Heart failure (HF) accounts for an increasing proportion of heart disease cases in ageing populations worldwide, with particularly poor treatment outcomes observed in Asia12. As HF is associated with high morbidity and mortality, an effective prediction scheme is urgently needed to identify patients at risk of adverse events, which would facilitate the development of preventive strategies. The AHEAD score (A: atrial fibrillation; H: haemoglobin; E: elderly; A: abnormal renal parameters; D: diabetes mellitus), a simple bedside clinical prognostic model that has been validated in both European and Asian populations, has been widely used to predict all-cause mortality in patients with HF34. Due to the high success rate of primary percutaneous coronary intervention (PCI) and the increasing availability of ambulance services in Japan, in-hospital mortality has declined among patients with acute myocardial infarction (MI) in recent decades56. In particular, the combination of primary PCI with drug-eluting stents and antithrombotic therapies, including P2Y12 receptor inhibitors, has contributed to reductions in stent thrombosis and bleeding events. Moreover, the introduction of novel mechanical support devices has decreased the rate of periprocedural mechanical complications, resulting in more stable haemodynamics and shorter hospital stays following acute MI78910. However, clinical outcomes remain poor in patients with severe acute HF5. Although the AHEAD scoring system is known to be useful in patients with acute HF, limited data are available regarding its prognostic utility in large cohorts of patients with acute MI. Therefore, the aim of this study was to investigate whether the AHEAD score, as a surrogate marker of HF, could predict all-cause mortality – particularly in patients with acute MI, regardless of the presence or absence of acute HF.
Methods
Study population and design
The Japan Acute Myocardial Infarction Registry (JAMIR) has been described in detail previously1112. Briefly, the JAMIR is a multicentre, nationwide, prospective registry in which consecutive patients with spontaneous onset of acute MI were enrolled at 50 institutions between December 2015 and May 2017. The diagnosis of acute MI was made based on either the Universal Definition of MI or the Monitoring Trends and Determinants in Cardiovascular Disease (MONICA) criteria1314. The following patients were excluded from the present secondary analysis: those admitted ≥24 hours after symptom onset, those with no return of spontaneous circulation on admission after out-of-hospital cardiopulmonary arrest, and those with acute MI occurring as a complication of PCI or coronary artery bypass grafting. Patient management was at the discretion of the treating physician. Primary data – including patient demographics and outcomes – were collected from medical records. Data were entered into the JAMIR system by investigators, clinical research coordinators, and local data managers at each site. Follow-up data at 1 year after MI onset were obtained using medical records from each institution. When follow-up data were unavailable due to reasons such as hospital transfer, a letter was sent to the patient requesting updated information. This study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. The research protocol was approved by the Institutional Review Board of the National Cerebral and Cardiovascular Center as well as by the respective ethics committees or institutional review boards of all participating sites. Although informed consent was not obtained because of the observational design of the registry, an opt-out process was implemented through each institution’s website and onsite postings, informing patients of the study and giving them the opportunity to decline participation in the JAMIR. The research secretariat also confirmed that the opt-out procedures were followed at each study site. The JAMIR study was registered with the University Hospital Medical Information Network (UMIN) Clinical Trials Registry (UMIN000019479). This secondary analysis was conducted using data on recorded patient characteristics, procedural details, and clinical outcomes. The primary endpoint of this study was all-cause mortality within 1 year following hospitalisation for acute MI.
Statistical analysis
Continuous variables are presented as means±standard deviations or medians with interquartile ranges, and categorical variables are expressed as frequencies and percentages. The normality of continuous variables was assessed using the Shapiro-Wilk test. Statistical significance was defined as a two-sided p-value<0.05. For continuous variables with normal distributions, comparisons were made using analysis of variance; for non-normally distributed variables, the Kruskal-Wallis test was used. Categorical variables were compared using the chi-squared test, as appropriate. Cox proportional hazard regression analysis was performed to examine the association between the AHEAD score and all-cause mortality up to 1 year after discharge. Variables with p-values<0.15 in the univariate analysis (Table 1) were included in the multivariate models. The individual components of the AHEAD score (i.e., atrial fibrillation, haemoglobin concentration, age, serum creatinine, and diabetes mellitus) were not included in the models to avoid multicollinearity. In sensitivity analyses, variables related to medical therapy were added to the multivariate model. Kaplan-Meier survival analysis and the log-rank test were used to estimate and compare cumulative mortality rates. A forest plot was generated to assess the association between the AHEAD score and all-cause mortality in predefined subgroups of patients with acute MI. A restricted cubic spline Cox regression model was used to evaluate the continuous relationship between the AHEAD score and all-cause mortality. The median AHEAD score of 1 was used as the reference point (hazard ratio 1.0). All analyses were conducted using SPSS, version 26.0 (IBM) and R, version 4.4.1 (R Foundation for Statistical Computing).
