As indicated by its title, the purpose of this article is to present a general methodology for constructing psychopharmacologic treatment algorithms for individuals who have both developmental disabilities and behavioral disorders that are difficult to diagnose. In addition, we present specific algorithms for three categories of behavioral disorders that fit this overall description.
The validity of psychiatric diagnosis for individuals with mental retardation tends to decrease as intelligence decreases.1 This is due to both an increase in nonspecific organic factors and the relative inaccessibility of the individual's inner life, as productive speech becomes rudimentary or nonexistent in those with more severe-profound levels of impairment. Some of the behavioral disorders seen in individuals with profound levels of retardation (ie, impulsivity, irritability, and mood lability) may represent neurophysiologic dysregulation that is not covered by established diagnostic categories.
This is not to imply, however, that one cannot make a psychiatric diagnosis in some individuals with severe-profound levels of impairment who present with fairly clear-cut psychiatric syndromes. An example would be a young man who presented with a cyclic pattern of 2 weeks of withdrawn, hypersomnolent, lethargic behavior during which he would become aggressive only if one attempted to engage him in activities. This behavior state would be followed by 3 weeks of agitated behavior characterized by motoric hyperactivity, markedly decreased sleep, and unprovoked aggression. After this period, there would again be an abrupt transition to a few weeks of euthymic behavior with normal sleep patterns, accompanied by a general functioning consistent with his developmental level. This pattern, coupled with a response to mood-stabilizing agents that decreased the amplitude of the cycles and increased the duration of the euthymic phase, seemed to confirm the diagnosis of a bipolar disorder.
Unfortunately, the inability to make a reliable diagnosis in some individuals with significant developmental disabilities, combined with the professional desire and regulatory pressure to have a psychiatric diagnosis of record that corresponds to the pharmacologic interventions used, can lead a clinician to make a diagnosis that reaily cannot be substantiated or defended. One pattern can be that practitioners will over-use a particular diagnosis for these individuals. For example, the clinical vignette above could lead to an over-generalization that all or most individuals with behavioral disorders and significant levels of intellectual impairment have underlying affective disorders.
It will usually be possible to establish a reasonably reliable psychiatric diagnosis for those individuals who function in the borderline-mild and mild-moderate levels of mental retardation. The clinician searches for behavior "clues" that may link the symptoms to a known specific syndrome or places the symptoms in the spectrum of a known disorder. Examples of clues would be significant seasonal variation in frequency of target behaviors, disruptions in sleep cycle pointing toward an underlying affective disorder, or an individual's apparent motoric response to unseen stimuli, suggestive of hallucinatory perceptions that might be part of a psychotic disorder. There are several symptom checklist rating scales that have been specifically developed to aid clinicians in making a psychiatric diagnosis in individuals with developmental disabilities.2"8
For treating these individuals with clear diagnoses, we rely on existing practice guidelines and treatment algorithms for these disorders, some of which are described in this issue of Psychiatric Annals. In those cases where it is difficult to make a valid psychiatric diagnosis, the diagnostic ambiguity mitigates toward a symptom-based diagnosis and corresponding series of treatment algorithms.
Our experience over the years has been that at least 5% to 10% of negative behavioral changes, especially when there was a dramatic change from baseline, can ultimately be linked to a medical illness, a painful condition, or both. As with the other algorithms in this issue of Psychiatric Annals, the user is advised to review or reestablish the diagnosis after each sequential medication trial if the response remains unsatisfactory.
In a related manner, one should also thoroughly reassess for environmental or behavior factors that could be contributing to the maladaptive behavior. Too often, providers of services feel that the behavior of the individual with developmental disabilities should be changed to fit their program, rather than adapting their program to meet the needs of the individual. If the treating physician loses perspective on these issues, he or she may unwittingly be led into over-medicating the patient, inappropriately medicating the patient, or both. Clues to an environmental contribution to the behavior include a precipitous change in behavior without a medical cause, occurrence of the maladaptive behavior in only one environment (ie, day program and not residence or vice versa), or occurrence at a specific time or in a narrow range of times. The activity and confusion that can occur around shift changes is a common example, as is the individual who becomes aggressive when awakened from a sound sleep at an early hour to go to a day program that he or she does not like.
QUANTIFICATION OF BEHAVIORAL PROFILE
All psychopharmacologic interventions should take into account the risk of the medication as weighed against the potential benefit. One often sees situations where the medication prescribed fits well with the diagnosis, but does not seem to take into account the severity of the condition. Algorithms that are more symptom based need to be even more rigorously constructed from a risk-benefit standpoint.
