263
Psychiatry Research, 431263-216
Elsevier
Neuroendocrine Responses to Challenge With dlFenfluramine and Aggression in Disruptive Behavior Disorders of Children and Adolescents David M. Stoff, Abner P. Pasatiempo, Jupiter Wagner H. Bridger, and Harris Rabinovich
Yeung,
Thomas
B. Cooper,
Received February 12,1992; revised version received June 26, 1992; accepted August 9, 1992. Abstract.
Prolactin (PRL) and cortisol (CORT) responses to a single oral administration (1.0 mg/ kg) of the indirect serotonin agonist dl-fenfluramine were assessed in unmedicated prepubertal and adolescent males with disruptive behavior disorders (DBD). Neuroendocrine responses were correlated with scores on aggression rating scales in prepubertal and adolescent DBD patients and compared with those of matched adolescent normal control subjects. Net dlfenfluramine-induced PRL and CORT release was not correlated with aggression rating scores in prepubertal and adolescent DBD patients and did not differ significantly between adolescent DBD patients and normal control subjects. Although the present study does not demonstrate a serotonergic abnormality in aggression or DBD, this may be more a reflection of limitations of the neuroendocrine challenge test procedures or the methods used than evidence that serotonergic function in the central nervous system is normal in aggression. Key Words. Prolactin, cortisol, serotonin.
A role for serotonin (Shydroxytryptamine; 5HT) in the regulation of aggression and a relationship between decreased 5HT transmission and an increased propensity for aggressive-impulsive acts has been proposed (Coccaro, 1989). Numerous studies have shown reduced levels of 5-hydroxyindoleacetic acid (SHIAA) in the lumbar cerebrospinal fluid (CSF) of adult patients with a history of aggressive behavior, whether directed inwardly toward self (Asberg et al., 1976; Trhkman et al., 1981; van Praag, 1982, 1983; Banki and Arat6, 1983), outwardly toward others (Brown et al., 1979, 1982; Linnoila et al., 1983; Roy et al., 1988), or toward property (Virkunnen et al., 1987). Results from the only CSF SHIAA study in aggressive youngsters are consistent with a 5HT deficit hypothesis in that children and
An earlier version of this report was presented at the Fifth World Congress of Biological Psychiatry, Florence, Italy, June 9-14, 1991. At the time this work was done, David M. Stoff, Ph.D., was Associate Professor; Abner P. Pasatiempo, M.D., was Research Associate; Jupiter H. Yeung, Ph.D., was Assistant Professor; Wagner H. Bridger, M.D., was Professor; and Harris Rabinovich, M.D., was Associate Professor, Department of Psychiatry, Medical College of Pennsylvania at Eastern Pennsylvania Psychiatric Institute, Philadelphia, PA. Thomas B. Cooper, M.A., was Director, Analytical Psychopharmacology Laboratory, Nathan Kline Institute, Orangeburg, NY (Reprint requests to Dr. D.M. Stoff at his current address: Violence and Traumatic Stress Research Branch, Rm. 18C25, National Institute of Mental Health, 5600 Fishers Lane, Rockville, MD 20857, USA.) 0165-178 I/ 92/ $05.00 Q 1992 Elsevier Scientific
Publishers
Ireland
Ltd.
