The limiting of the reporting of maternal deaths to those that are included in the criteria of the World Health Organization excludes deaths which yield useful information for further improvements in clinical performance. In this series of 22 maternal deaths, six deaths would have been excluded from reporting: one "direct" obstetric death of pre-eclampsia; one "indirect" death as a result of renal and cardiac failure; two deaths as a result of postnatal depression which led to suicide three and four months postpartum, respectively; and two deaths of cancers, where diagnostic delay may have been a result of the coexistent pregnancy. The importance of primary pulmonary hypertension, cardiomyopathy and psychiatric illness is emphasized. We endorse the recent recommendation of the International Federation of Gynaecology and Obstetrics (FIGO) that all maternal deaths that occur more than 42 days after the end of a pregnancy should be assessed for possible relationships with childbirth, and suggest that a time limit of one year would include all deaths that are worthy of scrutiny.