To the Editor:In their study, Desai and colleagues identified clinical predictors of reduced lung-transplant-free survival in patients with adult-diagnosed cystic fibrosis (1). Older age at diagnosis (hazard ratio, 1.24 for a 5-year higher age at diagnosis) was mentioned as an important risk factor, in addition to diabetes and lung function at baseline. This finding was based on a multivariable Cox regression, using years since diagnosis as the time variable. However, by choosing years since diagnosis, the finding for age at diagnosis might merely reflect the trivial fact that older patients have a higher risk of dying. Consider, for example, two patients (A and B) being diagnosed at 20 and 25 years, respectively, and having identical values for all other covariates considered in their multivariable model. Then, 10 years after diagnosis, patient B’s hazard ratio is 1.24 higher than patient A’s hazard ratio. But this is not so surprising, as patient B is at that moment 5 years older (patient A, aged 30 yr; patient B, aged 35 yr). Redoing the analysis with age instead of years since diagnosis can reveal whether or not labeling older age at diagnosis as a risk factor results from this possible artifact.We feel it is important to clarify this issue. If not, physicians may well communicate the wrong information to patients or be left with the thought that a diagnosis late into adulthood is much more worrisome than it in fact is.