Results
A total of 3,411 patients were registered in the JAMIR. Among them, we excluded patients with insufficient data: 7 lacked haemoglobin values, 9 lacked creatinine values, 221 were missing door-to-device time, 104 were missing final Thrombolysis in Myocardial Infarction (TIMI) flow data, and 3 lacked peak creatine kinase values. The final analysis included 3,067 patients (Supplementary Figure 1). Table 1 presents the baseline characteristics of the study population. Patients with higher AHEAD scores were more likely to have hypertension, diabetes mellitus, peripheral artery disease, prior stroke, atrial fibrillation, ST-segment elevation MI (STEMI), and a higher Killip classification. They also had lower haemoglobin concentrations and estimated glomerular filtration rates and were less likely to have a door-to-device time of <90 minutes. Table 2 provides the angiographic and interventional characteristics of the patients. The overall in-hospital mortality rate was 4.9% (149/3,067), and a higher AHEAD score was associated with an increased in-hospital mortality rate (p<0.001). Following univariate analyses, multivariable Cox regression was performed using the AHEAD score as the main variable, adjusting for male sex, hypertension, dyslipidaemia, peripheral artery disease, prior coronary artery bypass grafting, prior stroke, malignancy, STEMI, Killip classification (per class increase), door-to-device time <90 minutes, final TIMI flow, peak creatine kinase >1,672 IU/L (median), and left ventricular ejection fraction (Table 3). This analysis revealed that the AHEAD score was independently associated with an increased risk of all-cause mortality up to 1 year after acute MI (adjusted hazard ratio 1.60, 95% confidence interval: 1.39-1.84; p<0.001). In a sensitivity analysis that included variables related to medical treatment (Supplementary Table 1), the AHEAD score remained independently associated with increased 1-year all-cause mortality. In Kaplan-Meier analysis, patients with higher AHEAD scores exhibited lower survival rates at 1 year following hospitalisation for acute MI (Figure 1). The forest plot demonstrated that the AHEAD score was associated with all-cause mortality across all subgroups (Figure 2). Restricted cubic spline Cox regression demonstrated that an increasing AHEAD score was associated with a progressively higher risk of all-cause mortality (Figure 3).
Table 1. Baseline patient characteristics.
AHEAD score (n=3,067) | |||||||
---|---|---|---|---|---|---|---|
0(n=931) | 1(n=1,082) | 2(n=650) | 3(n=314) | 4(n=77) | 5(n=13) | p-value | |
Age, years | 59 (51, 66) | 68 (59, 76) | 77 (72, 83) | 80 (74, 85) | 80 (75, 84) | 77 (74, 80) | <0.001 |
