DAPT Score (Dual Antiplatelet Therapy)

Clinical background

The decision to extend dual antiplatelet therapy (DAPT) beyond 12 months after PCI with a stent is a trade-off between two risks that are driven in part by the same patient factors. Longer treatment reduces stent thrombosis and myocardial infarction but increases bleeding risk, and patients at high ischaemic risk often also have a high bleeding risk. The DAPT score was developed precisely to resolve this: to identify patients in whom the ischaemic benefit of continuing a thienopyridine outweighs the bleeding harm, and vice versa.

The score is therefore not a pure risk stratification for ischaemia or bleeding separately, but an integrated benefit–harm calculation. It answers the question: for this particular patient, is it worth continuing?

Calculating the DAPT score

The score is calculated as follows:

$$\text{DAPT score} = \text{age}$${\text{points}} + \text{smoking} + \text{diabetes} + \text{MI}{\text{presentation}} + \text{previous PCI/MI} + \text{stent}{<3\text{mm}} + \text{paclitaxel stent} + 2 \times (\text{heart failure/EF}{<30%}) + 2 \times (\text{saphenous vein graft PCI})

where the age points are:

$$\text{age}_{\text{points}} = \begin{cases} -2 & \text{if age} \geq 75 \ -1 & \text{if } 65 \leq \text{age} < 75 \ 0 & \text{if age} < 65 \end{cases}$$

The other variables score 1 point each except heart failure/EF <30% and PCI of a saphenous vein graft, which score 2 points each. The range is −2 to 10.

The derivation cohort consisted of 11,648 randomised patients from the DAPT study, recruited in 11 countries between August 2009 and May 2014 [1]. The mean age was 61.3 years and 25.2% were women. All had undergone PCI with a drug-eluting stent (DES) or a bare metal stent (BMS) and had been treated with open-label thienopyridine plus aspirin for 12 months. At 12 months, those who were event-free and compliant were randomised to continued thienopyridine or placebo for 18 months. The primary outcomes were stent thrombosis or myocardial infarction (ischaemia) and GUSTO moderate/severe bleeding, measured between months 12 and 30.

The models for ischaemia and bleeding had c-statistics of 0.70 and 0.68 in the derivation cohort [1]. The score was constructed by calculating for each patient the predicted absolute risk reduction for ischaemia and the predicted absolute risk increase for bleeding with continued thienopyridine, and then regressing the difference on the variables included in both models. Variables that explained more than 1% of the variation in this benefit–harm difference were retained in the simplified score.

Interpretation in practice

The calculator's threshold is 2 points. In the derivation cohort, patients were divided into high score (≥2) and low score (<2), and the treatment effect of continued thienopyridine differed between the groups (interaction p < 0.001 for ischaemia, p = 0.02 for bleeding) [1].

Mortality did not differ between the groups at a high score (2.1% vs 2.1%), but at a low score there was a numerically higher mortality with continued thienopyridine (1.7% vs 0.9%, p = 0.02), although without a significant interaction (p = 0.14) [1].

It is important to note that the score is meaningful only in patients who have already tolerated 12 months of DAPT without an ischaemic or bleeding event. It should not be used at the time of PCI to determine the initial duration of treatment.

Validation and performance

External validation was carried out within the PROTECT trial (8,709 patients, 36 countries) with a c-statistic of 0.64 for both the ischaemia and the bleeding model, compared with 0.70 and 0.68 in the derivation cohort [1]. In the PROTECT cohort, patients with a high score did indeed have an increased ischaemic risk, but bleeding risk did not differ between the score groups, indicating that the ability of the score to separate ischaemic from bleeding risk is weaker outside the original trial population.

