We read with great interest the article ‘Mood stabilisers and risk of stroke in bipolar disorder’ by Chen et al, published in this esteemed journal.^{Reference Chen, Tsai, Pan, Chang, Su and Chen1} The authors have assessed the association between mood stabilisers and risk of stroke in patients with bipolar disorder. They have addressed multiple potential confounders such as age, gender, physical illness and concomitant medication. However, we would like to discuss some important factors that could have influenced the study findings.

First, the study did not consider the role of psychiatric comorbidity. Only admission for bipolar disorder was taken into account. Certain psychiatric illnesses are common in patients with bipolar disorder and are noted to be risk factors for stroke as well. For example, the lifetime prevalence of anxiety disorder in bipolar disorder has been found to be 42.7%.^{Reference Nabavi, Mitchell and Nutt2} Further, anxiety disorders have been associated with a 24% increased risk of stroke compared with the general population.^{Reference Perez-Pinar, Ayerbe, González, Mathur, Foguet-Boreu and Ayis3}

Second, the role of psychosocial factors in stroke is not addressed. It is important to note that bipolar disorder is associated with several psychosocial factors and a recent meta-analysis mentions that the risk of stroke is increased by psychological, vocational and interpersonal factors by almost 39%, 35% and 16%, respectively.^{Reference Lightbody, Clegg, Patel, Lucas, Storey and Hackett4}

Third, bipolar disorder is associated with high rates of non-adherence to medication. As the study focuses on the association between mood stabiliser and bipolar disorder, the adherence to mood stabilisers is an important variable to be considered. This point is worth highlighting as the mean prevalence of non-adherence to medication has been found to range from 41.5% to 43%.^{Reference Chakrabarti5}

Fourth, the role of oral contraceptive pills has not been discussed. The increased risk of stroke with oral contraceptives is well-known. About 42.7% of participants are women. Oral contraceptives are used commonly by women for contraception as well as prescribed for polycystic ovarian syndrome, which is not uncommonly seen with use of valproate.

Fifth the comorbidity of seizure disorder has not been considered. This is important to address mainly for two reasons. Both valproate and carbamazepine are prescribed in patients with seizure disorder or epilepsy as well as in patients with bipolar disorder. Further, literature exists that late-onset epilepsy has been associated with increased risk of stroke.^{Reference Ben-Menachem6} This is a point to be noted as almost 61% of the participants were recruited when more than 45 years old.

Finally, the term ‘any mood stabiliser’ is not clearly defined. For example, the number of patients on any mood stabiliser during the case period is 212. However, the sum of patients as per numbers given separately for carbamazepine (35), valproic acid (118), lithium (62) and lamotrigine (18) is 233.