By Sameer Azhak, MD, FACC
At 58 Robert was 4 years away from the age his dad died. Calcium score 340, LDL 165, he came in to be seen. I could feel the tension when I walked in to greet him. He still had his coat on. The printout was folded into quarters in his shirt pocket, and he had folded and unfolded it enough times that the crease had gone soft.
He had read the number to his wife over the phone from the parking lot of the imaging center. By the time he sat down across from me he had already done what most people do, which is search the number, land on a survival curve built from a population he has never met, and decide that 340 is a sentence with a date attached to it.
It is not. And the reason it is not is the whole subject of this letter. Robert is not one of the five patients I am going to walk you through today. He is every patient who has ever come into my office holding one of these reports.
I have been ordering coronary calcium scans for most of my career, and I have watched the same test do two different jobs depending on what the patient walks out of the office holding. Without context, the number is what they think about at two in the morning. With a plan around it, the same number can be the most useful piece of information they have ever received about their own body. Same scan, same score, different interpretation.
So let me do here what I do in the room. Five patients from my practice. Five scores. Five different plans. The patients are composites, with identifying details changed and the physiology kept honest.
What the number in Robert’s pocket actually counts
One paragraph on the picture itself, because it changes how the score lands. Think of arterial plaque the way you think about spackle. When the wall of an artery gets injured, the body patches it. Early on, that patch is wet spackle — soft, lipid-rich, unfinished, easy to knock loose. Over years, the body cures the patch. Calcium moves in, the patch hardens, and what was soft becomes stone. What the scanner sees is a repair reaching completion.
The calcium scan photographs the cured spackle. The scanner counts every coronary speck at 130 Hounsfield units or higher, multiplies the area of each speck by a weighting factor tied to its peak density, and adds it all up. That sum is the Agatston score, and the density weighting runs from 1 for a lesion in the 130 to 199 range up to 4 for anything 400 and above. The scan takes about ten seconds of breath-hold, uses no contrast, and delivers roughly the radiation of a mammogram.
So the number in Robert’s shirt pocket is a count of finished repairs. Hold that. Everything below follows from it.
Frank at 1,247, and the stress test that lied
Frank is 78. Retired postal supervisor, widowed two years, lives alone in the house he raised his kids in. His score came back at 1,247. He had walked six minutes on a Bruce protocol treadmill test the week before, and the nuclear images from that study had been read as normal. He came to see me because he had noticed, over the previous month, that walking the empty garbage cans back up to the street left him more winded than it used to. Not chest pain. Just a fatigue at the end of the driveway that hadn’t been there in June.
Six minutes on a Bruce is a short study, and a normal perfusion scan in a 78-year-old with a calcium score above 1,000 is a combination I do not take at face value. Nuclear perfusion imaging compares blood flow across territories of the heart. It looks for a relative defect — one wall darker than its neighbor. When all three coronary arteries are narrowed to roughly the same degree, every wall dims together, the relative comparison finds no defect, and the study reads out as normal. Cardiologists call this balanced ischemia. It is a well-documented false-negative pattern in triple-vessel disease, and it is exactly the kind of miss a very high calcium score is designed to catch.
A Journal of Nuclear Medicine study looked at exactly this scenario — patients with normal SPECT imaging and a calcium score above 1,000 who went on to cardiac catheterization. The added diagnostic yield was substantial, and the mechanism was not always balanced ischemia in three-vessel disease. Sometimes a single vessel was severely narrowed and the nuclear study had simply missed it. For me, a normal stress test in a patient with a 1,247 is not an ending. It is the next question.
I laid the two paths out for Frank in the office. One was a CT angiogram, non-invasive, contrast in a peripheral vein, a definitive answer for whether obstructive disease was there. The other was a diagnostic catheterization, invasive, definitive for both anatomy and treatment in the same setting if a lesion needed a fix. He chose the cath. He wanted the answer and, if there was a fix, he wanted it in the same visit. He had three-vessel disease. He is now three months out from bypass surgery and back to walking the trash cans up the driveway without stopping.
The calcium score and the stress test ask different questions. The stress test asks whether the heart muscle is short of blood right now. The calcium score asks how much repair work is written into the walls of the arteries. When the two disagree, the calcium score is often the one telling you the truth.
Tony walked thirteen minutes
Tony is 65. Two hundred and thirty pounds at five feet nine, built like a man who has been in a gym six days a week for forty years. Retired firefighter, part-time strength coach. He has been on testosterone for twelve years and used anabolic steroids in his forties. His calcium score came back at 1,159, which arrived in his inbox at 6 AM and cost him his morning.
