Both hyper- and hypokalemia increase cardiovascular risk. Modest hyperkalemia is common with angiotensin-converting enzyme inhibition. We studied post-hoc the association of an initial, on-treatment serum potassium measurement with subsequent cardiovascular outcomes over 4.5 years in 9297 individuals at high cardiovascular risk, randomized to an ACE inhibitor or to placebo. Post-hoc analysis of cardiovascular outcomes, as related to serum potassium levels, in the HOPE (Heart Outcomes and Prevention Evaluation) study which compared ramipril to placebo, and included 692 patients with a serum potassium level >5.0 mM and 137 with a serum potassium level 0.4, respectively), with hypokalemia, the primary event rate was higher (22.6% vs 15.5%, respectively, p = 0.023). The hazard ratio for the primary outcome associated with this initial hypokalemia was 1.44 (1.00-2.06) on multivariate analysis. The combined primary outcome (myocardial infarction, cardiovascular death, stroke) was not different throughout deciles of serum potassium but the lowest and highest deciles included many with normokalemia. Randomized treatment was withheld because of hyperkalemia in 8 and 6 people allocated to ramipril and placebo, respectively. The benefit of ramipril on cardiovascular outcomes was independent of serum potassium, but ramipril reduced hypokalemia in the entire cohort (1.15 vs 1.86% with placebo, p = 0.005), particularly in those participants on diuretics (3.8% vs 6.5%, p = 0.07). In patients at high cardiovascular risk, modest hypokalemia predicts a less favorable outcome while modest hyperkalemia does not. Ramipril reduces hypokalemia and decreases risk.