Reply

Abstract

We appreciate Dr. Lai's interest and comments on our article. [1] In our current study, we demonstrated an association between aspirin use and the risk of hepatocellular carcinoma (HCC) using a Korean nationwide population-based cohort of patients with chronic hepatitis B. Immortal time bias, which can often mislead to an overestimation of the effect size, can be involved in most observational studies evaluating drug effects. [2] Immortal time refers to a follow-up period during which outcomes could not have occurred. [3,4] It typically occurs when there is a gap between cohort entry and an initial exposure to the drug, in which the span of time is inevitably event-free and thus "immortal." In our study, because the index date of the aspirin-treated group was set to the 180th day after the date of the first aspirin prescription, immortal time could not have intervened. If the index date of a patient who was prescribed aspirin for 90 days had been set to the first prescription date, those 90 days would correspond to an immortal time, during which the patient necessarily had to be eventfree. In Case A, we believe the period after the patient has been exposed to aspirin for 90 days, which Dr. Lai suggests to be an "immortal time," should be regarded only as an "unexposed period" because an event can occur at any time point during that period. In addition, in Case B, Dr. Lai considered the period between stopping and resuming aspirin an immortal time. Because an effective treatment duration of aspirin for HCC chemoprevention remains unclear, we adopted a selection criterion of 90 consecutive days of aspirin use to define aspirin users. [5] Discontinuation or resumption of aspirin after a sufficient exposure to aspirin (≥90 consecutive days) was not considered in the analysis. Thus, the period that Dr. Lai designated as immortal time in Case B cannot be called immoral time in the same context as aforementioned.