The number, in isolation, put him in the same statistical bin as Frank. In the MESA analysis, a score of 1,000 or higher carried a 4.7-fold adjusted risk of cardiovascular events compared with a zero. Treated from the number alone, Tony would be on maximum medical therapy, a low-dose aspirin conversation, and a rescan schedule.
Then Tony walked thirteen minutes on the same Bruce protocol Frank had done six minutes of. Thirteen minutes on a Bruce is a full-length study. The nuclear imaging was normal. No perfusion defect, no post-stress wall motion change, ejection fraction preserved through recovery. In most practices Tony’s evaluation would have stopped there. I sat down with him and laid out the same two paths I had offered Frank. He knew his exercise capacity, and he knew that a 1,159 with a decade of steroid history was not a number he was going to sleep on. He chose the cath. The coronaries showed mild luminal irregularities. Diffuse calcification of the arterial wall. No narrowing that would justify a stent, no lesion that would justify a bypass.
Two facts about Tony matter. First, anabolic steroids and long-term supraphysiologic testosterone are associated with accelerated arterial calcification independent of the usual risk factors. His 1,159 was, in part, a signature of his pharmacology, not solely a signature of his cholesterol. Second, his exercise capacity was extraordinary and his nuclear study was a full-length, well-powered negative. Balanced ischemia is a threat when the stress test is short and inconclusive. It is a much smaller threat when the patient walks thirteen minutes and the images are clean.
Tony still owned every finished repair on that scan. His plan is aggressive lipid management, tightening of his testosterone protocol in coordination with the physician who prescribes it, and a rescan schedule that lets us watch how the number moves without treating him as if he were about to have an event tomorrow.
Frank and Tony had almost the same number and almost the same reading on their nuclear studies. Their catheterizations found nearly opposite anatomy. Tony’s story argued against the aggressive interpretation the number alone would have earned him.
Marcus, the zero, and the stent he needed anyway
Marcus is 42. Six feet, three hundred and ten pounds. Type 2 diabetes for four years, blood pressure 148/92 on two medications, LDL 113, triglycerides 240, waist circumference the size of a truck tire. Metabolic syndrome, by every classical definition. He had seen an advertisement on the radio for a $99 calcium score at a local imaging center, self-referred, paid out of pocket, and had the scan done on a Saturday morning. It came back zero. The report landed in his inbox with a one-line note that the study was normal.
He came to see me two weeks later because his zero had not made the other thing go away. For the past month he had been getting a pressure in his chest, mid-sternum, when he walked up the two flights of stairs to his office. It went away by the time he sat down at his desk. He had not told his wife. He had not told his primary care doctor. He told me because his cousin, who was his age and heavier than he was, had died of a heart attack four months earlier, and because the zero on his printout and the pressure in his chest did not fit together in a way he could reconcile alone.
He was right that they did not fit. He was wrong about which one to trust. If the calcium score alone had guided the plan, Marcus would have walked out of that imaging center reassured. He had none of the finished repairs the machine was built to detect. His arteries, from the scanner’s point of view, looked clean.
A zero does not cover soft plaque. The scanner cannot see it. Soft plaque is the wet stage, before the calcium arrives, when the cap is thin and the core is lipid-rich and the lesion is the kind most likely to rupture and cause a heart attack. In a 42-year-old with metabolic syndrome, an LDL of 113, and new exertional symptoms, the pretest probability of soft plaque is high, and a calcium score is the wrong tool to rule it out. In the Western Denmark Heart Registry, among symptomatic patients age 40 or younger who turned out to have obstructive coronary disease on CT angiography, 58 percent had a calcium score of zero. In a young symptomatic patient, soft plaque hides from the scanner often.
So Marcus got a stress test. It was positive. Anterior ischemia, exactly where the LAD supplies the heart. Catheterization showed a ninety percent narrowing in the mid-LAD, non-calcified plaque, the widow-maker distribution. He got a stent that afternoon.
Marcus’s zero was accurate. His arteries had not calcified any of the repair work yet. But the injury was there, the plaque was there, and the calcium score is not the test that finds it in a patient like him. The calcium score is a good test for asymptomatic risk stratification and a bad test for a symptomatic patient. No negative calcium score has ever bought a patient of mine an exit from a real symptom.
Marcus’s plan looks like every plan I write for a patient in his situation. Aggressive management of the metabolic syndrome — glucose, blood pressure, weight, sleep, inflammation, and lipids together, not one at a time. High-intensity statin. Dual antiplatelet therapy for the stent, aspirin and a P2Y12 inhibitor for the year the interventionalist and I agreed on. And a slower conversation, over the next year, about the deeper drivers of the injury that put a stent in the heart of a 42-year-old.
