In this report we discuss the insecurity with present implementations of the Ekert protocol for quantum-key distribution based on the Wigner Inequality. We propose a modified version of this inequality which guarantees safe quantum-key distribution. QKD offers the possibility that two remote parties, conventionally called Alice and Bob, exchange a secret random key to implement a secure encryption-decryption algorithm, without meeting [1,2,3]. QKD provides a significant advantage over the public-key cryptography because the security of the distributed key relies on the laws of quantum physics [1, 2, 3], i.e. the wave packet collapse prohibits gaining information from a quantum channel without disturbing it. Indeed, any attempt by a third party (Eve) to obtain information about the key is detected.Two main goals underly the implementation of QKD schemes. One is to create and preserve authentic quantum channels against decoherence effects induced by any interaction with the environment [4], eventually reducing or even destroying invulnerability of quantum channels against Eve's attack. The other is to provide a true guarantee of absolute security against any possible eavesdropping attack i.e. the security is not simply based on technological feasibility.