Fundamental to this process of risk-benefit analysis is obtaining solid data on the following behavioral parameters: frequency, severity, duration, interval data, distribution (both inter day and intra day), the behavioral topography, and antecedent analysis. (Mikkelsen has provided a more complete discussion of data collection issues.1) The function of the data collection is not only to provide information relevant to the construction of the psychopharmacologic algorithms, but also to ensure that medical or environmental contributions to the behavior have not been overlooked before proceeding with a pharmacologic intervention.
Frequency is usually one of the most readily accessible pieces of data. To make recording of frequency more uniform, we have developed a 10-point scale that collapses the range of frequencies to a more manageable continuum (Table 1).
Severity is also a key variable that affects decision making. As a general principle, we feel that, in most cases, considerations of frequency are subordinate to severity. Ln other words, one is willing to employ more powerful, risky interventions with a low-frequency, high-severity behavior than with a high-frequency, low-severity behavior. Obviously, one incident of aggression leading to death is more serious than 100 incidents of verbal aggression. For purposes of rating severity, we again use a 10-point scale. This scale has been published and discussed previously,1 and it is reproduced here.
Frequency Continuum Scale (a rating scale of 1 to 10)
Consolidating the frequency and severity to two 10-point scales allows one to construct a graph with severity-intensity on the horizontal axis and frequency on the vertical axis. This can then be subdivided into four quadrants as illustrated in Figure 1. These quadrants are labeled Profiles 1, 2, 3, and 4 to reflect ascending levels of risk, which take into account the subordination of frequency to severity. The Profiles may be characterized as follows:
Profile 1: Low frequency, low severity
Profile 2: High frequency, low severity
Profile 3: Low frequency, high severity
Profile 4: High frequency, high severity
Scattergrams that plot the frequency-severity for the specific behaviors of an individual are most useful for clinical team planning meetings. Abbreviations can be used to denote the behavior represented by the data point: "A" for aggression; "HB" for head-banging; "VA" for verbal aggression; "PD" for property destruction; and "HB" for hand-biting.
Figure 1 . Profiles of self -injurious behavior, aggressive behav ior, or both.
These Profiles help define breakpoints in the algorithms that are presented in this article. For example, one might not proceed to a medication with a high side effect profile for an individual with a Profile 1 or 2 disorder (Fig. 1).
The processes so far are shown in a flowchart in Figure 2.
FACTORS IN THE CONSTRUCTION OP SPECIFIC ALGORITHMS
A reasoned approach to the use of psychiatric medication in difficult-to-diagnose populations takes into account the probability of success of the pharmacologic intervention, the severity profile of the target symptoms, and the side effect profile of the medication.1 These factors can be represented by the equation below.
The Probability of Success can be calculated from a meta-analysis of all prospective case series and controlled studies, if any are available. Unfortunately, there is a paucity of relevant data involving patients with developmental disabilities. This is discussed below. Symptom Severity is assessed using the 10-point scale from Table 2.
The rating of Side Effect Profile can be somewhat more problematic. We have adopted a fourlevel system,1 numbered One through Four. (No medication has zero side effects). Level One would include agents with very low levels of reported side effects, such as buspirone hydrochloride. A Level Two designation is for medications that would have a somewhat higher level of side effects, but are relatively safe in terms of serious toxicity or lethality. The selective serotonin reuptake inhibitors (SSRIs) would fall into this classification. Level Three medications would be those that can have serious side effects or even be lethal if appropriate monitoring is not performed. Examples would be lithium, tricyclic antidepressants, valproic acid, and carbamazepine. The highest designation (Level Four) would be used for agents that have potentially irreversible side effects, such as tardive dyskinesia. Experimental agents would be in this category. Clozapine is in this group because of seizure risk, the need for weekly blood drawing to monitor for agranulocytosis, and other serious side effects.
Naturally, this equation cannot be calculated with the degree of mathematical precision that one would like. Clinicians may differ in their estimations of the side effect profiles of the various medications. Likewise, ratings of the severity of the target behaviors are somewhat subjective, even if the scale in Table 2 is used.
