LETTERS TO THE EDITOR Outcome of Inpatient Treatment
Taking the Boards
To the Editor:
To the Editor:
Pfeiffer and Strze lecki (1990) prov ide a com prehensive and meticulous review of inpatient treatment outco me studies . Their points are well made , and we, of course, apprecia te their citing our review and were interested to note their findings generally to be in agreeme nt. We would like to respond to a couple of issues. A maj or trust of this present ation was ' ' to develop and app ly a more precise and conserva tive appr oach to integrating findin gs . .. by employing a statistical procedure to pool the results of previo us research " (p. 847). Although the authors have clearly do ne that , we believe they would have done well to have noted for the reader the necessary ca ution we know the authors maintain when studying these findings. If the original data are co llecte d without using reliable or valid meas ures, if the samples vary widely and include all ages of youth, if follow-up time frames vary from 0 to 2 month s to 11 years, if the sources of outcome measurement vary from IQ tests and interviews to quest ionnaires and record reviews, then one remain s careful in entrusti ng too much confidence eve n in a metanalysis of this kind of pooled data . The authors ackn owledge that they were "forced to develop an alternative meth od for synthes izing the results" (p. 848) . Th e statistical approac h employe d appea rs straig htfo rwa rd, clea r, and therefore appealin g. It stands "on the face of it " for what it is, and it is certain ly a reasoned approac h. However , the authors do not cite a reference or provide an ex pla nation of the math em atical validity of their equation for determining a weighted predictive value (WPY) . One illustrat ion might be help ful in clarifyi ng the potent ial for the findings of any such analysis to be too confide ntly accepted. The authors omit studies includ ing youth suffer ing mental retardation and seve re developm ent al de lays. Then , applying the ir WP Y ana lysis, they conclude that IQ demonstrated a mode stly strong positive relati onship to postdischarge adaptation. Clearl y, the interac tions of this dimension are complex . Is there a basel ine IQ below whic h there is a significa nt negative corr elation to later postdi scharge coping? Does IQ affect prognosis more in some disorders than others ? The calculatio n of WPYs is o nly modestly helpful , is probably more interesting than ultim ate ly meanin gful in term s of validity, and has the potenti al of misleading us into a false sense of confidence abou t these findings. Pfeiffer and Strzelecki ( 1990) have done an exce llent jo b of bringing us up-to-date with the current state of know ledge from inpatient treatment assessment follow-up studies . But let 's not forget the greater efforts requi red by child psych iatrists in designing and implementing prospective treatment assess me nt projec ts . More sophisticated statistical analyses of our current data base will not provide the need ed information. We encourage child psychiatrists to devote increasi ng energy, mone y, and manpow er to designing and completing methodologically sound proje cts-a task that has not always been well supported by training institutions, funding sources , or ment ors .
Having rece ntly passed the oral Child and Ado lesce nt Psychiatry Boards after two previous unsuccessful atte mpts in 1985 and 1989 and after almost 20 years of private practice, I now feel so mewhat qualified to comment on how to pass and how not to pass the boards, especi ally the oral part. Being " fami liar" with DSM-lI/-R diagnoses is not sufficie nt. For my third attempt, I mem orized all the child and adolescent diagnostic criteria instead. I also mem orized all significa nt aspects of the Ment al Status Exa minatio n as we ll as all significant subjec ts to be cove red in a diagnostic formul ation . Thi s prov ides ment al " shelv ing" upon whic h the clinical information to be gathered can be stored and catego rized. It helps greatly to practice board-like inter views with other candidates in order to have a feel of what to ex pect, includ ing time pressur es. In additio n, it is also helpful to have previously taken the AACAP review course or a previous board exam . I found the video tape/vig nette section the most difficult beca use it was the hardest to prepare for, and the cases present ed often are difficult pat ient s from chaotic, socia lly dep rived famili es , less commonly seen in private practice . They also make for a more challenging and extensive differential diagnos is . It would have been helpful to have been able to purch ase a serie s of practice vignette s and pract ice video tapes to get a better feel for the format and type of patient s that I would face . Even though I took the AACAP annual revi ew course in 1985 and agai n in 1990, such additio nal pract ice opp ortunities would have made my 1990 exper ience eas ier and less anx iety provoking. Not knowing what to expect is one of the most common causes of debilitating anxiety . Overall , I fee l that taking the oral boards did force me to sharpen my diagnost ic skill s, wh ich, I feel, made me a better clinician . I am please d, however , at not having to take the board s again.
