In this paper, the use of continuous time charge amplification and shaping is proposed for processing the signal delivered by CMOS pixel sensors in charged particle tracking applications. Such a choice aims at exploiting the large scale of integration of modern deep submicron CMOS technologies to incorporate into the design of a single device both the potential for thin detector fabrication, inherent in the concept of monolithic active pixel sensors (MAPS), and the data sparsification capabilities featured by hybrid pixels. With respect to classical MAPS, adoption of the above readout method involves a substantial change in the guidelines for the design of the front-end electronics and of the whole elementary cell, in order not to jeopardize the collection efficiency of the sensitive electrode. For the purpose of supporting the feasibility of the proposed solution, the paper discusses some experimental data and simulation results relevant to monolithic CMOS sensor prototypes, fabricated in a 0.13 mum technology, which were designed according to the mentioned rules. Finally, the performances of an all NMOSFET charge preamplifier, suitable for improving charge collection efficiency, are investigated through circuit simulations