264
adolescents with disruptive behavior disorders (DBD) exhibit a negative correlation between CSF SHIAA concentration and self-rating of aggression toward people (Kruesi et al., 1990). In recent years neuroendocrine challenge tests have been widely used to assess the responsivity of serotonergic neuronal circuits in the central nervous system (CNS) and postsynaptic 5HT receptors. These tests are a practical clinical method since they offer a more ethically justifiable “window into the brain” than CSF analyses to evaluate the role of central monoamine neurotransmitters in youngsters. This method depends on the ability of agents with major effects on SHT function to elicit specific neuroendocrine responses. Such responses are taken to reflect the functional status of central 5HT systems. Negative correlations have been demonstrated between the prolactin (PRL) response to indirect- and direct-acting 5HT agonists and aggressive-impulsive behaviors in adult patients with personality disorders (Coccaro et al., 1989; Moss et al., 1990) but this correlation has not been found in substance abusers (Fishbein et al., 1989) or patients with panic disorder (Wetzler et al., 1991). To the best of our knowledge, there have been no neuroendocrine challenge studies of the 5HT system in children or adolescents with aggressive behavior. We were particularly interested in examining the neuroendocrine responses to dl-fenfluramine (&FEN), an anorectic agent that releases intraneuronal 5HT and inhibits 5HT uptake (Rowland and Carlton, 1986) as a possible 5HT probe in aggressive children and adolescents. Administered as a single dose, &FEN enhances serotonergic transmission by augmenting presynaptic serotonergic function, thereby permitting an assessment of the integrity of 5HT neuronal circuits as a functional unit. The release of PRL and cortisol (CORT) is considered to be under serotonergic control, and 5HT is stimulatory to PRL and CORT in human subjects (found in most, but not all, studies) (reviewed by Van de Kar, 1989). In addition, prepubertal boys with DBD exhibit dose-dependent stimulation of the PRL response to &FEN (Stoff et al., 1989), and this response remains relatively stable over time (Stoff et al., 1992). In this report we evaluated PRL and CORT responses to acute &FEN challenge and determined whether these neuroendocrine responses correlated with scores on aggression rating scales in prepubertal-adolescent patients with DBD and differentiated adolescent patients with DBD from matched control subjects. On the basis of studies in adults, it was hypothesized that the neuroendocrine responses to &FEN would be negatively correlated with aggression or blunted in patients with DBD. Methods Subjects. Study 1: Neuroendocrine
responses to d/-FEN
in male prepubertal
DBD subjects.
We selected 1.5 male prepubertal patients (Tanner stage 1 or 2, mean age = 10.2, SD = 2.5) from our inpatient and day hospital units (Division of Child Psychiatry at the Medical College of Pennsylvania, Eastern Pennsylvania Psychiatric Institute) who met DSM-III- R criteria (American Psychiatric Association, 1987) for one or more DBDs on the basis of the Diagnostic Interview for Children and Adolescents-Parent version (DICA-P; Herjanic and Reich, 1982). The prepubertal patient group consisted of seven subjects with conduct disorder
265 (CD) plus attention deficit hyperactivity disorder (ADHD), four with oppositional defiant disorder (ODD), and four with ODD plus ADHD. Study 2: Neuroendocrine responses to d/-FEN in male adolescent DBD subjects and matched control subjects. We selected eight male adolescent patients (Tanner stage 4, mean age = 14.7, SD = 1.4) from our outpatient unit (Division of Child Psychiatry at the Medical College of Pennsylvania, Eastern Pennsylvania Psychiatric Institute) who met DSMIII-R criteria for one or more DBDs on the basis of the DICA-P. The adolescent patient group consisted of four subjects with CD, two with CD plus ADHD, and two with ODD. Eight healthy volunteer adolescent males (Tanner 4, mean age = 15.3, SD = 1.2), who were matched to the adolescent patient group in age, race, and socioeconomic status, were recruited from the community by local advertisements and paid for their participation. The DICA-P was also administered to the healthy volunteers and confirmed that none of them had a DSM-III-R diagnosis. They were also screened for a personal and family (first-degree relative) history of mental disorders on the basis of a semistructured interview. Major medical or neurological illness was ruled out in all patients and control subjects by physical examination and laboratory tests. DBD patients were included in both studies only if they scored at least 2 standard deviations above norms for similarly aged males on the aggressive and