A Swedish registry-based validation (Ueda et al., 41,101 patients from SWEDEHEART/SCAAR, 2006 to 2014) showed poorer discrimination still: a c-statistic of 0.58 for myocardial infarction or stent thrombosis and 0.49 for fatal or major bleeding [2]. The bleeding rate in the Swedish population was approximately half that in the placebo arm of the DAPT study, and the relationship between score and ischaemic risk did not follow the proposed decision rule. The authors concluded that the score and its threshold are not generalisable to a real-world clinical population.

A meta-analysis of seven external validation studies (77,274 patients) found, on the other hand, that a high score (≥2) was associated with an increased ischaemic risk (OR 1.54) and a lower bleeding risk (OR 0.84) [3]. With extended DAPT (12 to 24 months), ischaemic risk fell significantly in the high-score group (OR 0.67) without an increase in bleeding, while the low-score group had no ischaemic benefit but a significant increase in bleeding (OR 1.58). The meta-analysis therefore supports the usefulness of the score as a decision tool, but notes that discrimination is modest and that the results vary between cohorts.

Limitations

Several factors limit the applicability of the score:

Paclitaxel-eluting stents form a variable in their own right, but this stent type has essentially disappeared from clinical practice. The variable is therefore redundant for patients stented with modern everolimus- or zotarolimus-eluting stents, and the score has not been recalibrated to handle this.

The derivation population excluded patients with planned procedures requiring a treatment interruption of >14 days, those with an expected survival of <3 years and those on long-term anticoagulation. The score does not apply to these groups.

Discrimination is modest even in the derivation cohort (c-statistic 0.70 for ischaemia, 0.68 for bleeding) and falls in external cohorts, particularly in Swedish registry data [2]. The score does not capture all the variation in ischaemic and bleeding risk, and clinical assessment of individual factors (for example impaired renal function, anaemia, a previous bleeding history) must be weighed in alongside the score.

The treatment landscape has shifted since the DAPT study. Modern P2Y12 inhibitors (ticagrelor, prasugrel) and de-escalation strategies after acute coronary syndromes are not handled by the score, which was derived in an era dominated by clopidogrel. ESC guidelines from 2017 onwards emphasise a structured assessment of ischaemic and bleeding risk (for example with PRECISE-DAPT and ARC-HBR) rather than a single score [4].

Bleeding risk calibration: in the Swedish validation the bleeding rate was considerably lower than in the DAPT study, which means that a threshold of 2 points may lead to DAPT being stopped too early in populations with a low underlying bleeding risk [2].

Registry validation in SWEDEHEART

The Swedish registry study by Ueda et al. is particularly relevant because it was based on SWEDEHEART data and reflects Swedish practice [2]. The results showed that the DAPT score did not discriminate adequately either for ischaemia or for bleeding in a Swedish real-world population, and that the bleeding risk was lower than in the original DAPT study. This means that the score's threshold of 2 points, if applied mechanically, may lead to incorrect decisions in Swedish practice. Current ESC guidelines, which Swedish cardiology largely follows, recommend an individual trade-off with separate bleeding and ischaemic risk assessments rather than the DAPT score as the sole decision aid [4].

References

1. Yeh RW, Secemsky EA, Kereiakes DJ et al. Development and Validation of a Prediction Rule for Benefit and Harm of Dual Antiplatelet Therapy Beyond 1 Year After Percutaneous Coronary Intervention. JAMA 2016;315(16):1735-1749. PMID: 27022822
2. Ueda P, Jernberg T, James S et al. External Validation of the DAPT Score in a Nationwide Population. J Am Coll Cardiol 2018;72(10):1069-1078. PMID: 30158058
3. Witberg G, Zusman O, Yahav D et al. Meta-analysis of studies examining the external validity of the dual antiplatelet therapy score. Eur Heart J Cardiovasc Pharmacother 2020;6(5):285-291. PMID: 31794022
4. Hudzik B, Błachut A, Lesiak M et al. Summary of the European Society of Cardiology guidelines on dual antiplatelet therapy in patients after percutaneous coronary interventions. Kardiol Pol 2022;80(10):974-989. PMID: 36036339