Platelet Imipramine Binding in Depressed Children and Adolescents PAUL 1. AMBROSINI, M.D., CLAUDIA METZ, M.D., RAMESH C. ARORA, PHD., JAR-CHI LEE, M.A., LISA KREGEL, B.A., AND HERBERT Y. MELTZER, M.D.
Abstract. Kinetic constants of platelet imipramine binding were determined in youths with major depression and a contrast group. Subjects actively depressed (N = 10) had significantly fewer imipramine binding sites (B m",) (877 ± 148 fmol/mg protein) than recovering depressives (N = 12) (1220 ± 428 fmol/mg protein) and contrasts (N = 10) (1270 ± 230 fmol/mg protein). Affinity constants (~) (1.14 ± 0.36 nM, 0.97 ± 0.31 nM, and 1.17 ± 0.39 nM, respectively) were similar among the groups. Actively depressed males but not females had fewer imipramine binding sites than both their sex-matched comparison groups. Although actively depressed females' Bm", was significantly lower than recovering depressed and nondepressed males, neither age, Tannner stage, nor circannual rhythms influenced Bm"" but suicidality may be associated with low Bm",. A decrease in Bm", may be a state-specific marker of major depression in boys or associated with a depressive disorder with a suicidal history. J. Am. Acad. Child Adolesc. Psychiatry, 1992, 31, 2:298-305. Key Words: imipramine binding, depression, suicidality. Since Briley et aI. (1980) and Paul et aI. (1981) first reported a decreased number of platelet 3H-imipramine binding (lB) sites in depressed adults compared with normal controls, numerous investigators have replicated these findings, but these results are controversial (Bech et aI., 1988; Kanof et aI., 1987; Langer et aI., 1986; Nanki et aI., 1986; Nemeroff et aI., 1988; Plenge et aI., 1988; Schneider et aI., 1985). Studies in bipolar disorder are inconsistent when comparisons were made among euthymic, hypomanic, or depressed bipolar states (Baron et aI., 1986; Berrettini et aI., 1982; Lewis et aI., 1985a; Mellerup et aI., 1982; Meltzer et aI., 1990; Muscettola et aI., 1986). Besides methodological differences in the IE assay, several variables could account for these differences across studies. For example, the patient populations investigated are frequently a diagnostically mixed group of depressive subtypes at various stages of illness, treatment, or recovery. In some studies, IB was analyzed without a sufficiently long washout period from previous tricyclic or benzodiazepine treatment (Braddock et aI., 1984; Nutt et aI., 1985; Poirier et aI., 1985; Wagner et aI., 1987). Age ( Baron et aI., 1984; Langer et aI., 1980; Marazziti et aI., 1987; Schneider et aI., 1985) and seasonal variations may also influence IB (Arora et aI., 1988; Baron et aI., 1988; Egrise et aI., 1983; Galzin et aI., 1986; Goziotis et aI., 1988; Tang et aI., 1985; Whitaker et aI., 1984). Nevertheless, a body of evidence has emerged suggesting
Accepted July 17, 1991. From the Department of Psychiatry of the Medical College of Pennsylvania/Eastern Pennsylvania Psychiatric Institute, Philadelphia, Pennsylvania (Dr. Ambrosini) and the Division of Biological Psychiatry, Case Western Reserve University, Cleveland, Ohio (Drs. Metz, Arora, Meltzer, Ms. Lee, and Ms. Kregel). The authors would like to thank James Leckman, M.D., and Wagner Bridger, M. D., for their helpful comments in the preparation of this manuscript. This study was funded in part by NIMH grant MH41684 and presented in part at the 1988 ANCP meeting in Puerto Rico. Reprint requests to Dr. Ambrosini, MCPfEPPI, 3200 Henry Avenue, Philadelphia, PA 19129. 0890-8567/92/3102-0298$03.00/0© 1992 by the American Academy of Child and Adolescent Psychiatry.
298
that IB is abnormal in subsets of patients with affective disorder. Whether IB is an age dependent marker of subtypes of depression (Meltzer et aI., 1990), an age independent marker in familial pure depressive cohorts (Lewis et aI., 1985b), or is associated with specific symptom complexes (Carstens et aI., 1986; Meltzer et aI., 1986; Wagner et aI., 1987) awaits further investigation. It also is uncertain whether IB is a state or a trait marker of depressive illness because suggestive evidence exists for both positions, yet follow-up studies are marginally designed (Arora et aI., 1985; Baron et aI., 1987; Kanof et aI., 1987; Langer et aI., 1986; Maj et aI., 1988; Marazziti et aI., 1988a; Raisman et aI., 1981; Suranyi-Cadotte et aI., 1982; 1985; Wagner et aI., 1987). Platelet IE investigations in children and adolescents suggest that low IE occurs in enuresis (Weizman, A. et aI., 1985; Weizman, R. et aI., 1986a), obsessive-compulsive disorder (Weizman, A. et aI., 1986), and anorexia nervosa (Weizman, R. et aI., 1986b), but normal values exist in autism (Anderson et aI., 1984; Weizman, A. et aI., 1987) or attention deficit disorder (Weizman et aI., 1988). The two studies of IE in juvenile major affective disorder reported conflicting results. Rehavi et aI. (1984) did not find significant differences between depressed youngsters and psychiatric controls consisting exclusively of conduct disordered and schizophrenic subjects, whereas Carstens et aI. (1988) found significantly increased IE in the depressed group compared with normal controls. The reasons for these discrepant results are not clear, particularly given that both groups used a similar IE assay previously described by Paul et aI. (1981). The current study was designed to test the hypothesis that the number of platelet IE sites was low in children and adolescents in an active phase of unipolar major depression compared with a contrast group of nondepressed emotionally disordered patients or nonpsychiatric normal children. The preliminary findings suggest that platelet IE is significantly lower in the depressed group during an active episode of major depression and that IE may be a state-dependent marker of major depressive disorder (MDD) and/or associated with suicidality in this age group. J. Am. Acad. Child Adolesc. Psychiatry, 31:2, March 1992
IMIPRAMINE BINDING IN DEPRESSED YOUTHS
