To the Editor:I read with interest the recent article by Lai et al on risk of acute pancreatitis after discharge from hospitalization for pyogenic liver abscess (PLA).[1] In this cohort study, the authors found that PLA was associated with an increased risk of acute pancreatitis compared to non-PLA (adjusted hazard ratio 3.00; 95% confidence interval, 2.62–3.43).[1] However, the estimate may not be valid because of a violated proportionality assumption, which was acknowledged by the authors.[1] To deal with non-proportional hazards, it would be more appropriate to use time partitioning or other methods for time-varying covariates,[2] instead of modeling using different follow-up periods (as shown in Table 2 of the article).[1] Moreover, the interaction effects of PLA and comorbidity on risk of acute pancreatitis (Table 3 of the article)[1] may differ by time-partitioned period.[2]