Science Independently confirmed 2 publishers3 min readPublished Aug 26, 2026, 2:52 AM PDT

Oxford's Zostavax-to-Shingrix switch closes the healthy-vaccinee loophole, and opens another

A natural experiment at the October 2017 vaccine changeover gives the cardiovascular claim a cleaner comparison group. It also means nobody has yet measured Shingrix against no vaccine.

The Scientist · Science desk

How we use AI Send a correction

Photograph accompanying Oxford's Zostavax-to-Shingrix switch closes the healthy-vaccinee loophole, and opens another — Photo: statnews.com

What happened

Oxford researchers compared 72,000 vaccinated people aged 60 and over on either side of the October 2017 US switch from Zostavax to Shingrix.
Stroke was 12 per cent lower in men but not significantly lower in women; atrial fibrillation was 7 per cent lower across the cohort.
A Danish randomised trial of about 162,000 people is due to report its first results in 2027.

Why it matters

constraint Because both arms were vaccinated, the study can rank two products but leaves the benefit of Shingrix against no vaccine unquantified, which is the number most immunisation arguments actually use.
contradiction Outside commentary turns the result into advice to get vaccinated for heart protection, while the compared groups were all vaccinated, so the evidence and the advice are answering different questions.
cost The population-scale benefit has to be bought with coverage roughly three times what the US now achieves, which is outreach spending rather than anything the vaccine science can supply.
decision Anyone assembling a funding case for adult shingles vaccination has to decide now whether a benefit of about one avoided event per 100 people over seven years belongs in it, or hold until...

The design's strength and its limit are the same fact: everyone in both arms had a needle in their arm. Matching people who received Zostavax in the six months before October 2017 against those who received Shingrix in the six months after ^[2] removes the objection that dogged earlier work, where vaccinated people were compared with unvaccinated people who tended to be less healthy and less well connected to care ^[6]. It also changes the question being answered. Zostavax is not a placebo, so the reported 9 per cent is the gap between two vaccines ^[16], not the benefit of being vaccinated. The Danish trial that Maxime Taquet points to randomises Shingrix against no shingles vaccine at all ^[8], so its first results will speak to a different contrast and cannot confirm this figure.

Then there is the arithmetic across the two write-ups. New Scientist gives 10.9 per cent of the Zostavax group and 9.6 per cent of the Shingrix group with heart failure, stroke or clogged heart arteries by 3.5 years ^[20]. That is 1.3 percentage points ^[29], a relative reduction near 12 per cent ^[30], one event avoided per 77 people vaccinated ^[31]. STAT reports the seven-year figures as 9 per cent relative and about one percentage point absolute ^[16]^[19], which is one per 100 ^[32]. The relative effect is therefore smaller at seven years than at 3.5 ^[28], matching the researchers' statement that the benefit waned over the second half of follow-up while still favouring Shingrix ^[5]. An effect that peaks early and erodes fits protection against acute events better than a durable change in vascular biology.

That matters because the mechanism is unsettled. Betty Raman of Oxford pointed to AS01, the immune-stimulating component of Shingrix, reprogramming monocytes to make fewer inflammatory cytokines ^[9]^[10]. John Tregoning at Imperial College London offered the plainer route: Shingrix prevents shingles better, and infection brings inflammation that stresses organs including the heart ^[22]. If the adjuvant is doing the work, the result travels to other AS01 products; if it is simply better shingles prevention, the cardiac gain tracks vaccine efficacy and stops there. The authors say they can only speculate ^[11].

The population claim rests on neither. Taquet's estimate of hundreds of thousands of cardiovascular cases delayed or prevented within a decade assumes everyone aged 50 and over in the US receives Shingrix ^[12]; about a third of eligible people currently do ^[23], so the projection is priced at roughly triple today's coverage ^[33]. Two caveats sit in the data itself. TriNetX draws on 60 health systems but excludes people with public insurance ^[4], which thins the Medicare-age population in a study restricted to people 60 and older ^[3]. And October 2017 was also when US eligibility dropped from 60 to 50 ^[24]; the cohort's age floor keeps most of that expansion outside the comparison, though the two publishers describe the enrolment windows differently enough (six months either side, versus April to September 2017 and then the following year) ^[2]^[13] that the boundaries are worth seeing in the paper.