Patients and Methods The subjects studied were predominantly outpatients attending a specialty child depression clinic located at University Hospital of Cleveland. The clinic specialized in the assessment and treatment of depressed and anxious youngsters. Referrals came from both the general child outpatient services and from school and medical personnel who had been informed of the clinic's function by mailed brochures or direct presentation to professional groups. Individuals were excluded from the study if they were physically ill, severely psychotic, mentally retarded, or had symptoms clearly suggesting anorexia/bulimia nervosa, schizophrenia, or autism. Of the 32 subjects entered into this study, 30 were outpatients and two were adolescent inpatients. No subject received psychotropics nor was any ever treated with antidepressants before entry into the study. Diagnostic Assessment All subjects were screened initially for signs or symptoms of depressive or anxiety disorder during the first phone contact with the clinic's psychiatrists. Those callers who clearly had primary behavioral or exclusionary problems as described above were referred to the general child psychiatry outpatient service. The remainder of the contacts were scheduled for an extensive history gathering session with the parents. After this meeting, two sessions were scheduled within 24 to 72 hours for a diagnostic assessment with the third revised edition of the child version of the Schedule for Affective Disorders and Schizophrenia (K-SADSIIIR) (Ambrosini et al., 1989). The K-SADS-IIIR, which is an updated edition of the 1978 K-SADS, covers symptoms necessary not only for the affective syndromes but also for all the major anxiety and behavioral disorders first occurring in childhood and adolescence according to the third edition of the Diagnostic and Statistical Manual (DSM-III). Initially, the parent is interviewed alone about the child's symptomatology, and, in the second session, the child is given the same interview. A summary diagnosis is immediately generated from this information. The two senior authors achieved a K of 0.77 to 0.80 for the diagnosis of MDD in live test-retest (completed within 72 hours) or videotape reliability studies, respectively, with the K-SADS IIIR (Ambrosini et aI., 1989). Subjects were scheduled for a 2-week K-SADS in 10 to 14 days (i.e., K-SADS Time 2) if the summary K-SADS diagnoses generated at Time 1 met one of the following diagnostic patterns: (1) were actively in a episode of MDD; (2) were actively in an episode of a nondepressed anxiety disorder with or without oppositional disorder; (3) had experienced an MDD episode within the last 12 months, but did not meet diagnostic criteria for MDD in the last week preceding the assessment, yet may have met Research Diagnostic Criteria (RDC) (Spitzer et aI., 1978) for minor depression. All subjects included in this latter group were symptom free for less than 2 months. This second interview was done blind to the diagnostic status found at K-SADS Time 1. After the K-SADS Time 2, the following diagnostic groups emerged: (1) those persistently meeting criteria for J. Am. A cad. Child Adolesc. Psychiatry, 31:2, March 1992
a major depressive episode; these subjects were defined as being in an active phase of major depression (MDD-A); (2) those initially diagnosed MDD but now no longer meeting criteria for MDD or just meeting criteria for minor depression; (3) those initially diagnosed with a prior MDD, with or without a residual minor depression, who continued in the same diagnostic pattern; these latter two groups were defined as being in a recovering phase of their major depression (MDD-R) because all subjects still were not more than 2 months free of meeting Research Diagnostic Criteria for MDD; and (4) those persistently meeting diagnostic criteria for nondepressed anxiety syndromes formed the control group. No patients were psychotically depressed or met criteria for bipolar disorder. Of the two teenage inpatients, one male was in the MDD-A group and one female in the MDD-R cell. Four normal subjects whose parents were hospital staff members were also included in the nondepressed cell. These youngsters were assessed once with the epidemiological version of the K-SADS (Orvaschel, 1985) and found to have no lifetime history or current episode of psychopathology. Within the MDD-R group, one girl was not given a subsequent K-SADS but was followed clinically for several months and remained well. Among the nondepressed psychiatric controls, two girls with emotional and behavioral disorders were not seen for a subsequent K-SADS. Platelet Preparation and Binding Assay All blood samples were drawn between 8 A.M. and 9:30 A.M. before any pharmacotherapy. No subject had taken any psychotropic medication, although several subjects evenly distributed among the three study groups took analgesics within the last week. If they were taking over-the-counter preparations, these were discontinued at least 1 week before sampling. Thirty ml of whole blood was drawn with ethylenediamine tetraacetic acid as the anticoagulant. Platelet-rich plasma (PRP) was obtained by centrifugation, and platelet membranes were prepared by lysis and polytron homogenization (Langer et aI., 1980; Sahai et al., 1981). Platelet membranes were incubated with six concentrations of 3H-imipramine (sp. act., 52.6 Ci/mmole) between 0.36nM in the presence and absence of desipramine (100 l!M). Specific binding of the tritiated imipramine as calculated from Scatchard analysis of the data was defined as the difference in binding in the presence and absence of desipramine. Protein in the membrane preparation was determined following the method of Lowry et aI. (1951). The coefficients of variations for ~ and Bmax were 5.2% and 5.4% for split duplicates obtained from 15 normal controls assayed the same day and 7.5% and 8.2% at an interval of 3 to 4 weeks. Data Analysis Differences among independent subject groups were compared by computing X2 tests for categorical or proportional data and analysis of variance (ANOVA) for continuous data with post-hoc comparisons. The Mann-Whitney U test was used to test for significant differences between two independent groups with non-normally distributed data. The effects of sex and diagnostic status were analyzed using a twoway ANOVA; Tukey's multiple comparison procedure was
299
AMBROSINI ET AL.
TABLE 1. Demographic Statistics Major Depressive Disorder-Active (N = 10) Age (yr)a
Major Depressive Disorder-Recovering (N = 12)
15.5 :':: 1.7
13.3 :':: 3.1
Nondepressed Contrast (N = 10) 13.1 :':: 4.2
Sex (MIF)
4/6
5/7
5/5
PrepubertaVadolescent
1/9
6/6
4/6
4.5 :':: 1.0
3.2 :':: 1.9
3.2 :':: 1.6
49.4 :':: 14
43.5 :':: 17
43.4 :':: 15
Height (cm)a
166.2 :':: 12
150.3 :':: 14
156.5 :':: 24
Weight (kg)a
62.2 :':: 21
44.9 :':: 15
52.2 :':: 18
Tanner" SEsa,b
a b
Statistical Values
= 1.9 = 0.17 = 0.24, = 0.89 = 4.07, P = 0.13 F(2, 29) = 2.51 p = 0.10 F(2, 29) = 0.50 p = 0.61 F(2, 29) = 2.4 p = 0.11 F(2, 29) = 2.6 p = 0.09 F(2, 29)
p X2 p X2
Mean :':: SD. Hollingshead's two-factor socioeconomic status.