externalizing factors of the Child Behavior Checklist, which was completed by the parent (Achenbach, 1978; Achenbach and Edelbrock, 1979). Exclusion criteria were (1) receiving psychoactive medication during the 2-week period preceding the &FEN challenge test; (2) severe medical illness, endocrinopathies, or heart disease; (3) clinical seizures or other neurological illness; (4) fulfilling DSM-III-R criteria for autism or schizophrenia; (5) obesity (weight/ height ratio > the 95th percentile on the National Center for Health Statistics curve) or chronically malnourished (weight or height below the 3rd percentile); and (6) IQ < 70. Study participation was approved by the Committee for the Protection of Human Subjects of the Medical College of Pennsylvania; children and parents gave informed assent and consent, respectively. d/-FEN Challenge Test. Prepubertal subjects. After subjects fasted overnight (except for water intake), an intravenous catheter (kept open with normal saline solution) was inserted into the forearm vein at 9 a.m. Prechallenge blood samples for measurement of baseline plasma PRL and CORT levels were obtained immediately after catheter insertion (-30 min) in different groups of prepubertal patients who subsequently received at 9:30 a.m. (0 min) single-blind oral administration of either 1.0 mg/kg dl-fenfluramine hydrochloride, Pondimin (&FEN) (n = 15) or placebo (PBO) (n = 5). Postchallenge blood samples for PRL and CORT were obtained at lo:30 a.m. (l-60 min) and every 30 minutes thereafter until I:30 p.m. (+240 min). Due to practical considerations, patients were given a low monoamine breakfast IO-15 minutes after administration of &FEN or PBO, and a low monoamine lunch was given after the i-150-minute blood sample was taken. Samples for plasma dl-fenfluramine and dl-norfenfluramine, the N-dimethyl active metabolite, were obtained hourly from i-60 to i-240 minutes in patients who received &FEN. Adolescent subjects. In a randomized single-blind design, adolescent patients and matched control subjects participated in two challenge tests-PBO and &FEN (1 .O mg/ kg, p.o.)-which were separated by a l-week interval. An intravenous catheter (kept open with normal saline solution) was inserted into the forearm vein at 8:30 a.m. Subjects remained awake, semi-supine, and fasting until 2 p.m. Prechallenge blood samples for baseline PRL and CORT levels were obtained at 9:45 and 9:55 a.m. (-15 and -5 min). After a 90-minute adaptation period, subjects received identical tablets that contained either &FEN (1.0 mg/ kg) or PBO at 10 a.m. (0 min). Postchallenge blood samples for PRL and CORT were obtained every 30 minutes thereafter until 2 p.m. (+240 min). Samples for plasma dl-fenfluramine and dl-norfenfluramine were obtained hourly from +60 to +240 minutes in patients and control subjects who received &FEN.
266 Biochemical Assays. Prepubertal subjects. Plasma PRL levels were determined by a homologous double antibody radioimmunoassay (CIBA Corning, Medfield, MA) and plasma CORT levels by a solid phase competitive radioimmunoassay (Becton-Dickinson and Co., Orangeburg, NY). The interassay coefficients of variation for PRL and CORT were 3.8% and 4.2% respectively, and the intra-assay coefficients of variation for PRL and CORT were 3.7% and 5.7%, respectively. Plasma dl-fenfluramine and dl-norfenfluramine levels were assayed by high performance liquid chromatrography with uv detection after minor modifications (Suckow and Cooper, 1982). The detection limit was 5 ng/ ml for both dl-fenfluramine and dl-norfenfluramine, and the interassay and intra-assay coefficients of variation were < 5% for both compounds. Adolescent subjects. Plasma PRL levels were determined by a homologous double antibody radioimmunoassay (Sinha et al., 1973) and plasma CORT levels by a solid phase competitive radioimmunoassay (Micromedic RIA kit: Micromedic Systems, Inc.). The respective interassay and intra-assay coefficients of variation for PRL were 6.9% and 6.4% at 3.8 ng/ml, 5.7% and 3.8% at 27.0 ng/ml, and 3.5% and 2.0% at 46.3 ng/ml; the respective interassay and intra-assay coefficients of variation for CORT were 7.0% and 5.8% at 3.7 pg/dl, 3.1% and 4.6% at 12.5pg/dl, and 5.3% and 3.3% at 21.3 ,ug/dl. Gas chromatography fitted with a nitrogen detection system was used for quantification of plasma dl-fenfluramine and dl-norfenfluramine levels. This method is highly sensitive with a lower detection limit < 2 ng/ ml and a coefficient of variation of 5.2% (Cooper, personal communication, September 16, 1991). All samples from an individual subject were measured in duplicate in a single assay run and were accepted for the final analysis only if their intra-assay coefficient of variation was < 10%. Aggression Assessment. Because aggression is not a unitary concept, a