Male | 825 (88.6) | 837 (77.4) | 431 (66.3) | 211 (67.2) | 53 (68.8) | 11 (84.6) | <0.001 |
Hypertension | 584 (62.7) | 784 (72.5) | 494 (76.0) | 251 (79.9) | 71 (92.2) | 12 (92.3) | <0.001 |
Diabetes mellitus | 0 (0) | 447 (41.3) | 316 (48.6) | 221 (70.4) | 64 (83.1) | 13 (100) | <0.001 |
Dyslipidaemia | 670 (72.0) | 766 (70.8) | 424 (65.2) | 204 (65.0) | 47 (61.0) | 10 (76.9) | 0.011 |
Peripheral artery disease | 8 (0.9) | 31 (2.9) | 36 (5.5) | 25 (8.0) | 10 (13.0) | 3 (23.1) | <0.001 |
Atrial fibrillation | 0 (0) | 26 (2.4) | 56 (8.6) | 72 (22.9) | 34 (44.2) | 13 (100) | <0.001 |
Prior MI | 49 (5.3) | 91 (8.4) | 69 (10.6) | 47 (15.0) | 31 (40.3) | 3 (23.1) | <0.001 |
Prior CABG | 7 (0.8) | 20 (1.8) | 15 (2.3) | 18 (5.7) | 12 (15.6) | 1 (7.7) | <0.001 |
Prior stroke | 30 (3.2) | 96 (8.9) | 81 (12.5) | 59 (18.8) | 17 (22.1) | 4 (30.8) | <0.001 |
Atrial fibrillation | 0 (0) | 26 (2.4) | 56 (8.6) | 72 (22.9) | 34 (44.2) | 13 (100) | <0.001 |
Malignancy | 31 (3.3) | 83 (7.7) | 75 (11.5) | 48 (15.3) | 18 (23.4) | 2 (15.4) | <0.001 |
STEMI | 154 (16.5) | 205 (18.9) | 135 (20.8) | 74 (23.6) | 19 (24.7) | 6 (46.2) | 0.005 |
Killip classification | |||||||
I | 803 (86.3) | 884 (81.7) | 462 (71.1) | 181 (57.6) | 41 (53.2) | 5 (38.5) | <0.001 |
II | 55 (5.9) | 77 (7.1) | 72 (11.1) | 45 (14.3) | 9 (11.7) | 2 (15.4) | – |
III | 25 (2.7) | 43 (4.0) | 30 (4.6) | 40 (12.7) | 12 (15.6) | 4 (30.8) | – |
IV | 48 (5.2) | 78 (7.2) | 86 (13.2) | 48 (15.3) | 15 (19.5) | 2 (15.4) | – |
Haemoglobin, g/dL | 15.1 (14.2, 16.0) |
14.4 (13.3, 15.5) |
12.8 (11.5, 14.3) |
11.7 (10.5, 12.6) |
11.3 (10.2, 11.9) |
10.4 (8.6, 11.4) |
<0.001 |
eGFR, mL/min/1.73 m2 | 73.5 (63.5, 85.2) |
67.8 (54.6, 82.2) |
58.2 (43.9, 72.6) |
43.7 (27.3, 60.6) |
22.7 (11.4, 31.8) |
15.5 (6.96, 30.7) |
<0.001 |
LVEF, % | 53 (46, 60) | 53 (45, 60) | 50 (42, 60) | 50 (40, 60) | 41 (30, 57) | 44 (36, 55) | <0.001 |
Door-to-balloon time <90 min |
685 (73.6) | 772 (71.3) | 437 (67.2) | 196 (62.4) | 46 (59.7) | 6 (46.2) | <0.001 |
Final TIMI flow=3 | 862 (92.6) | 1,003 (92.7) | 601 (92.5) | 275 (87.6) | 74 (96.1) | 12 (92.3) | 0.042 |
Peak CK concentration, IU/L | 1,978 (878, 3,887) |
1,750 (678, 3,627) |
1,433 (546, 3,005) |
1,135 (395, 3,015) |
1,312 (446, 2,521) |
1,286 (486, 2,580) |
0.001 |
Values are presented as medians (interquartile ranges) or n (%). CABG: coronary artery bypass grafting; CK: creatine kinase; eGFR: estimated glomerular filtration rate; LVEF: left ventricular ejection fraction; MI: myocardial infarction; STEMI: ST-segment elevation myocardial infarction; TIMI: Thrombolysis in Myocardial Infarction |
Table 2. In-hospital outcomes and medical therapy during hospitalisation.
