OBJECTIVES: High blood pressure (BP) has been associated with increased risk of dementia. Concerns have been raised about lowering BP too far in the very elderly and thereby increasing risk. There is some evidence to suggest a potential 'J'-shaped relationship between DBP and risk of cognitive impairment. This was investigated using data from the HYpertension in the Very Elderly Trial (HYVET). METHODS: HYVET was a double-blind, placebo-controlled trial of antihypertensives in patients aged at least 80 years with an untreated SBP of 160-199 mmHg. Active medication was indapamide sustained release 1.5 mg+/- perindopril 2-4 mg to reach goal pressure of less than 150/80 mmHg. Incident dementia was a secondary endpoint and was not significantly different between the two treatment groups. The relationship between pressure and incident dementia was assessed using Cox proportional hazards regression with BP entered as either a discrete (quartile analysis) or continuous predictor variable. Achieved BP was calculated as the mean of all pressures from the 9 month visit onwards. RESULTS: During a mean follow-up of 2.2 years 263 incident cases of dementia were diagnosed. After adjustment for various covariates, baseline DBP was inversely related to incident dementia (P=0.0064). Achieved DBP did not predict later dementia in the placebo group (P=0.43), but showed a U-shaped relationship in the active treatment group (P=0.0195). The relationship between incident dementia and DBP did however not differ significantly between the placebo and active treatment groups (P=0.38). SBP was not associated with incident dementia, at baseline (P=0.62) or during follow-up (placebo group P=0.13, active group P=0.36). Wider achieved pulse pressure (PP) was associated with increased risk of dementia in both treatment groups (placebo P=0.032, active P=0.0046). The same tendency was observed for baseline PP (P=0.095). CONCLUSION: Wider PP may possibly indicate an increased risk for dementia. Active treatment may act to change the shape of the relationship between DBP and dementia. Future studies need to focus on exploring the ideal goal pressure for this age group.