Risk factors for cardiac death in women with coronary artery disease

Abstract

Aims

Coronary artery disease affects a substantial proportion of women, who may exhibit distinct disease progression and outcomes compared with men. This study aimed to investigate potential markers linked to cardiac death in women with coronary artery disease.

Methods and results

This study utilized data from the ARTEMIS cohort, comprising 1946 patients (619 women) with angiographically confirmed coronary artery disease. The primary endpoint during the 10-year follow-up was cardiac death. We examined associations between cardiac death and traditional coronary artery disease risk factors, echocardiographic and electrocardiographic parameters, biomarkers, and angiographical parameters. During follow-up, 7.6% (n = 47) of women experienced cardiac death. In multivariate analysis, factors with significant associations with cardiac death included age (HR 1.1 per year, 95% CI 1.0–1.2, P < 0.001), post-revascularization SYNTAX Score (HR 1.05 per unit increase, 95% CI 1.0–1.1, P = 0.005), systolic blood pressure (HR 1.1 per 10-unit increase, 95% CI 1.0–1.3, P = 0.039), high-sensitivity troponin T (HR 1.5 per 10-unit increase, 95% CI 1.1–1.9, P = 0.006), haemoglobin A1c (HR 1.4 per one-unit increase, 95% CI 1.1–1.9, P = 0.016), and permanent atrial fibrillation (HR 4.9, 95% CI 1.6–15.0, P = 0.005).

Conclusion

Increasing post-revascularization SYNTAX Score was associated with cardiac mortality in women with coronary artery disease. Alongside age and systolic blood pressure, high-sensitivity troponin T and haemoglobin A1c were also associated with risk. In addition, permanent atrial fibrillation may have prognostic significance in this population and warrants further investigation. Overall, these findings highlight markers that may help refine risk stratification among women with coronary artery disease.

Introduction

Coronary artery disease (CAD) remains a significant cause of death, causing a significant burden on mortality and person-years of life lost.¹ In 2021, among European Union residents aged ≥65 years, CAD was the leading cause of death, with a standardized death rate of 673.0 per 1 000 000 men and 399.5 per 100 000 women.² However, CAD concerns also younger populations, as about 20% of CAD-related deaths occur in adults under 65 years old.³ Half of all CAD-related deaths are sudden, while women comprise one-third of sudden cardiac death (SCD) cases overall.⁴

Approximately 40% of CAD patients are women.¹ Growing evidence suggests that CAD presents, progresses, and impacts women differently than men, with women often conveying unique risk factors, symptoms, and prognosis.^5,6 For instance, after myocardial infarction (MI), women tend to have worse outcomes, including higher mortality and adverse event rates.⁶ Understanding the sex-specific risk factors is essential for advancing treatment strategies and improving survival rates among women with CAD. This requires more study data on sex-specific aspects on women, as historically, women have often been underrepresented in CAD studies.⁷ This study aimed to investigate potential markers linked to cardiac death in women with CAD.

Methods

Study population

This study is based on the ARTEMIS (Innovation to Reduce Cardiovascular Complications of Diabetes at the Intersection) Study conducted in the Division of Cardiology of Oulu University Hospital (ClinicalTrials.gov identifier NCT01426685). The ARTEMIS population consists of 1946 (619 women) subjects with angiographically verified CAD with or without type 2 diabetes mellitus (T2DM). The original aim of ARTEMIS study was to study the prognostic value of cardiovascular risk markers in patients with CAD. Participants in the ARTEMIS study were recruited between 2007–14, 3–6 months following coronary angiography and revascularization, or after receiving contemporary medical treatment if revascularization was not indicated. The time point zero of data collection was defined as the first study visit after angiography or revascularization. Significant CAD was defined angiographically as an occlusion of >50% in at least one major epicardial artery, regardless of whether the patient had a prior CAD diagnosis. In the total ARTEMIS cohort, revascularization [percutaneous coronary intervention (PCI) or coronary artery bypass crafting (CABG)] had been performed in 80.3% of women and 86.8% of men. The inclusion criteria also required participants to be between 18 and 85 years of age. Exclusion criteria included New York Heart Association or Canadian Cardiovascular Society (CCS) class IV conditions, significant valvular disease, a planned or current implantable cardioverter-defibrillator (ICD), end-stage renal failure requiring dialysis, life expectancy under 1 year, or any illness significantly affecting psychological or physical fitness. A more detailed study protocol has been described in previous studies concerning this population.^8,9

