We read with interest the important and clinically relevant study by Krishnadas et al. ^{Reference Krishnadas, Jauhar, Telfer, Shivashankar and McCreadie1} Patients with severe nicotine dependence had greater scores on the positive subscale of the Positive and Negative Syndrome Scale (PANSS) and patients with mild– moderate dependence had greater scores on the PANSS negative subscale compared with non-smokers. As rightly pointed out by the authors, this finding is in contrast to a previous study ^{Reference Kelly and McCreadie2} carried out in the same area and to other similar studies. ^{Reference Smith, Singh, Infante, Khandat and Kloos3} The reason appears to be the use of the Fagerström Test for Nicotine Dependence (FTND), which has limited psychometric properties in patients with schizophrenia. The authors argue for the validity of the FTND in schizophrenia, by citing the article by Weiberger et al. ^{Reference Weinberger, Reutenauer, Allen, Termine, Vessicchio and Sacco4} Notwithstanding the methodological superiority of the study in terms of presence of a proper control group, the difficulty of using FTND in people with schizophrenia cannot be denied on a pragmatic basis. It has been widely accepted over the past few decades that dependence is a more holistic concept and cannot be attributed only to the amount or duration of smoking. However, Steinberg et al ^{Reference Steinberg, Williams, Steinberg, Krejci and Ziedonis5} has questioned the relevance of the items of the FTND – such as time to first smoking, difficulty abstaining in forbidden places and frequency of smoking in the first hours after waking up – by means of a factor analysis study in patients with schizophrenia. In fact, modification of the FTND for the serious mentally ill population has been suggested in view of individuals' frequent impairment in judgement and insight. Such a scale was also useful in Krishnadas et al's study because all the patients were residents of supported accommodation and there was lack of any objective assessment of nicotine use. Moreover, the emphasis on the amount smoked even in a cross-sectional study like this would have better helped to verify the authors' statement that ‘those with severe dependence have successfully overcome negative symptoms by increasing their level of nicotine dependence’ (pp. 309–310), although a longitudinal study is essential in settling this issue. We advocate the concept of pack-years in this regard.

The authors adjusted the results for many covariants but left out several important variables which may act as important confounders, such as use of smokeless nicotine, other substance use, presence of physical disorders, type of antipsychotics and other psychotropic medications. They have also not mentioned whether the consent from participants was taken or not. The fact that daily dose of medication was greater in the severely dependent group raises the possibility of a pharmacokinetic interaction or indicates the presence of a poor prognosis subtype with neuro-biological underpinnings, which should be clarified in future studies. In Krishnadas et al's study, the majority of patients were smoking to relax, to socialise better or to alleviate their loneliness, anxiety and depressive symptoms. This makes a strong case for a holistic treatment approach, rather than just prescribing antipsychotic medication, as many of the mentioned attributing factors can be addressed with a multimodal treatment approach.