Another difficulty with regard to the rating of Symptom Severity has to do with the interpretation and weighing of "quality of life" issues. An individual with chronic psychosis who is continually delusional and hallucmating and yet does not harm others or himself or herself, may be rated as mildly symptomatic. Another example would be individuals who are withdrawn and depressed, but appear to have no capacity to hurt themselves or others. These are obviously clinical states associated with impaired quality of life, but the behavior will not lead to a high score on the 10-point severity continuum. For the most part, we find that these are usually individuals with mild levels of mental retardation for whom one can make a valid psychiatric diagnosis, and thus one would proceed with diagnosis-based algorithms.
Behavioral Severity Rating Continuum
The other significant problem with the equation is the difficulty in deterrrdning the Probability of Success factor. Clinicians should keep track of their own success or failure rate with different medications. Multiple case reports in the literature can be of use if the authors report treatment failures as well as successes so that one can calculate a differential response rate for the series. More commonly, one finds single case reports or reports that describe only successes. These are useful for addressing the question: "Has anyone with similar symptoms ever responded to this medication?" But, they do not help with weighing different pharmacologic options.
The literature review process is greatly facilitated by two recent publications. The first is Psychotropic Medications and Developmental Disabilities: The International Consensus Handbook, edited by Steven Reiss and Michael Aman.9 This comprehensive text, published in 1998, provides a thorough review of the use of psychopharmacologic agents in this population, as well as a review of specialized diagnostic instruments. The second publication is Habilitative Neuropsychiatry: Psychopharmacology (1985-1996): A Reference Guide, by Jeffrey Fahs.10 This is an extremely valuable reference that provides an annotated bibliography concerning the use of psychopharmacologic agents in those with developmental disabilities.
These references should be on the shelf of anyone who is actively engaged in pharmacologic work with this population. In addition, one may obtain updated computerized literature reviews of the specific medications being considered. These sources are important because none of these drugs have received Food and Drug Administration (FDA) approval for the indications about to be discussed, and in obtaining informed consent one should be prepared to make reference to what has been reported.
In summary, at this point in time, our algorithm sequence equation cannot be "solved." It serves more as a guide to clinical thinking and a tool for algorithm construction and improvement than as an established methodology. With the progression of time, and the acquisition of new data, decision making will become more evidence based.
Finally, we want to comment on the use of two or more medications at the same time. Our observations led us to conclude that we are more conservative in this regard than some other clinicians who treat patients with developmental disabilities. We will usually not proceed to augmentation with a second psychotropic medication unless there are clear-cut data that there has been some improvement with the first medication. It is easy to fall into the habit of polypharmacy with this difficult-to-diagnose population, with results that can be counterproductive. To assist clinical teams in determining whether there has been improvement with a specific pharmacologic intervention, we have developed criteria for estimation of efficacy.1 There are four categoric levels of response: Category I - 30% to 50% reduction in symptom frequency; Category II - 50% to 70% reduction in symptom frequency; Category ILI - 70% to 90% reduction in frequency; and Category IV - 90% to 100% reduction in target symptom frequency.1 A 0% to 30% reduction in symptom frequency is considered to be within the range of random variation and would not be considered a response without strong mitigating factors, such as a dramatic decrease in intensity.1
In this section, we propose three specific algorithms for some of the more difficult-to-diagnose individuals with more severe levels of intellectual impairment, decreased or absent verbal abilities, and the suggestion of nonspecific organic factors.
Algorithm 1 : Intermittent Explosive Disorder (IED)
These individuals often present with aggressive behavior, self -injurious behavior, or both. The range of symptoms seen will often fit, from a descriptive standpoint, with the DSM-IV diagnostic category of mtermittent explosive disorder." The DSM-IV states, "The essential feature of intermittent explosive disorder is the occurrence of discrete episodes of failure to resist aggressive impulses that result in serious assaultive acts or disruption of property."" Another criterion is that the intensity "is grossly out of proportion to any provocation or precipitating psychosocial stressor." The criteria also indicate that other diagnoses such as psychotic disorder, manic episode, or antisocial personality should be ruled out, as well as dementia, head trauma, and substance abuse. The DSM-IV alludes to neurophysiologic factors by noting that "there may be non-specific EEG findings" and /or "evidence of abnormalities on neuropsychological testing," as well as "nonspecific or 'soft' findings on neurologic examinations."
Mental retardation is not mentioned as a specific rule-out in IED, although the DSM notes that aggression related to dementia or a head injury should more appropriately be considered as a "personality change due to a general medical condition, aggressive type." Thus, IED appears to be an appropriate descriptive diagnosis for many of those who present with these behaviors and who have mental retardation.
