99% of Heart Attacks Had This Prior Warning

On July 20, 1984, Jim Fixx — the man whose bestselling book on running had inspired millions of Americans to get in shape — dropped dead of a massive heart attack during his daily run on a quiet Vermont road. He was 52. He had quit smoking. He'd lost 60 pounds. He ran 10 miles a day [1].

It's the kind of story that makes all of us quietly wonder: is it possible to do everything right and still have a heart attack come out of nowhere?

And for a long time, the research seemed to say: yes. Between 11% and 23% of the most severe heart attacks occur in people with none of the standard documented risk factors — no high blood pressure, no high cholesterol, no diabetes, non-smoker. That number was alarming enough that cardiologists gave it a name: the "SMuRF-less" phenomenon. SMuRF stands for Standard Modifiable Risk Factors.

But a major new study of more than 9 million people just dismantled what we thought we knew. And the way they did it reveals something important about how we've been approaching heart attack risk all along.

Table of Contents

• The SMuRF-Less Claim
• The New Study
• The Methodological Reveal
• Jim Fixx, Revisited
• The Empowerment Reframe
• The ApoB Problem
• The Honest Caveat
• Final Thoughts
• References

The SMuRF-Less Claim

So first, let's take the scary finding seriously. Because it comes from real research, published in credible journals, and it has genuinely alarmed the cardiology community.

In 2023, researchers pooled data from 15 studies covering more than 1.28 million patients who had experienced an acute coronary syndrome. What they found was that 11.6% of those patients had none of the four standard modifiable risk factors: no diagnosed hypertension, no diagnosed high cholesterol, no diabetes, and no smoking history [2].

In Australia specifically, that number climbed as high as 23% [3].

And there's a troubling twist that made these findings even harder to ignore: the patients who had no documented risk factors didn't do better after their heart attacks. They did worse. Short-term death rates were higher in the SMuRF-less group than in patients who had the traditional risk factors [2].

The working hypothesis was that these patients were flying under the radar of preventive medicine. They weren't being treated with statins or antihypertensives because they didn't appear to need them. And so they arrived at the hospital in worse shape.

The implication of all this was uncomfortable. If up to nearly 1 in 4 of the most severe heart attacks happen in people with no detectable risk factors, maybe our entire prevention framework was missing something. Maybe heart attacks really could come out of nowhere.

It's an idea Jim Fixx explicitly rejected. He was confident that if a person didn't smoke and exercised enough, heart attacks were essentially ruled out [1]. Sadly, his own story proved he was wrong. But was Fixx an unlucky outlier — or was something hiding in plain sight?

The New Study

Recently, a team led by Dr. Hokyou Lee at Yonsei University in Seoul decided to approach the question of SMuRF-less heart attacks from the opposite direction.

Instead of looking at which risk factors patients had documented at the time they arrived in the emergency room, they looked backwards — tracking people through years of repeated health examinations before any cardiovascular event occurred.

The scale of this study is remarkable. They drew on two independent cohorts: the Korean National Health Insurance Service database — which covers essentially the entire South Korean population — and the Multi-Ethnic Study of Atherosclerosis, a long-running American cohort. Together, that's more than 9.3 million people followed for up to 13 years in Korea and 19 years in the US. And across those years of follow-up, over 600,000 cardiovascular events — like heart attacks and strokes — occurred [4].

For each person who had a cardiovascular event, the researchers went back through all their available health data — every recorded blood pressure reading, every cholesterol measurement, every glucose test, every documented smoking history — and asked: did this person ever have a nonoptimal level of any of these four risk factors, at any point in the years before their event?

And here's what they found.

More than 99% of people who had a first heart attack had at least one nonoptimal risk factor beforehand. Consistent across both countries. Consistent in men and women. Consistent across age groups from under 60 to over 80 [4].

And it wasn't just one risk factor hiding in the background. More than 93% had at least two [4].

So we have two bodies of research telling us opposite things. One says up to 23% of serious heart attacks happen in people with no risk factors. The other says more than 99% had at least one. Both published in top journals. Both using large datasets. How can they possibly both be right?