AHEAD score (n=3,067) | |||||||
---|---|---|---|---|---|---|---|
0(n=931) | 1(n=1,082) | 2(n=650) | 3(n=314) | 4(n=77) | 5(n=13) | p-value | |
Culprit lesion – left main | 10 (1.0) | 22 (2.0) | 18 (2.7) | 11 (3.5) | 0 (0) | 0 (0) | 0.041 |
Culprit lesion – left anterior descending | 487 (52.3) | 537 (49.6) | 300 (46.1) | 133 (42.3) | 32 (41.5) | 8 (61.5) | 0.013 |
Culprit lesion – left circumflex | 130 (13.9) | 164 (15.1) | 109 (16.7) | 46 (14.6) | 10 (12.9) | 0 (0) | 0.419 |
Culprit lesion – RCA | 321 (34.4) | 383 (35.3) | 247 (38.0) | 137 (43.6) | 33 (42.8) | 5 (38.4) | 0.051 |
PCI with drug-eluting stent | 840 (90.2) | 992 (91.6) | 563 (86.6) | 256 (81.5) | 64 (83.1) | 8 (61.5) | <0.001 |
CABG during hospitalisation | 4 (0.4) | 12 (1.1) | 5 (0.7) | 5 (1.5) | 0 (0) | 0 (0) | 0.314 |
Mechanical support – IABP | 87 (9.3) | 117 (10.8) | 106 (16.3) | 64 (20.3) | 10 (12.9) | 4 (30.7) | <0.001 |
Mechanical support – PCPS | 11 (1.1) | 14 (1.2) | 18 (2.7) | 15 (4.7) | 1 (1.2) | 2 (15.3) | <0.001 |
In-hospital mortality | 12 (1.2) | 32 (2.9) | 52 (8.0) | 38 (12.1) | 11 (14.2) | 4 (30.7) | <0.001 |
Cardiovascular death, non-fatal MI, or non-fatal stroke | 16 (1.7) | 32 (2.9) | 42 (6.4) | 38 (12.1) | 9 (11.6) | 4 (30.7) | <0.001 |
Myocardial infarction | 8 (0.8) | 7 (0.6) | 7 (1.0) | 4 (1.2) | 1 (1.2) | 1 (7.6) | 0.184 |
Stent thrombosis | 4 (0.4) | 6 (0.5) | 4 (0.6) | 0 (0) | 1 (1.2) | 0 (0) | 0.480 |
BARC Type 3 or 5 bleeding | 16 (1.7) | 31 (2.8) | 40 (6.1) | 37 (11.7) | 8 (10.3) | 2 (15.3) | <0.001 |
Dual antiplatelet therapy | 918 (98.6) | 1,071 (98.9) | 626 (96.3) | 304 (96.8) | 72 (93.5) | 11 (84.6) | <0.001 |
ACEi/ARB | 761 (81.7) | 848 (78.3) | 494 (76.0) | 208 (66.2) | 48 (62.3) | 6 (46.1) | <0.001 |
Beta blocker | 646 (69.3) | 719 (66.4) | 402 (61.8) | 188 (59.8) | 47 (61.0) | 9 (69.2) | <0.001 |
Statin | 884 (94.9) | 1,015 (98.3) | 563 (86.6) | 254 (80.8) | 63 (81.8) | 12 (92.3) | <0.001 |
Values are presented as n (%). ACEi: angiotensin-converting enzyme inhibitor; ARB: angiotensin II type 1 receptor blocker; BARC: Bleeding Academic Research Consortium; CABG: coronary artery bypass grafting; IABP: intra-aortic balloon pump; MI: myocardial infarction; PCI: percutaneous coronary intervention; PCPS: percutaneous cardiopulmonary support; RCA: right coronary artery |
Table 3. Cox regression analysis for the prediction of 1-year mortality in patients with acute myocardial infarction.
