LAUSR

www.thelancet.com Vol 389 March 11, 2017 987 Dual antiplatelet therapy (DAPT) reduces stent thrombosis and ischaemic events after percutaneous coronary intervention (PCI), but increases the risk of bleeding. Current guidelines recommend default DAPT durations of 6 months and 12 months for patients undergoing elective PCI and those presenting with acute coronary syndromes, respectively, but also leave ample room for shortening or extending the duration of DAPT based on the individual’s risks of ischaemia and bleeding. Ideally, it would be desirable for physicians to personalise the duration of DAPT based on a prediction rule that easily identifies patients at high bleeding risk and separates patients who benefit from shortening DAPT (eg, high bleeding risk and low risk of ischaemia) versus prolonging DAPT (eg, non-high bleeding risk and high risk of ischaemia). However, because risk factors for ischaemia and bleeding largely overlap, modelling of such an algorithm is challenging. Moreover, a risk model should be both precise in allowing for reproducible results across different patient populations and accurate in matching its predictions with true outcomes. Ultimately, despite multiple recent attempts to develop prognostic models for DAPT duration (table), a risk score is useful only if applied on a regular basis: therefore, ease of use is essential for a prognostic index to become part of routine clinical practice. In The Lancet, Francesco Costa and colleagues introduce PRECISE-DAPT, a parsimonious prognostic index to determine the risk of out-of-hospital bleeding in PCI patients on DAPT. The score is made of five domains that are relatively easy to obtain at the bedside (age, history of previous spontaneous bleeding, haemoglobin, white blood cell count, and creatinine clearance). Also, computation at the time of PCI and exploitation of the logistic formula to derive individual 1-year bleeding probabilities is usefully assisted by a nomogram, a website, and a smartphone app. The researchers should be commended for providing clinicians with a practical and user-friendly tool obtained from a large (n=14 963) pooled dataset of eight contemporary multicentre randomised clinical trials and validated in two external cohorts (n=8595 and n=6172, respectively). Anticipating the risk of out-of-hospital bleeding with in-hospital preprocedural variables has important implications that go even beyond the mere determination of DAPT duration. This kind of early risk assessment might be a timely factor in decision making for other bleeding avoidance strategies such as selection of vascular access, stent, or drugs. Patients at high bleeding risk are an increasingly recognised category within the interventional community, but correct identification of this heterogeneous subset remains elusive. Costa and colleagues now suggest that patients at high bleeding risk, such as those identified by a PRECISE-DAPT score of 25 or greater, are exposed to more bleeding with long DAPT duration (number needed to treat=38) but without significant ischaemic gain. Conversely, patients not at high bleeding risk might derive significant ischaemic benefits with long DAPT duration (number needed to treat=65) with no significant bleeding issues. Taken together, these figures suggest that the net clinical benefit of long DAPT duration belongs only to patients not at high bleeding risk, because in the remainder of the PCI population the consequences of bleeding offset the benefits of prolonged DAPT. Importantly, the principle behind the creation of PRECISE-DAPT (ie, defining DAPT duration at the time of PCI based on identification of patients Tailoring duration of DAPT with risk scores