I would like to congratulate Johnston & Eagles (Reference Johnston and Eagles1999) on the scale of their study and for providing a clearer idea regarding the prevalence of hypothyroidism related to lithium and the potential risk factors. I would like to comment on several important issues, including the presence of preexisting thyroid disease prior to initiation of lithium, the presence of antibodies and the possibility of diagnosis as a risk factor.

With regard to the presence of preexisting hypothyroidism (as suggested by patients receiving replacement thyroxine) the study excluded a total of 18 cases, which, given the total number of cases on thyroxine, is a substantial proportion of the sample. This would suggest that patients who have hypothyroidism are more likely to suffer from conditions that require treatment with lithium.

It is unfortunate that thyroid antibodies were not measured more frequently. The finding that 13 out of 15 patients with hypothyroidism were positive for antibodies would suggest that the prevalence of antibodies would have been high and would have clarified the role played by autoimmunity in contributing to hypothyroidism. Thyroid autoimmunity mediating the hypothyroid effect of lithium has been studied extensively and there is considerable evidence to support this. As early as 1973, Crowe et al (Reference Crowe, Lloyd and Bioch1973) suggested that two different types of hypothyroidism occur with lithium, one with evidence of underlying autoimmune hypothyroidism and one without, based on a review of cases reported. Studies by Lazarus et al (Reference Lazarus, John and Bennie1981) and Leroy et al (Reference Leroy, Villeneuve and Lajeunesse1988) suggest a high prevalence for antithyroid antibodies in patients who are hypothyroid on lithium, thus supporting the role of autoimmmunity mediating this effect. Indirect evidence that autoimmune factors may mediate the actions of lithium on the thyroid comes from cases of hyperthyroidism, a well-documented side-effect of lithium, that cannot be explained on the basis of a direct pharmacological effect of lithium on the thyroid.

The issue of a particular diagnosis being a potential risk factor for lithium-induced hypothyroidism has not been highlighted in the literature although it has been studied, albeit indirectly. It is reasonable to conclude from the literature that thyroid autoimmunity is increased in conditions in which lithium is likely to be prescribed (i.e. bipolar affective disorder and depressive disorders). A study by Lazarus et al (Reference Lazarus, McGregor and Ludgate1986), in which thyroid antibodies were investigated prior to the prescription of lithium, reported a prevalence of 43%. Importantly, the entire group had a diagnosis of bipolar affective disorder. This compares with only 8.6% in a study of unipolar depression (Reference JoffeJoffe, 1987). This would indicate that there is a case for studying the relationship between the psychiatric diagnosis, thyroid autoimmunity and the hypothyroid effect of lithium. This could answer the question raised by the authors as to why hypothyroid patients on lithium are selected to remain on both treatments. It is possible that lithium is more likely to be continued when the diagnosis is that of bipolar affective disorder than depression alone.

I hope that further studies in this area will help to dissect out the factors that play a role in lithium-induced hypothyroidism.