The Methodological Reveal

This is where the detective work gets interesting. Because the disagreement between these two findings isn't really a scientific contradiction — it's a methodological one. The two sets of studies are answering different questions.

The standard SMuRF-less research asks: "Did this patient have a clinical diagnosis of a risk factor?" That means: were they ever told by a doctor that they had hypertension? Was "high cholesterol" ever written in their chart? Did they have a formal diabetes diagnosis?

That sounds reasonable. But there are two massive problems with it.

The first is underdiagnosis. Getting a clinical diagnosis requires that you've seen a doctor, been screened, and had a threshold crossed that prompted a formal label. In the real world, that often doesn't happen. Around half of those with hypertension — the single most common modifiable risk factor for cardiovascular disease — haven't been diagnosed [5].

Low diagnosis rates for high cholesterol and diabetes are also well documented. So if your blood pressure has been quietly running at 138/88 for a decade but you've never had it flagged, you show up in the SMuRF-less data as "no hypertension." But you did have it.

The second problem relates to the thresholds researchers select to define risk. Cardiovascular risk doesn't have an on/off switch. These risk factors operate on a continuous, cumulative gradient — meaning damage accumulates even at levels that fall below the clinical diagnosis threshold.

A blood pressure of 126/82 is not "clinical hypertension" by any guideline. But it is elevated above what the American Heart Association considers optimal. It can promote plaque buildup in the arteries [6]. And if you've had it for 15 years, that's 15 years of cumulative vascular damage that won't show up in a SMuRF-less study because you were never officially diagnosed.

In fact, the thresholds used in the SMuRF-less literature have often been high by current standards. One study in the Journal of Cardiology defines hypertension as a blood pressure of 140 or above [7].

This is exactly why the new study used stricter cut points derived from the AHA's Life's Essential 8 framework — the targets associated with ideal cardiovascular health, not just the thresholds that trigger clinical treatment. Nonoptimal blood pressure was defined as systolic ≥120 mm Hg or diastolic ≥80 mm Hg. Nonoptimal cholesterol as total cholesterol ≥200 mg/dL. Nonoptimal fasting glucose as ≥100 mg/dL. And past or current smoking counted — not just current [4].

Dr. Lee put it plainly. He said that when they looked closely, those earlier reports of risk-factor-free heart attacks "might actually reflect missed risk factor diagnoses or risk factor levels that were nonoptimal in terms of cardiovascular risk, since many of these risk factors have a continuous rather than a yes-or-no relationship with risk of developing CVD" [8].

And when asked about the striking results, Lee pointed to the consistency: "Surprisingly, the results were almost identical between the Korean and US cohorts. That consistency across very different populations and health systems makes our findings especially strong" [8].

The two literatures aren't contradicting each other. They're answering different questions. "Was the patient diagnosed?" versus "Did the patient have the exposure?" Different questions. Very different answers.

Jim Fixx, Revisited

Which brings us back to Jim Fixx. Because his story turns out to be almost a perfect case study in this exact methodological gap.

A SMuRF-less study conducted at the moment of his death might have coded him as risk-factor-free. He had quit smoking 17 years earlier, so not a current smoker. He wasn't a documented diabetic, nor formally hypertensive by clinical record. It looked like a clean chart, at least with 3 out of the 4 standard risk factors.

But apply the new study's methodology — look backwards, longitudinally, with the stricter cut points — and the picture is completely different.

Fixx was smoking two packs of cigarettes per day when he took up running at age 35. And he weighed 214 pounds — significantly overweight by any standard [9].

His father had died of a heart attack at 43, after a first heart attack at 35 — a family history of premature coronary disease that should have put him in a high-risk category [9].

Then there was the fourth traditional risk factor. His cholesterol was above 250 mg/dL — well above the 200 mg/dL nonoptimal threshold, and elevated even by clinical diagnosis standards [10].

And his autopsy confirmed what all of this might have predicted. One coronary artery was 95% blocked. A second was 85% blocked. A third, 70% [9].