Mark 1. Blotcky, M .D. Th om as L. Dimperio , Ph .D . Timberlawn Psychi atric Hospital Dallas, Texas REFERENCE
Pfeiffer , S . 1. & Strzeleck i, B. A. (1990), Inpatient psychiatr ic treatment of children and adolesce nts: a review of outcome studie s. I.
Am. Acad. Child Adolesc. Psychiatry, 29:847-853. I .Am .Acad. Child Adolesc. Psychiatry , 30 :3, May 199/
Dirk E . Hutt enbach , M .D . Mari etta , Georgia
Fenfluramine and Mental Retardation To the Editor: Recent ly, we undert ook a doub le-blind, placebo -controlled crossover study comparing the cl inical ef fects of meth ylphenidate and fen fluram ine in ment ally retarded childr en with sy mpto ms of attentio n deficit hyper acti vity dis order. Twenty ch ildren have been tested thus far , 19 of whom are not autistic. Th is is unique in tha t the large majorit y of previous fenflura mine studies have employed only autistic children as subje cts (Aman and Kern , 1989), and there is only one other clinical trial known to us that com pares fcnflur amin e to another acti ve medication in children (Donnelly et al., 1989). We would like to take this opportunity to present some preliminary find ings regarding the side effects of fenflura mine . We assessed side effects in our subjects through parent reports using a modified version of the Dosage and Treatm ent Emerge nt Sympt om s Scale (DOTES) developed for psychopharmacological trials by the Nationa l Institut es of Menta l Health (Guy, 1976; Rapoport and Conners, 1985). Parents , observi ng red uctions in motor beh avior , reported on the item, " Increase d motor acti vity," where fenfluramine appeared to be causing significant reductions in comparis on with placebo (p 0= 0 .0 14, Wilcoxon s igned ranks tests). The se reductions in mot or be-
507
LEITERS TO THE EDITOR havior were more co nsistent than those seen with meth ylphen idate , which differs from placeb o only at marginal levels (p = 0.056) at this stage of the study . As in other studi es, fenflu ramine caused drowsiness in some children (although not seve re in any case), and the differences appro ached significance both in comparison with placebo (p = 0.052 ) and methylphenidate (p = 0.052) . We observed a nonsignificant trend for fenfluramine to cause anorexia, but so far, this is less apparent than the differences between methylphen idate and placebo (p < 0.10). We have also a seen a nonsignific ant tenden cy for less akathisia with fenfluram ine than with placebo (p = 0 .068) , which may be as a result of a mild anx iolytic effect or may be secondary to the reduction in activity level. One hereto fore unreported side effect is a very vivid report of a 7-year-old girl who developed a peculi ar body odor and halitosis that coincided exactly with the times of fenfluramine treatment. We observed no increa ses in insomnia , dizziness , or diarrhea with fenflur amine, although these are sometimes reported to occur with the drug (Aman and Kern , 1989). We also looked at heart rate, blood pressure , and body weight in these children. Analysis of varian ce indic ated that all three cardiovascular measures were significantly affected by the medications. Fenfluramine was associ ated with reductions in each , but all significant comparisons were with respect to meth ylphenidate , which tended to eleva te heart rate and blood pressure. As in most studies, fenfluramin e produced significant weight losses in these children (p = 0 .01 ). Our subjects lost about 1.8 pound s (0 .82 kg) relative to their placebo weights during the 4 weeks they received medication. Finally, two groups of symptom s that do not appear on the DOTES Scale occurred fairly often in our sampl e. Seven of our subj ects were reported to have difficulties with enure sis and encopresis, and another four demon strated stereotypy and self-injurious behavior. Enure sis and encopresis were lowest when the children were taking methylphenidate, appe aring to worse n both with placebo (p = 0. 