Two large studies published this year suggest the same handful of numbers your doctor already collects predict far more than heart attacks. One links them to cancer. The other puts a figure on what they cost you in years of clear thinking.
If you are somewhere between 40 and 75 and have had a physical in the last few years, your doctor almost certainly ran a number on you that you may never have seen. It is your ten-year risk of a heart attack or stroke, and it comes from a calculator built into the electronic health record, most often the Pooled Cohort Equations behind the American Heart Association’s risk estimator. It uses nothing exotic: your age, sex, race, blood pressure, cholesterol, smoking status, and whether you have diabetes. Cross a threshold of roughly 7.5%, and the conversation turns to statins.
Two studies published this year suggest that number, and the ingredients behind it, tell us about a good deal more than the heart. One asks whether the calculator can predict cancer. The other asks what the same risk factors do to the years you live with your mind intact. Together they make a case that the heart risk score is the most underused piece of information in American primary care.
A gap in how we handle the biggest killers
Cancer and cardiovascular disease together account for around half of all deaths worldwide, and they share a long list of risk factors: smoking, excess weight, inactivity, alcohol, deprivation. Yet the health system treats them in almost opposite ways.
For heart disease, the approach is predictive. A risk score estimates what might happen years from now, and if the number is high enough, the conversation turns to statins, blood pressure treatment, and lifestyle change, before anything has gone wrong.
For cancer, the approach is detective. Screening programs exist for only a handful of cancers (cervical, colorectal, breast, and lung for heavy smokers), and they look for disease that is already present, just early. That finds people who need treatment, but it does not identify healthy people heading toward a diagnosis who might still avoid one. Cancer risk models do exist, notably a tool called QCancer, but surveys show they are rarely used.
This matters more than it used to. In both datasets the cancer researchers examined, more people developed cancer than cardiovascular disease over ten years: 11.7% versus 5.9% in one, 17.8% versus 6.7% in the other. Cancer has overtaken heart disease as a cause of death in many wealthy countries. Prevention has not caught up.
What the cancer study did
A team at University College London, publishing in the European Journal of Cancer, took four cardiovascular risk models in routine clinical use (the Pooled Cohort Equations in the US, QRISK3 in the UK, and SCORE2 and SCORE2-OP across Europe) and tested how well they predicted fifteen types of cancer over ten years. They used data from more than 500,000 UK Biobank participants, then repeated the exercise in anonymized primary care records covering 4.8 million people. Because the models were never built for this, they were recalibrated so their predicted percentages line up with reality.
For their original job, the heart models scored around 0.72 to 0.74 on a measure statisticians call the c-statistic, where 0.5 is a coin flip, and 1.0 is perfect. For “any cancer,” they scored 0.63, modest but close behind the purpose-built QCancer model at 0.65.
The average hides the useful part. For several specific cancers, the repurposed models did as well as they do for heart disease itself, scoring 0.70 or higher: stomach and esophagus, liver and bile ducts, larynx, kidney and bladder, and lung. Most of these have no national screening program. Fed a deliberately meaningless outcome, accidental injury, the models scored barely above chance, confirming they were not simply detecting general frailty.
After age, the two most influential ingredients were smoking status and systolic blood pressure. Smoking is no surprise. Blood pressure is more subtle, and the authors are careful: trial and genetic evidence shows that lowering blood pressure does not lower cancer risk. It is a marker traveling alongside the real causes, such as inactivity, alcohol, and weight.
And then there is your brain
The second study comes at the same idea from a different direction, and with American data. Roughly 42% of US middle-aged adults may develop dementia at some point in their lives, and with no disease-modifying treatment worth the name, prevention is most of what we have. Researchers using the Atherosclerosis Risk in Communities study followed 12,409 adults from four US communities for a median of 26 years, publishing their results in Neurology Open Access.
They did not use a risk score. They simply counted how many of three things a person had in midlife, at an average age of 56: high blood pressure, diabetes, and current smoking. Then they asked not how likely dementia is, but how long a person lives before it arrives.
The gap is startling. People with none of the three lived an average of 30.1 years dementia-free from age 55, taking them to about 85. People with all three averaged 17.5 years, taking them to about 73. That is a difference of 12.6 years of cognitively intact life, attached to three measurements that are already in your chart, two of which feed directly into the heart calculator.
One finding needs careful handling. By age 95, the group with all three risk factors had a lower cumulative rate of dementia, 23%, than the group with none, at 42%. That is not protection. It is competing mortality: they were 5.6 times likelier to die before dementia could develop, even though their hazard of dementia itself was 2.69 times higher. Avoiding dementia by not living long enough to get it is not a plan.
What makes the study useful is the unit of measurement. Percentages are abstract; years are not. Telling someone that treating their blood pressure and quitting smoking is worth a decade of additional life with their memory intact is a different conversation from quoting them a 12% risk. The differences also fall along familiar lines: among people with all three risk factors, women averaged 18.1 dementia-free years to men’s 16.6, and Black participants 16.0 against 19.6 for White participants.
What could actually change
Put the two studies side by side, and the heart risk score stops looking like a cardiology metric and starts looking like a general-purpose forecast. The infrastructure is already there. Cardiovascular risk calculators are embedded in essentially every major electronic health record system in the country, and the Pooled Cohort Equations, the American model, held up particularly well in the cancer analysis. No new test, blood draw or appointment is required. The UCL researchers have put a free web version of their recalibrated models online.
Three things become possible. First, a single conversation could cover all three risks at once, which is far more persuasive than any of them alone. “This lowers your chance of a heart attack, lowers your risk of several cancers, and is worth years of clear thinking later on” lands differently from a lecture about cholesterol. Second, a high score could trigger the same structured referrals that heart risk already does, to exercise or weight management programs. Third, screening could be aimed better: lung CT scanning is currently offered by age and smoking history but could be targeted by individual risk, and expensive multi-cancer blood tests become far more cost-effective when aimed at people likelier to have something to find.
The honest caveats
A score of 0.63 for cancer overall is a population-level sorting tool, not a diagnosis. It will misjudge plenty of individuals, and a reassuring number should never override a worrying symptom. The cancer study covered ages 40 to 84, leaving out young-onset cancers, which are rising and biologically different, and its data is British. Smoking and blood pressure do not respect borders, but US populations differ enough that the models need checking here.
The dementia findings carry their own limits. The study is observational, so it cannot prove that treating blood pressure at 55 buys back those 12.6 years, only that people who arrive at 55 without these risk factors go on to live far longer with their cognition intact. Just 176 of the 12,409 participants had all three, which makes that estimate the least precise in the paper. And a simple count of three conditions is not the same instrument as the Pooled Cohort Equations. Most importantly, for both papers, nobody has yet run a trial showing that using risk scores this way leads to fewer cancers, less dementia, or longer lives.
The bottom line
Neither study proposes a new technology or a new test. Between them, they point out something better: a calculation the health system already performs, millions of times a year, carries information about the two diseases people fear most, and the same short list of midlife numbers shapes not just how long you live but how many of those years you spend as yourself.
References