Study protocol

The present study focused on evaluating the effect of risk factors for cardiac death listed in Table 1. The initial study examinations of the ARTEMIS population took place at least 3 months after coronary angiography or the most recent revascularization, and the patient was stable regarding CAD. At baseline, medical therapy for both DM and coronary atherosclerosis was optimized in accordance with the 2007 American Diabetes Association Standards of Medical Care in Diabetes.¹⁰ After baseline examinations, patients were managed by their local physicians, who adjusted therapy as clinically indicated. For patients without a previous T2DM diagnosis, an oral glucose tolerance test was performed, along with measurements of fasting plasma glucose and haemoglobin A1c (HbA1c). Type 2 diabetes mellitus was diagnosed according to WHO criteria. Alongside the tests for T2DM diagnosis, various laboratory assessments were obtained, including haemoglobin, thyroid-stimulating hormone (TSH), creatinine clearance (CrCl), high-sensitivity C-reactive protein (hs-CRP), high-sensitivity troponin T (hs-TnT), blood lipid levels, and B-type natriuretic peptide (BNP). Each patient’s blood pressure (BP) was measured in a supine position following 10 min of rest. High-sensitivity troponin T was categorized based on the established normal threshold of 14 ng/L (99th percentile), and BP was categorized using the commonly accepted cut-off value of 140 mmHg.

Table 1.

Baseline characteristics of subjects with cardiac death during 10-year follow-up

	Women (n = 47)	Men (n = 126)	P-value
Constant variables
Age (years)	74.4 ± 6.5	70.5 ± 7.8	0.003
SYNTAX Score (pre)a	15.3 ± 11.9	17.6 ± 13.2	0.318
SYNTAX Score (post)a	2.0 (0.0, 9.8)	5.0 (0.0, 12.0)	0.712
BMI (kg/m2)	28.6 ± 4.8	28.9 ± 5.0	0.756
Waist circumference (cm)	95.0 ± 13.6	104.2 ± 13.2	<0.001
Systolic BP (mmHg)	163.1 ± 24.6	142.7 ± 21.0	<0.001
Diastolic BP (mmHg)	77.1 ± 12.0	76.5 ± 10.0	0.713
Mean heart rate (b.p.m.)	65.9 ± 8.0	68.1 ± 11.3	0.248
Creatinine clearance (mL/min)	61.9 (49.1, 77.4)	80.2 (62.9, 100.4)	<0.001
LDL (mmol/L)	2.2 (1.9, 3.0)	2.2 (1.8, 2.7)	0.743
BNP (ng/L)	120 (75.0, 211.0)	94.5 (47.8, 181.0)	0.079
hs-CRP (mg/L)	1.1 (0.7, 2.4)	1.6 (0.5, 5.2)	0.665
hs-TnT (ng/L)	11.0 (6.0, 20.3)	15.0 (9.0, 24.0)	0.011
HbA1c (%)	6.5 (5.9, 7.3)	6.4 (5.9, 7.2)	0.943
TSH (mU/L)	2.0 (1.1, 2.1)	1.9 (1.3, 2.7)	0.939
Haemoglobin (g/L)	129.2 ± 13.7	136.4 ± 14.5	0.005
LV ejection fraction (%)	65.0 (60.0, 71.4)	57.4 ± 14.6	0.002
LV mass (g)	184.1 ± 51.1	257.8 ± 75.6	<0.001
Categorical variables, % (n)
Type 2 diabetes	57.4 (27)	58.7 (74)	1.000
Oestrogen or testosterone therapy	27.7 (13)	7.9 (10)	0.001
Smoker	4.3 (2)	12.7 (16)	0.160
LBBB	6.7 (3)	13.6 (17)	0.286
RBBB	4.4 (2)	6.4 (8)	0.733
Q waves	18.6 (8)	21.2 (25)	0.827
T inversions	54.8 (23)	55.8 (67)	1.000
ECG-LVH	19.0 (8)	16.7 (20)	0.813
Prolonged QTc	16.7 (7)	41.7 (50)	0.004
Permanent atrial fibrillation	13.0 (6)	12.8 (16)	1.000
CCS gradeb
1	21.3 (10)	37.3 (37)	0.068
2	53.2 (25)	42.1 (53)	0.230
3	25.5 (12)	20.6 (26)	0.537

Constant variables are expressed as either mean ± standard deviation or median (interquartile range).

BMI, body mass index; BP, blood pressure; LDL, low-density lipoprotein; BNP, B-type natriuretic peptide; hs-TnT, high-sensitivity troponin T; hs-CRP, high-sensitivity C-reactive protein; HbA1c, haemoglobin A1c; TSH, thyroid-stimulating hormone; LV, left ventricular; LBBB, left bundle branch block; RBBB, right bundle branch block; ECG, electrocardiogram; LVH, left ventricular hypertrophy; CCS, Canadian Cardiovascular Society.

