Should a healthy 73-year-old start a statin after the STAREE trial?

The STAREE trial randomized 9,971 community-dwelling adults aged 70 and older, with no prior cardiovascular disease, diabetes, or dementia, to atorvastatin 40 mg or placebo for a median of 5.9 years. Major adverse cardiovascular events fell from 8.3 percent to 6.0 percent, a hazard ratio of 0.70 and a 30 percent relative reduction, equivalent by simple arithmetic to roughly one event prevented for every 44 people treated for six years, though that number-needed-to-treat estimate was not itself reported by the trial. However, the trial's primary endpoint of disability-free survival, meaning living free of death, dementia, and persistent physical disability, showed no significant difference: the composite event occurred in 12.8 percent of the atorvastatin group and 13.6 percent of the placebo group, so about 87 percent of both groups remained alive and independent. Adverse events were more common with atorvastatin, including musculoskeletal symptoms, hepatobiliary abnormalities, and new diabetes-related events. The decision should account for life expectancy, since the benefit curves separate at about two years, personal weighting of preventing a nonfatal heart attack or stroke, and whether atherosclerosis is actually present. A coronary artery calcium score is the most useful single test to clarify that, and a score of zero is meaningful information in an older adult.

Why do I need a urine albumin test if my kidney blood test was normal?

Blood creatinine and the estimated glomerular filtration rate calculated from it measure how well the kidney filters, while the urine albumin-to-creatinine ratio measures whether the kidney is leaking protein, which often occurs earlier while filtration still appears normal. The 2026 ESC guidelines on cardiovascular disease and chronic kidney disease recommend both an eGFR and a urine albumin-to-creatinine ratio at the time of cardiovascular diagnosis, under a framework called STAMP: Screen, Triage, Address CKD risk, Modify CVD management, Plan health services. Approximately 100 million people in Europe have chronic kidney disease, much of it undiagnosed. Albuminuria is also an independent marker of vascular injury and cardiovascular risk. If elevated, treatments including renin-angiotensin system inhibitors and SGLT2 inhibitors slow progression and reduce cardiovascular events. A cystatin C-based eGFR is often more accurate than creatinine alone in people with high or low muscle mass.

EverHealth Weekly Issue 07 - EverHealth Institute

Good morning,

We promised you Munich, and Munich delivered. Over four days the European Society of Cardiology released 59 trials, three new clinical practice guidelines, and a new universal definition of heart attack — and three of those trials answered questions we have been unable to answer for a decade.

The first told us what happens when you give a statin to a healthy 70-year-old — and the answer is both more encouraging and more sobering than either camp expected. The second put an imaging scanner on 16,808 people with no known heart disease and found plaque in 57% of them, including roughly one in thirteen adults aged 18 to 29. The third followed 33,995 rural Chinese adults for seven years and showed that lowering blood pressure lowers dementia — not in a model, in a randomized trial.

Read together, they make one argument. Prevention has spent thirty years estimating who has disease. We now have the tools to look. That’s Section 3.

— The Everhealth Institute team

This Week in the Evidence

Dispatch from ESC Congress 2026 · Munich, August 28–31

Strong — randomized controlled trial

STAREE: statins cut cardiovascular events by 30% in healthy adults over 70 — but did not extend disability-free life

Double-blind randomized controlled trial · 9,971 adults aged 70+ · median 5.9 years · presented ESC Congress 2026, August 29 · simultaneously published in the New England Journal of Medicine

This is the trial we flagged last week, and it is the most consequential primary-prevention result of the year. Australian investigators randomized 9,971 community-dwelling adults aged 70 and older — mean age 74.7, 52% women, none with prior cardiovascular disease, diabetes, or dementia — to atorvastatin 40 mg or placebo, and followed them for a median of 5.9 years. Nearly every statin trial before this one studied people under 70; the practice of prescribing statins to older adults has been an extrapolation, not a finding.

On cardiovascular events, the answer was unambiguous. Major adverse cardiovascular events — cardiovascular death, non-fatal heart attack, stroke, or coronary revascularization — occurred in 6.0% of the atorvastatin group versus 8.3% on placebo: HR 0.70 (95% CI 0.61–0.82), p<0.001. That is a 30% relative reduction and roughly 2.3 fewer events per 100 people over six years — by our arithmetic, about 44 people treated for six years to prevent one event (a derived estimate, not a figure the trial reported). The benefit held across age bands, including those over 75.

