Information is the keystone to the participation of subjects in biomedical research. Clear comprehension of the informed consent documents (ICDs) is primordial and a necessary requirement is that they are readable. While submission of a protocol to a French 'Comités de Protection des Personnes' (CPP) is a mandatory step with regard to the French legislation on biomedical research, no published data are available concerning its influence on ICDs readability. The aim of our study was to determine the impact of French CPP on the readability of ICDs, using lexico-syntactic readability indexes and ICDs from four clinical research centres and one clinical research unit. Twenty-five ICDs were analysed. The Flesch score was not modified after CPP review, while the Cordial score was significantly lower [from 4 (1-14) to 1 (1-13), P = 0.014]. The information was longer and more complex following CPP review. No protocol characteristics had any impact on the variation before and after review for either the Flesch or the Cordial indexes, nor on the number of syllables per word. Changes in the total number of words before and after review varied considerably between study centre, supporting heterogeneity of CPP review. Since August 2004, French CPP have to study the intelligibility of ICDs in addition to the scientific and ethic aspects of a research. We show that their current reviews do not increase the readability, while increasing the length of ICDs.