Univariate | Multivariate | |||||
---|---|---|---|---|---|---|
HR | 95% CI | p-value | aHR | 95% CI | p-value | |
AHEAD score (per 1-point increase) | 1.98 | 1.79-2.20 | <0.001 | 1.60 | 1.39-1.84 | <0.001 |
Variables | ||||||
Male | 1.37 | 1.02-1.82 | 0.032 | 1.15 | 0.79-1.67 | 0.455 |
Hypertension | 0.72 | 0.55-0.95 | 0.022 | 0.69 | 0.48-1.01 | 0.055 |
Dyslipidaemia | 0.52 | 0.40-0.67 | <0.001 | 0.43 | 0.31-0.60 | <0.001 |
Peripheral artery disease | 2.53 | 1.58-4.05 | <0.001 | 1.73 | 1.00-3.02 | 0.050 |
Prior MI | 1.28 | 0.84-1.94 | 0.243 | |||
Prior coronary artery bypass grafting | 1.91 | 1.01-3.60 | 0.045 | 1.18 | 0.54-2.60 | 0.655 |
Prior stroke | 2.53 | 1.82-3.52 | <0.001 | 1.22 | 0.79-1.89 | 0.360 |
Malignancy | 2.64 | 1.88-3.69 | <0.001 | 1.86 | 1.22-2.83 | 0.004 |
STEMI | 0.55 | 0.48-1.00 | 0.055 | 0.57 | 0.35-0.93 | 0.026 |
Killip classification (per class increase) | 2.28 | 2.07-2.52 | <0.001 | 1.64 | 1.42-1.90 | <0.001 |
Door-to-balloon time <90 min | 0.52 | 0.40-0.68 | <0.001 | 0.70 | 0.50-0.98 | 0.040 |
Final TIMI flow (per class increase) | 0.67 | 0.55-0.82 | <0.001 | 0.70 | 0.54-0.89 | 0.005 |
Peak creatine kinase concentration >1,672 IU/L | 1.93 | 1.47-2.54 | <0.001 | 1.33 | 0.93-1.89 | 0.112 |
Left ventricular ejection fraction, % | 0.92 | 0.91-0.93 | <0.001 | 0.95 | 0.93-0.96 | <0.001 |
aHR: adjusted hazard ratio; CI: confidence interval; HR: hazard ratio; MI: myocardial infarction; STEMI: ST-segment elevation myocardial infarction; TIMI: Thrombolysis in Myocardial Infarction |
Figure 1. Kaplan-Meier analysis stratified by AHEAD score. Patients with higher AHEAD scores had a lower survival rate at 1 year following hospitalisation for acute myocardial infarction.
Figure 2. Forest plots of hazard ratios for all-cause mortality in different subgroups. The forest plot demonstrates that the AHEAD score was a consistent predictor of all-cause mortality across all subgroups. CI: confidence interval; CK: creatine kinase; DTB: door-to-balloon time; HR: hazard ratio; LVEF: left ventricular ejection fraction; STEMI: ST-segment elevation myocardial infarction
Figure 3. Restricted cubic spline Cox regression showing the association between the AHEAD score and all-cause mortality. An increasing AHEAD score was associated with a progressively higher risk of all-cause mortality. The median AHEAD score of 1 was used as the reference point (hazard ratio 1.0). The blue line represents the hazard ratio, and the shaded light blue area indicates the 95% confidence interval.
Discussion
In this multicentre study, we demonstrated that the AHEAD score, as a marker of HF, is predictive of all-cause mortality among patients with acute MI. Our results were consistent across multivariable Cox regression and stratified analyses in the forest plot (Central illustration). According to their baseline characteristics, patients with a higher AHEAD score appeared to have more severe cardiovascular disease. In addition, late presentation – probably due to less prominent symptoms resulting from diabetic neuropathy – may have contributed to the longer door-to-device times observed in patients with higher AHEAD scores, thereby complicating their systemic condition. Notably, patients with higher AHEAD scores were more likely to experience Bleeding Academic Research Consortium Type 3 or 5 bleeding during hospitalisation for acute MI, which may have led to a lower rate of PCI with drug-eluting stent use, as well as a reduced rate of dual antiplatelet therapy. In addition, they were less likely to receive angiotensin-converting enzyme inhibitors or angiotensin II type 1 receptor blockers, possibly due to unstable haemodynamics or lower blood pressure during the periprocedural period. This finding is consistent with the fact that these patients had a higher incidence of Killip class III/IV and greater use of mechanical circulatory support during the same period. Multivariate analysis demonstrated that the AHEAD score was independently associated with 1-year all-cause mortality following hospitalisation for acute MI. This result is comparable to that of a previous study in which patients with acute MI were excluded4. To the best of our knowledge, this is the first study in which the predictive value of the AHEAD score has been externally validated in a large cohort of Asian patients with acute MI, regardless of the presence or absence of acute HF. We demonstrated that the AHEAD score can be used to accurately stratify the risk of all-cause mortality at 1 year in patients with acute MI, not only those with HF4. In subgroup analysis, the AHEAD score was found to be an independent predictor of all-cause mortality across all subgroups, suggesting that the AHEAD score is useful for predicting poor outcomes in acute MI, irrespective of the presence of HF. Furthermore, we examined the association between the AHEAD score and all-cause mortality