combined pharmacodynamic properties of medications. An example is the use of 2 SSRIs, both at suboptimal doses. Another is the addition of a serotonin- norepinephrine reuptake inhibitor (SNRI), such as venlafaxine, to an SSRI when the serotonin reuptake inhibition of the SNRI alone may be sufficient.

Another recently observed example was the combined use of an anticholinergic agent with donepezil. Obviously, the cholinergic antagonist should cancel out any benefits of a cholinesterase inhibitor. This, again, shows inattention to the combined pharmacodynamic effects.


In contrast to polypharmacy that is mostly wasteful or ineffective, as described in the previous section, the "ugly" refers to polypharmacy that may be harmful. Two major errors may lead to harm: inattention to pharmacokinetic interactions and inattention to side-effect profiles of medication combinations.

Much has been learned in recent years about the cytochrome P-450 (CYP450) system as well as medications that may induce or inhibit specific isoenzymes. We have repeatedly observed patients taking a number of medications receive an additional medication with potent CYP450 induction or inhibition properties, which potentially leads to adverse effects or decreased effectiveness of the original medications. Conversely, we have observed detrimental effects following the removal of a medication that was inducing or inhibiting concomitant medications. As simple examples, although the potent CYP450 inhibition of paroxetine and fluoxetine are well known, less attention is typically paid to the metabolism of medications used with them.

Similarly, inattention to the use of multiple medications with similar side-effect profiles may lead to greater problems than the use of either alone. For instance, many psychiatric medications may lead to weight gain. Nevertheless, several such medications are often freely combined without considering alternatives that are more weight-neutral.


As more therapeutic agents are introduced with varying pharmacological profiles, it becomes increasingly important for practitioners to be familiar with existing research, the mechanisms of action of medications, and metabolic pathways in order to use rational polypharmacy. Various interventions have been described that lessen irrational polypharmacy, including peer review processes, continuing education on rational polypharmacy approaches, and clinical pharmacy involvement in patient care.

If future trends in psychiatric disease management follow along their current route, polypharmacy will continue to be necessary in order to obtain optimal therapeutic goals. With this in mind, it becomes imperative for providers to remain judicious when using polypharmacy strategies in the treatment of psychiatric disorders.

Dr Kingsbury is chief of outpatient mental health at the VA Southern Nevada Healthcare System in Las Vegas and clinical professor of psychiatry in the department of Psychiatry and Behavioral Science at the University of Nevada. Dr Lotito is a clinical pharmacy specialist in psychiatry and director of the PGY1 pharmacy residency program at the VA Southern Nevada Healthcare System in Las Vegas. She is also assistant professor of pharmacy practice at the University of Southern Nevada College of Pharmacy. Dr Kingsbury reports that he is on the speakers' bureau for AstraZeneca, Bristol-Myers Squibb, Forest, Janssen, and Pfizer. Dr Lotito reports that she has no conflicts of interest concerning the subject matter of this article.

 Preskorn SH, Silkey B, Shah R, et al. Complexity of medication use in the Veterans Affairs healthcare system, part 1: outpatient use in relation to age and number of prescribers. 

Joukamaa M, Heliovaara M, Knekt P, et al. Schizophrenia, neuroleptic medication and mortality. 

Parks ED, Josef N. A retrospective study of determinants of length of stay in a geropsychiatric state hospital. 

Bates DE, Herman RJ. Carbamazepine toxicity induced by lopinavir/ritonavir and nelfinavir. 

Karp JF, Buysse DJ, Houck PR, et al. Relationship of variability in residual symptoms with recurrence of major depressive disorder during maintenance treatment. 

Pintor L, Torres X, Navarro V, et al. Is the type of remission after a major depressive episode an important risk factor to relapse in a 4-year follow up? 

Frye MA, Ketter TA, Leverich GS, et al. The increasing use of polypharmacy for refractory mood disorders: 22 years of study. 

Silkey B, Preskorn SH, Golbeck A, et al. Complexity of medication use in the Veterans Affairs healthcare system, part 2: antidepressant use among younger and older outpatients. 

West JC, Wilk JE, Olfson M, et al. Patterns and quality of treatment for patients with schizophrenia in routine psychiatric practice. 

Yatham LN, Paulsson B, Mullen J, Vagero AM. Quetiapine versus placebo in combination with lithium or divalproex for the treatment of bipolar mania. 