used for the between-group comparisons. Correlations were calculated using either a Pearson's or Spearman rho coefficient depending on the normality of the data distribution. Bonferroni inequalities were applied to decrease the risk of obtaining significant associations when multiple comparisons were made with the data set (Grove et aI., 1982). Results
Sample characteristics
Thirty-two subjects were accepted into the study. Demographic statistics for the three diagnostic groups are presented in Table 1. The trend for weight differences among the three groups is accounted for by one 15 year-8-monthold male in the MDD-A cell who measured 184.5 cm and weighed 104 kg and the fact that the actively depressed subjects were slightly older. Because Tanner Stage and frequency of prepubertal versus adolescent status was similar among the three groups, the data set was not analyzed separately for levels of sexual maturation. Parameters assessing depression severity and dysfunction are compared across the three experimental groups in Table 2. A reliable l7-item depression scale derived from the K-SADS (Ambrosini et aI., 1989), which is similar to a Hamilton Depression Rating Scale (Endicott et aI., 1981), was compared at three time points. These time points are the summary depression scores rated at the initial and 2week K-SADS. The Summary Present Episode (SPE[l]) score refers to the severity of depressive symptomatology during the worst period of the episode in the last year, whereas the Summary Last Week (SLW[l]) score is the depression severity rating during the last week before the first K-SADS interview. There is only a Summary Last Week rating at the second K-SADS (SLW[2]) because the evaluation at that time assessed depression severity just since the prior interview. As four normal contrast subjects were evaluated with the K-SADS-E, a l7-item depression score could not be generated; nor were such scores available 300
at SLW(2) for the subjects not seen for a 2-week K-SADS. The duration ofMDD and the age of onset were not applicable for the contrast group. Also the Global Assessment Scores (GAS) (Shaffer et aI., 1983) were not available for the four normal subjects assessed with the K-SADS-E. The MDD-A group was significantly more severely depressed at all three time frames compared with the MDD-R or contrast groups. The groups differed only in their depression severity and not in their level of functional impairment because the GAS scores were not different among the groups. Although duration of illness was absolutely longer in the MDD-R cohort, by definition the MDD-A subjects were still ill and so it is unclear how long their illness would be once they were in a recovering state. Imipramine Binding The maximal number of binding sites (B max ) and the dissociation constant (:KI) were similar in the normal control group and the nondepressed psychiatric controls (:KI = 1.39 ± 0.23 vs. 1.03 ± 0.43 nM; t = 1.52, p = NS; Bmax = 1320 ± 323 vs. 1236 ± 170 fmole/mg protein, t = 0.55, p = NS). When these parameters were compared in the three study cells, there was no difference in the :KI; however, the Bmax was significantly lower (p < 0.014) in the MDD-A cohort compared with the MDD-R and nondepressed contrast groups who had similar Bmax values. The mean Bmax value for the actively depressed sample was approximately 30% less than that of the recovering depressed subjects or those not depressed (Table 3). These differences are graphically displayed in Figure 1. Circannual influences were also analyzed among the three study groups by comparing the proportion of subjects within each diagnostic cell whose blood samples were obtained during each of the four seasons. Although the MDD-R cohort had a skewed distribution with 11 samples obtained in the spring and one during the winter, the proportions were more evenly distributed within both the MDD-A and contrast cases and were not significantly different (X 2 = 1.18, J. Am. Acad. Child Adolesc. Psychiatry, 31:2, March 1992
IMIPRAMINE BINDING IN DEPRESSED YOUTHS TABLE
2. Severity Parameters
Major Depressive Disorder-Active (N = 10)
Major Depressive Disorder-Recovering (N = 12)
Nondepressed Contrast (N = 6
81.1 ± 6.5
67.5 ± 8.5
56.7 ± 12.6
Depression Scale: 17 item SPE (1)0
Statistical Values
SLW (l)b
74.0 ± 9.8
54.4 ± 10.3
52.3 ± 11.9
SLW (2)'
71.0 ± 11.2
48.4 ± 7.5
58.5 ± 21.0 (N = 4)
= 15.0 = 0.0001 F(2, 25) = 12.1 p = 0.0002 F(2,22) = 9.78 p = 0.0009
55.0 ± 9.1
F(2, 25)
F(2, 25)
p
(N
Global Assessment Scale SPE (1)0
38.7 ± 16.4
=
11)
43.5 ± 19.3
= 0.80 = 0.46 F(2, 25) = 0.87 p = 0.43 F(2, 22) = 0.11 P = 0.89 Z = 0.09 p = 0.92 Z = 1.22 p = 0.22 p
SLW (l)b
42.8 ± 16.2
51.1 ± 23.2
57.3 ± 9.4
, SLW (2)'
54.1 ± 7.6
61.2 ± 22.8
47,0 ± 10.1 (N = 4)
(N
55.2 + 50 (4-168) 14.5 ± 1.2
Duration (wks) (range) Age at onset
11)
73 ± 98 (8-342) 11.9 ± 4.1
d
d
SPE(I) summary present episode, first interview. SLW(I) summary last week, first interview. c SLW(2) summary last week, second interview. d Not applicable. o
b
TABLE
3. Platelet Imipramine Binding Parameters in Child and Adolescent Major Depression
Major Depressive Disorder-Active (N = 10)
Major Depressive Disorder-Recovering (N = 12)
Nondepressed Contrast (N = 6
1.14 ± 0.36
0.97 ± 0.31
1.17 ± 0.39
877 ± 1480
1,220 ± 428
1,270 ± 230
Statistical Values
= 1.08 p = 0.35 F(2, 29) = 5.01 p < 0.014 F(2, 29)
Note: ~ = nM ± SD; Bmax = fmole/mg protein ± SD. o MDD-A < MDD-R (p < 0.02); Contrast (p < 0.001).
p = NS). Because a major finding is a B max difference be2000
• •
Smax
Kd
1600
• •
1400
0
•
1000 800
i
0
9 0
8
•
0 0
• • ••
0
1.6
•0
0
-
~
.0..
