We are pleased that Adetunji et al read our paper and saw fit to comment upon it. However, we are surprised at the nature of their remarks, which suggest that not only did they not read our piece with particular care, they have perhaps also not thoroughly read (or perhaps understood) the paper they quote by Smith et al. ^{Reference Smith, Cornelius, Warnock, Bell and Young1} One of us (A.H.Y.) is the senior and corresponding author on this meta-analysis and thus very familiar with the content! Close perusal of our meta-analysis does not show our case for lithium to be one-sided, unbalanced and misleading. Indeed, we conclude ‘mood stabilisers have differing profiles of efficacy and tolerability’ and demonstrate that lithium has clear evidence of both tolerability and efficacy. Nowhere do we suggest that lithium is the best treatment for every patient with bipolar disorder, nor is the purpose of the article to review the evidence for all bipolar medications. Rather, as we state in our conclusion, our argument is that lithium remains the best treatment in a significant portion of cases and must be included in any psychiatrist's treatment arsenal. The reason this message is so important is that lithium is increasingly being neglected as a treatment option in several countries, resulting in inadequate treatment of some patients with the disorder, who are then labelled ‘treatment-resistant’ without having ever tried lithium. ^{Reference Jefferson2}

Prescribing patterns are influenced by pharmaceutical company promotion – or why would companies spend this money? ^{Reference Huang, Weiss, Fenimore, Fleming, Haller, Lichtmacher and Eisendrath3} Jefferson ^{Reference Jefferson4} and Chan ^{Reference Chan5} both report declining psychiatry resident knowledge about, and use of, older medications (including lithium) despite evidence supporting their continued use; we are unclear why Adetunji et al find this literature ‘patronising’. We agree with them, however, that psychiatrists should base treatment choices on individual patient characteristics as well as the profile of medicines. Applying this approach to the wide array of available agents will undoubtedly ensure that lithium continues to be widely used for the foreseeable future.