Same answer as Sahib's. In addition, spontaneous or enzymatic removal (by phosphatases) of the 5' triphosphate would also preclude further elongation of the poly(deoxy)nucleotide. The situation would have been different if polymerization had involved (deoxy)nucleotide 3' triphosphates, but to my knowledge, these don't exist in Nature.
The fault is in the dNTP. Deoxy-nucleotides have the P-P-P group at C5’ and OH at C3’. The union of an additional dNTP requires the energy from removing P-P and using that energy for the linkage of the resulting dNMP to the only possible place, the C3’ of the polynucleotide. If in other planets nucleotides had P-P-P at the C3’, the growth of the chain should be from 3’ to 5’.