- Design
- prospective observational cohort, Copenhagen General Population Study, median 12.5 years of follow-up (maximum 19.1)
- Population
- 103,341 adults aged 20 to 100 years (57,899 women, 45,442 men), 18,322 of them aged 70-100
- Primary outcome
- incident atherosclerotic cardiovascular disease, by age band and sex
- Effect
- hazard ratio 1.12-1.18 per 50 mg/dL higher Lp(a) at every age (P interaction by age 0.48); absolute excess incidence 0.3 per 1,000 person-years at 20-49 years rising to 4.5 at 80-100 years
The Copenhagen General Population Study followed 57,899 women and 45,442 men aged 20 to 100 years, 18% of them aged 70 or above, for a median of 12.5 years. Some 9,091 developed atherosclerotic cardiovascular disease. The analysis asked whether the association between lipoprotein(a) and that disease weakens with age, as risk-factor associations often appear to.
It did not. The hazard ratio per 50 mg/dL higher lipoprotein(a) ranged from 1.12 to 1.18 across every age band and both sexes, with no interaction by age (P=0.48) or sex (P=0.74). What changed was the absolute arithmetic. The extra incidence per 50 mg/dL rose from 0.3 per 1,000 person-years at ages 20 to 49, to 2.7 at 70 to 79, and 4.5 at 80 to 100. For those with lipoprotein(a) of 90 mg/dL or more, the excess ran from 3.2 per 1,000 person-years in the youngest group to 48 in the oldest.
The practical consequence is the opposite of the usual instinct. Clinicians commonly stop measuring lipoprotein(a) in older patients on the grounds that it is a marker for premature disease. On these data an older patient with high lipoprotein(a) carries far more excess events than a younger one with the same level, because the same relative risk is applied to a much larger baseline.
Measure it once, in anyone in whom the result would change how hard you push everything else that is modifiable. Lipoprotein(a) is a single non-fasting sample and does not need repeating. This is an observational cohort, so it establishes association and consistency of association, not that lowering the level helps; the trials answering that are still running. Note too that the cohort is Danish, so the absolute rates should not be carried across to Indian patients without adjustment, even though the relative associations plausibly travel.
- Measure lipoprotein(a) once in adults with premature or recurrent atherosclerotic disease, or a strong family history, whatever their age
- Do not repeat the test: the level is largely genetically set and stable, and one non-fasting sample is enough
- Record whether your laboratory reports mg/dL or nmol/L, because thresholds differ and the two are not interchangeable by a fixed factor
- Use a high result to intensify what can be treated, LDL cholesterol, blood pressure, smoking and diabetes, not as a reason for therapeutic nihilism
- Say plainly that no approved therapy lowers lipoprotein(a) for outcomes yet, so the number guides intensity rather than prompting a new drug
The statistics, in plain English
Hazard ratios of 1.12 to 1.18 per 50 mg/dL are modest and would look unimpressive alone, but they are applied to a background risk that multiplies with age, which is why the same ratio produces 0.3 extra events per 1,000 person-years in the under-fifties and 4.5 in the over-eighties. The interaction P values of 0.48 and 0.74 mean the data give no reason to believe the relationship differs by age or sex, which is the point of the paper. Being observational, it cannot show that lowering lipoprotein(a) lowers risk, only that the two track together consistently across the lifespan.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for cardiology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free