The zero was true. It just answered a question I had not fully asked.
Michael, 275, and a spreadsheet
Michael is 54. Real-estate appraiser, thin, runs three miles most Saturdays, LDL 168, HDL 51, triglycerides 118, blood pressure 132/84, hs-CRP 2.4, non-smoker, no diabetes. His primary care doctor calculated a 10-year PREVENT risk of 6.2 percent, which is intermediate, and the two of them decided a scan would settle the question. It came back at 275.
Michael has never taken a statin. He came in with a printout, a spreadsheet of his last five lipid panels, two weeks of home blood pressure readings taken the way I teach it — sit five minutes, rest and breathe, check three times in a row, log the lowest of the three — and a specific ask, which was to be talked out of medication. That home BP method, along with the rest of the natural-first approach to blood pressure, is written up in my book, Lower Your Blood Pressure, available on Kindle now with the paperback releasing October 20.
His 275 sits in the 100 to 299 band, and the 2026 ACC/AHA dyslipidemia guideline treats a score in that band as a marker of established atherosclerosis and moves the LDL target down accordingly. His percentile in the MESA reference tool is above the 75th for a 54-year-old man, which by itself crosses the same threshold. Two independent readings of his scan land in the same place.
Then I told him what the number meant in real terms. In the CAC Consortium follow-up, participants with a score in the 1 to 399 range carried roughly a twofold higher risk of coronary heart disease death compared with a zero after adjustment for the usual risk factors. That is Michael’s trajectory without treatment. The higher the risk you are starting from, the more a statin saves you. A 275 has more to gain from the same LDL reduction than a 25 does.
Michael pushed back where I expected him to. He runs, his weight is fine, and what he wanted to know was whether the running could carry the number the rest of the way. Lifestyle bends risk. It does not close a hundred-point LDL gap. The distance from 168 to where a 275 needs him to be is not a distance running has ever moved anyone.
We added the two together. He started a moderate-intensity statin, kept the running, added two days of resistance work, and we tightened the anti-inflammatory side of the plate. His three-month LDL came back at 71. His hs-CRP came down to 1.1. His real question was never about pills. His real question was whether the number meant he had a disease. It did. So we treated it.
A quick piece of history, because it changes the next patient
Before I introduce Ellen, I want to tell you where the calcium score came from, because the timing explains why the same number now means two different things.
In March of 1990, a Miami Beach cardiologist named Arthur Agatston and a radiologist named Warren Janowitz published a short paper in the Journal of the American College of Cardiology titled “Quantification of coronary artery calcium using ultrafast computed tomography”. The scanner they used was electron-beam CT, a machine fast enough to freeze the moving heart in an era when ordinary CT could not. Multidetector CT eventually replaced it, but the formula they proposed — area of each speck multiplied by a density factor, summed across all three coronary territories — survived intact. Every calcium score you have ever seen is still an Agatston score.
Lovastatin, the first statin the FDA approved, cleared in 1987. Agatston’s paper appeared in 1990. The scan that shows us plaque was invented into a world where the drug that stabilizes plaque had just barely arrived. For the first decade of calcium scoring, we were quantifying a disease process we could only partly treat. The aggressive treatment thresholds in the 2026 guideline are the product of thirty-six years of watching what happens to the number when a real drug meets a real repair. The scan is old technology. The interpretation is where the field has moved, and the interpretation is what the next patient’s story turns on.
Ellen, on treatment for fifteen years
Ellen is also 54. Similar build and roughly the same weekly mileage on the treadmill. Her first-ever calcium score came back at 275 too. She has been on atorvastatin 40 mg since her late thirties because her LDL ran above 220 in her twenties and a strong family history landed her under a lipid specialist’s care early. Her LDL on treatment is 62, her hs-CRP is 0.6, and her Lp(a) is 40 nmol/L.
Same number as Michael, same demographic, different plan.
The first thing I did was slow the room down, because a 275 on a treated patient looks like a treatment failure and it almost never is. I gave her the two sentences I give every long-term statin patient who gets a first-time positive scan.
One: your score is a count of the repairs your body finished. It is not a count of the repairs your body started.
Two: those repairs cured harder on a statin than they would have without one.
That is the statin paradox in one exchange. Statins pull lipid out of the soft core of a plaque and let the repair finish. Serial imaging studies show statin therapy associated with higher plaque density and lower low-attenuation and fibro-fatty plaque volume, and the Heinz Nixdorf Recall investigators documented the same accelerated calcification signal in statin-treated participants. The Agatston score, remember, weights each lesion by peak density. A patch that hardens without growing at all can score higher than the same patch would have scored while it was still soft. Ellen is not looking at fifteen years of runaway disease. She is looking at fifteen years of successful repair that the scanner can now see because it has cured to the density the machine was built to detect.
