Evidence suggests that inflammation may be associated with increased risk of ovarian cancer but there is paucity of studies investigating this association, especially using over-time changes in inflammatory biomarkers. We conducted a prospective population-based case–control study nested within the Finnish Maternity Cohort (FMC). Within the FMC, 170 women with ovarian cancer who had donated serum samples to the cohort twice, ≥1 year apart, before cancer diagnoses were identified. One control per case was matched for age, parity and sampling date. Comparing the highest with lowest tertiles, the odds ratio (OR) of ovarian cancer using the first set of serum samples (mean lag time to cancer diagnosis 9.0 years) was 1.62 95% confidence interval (CI) 0.93–2.83. However, analysis conducted using the second set of serum samples donated closer to cancer diagnosis (mean lag time 6.4 years) revealed a significantly increased risk of ovarian cancer comparing extreme tertiles of C-reactive protein (CRP) concentrations; OR 1.96 (95% CI 1.11–3.4). Over time, increases in individuals' CRP concentrations were also associated with increased risk; OR 1.90 (95% CI 1.12–3.23). The results suggest that inflammation may precede ovarian cancer since increasing CRP concentrations, both across tertiles and longitudinally at the individual level, were associated with increased risk.