068) and fenfluramin e (p = 0.068) . Th e emer gence of these symptoms (usually in children who had previou sly been maintained on methylphenidate) has probably caused more distress than any other side effect in this study . Furthermore , it is a well-est ablished fact that moderate-to-high doses of psychostimul ants can induce stereotypic behavior in laboratory animal s (Randrup and Munk vad, 1974 ), and stero typies may occur de novo (or existing stereotypies may be exacerbat ed) at lower doses of stimulants in children with developmental disabilities (Aman, 1982). Both enure sis/encopresis and stere otypy /self-injur y are obviously of relevance to researcher s and clinicians working with developmentally delayed clinical populations. Other research ers working with development ally delayed children may wish to add these symptoms to their inventory of side effects if they are using the DOTES to monitor side effec ts. Furthe rmore , if the DOTES is revised in the future , its developers should seriously consider addin g these and other relevant items to document such potenti al side effects. Michael G. Aman , Ph.D. Richard A . Kern , M.D . L. Eugene Arnold, M .D . Debbie E. McGhee , B.A . The Nisonger Cent er for Ment al Retardati on and Developmental Disabil ities Ohio State University , Columbus, Ohio REFER ENCES
Aman, M. G. (1982) , Stimul ant drug effects in developmental disorders and hyper activit y-toward a resolution of disparate findings. J . Autism Dev, Disord. , 12:385-398. Aman, M . G . & Kern , R. A. (1989) , Review of fenfluramine in the treatment of the devel opmental disabilitie s. J. Am . A cad . Child Ado/esc. Psychiat ry , 28 :549-565 . Donnell y , M., Rapoport , J . L. , Potter , W. Z., Oliver , J . , Keysor ,
508
C. S . & Murphy, D. L. (1989), Fenfluramine and dextroamphetamine treatment of childhood hyperactivit y . Arch . Gen . Psychiatry , 46:205-21 2. Guy, W . (1976) , ECDEU Assessm ent Manual f or Psychoph armacology. (Rev . Ed.). Rockville , MD : U.S. Department of Health , Education , and Welfare. Randrup , A. & Munkvad , 1. (1974) , Pharm acology and physiolog y of stereotyped behavior. J , Psychiat r . Res. , 11:1-10 . Rapoport , J. L. & Conner s, C . K. (eds .) (1985 ), Special feature: rating scales and assessment instrum ents for use in pediatric psychopharmacology research . Psychoph armaco/ . Bull ., 21:7 19-1125.
Adverse Effects of Fluoxetine To the Editor: Increased anxiety and stereo typical beha viors in response to changing fluoxetine dosage from 20 mg every other day after 2 weeks to 20 mg every day in a 26-year-old white autistic female (Mehlinger et aI., 1990) may have been due to the serotonergic-mediated inhibition of dopamine lateraliz ed to the right hemisphere of the brain in which the metabolic rate is higher in femal es (Friedman and Janko vic, 1990). This hypothesis is supported by the stereotyped, rituali stic behaviors in autisti c individual s resemblin g compul sive behavi or (Mehlinger et aI., 1990) linked to right frontal brain dysfunction (Friedman and Jankovic, 1990). Adver se respon ses to fluoxetine may be the result of a sudden increa se in serotonin that might be avoided by a more gradual dose titration (Papp et aI., 1990). Ernest H . Friedman , M.D. East Cleveland, Ohio REFERENCES
Friedman, E. H. & Jankovic, J. (1990), Tardi ve oculogyria (letter and reply). Neurology . 40 :728. Mehling er , R. , Scheftner, W . A. & Poznanski , E. (1990) , Fluoxetine and autism (letter ). J . Am . Ac ad . ChildAdolesc. Psychiatry, 29:985. Papp , L. A. , Gorman , J. M. , Teicher , M . H . , Glod , C. & Cole, J . O. (1990), Suicidal preoccup ation durin g fluoxetine treatment (letter and reply). Am . J . Psychiatry. 147:1380--1381 .
