Of 1665 consecutive patients undergoing PCI, after excluding 57 patients not available for BMI, 1608 patients were analyzed for the current study. Patients were divided into two groups according to the median value of the HELT-E2S2 score (low HELT-E2S2 score [< 2, n = 915] and high HELT-E2S2 score [≥ 2, n = 693]) (Fig. 1A). Major criteria of the HELT-E2S2 score were hypertension (74.7%), elderly age 75–84 years (31.3%), and type of AF (11.1%) (Fig. 1B). Low BMI, extreme elderly age ≥ 85, and previous stroke frequently overlapped with other criteria as illustrated in Supplementary Fig. 1.

Distribution (A) and prevalence (B) of HELT-E2S2 score. AF atrial fibrillation, BMI body mass index

Clinical and procedural characteristics are summarized in Tables 1 and 2. Patients in the high HELT-E2S2 score group were older and had more atherosclerotic risk factors, comorbidities, chronic coronary syndrome as an indication for PCI, and higher CHADS2 and CHA2DS2-VASc scores compared with those in the low HELT-E2S2 score group. There were no differences in almost all procedural characteristics between the two groups, except for the use of new-generation drug eluting stent (DES) and mean stent diameter.

During a median follow-up period of 5 years (inter-quartile range: 3.4–5.0 years), there were 367 events of MACE (incidence rate: 5.22 per 100 patient-years), 280 all-cause deaths (incidence rate: 4.30 per 100 patient-years), 68 non-fatal MIs (incidence rate: 0.49 per 100 patient-years), 69 non-fatal strokes (incidence rate: 1.02 per 100 patient-years), 124 BARC 2, 3 or 5 bleeding events (incidence rate: 1.92 per 100 patient-years) (Table 3). The high HELT-E2S2 score group had an increased risk of MACE (33.4% vs. 14.8%, P < 0.001), all-cause death (27.6% vs. 9.7%, P < 0.001), non-fatal ischemic stroke (8.0% vs. 2.4%, P < 0.001), and BARC 2, 3 or 5 bleeding (12.7% vs. 5.3%, P < 0.001) (Figs. 2 and 3). There was no significant difference in myocardial infarction (4.1% vs. 4.5%, P = 0.524).

Kaplan–Meier curve for MACE at 5 years. MACE major adverse cardiovascular events

Kaplan–Meier curves for A all-cause death, B non-fatal myocardial infarction, C non-fatal ischemic stroke, and D BARC 2, 3 or 5 bleeding events. BARC Bleeding Academic Research Consortium

As a sensitivity analysis, we divided patients into 4 groups based on HELT-E2S2 and CHADS2 score and assessed clinical outcomes. The HELT-E2S2 and CHADS2 scores were concordant in 77.6% of patients (HELT-E2S2 score ≥ 2 and CHADS2 score ≥ 2 [n = 641, 39.9%], HELT-E2S2 score < 2 and CHADS2 score < 2 [n = 607, 37.7%], HELT-E2S2 score ≥ 2 and CHADS2 score < 2 [n = 52, 3.2%], HELT-E2S2 score < 2 and CHADS2 score ≥ 2 [n = 308, 19.2%]). Kaplan–Meier analysis revealed the high HELT-E2S2 score and high CHADS2 score group had a higher incidence of MACE compared with other groups (Supplementary Fig. 2). Patients with discordance between the HELT-E2S2 and CHADS2 scores had an increased risk of MACE compared with those with low HELT-E2S2 score and low CHADS2 score.

The results of Cox proportional hazards analyses for clinical outcomes are summarized in Table 3. Multivariable Cox analysis demonstrated that, compared with the low HELT-E2S2 score (< 2), the high HELT-E2S2 score (≥ 2) was associated with an increased risk of MACE (HR, 1.73; 95% CI, 1.11–2.69, P = 0.015), all-cause death (HR, 1.65; 95% CI, 1.01–2.72, P = 0.047), and non-fatal stroke (HR, 4.67; 95% CI, 1.54–14.17, P = 0.006). Similarly, the HELT-E2S2 score as a continuous value (i.e. 0–7) emerged as an independent predictor for MACE, all-cause death, and non-fatal stroke (all, P < 0.05).

The C-statistics of the HELT-E2S2 score, CHADS₂ score, and CREDO-Kyoto thrombotic risk score for MACE at 5 years were 0.646, 0.642, and 0.696, respectively, and for ischemic stroke at 5 years were 0.667, 0.612, and 0.630, respectively.