And there's one more detail that makes Fixx's story particularly striking. In December 1983 — seven months before his death — he visited the Cooper Aerobics Center in Dallas. Dr. Kenneth Cooper, the pioneering exercise physiologist and Fixx's personal friend, urged him to take a cardiovascular stress test.

Fixx refused [11].

Cooper was haunted by the refusal. He told reporters afterward: "He adamantly refused. That stands out in my mind now. I don't know why he refused. Maybe he had an underlying feeling he was ill" [12].

Friends and family had also reported that Fixx complained of chest pains during runs in the months before his death — what physicians reviewing his case later suggested may have been angina. He told his family that he felt a tightness in his throat while running, which was probably a telltale sign of coronary trouble [11].

And Fixx's own journal entries, recovered after his death, reveal the denial at work. That spring — just months before the fatal run — he wrote: "My neurotic blood pressure anxieties, occasioned by feeling tense these past two months are, apparently without foundation. Maybe the problem is nothing more than hypochondria" [13].

A cardiologist commenting on the case at the time put it bluntly: "The second most common symptom of coronary disease, after angina, is denial" [11].

Fixx had at least three silent heart attacks in the weeks before the fatal one, according to his autopsy [1].

Jim Fixx did not lack risk factors. He lacked detected risk factors — and he actively avoided the medical engagement that would have found them. His story is not about the failure of healthy living. It's about what happens when you assume that the visible transformation — the weight lost, the cigarettes quit, the miles run — is the whole story. And when you decline to look underneath.

The Empowerment Reframe

Here's what genuinely struck me about this data. Because I think the correct interpretation is empowering, not frightening.

If cardiovascular events are nearly always preceded by detectable, nonoptimal risk factors — if the 99% number holds across 9 million people in two different countries — then the scenario of a heart attack truly coming out of nowhere, in someone with long-term optimal risk factor levels, is vanishingly rare. Not common. Not increasing. Rare.

That's good news. It means the framework for prevention actually works, when it's applied correctly and early enough.

The key word there is early enough. The damage from nonoptimal risk factor levels accumulates over years and decades — the new study's data proves this by tracking people longitudinally rather than catching them at the moment of crisis.

As Nathan Wong, who directs the heart disease prevention program at UC Irvine, put it when commenting on this study: "The problem is that we practice reactive medicine rather than preventive medicine" [8].

That's the real villain in this story — not heart disease itself, but the way clinical medicine waits for risk factors to cross a treatment threshold before acting. The model is structurally too late.

The goal isn't to get your blood pressure below 140. The goal is to keep it optimal — below 120/80 — for as long as possible. For older adults, we may accept higher blood pressure readings to make sure that they don't faint or feel dizzy. It's not to get your LDL-cholesterol below 100; in my view, it's to get it below 55. To manage your weight and blood sugar levels. Use medications like Tirzepatide if needed.

The targets I aim for in clinic:

• Blood pressure: below 120/80
• LDL cholesterol: below 55 mg/dL
• Manage insulin resistance and weight

It's a higher bar than most clinical practice currently aims for. But it's the bar the data actually supports.

The ApoB Problem

But there's a catch — because even if your LDL-cholesterol looks fine on that standard blood test, there's a reason it might be misleading you.

LDL-cholesterol <55 mg/dL is a real threshold, and it captures a lot. But it's a blunt instrument. What actually drives atherosclerosis — the plaque buildup in your arteries that eventually causes a heart attack — isn't cholesterol mass. It's the number of atherogenic lipoprotein particles. Each particle that can penetrate the arterial wall and deposit cholesterol is carried by one molecule of apolipoprotein B, or ApoB. ApoB is the direct count of those particles.

The problem is that the standard test most doctors order — LDL-C — measures the cholesterol content carried by those particles, not the number of particles themselves. And those two things can diverge significantly, particularly in people with metabolic dysfunction, insulin resistance, or elevated triglycerides. You can have an LDL-C that looks fine while your ApoB remains elevated — meaning your atherogenic particle burden is still high, your risk is still elevated, and you and your doctor might not know it [14].