variety of standardized rating scales were used in an attempt to obtain more homogeneous subtypes of aggression that might correlate with the nemoendocrine responses to &FEN. Before the &FEN challenge test, aggression of prepubertal patients was ascertained by the parent-rated Child Hostility Inventory (Kazdin et al., 1987), the children’s version of the Buss-Durkee Hostility Inventory (Buss and Durkee, 1957), and the parent-rated Interview for Antisocial 1986); aggression of adolescent subjects was Behavior (Kazdin and Esveldt-Dawson, determined using the self-rated Buss-Durkee Hostility Inventory (Buss and Durkee, 1957) the self-rated aggression subscale of the Multidimensional Personality Questionnaire (Tellegen, 1982), and the clinician-rated Brown-Goodwin Assessment for History of Lifetime Aggression (Brown et al., 1979, 1982). Side Effects and Subjective Responses to (II-FEN. Side effects were rated at baseline and hourly thereafter from a list of 32 potential side effects to &FEN (Realmuto et al., 1986). Visual analogue scales (lOO-mm lines in which 0 mm = “not at all” and 100 mm = “most ever”) were scored (by adolescent subjects only) at baseline and at +60 and +240 minutes after &FEN administration. The scales were used to separate 12 pairs of items that characterized opposite mood states, as described by Norris (1971). Scales of mental sedation (alert/drowsy, clear-headed/fuzzy, quick-witted/mentally slow, and attentive/dreamy), physical sedation (strong/feeble, well-coordinated/clumsy, energetic/lethargic, proficient/incompetent), and tranquilization (calm/excited, contented/discontented, tranquil/ troubled, and relaxed/ tense) were constructed by grouping the scores from four similar items. While the patients were in a sitting position, pulse and blood pressure were recorded with an automated vital signs monitor (Dynamap) at baseline and on the hour. Data Analysis. The baseline score was defined as the prechallenge PRL/ CORT 1) or the average of two prechallenge PRL/CORT values (study 2). In the first and CORT responses to &FEN were assessed using two A (change) values measures of the &FEN challenge test: (1) subtracting the baseline score from
value (study study, PRL as outcome the PRL or
267 CORT values at each subsequent time point and selecting the maximal elevation as the peak the baseline score from the change after &FEN (peak APaL, peak A co&, (2) subtracting total area under the curve, using Simpson’s rule, for PRL or CORT (area under the curve A rRL, area under the curve ACOsT). In the second study, a placebo-corrected change from baseline was obtained by subtracting at each time point the value on PBO from the value on &FEN for each individual and selecting the maximal elevation as the placebo-corrected peak change after &FEN (placebo-corrected peak APRL, placebo-corrected peak ACOaT). One-way analysis of variance (ANOVA) for repeated measures was used to determine the significance of time changes in PRL and CORT concentrations and levels of dl-fenfluramine and dlnorfenfluramine. Two-tailed Student’s t tests were performed for intergroup variations and two-tailed paired t tests for intragroup variations. Pearson’s product-moment correlations were calculated by simple linear regression analysis for relationships between neuroendocrine response measures and aggression measures. All data are presented as mean and SD unless otherwise stated. The significance level was set at p < 0.05.
Results Prepubertal Patients. There were no significant differences (by independent z tests) in baseline levels of PRL and CORT among subjects who subsequently received &-FEN and PBO challenges (PRL: pre-FEN vs. pre-PBO mean + SD = 2.7 + 0.7 vs. 2.8 f 0.4 ng/ml; CORT: pre-FEN vs. pre-PBO mean + SD = 7.6 + 3.6 vs. 4.4 +_ 1.7 pg/ dl). After placebo administration, there was negligible change in PRL values (peak APRL ranged from -1.1 to 0.7 ng/ml) and slight variation in CORT values (peak ACoRT ranged from -2.4 to 3.6 pg/dl). Following placebo, delta PRL values remained relatively uniform and stable over time (SDS at different time points ranged from 0.2 to 0.4), but ACoaT values exhibited variability (SDS at different time points ranged from 1.1 to 2.1). Thus, it appears that the CORT (and not the PRL) response to &FEN was subject to nonspecific stress factors. Repeated measures ANOVAs revealed a significant time effect on A nRr_(F= 10.61; df = 6,66;p
Psychiatry Research, 431263-216
Elsevier
Neuroendocrine Responses to Challenge With dlFenfluramine and Aggression in Disruptive Behavior Disorders of Children and Adolescents David M. Stoff, Abner P. Pasatiempo, Jupiter Wagner H. Bridger, and Harris Rabinovich
Yeung,
Thomas
B. Cooper,
Received February 12,1992; revised version received June 26, 1992; accepted August 9, 1992. Abstract.
