IntroductionVariable D-dimer trends during hospitalization reportedly result in distinct in-hospital mortality. In this multinational case series from the first and second waves, we show the universality of such D-dimer trends.MethodsWe reviewed 405 patients with COVID-19 during the first wave admitted to three institutions in the United States, Italy, and Colombia, and 111 patients admitted to the U.S. site during the second wave and 55 patients during the third wave. D-dimer was serially followed during hospitalization.ResultsDuring the first wave, 66 (15%) patients had a persistently-low pattern, 33 (8%) had early-peaking, 70 (16%) had mid-peaking, 94 (22%) had fluctuating, 30 (7%) had late-peaking, and 112 (26%) had a persistently-high pattern. During the second and third waves, similar patterns were observed. D-dimer patterns were significantly different in terms of in-hospital mortality similarly in all waves. Patterns were then classified into low-risk patterns (persistently-low and early-peaking), where no deaths were observed in both waves, high-risk patterns (mid-peaking and fluctuating), and malignant patterns (late-peaking and persistently-high). Overall, D-dimer trends were associated with an increased risk for in-hospital mortality in the first wave (overall: HR: 1.73) and stayed the same during the second (HR: 1.67, p < 0.001) and the third (HR: 4.4, p = 0.001) waves.ConclusionD-dimer behavior during COVID-19 hospitalization yielded universal categories with distinct mortality risks that persisted throughout all studied waves of infection. Monitoring D-dimer behavior may be useful in the management of these patients.