Methodological considerations for smoking cessation, weight change, and psoriasis

We read with great interest the nationwide cohort study by Kim et al.,1 which examined the joint association of smoking cessation and post-cessation weight change with psoriasis and its subtypes. The authors should be commended for addressing a clinically relevant question, as concern about weight gain remains a common barrier to quitting smoking. Their findings are consistent with broader epidemiological and genetic evidence linking smoking to psoriasis risk and support the public health message that smoking cessation should be encouraged.2, 3 Nevertheless, several methodological issues deserve further clarification, and additional analyses could strengthen the causal interpretation of this important study. First, the analyses do not appear to have fully accounted for the pre-cessation smoking burden. The adjusted Cox models included demographic, metabolic and lifestyle covariates, but not cigarettes smoked per day, smoking duration or pack-years. This is important because successful quitters and sustained smokers may differ substantially before cessation. Individuals who quit may have had lower cumulative tobacco exposure, lighter nicotine dependence or greater health consciousness than those who continued smoking. Therefore, the lower risk observed among quitters, particularly for palmoplantar pustulosis, may partly reflect residual confounding by prior smoking burden rather than the independent effect of cessation itself. Additional adjustment for baseline smoking intensity, smoking duration or pack-years, together with stratified analyses by cumulative smoking burden, would help clarify whether the observed associations persist across different levels of tobacco exposure. Second, the temporal interpretation of “post-cessation weight gain” requires caution. Smoking cessation was inferred from smoking at the first health examination in 2004–2005 and non-smoking at the second examination in 2006–2007, while weight change was defined by the change in BMI between the same two examinations. Because the exact date of cessation was unavailable, BMI gain during this interval cannot necessarily be interpreted as weight gain occurring after cessation. Some participants may have gained weight before quitting, whereas others may have changed weight for reasons unrelated to cessation. Although the supplementary analyses using stable smoking status and washout periods are helpful, they do not fully resolve the chronology of cessation and weight change within the exposure-assessment window. A more conservative interpretation would be to describe the exposure as interval BMI change among quitters. Analytically, future work could emulate a clearer target trial by defining cessation first and assessing subsequent BMI change thereafter, using repeated health-screening data, time-updated exposure models or landmark analyses.4 Third, death was treated as a censoring event in the Cox models. In a long-term cohort, this assumption may be consequential because smoking is strongly associated with mortality, and death precludes subsequent clinical diagnosis of psoriasis. If sustained smokers die earlier, they may have less opportunity to receive the repeated documented consultations required for case ascertainment. We suggest that the authors present cumulative incidence functions and conduct competing-risk analyses, such as Fine–Gray subdistribution models, to evaluate whether the reported associations remain robust when death is treated as a competing event rather than non-informative censoring.5 These considerations do not weaken the central recommendation that dermatologists should encourage smoking cessation. Rather, they suggest that the findings would be more clinically and causally informative if prior smoking burden, within-window exposure timing and competing mortality were more explicitly addressed. Such refinements would help distinguish the benefits of smoking cessation itself from differences in cumulative tobacco exposure, interval weight change and differential survival during follow-up. This work received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. The authors declare no conflicts of interest. Not applicable. Not applicable. Data sharing is not applicable to this article as no new data were generated or analysed.