Abstract.We analyze the consequences of the hypothesis that the formation of the Broad Line Region is intrinsically connected to the existence of the cold accretion disk. We assume that the Broad Line Region radius is reliably estimated by the formula of Kaspi et al. (2000, ApJ, 533, 631). We consider three models of the disappearance of the inner disk that limit the existence of the Broad Line Region: (i) the classical ADAF approach, i.e. the inner hot flow develops whenever it can exist; (ii) the disk evaporation model of Meyer & Meyer-Hofmeister (2002, A&A, 392, L5); (iii) the generalized disk evaporation model of Różańska & Czerny (2000b, A&A, 360, 1170. For each of the models, we determine the minimum value of the Eddington ratio and the maximum value of the broad line widths as functions of the viscosity parameter α and the magnetic field parameter β. We compare the predicted parameter space with observations of several AGN. Weak dependence of the maximum value of the FWHM and minimum value of the Eddington ratio on the black hole mass in our sample is noticeable. It seems to favor the description of the cold disk/hot inner flow transition as in the classical ADAF approach rather than with the model of disk evaporation due to conduction between the disk and accreting corona.