The trial’s primary endpoint, however, was not cardiovascular events. It was disability-free survival — living free of death, dementia, and persistent physical disability. On that measure there was no difference: the composite event occurred in 12.8% of the atorvastatin group and 13.6% of placebo — meaning roughly 87% of both groups remained alive and independent — HR 0.94 (95% CI 0.84–1.05), p=0.25. Mortality and dementia did not differ substantively either. Serious adverse events were uncommon and equal (2.7% in both arms), but the atorvastatin group had more musculoskeletal complaints (32.0% vs 29.4%), more hepatobiliary abnormalities (3.3% vs 0.9%), and more new diabetes-related events (4.0% vs 2.7%).

Clinical takeaway: Both headlines are true, and honest practice requires holding both. A statin in a healthy 70-something clearly prevents heart attacks and strokes — that is now proven, not assumed, and age alone is no longer a reason to withhold one. But it did not, over six years, buy additional years of independent living. The reconciliation is arithmetic: the events prevented were largely non-fatal, and in this age group dementia and frailty are driven by much more than lipids. Two practical points. First, the benefit curves separate at roughly two years — so life expectancy genuinely matters, and a patient unlikely to have five good years ahead may gain little. Second, this trial excluded people with diabetes and prior cardiovascular disease, so it says nothing about those groups, in whom the case for treatment was already strong. For the healthy older adult sitting in a coin-flip decision, a coronary artery calcium score — or, where more detail is warranted, CT coronary angiography — remains the single best tiebreaker in either direction.

Source: Zoungas S, et al. STAREE: Statins in Reducing Events in the Elderly. Presented at ESC Congress 2026, Munich, August 29, 2026; published simultaneously in the New England Journal of Medicine. ESC press release → · Detailed report and expert commentary →

Strong — large imaging cohort

REACT: silent atherosclerosis in 57% of healthy adults — and standard risk scores missed most of it

Cross-sectional imaging study · 16,808 adults aged 18–70 without known cardiovascular disease · Denmark and Spain · New England Journal of Medicine, August 29, 2026

If STAREE was the most consequential trial in Munich, REACT was the most unsettling study. Investigators imaged 16,808 adults aged 18 to 70 (mean age 45) with no known cardiovascular disease, balanced by sex and age decade, using 3D vascular ultrasound of the carotid and femoral arteries plus coronary CT angiography with assessment of calcification. This is not a risk model. It is a photograph of the arterial tree in a general adult population.

57.1% had atherosclerotic plaque. The prevalence traced an S-shaped curve across the lifespan: roughly 1 in 13 adults aged 18–29 already had detectable plaque, rising to about 9 in 10 by ages 60–70. Men ran five to ten years ahead of women — but women showed a notably steep climb between ages 40 and 60, coinciding with the menopausal transition. Most people with coronary plaque also had it in the carotid or femoral arteries, meaning a simpler peripheral ultrasound often flags the same disease.

The finding with the sharpest clinical edge concerns risk calculators. SCORE2 — Europe’s standard 10-year risk equation, conceptually analogous to the US Pooled Cohort Equations and PREVENT — classified only a small minority of the people who actually had plaque as high risk, and performed especially poorly in younger adults. As lead investigator Henning Bundgaard put it: “This approach relies on an estimation — not on knowledge of the presence of the actual disease.”

Clinical takeaway: This is the most important number in the issue: a low risk score is not the same as an absent disease. Risk equations were built to allocate statins efficiently across populations, and they do that job reasonably well. They were never designed to tell an individual whether they have atherosclerosis — and REACT shows how often they get that specific question wrong, especially before age 50, and especially in women approaching menopause, whose risk scores stay reassuringly low while their arteries change. Two honest limits: this is a single snapshot, so it cannot prove that imaging-guided treatment improves outcomes — the authors explicitly call for that trial next — and finding plaque in a 28-year-old raises real questions about overtreatment and anxiety that the study does not answer. One clarification worth making: REACT used research-grade 3D ultrasound and CT angiography, which is not the same as the inexpensive coronary calcium score most patients can access. The calcium score is the practical, well-validated entry point; REACT is the argument for looking at the artery rather than estimating it, not a prescription for a specific scan.