Sachs G, Grossman F, Okamoto A, et al. Risperidone plus mood stabilizer versus placebo plus mood stabilizer for acute mania of bipolar disorder: a double-blind comparison of efficacy and safety. 

Tohen M, Chengappa KN, Suppes T, et al, Efficacy of olanzapine in combination with valproate or lithium in the treatment of mania in patients partially nonresponsive to valproate or lithium monotherapy. 

Post RM, Speer AM, Leverich GS. Complex combination therapy: the evolution toward rational polypharmacy in lithium-resistant bipolar illness. In: Akiskal HS, Tohen M, eds. 

Bipolar Psychopharmacotherapy Caring for the Patient.

 Chichester, UK: John Wiley & Sons Ltd; 2006:135-167.


Honer WG, Thornton AE, Chen EY, et al. Clozapine alone versus clozapine and risperidone with refractory schizophrenia. 

Josiassen RC, Joseph A, Kohegyi E, et al. Clozapine augmented with risperidone in the treatment of schizophrenia: a randomized, double-blind, placebo-controlled trial. 

Anil Yagcioglu AE, Kivircik Akdede BB, Turgut TI, et al. A double-blind controlled study of adjunctive treatment with risperidone in schizophrenic patients partially responsive to clozapine: efficacy and safety. 

Kingsbury SJ, Yi D, Simpson GM. Rational and irrational polypharmacy. 

Patrick V, Schleifer SJ, Nurenberg JR, Gill KJ. Best practices: an initiative to curtail the use of antipsychotic polypharmacy in a state psychiatric hospital. 

Wu JY, Leung WY, Chang S, et al. Effectiveness of telephone counseling by a pharmacist in reducing mortality in patients receiving polypharmacy: randomized controlled trial. 

Articles

Psychiatric Polypharmacy: The Good, the Bad, and the Ugly

Before deciding that a patient has treatment-resistant schizophrenia, one must first determine that the patient is compliant with medication. Covert noncompliance or partial noncompliance can be missed. A trial with a depot antipsychotic can be helpful in ruling out noncompliance. Similarly, there can be large variations in plasma levels of antipsychotics, even with the same dose. Therapeutic levels of most typical and atypical antipsychotics have not been determined, but blood levels of medications may help determine whether a patient is a rapid metabolizer (or noncompliant).

Effective treatment also presupposes an accurate diagnosis. Conditions such as temporal-lobe epilepsy may mimic symptoms of schizophrenia, while comorbid conditions such as endocrinopathies can impact response. Before determining that a patient has treatment-resistant schizophrenia, it is always useful to carefully review the diagnosis.

Beyond these considerations, discussion of treatment-resistant schizophrenia must start with definitions of what constitutes an adequate trial of an antipsychotic medication, how many trials are necessary and whether certain medications or classes of medications are needed before a patient is determined to be resistant to treatment. Defining what constitutes an inadequate response is also necessary. Unfortunately, there is no unanimity concerning any of these issues.

Research protocols typically define an inadequate response as scores above a certain level on commonly used validated inventories such as the Brief Psychiatric Rating Scale (BPRS) or the Positive and Negative Syndrome Scale (PANSS) and/or an inadequate decrease in score after a defined course of treatment. Although more formal, these research definitions overlap the informal, clinical criteria of persistent target problems such as positive, negative or disorganized symptoms; persistent poor functioning; or frequent relapses despite adequate doses of medication (determined, for example, by using depot antipsychotics). For purposes of this discussion, treatment resistance is defined as an inadequate response from one adequate trial of a typical antipsychotic medication and two of the atypical antipsychotics, excluding clozapine (Clozaril). Since an adequate trial of a typical antipsychotic is defined differently than for the atypicals, each will be discussed separately.

Accumulated evidence suggests that response to each of the typical or first-generation antipsychotics will be similar if they are given at equivalent doses. For example, Kane et al. (1998) showed that fewer than 5% of patients with a documented history of treatment resistance showed any response when treated with haloperidol (Haldol). Further, other research has consistently shown that high-dose strategies lead to greater side effects without measurably improving the response (Thompson, 1994).

An adequate trial of a typical antipsychotic is often defined as six weeks at a dosage level of 1000 mg/day in chlorpromazine (Thorazine) equivalents. Based on the previously cited research, we recommend only one adequate trial of a typical antipsychotic, and we do not recommend heroic doses. As an example, little or no response to 20 mg/day of haloperidol after six to eight weeks should be considered a failure. An inadequate response should lead to trials of the atypical antipsychotics.

