Oxford Tool Predicts Statin Severe Muscle Risk

Researchers at the University of Oxford published a clinical prediction model in The Lancet Digital Health on June 25, 2026, that estimates individual patients' risk of serious muscle disorders, defined as muscle conditions leading to hospital admission or death, from statin therapy, finding that more than 98% of GP-eligible patients in England fall into the low-risk category over a ten-year horizon.

What the Model Does

The tool, developed at Oxford's Nuffield Department of Primary Care Health Sciences, draws on anonymised health records from more than 5.6 million people registered with GP practices across England. Researchers used data from 1.7 million people to build the model and a further 3.9 million to validate it.

The prediction model incorporates 22 routinely recorded clinical factors: age, sex, ethnicity, body mass index, smoking status, existing health conditions, prior muscle problems, vitamin D deficiency, concurrent medication use, and the specific statin prescribed. It generates risk estimates across one, five, and ten-year time horizons.

The study's validation was conducted in men aged over 50 and women aged over 60 in England who were potentially eligible for statin treatment. The calculator is currently available through the Oxford University Innovation software store for academic and scientific research use, and the researchers envisage it being used alongside existing cardiovascular risk tools such as QRISK, allowing clinicians to weigh the benefit of preventing a heart attack or stroke against the personalised risk of a serious muscle event.

The Treatment Gap the Study Identified

More than 60% of people identified by their GPs as eligible for statin therapy were not taking the medication, including a proportion assessed as being at high risk of heart attack or stroke. The Oxford modeling study noted that concerns about side effects can be a barrier to uptake, though it did not establish the precise share of non-users whose decision was driven primarily by muscle-related fears.

Cardiovascular diseases are the leading cause of death globally, according to the World Health Organization, and high-intensity statin therapy can reduce LDL cholesterol by 50% or more. Among people who do start a statin, persistence falls over time: a large retrospective study of 411,956 patients found persistence declined from 92.65% at three months to 78.28% at twelve months.

A separate analysis estimated that 65.8 million US adults, representing 44% of the adult population studied, had an indication for statin therapy, with only 45% of that group currently using one.

What the Evidence Shows on Muscle Risk

The Oxford model's primary outcome, hospitalisation or death from a muscle disorder, is rarer than the muscle aches and cramps that patients commonly report. The 98%-plus low-risk finding applies specifically to that severe endpoint, not to ordinary statin-associated muscle symptoms.

The National Lipid Association estimates that clinical rhabdomyolysis occurs in fewer than 1 in 10,000 patients over five years of statin treatment. The condition, which involves severe breakdown of skeletal muscle, can escalate to acute kidney injury requiring dialysis, but this outcome is associated with compounding risk factors, older age, high statin doses, kidney disease, and conditions such as uncontrolled hypothyroidism.

A February 2026 meta-analysis published in The Lancet, drawing on data from 19 double-blind randomised trials and 123,940 participants, found that the rate of muscle pain or weakness was similar between participants taking statins and those taking a placebo, 27.1% versus 26.6%. The authors identified a limited set of outcomes as having a statistically supported excess attributable to statins, and concluded that the majority of conditions listed in product labelling were not causally established.

The editorial accompanying that analysis noted that fear of side effects, muscle pain, fatigue, sleep disturbances, has been reinforced by long lists of potential adverse effects in product labelling, often without clear distinction between established causal harms and safety signals that have not been confirmed.

Who Faces Higher Risk

While the Oxford model confirms that severe muscle events are uncommon across the broad eligible population, the 22-factor approach exists because risk is not evenly distributed.

Older patients carry a meaningfully higher exposure. A nested case-control study across 252,460 new users of lipid-lowering medications in the United States found that statin users aged 65 and over had roughly four times the hospitalisation risk for rhabdomyolysis compared with those under 65, though that finding rested on 21 confirmed rhabdomyolysis cases, and the confidence interval was wide (OR 4.36, 95% CI 1.5–14.1). High statin doses combined with kidney disease represented a further compounding factor in that analysis.

Individual statin type also matters. Spontaneous-report data from the FDA Adverse Event Reporting System show a higher reporting signal for simvastatin and lovastatin relative to other statins for rhabdomyolysis, though pharmacovigilance signals reflect reporting patterns rather than confirmed incidence rates. The same dataset identified possible interaction signals with several concurrent medications, including furosemide, allopurinol, and pantoprazole, which may influence statin metabolism.

The Oxford model brings these variables into a single patient-level estimate, quantifying a ten-year risk number rather than leaving clinicians to weigh isolated population averages.

What the Researchers Expect the Tool to Do

The team envisions the calculator being used alongside QRISK in primary care consultations, giving clinicians and patients a side-by-side picture of cardiovascular benefit and muscle-disorder risk specific to that individual. As currently distributed through the Oxford University Innovation software store, the algorithm is designated for academic and scientific research use, and has been validated in UK men over 50 and women over 60. Use in younger patients or those outside England has not been validated.

The Lancet Digital Health paper's full citation: Cai, T., et al. (2026). Predicting the risk of serious muscle disorders in individuals eligible for statin treatment in England: derivation and validation of a clinical prediction model. The Lancet Digital Health. DOI: 10.1016/j.landig.2026.101024.