What to watch

Whether the male-only stroke signal replicates elsewhere or dissolves as a subgroup artefact.
First DAN ZOSTER results in 2027, which test Shingrix against no vaccine rather than against Zostavax.
Whether any effect is traced to the AS01 component itself, which would implicate other adjuvanted vaccines.

Clarity's read

What the record supports and how the coverage leans. The claims behind it follow.

Reality

Evidence63
Adoption58
Hype gap+20
Incentives42
Confidence74

Why these scores

Claim ledger

Ranked by verification strength, evidence, and original report placement.

[1]
A study by Oxford researchers, the same team that reported a link between Shingrix and lower dementia rates two years earlier, was published on Wednesday in Nature Medicine.
ReportedSupported2 sources— create a free account to open them View cited source
[2]
The study compared matched groups of vaccinated people who received their shots in the six months before or the six months after October 2017, when Shingrix was approved in the US and displaced the live vaccine Zostavax.
ReportedSupported2 sources— create a free account to open them View cited source
[3]
The analysis used the TriNetX network's electronic health records of 72,000 vaccinated people aged 60 and older.
ReportedSupported2 sources— create a free account to open them View cited source
[4]
TriNetX holds data from 60 health systems but does not include people with public health insurance.
ReportedSupported2 sources— create a free account to open them View cited source
[5]
The benefit waned over the next 3.5 years but still favoured the Shingrix group.
ReportedSupported2 sources— create a free account to open them View cited source
[6]
A 2025 study linked shingles vaccination to reduced cardiovascular risk, but compared people who chose to be vaccinated with those who did not; Taquet said vaccinated people tend to have healthier lifestyles, making it unclear whether vaccination itself was responsible.
ReportedSupported2 sources— create a free account to open them View cited source
[7]
The DAN ZOSTER study involves about 162,000 people in Denmark, with first results anticipated in 2027.
ReportedSupported2 sources— create a free account to open them View cited source
[8]
In the ongoing Danish trial, more than 160,000 people aged 65 and older will receive either Shingrix or no shingles vaccine.
ReportedSupported2 sources— create a free account to open them View cited source
[9]
Zostavax contains a live, weakened varicella-zoster virus, while Shingrix is made of a viral protein plus a strong immune stimulant called AS01.
ReportedSupported2 sources— create a free account to open them View cited source
[10]
Betty Raman of Oxford said prior research suggests AS01 reprogrammes monocytes to produce fewer inflammatory cytokines, which promote clogging of blood vessels.
ReportedSupported2 sources— create a free account to open them View cited source
[11]
The study authors said they can only speculate on how Shingrix exerts this protection because they do not know the exact mechanism; prior research has shown the virus that causes chickenpox and shingles heightens cardiovascular risk.
ReportedSupported2 sources— create a free account to open them View cited source
[12]
The researchers estimate that if everyone aged 50 and older in the US received Shingrix, it could delay or prevent hundreds of thousands of cases of cardiovascular problems within a decade, Taquet said.
ReportedSupported2 sources— create a free account to open them View cited source
[13]
New Scientist describes the cohort as about half receiving Zostavax between April and September 2017, with nearly all the remaining participants receiving Shingrix the following year.
ReportedSupported2 sources— create a free account to open them View cited source
[14]
Mark Russell of King's College London said that although benefits are relatively small on an individual basis, millions of people receive shingles vaccination, so even a small cardiovascular benefit could translate into a substantial number of events prevented at population level.
ReportedSupported2 sources— create a free account to open them View cited source
[15]
Taquet said that if confirmed, Shingrix could prevent hundreds of thousands of cardiovascular events in the USA alone and may be the first vaccine that protects the heart and the brain.
ReportedSupported2 sources— create a free account to open them View cited source
[16]
After seven years, receiving Shingrix rather than Zostavax was associated with a 9% lower overall cardiovascular burden, defined as a composite of coronary heart disease, heart failure and stroke.
ReportedSupportedView cited source
[17]
Coronary heart disease was 10% lower and heart failure 12% lower in the Shingrix cohort.
ReportedSupportedView cited source
[18]
Strokes were 12% lower in men but not significantly lower in women; atrial fibrillation was 7% lower in all participants.
ReportedSupportedView cited source
[19]
The relative risk figures translate in absolute terms to 1% fewer cardiovascular events after seven years.
ReportedSupportedView cited source
[20]
By 3.5 years after vaccination, 10.9% of participants who had Zostavax had developed at least one of heart failure, stroke or clogged arteries in the heart, compared with 9.6% in the Shingrix group, a small but statistically significant difference.
ReportedSupportedView cited source
[21]
Taquet said the study leverages a natural experiment that mitigates many of the biases associated with observational research and gives a more reliable estimate of the effect of Shingrix on cardiovascular disease.
ReportedSupportedView cited source
[22]
John Tregoning of Imperial College London said Shingrix also prevents shingles more effectively than Zostavax, and that viral infection causes inflammation that stresses other organs including the heart.
ReportedSupportedView cited source
[23]
Currently only about a third of those eligible for the shingles vaccine in the US get it.
ReportedSupportedView cited source
[24]
The availability of Shingrix in the US in October 2017 coincided with shingles vaccine eligibility being lowered from age 60 to 50.
ReportedSupportedView cited source
[25]
Tregoning said people should get their shingles vaccine not only because it stops shingles but because it potentially gives protection against dementia and heart disease.
ReportedSupportedSource: John Tregoning, Imperial College London, quoted by New ScientistView cited source
[26]
Kaleen Hayes of Brown University School of Public Health, who was not involved in the study, said there had been some but very conflicting evidence on whether the vaccine has cardioprotective as well as potential neuroprotective effects.
ReportedSupportedView cited source
[27]
Taquet said the study remains observational even though it is a natural experiment, and that the team is keen to see results from randomised controlled trials.
ReportedSupportedView cited source
[28]
The relative benefit is smaller at seven years (9%) than at 3.5 years (about 12%).
DerivedSupported2 sources— create a free account to open them
[29]
The 3.5-year absolute difference between the two vaccine groups is 1.3 percentage points.
DerivedSupported1 source— create a free account to open them
[30]
That 3.5-year gap is a relative reduction of about 12%.
DerivedSupported1 source— create a free account to open them
[31]
At 3.5 years the difference amounts to roughly one event avoided per 77 people given Shingrix instead of Zostavax.
DerivedSupported1 source— create a free account to open them
[32]
The reported seven-year absolute difference of 1% amounts to about one event avoided per 100 people vaccinated with Shingrix instead of Zostavax.
DerivedSupported1 source— create a free account to open them
[33]
Reaching the modelled coverage of everyone aged 50 and over would require roughly three times current US uptake.
DerivedSupported1 source— create a free account to open them

