August 30, 2026 · Science · Biomarkers · Men's Health
What the Referral Trial Changed about Cholesterol
Sending men with prostate cancer to a cardiologist lowered their cholesterol and did not lower their rate of heart attacks. The RADICAL PC 2 trial randomized 2,501 men either to a referral to an internist or cardiologist or to usual care, then followed them for nearly six years. Statin use reached 63% in the referred group against 40% under usual care, and blood pressure medication reached 56% against 49%.
Cholesterol came down by a statistically significant margin in the referred group. Cardiovascular death, heart attack, stroke and heart failure did not differ between the two groups. "I was surprised that we didn't see a reduction in heart attacks," said Darryl Leong, the trial's principal investigator. Leong is a senior scientist at the Population Health Research Institute, a joint institute of Hamilton Health Sciences and McMaster University.
Enrollment ran from 2015 to 2025 across eight countries: Canada, Brazil, Singapore, Colombia, India, Australia, Israel and the United States. Participants were men with newly diagnosed prostate cancer or men starting androgen deprivation therapy for the first time. The results appeared in JAMA Internal Medicine and were presented at the European Society of Cardiology Congress on August 30, 2026. A companion paper ran in JACC CardioOncology, according to MedicalXpress.
The short version
Why Cholesterol Fell and Events Did Not
The gap between the two results has a mechanical explanation that does not require the treatment to have failed. Cardiovascular events are counted, and a trial can only detect a difference in them if enough of them happen. Leong states that fewer participants experienced these complications than expected, which he describes as making it harder to determine whether the referral strategy affected those outcomes. Cholesterol, by contrast, is measured in everyone at every visit, so a difference in it shows up with far less data.
The intervention itself was also diluted by design. This trial randomized access to a specialist rather than a specific therapy, and 40% of the usual care group ended up on a statin anyway. The contrast being tested was therefore 63% treated against 40% treated. Neither arm was untreated. A referral strategy that moves a quarter of a population onto a drug is a weaker intervention than the drug itself, and it dilutes any event difference the drug would produce.
Leong did note a signal suggesting a reduction in heart attacks that he described as consistent with the cholesterol-lowering effect. A signal of that kind does not carry the weight of a result. The trial's stated conclusion is that no statistically significant reduction in major cardiovascular events was found, and that is the finding that stands until a larger or longer study says otherwise.
Why a Cholesterol Result Is Only Half the Sequence
Our guides keep returning to the distance between a marker that moves and an outcome that changes, and this trial is the cleanest illustration of it we have covered. A biomarker panel sells the first half of the sequence. It finds the abnormal value, and finding it is genuinely useful, because the men in the referred arm did end up on treatment their usual care counterparts did not get. What a panel cannot deliver on its own is the second half, where the treated number turns into a person who does not have a heart attack. Our page on normal versus optimal lab ranges makes the same point from the reference-range side, and our ApoB guide covers why particle count separates people whose standard cholesterol panels look alike.
The commercial version of this gap is worth naming plainly. A subscription that reports a hundred markers and hands back a dashboard has completed the cheap half of the work. The expensive half is a clinician who decides what to treat, prescribes it, and follows the person for years. RADICAL PC 2 put resources behind exactly that second half, with a referral to a specialist rather than a report, and six years later the event count still had not separated. That should temper what anyone expects from a panel alone.
Who Should Not Read This as a Reason to Stop Treatment
Anyone currently taking a statin or a blood pressure medication should treat this trial as background and change nothing on the strength of it. The trial did not randomize men off their medication, and it reports no comparison that speaks to stopping one. Statin benefit in populations at cardiovascular risk rests on a body of randomized evidence that a single underpowered referral trial does not displace.
Men on androgen deprivation therapy are the group with the most at stake here and the least reason to read this as reassurance. The therapy raises cardiovascular risk, which is the reason this trial existed. A null result on a referral strategy says the strategy did not prove itself in six years. It does not say the underlying risk went away. That conversation belongs with the oncologist and the physician managing the cardiovascular side. The page on this site covering hormone therapy risks sets out why hormonal treatment and cardiovascular monitoring get planned together.
What Would Change This Read
A longer follow-up would change it first. Cardiovascular events accumulate with time, and a cohort enrolled as late as 2025 has contributed only a fraction of the person-years the earliest participants have. An extended analysis on the same cohort could separate the curves that had not separated at six years, or confirm that they do not.
