BACKGROUND: Rates of prenatal and postnatal depression in developing countries are high. Prolonged depression during the postnatal period is associated with impaired infant growth and development. Little is known about the factors predicting the persistence of prenatal depression beyond the first few postnatal months. METHODS: From a sample of 701 women in a rural sub-district of Pakistan, the Schedule for Clinical Assessment in Neuropsychiatry (SCAN) was used to identify those with ICD-10 depressive disorder in the third trimester of pregnancy (n=160). Depressed women were re-assessed at 3, 6 and 12 months postnatal. Persistently depressed women (depressed at all time points) were compared with the remainder. Psychiatric symptoms, disability and life events were measured using the Self-Reporting Questionnaire (SRQ), Brief Disability Questionnaire (BDQ), and a modified Life Events Checklist. RESULTS: Of 129 mothers who completed follow-up, 73 (56%) were depressed at all points of assessment. These persistently depressed mothers had higher SRQ and BDQ scores prenatally and had experienced more life events in the year preceding the third pregnancy trimester than the mothers whose depressive disorder resolved (none received treatment). Persistent depression was significantly associated also with poverty, having 5 or more children, an uneducated husband and lack of a confidant or friend. On multivariate analysis, higher SRQ score and poverty during pregnancy predicted persistent depression. LIMITATIONS: The sample was from one rural sub-district only. We did not assess the women for physical conditions such as anaemia and thyroid-deficiency. CONCLUSION: Women who are poor and have more psychological symptoms during pregnancy are more likely to remain depressed one year after giving birth. This study highlights the need for developing mechanisms of early identification and suitable psychosocial interventions to minimise the damaging effects of persistent postnatal depression in poor communities.