1200
•
0
1800
••
0
•
-i-
•
CD
0
•
~
9
MDD-A
MDD-R
1.4 1.2 1.0
••
0
0
0 0
1.8
.8 .6
• Contrast
=fmoles/mg
MDD-A
MDD-R
Smax
• Non-suicidal
MDD-A'Major depressive disorder-Active
protein
Kd
Contrast
=nM
o Suicidal
-R'Recovering
FIG. 1. Bmax and ~ in depressed and nondepressed children and adolescents. J. Am. Acad. Child Adolesc. Psychiatry, 31: 2, March 1992
tween these two latter groups, circannual rhythm disturbances between these diagnostic groups could not account for this result. Within-group correlations of K.i and B max with age, sex, Tanner stage, duration of disorder, age of onset, 17-item depression scale scores, and GAS were not significant after Bonferroni correction (Table 4). It was suggested that in the MDD-R cohort, an earlier age of onset was associated with a higher Bmax • However, when the analysis was controlled for age with an analysis of covariance, B max was still significantly different in the three groups (P < 0.015). B max in the MDD-A remained lower than in the MDD-R (p < 0.04) or in the contrast samples. Severity of depression in the last week preceding the blood sampling showed a positive association with B max in the MDD-A group. This suggested that individuals with a more severe depression had a higher number of binding sites so that low B max is not an index of depression severity. An exploratory group analysis of B max in the three study 301
AMBROSINI ET AL. TABLE
4. Within-group Correlations among Actively Depressed (MDD-A), Recovering Depressed (MDD-R), and Nondepressed Contrast MDD-A (N = 10)
MDD-R (N = 12)
Contrast (N = 10)
KJB"""
KJBrnax
KJB"""
-0.23/-0.51 -0.06/0.71" 0.40/-0.29 -0.25/-0.20 0.26/-0.50 -0.09/0.72"
0.43/-0.62" 0.22/-0.41 0.31/-0.53 -0.20/0.38" 0.28/-0.65 0.31/0.27
0.37/0.25 -0.03/-0.17 0.54/0.24
(N = II)
(N = 4)
0.09/-0.21
0.21/-0.21
0.86/0.32
0.13
Age Sexb Tanner Duration Age at onset 17-item depression scale (SLW[2]) GAS (SLW[2])
-0.17
(N
0.10
-0.98 a
= II)
(N
= 4)
" p < 0.05 (without Bonferroni). b Male = 0, female = 1. c
Not applicable.
TABLE
5. Imipramine Binding (B""",): Sex and Group Interactions (p values) Male MDD-A (N = 4)
Male MDD-A MDD-R
CaNT Female MDD-A MDD-R
CaNT
753 1461 1324
caNT (N = 5)
MDD-A (N = 6)
MDD-R
CaNT
(N = 7)
(N = 5)
0.001
0.006 0.45
0.26 0.007 0.04
0.11 0.02 0.11
0.02 0.18 0.55
0.58
0.15 0.32
960 1048 1215
groups was also conducted because in the MDD-A group female sex was correlated with high Bmax values. The sex by diagnostic status ANOVA showed a significant overall trend, F(5,26) = 3.21, p < 0.057 (Table 5). Actively depressed males and females had similarly low B max ; however, whereas the male MDD-A group differed from their two sex-matched comparison groups (but only from the MDD-R after Bonferroni correction), the female MDD-A group did not differ from the MDD-R and nondepressed contrast females. The MDD-A females, however, are significantly different than the MDD-R and the nondepressed males. These data suggest that decreased platelet IE sites may be more specific as a mark of illness in depressed boys rather than girls, or that boys attain a more rapid return to normal platelet IE kinetics. The IE data were also analyzed according to the suicide status of the patient. Individuals were defined as suicidal using K-SADS data if within the last year before either the first or second K-SADS, they acknowledged suicidal ideation, planning, gestures, or attempts. Subjects were thus classified dichotomously as suicidal or nonsuicidal or as attempters or nonattempters. Those with nonspecific wishes to be dead or not be born were not considered suicidal. Of the 10 MDD-A subjects, five were suicidal ideators, four were attempters, and one was nonsuicidal. Among the MDD-R group, three were ideators, two were attempters, and seven nonsuicidal. All contrast cases were nonsuicidal.
302
Female
MDD-R (N = 5)
The Bmax did not differ between the suicidal ideators and the attemptors (N = 9) within either the MDD-A or MDD-R groups. The ideators and attempters thus were combined and a comparison made among the following four cells: MDD-A suicidal (MDD-NS), MDD-R nonsuicidal (MDD-RlNS), MDD-R suicidal (MDD-RlS) , and controls. The number of IE sites was significantly lower in the MDDNS group compared with the MDD-RINS (p < 0.002) and contrast cells (p < 0.005). The MDD-RiS were significantly different from the MDD-RINS (p < 0.029) and showed a trend for a lower B max compared with the contrast group (p < 0.097). The B max values did not differ significantly between either the MDD-NS and MDD-RiS or between the MDDRlNS and controls. There was no difference among these four groups in :Ki values. Discussion The results of this study support the primary hypothesis that the number of platelet IE sites are decreased in youngsters during an active phase of MDD. The two previous studies of platelet IE in MDD in children and adolescents contradict each other, and both are at odds with these present findings. In Rehavi et al.' s study (1984), although a K-SADS assessment was administered twice over 2 weeks, the active depressives were a mixed bipolar and unipolar sample, whereas the controls were exclusively conduct disordered or schizophrenic. Prior reports suggest that these J. Am. Acad. Child Adolesc. Psychiatry, 31 :2, March 1992
IMIPRAMINE BINDING IN DEPRESSED YOUTHS
may be inappropriate comparison groups, because the latter two diagnostic groups may have low platelet IE (Arora et aI., 1986; Birmaher et aI., 1990; Stoff et al., 1987). Carstens et aI. (1988) found elevated IB in depressed juvenile depressives, compared with normal controls. However, whereas this group assessed their subjects with a reliable interview schedule (i.e., Interview Schedule for Children), only one assessment was completed, so it remains uncertain how many of these depressed adolescents were truly in an active phase of illness. Carsten's control sample was composed of normal school children, and this also may not be an appropriate group for biological validation studies because a comparison between pathological cases and normal controls reveals nothing about the specificity of a diagnosis with respect to other disorders (Werry et aI., 1987). The decision to diagnostically differentiate patients into those in active illness or in a recovering state was made before the collection of the data and was based on the clinical characteristics of the patient flow. The authors have found that approximately 44% of those who completed two K-SADS assessments had remission of their syndromal major depression between the two interviews. This clinical dichotomy received further validation as a distinct illness status in an expanded sample of 122 adolescents (which includes the adolescents in this present study) assessed and diagnosed by the authors' group in a similar manner, who also completed a Beck's Depression Inventory (BDI) (Beck et aI., 1961) before each K-SADS. Those adolescents defined as MDD at K-SADS Time 1, who were then also depressed at K-SADS Time 2 (i.e., MDD-A), had significantly higher mean BDI scores than the MDD-R at both KSADS Time 1 (26.3 vs. 18.4) and K-SADS Time 2 (24.3 vs. 7.7), respectively (Ambrosini et aI., 1991). The findings from this present study add biological data in support of this clinical distinction and syndromal validation vis-a-vis nondepressed psychiatric and normal controls. One adult report of IB identified patients into clinical diagnostic states approaching the classification used in this present study. Kanof et a!. (1987) studied 38 male major depressives, of whom 22 were actively depressed and 16 were a group of "remitted depressives" who met RDC criteria for MDD in the past but not at the study date. These investigators found no differences between these groups for either Bmax or I
