Abstract There are unresolved questions regarding the association between persistent leukocytosis and risk of thrombosis and disease evolution in polycythemia vera (PV), as much of the published literature on the topic does not appropriately use repeated-measures data or time-dependent modeling to answer these questions. To address this knowledge gap, we analyzed a retrospective database of 520 PV patients seen at 10 academic institutions across the United States. Taking hematologic laboratory data at ∼3-month intervals (or as available) for all patients for duration of follow-up, we used group-based trajectory modeling to identify latent clusters of patients who follow distinct trajectories with regard to their leukocyte, hematocrit, and platelet counts over time. We then tested the association between trajectory membership and hazard of 2 major outcomes: thrombosis and disease evolution to myelofibrosis, myelodysplastic syndrome, or acute myeloid leukemia. Controlling for relevant covariates, we found that persistently elevated leukocyte trajectories were not associated with the hazard of a thrombotic event (P = .4163), but were significantly associated with increased hazard of disease evolution in an ascending stepwise manner (overall P = .0002). In addition, we found that neither hematocrit nor platelet count was significantly associated with the hazard of thrombosis or disease evolution.