Sources

2 independent publishers whose own reporting we read for this story.

newscientist.com
1 article · August 26, 2026
Shingles vaccine may lower risk of stroke and heart failure
statnews.com
1 article · August 26, 2026
Newer shingles vaccine linked to lower heart disease risk, ‘natural experiment’ shows

Ask Clarity

Have a question this piece left open? The answer comes from its reporting and claim ledger — nothing else.

Clarity analysis

How the coverage leans and what the record actually supports. Read the methodology

Perspective Coverage

How much of the coverage speaks to builders, operators and investors. It measures emphasis, not whether the story is right.

Builder

For anyone who builds or evaluates causal-inference work on real-world data, the interesting object here is the identification strategy, not the effect size. The Oxford team exploited a policy discontinuity - the October 2017 US switch from live Zostavax to adjuvanted Shingrix - to compare two cohorts of people who all chose to be vaccinated, which removes the healthy-vaccinee selection that undermined the 2025 vaccinated-versus-unvaccinated study. STAT documents the mechanics in most detail: 72,000 TriNetX records for people 60 and older, matched six-month windows either side of the approval, and a dataset spanning 60 health systems that excludes publicly insured patients - a material coverage hole in an over-60 population. Two design residues remain visible in the sources: the same date saw eligibility drop from 60 to 50, and by construction the design can compare only vaccine against vaccine, never vaccine against none. The reported effects are internally consistent but modest and time-varying: 1.3 percentage points at 3.5 years shrinking to a 9% relative, 1% absolute difference at seven years. The authors themselves label the work observational and point to the randomised DAN ZOSTER trial.