The second-generation or atypical antipsychotics currently available include risperidone (Risperdal), olanzapine (Zyprexa), quetiapine (Seroquel) and ziprasidone (Geodon). Clozapine will be discussed separately. These agents all appear to have superior efficacy in several symptom and side-effect dimensions than do the typical antipsychotics. Although these four atypicals differ on important side effects such as extrapyramidal symptoms (EPS), weight gain, QT

 widening, and glucose and lipid control, there is no convincing evidence that any of these medications have superiority to the others. Also, unlike the typical agents, it is unclear whether these atypicals are basically interchangeable in their therapeutic response, so failure of one does not preclude a trial with another one. At this point, more research exists for the increased efficacy of clozapine, risperidone and olanzapine, although evidence of the increased efficacy of ziprasidone is beginning to appear.

An adequate trial of an atypical antipsychotic should be at least eight to 12 weeks on a higher dose of the medication -- for example 12 mg/day of risperidone, 40 mg/day of olanzapine, 800 mg/day of quetiapine or 200 mg/day of ziprasidone. We recommend two full trials with atypicals before abandoning this group of medications as the single therapeutic agent. However, we do not have a hierarchical ordering of preferences, believing that side effects should influence selections. We would stress two trials since there seems little to be gained in trying a third before moving on to clozapine. Some would suggest that a trial with quetiapine should not be counted as one of the two trials due to the lack of supportive clinical data for quetiapine.

Failure from these trials should lead to a trial with clozapine, which has generally been found superior to other available agents. Because of its many side effects -- including agranulocytosis and lowered seizure threshold -- and the requirement for weekly blood monitoring (which may enhance compliance) for at least the first six months (with biweekly monitoring thereafter), clozapine is only approved for treatment-resistant schizophrenia despite its greater efficacy.

The superiority of clozapine may take time to emerge. Therefore, a minimum trial of three or four months should be undertaken, although some would recommend longer trials of six months or a year before concluding that a person is unresponsive to clozapine. Typical clozapine doses range from 300 mg/day to 500 mg/day, and doses of up to 900 mg/day can be used. Some data suggest that minimum plasma levels between 350 ng/ml and 500 ng/ml are associated with increased therapeutic response (Kronig et al., 1995). A lack of noticeable response at two months may be better assessed with a clozapine blood level.

Strategies to use in the face of failure to respond adequately to clozapine have received little empirical evaluation, although published case studies and expert opinions have suggested some therapeutic possibilities. Treatment options after clozapine trials typically involve an augmentation strategy where a second medication or therapeutic strategy is used concurrently with another antipsychotic medication. Multiple different combinations are possible; most have little empirical support; and none have been shown superior to clozapine as of yet. Therefore, as stated, a clozapine trial is recommended before using augmentation strategies, except when patient objections or medical contraindications rule out the use of clozapine. 

Because there is little research supporting augmentation strategies, such trials should be viewed as mini-experiments in which specific target symptoms are tracked. If there is little or no improvement after a reasonable trial, the augmenting agent should be withdrawn and another strategy pursued. For example, if divalproex sodium (Depakote) is added at therapeutic plasma levels for six to eight weeks and minimal improvement is noted, it should be discontinued before a different strategy is tried. Otherwise, there is the danger of having patients on ever-increasing lists of medications with few benefits and the possibilities of increased side effects and interactions. We have seen many cases of patients with schizophrenia taking three different antipsychotics, several mood stabilizers, benzodiazepines and an antidepressant without any clear documentation of improvement -- a strategy we have termed irrational polypharmacy (Kingsbury et al., 2001).

A number of augmentation strategies have been proposed, which include the addition of lithium, an anticonvulsant, a second antipsychotic, a benzodiazepine or a course of electroconvulsive therapy (ECT). Psychosocial treatment may also be used as an augmentation strategy. No data support a specific ordering of these choices, but each deserves specific comment.

Lithium has not been shown to be an effective augmenting agent for schizophrenia in the few trials conducted, although some case studies have suggested its effectiveness. With clozapine, lithium may raise low leukocyte counts but mask other myeloid processes associated with agranulocytosis. This strategy appears to be used less frequently at present, and we would rate this as a less useful strategy.

