Many articles tried to address the question of the " optimal " urban form regarding sustainable mobility (Newman & Kenworthy, 1999 ; a, 2008). Conclusions however remain controversial for very large metropolitan regions given complexity of multiscale organization of the metropolitan area (Schwanen et al., 2001 ; Banister, 2008 ; Le Néchet & Aguilera, 2012a) and the diversity of their associated mobility patterns, partly due to the inertia of both urban form and mobility habits (Klinger, 2016). In particular, the choice between individual (car, bike, walking) and collective (bus, train, carpool) forms of transport depend not only on the supply side (Kenworthy & Laube, 1996), but also on cultural and individual preference aspects (Donald et al., 2014). In the context of increasingly polycentric metropolis, we emphasize the importance of taking into account spatial organisation of flows to explain mode choice. This is important because commuting flows that start and end outside city centers are now an overwhelming majority are the most difficult to channel through transit. Moreover, with the emergence of Mega City Regions (Hall & Pain, 2006), exchange between activity cores might be an increasing important dimension to assess the sustainability of the global region. Hence, if the level of functional integration is weaker than classical " Daily Urban System " , the commuting distance resulting from the proximity of several cities is high and typically not addressed by the planners (Conti, 2015). We analyse harmonized commuting data in Paris and Rhine-Ruhr regions two metropolitan regions very similar in size (both 12,000 km 2) and population (both 12 M inh.) but with contrasted spatial and functional organisation. Paris is primarily monocentric, and the Rhine-Ruhr region polycentric. We showed a paradox that establishing a better self-containment of labour zones could lead to less transportation pooling and in the end a less sustainable mobility system, in so far as medium and long-range flows might not be effectively channelled via heavy transportation infrastructures. There is therefore a serious challenge in Paris, where polycentricity is an emerging phenomenon (Berroir & al., 2008), to maintain public transit's efficiency, hence/that is to ensure that journeys to employment centers are not undertaken, as is currently often the case, exclusively by car; in other words, as they grow, these flows could/should be channelled by effective public transport infrastructures.