Achieving consensus in a network is one of the most important performance bottlenecks in distributed computing. This paper takes a look at the existing protocols for achieving Byzantine Reliable Broadcast on asynchronous partially connected networks and how these protocols change to leverage the fact that some nodes have access to Trusted Execution Environments. Modeling some nodes to be completely trusted improves the throughput and reduces latency but the impact changes heavily depending on the placement of these nodes. The second, more realistic approach is having all processes use a local trusted subsystem implemented in a TEE. We show that this reduces the upper bound of faulty nodes from f<N/3 to f<N/2 and reducing the amount of messages sent by up to 64\% (N=30, f=5).