The DSM clearly refers to those who are aggressive to others or to property. In the population of individuals with mental retardation there is a group who share many of the "intermittent explosive" characteristics, but the aggression is expressed toward themselves. Clinically, it is important to differentiate between those who are spontaneously aggressive toward others as well as themselves and those who are almost exclusively self-injurious and become aggressive only when staff intervene to interrupt the self-injurious behavior. For the purposes of this algorithm, however, we are including all these forms of aggression under the heading of IED, even if the aggression is directed primarily toward themselves.
Our algorithm for IED is displayed in Figure 3. There is considerable room for debate concerning the relative rankings due to the lack of controlled trials. Note particularly the role of the severity Profile rankings, as described below.
Buspirone hydrochloride12 has a low profile of side effects arguing for its first-line consideration in individuals with a low-severity /frequency Profile. Results can usually be determined in 6 to 8 weeks, depending on the baseline frequency of the behavior. (The more intermittent the behavior, the longer it takes to determine efficacy.) The rate of improvement will be less than 50%, and when it is effective, a common clinical description is that "it took the edge off." Presumably, these are individuals for whom anxiety was contributing to the episodes. However, buspirone hydrochloride would not be our first choice for individuals with Profile 3 or 4 behaviors, given the severity of the behavior and the relatively low probability of success.
Our second-line recommendation would be a trial of an SSRl," the rationales being the relatively low level of side effects, the link between aggression and serotonin,14 and the observation that some of these individuals may have a mood disorder contribution to their behavioral disorder. The probability of response is moderate. We have the most experience with fluoxetine hydrochloride and fluvoxamine maléate, and most of the published literature concerns fluoxetine hydrochloride.13 Fluoxetine hydrochloride seems to be more activating, whereas fluvoxamine maléate has more calming properties and for this reason is often our preferred SSRI.
There is a reasonable amount of clinical literature concerning the efficacy of beta-blockers for individuals with developmental disabilities who present with impulsive aggression, self-injurious behavior, or both.15"17 We consider this the thirdline recommendation due to the cardiovascular considerations and the need for monitoring of blood pressure and pulse. This may be particularly inconvenient if the patient lives in a nonmedical residential setting. Our experience is that patients vulnerable to these blood pressure and pulse changes usually declare themselves at relatively low doses, and the medication can be stopped before a significant adverse event occurs.
The next recommendation is an alpha-blocking agent. The use of Clonidine and, more recently, the longer-acting guanfacine hydrochloride is primarily borrowed from child psychiatry where they are often used as adjunctive treatment for children and adolescents who have extreme impulsi vity in addition to motoric hyperactivity and inattention.18 The literature regarding the use of these agents in those with developmental disabilities is more limited compared with the support for beta-blockers.19,20 We ranked them ahead of the mood-stabilizing anticonvulsants due to their lower side effect profiles, and the fact that one does not have to monitor blood levels. Of interest are two case reports by Sovner describing the use of Clonidine to facilitate withdrawal from long-term use of thioridazine hydrochloride.21
The next choice is an anticonvulsant. As indicated above, there are several aspects of this behavior disorder that are suggestive of neurophysiologic dysfunction, even in the absence of a definite correlation with a documented seizure disorder. These drugs also have mood-stabilizing properties. One could argue that these agents should be introduced earlier in the algorithm, and this would be reasonable for individuals with more impressive evidence of neurophysioIogic or mood dysregulation. However, the need for periodic blood tests can be a major drawback for some individuals with developmental disabilities due to fear of venipuncture. Particular side effects that we have occasionally encountered include hyponatremia with carbamazepine, thrombocytopenia with valproic acid, and disabilhibition with gabapentin. Lamotrigine is also a member of this category, but we have not included it because we have no experience with it and limited anecdotal experience from colleagues has been unfavorable so far. As these anticonvulsant agents also have mood-stabilizing properties, a natural question would arise with regard to lithium. Our experience with hthium in this group has been far less favorable than that with the anticonvulsant agents and thus we have not included it in this particular algorithm.
It is difficult to rank order the agents within this category. Earlier positive reports were with carbamazepine.22 Recently, there has been more interest in valproic acid23"25 and it has achieved FDA approval for bipolar disorder. Nevertheless, we still prefer carbamazepine at this time due to an impression of greater efficacy. We continue to gain experience with gabapentin, which appears to have a lower level of side effects, but its efficacy over the long term is unclear. Its advantage is that blood levels are unnecessary.