Prolactin (PRL) and cortisol (CORT) responses to a single oral administration (1.0 mg/ kg) of the indirect serotonin agonist dl-fenfluramine were assessed in unmedicated prepubertal and adolescent males with disruptive behavior disorders (DBD). Neuroendocrine responses were correlated with scores on aggression rating scales in prepubertal and adolescent DBD patients and compared with those of matched adolescent normal control subjects. Net dlfenfluramine-induced PRL and CORT release was not correlated with aggression rating scores in prepubertal and adolescent DBD patients and did not differ significantly between adolescent DBD patients and normal control subjects. Although the present study does not demonstrate a serotonergic abnormality in aggression or DBD, this may be more a reflection of limitations of the neuroendocrine challenge test procedures or the methods used than evidence that serotonergic function in the central nervous system is normal in aggression. Key Words. Prolactin, cortisol, serotonin.
A role for serotonin (Shydroxytryptamine; 5HT) in the regulation of aggression and a relationship between decreased 5HT transmission and an increased propensity for aggressive-impulsive acts has been proposed (Coccaro, 1989). Numerous studies have shown reduced levels of 5-hydroxyindoleacetic acid (SHIAA) in the lumbar cerebrospinal fluid (CSF) of adult patients with a history of aggressive behavior, whether directed inwardly toward self (Asberg et al., 1976; Trhkman et al., 1981; van Praag, 1982, 1983; Banki and Arat6, 1983), outwardly toward others (Brown et al., 1979, 1982; Linnoila et al., 1983; Roy et al., 1988), or toward property (Virkunnen et al., 1987). Results from the only CSF SHIAA study in aggressive youngsters are consistent with a 5HT deficit hypothesis in that children and
An earlier version of this report was presented at the Fifth World Congress of Biological Psychiatry, Florence, Italy, June 9-14, 1991. At the time this work was done, David M. Stoff, Ph.D., was Associate Professor; Abner P. Pasatiempo, M.D., was Research Associate; Jupiter H. Yeung, Ph.D., was Assistant Professor; Wagner H. Bridger, M.D., was Professor; and Harris Rabinovich, M.D., was Associate Professor, Department of Psychiatry, Medical College of Pennsylvania at Eastern Pennsylvania Psychiatric Institute, Philadelphia, PA. Thomas B. Cooper, M.A., was Director, Analytical Psychopharmacology Laboratory, Nathan Kline Institute, Orangeburg, NY (Reprint requests to Dr. D.M. Stoff at his current address: Violence and Traumatic Stress Research Branch, Rm. 18C25, National Institute of Mental Health, 5600 Fishers Lane, Rockville, MD 20857, USA.) 0165-178 I/ 92/ $05.00 Q 1992 Elsevier Scientific
Publishers
Ireland
Ltd.