Although there are a number of anticonvulsants, there are few published reports of the use of many of the newer ones for augmenting response in schizophrenia. Divalproex is commonly used, although there are inconsistent reports of its effectiveness. Casey et al. (2001) published a preliminary report of a large multicenter study that used this strategy and found it effective early in treatment. It should be noted, however, that the improvement had dissipated by week 4 of the study. Divalproex may affect the levels of other medications and is frequently used as protection for clozapine-induced seizures. Because of its independent risk of agranulocytosis, carbamazepine (Tegretol) use with clozapine is contraindicated. At present, a trial testing the addition of divalproex would seem preferred to the other anticonvulsants until more data accumulate.

Using more than one antipsychotic simultaneously is a common practice, although evidence supporting this practice comes only from case studies and case series. There appears to be no justification for using more than one typical antipsychotic at the same time. Since some of the atypical antipsychotics such as clozapine and quetiapine have relatively weak dopamine (D2) antagonism, there may be some justification for adding a more potent D2 antagonist such as haloperidol, risperidone or olanzapine to augment an inadequate response. Adding even a relatively small dose of a typical antipsychotic, such as 2 mg to 4 mg of haloperidol, to an atypical antipsychotic may augment the therapeutic response at the cost of losing some of the advantages of the atypicals such as low EPS or lack of prolactin elevation.

Although g-aminobutyric acid circuits have been discussed as affecting D2 circuits and some research has been conducted, benzodiazepines have generally not been found to be of value, beyond their nonspecific calming and sedating properties, in treating schizophrenia when used alone or as augmenting agents. Further, there are concerns about their addiction potential and their possible pharmacodynamic interactions with clozapine. Only short-term use of the benzodiazepines in schizophrenia is recommended.

Giving ECT concomitantly with antipsychotic medications has been found to be an effective augmentation strategy (Krueger and Sackeim, 1995). Results appear better when the antipsychotic medication is not discontinued during the course of ECT. This strategy has some research support but appears to be underutilized. We recommend a trial of ECT be considered as a possibly preferred augmentation strategy.

Unlike the other augmentation strategies discussed here, psychosocial treatments are useful and recommended whether or not the patient is resistant to medication treatment. Research clearly supports such interventions as ameliorative for a number of problems, including some aspects of negative symptoms. Like the medications, effective approaches continue to be developed.

Finally, it must be realized that some patients with schizophrenia may remain poorly responsive to treatment despite optimal trials of medications and artful use of augmentation strategies. At such times, it may be necessary to settle for a suboptimal response. Heroic efforts with long lists of medications at higher than useful doses can cause more harm than good. Help in such cases will await further progress in psychopharmacology and in our understanding of the pathophysiology of schizophrenia.

 Casey DE, Daniel D, Tracy K et al. (2001), Improved antipsychotic effect of divalproex combined with risperidone or olanzapine for schizophrenia. Poster No. 437. Presented at the 26th Biennial Congress of the World Federation for Mental Health. Vancouver, British Columbia; July 24.
 

 Kane JM, Aguglia E, Altamura AC et al. (1998), Guidelines for depot antipsychotic treatment in schizophrenia. European Neuropsychopharmacology Consensus Conference in Siena, England. Eur Neuropsychopharmacol 8(1):55-66.
 

 Kingsbury SJ, Yi D, Simpson GM (2001), Psychopharmacology: rational and irrational polypharmacy. Psychiatr Serv 52(8):1033-1036.
 

 Kronig MH, Munne RA, Szymanski S et al. (1995), Plasma clozapine levels and clinical response for treatment-refractory schizophrenic patients. Am J Psychiatry 152(2):179-182.
 

 Krueger RB, Sackeim HA (1995), Electroconvulsive therapy and schizophrenia. In: Schizophrenia, Hirsch SR, Weinberger DR, eds. Oxford, England: Blackwell Science, pp503-545.
 

 Thompson C (1994), The use of high-dose antipsychotic medication. Br J Psychiatry 164(4):448-458 [see comments].

Several confounding factors need to be taken into account before labeling a patient as treatment-resistant. What options are available for treating a patient with schizophrenia who has been diagnosed as such?

Steven J. Kingsbury, MD, PhD

Articles

Psychiatric Polypharmacy: The Good, the Bad, and the Ugly

Treatment-Resistant Schizophrenia: Making the Determination