As indicated in Figure 3, if an individual with Profile 1 or 2 behavior severity has still not shown any response at this point, one has to seriously consider whether to proceed further. A breakpoint where the "cure is worse than the disease" may have been reached. Also, as mentioned earlier, with each progression down the algorithm one should reevaluate diagnosis.
However, if the individual has Profile 3 or 4 severity of behavior and the response is still unsatisfactory, one questions whether there is a psychotic component to the disorder. The introduction of the newer generation of antipsychotics, which are reported to have lower frequencies of movement disorder and other side effects than the traditional agents, has made it easier to justify an empiric trial of an antipsychotic. These agents are risperidone, olanzapine, quetiapine, and clozapine. The latter is in a class by itself in terms of efficacy and side effects and is discussed later.
The newer agent associated with the most published reports concerning its use in individuals with developmental disabilities is risperidone.26"28 In a recent brief review of the use of atypical antipsychotic agents in individuals with developmental disabilities, Connor and Posever noted that they could find only case reports regarding the use of olanzapine in children with developmental disabilities and could not find published reports concerning the use of quetiapine in individuals with mental retardation.29 At this point, we would make a selection based primarily on the patient's likely susceptibility to the different side effects of the newer agents.
If two of the atypicals have been unsuccessful, we would strongly suggest consideration of clozapine.30"32 If the behavior disorder is extremely severe, one may consider going directly to clozapine, or doing so after a trial of only one other antipsychotic. We have analyzed our experience with clozapine and can say that its probability of being effective for individuals with severe behavioral disturbances who have not responded to other pharmacologic agents, including multiple antipsychotic agents, exceeds 90%, for those who can tolerate its side effects. Although clozapine does have FDA approval for the treatment of schizophrenia, it does not have this approval for those who do not meet the criteria for schizophrenia. The FDA also has stipulated that, for schizophrenia, it may be used only after adequate trials of at least two other antipsychotics.
Algorithm Z: Atypical Psychotic Disorders
Among the severely cogmtively impaired, diificult-to-diagnose individuals with significant behavioral problems, one will encounter many individuals who are already receiving antipsychotic agents. Some of these individuals will have fairly classic symptoms of schizophrenia, although the symptoms may not be easily detected due to the cognitive limitations, lack of developed speech, or both in these individuals. For example, one may not be able to elicit clear evidence of paranoid delusional thoughts, but may note irrational tearfulness, irritability, and aggression in situations that would not be expected to provoke such a response. Similarly, with hallucinations one may not be able to elicit verbal descriptions, but may infer their presence from physical actions (eg, turning of the head in response to stimuli that are not visible, talking to persons who are not visible, and /or striking out at objects or individuals who are not seen by others). Naturally, a diagnosis of psychosis based on inferences such as these may be incorrect (eg, an anxiety disorder could produce irrational fearfulness). Also, some individuals with mental retardation may have been given antipsychotic agents in the past for inappropriate reasons such as general sedation, leading to significant over-use of antipsychotic agents in this population.
We start with an algorithm for the individual who is already receiving a neuroleptic, given the large numbers of such individuals that one encounters. Within this large group there are two subgroups: those receiving neuroleptics and doing well with no side effects, and those receiving antipsychotic agents and not doing well behaviorally and /or experiencing significant side effects. These algorithms are presented in Figures 4 and 5.
Patient's Condition Is Stable With a Neuroleptic Without Significant Manifest Side Effects? (Fig. 4) If the patient presents with no clear history that the antipsychotic agent is necessary, we recommend a trial of slow tapering of this medication. A slow tapering will hopefully allow one to pinpoint the minimum effective dosage of neuroleptic by noting the dosage at which behavioral regression begins to develop. Usually, regression will occur at some point. However, we have been able to get many individuals completely off neuroleptics by this approach. A slow taper also helps distinguish temporary withdrawal-emergent effects from true threshold effects. Withdrawal-emergent problems may involve a transient increase in aggression, self-injurious behavior, or both, which soon improves. Going below a mimmxim effective (threshold) dose more often results in a flattening of affect, a decrease in hygiene, a loss of interest in activities, withdrawal, confusion, incontinence, and disturbed sleep. This behavioral regression can be severe and may include smearing of feces and coprophagia. A syndrome of significant weight loss and cachexia can also be seen, which can lead to extensive medical workups. This syndrome has been referred to as neuroleptic-withdrawal cachexia.33
If it is established that the individual is doing well with a neuroleptic, there are no side effects, and the dose is relatively low34 (as it usually is), then at this time we would not recommend routinely trying to switch the individual to one of the newer-generation antipsychotics. During the years since these medications have been available, we have attempted to convert a large number of individuals whose conditions were stable with low dosages of traditional antipsychotic agents to one of these newer agents to lower the long-term risk of tardive dyskinesia. Our success rate has been low, and many of these patients have been destabilized for up to 1 to 2 years, until they are returned to the original antipsychotic. Thus, we no longer attempt to convert these individuals unless there is a known compelling clinical reason to do so. The one exception to this is individuals whose conditions have been maintained at relatively high doses of traditional antipsychotic agents. We will often try adding risperidone or olanzapine, which may allow us to significantly lower the dose of the original agent, producing a brightening of affect, increased verbal expressivity, and, presumably, a decrease in the long-term risk of tardive dyskinesia. We do not yet have enough experience with quetiapine to make a comment on its use in this regard: we are waiting for evidence to appear of its efficacy in individuals whose conditions have been refractory to neuroleptics.