Source: Bundgaard H, García-Lunar I, Kofoed KF, et al. Prevalence of Silent Atherosclerosis Across Adult Life. N Engl J Med. 2026. doi:10.1056/NEJMoa2609059 → · ESC press release →

Strong — cluster-randomized trial, 7 years

CRHCP: seven years of intensive blood-pressure control cut dementia by 15%

Cluster-randomized trial · 326 villages · 33,995 adults · mean age 63 · 4 years intervention + 3 years observation · presented ESC Congress 2026, August 30

The China Rural Hypertension Control Program randomized 326 villages — 33,995 adults, mean age 63, 61% women — to usual care or to a stepped-care antihypertensive protocol delivered by non-physician community health providers under primary-care supervision, targeting below 130/80 mmHg. The intervention ran four years; participants were then followed a further three.

The blood-pressure separation was enormous: systolic pressure fell 17.6 mmHg in the intervention villages versus 2.4 mmHg in usual care. Over seven years, dementia occurred in 8.85% of the intervention group versus 10.55% of usual care — an adjusted risk ratio of 0.85 (95% CI 0.78–0.91, p<0.001), a 15% relative reduction. Cognitive impairment without dementia fell 13%. Serious adverse events were, if anything, lower in the intervention arm (47.15% vs 49.78%). The same program’s cardiovascular outcomes were published in The Lancet in 2023 by He and colleagues; the dementia analysis was presented in Munich by Shanshan Zhong, with Yingxian Sun as principal investigator.

Clinical takeaway: Last week we reported the ARIC observational finding that three midlife vascular risk factors tracked with 13 dementia-free years. This is the randomized test of the same hypothesis, and it is positive. Blood pressure is not merely associated with dementia — lowering it lowers dementia risk. That moves hypertension control from cardiovascular housekeeping to the most evidence-backed brain-protective intervention we currently possess. Three caveats worth stating plainly: the effect size is modest in absolute terms (about 1.7 fewer dementia cases per 100 people over seven years), the population was rural Chinese adults with substantial untreated hypertension at baseline — the effect may be smaller where treatment is already good — and dementia adjudication in a village-based trial is harder than in an academic memory clinic. None of that undoes the central result. The other quietly radical finding is who delivered the care: trained non-physicians, at scale, achieved a 17.6 mmHg drop. That is better than most specialty clinics manage.

Source: Zhong S, Sun Y, et al. China Rural Hypertension Control Program: long-term effects of intensive blood pressure control on dementia. Presented at ESC Congress 2026, Munich, August 30, 2026. Prior cardiovascular outcomes: He J, Ouyang N, Guo X, et al. Lancet. 2023;401:928–938. ESC press release → · Trial report →

Early — phase 2, biomarker endpoint

TRANQUILITY: an IL-6 antibody dropped hs-CRP by up to 89% — and by 85% on quarterly dosing

Phase 2 randomized placebo-controlled trial · 143 adults with stage 3–4 chronic kidney disease and hs-CRP 2–<15 mg/L · 49 US centers · 6 months · presented ESC Congress 2026, August 29

The inflammation hypothesis — the reason we measure hs-CRP at all — holds that residual inflammatory risk drives cardiovascular events even when cholesterol is controlled. TRANQUILITY tested pacibekitug, a monoclonal antibody against interleukin-6, in 143 patients with chronic kidney disease and elevated inflammation (mean age 69, 64% women, 72% already on statins, 59% with diabetes), across three dosing schedules versus placebo for six months.

Median time-averaged hs-CRP change through day 180: +7% on placebo, versus −76% (25 mg every 90 days), −85% (50 mg every 90 days), and −89% (15 mg every 30 days) — all p<0.0001. Fibrinogen and lipoprotein(a) also fell. Discontinuations were rare (1.9%) with no clear dose-related safety signal. A phase 3 cardiovascular outcomes trial is planned.