Fluoxetine in Autism with Depression To the Editor: We enjo yed reading the report on the use of fluoxetine in autism by Mehlin ger et al. (1990). We agree that fluoxetine may be useful in the treatment of persons with autism . Our expe rience with this drug has been simila r to that of Mehlinger et al. with some exceptions as described below: Case 1 J. F. , a 16-year-old girl with autism and moderate mental reta rdation, was referred with a 3-week histor y of chan ge of behav ior that consisted of social withdrawal, povert y of speech, a general lack of interest in hobbie s, and disturbance of sleep and of appetite. Her other behaviors consisted of lining up objects and compulsively carrying obje cts with her. There was a strong famil y history of affective disorder. After a careful assessment, a diagnosis of major depre ssion in addition to her developmental disorder s was made , and treatment with fluoxet ine, 20 mg/d , was started . About 2 weeks after starting the medic ation , her parents reported that the patient was much better . She looked brighter, and her vege tative features also improved. Howe ver, there was no decrease in her compulsive behaviors . In view of the respon se, fluoxetine was increased to 20 mg twice a day . Two week s later , her parents reported l .Am . A cad. Child Ado/esc. Psychiatry, 30:3, May 1991
Taking the Boards
To the Editor:
To the Editor:
Pfeiffer and Strze lecki (1990) prov ide a com prehensive and meticulous review of inpatient treatment outco me studies . Their points are well made , and we, of course, apprecia te their citing our review and were interested to note their findings generally to be in agreeme nt. We would like to respond to a couple of issues. A maj or trust of this present ation was ' ' to develop and app ly a more precise and conserva tive appr oach to integrating findin gs . .. by employing a statistical procedure to pool the results of previo us research " (p. 847). Although the authors have clearly do ne that , we believe they would have done well to have noted for the reader the necessary ca ution we know the authors maintain when studying these findings. If the original data are co llecte d without using reliable or valid meas ures, if the samples vary widely and include all ages of youth, if follow-up time frames vary from 0 to 2 month s to 11 years, if the sources of outcome measurement vary from IQ tests and interviews to quest ionnaires and record reviews, then one remain s careful in entrusti ng too much confidence eve n in a metanalysis of this kind of pooled data . The authors ackn owledge that they were "forced to develop an alternative meth od for synthes izing the results" (p. 848) . Th e statistical approac h employe d appea rs straig htfo rwa rd, clea r, and therefore appealin g. It stands "on the face of it " for what it is, and it is certain ly a reasoned approac h. However , the authors do not cite a reference or provide an ex pla nation of the math em atical validity of their equation for determining a weighted predictive value (WPY) . One illustrat ion might be help ful in clarifyi ng the potent ial for the findings of any such analysis to be too confide ntly accepted. The authors omit studies includ ing youth suffer ing mental retardation and seve re developm ent al de lays. Then , applying the ir WP Y ana lysis, they conclude that IQ demonstrated a mode stly strong positive relati onship to postdischarge adaptation. Clearl y, the interac tions of this dimension are complex . Is there a basel ine IQ below whic h there is a significa nt negative corr elation to later postdi scharge coping? Does IQ affect prognosis more in some disorders than others ? The calculatio n of WPYs is o nly modestly helpful , is probably more interesting than ultim ate ly meanin gful in term s of validity, and has the potenti al of misleading us into a false sense of confidence abou t these findings. Pfeiffer and Strzelecki ( 1990) have done an exce llent jo b of bringing us up-to-date with the current state of know ledge from inpatient treatment assessment follow-up studies . But let 's not forget the greater efforts requi red by child psych iatrists in designing and implementing prospective treatment assess me nt projec ts . More sophisticated statistical analyses of our current data base will not provide the need ed information. We encourage child psychiatrists to devote increasi ng energy, mone y, and manpow er to designing and completing methodologically sound proje cts-a task that has not always been well supported by training institutions, funding sources , or ment ors .
Having rece ntly passed the oral Child and Ado lesce nt Psychiatry Boards after two previous unsuccessful atte mpts in 1985 and 1989 and after almost 20 years of private practice, I now feel so mewhat qualified to comment on how to pass and how not to pass the boards, especi ally the oral part. Being " fami liar" with DSM-lI/-R diagnoses is not sufficie nt. For my third attempt, I mem orized all the child and adolescent diagnostic criteria instead. I also mem orized all significa nt aspects of the Ment al Status Exa minatio n as we ll as all significant subjec ts to be cove red in a diagnostic formul ation . Thi s prov ides ment al " shelv ing" upon whic h the clinical information to be gathered can be stored and catego rized. It helps greatly to practice board-like inter views with other candidates in order to have a feel of what to ex pect, includ ing time pressur es. In additio n, it is also helpful to have previously taken the AACAP review course or a previous board exam . I found the video tape/vig nette section the most difficult beca use it was the hardest to prepare for, and the cases present ed often are difficult pat ient s from chaotic, socia lly dep rived famili es , less commonly seen in private practice . They also make for a more challenging and extensive differential diagnos is . It would have been helpful to have been able to purch ase a serie s of practice vignette s and pract ice video tapes to get a better feel for the format and type of patient s that I would face . Even though I took the AACAP annual revi ew course in 1985 and agai n in 1990, such additio nal pract ice opp ortunities would have made my 1990 exper ience eas ier and less anx iety provoking. Not knowing what to expect is one of the most common causes of debilitating anxiety . Overall , I fee l that taking the oral boards did force me to sharpen my diagnost ic skill s, wh ich, I feel, made me a better clinician . I am please d, however , at not having to take the board s again.
