Data from 350 herds enrolled in the Norwegian Dairy Herd Recording System (NDHRS) were used to investigate the associations between the first two cow-milk somatic cell counts (SCC) test-days’ results after calving or the three last SCC test-days prior to drying off in the first lactation and the hazard ratio (HR) of clinical mastitis (CM) during the remaining first or the subsequent second lactation respectively. Altogether, 9519 first lactations and 6046 second lactations were included. Cox regression analyses adjusted for herd frailty effect were used. In the first lactation, SCC>40000 cells/ml on the first or second test-day was significantly associated with an increased risk of a CM event in the remaining first lactation. HR, compared with 10000 cells/ml, increased from 1·6 (1·4) for SCC of 40000–60000 cells/ml to 6·9 (4·2) for SCC >800000 cells/ml, when using the first (second) SCC test-day in the first lactation. Cows with a geometric mean of the three last SCC test-days between 50000 and 100000 cells/ml and between 401000 and 800000 cells/ml in the first lactation had HR of CM during the second lactation of 1·3 and 2·8 respectively compared with a reference group of 10000–20000 cells/ml. If a CM episode in the first lactation occurred, the HR for having a CM event during the second lactation was 1·5. There was a significant frailty effect which disappeared if the incidence rate of CM at herd level was included in the model.