264
adolescents with disruptive behavior disorders (DBD) exhibit a negative correlation between CSF SHIAA concentration and self-rating of aggression toward people (Kruesi et al., 1990). In recent years neuroendocrine challenge tests have been widely used to assess the responsivity of serotonergic neuronal circuits in the central nervous system (CNS) and postsynaptic 5HT receptors. These tests are a practical clinical method since they offer a more ethically justifiable “window into the brain” than CSF analyses to evaluate the role of central monoamine neurotransmitters in youngsters. This method depends on the ability of agents with major effects on SHT function to elicit specific neuroendocrine responses. Such responses are taken to reflect the functional status of central 5HT systems. Negative correlations have been demonstrated between the prolactin (PRL) response to indirect- and direct-acting 5HT agonists and aggressive-impulsive behaviors in adult patients with personality disorders (Coccaro et al., 1989; Moss et al., 1990) but this correlation has not been found in substance abusers (Fishbein et al., 1989) or patients with panic disorder (Wetzler et al., 1991). To the best of our knowledge, there have been no neuroendocrine challenge studies of the 5HT system in children or adolescents with aggressive behavior. We were particularly interested in examining the neuroendocrine responses to dl-fenfluramine (&FEN), an anorectic agent that releases intraneuronal 5HT and inhibits 5HT uptake (Rowland and Carlton, 1986) as a possible 5HT probe in aggressive children and adolescents. Administered as a single dose, &FEN enhances serotonergic transmission by augmenting presynaptic serotonergic function, thereby permitting an assessment of the integrity of 5HT neuronal circuits as a functional unit. The release of PRL and cortisol (CORT) is considered to be under serotonergic control, and 5HT is stimulatory to PRL and CORT in human subjects (found in most, but not all, studies) (reviewed by Van de Kar, 1989). In addition, prepubertal boys with DBD exhibit dose-dependent stimulation of the PRL response to &FEN (Stoff et al., 1989), and this response remains relatively stable over time (Stoff et al., 1992). In this report we evaluated PRL and CORT responses to acute &FEN challenge and determined whether these neuroendocrine responses correlated with scores on aggression rating scales in prepubertal-adolescent patients with DBD and differentiated adolescent patients with DBD from matched control subjects. On the basis of studies in adults, it was hypothesized that the neuroendocrine responses to &FEN would be negatively correlated with aggression or blunted in patients with DBD. Methods Subjects. Study 1: Neuroendocrine
responses to d/-FEN
in male prepubertal
DBD subjects.
We selected 1.5 male prepubertal patients (Tanner stage 1 or 2, mean age = 10.2, SD = 2.5) from our inpatient and day hospital units (Division of Child Psychiatry at the Medical College of Pennsylvania, Eastern Pennsylvania Psychiatric Institute) who met DSM-III- R criteria (American Psychiatric Association, 1987) for one or more DBDs on the basis of the Diagnostic Interview for Children and Adolescents-Parent version (DICA-P; Herjanic and Reich, 1982). The prepubertal patient group consisted of seven subjects with conduct disorder
265 (CD) plus attention deficit hyperactivity disorder (ADHD), four with oppositional defiant disorder (ODD), and four with ODD plus ADHD. Study 2: Neuroendocrine responses to d/-FEN in male adolescent DBD subjects and matched control subjects. We selected eight male adolescent patients (Tanner stage 4, mean age = 14.7, SD = 1.4) from our outpatient unit (Division of Child Psychiatry at the Medical College of Pennsylvania, Eastern Pennsylvania Psychiatric Institute) who met DSMIII-R criteria for one or more DBDs on the basis of the DICA-P. The adolescent patient group consisted of four subjects with CD, two with CD plus ADHD, and two with ODD. Eight healthy volunteer adolescent males (Tanner 4, mean age = 15.3, SD = 1.2), who were matched to the adolescent patient group in age, race, and socioeconomic status, were recruited from the community by local advertisements and paid for their participation. The DICA-P was also administered to the healthy volunteers and confirmed that none of them had a DSM-III-R diagnosis. They were also screened for a personal and family (first-degree relative) history of mental disorders on the basis of a semistructured interview. Major medical or neurological illness was ruled out in all patients and control subjects by physical examination and laboratory tests. DBD patients were included