Patient's Condition Is Refractory to Neuroleptics, Patient Is Experiencing Significant Side Effects, or Both? (Fig. 5) One encounters many individuals who have taken high dosages of multiple antipsychotic agents in the past without clinical response or with only a partial response. We have recently reviewed our experience with individuals with level 3 and 4 Profiles of aggression, self-injury, or both and who have had years of unsatisfactory response to high doses of neuroleptics.
As has been our experience with the stabilized patients, only a few of these refractory patients seem to do better with risperidone or olanzapine, although slightly more show improvement if the original antipsychotic is continued (usually at a lower dose) while the newer medication is added. We have greater confidence in clozapine. However, the need for weekly blood tests is an extreme problem for some individuals and is also often a concern to guardians and family members. Therefore, we usually try more than one of the new-generation antipsychotics, anyway. The only instance in which we would strongly suggest going directly to clozapine would be a severe life-threatening situation or one that presents the risk of serious irreversible physical damage.
Algorithm 3: Chronic Hypomanla (Fig. 6)
A third group of difficult-to-diagnose patients who do not fit readily into existing diagnostic categories are those who present with what appears to be chronic non-cyclic hypomanic behavior. The symptom cluster usually consists of extreme motoric hyperactivity, impulsive aggression, thin body habitus despite increased appetite, and decreased sleep. Sleep ranges from 4 to 6 hours per night, and on many nights these individuals sleep even less. A high frequency of intense obsessive-compulsive (OC) behaviors often accompanies the above profile, but not always. We have not been able to find a published description of this clinical entity, but clinicians who work with this population tell us of further examples.
Figure 6. Algorithm for chronic hypomania. VPA = valproic acid; CBZ = carbamazepine.
Chronically Hypomanic Without OC Symptoms? As summarized in the algorithm in Figure 6, the first-line recommendation is a moodstabilizing agent. We provided the references for valproic acid and carbamazapine earlier in this article.22*25 There is published literature concerning lithium35"38 that is comparable to or perhaps more impressive than that for the other mood-stabilizing agents; however, it does not address this specific chronically hypomanic group.
As this algorithm indicates, we would begin with lithium or valproic acid, as both have FDA approval for bipolar disorder and, presumably, this syndrome is a variant of mania. If the individual also had a seizure disorder, we might lean toward valproic acid in hopes that it might be able to address both conditions.
If there was no response to these agents, we then would move on to carbamazapine, followed by consideration of gabapentin. However, our experience with gabapentin monotherapy in this group has been disappointing. It may be more useful to augment a partial response to one of the earlier agents by adding gabapentin, as indicated in the footnote to Figure 6. A similar pattern of experience has been reported with the use of gabapentin in nondevelopmentally impaired individuals with bipolar disorder.39
For Profile 3 and 4 individuals, one would consider an antipsychotic agent as adjunctive treatment if there had been only a partial response to one or two mood stabilizers, or as monotherapy if the mood stabilizers had no effect. If clozapine is added and produces an excellent outcome, we recommend trying to remove the other medications one by one and to determine whether clozapine could be effective as monotherapy.
Chronically Hypomanic With OC Symptoms? We find that the core hypomanic motoric hyperactivity and impulsive aggression symptoms will be more problematic than the OC symptoms, so we would not pursue these until the hypomanic symptoms have been stabilized. The OC symptoms may be treated with SSRIs. If two trials tan, domipramine hydrochloride4*42 might be considered next, due to its higher level of side effects.
