News · Brain & Mental Health
Blood pressure that swings tracked with faster memory loss
Pooling 11,104 participants from two major blood pressure trials, researchers in Alzheimer's & Dementia found that how much systolic pressure moved between visits tracked with cognitive decline, independently of the average.
Based on a peer-reviewed pooled analysis of two randomized trials in Alzheimer's & Dementia
- 11,104 participants pooled from the ACCORD-MIND and SPRINT-MIND blood pressure trials.
- Each 10% rise in visit-to-visit variability tracked with faster annual cognitive decline.
- The effect held after adjusting for mean systolic pressure, so it is not just high blood pressure.
- It showed up in the intensive treatment arms but not the standard ones, though that split was not confirmed.
- Abstract only, and the treatment-arm pattern is exploratory by the authors' own description.
Two people can walk out of a clinic with the same number written on the chart and have arrived there very differently. One has sat at 130 for a decade. The other has been swinging between 110 and 150 and happens to average 130. Standard care treats them as the same patient.
A pooled analysis in the journal Alzheimer’s & Dementia suggests they may not be. Higher systolic blood pressure variability was associated with accelerated cognitive decline independently of mean blood pressure.
The word doing the work there is independently. Everyone already knew that sustained high pressure is bad for the brain. The claim here is that the wobble carries its own signal on top of that.
What counts as variability, and why the average hides it
Blood pressure is the force of your blood pushing against the walls of your arteries, and it is highest when your heart beats, pumping the blood. That peak is the systolic number, the one people quote.
It is never a fixed quantity. Clinics already know this, which is why they will take two or more readings at separate appointments before making a diagnosis. The usual purpose of repeat readings is to find a truer average.
This study inverted that. Instead of averaging the readings away, it treated the spread between them as the exposure, using a measure the authors call systolic BP variation independent of the mean. The construction matters: it is built so that a person with wildly unstable pressure does not simply score high because their average is high.
What the pooled trials found
The researchers did not run a new study. They pooled individual-participant data from two of the largest blood pressure trials ever run in this area, ACCORD-MIND and SPRINT-MIND, giving 11,104 participants. Everyone included had baseline and follow-up cognitive testing and at least three blood pressure measurements from three months onwards.
The outcome was a processing-speed test, scored as annual change.
Each 10% increment in variability was independently associated with faster annual cognitive decline, and the association survived adjustment including mean systolic pressure.
The size is worth stating plainly, because it is small. This is a slope on a test score, measured across thousands of people over years. It is the kind of effect that shows up in a population and would be invisible inside any one person’s life.
The treatment split the authors will not lean on
The most provocative result is also the least secure. Associations were observed in the intensive but not standard BP treatment arms of both trials, appearing among people whose pressure was being pushed hardest toward a low target.
There is an obvious story to tell about that, and the authors decline to tell it. The pooled interaction was not firm enough to lean on, which means the apparent gap between the two arms could reasonably be noise.
Their own summary is careful: the exploratory treatment-context pattern warrants prospective confirmation. That phrasing is an instruction to hold the finding open.
It bears repeating that both of those trials tested intensive treatment and found it worthwhile. Nothing in this analysis overturns that, and nobody should read a reason to loosen blood pressure control into it.
What a pooled analysis of test scores cannot settle
The outcome was cognitive test performance, not a dementia diagnosis. Slower scores on a timed task and a clinical diagnosis are related but not interchangeable.
Both source trials also enrolled selected populations, one in type 2 diabetes and one in hypertension at elevated cardiovascular risk. People recruited into a trial differ from the general population by design.
Reading the abstract alone, as here, means the authors’ full account of their own limitations is not available to quote, so these caveats are the conservative ones rather than the complete set.
And the direction of travel is unproven. An analysis of this shape can show that unstable pressure and declining scores travel together after adjustment. It cannot show which one moved first.
Why it is worth watching
When your blood pressure stays high over time, it causes the heart to pump harder and work overtime, and the damage that follows is well mapped. Variability is a different proposition: a property of the shape a series makes, invisible on any single reading, and currently attached to no treatment target at all.
If it holds up, the useful shift is in what gets looked at. A row of numbers in a file is currently read for its average. This suggests the shape of the row might matter too.
People also ask
What is visit-to-visit blood pressure variability?
It is how much your reading moves between separate appointments, rather than how high it is on any one of them. Someone averaging 130 could reach that by sitting at 130 every time, or by bouncing between 110 and 150. Conventional care looks at the average and treats those two people identically. This study measured the bouncing instead, using a statistic designed so that the result does not simply track the average.
How big was the effect?
Small per unit and measured on a test score rather than a diagnosis. Each 10% increment in variability was associated with faster annual decline of 0.008 standardized points a year (95% CI, -0.014 to -0.003). That is a group-level slope, not something an individual would notice from one year to the next, and nobody here was diagnosed with dementia as the outcome.
Why would swings matter if the average is fine?
The study does not establish a mechanism, so anything here is a hypothesis. The usual proposal is that the brain's small vessels cope poorly with repeated surges and drops, and that unstable perfusion damages tissue in ways a steady pressure would not. That reasoning is plausible and untested by this analysis.
Does this mean intensive blood pressure treatment is harmful?
No, and it would be a serious misreading. The association appeared in the intensive arms and not the standard ones, but the interaction between them was not statistically significant, and the authors call the pattern exploratory and in need of prospective confirmation. Both trials found intensive treatment beneficial overall. Nobody should change blood pressure medication on the strength of this.
Can I find out my own variability?
Only roughly, and it is not a clinical measure. Any record of repeated readings over time would show whether yours move a lot, and home monitors make that easy to collect. But there is no established threshold for what counts as too much variability, and no treatment aimed at reducing it. Discuss any blood pressure concern with your doctor.
References
- Qiao, Y., Sun, Z., Ji, X., Markus, H. S., Zhao, W. Visit-to-visit systolic blood pressure variability, blood pressure treatment intensity, and cognitive decline. Alzheimer's & Dementia, 2026.
- MedlinePlus. High Blood Pressure. US National Library of Medicine.
- National Institute of Neurological Disorders and Stroke. Vascular Dementia.