BACKGROUND: Inhibitors of histone deacetylases (HDAC) have shown promise as targeted cancer therapy. Valproate, an older anticonvulsant, has been shown to possess HDAC inhibitory activity. We undertook this case-control study to clarify whether long-term users of valproate had a reduced cancer incidence. If so, it would support HDAC inhibition as a pharmacologic principle in chemoprevention. METHODS: We identified 149,417 incident cancer cases in Denmark during the study period 2000 through 2005, and 597,668 age- and gender-matched controls. Data on history of cancer, past hospital admission diagnoses, and prescription history were obtained from the Danish Cancer Registry, the Danish National Patient Registry, and the Danish Prescription Registry. Primary exposure to valproate was defined as a cumulative dose of minimum 1,500 g within the past 5 years. Confounders were controlled by conditional logistic regression. RESULTS: Among the cases and controls, 81 (0.05%) and 260 (0.04%), respectively, were long-term users of valproate. For cancer overall, the crude and adjusted odds ratios were 1.25 [95% confidence interval (95% CI), 0.97-1.60] and 1.21 (95% CI, 0.95-1.56), respectively. Subgroup analyses revealed no dose or duration effect for overall cancer incidence, and no specific cancer site was found to be inversely associated with long-term use of valproate. For lung cancer, we found a positive but imprecise association (adjusted odds ratio, 2.32; 95% CI, 1.12-4.79). CONCLUSION: Long-term valproate use is not associated with a reduced cancer risk. Our study does not support HDAC inhibition as a pharmacologic principle for general chemoprevention.