Zero on the Scan: A Borderline Risk Score Against a Brother's Heart Attack
A 58-year-old man with a borderline pooled-cohort risk score and a calcium score of zero, weighed against a brother's heart attack four months ago — whether a single clean scan should outrank a family history the scan cannot fully see.
Daniel R., a 58-year-old high school shop teacher, spent this spring rebuilding a 1967 pickup truck with his advanced metalworking class — six hours a day on his feet, which is part of why his own health had slipped so far down his list of priorities that he hadn't seen a doctor in three years. What actually got him into clinic wasn't a symptom of his own; it was a phone call in March, when his younger brother, 54, had an anterior STEMI while shoveling snow. Daniel came in the following week with a checklist his brother's cardiologist had given him and a question he clearly hadn't expected to be complicated: was he next?
His labs came back with an LDL of 152mg/dL and a calculated 10-year pooled-cohort ASCVD risk of 6.4% — squarely inside the 5%-to-under-7.5% band the 2018 ACC/AHA cholesterol guideline built specifically because a score in that range doesn't reliably sort a patient into treat-or-don't on its own. He doesn't smoke, isn't diabetic, and his blood pressure has never needed medication, so the risk-enhancer conversation comes down almost entirely to his brother, whose own labs from the hospitalization — an LDL of 168mg/dL, no diabetes either — look enough like Daniel's own numbers that the resemblance is hard to wave off as coincidence.
A coronary artery calcium scan, ordered specifically to settle the ambiguity, came back at zero — a result that, read one way, is the most reassuring number he could have gotten, and read another way, is a test that was never built to see the kind of soft, non-calcified plaque that put his brother on a cath lab table at 54. The MESA cohort's own long-running data associate a zero score with a meaningfully low near-term event rate, often described as a multi-year 'warranty period' before the risk needs to be reassessed — but that warranty period was characterized in a general population, not one already carrying a first-degree relative's premature MI on top of it, and Daniel's clinic visit is really an argument about how much that difference should matter.
In clinic, a scan against a phone call
Hold the statin and repeat the scan in a few years, not treat today. A calcium score of zero is not the absence of information — it is a direct measurement of zero calcified atherosclerotic burden, and the MESA-derived literature on exactly this scenario associates a zero score with a roughly decade-long low-risk period even in patients who carry a risk-enhancing factor like his.
If his brother's event had happened to him rather than to a sibling, or if today's scan had shown even a modest score, I would not be making this argument — the whole case for deferring rests on the scan being genuinely clean, not on discounting the family history itself.
I don't dispute what the scan shows. What I dispute is treating it as the more informative of the two facts in this room. The 2018 guideline lists a family history of premature ASCVD as its own named risk-enhancing factor precisely because it captures something a single-timepoint scan structurally cannot: a calcium score only detects plaque that has already calcified, and the plaque that ruptures and causes an MI in a 54-year-old is disproportionately the softer, non-calcified kind.
A clean scan in a 58-year-old whose brother just had exactly this outcome four months ago isn't nothing, but treating it as license to walk away for five years asks a single test to settle a question it wasn't designed to answer on its own.
Both of you are arguing about how to weigh two imperfect signals against each other. There's a third test that could actually adjudicate instead: a one-time lipoprotein(a). It's substantially genetically determined, it's independent of LDL, and the ESC/EAS guidance specifically recommends checking it once in adults with a family history of premature disease — which is exactly the situation in front of us.
An elevated Lp(a) would give the internal medicine physician's position a concrete mechanism, not just a worry; a normal one would give real weight back to the clean scan. Either way, we'd be deciding on a third data point instead of arguing the first two past each other.
Agreed within the visit: order a one-time lipoprotein(a), and hold the statin decision until it results rather than defaulting to either extreme in the meantime. Both physicians accepted this not as a compromise but as genuinely the most information-dense next step available for the same cost as a blood draw.
Not agreed: what to do if the Lp(a) comes back unremarkable. The preventive cardiologist would treat a normal Lp(a) as removing the last plausible reason to override a clean scan, and would recommend a repeat CAC in five years rather than starting a statin today. The internal medicine physician would still weight the brother's age at event heavily enough to prefer starting a moderate-intensity statin regardless, arguing that a normal Lp(a) narrows the explanation for the family clustering without eliminating it. That disagreement was left explicitly open, to be revisited once the actual result is in hand rather than argued in the abstract.