An important approach to the design of antiallergic agents with reduced penetration into the central nervous system (CNS) is described. A series of 3-[(5,11- dihydro[1]benzoxepino[4,3-b]-pyridin-11-ylidene)piperidino]propion ic acid derivatives (31-47) and related compounds (48-54) were synthesized and evaluated for antiallergic activity and penetration of a compound into the CNS in comparison with the corresponding 6H-dibenz[b,e]oxepin derivative (3). Combination of zwitterionization and introduction of a pyridine component resulted in an increase in antiallergic activity and a great reduction of penetration into the CNS, which was evaluated by the selectivity (B/A) of antihistaminic activities in the central system [ID50 value (B) for ex vivo H1 binding to mouse brain membranes] and in the peripheral system [ED50 value (A) for inhibitory effect on histamine-induced increase in vascular permeability in mice]. This surprising reduction of penetration into the CNS could be considered on the basis of an increase in hydrophilicity caused by both of the zwitterionization and the introduction of a pyridine component. 3-[4-(8-Fluoro-5,11-dihydro[1]benzoxepino[4,3- b]pyridin-11-ylidene)piperidino]propionic acid (33) exhibited a strong antiallergic effect in various experimental models and very low penetration into the CNS. Compound 33 (HSR-609) is now under clinical trial as a promising antiallergic agent with greatly reduced penetration into the CNS.