THURSDAY, MAY 18 BIOL PSYCHIATRY 595 1995;37:593-683 8. PREDICTING ONSET AND COURSE OF ADOLESCENT DEPRESSION WITH SLEEP AND CORTISOL MEASURES U. Rao, R.E. Dahl, N. D. Ryan, B. Birmaher, D. E. Williamson, & R. Rao University of Pittsburgh School of Medicine, Western Psychiatric Institute and Clinic, Pittsburgh, PA 15101 This study examined EEG sleep and baseline plasma cortisol predictors of differential outcome in adulthood in 26 adolescent subjects with unipolar major depressive disorder (MDD) and 33 normal controls. The subjects (15.4 + 1.3 years) initially participated in sleep and neuroendocrine stud- ies. Clinical course was assessed for the interval period 6-8 years later using "blind" standardized assessments. The MDD sample was divided by clinical outcome into those with no further episodes (n = 6), those with unipolar recurrent course (n - 15), and those who switched to bipolar dis- order (n = 5). The controls were separated into those with no lifetime psy- chiatric disorder (n - 23) and those who developed depression (n = 7). Subjects with bipolar course were separated, and the remaining groups were compared on initial sleep and cortisol parameters as potential predic- tors of clinical course. MDD subjects with recurrent unipolar course had reduced REM latency (76.8 + 31.2 vs. 104.1 _+48.2, p _< 0.07), increased REM density (1.4 5:_ 0.2 vs. 1.1 + 0.3, p <_ 0.01) and elevated cortisol around sleep onset (10.8 +_4.7 vs. 6.7 +_2.7, p _<0.01) compared to con- trols with no disorder. Controls who developed depression had reduced REM latency (67.3 +_20.2 vs. 104.1 + 48.2,p <_ 0.08) and increased REM density (1.5 5:0.3 vs. 1.1 +_0.3, p < 0.005) compared to controls with no disorder. MDD subjects with no further episodes were similar to MDD subjects with recurrent unipolar course on REM latency (81.0 5: 45.4) and REM density (1.3 5: 0.4); however, cortisol around sleep onset discrimi- nated unipolar subjects with no further episodes from those with recurrent unipolar course (7.3 + 5.3 vs. 10.8 + 4.7, p _<0.06). EEG sleep and ba- seline cortisol may be helpful in predicting the differential clinical course of adolescent MDD. Reduced REM latency and increased REM density may predict a new onset of depression. Elevated cortisol around sleep onset may predict recurrences in adolescents with unipolar MDD. 9. PHARMACOLOGICAL ALGORITHMS FOR TREATMENT RESISTANT DEPRESSION IN CHILDREN AND ADOLESCENTS N. Alessi University of Michigan, Child and Adolescent Psychiatry, Ann Arbor, MI 48109 Treatment-resistant depression inw)lves up to 40% of adults with major de- pression. Medication trials, their duration, and augmentation strategies in- volving adults have been discussed at length. The prevalence and nature of treatment-resistant depression in children and adolescents is just beginning to be appreciated. For children and adolescents, the same experience does not exist in either the identification or treatment of treatment-resistant de- pression. A review of current treatment and follow-up studies would sug- gest that there exists such a population, although its frequency is study de- pendent. This paper will deal with the topic of treatment-resistant depres- sion in children and adolescence, their identification, and treatment. Postu- lated algorithms for the selection of medications and augmentation strate- gies will be presented and compared to those offered in the treatment of adults with treatment-resistant depression. Case examples and case series utilizing a wide range of strategies will be presented and critiqued. Particu- lar emphasis will be placed on the identification of comorbid conditions as a means to determine medication selection among this group. 10. ELECTRO-CONVULSIVE TREATMENT (ECT) IN REFRACTORY ADOLESCENT DEPRESSION N. Ghaziuddin, C. King, M. Naylor, M. Ghaziuddin, N. Chaudhary, & J. Greden University of Michigan, Ann Arbor, MI 48109-0390 Infrequent use of ECT in adolescents with affective disorders indicates absence of objective data to establish its efficacy and safety. Because of lack of well established efficacious pharmacological treatments, ECT is an option which should be considered. We studied 7 hospitalized adoles- cent females with a primary DSM Ill-R diagnosis of affective disorder (MDD = 5; bipolar, depressed - 1; organic mood disorder - 1), and mean age 16 years (SD = 1.2). All subjects had failed to respond to three or more adequate trials of antidepressant pharmacotherapy. After thorough diag- nostic evaluation including diagnostic interview, neuropsychological evaluation, neurology consultation, and 24-hour observation, subjects were offered treatment with ECT. Consent from parent/legal guardian and assent of the adolescent were obtained. ECT treatment was commenced with nondominant right electrode placement (six subjects) or bilateral (one subject with organic mood disorder). Due to lack of adequate re- sponse, electrode placement was changed to bilateral for three subjects. Significant improvement was noted on Children Depression Rating Scores, t - 2.83, p < 0.05, and the Global Assessment of Functioning Scale, t = -7.17, p < 0.05. Assessment of cognitive functioning using the Mini Mental State Examination showed no significant decline in cogni- tive functioning. Side effects noted were headache (100%), prolonged sei- zures (75%), after recovery seizures (14%). Minor side effects were com- mon. After recovery, seizures were rare but are a potentially serious com- plication. We conclude that ECT can be an effective treatment for de- pressed adolescents who have failed to respond to antidepressant pharma- cotherapy. 11. MELATONIN AFFECTS THE PROCESSING OF [3-AMYLOID PRECURSOR PROTEIN IN CELL LINES D.K. Lahiri, W. Song, & J.I. Numberger, Jr. Laboratory of Molecular Neurogenetics, Institute of Psychiatric Research, Program of Medical Neurobiology, Department of Psychiatry, Indiana University School of Medicine, Indianapolis, IN 46202 Alzheimer disease (AD) patients display irregularities in the pattern of normal circadian rhythms. In addition to the hypothalamus, Alzheimer- type pathology has been reported in the superchiasmatic nuclei, the site of the circadian oscillator. The disruption of circadian regulation by brain grafts that overexpress amyloid beta-peptide (A[~P) has been reported. We examined the relationship between the pineal hormone melatonin, which is secreted in mammals during the dark phase of the circadian cycle, and the processing of beta-amyloid precursor protein ([3APP). AD is charac- terized by depositions of A[~P in the brain in the form of extracellular plaques and cerebrovascular amyloid. AI3P (Mr.-4 kDa) is derived from a family of large [~APP molecules (Mr.-110 kDa) which are integral membrane glycoproteins. Soluble derivatives of AI3P and lAPP lacking the cytoplasmic tail, transmembrane domain, and a small portion of the extracellular domain are generated by [3APP secretases. Using cell cul- tures, we analyzed the level of [3APP in glial, fibroblast, pheochromocy- toma, and neuroblastoma cells by immunoblotting cell lysates and condi- tioned media. Normal levels of secretion of soluble 13APPderivatives by cells into conditioned media were severely inhibited by treating cells with