in both studies only if they scored at least 2 standard deviations above norms for similarly aged males on the aggressive and externalizing factors of the Child Behavior Checklist, which was completed by the parent (Achenbach, 1978; Achenbach and Edelbrock, 1979). Exclusion criteria were (1) receiving psychoactive medication during the 2-week period preceding the &FEN challenge test; (2) severe medical illness, endocrinopathies, or heart disease; (3) clinical seizures or other neurological illness; (4) fulfilling DSM-III-R criteria for autism or schizophrenia; (5) obesity (weight/ height ratio > the 95th percentile on the National Center for Health Statistics curve) or chronically malnourished (weight or height below the 3rd percentile); and (6) IQ < 70. Study participation was approved by the Committee for the Protection of Human Subjects of the Medical College of Pennsylvania; children and parents gave informed assent and consent, respectively. d/-FEN Challenge Test. Prepubertal subjects. After subjects fasted overnight (except for water intake), an intravenous catheter (kept open with normal saline solution) was inserted into the forearm vein at 9 a.m. Prechallenge blood samples for measurement of baseline plasma PRL and CORT levels were obtained immediately after catheter insertion (-30 min) in different groups of prepubertal patients who subsequently received at 9:30 a.m. (0 min) single-blind oral administration of either 1.0 mg/kg dl-fenfluramine hydrochloride, Pondimin (&FEN) (n = 15) or placebo (PBO) (n = 5). Postchallenge blood samples for PRL and CORT were obtained at lo:30 a.m. (l-60 min) and every 30 minutes thereafter until I:30 p.m. (+240 min). Due to practical considerations, patients were given a low monoamine breakfast IO-15 minutes after administration of &FEN or PBO, and a low monoamine lunch was given after the i-150-minute blood sample was taken. Samples for plasma dl-fenfluramine and dl-norfenfluramine, the N-dimethyl active metabolite, were obtained hourly from i-60 to i-240 minutes in patients who received &FEN. Adolescent subjects. In a randomized single-blind design, adolescent patients and matched control subjects participated in two challenge tests-PBO and &FEN (1 .O mg/ kg, p.o.)-which were separated by a l-week interval. An intravenous catheter (kept open with normal saline solution) was inserted into the forearm vein at 8:30 a.m. Subjects remained awake, semi-supine, and fasting until 2 p.m. Prechallenge blood samples for baseline PRL and CORT levels were obtained at 9:45 and 9:55 a.m. (-15 and -5 min). After a 90-minute adaptation period, subjects received identical tablets that contained either &FEN (1.0 mg/ kg) or PBO at 10 a.m. (0 min). Postchallenge blood samples for PRL and CORT were obtained every 30 minutes thereafter until 2 p.m. (+240 min). Samples for plasma dl-fenfluramine and dl-norfenfluramine were obtained hourly from +60 to +240 minutes in patients and control subjects who received &FEN.
266 Biochemical Assays. Prepubertal subjects. Plasma PRL levels were determined by a homologous double antibody radioimmunoassay (CIBA Corning, Medfield, MA) and plasma CORT levels by a solid phase competitive radioimmunoassay (Becton-Dickinson and Co., Orangeburg, NY). The interassay coefficients of variation for PRL and CORT were 3.8% and 4.2% respectively, and the intra-assay coefficients of variation for PRL and CORT were 3.7% and 5.7%, respectively. Plasma dl-fenfluramine and dl-norfenfluramine levels were assayed by high performance liquid chromatrography with uv detection after minor modifications (Suckow and Cooper, 1982). The detection limit was 5 ng/ ml for both dl-fenfluramine and dl-norfenfluramine, and the interassay and intra-assay coefficients of variation were < 5% for both compounds. Adolescent subjects. Plasma PRL levels were determined by a homologous double antibody radioimmunoassay (Sinha et al., 1973) and plasma CORT levels by a solid phase competitive radioimmunoassay (Micromedic RIA kit: Micromedic Systems, Inc.). The respective interassay and intra-assay coefficients of variation for PRL were 6.9% and 6.4% at 3.8 ng/ml, 5.7% and 3.8% at 27.0 ng/ml, and 3.5% and 2.0% at 46.3 ng/ml; the respective interassay and intra-assay coefficients of variation for CORT were 7.0% and 5.8% at 3.7 pg/dl, 3.1% and 4.6% at 12.5pg/dl, and 5.3% and 3.3% at 21.3 ,ug/dl. Gas chromatography fitted with a nitrogen detection system was used for quantification of plasma dl-fenfluramine and dl-norfenfluramine levels. This method is highly sensitive with a lower detection limit < 2 ng/ ml and a coefficient of variation of 5.2% (Cooper, personal communication, September 16, 