The bioavailability of propranolol from a matrix-in-cylinder system for sustained drug delivery, consisting of a hot-melt extruded ethylcellulose pipe surrounding a drug-containing HPMC-Gelucire 44/14 core, was determined. An oral dose of 80 mg propranolol hydrochloride was administered to healthy volunteers (n = 10) in a randomized cross-over study design either as a commercial pellet formulation (Inderal retard mitis) or as a matrix-in-cylinder system. The influence of concomitant food intake on drug release from the matrix-in-cylinder system was also studied. During the first 10 h after administration, the matrix-in-cylinder system resulted in similar plasma levels as the reference formulation Inderal. The concomitant intake of a high-fat, high-calorie breakfast did not cause dose-dumping. Between 10 h and 24 h after administration of the matrix-in-cylinder system, a remarkable increase of the propranolol plasma levels was noticed (compared to Inderal). This effect was even more pronounced under fed conditions. The matrix-in-cylinder system had a relative bioavailability of 156% (fasted conditions) and 222% (fed conditions) compared to the marketed reference product. In order to elucidate the origin of this increased bioavailability, Caco-2 experiments and dog lymph studies were performed. However, none of these experiments was able to provide a conclusive answer.