Mark 1. Blotcky, M .D. Th om as L. Dimperio , Ph .D . Timberlawn Psychi atric Hospital Dallas, Texas REFERENCE
Pfeiffer , S . 1. & Strzeleck i, B. A. (1990), Inpatient psychiatr ic treatment of children and adolesce nts: a review of outcome studie s. I.
Am. Acad. Child Adolesc. Psychiatry, 29:847-853. I .Am .Acad. Child Adolesc. Psychiatry , 30 :3, May 199/
Dirk E . Hutt enbach , M .D . Mari etta , Georgia
Fenfluramine and Mental Retardation To the Editor: Recent ly, we undert ook a doub le-blind, placebo -controlled crossover study comparing the cl inical ef fects of meth ylphenidate and fen fluram ine in ment ally retarded childr en with sy mpto ms of attentio n deficit hyper acti vity dis order. Twenty ch ildren have been tested thus far , 19 of whom are not autistic. Th is is unique in tha t the large majorit y of previous fenflura mine studies have employed only autistic children as subje cts (Aman and Kern , 1989), and there is only one other clinical trial known to us that com pares fcnflur amin e to another acti ve medication in children (Donnelly et al., 1989). We would like to take this opportunity to present some preliminary find ings regarding the side effects of fenflura mine . We assessed side effects in our subjects through parent reports using a modified version of the Dosage and Treatm ent Emerge nt Sympt om s Scale (DOTES) developed for psychopharmacological trials by the Nationa l Institut es of Menta l Health (Guy, 1976; Rapoport and Conners, 1985). Parents , observi ng red uctions in motor beh avior , reported on the item, " Increase d motor acti vity," where fenfluramine appeared to be causing significant reductions in comparis on with placebo (p 0= 0 .0 14, Wilcoxon s igned ranks tests). The se reductions in mot or be-
507
LEITERS TO THE EDITOR havior were more co nsistent than those seen with meth ylphen idate , which differs from placeb o only at marginal levels (p = 0.056) at this stage of the study . As in other studi es, fenflu ramine caused drowsiness in some children (although not seve re in any case), and the differences appro ached significance both in comparison with placebo (p = 0.052 ) and methylphenidate (p = 0.052) . We observed a nonsignificant trend for fenfluramine to cause anorexia, but so far, this is less apparent than the differences between methylphen idate and placebo (p < 0.10). We have also a seen a nonsignific ant tenden cy for less akathisia with fenfluram ine than with placebo (p = 0 .068) , which may be as a result of a mild anx iolytic effect or may be secondary to the reduction in activity level. One hereto fore unreported side effect is a very vivid report of a 7-year-old girl who developed a peculi ar body odor and halitosis that coincided exactly with the times of fenfluramine treatment. We observed no increa ses in insomnia , dizziness , or diarrhea with fenflur amine, although these are sometimes reported to occur with the drug (Aman and Kern , 1989). We also looked at heart rate, blood pressure , and body weight in these children. Analysis of varian ce indic ated that all three cardiovascular measures were significantly affected by the medications. Fenfluramine was associ ated with reductions in each , but all significant comparisons were with respect to meth ylphenidate , which tended to eleva te heart rate and blood pressure. As in most studies, fenfluramin e produced significant weight losses in these children (p = 0 .01 ). Our subjects lost about 1.8 pound s (0 .82 kg) relative to their placebo weights during the 4 weeks they received medication. Finally, two groups of symptom s that do not appear on the DOTES Scale occurred fairly often in our sampl e. Seven of our subj ects were reported to have difficulties with enure sis and encopresis, and another four demon strated stereotypy and self-injurious behavior. Enure sis and encopresis were lowest when the children were taking methylphenidate, appe aring to worse n both with placebo (p = 0. 