River widening, defined as a lateral expansion of the channel, is a critical process that maintains fluvial ecosystems and is part of the regular functioning of rivers. However, in areas with high population density, channel widening can cause damage during floods. Therefore, for effective flood risk management it is essential to identify river reaches where abrupt channel widening may occur. Despite numerous efforts to predict channel widening, most studies have been limited to single rivers and single flood events, which may not be representative of other conditions. Moreover, a multi-catchment scale approach that covers various settings and flood magnitudes has been lacking. In this study, we fill this gap by compiling a large database comprising 1564 river reaches in several mountain regions in Europe affected by floods of varying magnitudes in the last six decades. By applying a meta-analysis, we aimed to identify the types of floods responsible for more extensive widening, the river reach types where intense widening is more likely to occur, and the hydraulic and morphological variables that explain widening and can aid in predicting widening. Our analysis revealed seven groups of reaches with significantly different responses to floods regarding width ratios (i.e., the ratio between channel width after and before a flood). Among these groups, the river reaches located in the Mediterranean region and affected by extreme floods triggered by short and intense precipitation events showed significantly larger widening than other river reaches in other regions. Additionally, the meta-analysis confirmed valley confinement as a critical morphological variable that controls channel widening but showed that it is not the only controlling factor. We proposed new statistical models to identify river reaches prone to widening, estimate potential channel width after a flood, and compute upper bound width ratios. These findings can inform flood hazard evaluations and the design of mitigation measures.