Clinical takeaway: Keep this in the “promising, unproven” column — and understand precisely why. Every number here is a biomarker, not an outcome. We have been burned before: lowering a number is not the same as preventing an event, and the IL-6 pathway carries theoretical infection risk that a 143-patient six-month trial cannot exclude. What makes this worth watching is that the inflammation hypothesis already has one positive outcomes trial behind it (CANTOS, with canakinumab), and that quarterly dosing — which achieved an 85% hs-CRP reduction — would make chronic anti-inflammatory therapy practical in a way daily drugs are not. The immediately actionable point is not the drug — it is the marker. If your cholesterol is well controlled and your risk still feels unexplained, hs-CRP and Lp(a) are inexpensive, widely available, and still under-ordered.

Source: TRANQUILITY investigators. Pacibekitug in patients with chronic kidney disease and elevated hs-CRP. Presented at ESC Congress 2026, Munich, August 29, 2026. ESC press release → · ACC trial summary →

Guidelines That Changed in Munich

Four new ESC documents · what actually changes for prevention

Alongside the trials, the ESC released three new clinical practice guidelines — cardiovascular disease in chronic kidney disease, heart failure, and cardiac rehabilitation (the society’s first dedicated document on the subject) — plus the Fifth Universal Definition of Myocardial Infarction, issued jointly with the ACC, AHA, and World Heart Federation. Two of them change what should be on your lab order.

KidneyEvery patient with cardiovascular disease should be screened for kidney disease — with two tests, not one. The new guideline recommends both an eGFR (from blood creatinine) and a urine albumin-to-creatinine ratio (UACR) at the time of cardiovascular diagnosis, under a framework the task force calls STAMP — Screen, Triage, Address CKD risk, Modify CVD management, Plan services. Roughly 100 million people in Europe have chronic kidney disease, much of it undiagnosed. It emphasizes early RAS inhibitors, early and cost-effective SGLT2 inhibitors, and statin-based therapy. Why this matters to you: the urine test is the one that gets skipped, and it is frequently abnormal while creatinine still looks normal. Albuminuria is both a kidney finding and an independent cardiovascular risk marker.
Heart failureHeart failure is now staged from prevention onward. The 2026 guideline adopts an A-through-D staging framework that begins with at-risk patients before any structural change — an explicit shift toward prevention. It eliminates the confusing “mildly reduced ejection fraction” category (consolidating everything below 50%), replaces “acute” heart failure with “decompensated” to reflect gradual decline, and reorganizes drugs into foundational, additional, and interventional therapy. Mineralocorticoid receptor antagonists become a Class I recommendation regardless of ejection fraction, and semaglutide or tirzepatide earn a Class IIa recommendation in preserved-ejection-fraction heart failure with obesity — formal guideline recognition that incretin-based therapy (semaglutide is a GLP-1 receptor agonist; tirzepatide a dual GIP/GLP-1 agonist) now belongs in the heart failure toolkit.
Also reportedThree more results worth knowing. Plozasiran (SHASTA-3 and SHASTA-4) reduced triglycerides by roughly 80% in severe hypertriglyceridemia and reduced acute pancreatitis events — meaningful for patients with triglycerides in the many hundreds. Milvexian, a factor XIa inhibitor, showed no benefit after acute coronary syndrome in LIBREXIA-ACS (5.4% vs 5.1%). And ENRICH-AF found that edoxaban in atrial fibrillation patients with prior intracranial hemorrhage did not reduce stroke but more than doubled major bleeding (11.6% vs 5.2%) — a clear negative that will change practice.

Sources: ESC CKD guideline announcement → · 2026 ESC Guidelines: CVD and CKD → · ESC heart failure guideline announcement → · Fifth Universal Definition of MI → · SHASTA-3/4 →

The Long View

Original perspective · one idea, examined

The Long View

The estimate and the photograph

Everhealth Institute · ~4 minute read

There is a sentence buried in the REACT press coverage that deserves to be the epigraph for the next decade of preventive medicine. Describing the risk calculators that currently govern who gets treated, the study’s lead investigator said: “This approach relies on an estimation — not on knowledge of the presence of the actual disease.”

Consider what that means in practice. A forty-four-year-old woman walks into a clinic. Her cholesterol is unremarkable, her blood pressure is 124/78, she doesn’t smoke. A risk equation returns a ten-year risk of 2%, and she is told, correctly by the standards of 2026, that she is low risk. REACT suggests there is a meaningful chance she already has plaque in her carotid, femoral, or coronary arteries — and that her risk score would not have known, because risk scores are built from age, sex, and a handful of blood values, and none of those are the disease. They are correlates of the disease. The disease itself is a physical structure in an artery wall, and it can be photographed.