Abstract. Kinetic constants of platelet imipramine binding were determined in youths with major depression and a contrast group. Subjects actively depressed (N = 10) had significantly fewer imipramine binding sites (B m",) (877 ± 148 fmol/mg protein) than recovering depressives (N = 12) (1220 ± 428 fmol/mg protein) and contrasts (N = 10) (1270 ± 230 fmol/mg protein). Affinity constants (~) (1.14 ± 0.36 nM, 0.97 ± 0.31 nM, and 1.17 ± 0.39 nM, respectively) were similar among the groups. Actively depressed males but not females had fewer imipramine binding sites than both their sex-matched comparison groups. Although actively depressed females' Bm", was significantly lower than recovering depressed and nondepressed males, neither age, Tannner stage, nor circannual rhythms influenced Bm"" but suicidality may be associated with low Bm",. A decrease in Bm", may be a state-specific marker of major depression in boys or associated with a depressive disorder with a suicidal history. J. Am. Acad. Child Adolesc. Psychiatry, 1992, 31, 2:298-305. Key Words: imipramine binding, depression, suicidality. Since Briley et aI. (1980) and Paul et aI. (1981) first reported a decreased number of platelet 3H-imipramine binding (lB) sites in depressed adults compared with normal controls, numerous investigators have replicated these findings, but these results are controversial (Bech et aI., 1988; Kanof et aI., 1987; Langer et aI., 1986; Nanki et aI., 1986; Nemeroff et aI., 1988; Plenge et aI., 1988; Schneider et aI., 1985). Studies in bipolar disorder are inconsistent when comparisons were made among euthymic, hypomanic, or depressed bipolar states (Baron et aI., 1986; Berrettini et aI., 1982; Lewis et aI., 1985a; Mellerup et aI., 1982; Meltzer et aI., 1990; Muscettola et aI., 1986). Besides methodological differences in the IE assay, several variables could account for these differences across studies. For example, the patient populations investigated are frequently a diagnostically mixed group of depressive subtypes at various stages of illness, treatment, or recovery. In some studies, IB was analyzed without a sufficiently long washout period from previous tricyclic or benzodiazepine treatment (Braddock et aI., 1984; Nutt et aI., 1985; Poirier et aI., 1985; Wagner et aI., 1987). Age ( Baron et aI., 1984; Langer et aI., 1980; Marazziti et aI., 1987; Schneider et aI., 1985) and seasonal variations may also influence IB (Arora et aI., 1988; Baron et aI., 1988; Egrise et aI., 1983; Galzin et aI., 1986; Goziotis et aI., 1988; Tang et aI., 1985; Whitaker et aI., 1984). Nevertheless, a body of evidence has emerged suggesting
Accepted July 17, 1991. From the Department of Psychiatry of the Medical College of Pennsylvania/Eastern Pennsylvania Psychiatric Institute, Philadelphia, Pennsylvania (Dr. Ambrosini) and the Division of Biological Psychiatry, Case Western Reserve University, Cleveland, Ohio (Drs. Metz, Arora, Meltzer, Ms. Lee, and Ms. Kregel). The authors would like to thank James Leckman, M.D., and Wagner Bridger, M. D., for their helpful comments in the preparation of this manuscript. This study was funded in part by NIMH grant MH41684 and presented in part at the 1988 ANCP meeting in Puerto Rico. Reprint requests to Dr. Ambrosini, MCPfEPPI, 3200 Henry Avenue, Philadelphia, PA 19129. 0890-8567/92/3102-0298$03.00/0© 1992 by the American Academy of Child and Adolescent Psychiatry.
298
that IB is abnormal in subsets of patients with affective disorder. Whether IB is an age dependent marker of subtypes of depression (Meltzer et aI., 1990), an age independent marker in familial pure depressive cohorts (Lewis et aI., 1985b), or is associated with specific symptom complexes (Carstens et aI., 1986; Meltzer et aI., 1986; Wagner et aI., 1987) awaits further investigation. It also is uncertain whether IB is a state or a trait marker of depressive illness because suggestive evidence exists for both positions, yet follow-up studies are marginally designed (Arora et aI., 1985; Baron et aI., 1987; Kanof et aI., 1987; Langer et aI., 1986; Maj et aI., 1988; Marazziti et aI., 1988a; Raisman et aI., 1981; Suranyi-Cadotte et aI., 1982; 1985; Wagner et aI., 1987). Platelet IE investigations in children and adolescents suggest that low IE occurs in enuresis (Weizman, A. et aI., 1985; Weizman, R. et aI., 1986a), obsessive-compulsive disorder (Weizman, A. et aI., 1986), and anorexia nervosa (Weizman, R. et aI., 1986b), but normal values exist in autism (Anderson et aI., 1984; Weizman, A. et aI., 1987) or attention deficit disorder (Weizman et aI., 1988). The two studies of IE in juvenile major affective disorder reported conflicting results. Rehavi et aI. (1984) did not find significant differences between depressed youngsters and psychiatric controls consisting exclusively of conduct disordered and schizophrenic subjects, whereas Carstens et aI. (1988) found significantly increased IE in the depressed group compared with normal controls. The reasons for these discrepant results are not clear, particularly given that both groups used a similar IE assay previously described by Paul et aI. (1981). The current study was designed to test the hypothesis that the number of platelet IE sites was low in children and adolescents in an active phase of unipolar major depression compared with a contrast group of nondepressed emotionally disordered patients or nonpsychiatric normal children. The preliminary findings suggest that platelet IE is significantly lower in the depressed group during an active episode of major depression and that IE may be a state-dependent marker of major depressive disorder (MDD) and/or associated with suicidality in this age group. J. Am. Acad. Child Adolesc. Psychiatry, 31:2, March 1992
IMIPRAMINE BINDING IN DEPRESSED YOUTHS