1991). All samples from an individual subject were measured in duplicate in a single assay run and were accepted for the final analysis only if their intra-assay coefficient of variation was < 10%. Aggression Assessment. Because aggression is not a unitary concept, a variety of standardized rating scales were used in an attempt to obtain more homogeneous subtypes of aggression that might correlate with the nemoendocrine responses to &FEN. Before the &FEN challenge test, aggression of prepubertal patients was ascertained by the parent-rated Child Hostility Inventory (Kazdin et al., 1987), the children’s version of the Buss-Durkee Hostility Inventory (Buss and Durkee, 1957), and the parent-rated Interview for Antisocial 1986); aggression of adolescent subjects was Behavior (Kazdin and Esveldt-Dawson, determined using the self-rated Buss-Durkee Hostility Inventory (Buss and Durkee, 1957) the self-rated aggression subscale of the Multidimensional Personality Questionnaire (Tellegen, 1982), and the clinician-rated Brown-Goodwin Assessment for History of Lifetime Aggression (Brown et al., 1979, 1982). Side Effects and Subjective Responses to (II-FEN. Side effects were rated at baseline and hourly thereafter from a list of 32 potential side effects to &FEN (Realmuto et al., 1986). Visual analogue scales (lOO-mm lines in which 0 mm = “not at all” and 100 mm = “most ever”) were scored (by adolescent subjects only) at baseline and at +60 and +240 minutes after &FEN administration. The scales were used to separate 12 pairs of items that characterized opposite mood states, as described by Norris (1971). Scales of mental sedation (alert/drowsy, clear-headed/fuzzy, quick-witted/mentally slow, and attentive/dreamy), physical sedation (strong/feeble, well-coordinated/clumsy, energetic/lethargic, proficient/incompetent), and tranquilization (calm/excited, contented/discontented, tranquil/ troubled, and relaxed/ tense) were constructed by grouping the scores from four similar items. While the patients were in a sitting position, pulse and blood pressure were recorded with an automated vital signs monitor (Dynamap) at baseline and on the hour. Data Analysis. The baseline score was defined as the prechallenge PRL/ CORT 1) or the average of two prechallenge PRL/CORT values (study 2). In the first and CORT responses to &FEN were assessed using two A (change) values measures of the &FEN challenge test: (1) subtracting the baseline score from
value (study study, PRL as outcome the PRL or
267 CORT values at each subsequent time point and selecting the maximal elevation as the peak the baseline score from the change after &FEN (peak APaL, peak A co&, (2) subtracting total area under the curve, using Simpson’s rule, for PRL or CORT (area under the curve A rRL, area under the curve ACOsT). In the second study, a placebo-corrected change from baseline was obtained by subtracting at each time point the value on PBO from the value on &FEN for each individual and selecting the maximal elevation as the placebo-corrected peak change after &FEN (placebo-corrected peak APRL, placebo-corrected peak ACOaT). One-way analysis of variance (ANOVA) for repeated measures was used to determine the significance of time changes in PRL and CORT concentrations and levels of dl-fenfluramine and dlnorfenfluramine. Two-tailed Student’s t tests were performed for intergroup variations and two-tailed paired t tests for intragroup variations. Pearson’s product-moment correlations were calculated by simple linear regression analysis for relationships between neuroendocrine response measures and aggression measures. All data are presented as mean and SD unless otherwise stated. The significance level was set at p < 0.05.
Results Prepubertal Patients. There were no significant differences (by independent z tests) in baseline levels of PRL and CORT among subjects who subsequently received &-FEN and PBO challenges (PRL: pre-FEN vs. pre-PBO mean + SD = 2.7 + 0.7 vs. 2.8 f 0.4 ng/ml; CORT: pre-FEN vs. pre-PBO mean + SD = 7.6 + 3.6 vs. 4.4 +_ 1.7 pg/ dl). After placebo administration, there was negligible change in PRL values (peak APRL ranged from -1.1 to 0.7 ng/ml) and slight variation in CORT values (peak ACoRT ranged from -2.4 to 3.6 pg/dl). Following placebo, delta PRL values remained relatively uniform and stable over time (SDS at different time points ranged from 0.2 to 0.4), but ACoaT values exhibited variability (SDS at different time points ranged from 1.1 to 2.1). Thus, it appears that the CORT (and not the PRL) response to &FEN was subject to nonspecific stress factors. Repeated measures ANOVAs revealed a significant time effect on A nRr_(F= 10.61; df = 6,66;p