Published effect sizes would change how much weight the result carries. A trial that lowered cholesterol by a large margin and still produced no event difference is a stronger negative than one that moved it slightly. Those numbers exist in the JAMA Internal Medicine paper and are not in the summaries. The companion analysis in JACC CardioOncology reaches a narrower conclusion. It names men with blood pressure above 130/80, diabetes, or cholesterol above 4 mmol/L as those most likely to benefit. That account does not state which cholesterol measure the threshold describes, so the figure needs the paper itself before anyone reads their own result against it.
Related Coverage
Three briefings on this site now land on the same structural point. A Finnish cohort found that conventional risk factors outpredicted epigenetic clocks, and a federal guideline reviewed the body fat measurements and endorsed none of them. RADICAL PC 2 adds the randomized version: even acting on a risk factor correctly, through a specialist, did not produce a measurable outcome difference in this population within six years. Anyone weighing a testing subscription on cardiovascular grounds should ask what happens after the cholesterol result comes back, and who is going to follow it for the next decade.
Sources
- MedicalXpress, "Cardiologist referrals improve cholesterol levels in prostate cancer patients, but may not be necessary for all," August 30, 2026, medicalxpress.com (accessed August 30, 2026). Source of the 2,501 participants, the eight countries, the 2015 to 2025 enrollment and the nearly six years of follow-up. Also the source of the 63% and 40% statin figures, the 56% and 49% blood pressure medication figures, and the absence of a difference in cardiovascular outcomes. The Leong quotation and the companion study thresholds come from the same piece.
- JAMA Internal Medicine, RADICAL PC 2 primary results, published August 30, 2026, jamanetwork.com. Named as the publication venue by the release above. The full text was not retrievable at the time of writing, and no figure quoted on this page is taken from it.
- Population Health Research Institute, RADICAL PC trial programme, phri.ca (accessed August 30, 2026). Source of the trial's stated objective and design. The page carries no results.
Frequently Asked Questions
Does this trial show that lowering cholesterol does not work?
No, and reading it that way inverts what was tested. RADICAL PC 2 randomized a referral rather than a drug. Men were assigned to see an internist or a cardiologist, or to usual care, and the referral produced more statin prescriptions and lower cholesterol. The trial then found no statistically significant difference in cardiovascular death, heart attack, stroke or heart failure over nearly six years. Darryl Leong, the principal investigator, notes that fewer participants had those complications than expected, which makes the comparison harder to interpret rather than negative.
How many men were in the trial and where were they?
MedicalXpress reports 2,501 participants across eight countries: Canada, Brazil, Singapore, Colombia, India, Australia, Israel and the United States. Enrollment ran from 2015 to 2025, with follow-up of nearly six years. Participants were men with newly diagnosed prostate cancer or men starting androgen deprivation therapy for the first time.
What did the referral actually change?
Two things were reported in numbers. Statin use reached 63% in the referred group against 40% under usual care, and blood pressure medication reached 56% against 49%. Cholesterol was reported as lower in the referred group by a statistically significant margin, though no figure for the size of that difference appears in the coverage published so far.
Who did the companion study say benefits most?
A companion paper in JACC CardioOncology identified men with blood pressure above 130/80, diabetes, or cholesterol above 4 mmol/L as those most likely to benefit. The account naming those thresholds does not say which cholesterol measure the 4 mmol/L figure refers to, and no conversion to milligrams per deciliter appears in it either. Anyone applying that threshold to their own results needs the paper itself and a clinician reading it.
Does this apply to men who do not have prostate cancer?
Not directly. Every participant had newly diagnosed prostate cancer or was starting androgen deprivation therapy, a treatment that changes body composition and metabolic risk. The wider interest is structural rather than clinical. A randomized test of finding a risk factor and sending someone to a specialist moved the laboratory numbers. It did not move the event count in this population over this period.
Was the trial peer reviewed?
Yes. The results were published in JAMA Internal Medicine and presented at the European Society of Cardiology Congress on August 30, 2026, with a companion paper in JACC CardioOncology. That is a stronger evidence base than a conference abstract alone, because the full methods and statistics went through journal review before release.