Patients and Methods The subjects studied were predominantly outpatients attending a specialty child depression clinic located at University Hospital of Cleveland. The clinic specialized in the assessment and treatment of depressed and anxious youngsters. Referrals came from both the general child outpatient services and from school and medical personnel who had been informed of the clinic's function by mailed brochures or direct presentation to professional groups. Individuals were excluded from the study if they were physically ill, severely psychotic, mentally retarded, or had symptoms clearly suggesting anorexia/bulimia nervosa, schizophrenia, or autism. Of the 32 subjects entered into this study, 30 were outpatients and two were adolescent inpatients. No subject received psychotropics nor was any ever treated with antidepressants before entry into the study. Diagnostic Assessment All subjects were screened initially for signs or symptoms of depressive or anxiety disorder during the first phone contact with the clinic's psychiatrists. Those callers who clearly had primary behavioral or exclusionary problems as described above were referred to the general child psychiatry outpatient service. The remainder of the contacts were scheduled for an extensive history gathering session with the parents. After this meeting, two sessions were scheduled within 24 to 72 hours for a diagnostic assessment with the third revised edition of the child version of the Schedule for Affective Disorders and Schizophrenia (K-SADSIIIR) (Ambrosini et al., 1989). The K-SADS-IIIR, which is an updated edition of the 1978 K-SADS, covers symptoms necessary not only for the affective syndromes but also for all the major anxiety and behavioral disorders first occurring in childhood and adolescence according to the third edition of the Diagnostic and Statistical Manual (DSM-III). Initially, the parent is interviewed alone about the child's symptomatology, and, in the second session, the child is given the same interview. A summary diagnosis is immediately generated from this information. The two senior authors achieved a K of 0.77 to 0.80 for the diagnosis of MDD in live test-retest (completed within 72 hours) or videotape reliability studies, respectively, with the K-SADS IIIR (Ambrosini et aI., 1989). Subjects were scheduled for a 2-week K-SADS in 10 to 14 days (i.e., K-SADS Time 2) if the summary K-SADS diagnoses generated at Time 1 met one of the following diagnostic patterns: (1) were actively in a episode of MDD; (2) were actively in an episode of a nondepressed anxiety disorder with or without oppositional disorder; (3) had experienced an MDD episode within the last 12 months, but did not meet diagnostic criteria for MDD in the last week preceding the assessment, yet may have met Research Diagnostic Criteria (RDC) (Spitzer et aI., 1978) for minor depression. All subjects included in this latter group were symptom free for less than 2 months. This second interview was done blind to the diagnostic status found at K-SADS Time 1. After the K-SADS Time 2, the following diagnostic groups emerged: (1) those persistently meeting criteria for J. Am. A cad. Child Adolesc. Psychiatry, 31:2, March 1992
a major depressive episode; these subjects were defined as being in an active phase of major depression (MDD-A); (2) those initially diagnosed MDD but now no longer meeting criteria for MDD or just meeting criteria for minor depression; (3) those initially diagnosed with a prior MDD, with or without a residual minor depression, who continued in the same diagnostic pattern; these latter two groups were defined as being in a recovering phase of their major depression (MDD-R) because all subjects still were not more than 2 months free of meeting Research Diagnostic Criteria for MDD; and (4) those persistently meeting diagnostic criteria for nondepressed anxiety syndromes formed the control group. No patients were psychotically depressed or met criteria for bipolar disorder. Of the two teenage inpatients, one male was in the MDD-A group and one female in the MDD-R cell. Four normal subjects whose parents were hospital staff members were also included in the nondepressed cell. These youngsters were assessed once with the epidemiological version of the K-SADS (Orvaschel, 1985) and found to have no lifetime history or current episode of psychopathology. Within the MDD-R group, one girl was not given a subsequent K-SADS but was followed clinically for several months and remained well. Among the nondepressed psychiatric controls, two girls with emotional and behavioral disorders were not seen for a subsequent K-SADS. Platelet Preparation and Binding Assay All blood samples were drawn between 8 A.M. and 9:30 A.M. before any pharmacotherapy. No subject had taken any psychotropic medication, although several subjects evenly distributed among the three study groups took analgesics within the last week. If they were taking over-the-counter preparations, these were discontinued at least 1 week before sampling. Thirty ml of whole blood was drawn with ethylenediamine tetraacetic acid as the anticoagulant. Platelet-rich plasma (PRP) was obtained by centrifugation, and platelet membranes were prepared by lysis and polytron homogenization (Langer et aI., 1980; Sahai et al., 1981). Platelet membranes were incubated with six concentrations of 3H-imipramine (sp. act., 52.6 Ci/mmole) between 0.36nM in the presence and absence of desipramine (100 l!M). Specific binding of the tritiated imipramine as calculated from Scatchard analysis of the data was defined as the difference in binding in the presence and absence of desipramine. Protein in the membrane preparation was determined following the method of Lowry et aI. (1951). The coefficients of variations for ~ and Bmax were 5.2% and 5.4% for split duplicates obtained from 15 normal controls assayed the same day and 7.5% and 8.2% at an interval of 3 to 4 weeks. Data Analysis Differences among independent subject groups were compared by computing X2 tests for categorical or proportional data and analysis of variance (ANOVA) for continuous data with post-hoc comparisons. The Mann-Whitney U test was used to test for significant differences between two independent groups with non-normally distributed data. The effects of sex and diagnostic status were analyzed using a twoway ANOVA; Tukey's multiple comparison procedure was
299
AMBROSINI ET AL.
TABLE 1. Demographic Statistics Major Depressive Disorder-Active (N = 10) Age (yr)a
Major Depressive Disorder-Recovering (N = 12)
15.5 :':: 1.7
13.3 :':: 3.1
Nondepressed Contrast (N = 10) 13.1 :':: 4.2
Sex (MIF)
4/6
5/7
5/5
PrepubertaVadolescent
1/9
6/6
4/6
4.5 :':: 1.0
3.2 :':: 1.9
3.2 :':: 1.6
49.4 :':: 14
43.5 :':: 17
43.4 :':: 15
Height (cm)a
166.2 :':: 12
150.3 :':: 14
156.5 :':: 24
Weight (kg)a
62.2 :':: 21
44.9 :':: 15
52.2 :':: 18
Tanner" SEsa,b
a b
Statistical Values
= 1.9 = 0.17 = 0.24, = 0.89 = 4.07, P = 0.13 F(2, 29) = 2.51 p = 0.10 F(2, 29) = 0.50 p = 0.61 F(2, 29) = 2.4 p = 0.11 F(2, 29) = 2.6 p = 0.09 F(2, 29)
p X2 p X2
Mean :':: SD. Hollingshead's two-factor socioeconomic status.
