Potassium and blood pressure: separate the trials from the stroke cohorts
A systematic review found a blood-pressure effect in hypertension without proving an event benefit through randomized trials.

Illustrative food image, not a photograph of the study intervention.
Two kinds of evidence entered the review
This review included 22 randomized controlled trials with 1,606 adults and 11 cohort studies with 127,038 participants. The trials reported outcomes including blood pressure, renal function, and blood lipids; the cohorts examined mortality and cardiovascular disease outcomes. That division is essential to interpretation. A pooled paper can contain randomized evidence for one endpoint and observational evidence for another. It should not be described as a single experiment proving that increasing potassium prevents strokes. The strongest design label must stay attached to the particular outcome it actually supports.
Aburto NJ, Hanson S, Gutierrez H, Hooper L, et al.. BMJ (Clinical research ed.), 2013. PMID 23558164 · DOI
Blood-pressure benefit depended on the population
Across adults, higher potassium intake reduced systolic pressure by 3.49 millimeters of mercury and diastolic pressure by 1.96 millimeters of mercury. The reported confidence intervals for the reductions were 1.82 to 5.15 and 0.86 to 3.06, respectively. The effect appeared in people with hypertension but not in those without hypertension. This population distinction prevents a universal promise that everyone will experience the pooled decrease. The result is an average across included trials, not a measured response to one potassium-containing meal or an individualized target for a reader.
Aburto NJ, Hanson S, Gutierrez H, Hooper L, et al.. BMJ (Clinical research ed.), 2013. PMID 23558164 · DOI
Stroke was not the same kind of finding
Higher potassium intake was associated with lower incident stroke risk, with a relative risk of 0.76 and a 95% confidence interval from 0.66 to 0.89. Associations with incident cardiovascular disease overall and coronary heart disease were not statistically significant. Those event findings came from cohort evidence rather than the blood-pressure trials. It would therefore overstate the review to claim that randomized potassium assignment prevented strokes or reduced every cardiovascular endpoint. The evidence supports keeping the trial-based blood-pressure result and the observational stroke association in separate sentences with separate design labels.
Aburto NJ, Hanson S, Gutierrez H, Hooper L, et al.. BMJ (Clinical research ed.), 2013. PMID 23558164 · DOI
Do not extend the safety conclusion universally
The abstract’s adult conclusion applies to people without impaired renal handling of potassium. Its report of no significant adverse renal effect in the included evidence is not a safety certificate for every patient or a rationale for unsupervised supplementation. A larger blood-pressure decrease appeared within one higher-intake category, but the review did not find a dose-response relationship. Neither that subgroup nor the overall average establishes a personal potassium prescription. The linked bean-and-sweet-potato meal is food context, not a verified dose or substitute for reviewing kidney status, supplements, and medication with a clinician.
Aburto NJ, Hanson S, Gutierrez H, Hooper L, et al.. BMJ (Clinical research ed.), 2013. PMID 23558164 · DOI