The aim of the study was to evaluate the justification of Croatian modification of Bethesda classification after thirteen years of its application, answering the question if the subclassification of high-grade squamous intraepithelial lesion (HSIL) into cervical intraepithelial lesion (CIN) grade 2 (CIN2) and grade 3 (CIN3) is cytologically real and clinically justified. The retrospective study included 3110 women to whom cervical intraepithelial lesion of different grade was diagnosed by cytological examination of vaginal-cervical-endocervical (VCE) smear at Department of Clinical Cytology, Clinical hospital Osijek in period from 1993 to 2005. 57.1% of women were monitored cytologically and colposcopically, while 42.9% of them had also pathohistological examination. The spontaneous regression of cytological finding was noted in 66.3% of the cases. Moderate dysplasia regressed more often (50.98%) than severe dysplasia (31.3%) and more rarely than mild dysplasia (70.1%), which was statistically significant (p 0.05) in their biological behaviour and histological finding. In fact, 50.9% of moderate dysplasia spontaneously regressed, 14.4% persisted during follow-up, and 59.7% had a histological finding milder than CIN3. Therefore, in almost 65% of moderate dysplasia lesions it is not justified to apply the same diagnostic therapeutic procedures as for severe lesions, which means that cytological subclassification of HSIL into moderate dysplasia and severe dysplasia lesions is clinically justified. Positive predictive value of differential cytological diagnoses mild, moderate and severe dysplasia calculated in relation to histological CIN3+ statistically significantly increases for every single diagnosis, which also confirms that moderate dysplasia can be individual diagnostic category, thus the subclassification of HSIL is cytologically possible.