We are very pleased to hear of the interest of Griffiths and Morgan in our paper on antidepressant toxicity using the np-SAD database. It is intriguing that even though our paper was published earlier this year and prior to the Office for National Statistics report (Morgan et al, 2004), there is no mention in their letter of the similarities between the two studies. For example, the proportion of deaths from TCAs (amitriptyline and dothiepin being the most frequently implicated compounds) were very similar (85% and 89%), and both studies found that approximately 80% of deaths from antidepressants were suicides. Furthermore, one of the main implications of our study for clinical practice (and a result not previously reported in the UK) was the risk of SSRI-related fatality when these drugs are ingested in combination with TCAs, with or without other illicit drugs. Griffiths and Morgan appear erroneously to assume that np-SAD only collects information on illicit drugs. Rather to the contrary, the programme protocol, which is published bi-annually with surveillance reports (Reference Ghodse, Schifano and OyefesoGhodse et al, 2003), defines a case as one where psychoactive substances are directly implicated in the fatality, and this includes antidepressant-related deaths. Consequently, Griffiths and Morgan's findings de facto suggest that some of the ‘culpable’ antidepressants in their data-set might not have been prescribed. It is unlikely that the Office for National Statistics data-set will contain information on prescriptions written as this is often not required on death certificates.