068) and fenfluramin e (p = 0.068) . Th e emer gence of these symptoms (usually in children who had previou sly been maintained on methylphenidate) has probably caused more distress than any other side effect in this study . Furthermore , it is a well-est ablished fact that moderate-to-high doses of psychostimul ants can induce stereotypic behavior in laboratory animal s (Randrup and Munk vad, 1974 ), and stero typies may occur de novo (or existing stereotypies may be exacerbat ed) at lower doses of stimulants in children with developmental disabilities (Aman, 1982). Both enure sis/encopresis and stere otypy /self-injur y are obviously of relevance to researcher s and clinicians working with developmentally delayed clinical populations. Other research ers working with development ally delayed children may wish to add these symptoms to their inventory of side effects if they are using the DOTES to monitor side effec ts. Furthe rmore , if the DOTES is revised in the future , its developers should seriously consider addin g these and other relevant items to document such potenti al side effects. Michael G. Aman , Ph.D. Richard A . Kern , M.D . L. Eugene Arnold, M .D . Debbie E. McGhee , B.A . The Nisonger Cent er for Ment al Retardati on and Developmental Disabil ities Ohio State University , Columbus, Ohio REFER ENCES
Aman, M. G. (1982) , Stimul ant drug effects in developmental disorders and hyper activit y-toward a resolution of disparate findings. J . Autism Dev, Disord. , 12:385-398. Aman, M . G . & Kern , R. A. (1989) , Review of fenfluramine in the treatment of the devel opmental disabilitie s. J. Am . A cad . Child Ado/esc. Psychiat ry , 28 :549-565 . Donnell y , M., Rapoport , J . L. , Potter , W. Z., Oliver , J . , Keysor ,
508
C. S . & Murphy, D. L. (1989), Fenfluramine and dextroamphetamine treatment of childhood hyperactivit y . Arch . Gen . Psychiatry , 46:205-21 2. Guy, W . (1976) , ECDEU Assessm ent Manual f or Psychoph armacology. (Rev . Ed.). Rockville , MD : U.S. Department of Health , Education , and Welfare. Randrup , A. & Munkvad , 1. (1974) , Pharm acology and physiolog y of stereotyped behavior. J , Psychiat r . Res. , 11:1-10 . Rapoport , J. L. & Conner s, C . K. (eds .) (1985 ), Special feature: rating scales and assessment instrum ents for use in pediatric psychopharmacology research . Psychoph armaco/ . Bull ., 21:7 19-1125.
Adverse Effects of Fluoxetine To the Editor: Increased anxiety and stereo typical beha viors in response to changing fluoxetine dosage from 20 mg every other day after 2 weeks to 20 mg every day in a 26-year-old white autistic female (Mehlinger et aI., 1990) may have been due to the serotonergic-mediated inhibition of dopamine lateraliz ed to the right hemisphere of the brain in which the metabolic rate is higher in femal es (Friedman and Janko vic, 1990). This hypothesis is supported by the stereotyped, rituali stic behaviors in autisti c individual s resemblin g compul sive behavi or (Mehlinger et aI., 1990) linked to right frontal brain dysfunction (Friedman and Jankovic, 1990). Adver se respon ses to fluoxetine may be the result of a sudden increa se in serotonin that might be avoided by a more gradual dose titration (Papp et aI., 1990). Ernest H . Friedman , M.D. East Cleveland, Ohio REFERENCES
Friedman, E. H. & Jankovic, J. (1990), Tardi ve oculogyria (letter and reply). Neurology . 40 :728. Mehling er , R. , Scheftner, W . A. & Poznanski , E. (1990) , Fluoxetine and autism (letter ). J . Am . Ac ad . ChildAdolesc. Psychiatry, 29:985. Papp , L. A. , Gorman , J. M. , Teicher , M . H . , Glod , C. & Cole, J . O. (1990), Suicidal preoccup ation durin g fluoxetine treatment (letter and reply). Am . J . Psychiatry. 147:1380--1381 .
Fluoxetine in Autism with Depression To the Editor: We enjo yed reading the report on the use of fluoxetine in autism by Mehlin ger et al. (1990). We agree that fluoxetine may be useful in the treatment of persons with autism . Our expe rience with this drug has been simila r to that of Mehlinger et al. with some exceptions as described below: Case 1 J. F. , a 16-year-old girl with autism and moderate mental reta rdation, was referred with a 3-week histor y of chan ge of behav ior that consisted of social withdrawal, povert y of speech, a general lack of interest in hobbie s, and disturbance of sleep and of appetite. Her other behaviors consisted of lining up objects and compulsively carrying obje cts with her. There was a strong famil y history of affective disorder. After a careful assessment, a diagnosis of major depre ssion in addition to her developmental disorder s was made , and treatment with fluoxet ine, 20 mg/d , was started . About 2 weeks after starting the medic ation , her parents reported that the patient was much better . She looked brighter, and her vege tative features also improved. Howe ver, there was no decrease in her compulsive behaviors . In view of the respon se, fluoxetine was increased to 20 mg twice a day . Two week s later , her parents reported l .Am . A cad. Child Ado/esc. Psychiatry, 30:3, May 1991