Reality

Evidence, adoption, hype gap, incentives and confidence, each read independently from the current record on the scale shown beneath it. How these are measured.

Evidence63: 0100
Peer-reviewed natural experiment, still observational with a known data gap
The finding rests on a Nature Medicine analysis of about 72,000 matched TriNetX records using a policy discontinuity that removes healthy-vaccinee selection, and effects are consistent across the composite and most components. It is capped short of strong because the authors themselves call it observational, the dataset excludes publicly insured patients in an over-60 population, the same October 2017 date carried an eligibility change from 60 to 50, stroke benefit was not significant in women, the absolute difference is about 1% at seven years and shrinking, and the confirmatory randomised trial does not report until 2027.
Adoption58: 0100
Vaccine widely deployed, cardiovascular rationale not yet acted on
Shingrix is already the standard US shingles vaccine since the 2017 displacement of Zostavax and is offered by the UK NHS to those turning 65 from September 2023, so the product itself is broadly deployed. But only about a third of eligible Americans take it, the modelled cardiovascular benefit assumes roughly triple that uptake, and the reports show no guideline, indication, or programme change made on the basis of the cardiovascular finding.
Hype gap+20: −100 understated0 aligned+100 overstated
Framing outpaces a 1% absolute, waning effect
Headline framing of 'significantly reduce your risk' and projections of hundreds of thousands of prevented US events sit above what the sources measure: a 9% relative and roughly 1% absolute difference at seven years, narrowing from 1.3 percentage points at 3.5 years, in a vaccine-versus-vaccine comparison that never tested Shingrix against no vaccine. The gap is moderate rather than large because both publishers carry counterweights - the authors' observational caveat, the small per-person benefit, the conflicting prior evidence, and the pending randomised trial.
Incentives42: 0100
Author-briefed findings, partly offset by independent commentary
The most expansive statements - the best-evidence-yet claim, the hundreds-of-thousands projection, the first-vaccine-for-heart-and-brain line, and the booster cost-saving suggestion - come from the study team at its own press briefing, and the team has a track record to defend from its earlier dementia finding. That is partly offset by three commentators not involved in the study, one of whom stresses conflicting prior evidence, and by the authors' own observational caveat. The sources disclose no funding, manufacturer, or competing-interest information, so this reads as ordinary researcher promotion rather than a documented commercial interest.
Confidence74: 0100
Two independent accounts of one peer-reviewed study
Both sources report the same peer-reviewed analysis on the same day with consistent cohort size, design, and direction of effect, and each adds numbers the other lacks, so the factual core is well pinned. Confidence is not higher because the entire cluster traces to a single study and press briefing, the sources give partly non-overlapping effect metrics at different time horizons, and no funding disclosures or trial-registration details are supplied.

Topics and entities

Follow any of these and your For You feed starts watching them — no settings page required.

Topics

Shingles vaccinationFollow
Pending randomised trial readoutsFollow
Vaccine adjuvants and trained immunityFollow
Cardiovascular preventionFollow
Natural-experiment epidemiologyFollow
Real-world evidence from EHR networksFollow

Entities

ShingrixFollow
ZostavaxFollow
AS01Follow
TriNetXFollow
University of Oxford Follow
Nature Medicine Follow
DAN ZOSTERFollow
Maxime TaquetFollow
Betty RamanFollow
Fabiana Corsi-ZuelliFollow
John TregoningFollow
Mark RussellFollow
Kaleen N. HayesFollow
Brown University School of Public HealthFollow
Imperial College London Follow
King's College London Follow
Science Media CentreFollow
National Health Service Follow