In the context of pattern classification with two-fold reject options (ambiguity and distance), we have identified different strategies leading to two-stage classifiers. Strategies differ on managing both types of rejection (independently or not). This paper addresses the problem of combining the first stage of rejection-based classifiers of each strategy. A Dempster-Shafer's model is designed. Its main characteristic lies in basic probabilities assignments to reject classes before using the combination rule. After combination, a decision rule is proposed for classifying or rejecting patterns. We emphasize that such rejection is not related to a lack of consensus between the classifiers but to the initial reject options. In case of ambiguity rejection, a class-selective approach has been used. As illustration, results obtained on artificial data are given.