How do structure and function in European old-growth forests compare to other temperate regions? Are old-growth characteristics shared universally or are there regional differences reflecting variation in growth, stand dynamics, and disturbance history? We tested the hypothesis that important ecological functions are provided universally by old-growth temperate forests, such as high carbon storage and co-varying habitat complexity. We pooled site-specific (n =501) forest inventory datasets representative of eight ecoregions: U.S. Pacific Northwest; U.S. Northeast; U.S. Mid-Atlantic; Southern Europe; Central Europe; Southern Chile; Northeastern China; and Temperate Australia. We used NMDS to analyze similarity/dissimilarity in stand structure among and within the ecoregions, broadly defined forest types, by mature vs. old-growth age class. Dimensionality of the data was assessed using the Monte Carlo permutation procedure with 50 runs of randomized data. Post-hoc tests employed bi-plots and Kendall’s Tau. The results reveal a wide range of structural variability within and among late-successional temperate forest systems. There is structural dissimilarity between mature and old-growth, but also a very high degree of overlap among both age classes and sub-regions. Ordination results did show distinct inter-regional variation in old-growth forest structure, with the Carpathians and Alps most similar to the U.S. Northeast and least similar to the U.S. Pacific Northwest and the southern hemisphere. When old-growth is assessed by forest type, a clear gradient of structural complexity emerges. Evergreen broadleaf and needleleaf forests occupy one end of this spectrum, mixed conifer-deciduous systems are intermediate, and deciduous broadleaved forests exhibit lower relative values for structural indicators. Live aboveground biomass, downed coarse woody debris, and tree density are the top ranked indicators explaining directionality in the data both for inter and intra-regional variability. Compared to other regions, European forests exhibit greater contrast between age classes for mean basal area, downed coarse woody debris, and aboveground biomass (live and dead), but have similar variability around means as that exhibited globally. As is true globally and for Europe specifically, conservation of late-successional forests provides both carbon storage benefits and structurally complex habitats. Yet the results also show tremendous variability within and among systems. In some cases, structural characteristics typically associated with old-growth may be equally if not more pronounced in mature forests, depending on site characteristics and disturbance history. This suggests a need to broaden our conception of late-successional dynamics, and to move away from overly narrow or rigid classification criteria not accommodating of this range of variability. Primary forest mapping needs to account for variability in stand development processes.