Smoking can affect several disease outcomes and it is an important risk factor for morbidity and mortality. Lee and colleagues use the IRIS registry database to examine the relationship between smoking and intraocular pressure (IOP). The sample size included in the study is enormous, and the challenges typically encountered in the analysis of big data were well accounted for by the authors.
Real-world datasets, such as the IRIS registry, provide opportunities to examine hypotheses with unprecedented statistical power.1 Although the age distribution among the three groups by smoking status in the study of Lee et al. was different, with a big sample size of more than 12 million patients, the study successfully showed the difference in IOP among the three groups in each of age subgroups from 20 to 85 years in both glaucoma and non-glaucoma patients. Without a big dataset, such a difference across age subgroups would not be easily detected. The results suggested that IOP varied by smoking status, with the highest IOP in current smokers, in patients aged between 30 and 60 years. It is, however, important to note that the magnitude of difference in IOP between the three groups was small and less than 1 mmHg. Although highly statistically significant, it may have limited clinical implications.
In addition to examining the difference in clinical outcomes by smoking status, it is vital to measure lifetime smoking exposure. For example, results from the NHANES study found that current smokers had a lower odds of glaucoma compared to non-smokers (OR = 0.61) and ex-smokers (OR = 0.46).2 Nevertheless, among the smokers, greater pack/day smoking was associated with higher odds of glaucoma (OR = 1.70). However, information on smoking burden was not available in the IRIS registry.
Another limitation of the study is its inability to account for ethnicity and corneal biochemical properties in the analysis
Another limitation of the study is its inability to account for ethnicity and corneal biochemical properties in the analysis. Results from another big dataset, the UK Biobank, showed that regular smokers had 0.19 mmHg higher IOP on average, compared to non-smokers.3 However, the direction of association shifted when corneal-compensated IOP was used for analysis; regular smokers had 0.35 mmHg lower IOP, compared to non-smokers. In addition, in the Singapore Epidemiology of Eye Diseases Study which included Malays, Indians and Chinese, current smokers had 0.47 mmHg and 0.26 mmHg lower IOP, respectively, before and after adjustment for potential confounders.4 Further research is needed to evaluate whether smoking has differential effects on IOP among ethnic groups.