Clinical background
The decision on primary prevention in older adults differs fundamentally from the corresponding decision in middle age. The absolute risk of cardiovascular disease rises with age, but so does the competing risk of non-cardiovascular death, and the benefit of lipid-lowering and blood pressure-lowering drugs must be weighed against adverse effects, polypharmacy and interactions. SCORE2-OP was developed to fill the gap between SCORE2 (which applies to ages 40 to 69 years) and the clinical need for a structured risk estimate in patients from the age of 70. The tool estimates the 10-year risk of fatal and non-fatal cardiovascular disease, and explicitly models the competing risk of non-cardiovascular death, which is necessary in a population with a high background mortality [1].
Calculating SCORE2-OP
SCORE2-OP rests on a sex-specific Cox model adjusted for competing risks. The risk is calculated as:
where is the sex-specific baseline survival at 10 years, LP is the linear predictor and is a regional calibration adjustment. The linear predictor is a weighted sum of age, smoking status, diabetes, systolic blood pressure, total cholesterol and HDL cholesterol, with sex-specific coefficients. The result is recalibrated to one of four risk regions based on the country's cardiovascular mortality: low risk, moderate risk, high risk and very high risk [1].
The derivation cohort consisted of 28,503 individuals over 65 years without established atherosclerotic cardiovascular disease in the Cohort of Norway (CONOR), with 10,089 cardiovascular events during follow-up. The model was validated externally in six independent cohorts totalling 338,615 individuals and 33,219 events [1].
Interpretation in practice
The 2021 ESC guidelines define risk categories for people aged 70 years and over from SCORE2-OP:
| Risk category | 10-year risk | Clinical action |
|---|---|---|
| Low to moderate risk | <15% | Lifestyle advice; treatment of individual risk factors to target, but routine pharmacological primary prevention is discussed individually |
| High risk | 15 to <20% | Active pharmacological treatment of lipids and blood pressure to target; intensified lifestyle intervention |
| Very high risk | ≥20% | Full risk factor modification; a statin and blood pressure-lowering treatment are indicated unless contraindicated |
The thresholds are lower than for younger age groups, reflecting the shorter expected remaining life span and the higher competing risk. An important addition is that the decision must not rest on the percentage alone: in an 85-year-old patient with frailty and polypharmacy, a calculated risk of 22 per cent may lead to a different decision from that in a robust 72-year-old with the same score.
Validation and performance
In the derivation study, the c-index in the external validation cohorts ranged from 0.63 (95 per cent confidence interval 0.61 to 0.65) to 0.67 (0.64 to 0.69), lower than is usually seen for SCORE2 in younger age groups [1]. The lower discrimination is expected: age becomes an increasingly dominant risk factor at older ages, and the other variables contribute less differentiation between individuals.
In a Canadian primary care cohort of 9,885 individuals aged 70 to 89 years, the c-statistic for SCORE2-OP was 0.64 in women and 0.62 in men. The uncalibrated model overestimated risk by more than 100 per cent in both sexes. When the low-risk region calibration was applied, calibration improved, but risk was still overestimated by 60 per cent in women and 13 per cent in men [2]. This underlines that regional calibration is necessary but not always sufficient when the model is applied in a population that differs from the European cohorts in which it was derived.
An analysis of real-world clinical use in four southern European countries showed that SCORE2-OP was in practice most often used to classify patients into the high-risk or very high risk categories, and that a substantial proportion of older patients in countries with low cardiovascular mortality nonetheless fall into these categories because of the strong contribution of age to the score [3].
Limitations
SCORE2-OP applies only to people without established cardiovascular disease. With a previous myocardial infarction, stroke, peripheral arterial disease or established coronary artery disease, the patient is by definition at very high risk whatever the score, and SCORE2-OP must not be used for decisions on primary prevention.
The model was derived in an otherwise healthy population. It has not been validated for patients with severe frailty, multiple chronic diseases or high health care use, and caution is required when applying it in specialist geriatric care. An external validation in a cohort of geriatric outpatients has shown that discrimination falls in this population, particularly when all patients are included irrespective of previous CVD [4].
Diabetes is included as a binary variable without distinguishing type 1 from type 2 or accounting for the degree of metabolic control. Renal function, familial hypercholesterolaemia and lipoprotein(a) are not included, even though these factors can substantially affect risk in the older population. Creatinine and eGFR are absent specifically because SCORE2-OP is built on variables available in large population registries, but this means that patients with impaired renal function may be misclassified.
A practical pitfall is to confuse SCORE2-OP with SCORE2. The tables and algorithms for the two age groups use different baseline survivals, adjustment constants and calibration factors. Applying the SCORE2 tables to a 74-year-old gives incorrect results.
Choosing the right risk region
Much of western and southern Europe belongs to the ESC low-risk region, including, among others, Spain, France and Italy. This means that the low-risk region calibration factors should be selected when the tool is used in these populations. The Canadian validation study showed, however, that the low-risk calibration does not eliminate overestimation entirely. Clinical decisions should therefore take into account that the calculated risk may be somewhat overestimated in a low-risk population, particularly in women [2].
References
- SCORE2-OP working group and ESC Cardiovascular risk collaboration. SCORE2-OP risk prediction algorithms: estimating incident cardiovascular event risk in older persons in four geographical risk regions. Eur Heart J 2021;42(25):2455-67. PMID: 34120185
- Sud M, Sivaswamy A, Austin PC et al. Validation of the European SCORE2 models in a Canadian primary care cohort. Eur J Prev Cardiol 2024;31(6):668-76. PMID: 37946603
- Fontainhas M, Gavina C, Miranda J et al. Cardiovascular risk profile with SCORE2 and SCORE2-OP: comparing Portugal, Spain, Italy, and France using the new European predictive models. Front Cardiovasc Med 2024;11:1509240. PMID: 39744208
- Nijskens CM, Wiersinga JHI, MacNeil-Vroomen JL et al. External Validation of SCORE2-OP for Predicting Cardiovascular Mortality in Frail Geriatric Outpatients. J Am Med Dir Assoc 2026;27(4):106102. PMID: 41628900
