Deuterium at high redshift: primordial or evolved ?

On the basis of arguments from galactic chemical evolution we suggest that the recent observations of D/H vs. metallicity in several high redshift absorbers are best understood if the primordial D value is in the range 2-3 10$^{-5}$. This range points to a rather high baryonic density ($\Omega_Bh^2$=0.019-0.026) compatible to the one obtained by recent estimates based on the Cosmic Microwave Background (CMB) anisotropy measurements. Slightly higher values (D/H=4 10$^{-5}$) are found in Lyman limit systems. Such values are still compatible with CMB estimates but, if taken at face value, they suggest a trend of decreasing D abundance with metallicity. We argue that special assumptions, like differential enrichment, are required to explain the data in that case. A clear test of such a differential enrichment would be an excess of products of low mass stars like C and/or N in those systems, but currently available data of N/Si in DLAs do not favour such a ``non-standard'' senario.