used for the between-group comparisons. Correlations were calculated using either a Pearson's or Spearman rho coefficient depending on the normality of the data distribution. Bonferroni inequalities were applied to decrease the risk of obtaining significant associations when multiple comparisons were made with the data set (Grove et aI., 1982). Results
Sample characteristics
Thirty-two subjects were accepted into the study. Demographic statistics for the three diagnostic groups are presented in Table 1. The trend for weight differences among the three groups is accounted for by one 15 year-8-monthold male in the MDD-A cell who measured 184.5 cm and weighed 104 kg and the fact that the actively depressed subjects were slightly older. Because Tanner Stage and frequency of prepubertal versus adolescent status was similar among the three groups, the data set was not analyzed separately for levels of sexual maturation. Parameters assessing depression severity and dysfunction are compared across the three experimental groups in Table 2. A reliable l7-item depression scale derived from the K-SADS (Ambrosini et aI., 1989), which is similar to a Hamilton Depression Rating Scale (Endicott et aI., 1981), was compared at three time points. These time points are the summary depression scores rated at the initial and 2week K-SADS. The Summary Present Episode (SPE[l]) score refers to the severity of depressive symptomatology during the worst period of the episode in the last year, whereas the Summary Last Week (SLW[l]) score is the depression severity rating during the last week before the first K-SADS interview. There is only a Summary Last Week rating at the second K-SADS (SLW[2]) because the evaluation at that time assessed depression severity just since the prior interview. As four normal contrast subjects were evaluated with the K-SADS-E, a l7-item depression score could not be generated; nor were such scores available 300
at SLW(2) for the subjects not seen for a 2-week K-SADS. The duration ofMDD and the age of onset were not applicable for the contrast group. Also the Global Assessment Scores (GAS) (Shaffer et aI., 1983) were not available for the four normal subjects assessed with the K-SADS-E. The MDD-A group was significantly more severely depressed at all three time frames compared with the MDD-R or contrast groups. The groups differed only in their depression severity and not in their level of functional impairment because the GAS scores were not different among the groups. Although duration of illness was absolutely longer in the MDD-R cohort, by definition the MDD-A subjects were still ill and so it is unclear how long their illness would be once they were in a recovering state. Imipramine Binding The maximal number of binding sites (B max ) and the dissociation constant (:KI) were similar in the normal control group and the nondepressed psychiatric controls (:KI = 1.39 ± 0.23 vs. 1.03 ± 0.43 nM; t = 1.52, p = NS; Bmax = 1320 ± 323 vs. 1236 ± 170 fmole/mg protein, t = 0.55, p = NS). When these parameters were compared in the three study cells, there was no difference in the :KI; however, the Bmax was significantly lower (p < 0.014) in the MDD-A cohort compared with the MDD-R and nondepressed contrast groups who had similar Bmax values. The mean Bmax value for the actively depressed sample was approximately 30% less than that of the recovering depressed subjects or those not depressed (Table 3). These differences are graphically displayed in Figure 1. Circannual influences were also analyzed among the three study groups by comparing the proportion of subjects within each diagnostic cell whose blood samples were obtained during each of the four seasons. Although the MDD-R cohort had a skewed distribution with 11 samples obtained in the spring and one during the winter, the proportions were more evenly distributed within both the MDD-A and contrast cases and were not significantly different (X 2 = 1.18, J. Am. Acad. Child Adolesc. Psychiatry, 31:2, March 1992
IMIPRAMINE BINDING IN DEPRESSED YOUTHS TABLE
2. Severity Parameters
Major Depressive Disorder-Active (N = 10)
Major Depressive Disorder-Recovering (N = 12)
Nondepressed Contrast (N = 6
81.1 ± 6.5
67.5 ± 8.5
56.7 ± 12.6
Depression Scale: 17 item SPE (1)0
Statistical Values
SLW (l)b
74.0 ± 9.8
54.4 ± 10.3
52.3 ± 11.9
SLW (2)'
71.0 ± 11.2
48.4 ± 7.5
58.5 ± 21.0 (N = 4)
= 15.0 = 0.0001 F(2, 25) = 12.1 p = 0.0002 F(2,22) = 9.78 p = 0.0009
55.0 ± 9.1
F(2, 25)
F(2, 25)
p
(N
Global Assessment Scale SPE (1)0
38.7 ± 16.4
=
11)
43.5 ± 19.3
= 0.80 = 0.46 F(2, 25) = 0.87 p = 0.43 F(2, 22) = 0.11 P = 0.89 Z = 0.09 p = 0.92 Z = 1.22 p = 0.22 p
SLW (l)b
42.8 ± 16.2
51.1 ± 23.2
57.3 ± 9.4
, SLW (2)'
54.1 ± 7.6
61.2 ± 22.8
47,0 ± 10.1 (N = 4)
(N
55.2 + 50 (4-168) 14.5 ± 1.2
Duration (wks) (range) Age at onset
11)
73 ± 98 (8-342) 11.9 ± 4.1
d
d
SPE(I) summary present episode, first interview. SLW(I) summary last week, first interview. c SLW(2) summary last week, second interview. d Not applicable. o
b
TABLE
3. Platelet Imipramine Binding Parameters in Child and Adolescent Major Depression
Major Depressive Disorder-Active (N = 10)
Major Depressive Disorder-Recovering (N = 12)
Nondepressed Contrast (N = 6
1.14 ± 0.36
0.97 ± 0.31
1.17 ± 0.39
877 ± 1480
1,220 ± 428
1,270 ± 230
Statistical Values
= 1.08 p = 0.35 F(2, 29) = 5.01 p < 0.014 F(2, 29)
Note: ~ = nM ± SD; Bmax = fmole/mg protein ± SD. o MDD-A < MDD-R (p < 0.02); Contrast (p < 0.001).
p = NS). Because a major finding is a B max difference be2000
• •
Smax
Kd
1600
• •
1400
0
•
1000 800
i
0
9 0
8
•
0 0
• • ••
0
1.6
•0
0
-
~
.0..
