Many contemporary ecosystems are likely to be affected by multiple environmental drivers, complicating efforts to predict future changes in those ecosystems. We studied long-term changes (1980–2012) in forest dynamics and liana (woody vine) abundance and biomass in fragmented and intact forests of the central Amazon. We did so by contrasting trends in 33 permanent 1-ha plots near forest edges (plot center <100 m from the nearest edge) with those in 36 1-ha plots in intact-forest interiors (150–3300 m from nearest edge). In fragmented and edge-affected forests, rates of tree (≥10 cm diameter at breast height) mortality and recruitment were often sharply elevated, especially in the first 10–15 years after fragmentation. Lianas (≥2 cm stem diameter) also increased markedly in abundance (mean ± SD = 1.78 ± 1.23% per yr) and biomass (1.30 ± 1.39% per yr) over time, especially in plots with high edge-related tree mortality. However, plots in undisturbed forest interiors, which were originally established as experimental controls, also experienced long-term changes. In these plots, tree mortality and recruitment rose significantly over time, as did liana abundance (1.00 ± 0.88% per yr) and biomass (0.32 ± 1.37% per yr). These changes were smaller in magnitude than those in fragments but were nonetheless concerted in nature and highly statistically significant. The causes of these changes in forest interiors are unknown, but are broadly consistent with those expected from rising atmospheric CO2 or regional climate drivers that influence forest dynamics. Hence, the dynamics of Amazonian forest fragments cannot be understood simply as a consequence of forest fragmentation. Rather, the changes we observed appear to arise from an interaction of fragmentation with one or more global- or regional-scale drivers affecting forest dynamics. Both sets of phenomena are evidently increasing forest dynamics and liana abundances in fragmented forests, changes that could reduce carbon storage and alter many aspects of forest ecology.