Diagnostic criteria for prolonged grief have appeared in the Diagnostic and Statistical Manual of Mental Disorders ( DSM-5; persistent complex bereavement disorder, PCBD) and in the ICD-11 (prolonged grief disorder, PGD), and the question of which diagnosis is most clinically useful has been hotly debated. This study provides the first longitudinal comparison of PCBD and PGD in their ability to capture symptom change over time and their relation to long-term outcomes. A community sample was recruited consisting of 282 individuals who had recently lost a spouse. Structured clinical interviews were conducted at 3, 14, and 25 months postloss for symptoms corresponding to PCBD and PGD criteria. Outcomes at 25 months included PCBD and PGD caseness, depression, global functioning, and interviewer ratings of participant suffering. PCBD and PGD trajectories determined by growth mixture modeling, each captured three primary outcomes: resilience, moderate-improving symptoms, and prolonged-stable symptoms. The PGD solution also identified trajectories of increasing and decreasing distress: prolonged-worsening and acute-recovering symptoms. Prediction of 25-month outcomes indicated differences conforming to the severity of PGD symptoms, and the prolonged-worsening trajectory was associated with the worst adjustment. PGD symptoms were more differentiated, better-captured psychopathology, and other outcomes and were more sensitive to change over time compared to PCBD. © 2019 Wiley Periodicals, Inc.