This is one group that often requires multiple medications. They usually require the OC medication to be added to their mood stabilizer, and they may need a low-dose antipsychotic agent as well. In those cases that proceed to clozapine, there is the hope of returning to monotherapy, although clozapine sometimes exacerbates OC symptoms in schizophrenia.
This article describes a systematic process îor evaluating and choosing pharmacotherapy for difBcult-to-diagnose patients with developmental disabilities. If a reliable psychiatric diagnosis can be established, one would proceed with standard algorithms, such as those presented elsewhere in this issue of Psychiatric Annals. We believe that the three specific algorithms presented here represent common clinical situations that will be encountered by those who work with this population.
The paucity of literature concerning the use of psychotropic agents in this population makes the construction of these algorithms most difficult and more subject to opinion than one would like. However, these patients are found in abundance throughout the United States and physicians grapple on a daily basis with the problems associated with their care. There are little systematic outcome data to inform discussion about how best to proceed. We have attempted to establish a reasonable, generic approach that could form the starting point for data collection on patient outcomes. These data can then be used to improve the algorithms. It is hoped that computerization of these algorithms and data collection over the Internet or over intranets, as a part of the Harvard Psychopharmacology Algorithm Project, will facilitate this goal.
1. Mikkelsen EJ. Risk-benefit analysis in the use of psychopharmacologic interventions for diffictút-to-diagnose behavioral disorders in individuals with mental retardation. Psichiatric Annals. 1997;27:207-212.
2. Aman MG, Singh NN. Aberrant Behavior Checklist Manual. East Aurora, NY: Slossón Educational Publications; 1986.
3. Aman MG, Singh NN. Aberrant Behavior Checklist Community Supplementary Manual. East Aurora, NY: Slosson Educational Publications; 1994.
4. Nihira K, Leland H, Lambert N. Adaptive Behavior Scale: Residential and Community. Washington, DC: American Association of Mental Retardation; 1993.
5. Matson JL, Coe A, Gardner WI, Sovner R. A factor analytic study of the diagnostic assessment for the severely handicapped scale. / Nerv Ment Dis. 1991;179:553-557.
6. Einfeld SL, Tonge BJ. The Developmental Behaviour Checklist: the development and validation of an instrument for the assessment of behavioral and emotional disturbance in children and adolescents with mental retardation. J Autism Deo Disord. 1995;25:81-104.
7. Strohmer DC Prout HT. The Emotional Problems Scales. Odessa, FL: Psychological Assessment Resources; 1991.
8. Matson JL. The PlMRA Manual, 2nd ed. Worthington, OH: IDS Publishing; 1997.
9. Reiss S, Aman MG, eds. Psychotropic Medications and Developmental Disabilities: The International Consensus Handbook. Columbus, OH: Ohio State University Nisonger Center; 1998.
10. Fahs JJ. Habilitative Neuropsychiatry: Psychopharmacology (1985-1996): A Reference Guide. Kingston, NY: National Association of Dual Diagnosis; 1996.
11. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th ed. Washington, DC: American Psychiatric Association; 1994.
12. Ratey JL Sovner R, Mikkelsen E, Chmielinski HE. Buspirone therapy for maladaptive behavior and anxiety in developmentally disabled persons. J Clin Psychiatry. 1989;50:382-384.
13. Mikkelsen EJ, Albert LG, Emens M, Rubin E. The efficacy of antidepressant medication for individuals with mental retardation. Psychiatric Annals. 1997;27:198-206.
14. Fuller RW. Fluoxetine effects on serotonin function and aggressive behavior. Ann N Y Acad Sci. 1996;794:90-97.
15. Ratey JJ, Mikkelsen E, Sorgi P, et al. Autism: the treatment of aggressive behaviors. J Gin Psychopharmacol. 1987; 7:35-41.
16. Ratey JJ, Mikkelsen EJ, Smith GB, et al. β-blockers in the severely and profoundly mentally retarded. J Clin Psychopharmacol. 1986;6:103-107.
17. Ruedrich SL. Beta-adrenergic blocking medications for treatment of rage outbursts in mentally retarded persons. Seminars in Clinical Neuropsychiatry. 1996;1:115-121.
18. Hunt RD, Capper L, O'Connell P. Clonidine in child and adolescent psychiatry. J Child Adolesc Psychopharmacol. 1990;1:87-102.
19. Jaselskis CA, Cook EH, Fletcher KE, Leventhal BL. Qonidine treatment of hyperactive and impulsive children with autistic disorder. J Clin Psychopharmacol. 1992;12:322-327.