I do not tell Ellen the calcification itself is what protects her. I tell her what I can defend at the bedside. Her LDL is 62. Her hs-CRP is 0.6. Her Lp(a) at 40 nmol/L sits on the reassuring side of the risk cutoff. The number on her scan is a picture of repairs her body finished on a drug that helped them finish. Her clinical risk profile is what drives her plan, and by that profile she is exactly where I want a 54-year-old with her genetics to be.
Michael and Ellen have the same score and different risks. Michael’s 275 is a snapshot of untreated disease. Ellen’s is a snapshot of treated disease. The scanner does not distinguish between the two, so the physician has to.
For the untreated patient the same number means start. For the treated patient it means do not confuse the density of a successful repair with the failure of one.
Ellen kept her atorvastatin. Ellen’s plan is Ellen’s plan: an NMR lipoprofile to resolve the LDL into particle number and size, an eye on the Lp(a) where it already sat on the reassuring side of the cutoff, and continuation of the work that had brought her hs-CRP to 0.6 — anti-inflammatory eating, sleep, and stress management alongside the drug. Her plan was written in her thirties. Her scan confirmed it.
What none of the five scores said
Every one of those numbers was silent on the same four things, and I would rather say them plainly than let a reader discover them the hard way.
The score does not measure narrowing. It counts calcium. Blockage is a separate question with separate tests. Frank had a very high score and severe blockage. Tony had a very high score and almost no blockage. Same number, different angiograms.
The score does not measure soft plaque. It is blind to exactly the lesions most likely to rupture, which is why Marcus’s zero did not close the question in a symptomatic 42-year-old with metabolic syndrome. A zero from an asymptomatic 60-year-old with good numbers is a strong negative finding. A zero from a symptomatic 42-year-old with metabolic risk factors is the wrong test at the wrong time.
The score does not overrule symptoms. A zero in an asymptomatic person is powerful. A zero in a person with chest pain or exertional dyspnea is useful but not absolute. In a pooled analysis, a zero calcium score ruled out obstructive disease in 97 percent of patients with stable chest pain and 98 percent with acute chest pain. Those are strong numbers. They are not 100. Symptoms get evaluated on their own terms.
And the score does not tell you what caused it. The number counts the repairs. It says nothing about what kept injuring the wall in the first place. That answer comes from a blood draw and a conversation: LDL and apoB, Lp(a), blood pressure, fasting insulin and A1c, hs-CRP, sleep, and how much of the day is spent in a low-grade inflammatory state. The calcium score is where the workup starts, not where it ends. It tells me the fire has been burning. It does not tell me where the pilot light is.
Before you close this
The score counts the patches, not the leaks. It does not tell you whether the artery you are looking at is a Frank or a Tony, whether a zero is a Marcus, or whether a 275 is a Michael or an Ellen.
Five numbers, five plans, one test that was never designed to deliver a verdict.
What I am after in every one of these plans is the same three things. Calm the fire. Shrink the spackle bucket. And if there is a lot of sticky spackle, shrink the bucket a lot and the fire absolutely has to be out.
If you have a score, look up two things next to it: your absolute number and your percentile for your age and sex. A zero with symptoms does not answer the symptom. A first-time positive scan on a long-term statin is not a treatment failure until Ellen has been read again. And a score above 1,000 with a normal stress test is a set of results that only makes sense when you look at all of it together, the way I looked at Frank’s.
A few questions patients ask me. Can I lower my score? Not reliably; stabilizing what it represents is the goal. Should I rescan? Not for at least 3 to 5 years, and only if the result would change what we do. Does a high score mean I need a stress test? Not by itself, and not in the absence of symptoms.
If you are sitting with a folded printout in your pocket right now, the feeling you are having is the correct feeling, and it does not have to last. It usually takes about two visits for a frightening number to become a working number.
There is a younger healthier version inside of you and we are going to set them free together.
I will be here when the numbers come back.
Sameer Azhak, MD, FACC is a board-certified cardiologist with fellowship training in integrative medicine through the Andrew Weil Center. He writes The Integrative Cardiologist and its sister publication, The Menopause Heart. Reader mail reaches him at hello@theintegrativecardiologist.com.
This article is for educational purposes only and does not constitute medical advice or create a physician-patient relationship. Please consult your own physician before making changes to your treatment.