Background: Numerous studies have shown a stronger relationship between ambulatory blood pressure (ABP), particularly night ABP, and cardiovascular events/mortality than for office blood pressure (OBP). A previous clinical trial (Syst-Eur) showed that pretreatment ABP was only a better predictor of outcome than OBP in placebo-treated participants. The current study in treated elderly hypertensives from the Second Australian National Blood Pressure study (ANBP2) examined whether pretreatment ABP was a better predictor of mortality than OBP over long-term (∼11 years) follow-up. Participants and methods: ANBP2 was a comparative outcome trial in 6083 off-treatment or previously untreated elderly hypertensives. In the ABP substudy, at study entry, participants had ABP and nurse-performed OBP measurements. Cox proportional hazards analysis assessed the relationships between both OBP and ABP at study entry and 11-year all-cause and cardiovascular mortality, with results pooled from both active treatment phases. Results: In 702 participants, over a median of 10.8 years, including 6.7 years after the trial, 167 died (82 cardiovascular). Pretreatment 'night' systolic ABP and pulse pressure were the best predictors of '11-year' cardiovascular mortality (hazard ratios: 1.26; 95% confidence intervals: 1.10-1.45, P=0.001 and 1.18; 1.06-1.31, P=0.003, respectively) and all-cause mortality (hazard ratios: 1.15; 95% confidence intervals:1.05-1.28, P=0.005 and 1.09; 1.10-1.31, P=0.03, respectively). OBP was not a significant predictor of mortality. Conclusion: In actively treated elderly hypertensives participating in ANBP2, all-cause or cardiovascular deaths were significantly related to pretreatment ABP, particularly to night-time systolic ABP and pulse pressure, but not to OBP.