Stopping statins in older adults without heart disease met a non-inferiority bar

Many people reach their seventies and eighties still taking a statin that was started years earlier to prevent a first heart attack or stroke they never had. Whether they should keep taking it is a genuinely open question: the evidence for statins as primary prevention is thinnest in exactly this age group, and the drugs carry their own burdens of pills, interactions and side effects. A trial presented at the American Heart Association's 2026 Scientific Sessions and published in the New England Journal of Medicine on 7 November tested what happens when such patients stop [s1].

The question

The trial addressed a narrow, practical situation: older adults who are taking statins purely for primary prevention — meaning they have no known cardiovascular disease — and who also carry at least two other chronic conditions, the kind of patient for whom cutting unnecessary medicines is most tempting [s1]. STREAM, registered as NCT05178420, asked whether discontinuing the statin is as safe as continuing it over the following two years [s1][s2].

How the trial was built

STREAM was an international, open-label, randomised controlled trial [s1]. It enrolled adults at least 70 years of age who were taking statins for primary prevention, and assigned them in a 1:1 ratio either to discontinue or to continue statin therapy and other lipid-lowering treatments [s1].

The primary outcome was a composite of non-fatal myocardial infarction, non-fatal ischaemic stroke, or death from any cause, whichever came first [s1]. Crucially, the trial was designed as a non-inferiority study: rather than asking whether stopping is better, it asked whether stopping is not meaningfully worse. The primary analysis compared the restricted mean event-free survival time at two years between the groups, against a prespecified non-inferiority margin of minus 5.2 weeks [s1]. In plain terms, stopping would count as non-inferior if, on average, it cost patients no more than 5.2 weeks of event-free time over the two years.

A total of 1,880 participants were randomised: 938 to discontinuation and 942 to continuation [s1]. Their mean age was 77.5 years, and 46.9% were women [s1].

What it found

By the trial's own yardstick, stopping passed. The between-group difference in restricted mean event-free survival time at two years was minus 2.0 weeks, with a 95% confidence interval of minus 3.7 to minus 0.3 weeks and a P value below 0.001 for non-inferiority [s1]. Because the entire confidence interval sat inside the minus 5.2-week margin, discontinuation was declared non-inferior to continuation [s1].

The raw event counts, however, deserve to be read alongside that verdict. Over a median follow-up of 103 weeks, 119 participants (12.7%) in the discontinuation group and 87 (9.2%) in the continuation group had a primary-outcome event [s1]. More events accrued in the group that stopped — the non-inferiority finding means that difference, measured as average event-free time, stayed within a margin the investigators had defined in advance as acceptable, not that the two arms were identical.

Serious adverse events were common in this older, multimorbid population and similar between groups: at least one occurred in 43.0% of the discontinuation group and 39.8% of the continuation group [s1].

What is, and is not, established

STREAM supports a specific, bounded claim: in adults aged 70 and over with at least two chronic conditions and no known cardiovascular disease, stopping a primary-prevention statin was non-inferior to continuing it for major cardiovascular events or death at two years, by a prespecified margin [s1]. That is a meaningful result for a question that previously rested largely on observational data and extrapolation.

The limits sit close to the surface. "Non-inferior within 5.2 weeks" is a judgement about an acceptable trade-off, not a finding of no difference, and the numerically higher event count in the discontinuation group is part of the picture [s1]. The follow-up was two years; statins are taken for decades, and a two-year horizon cannot speak to longer-term risk [s1]. The trial was open-label, and it enrolled a particular group — older, multimorbid, no established heart disease — so it does not apply to people taking statins because they already have cardiovascular disease, for whom the evidence to continue is strong [s1].

Starting or stopping a statin is a decision for a patient and their prescriber, who can weigh the full history; abrupt self-discontinuation of any medicine can carry its own risks. This article describes research and is not medical advice.