This distinction has been academic for most of the history of cardiology, because we could not photograph arteries cheaply or safely. That is no longer true. A coronary calcium score costs less than a month of many prescriptions and takes ten minutes. Carotid ultrasound involves no radiation at all. REACT’s incidental finding — that most people with coronary plaque also had it in the neck or leg — hints that the entry point may be even simpler than we assume.

Now hold STAREE next to it. Nearly ten thousand healthy older adults, six years, a rigorous randomized design — and the answer was split. Thirty percent fewer cardiovascular events, and no gain in disability-free survival. Population-level trials give population-level answers, and this one says: this drug prevents heart attacks in people over 70, and does not by itself make them live longer independently. Both halves are real. The half that applies to you depends on which of those outcomes your biology is actually heading toward — and a risk equation cannot tell you that, while an image of your arteries goes a long way.

CRHCP completes the argument from the opposite direction. Seven years, thirty-four thousand people, a 17.6 mmHg drop in systolic pressure, and 15% less dementia. Here the intervention was not exotic and the measurement was not sophisticated. It was a blood pressure cuff, used relentlessly, by people who were not physicians. The most powerful result of the meeting came from the cheapest instrument in the building.

So the honest synthesis is not “image everything.” It is a hierarchy. Do the cheap, proven things relentlessly — blood pressure to target and kept there for years, lipids including Lp(a) and hs-CRP measured once and acted on, kidney function assessed with both a blood and a urine test, cardiorespiratory fitness and body composition tracked. Then, where the decision is genuinely uncertain, stop estimating and look. That is what anatomic imaging is for: not to generate anxiety in the worried well, but to resolve the coin flips — the fifty-year-old with a family history and a normal panel, the postmenopausal woman whose risk score has not caught up to her physiology, the seventy-two-year-old deciding about a statin.

Medicine has spent a century getting very good at predicting who will get sick. The next decade belongs to clinicians willing to find out who already is. An estimate is a guess with math around it. A photograph is a fact.

Do This Week

Plain-language, practical prevention

1If a “low risk” score is the only reason you haven’t looked at your arteries, that reason just got weaker. REACT found plaque in 57% of adults with no known heart disease, and standard risk equations flagged only a small minority of them. You do not need imaging if your decision is already clear — if you are on treatment, or genuinely low risk by every measure and comfortable with that. You should consider a coronary artery calcium score — or, where a fuller picture is warranted, one of the advanced imaging studies such as CT coronary angiography — if you are 40+ with a family history, a borderline risk estimate, an elevated Lp(a), or a statin decision you keep postponing. A calcium score of zero is genuinely reassuring; a score above zero changes the conversation entirely.
2Add one urine test to your next lab order. The new ESC kidney guideline says every patient with cardiovascular disease should have both an eGFR and a urine albumin-to-creatinine ratio. The UACR is the one routinely omitted — and it can be abnormal while your creatinine and eGFR still read normal. It is inexpensive, it is a marker of both kidney and cardiovascular risk, and if it is elevated there are now several treatments that change the trajectory. Ask for it by name, alongside a cystatin C-based eGFR if available, which is often more accurate than creatinine alone in muscular or older adults.
3If you are over 70 and healthy, have the statin conversation with both STAREE numbers in hand. The trial showed 30% fewer heart attacks, strokes, and revascularizations — and no improvement in years lived free of dementia and disability. Two questions make the decision tractable: Do I have a reasonable expectation of five or more good years? (the benefit curves do not separate until about year two) and Do I actually have plaque? A calcium score can turn this from a values judgment into a data-driven answer, and a zero score in a 74-year-old is meaningful information.
4Treat your blood pressure as brain care, and measure it at home. CRHCP is now the randomized evidence that sustained control below 130/80 reduces dementia — 15% fewer cases over seven years. The operative word is sustained: this was four years of active treatment plus three more of follow-through, not a single good reading. Buy a validated upper-arm cuff, measure twice in the morning and twice in the evening for a week, and bring the average to your clinician. Home averages predict outcomes better than office readings, and most people are surprised by the gap.