using a restricted cubic spline Cox model. The analysis demonstrated an overall upward trend, reflecting an increase in mortality risk with higher AHEAD scores. This association shown in the spline curve is consistent with the Kaplan-Meier curve, suggesting that the AHEAD score and risk of all-cause mortality have a linear relation. As described above, the AHEAD score, originally developed as a simple risk stratification tool for patients with acute HF, incorporates five components: atrial fibrillation, haemoglobin level, elderly age, abnormal renal function, and diabetes mellitus. Previous studies have demonstrated its utility in predicting mortality and hospitalisation in heart failure cohorts34. In our analysis, the observed increase in hazard with higher AHEAD scores aligns with existing evidence showing that cumulative comorbidities adversely affect prognosis. However, the elevated risk observed among patients with lower AHEAD scores raises important questions. This subgroup may include individuals with unmeasured high-risk features such as frailty, active malignancy, or life-threatening bleeding events1516. Additionally, early mortality related to acute cardiovascular events may not be adequately captured by the components of the AHEAD score, which focus on chronic conditions. Such limitations emphasise that the AHEAD score, while useful, does not fully encompass the multidimensional nature of risk in contemporary patient populations. This finding supports the integration of complementary tools or biomarkers into risk assessment frameworks. Physicians should remain aware that risk prediction tools − while simple, accessible, and informative − require context-specific interpretation and should be augmented by clinical judgement and, where appropriate, additional risk markers.
Central illustration. The AHEAD score as a useful marker of all-cause mortality in both acute MI and heart failure populations. A) The AHEAD score points system; (B) all-cause mortality in different subgroups; (C) Kaplan-Meier analysis; (D) the association between the AHEAD score and all-cause mortality. CI: confidence interval; CK: creatine kinase; DTB: door-to-balloon time; HR: hazard ratio; LVEF: left ventricular ejection fraction; MI: myocardial infarction; STEMI: ST-segment elevation myocardial infarction
Limitations
First, residual or unmeasured confounding factors may have remained in our analyses due to the observational nature of the study. Second, although the original study was conducted across 50 institutions throughout Japan, the selection process for participating hospitals may have introduced selection bias. In addition, limited ethnic diversity and interinstitutional variability in healthcare practices may restrict the generalisability of the findings. Third, only patients with acute MI who were treated with early revascularisation therapy – the vast majority of whom underwent PCI – were included; hence, our results may not be applicable to patients receiving alternative treatments. Finally, because the study was observational, the timing of blood sampling for creatine kinase measurement was not standardised in the study protocol and was instead determined by local hospital practices. As a result, peak creatine kinase concentrations may have been underestimated in certain cases.
Conclusions
The AHEAD score, a simple and validated marker of all-cause mortality in patients with acute HF, was also found to be associated with all-cause mortality in patients with acute MI at 1 year. This score may be useful because of its simplicity and clinical applicability, particularly in patients with acute MI complicated by HF.
Impact on daily practice
The AHEAD score, a simple bedside tool, enables rapid risk stratification for patients with acute myocardial infarction. This study confirms its predictive value for 1-year all-cause mortality, even in patients without overt heart failure. Clinicians can apply the score at admission to identify high-risk patients, guide therapeutic decisions, and prioritise monitoring. Its simplicity allows for easy integration into clinical workflows and electronic health records. By enhancing early prognostic evaluation, the AHEAD score supports personalised care strategies, improves communication with patients and families, and contributes to more effective allocation of healthcare resources.
Acknowledgements
The authors thank all the investigators who participated in this registry. The authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation.
Funding
This work was planned by the Japan Cardiovascular Research Foundation and was funded by Daiichi Sankyo Co., Ltd. This study was supported in part by a Grant-in-Aid for Scientific Research (17K09542) from the Ministry of Education, Science, and Culture, Japan.
Conflict of interest statement
The authors have no conflicts of interest to declare.