This paper investigates further on Quantum Key Distribution (QKD) protocols based on classical error-correcting codes, previously proposed by the authors. In [1], the classic channel coding techniques were shown to be suitable, in particular conditions, instead of quantum error-correcting codes, for correcting errors that occur during the exchange of a cryptographic key between two authorized users. To this end, systematic parallel concatenated codes, also known as turbo codes, were proposed in [1]. In this paper we address the properties that the systematic parallel concatenated codes must have in order to be successfully used in a QKD protocol. In particular, we address the problem, not considered in [1], that the constituent codes used in the parallel concatenation scheme must be such that the information sequence cannot be easily recovered from the encoder output sequences. In fact, the parallel classical channel on which to transmit only the redundancy bits is highly reliable but also public, i.e., very vulnerable as far as the attack attempts of a possible attacker are concerned.
