Voice over IP (VoIP) is a key enabling technology for the migration of circuit-switched PSTN architectures to packet-based networks. The problem of spam in VoIP networks has to be solved in real time compared to e-mail systems. Many of the techniques devised for e-mail spam detection rely upon content analysis and in the case of VoIP it is too late to analyze the media after picking up the receiver. So we need to stop the spam calls before the telephone rings. From our observation, when it comes to receiving or rejecting a voice call people use social meaning of trust and reputation of the calling party. We describe a multi-stage spam filter based on trust, and reputation for detecting the spam. In particular we used closed loop feedback between different stages in deciding if the incoming call is a spam or not. For verifying the concepts, we used a laboratory setup of several thousand soft-phones and a commercial grade proxy server. We verified our filtering mechanisms by simulating the spam calls and measured the accuracy of the filter. Results show that multistage feedback loop fares better than any single stage. Also, the larger the network size, the harder to detect a spam call. Further work includes understanding the behavior of different controlling parameters in trust and reputation calculations and deriving meaningful relationships between them.