624 CORRESPONDENCE CALEV, A. (1988) Effect of emotion on the rate of forgetting in normals, schizophrenics and depressives. Paper presented at the First International Conference on Individual Differences, Tel Aviv, Israel. DUNBAR, G. C. & LI5HMAN, W. A. (1984) Depression, recognition memory and hedonic tone: a signal detection analysis. British Journal ofPsychiatry. 144, 376—382. Hout, C. & CASTLE, R. (1966)ContentAnalysisof Dreams.New York: Appleton-Century-Crofts. TEASDALE, J. D. & Dn@ii, J. (1987) Cognitive vulnerability to depression: an investigation oftwo hypotheses. British Journal of Psychiatry, 26, 113—126. Expressed emotion and lithium prophylaxis SIR: In a previous study (Priebe; Journal, March 1989, 154, 396—399), expressed emotion (EE) in key relatives of 21 patients with bipolar affective or schizoaffective psychoses was assessed by the Camberwell Family Interview (CFI). All patients had been on prophylactic lithium for at least three years. Patients living with high-EE relatives showed a significantly poorer response during the three years before interview and particularly during the nine month follow-up. Twenty-eight months after the initial CFI, the key relatives of 15 patients were re-interviewed. All 15 patients had continued on prophylactic lithium throughout the 28 months. Two critical remarks designated high EE. There were 10 relatives ident ified with high EE in the first interview and eight in the second, since two relatives changed from high- to low-EE status. The course of patients' illness was assessed by means of a morbidity index (Coppen et al, 1973) reflecting severity and length of recurrences (a recurrence was defined by hospital admission or a temporary additional antidepressive or neuroleptic medication). Regardless of whether EE status was defined according to first or second CFI, morbidity indices concerning the 28-month period were more than six times higher in patients living with high-EE relatives (regarding the first CFI: t=2.9l, P<0.05; regarding the second CFI: t = 3.83, P< 0.01). Four out of five patients with consistently low-EE relatives, and one patient with a relative who had changed from high to low-EE status were virtually without any recur rences during the 28-month period. This applied to none of the patients living with consistently high-EE relatives. A clear answer as to whether high-EE status of relatives leads to an unfavourable course of illness or vice versa was not found. Both course of illness and relatives' EE status might be influenced by inter actional patterns in the patients' families, and by changes in those patterns. As far as this small and highly selective sample is concerned, it may be con cluded that patients living with consistently low EE relatives rarely need therapeutic interventions of whatever kind in addition to prophylactic lithium. Department of Social Psychiatry Freie UniversitâtBerlin Platanenallee 19 D 1000 Berlin 19 Department of Clinical Psychiatry Freie UniversitätBerlin Reference S. PRIEBE C. WILDGRUBE COPPEN,H., PEST,M., BAILEY,J. etal(l973)Double-blind and open prospective studies of lithium prophylaxis in affective disorders. Psychiatry, NeurologyandNeurosurgery, 76, 501—510. Pseudocyesis followed by depressive psychosis SIR: Pseudocyesis is a false beliefin pregnancy associ ated with its symptoms and signs (Cohen, 1982). It may be associated with psychiatric disorder, most commonly depression, which usually does not reach case level (Brown & Barglow, 1971). Occasionally, it is associated with psychosis (Taylor & Kreeger, 1987; Mortimer & Banberry, 1988; Milner & Hayes, 1990). We would like to report the first case of depressive psychosis following pseudocyesis after an interlude of normality. Case report: A 38-year-old married domestic worker of stablepre-morbid personality had children of ages20, 14 and 10years.Shehad had one miscarriageat age35 years and one electiveabortion when 37 yearsold. Six months after this abortion shewas referred to the psychiatric ser vicesbecauseshebelievedherself to be pregnant,despite two negativepregnancytests.Shehadmissedtwo periods, had back-ache,breast swellingandtendernessandmorning nausea.Examination revealeda distendedabdomen but non-pregnantcervix and uterus.After threeweeksof sup portive psychotherapy all symptoms resolved. She remainedcompletelywell for threeweeksbut then became depressed, with early morning wakening,reducedenergy, poor appetite and weight loss. Shewas severelyagitated, believingherselfto be in danger.Shewasdeludedthat her facewasbeingdistorted and that her body wasrotting. She made three suicide attempts. Her symptoms failed to re spondto antidepressantsandphenothiazinesbut resolved rapidly with a short courseof electroconvulsivetherapy. She returned to her pre-morbid functioning and has remainedwell for two yearsoff all medication. Our patient's illness satisfied DSM—III—R criteria for a major depressive episode with psychotic features and ICD—9 guidelines for manic-depressive psy chosis, depressed type. To our knowledge, this is the first report of a depressive psychosis following