Questions We Got This Week

Straight answers to what patients actually asked

I’m 73, healthy, and not on a statin. Does STAREE mean I should start one?

It means the question finally has real evidence behind it, and the answer is genuinely individual. STAREE randomized 9,971 adults aged 70 and older to atorvastatin 40 mg or placebo for a median of 5.9 years. Major cardiovascular events fell from 8.3% to 6.0% — a 30% relative reduction, or about one event prevented for every 44 people treated for six years. But the trial’s main endpoint, living free of death, dementia, and disability, showed no difference (12.8% vs 13.6%). Side effects were mostly minor but real: more muscle complaints, more liver enzyme abnormalities, and more new diabetes-related events. Three factors should drive your decision — whether you have a reasonable expectation of at least five more good years (the benefit takes about two years to emerge), how you personally weigh preventing a non-fatal heart attack or stroke against taking a daily medication, and whether you actually have atherosclerosis. On that last point a coronary artery calcium score is the most useful single test available — a score of zero at 73 is meaningful information — and where more anatomic detail is needed, CT coronary angiography shows the artery wall directly.

My risk calculator says I’m low risk. Should I still get a coronary calcium score?

Possibly — and REACT is the reason. Imaging 16,808 adults aged 18 to 70 with no known heart disease found atherosclerotic plaque in 57.1%, including roughly 1 in 13 of those aged 18 to 29 and about 9 in 10 of those aged 60 to 70. Critically, the SCORE2 risk equation classified only a small minority of the people who actually had plaque as high risk, and it performed worst in younger adults. A risk score estimates the probability that you will have an event; it does not tell you whether disease is present. Those are different questions. Calcium scoring is most valuable when a decision hangs in the balance: a borderline risk estimate, a strong family history, an elevated Lp(a), or a woman in the perimenopausal window whose risk score has not caught up to her physiology. It is less useful if you are already on treatment, or if you would not act on the result. And be clear-eyed about the trade-off: REACT was a snapshot, not a treatment trial, so it has not yet been proven that imaging-guided therapy improves outcomes — that trial is the field’s next task.

Does controlling blood pressure really prevent dementia, or is that just an association?

As of this week, it is more than an association. The China Rural Hypertension Control Program randomized 326 villages — 33,995 adults — to usual care or to a stepped-care blood pressure program run by trained non-physician community health providers targeting below 130/80 mmHg. Systolic pressure fell 17.6 mmHg in the intervention group versus 2.4 mmHg with usual care. Over seven years, dementia occurred in 8.85% of the intervention group versus 10.55% with usual care, a 15% relative reduction (adjusted RR 0.85, 95% CI 0.78–0.91), and cognitive impairment short of dementia fell 13%. Because villages were randomly assigned, this is causal evidence rather than correlation. Two honest qualifications: the absolute benefit is modest — roughly 1.7 fewer dementia cases per 100 people over seven years — and this was a population with a great deal of untreated hypertension at baseline, so the gain may be smaller where blood pressure is already reasonably managed. Even so, sustained blood pressure control is now the best-evidenced dementia prevention strategy available.

My kidney function was normal on my last blood test. Why would I need a urine test too?

Because they detect different things. Blood creatinine and the eGFR calculated from it measure how well the kidney is filtering. The urine albumin-to-creatinine ratio (UACR) measures whether the kidney is leaking protein — which frequently happens earlier, while filtration still looks completely normal. The 2026 ESC guideline on cardiovascular disease and chronic kidney disease now recommends both tests at the time of any cardiovascular diagnosis, precisely because relying on creatinine alone misses a large share of early disease; roughly 100 million people in Europe have chronic kidney disease and much of it is undiagnosed. Albuminuria also matters independently of the kidneys — it is a marker of vascular injury throughout the body and an established cardiovascular risk factor. If it is elevated, several treatments including RAS inhibitors and SGLT2 inhibitors meaningfully alter the trajectory. It is worth adding to your next biomarker panel, along with a cystatin C-based eGFR, which is often more accurate than creatinine in people with high or low muscle mass.

Next week: what the new ESC cardiac rehabilitation guideline says about exercise dose — and why “how much” turns out to be the wrong question.