Now, clinical medicine has been slow to adopt ApoB testing because of concerns around cost-effectiveness. Adding ApoB roughly doubles the cost of a standard lipid panel. So is it worth routinely measuring something beyond the standard lipid panel?

That argument just lost its last leg. A study just published in JAMA modeled 250,000 statin-eligible adults across their lifetimes, comparing three strategies for deciding when to intensify lipid-lowering therapy: using LDL-C targets, using non-HDL-C targets, or using ApoB targets. The ApoB-guided strategy produced the most quality-adjusted life years and was highly cost-effective [15].

The 2026 ACC/AHA dyslipidemia guideline — the most recent update — has now formally incorporated ApoB into its recommendations for the first time in a major US guideline, particularly for patients where LDL-C may underestimate true risk [16].

At the clinic, I recommend to my patients that they have their ApoB tested. The cost-effectiveness case for it is now established. And it tells us something the standard lipid panel may be missing.

The Honest Caveat

I want to be clear about one thing. The new study tracked four risk factors: blood pressure, cholesterol, blood glucose, and smoking history. It did not assess family history, Lp(a), chronic inflammation, sleep apnea, or other nontraditional contributors to cardiovascular risk.

That under-1% of cardiovascular events that occurred without any of those four nonoptimal risk factors likely includes people with elevated Lp(a) or strong family histories of premature disease — drivers the study simply wasn't designed to measure. Those contributors are real, and if you have a family history like Jim Fixx's, they warrant more aggressive monitoring regardless of how your traditional numbers look.

This is precisely why the data argues for individualized medical assessment, not just lifestyle management. Knowing your numbers — all of them, including ApoB and ideally Lp(a) — and having a physician who understands your personal risk profile is the mechanism through which longitudinal monitoring actually works. A coronary artery calcium score, in the right patient, can reveal subclinical atherosclerosis that no blood test would catch.

Regular checkups aren't a bureaucratic checkbox. They're the clinical equivalent of what the new study did: looking backwards through your history to find what's there, before it becomes an event.

Final Thoughts

Jim Fixx was not unlucky. He was unmonitored — and he chose to stay that way.

Kenneth Cooper spent the rest of his career haunted by a single thought: he had the chance to force the issue, and he didn't. But the data from 9 million people tells us the same thing Cooper already knew — the fingerprints were there. They were always there. In Fixx's case, in the SMuRF-less patients, in almost every heart attack and stroke and heart failure event that followed.

The real story is not randomness, not bad luck, not a disease that strikes without warning. It's that we weren't looking — or we were looking with the wrong tools, at the wrong thresholds, at the wrong moment.

References

1. https://pmc.ncbi.nlm.nih.gov/articles/PMC7237076/

2. https://pubmed.ncbi.nlm.nih.gov/36179904/

3. https://pmc.ncbi.nlm.nih.gov/articles/PMC6898813/

4. https://www.jacc.org/doi/10.1016/j.jacc.2025.07.014

5. https://pubmed.ncbi.nlm.nih.gov/34450083/

6. https://www.ahajournals.org/doi/10.1161/01.hyp.34.1.51

7. https://www.journal-of-cardiology.com/article/S0914-5087(23)00286-1/fulltext

8. https://www.tctmd.com/news/risk-factors-present-nearly-all-patients-who-develop-cvd

9. https://en.wikipedia.org/wiki/Jim_Fixx

10. https://www.recordonline.com/story/sports/2004/08/25/runners-heed-cholesterol-counts/51131216007/

11. https://time.com/archive/6855620/medicine-why-joggers-are-running-scared/

12. https://www.upi.com/Archives/1984/07/23/Running-guru-Jim-Fixx-who-collapsed-and-died-of/2268459403200/

13. https://www.si.com/track-and-field/2020/05/21/jim-fixx-legacy-running-coronavirus

14. https://www.ahajournals.org/doi/10.1161/CIR.0000000000001423

15. https://jamanetwork.com/journals/jama/fullarticle/2847303

16. https://www.hcplive.com/view/expert-insights-lp-a-and-apob-in-the-2026-dyslipidemia-guidelines