20. Fankhauser MP, Karumanchi VC, German ML, Yates A, Karumanchi SD. A double-blind placebo-controlled study of the efficacy of transdermal Clonidine in autism. J Clin Psychiatry. 1992;53:77-82.
21. Sovner R. Thioridazine withdrawal-induced behavioral deterioration treated with Clonidine: two case reports. Ment Retard. 1995;33:221-225.
22. Langee HR. A retrospective study of mentally retarded patients with behavioral disorders who were treated with carbamazapine. Am J Ment Retard. 1989;93:640-643.
23. Kastnert T, Finesmith R, Walsh K. Long-term administration of valproic acid in the treatment of affective symptoms in people with mental retardation. J Clin Psychopharmacol. 1993;13:448-451.
24. Mattes JA. Valproic acid for nonaffective aggression in the mentally retarded. J Nerv Ment Dis, 1992;180:601-602.
25. Sovner R. The use of valproate in the treatment of mentally retarded persons with typical and atypical bipolar disorders. J Clin Psychiatry. 1989;50:40-43.
26. McDougle CJ, Brodkin ES, Teung PP, Naylor ST, Cohen DJ, Price LH. Risperidone in adults with autism or pervasive developmental disorder. J Clin Adolesc Psychopharmacol. 1995;5:273-282.
27. Purdon SE, Lit VV, LaBeHe A, Jones BDW. Risperidone in the treatment of pervasive developmental disorder. Can J Psychiatry. 1994;39:400-405.
28. Vanden Borre R, Vermote R, Buttiens M, et al. Risperidone as add-on therapy in behavioural disturbances in mental retardation: a double-blind placebo-controlled cross-over study, Acta Psychiatr Scand. 1993;87:167-171.
29. Connor DF, Posever TA. A brief review of atypical antipsychotics in individuals with developmental disability. Mental Health Aspects of Developmental Disabilities. 1998;1:93-102.
30. Cohen SA, Underwood MT. The use of clozapine in a mentally retarded and aggressive population. J Clin Psychiatry. 1994;55:440-444.
31. Rubin M, Langa A. Clozapine, mental retardation, and severe psychiatric illness: clinical response in the first year. Haw Rev Psychiatry. 1995;3:293-294.
32. Sajatovic M, Ramirez, LF, Kenny JT, Meltzer HY. The use of clozapine in borderline-intellectual-functioning and mentally retarded schizophrenic patients. Compr Psychiatry. 1994;35:29-33.
33. Mikkelsen EJ, Albert LG, Upadhyaya A. Neurolepticwithdrawal cachexia. N Engl J Med. 1988;318:929.
34. Mikkelsen EJ. Low-dose Haloperidol for stereotypic selfinjurious behavior in the mentally retarded. N Engl J Med. 1986;315:398-399.
35. Craft M, Ismail IA, Krishnamurth D, et al. Lithium in the treatment of aggression in mentally handicapped patients: a double-blind triaL Br J Psychiatry, 1987;150: 685-689.
36. Langee HR. Retrospective study of lithium use for institutionalized mentally retarded individuals with behavior disorders. Am J Ment Retard. 1990;94:448-452.
37. Pary RJ. Towards defining adequate lithium trials for individuals with mental retardation and mental illness. Am J Ment Retard. 1991;95:681-691.
38. Spreat S, Behar D, Reneski B, Miazzo P. Lithium carbonate for aggression in mentally retarded persons. Compr Psychiatry. 1989;30:505-511.
39. Ghaemi SN, Katzow JJ, Desai SP, Goodwin FK. Gabapentin treatment of mood disorders: a preliminary study. J Clin Psychiatry. 1998;59:426-429.
40. Barak Y, Ring A, Levy D, Granek I, Szor H, Elizur A. Disabling compulsions in 11 mentally retarded adults: an open trial of clomipramine SR. J Clin Psychiatry. 1995;56:459-461.
41. Garber HJ, McGorugle JJ, Slomka GT, Monteverde E. Clomipramine treatment of stereotypic behaviors and self-injury in patients with developmental disabilities. J Am Acad Child Adolesc Psychiatry. 1992;31:1157-1160.
42. Lewis MH, Bodfish JW, Powell SB, Golden RN. Clomipramine treatment for stereotypic and related repetitive movement disorders associated with mental retardation. Am J Ment Retard. 1995;100:299-312.
Frequency Continuum Scale (a rating scale of 1 to 10)
Behavioral Severity Rating Continuum