Pairing a statin with certain blood pressure drugs tracked 12% less dementia over 12 years

A very large number of people over sixty take two pills every morning: one for cholesterol and one for blood pressure. Which blood pressure pill, out of the six or seven classes available, is generally decided on how well it works and what it does to the kidneys.

Researchers writing in the International Journal of Geriatric Psychiatry, using a New South Wales cohort, asked whether that choice also tracks something else. Among 34,610 matched participants taking both drugs reliably, one pairing went with less dementia across a mean 12.4 years, by about 12%.

The pairing that came out ahead is not the one most people would pick from first principles.

Two ways to lower a blood pressure

High blood pressure, also called hypertension, is when blood puts too much pressure against the walls of your arteries, and almost half of American adults have it.

The medicines used against it work in different ways to lower blood pressure. Several of them act on the same hormone system, and this is where the study draws its dividing line.

Angiotensin II, the body’s own signal to tighten blood vessels, sits at the middle of it. Some medicines reduce how much of it gets made. Others lower blood pressure by a route that leaves the body making more of it, or making the same amount while blocking where it lands.

So the two groups here are not stronger and weaker drugs. They are drugs that increase Ang-II formation on one side, covering angiotensin receptor blockers, thiazide diuretics and one family of calcium channel blockers, and drugs that decrease it on the other, covering angiotensin-converting enzyme inhibitors, beta blockers and a different family of calcium channel blockers.

Why anyone would ask

Hypertension and dyslipidaemia are key modifiable risk factors of dementia, and treating both is already standard. What is not established is whether the particular combination matters, and evidence on specific combinations remains limited.

Given how many people are on some version of these two drugs, even a small difference between pairings would be worth a great deal at population scale. That is the prize, and it is why the question keeps getting asked of observational data while nobody funds the trial.

What the cohort showed

Everyone included was over 45, had both high blood pressure and a cholesterol problem, and was actually taking both medicines: the analysis required a proportion of days covered above 80%, which screens out people who filled a prescription once.

Across a mean follow-up of more than twelve years, the group on statins with angiotensin-II promoting medicines had roughly 12% less dementia, and about 13% lower death from any cause, than the group on statins with the suppressing kind.

Both figures are modest, and both are consistent in direction, which in a cohort this size is about as much as an association can offer.

The number that should not be believed

Then the subgroup analysis, and this is where care is needed.

Combinations of rosuvastatin or atorvastatin with the promoting group opened the widest gaps of all: 57% less dementia for the first, 26% less for the second.

A 57% reduction in dementia from choosing one common drug pairing over another would be the most important finding in preventive neurology for a decade. It would also be wildly out of line with everything else known about how much any drug moves dementia risk, and out of line with this study’s own headline figure of 12%.

When a subgroup produces an effect several times larger than the overall result, the likeliest explanation is not that the subgroup is special. It is that the people in it differ from the people they are being compared with in ways the matching did not capture.

What prescribing data cannot see

Nobody was randomly assigned here, and the assignment that did happen was made by clinicians with reasons.

A drug from the suppressing group tends to be chosen for someone with particular kidney or heart circumstances; a beta blocker often follows a heart attack or arrives with a rhythm problem. Those histories carry their own dementia risk, and they are the reason the two groups existed in the first place. Propensity matching balances what was recorded, and a prescribing decision contains more than what gets recorded.

There is also the question of who stays on a medicine for twelve years at high adherence. That is a group selected for stability, engagement and relative good health before any drug is considered.

What it is good for

Not a conversation with a pharmacist about switching. The blood pressure medicine that suits a person is the one that controls their blood pressure without wrecking anything else, and dementia risk twelve years out is not a criterion any prescriber can act on from a cohort study.

What a result like this is genuinely good for is justifying the trial. The question is answerable: take people already needing both drugs, randomize the pairing, follow them. It has not been done because dementia trials are long and expensive, and the case for funding one is built out of exactly this kind of evidence.

On that reading the useful finding is not the 12%. It is that the difference ran opposite to the intuitive direction, which is the sort of thing that makes a question worth settling properly rather than arguing about.