1200
•
0
1800
••
0
•
-i-
•
CD
0
•
~
9
MDD-A
MDD-R
1.4 1.2 1.0
••
0
0
0 0
1.8
.8 .6
• Contrast
=fmoles/mg
MDD-A
MDD-R
Smax
• Non-suicidal
MDD-A'Major depressive disorder-Active
protein
Kd
Contrast
=nM
o Suicidal
-R'Recovering
FIG. 1. Bmax and ~ in depressed and nondepressed children and adolescents. J. Am. Acad. Child Adolesc. Psychiatry, 31: 2, March 1992
tween these two latter groups, circannual rhythm disturbances between these diagnostic groups could not account for this result. Within-group correlations of K.i and B max with age, sex, Tanner stage, duration of disorder, age of onset, 17-item depression scale scores, and GAS were not significant after Bonferroni correction (Table 4). It was suggested that in the MDD-R cohort, an earlier age of onset was associated with a higher Bmax • However, when the analysis was controlled for age with an analysis of covariance, B max was still significantly different in the three groups (P < 0.015). B max in the MDD-A remained lower than in the MDD-R (p < 0.04) or in the contrast samples. Severity of depression in the last week preceding the blood sampling showed a positive association with B max in the MDD-A group. This suggested that individuals with a more severe depression had a higher number of binding sites so that low B max is not an index of depression severity. An exploratory group analysis of B max in the three study 301
AMBROSINI ET AL. TABLE
4. Within-group Correlations among Actively Depressed (MDD-A), Recovering Depressed (MDD-R), and Nondepressed Contrast MDD-A (N = 10)
MDD-R (N = 12)
Contrast (N = 10)
KJB"""
KJBrnax
KJB"""
-0.23/-0.51 -0.06/0.71" 0.40/-0.29 -0.25/-0.20 0.26/-0.50 -0.09/0.72"
0.43/-0.62" 0.22/-0.41 0.31/-0.53 -0.20/0.38" 0.28/-0.65 0.31/0.27
0.37/0.25 -0.03/-0.17 0.54/0.24
(N = II)
(N = 4)
0.09/-0.21
0.21/-0.21
0.86/0.32
0.13
Age Sexb Tanner Duration Age at onset 17-item depression scale (SLW[2]) GAS (SLW[2])
-0.17
(N
0.10
-0.98 a
= II)
(N
= 4)
" p < 0.05 (without Bonferroni). b Male = 0, female = 1. c
Not applicable.
TABLE
5. Imipramine Binding (B""",): Sex and Group Interactions (p values) Male MDD-A (N = 4)
Male MDD-A MDD-R
CaNT Female MDD-A MDD-R
CaNT
753 1461 1324
caNT (N = 5)
MDD-A (N = 6)
MDD-R
CaNT
(N = 7)
(N = 5)
0.001
0.006 0.45
0.26 0.007 0.04
0.11 0.02 0.11
0.02 0.18 0.55
0.58
0.15 0.32
960 1048 1215
groups was also conducted because in the MDD-A group female sex was correlated with high Bmax values. The sex by diagnostic status ANOVA showed a significant overall trend, F(5,26) = 3.21, p < 0.057 (Table 5). Actively depressed males and females had similarly low B max ; however, whereas the male MDD-A group differed from their two sex-matched comparison groups (but only from the MDD-R after Bonferroni correction), the female MDD-A group did not differ from the MDD-R and nondepressed contrast females. The MDD-A females, however, are significantly different than the MDD-R and the nondepressed males. These data suggest that decreased platelet IE sites may be more specific as a mark of illness in depressed boys rather than girls, or that boys attain a more rapid return to normal platelet IE kinetics. The IE data were also analyzed according to the suicide status of the patient. Individuals were defined as suicidal using K-SADS data if within the last year before either the first or second K-SADS, they acknowledged suicidal ideation, planning, gestures, or attempts. Subjects were thus classified dichotomously as suicidal or nonsuicidal or as attempters or nonattempters. Those with nonspecific wishes to be dead or not be born were not considered suicidal. Of the 10 MDD-A subjects, five were suicidal ideators, four were attempters, and one was nonsuicidal. Among the MDD-R group, three were ideators, two were attempters, and seven nonsuicidal. All contrast cases were nonsuicidal.
302
Female
MDD-R (N = 5)
The Bmax did not differ between the suicidal ideators and the attemptors (N = 9) within either the MDD-A or MDD-R groups. The ideators and attempters thus were combined and a comparison made among the following four cells: MDD-A suicidal (MDD-NS), MDD-R nonsuicidal (MDD-RlNS), MDD-R suicidal (MDD-RlS) , and controls. The number of IE sites was significantly lower in the MDDNS group compared with the MDD-RINS (p < 0.002) and contrast cells (p < 0.005). The MDD-RiS were significantly different from the MDD-RINS (p < 0.029) and showed a trend for a lower B max compared with the contrast group (p < 0.097). The B max values did not differ significantly between either the MDD-NS and MDD-RiS or between the MDDRlNS and controls. There was no difference among these four groups in :Ki values. Discussion The results of this study support the primary hypothesis that the number of platelet IE sites are decreased in youngsters during an active phase of MDD. The two previous studies of platelet IE in MDD in children and adolescents contradict each other, and both are at odds with these present findings. In Rehavi et al.' s study (1984), although a K-SADS assessment was administered twice over 2 weeks, the active depressives were a mixed bipolar and unipolar sample, whereas the controls were exclusively conduct disordered or schizophrenic. Prior reports suggest that these J. Am. Acad. Child Adolesc. Psychiatry, 31 :2, March 1992
IMIPRAMINE BINDING IN DEPRESSED YOUTHS
may be inappropriate comparison groups, because the latter two diagnostic groups may have low platelet IE (Arora et aI., 1986; Birmaher et aI., 1990; Stoff et al., 1987). Carstens et aI. (1988) found elevated IB in depressed juvenile depressives, compared with normal controls. However, whereas this group assessed their subjects with a reliable interview schedule (i.e., Interview Schedule for Children), only one assessment was completed, so it remains uncertain how many of these depressed adolescents were truly in an active phase of illness. Carsten's control sample was composed of normal school children, and this also may not be an appropriate group for biological validation studies because a comparison between pathological cases and normal controls reveals nothing about the specificity of a diagnosis with respect to other disorders (Werry et aI., 1987). The decision to diagnostically differentiate patients into those in active illness or in a recovering state was made before the collection of the data and was based on the clinical characteristics of the patient flow. The authors have found that approximately 44% of those who completed two K-SADS assessments had remission of their syndromal major depression between the two interviews. This clinical dichotomy received further validation as a distinct illness status in an expanded sample of 122 adolescents (which includes the adolescents in this present study) assessed and diagnosed by the authors' group in a similar manner, who also completed a Beck's Depression Inventory (BDI) (Beck et aI., 1961) before each K-SADS. Those adolescents defined as MDD at K-SADS Time 1, who were then also depressed at K-SADS Time 2 (i.e., MDD-A), had significantly higher mean BDI scores than the MDD-R at both KSADS Time 1 (26.3 vs. 18.4) and K-SADS Time 2 (24.3 vs. 7.7), respectively (Ambrosini et aI., 1991). The findings from this present study add biological data in support of this clinical distinction and syndromal validation vis-a-vis nondepressed psychiatric and normal controls. One adult report of IB identified patients into clinical diagnostic states approaching the classification used in this present study. Kanof et a!. (1987) studied 38 male major depressives, of whom 22 were actively depressed and 16 were a group of "remitted depressives" who met RDC criteria for MDD in the past but not at the study date. These investigators found no differences between these groups for either Bmax or I
