The knowledge of the proportion of the different cell types in milk, rather than just the overall somatic cell count (SCC), could provide valuable information for a more precise definition of the health status of the udder. Recent developments in milk-testing technology have allowed cell differentiation in milk in a high throughput manner using flow cytometry. In the present study, we defined cut-off values of a novel indicator of udder health, i.e. the milk differential somatic cell count (DSCC), to identify subclinical mastitis. This novel trait represents the percentage of neutrophils plus lymphocytes in the total SCC. In addition, we investigated the genetic background of DSCC to gain knowledge on its potential inclusion in selection programmes aimed to improve mastitis resistance in dairy cattle. A total of 21,043 test-day records of 10,106 Italian Holstein cows from 406 herds were considered for ROC analysis, whereas a subset including 10,709 test-day records of 5142 cows from 299 herds was used for genetic analysis. The ROC procedure allowed the identification of the best DSCC thresholds (given by the highest combination of sensitivity and specificity) among all the possible cut-off values. Different thresholds were obtained for primiparous and multiparous cows, in early or mid-late lactation. In addition, our results revealed, for the first time, that DSCC is a heritable trait (h2=0.08 ± 0.02), and heritability was lower in primiparous than in multiparous cows. Furthermore, bivariate analysis demonstrated that DSCC and traditional somatic cell score (SCS) are two different traits as their genetic correlation was less than one. This study provided DSCC cut-off values for a practical application of this novel trait to identify subclinical mastitis and findings on the genetic aspects of DSCC are a prelude to its potential use in breeding programmes aimed to select against mastitis occurrence.