This article is devoted to a study of the impact of using bound sets in biobjective dynamic programming. This notion, introduced by Villareal and Karwan [1981], has been independently revisited by Ehrgott and Gandibleux [2007], as well as by Sourd and Spanjaard [2008]. The idea behind it is very general and can, therefore, be adapted to a wide range of biobjective combinatorial problems. We focus here on the biobjective binary knapsack problem. We show that using bound sets to perform a hybrid dynamic programming procedure embedded in a two-phase method [Ulungu and Teghem 1995] yields numerical results that outperform previous dynamic programming approaches to the problem, both in execution times and memory requirements.