The problem considered in this paper consists of planning and routing rail-truck intermodal shipments from a set of suppliers to a set of clients. The delivered quantities have to follow a truck-rail-truck intermodal path. There are three constraints to satisfy: the satisfaction of the demand is mandatory, each request should be delivered by its due date, and the capacity of a train cannot be exceeded. The objective function to minimize is a weighted summation of the cost and the risk. The former component concerns the fixed and variable transportation costs associated with trucks and trains, while the latter associated with dangerous goods is non-linear for trains. We propose a graph model to represent this problem and a tabu search metaheuristic to tackle it. Numerical experiments on realistic data are provided and discussed.