^aIn the SYNTAX Score, ‘pre’ and ‘post’ specify whether the evaluation occurred before or after revascularization. b CCS class 4 is not represented, as it was an exclusion criterion in the original ARTEMIS study.

^bCCS class 4 is not represented, as it was an exclusion criterion in the original ARTEMIS study.

Based on angiography findings, the SYNTAX Score was determined via the SYNTAX Score website using version 2.11 of the web-based calculator (www.syntaxscore.com). The SYNTAX Score was assessed both pre- and post-revascularization. For PCI, the post-revascularization SYNTAX Score was calculated by removing lesions that were successfully treated from the original score. For CABG, the same principle applied: lesions that were successfully bypassed were excluded from the score. A graft was considered successful if it supplied the territory of the stenotic segment effectively. The left ventricular (LV) function and mass were assessed through echocardiography performed by experienced cardiologists. Left ventricular mass was calculated according to the following formula: LV mass = 0.8 × [1.04 {(LVIDd + PWTd + SWTd)³ − (LVIDd)³}] + 0.6 g.¹¹ Data on medication use, smoking status, comorbidities, previous strokes or transient ischaemic attacks, and prior coronary interventions were obtained from medical records and directly from the patients. Body weight, height, and waist circumference (measured at the midpoint between the lowest rib margin and the iliac crest) were also assessed.

Participants also underwent an exercise stress test, 24-h ambulatory electrocardiogram (ECG) monitoring, and a resting 12-lead ECG. A resting ECG was recorded for all patients and analysed by nine independent investigators. Various ECG measurements were collected, of which the current study focused specifically on T wave inversions, Q waves, bundle branch blocks (BBBs), heart rate, QTc intervals, permanent atrial fibrillation (AF), and signs of LV hypertrophy (LVH). Prolonged QTc was defined as a duration of ≥440 ms in men and ≥460 ms in women. T wave inversions were defined as inversions (>0.1 mV) present in at least two contiguous leads within the same anatomical region. Q waves were assessed based on the standard Minnesota Criteria. Left ventricular hypertrophy was identified if ECG findings met either the Cornell criteria (S wave in V3 and R wave in aVL > 20 mm) or the Sokolow–Lyon criteria (tallest R wave in V5 or V6 and S wave in V1 > 35 mm). All interval measurements were calculated as the mean intervals across all ECG leads.

Outcomes

The follow-up period for this study was 10 years. The follow-up concluded if death or aborted sudden cardiac arrest (SCA) occurred, whichever came first. For subjects who died during follow-up, the cause of death was categorized by MJ Junttila and HV Huikuri based on death certificates, which in turn were based on autopsy data, hospital records, and interviews with next of kin. Death certificates were obtained for all subjects who died during follow-up. In Finland, in cases of natural or expected death, the death certificate is issued by the attending physician, who compiles the available clinical and post-mortem information to determine the cause of death. Each certificate is subsequently reviewed by a forensic pathologist to ensure that the chain of events and diagnoses are internally consistent and valid. In cases of unexpected death, a medico-legal autopsy is performed by a forensic pathologist, and the cause of death is determined and death certificate written based on these findings.

The primary endpoint of this study was cardiac death, which included both SCDs and non-sudden cardiac deaths (NSCDs). Sudden cardiac deaths also encompassed SCAs. As secondary endpoints, SCDs and NSCDs were analysed separately. Death was classified as SCD if it was a witnessed sudden death occurring within 1 h of symptom onset or in unwitnessed cases, within 24 h of the last known alive time in a normal state of health.

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics version 21 (IBM Corp., New York, USA). Continuous variables were analysed using a two-tailed t-test for normally distributed data (skewness between −1 and +1) and the Mann–Whitney U-test for non-normally distributed data. Group comparisons for categorical variables were conducted using the χ² test. Univariate Cox regression analysis was conducted to evaluate the hazard ratios (HRs) of individual risk factors, and sex × risk factor interaction terms were included to evaluate sex differences in their associations with the outcome. Statistically significant risk factors from the univariate analysis were then included in a multivariate Cox regression model with backward stepwise elimination, where variables were sequentially removed until only those with significant associations (P < 0.20) remained. To assess the stability of the multivariable model, internal validation was performed using bootstrap resampling (1000 iterations), and the resulting estimates were compared with those of the original model. To specifically confirm the associations with cardiac death, a competing risk Cox regression analysis was also performed, treating non-cardiac death as a competing event. Results from the Cox regression were presented as HRs with 95% confidence intervals (CIs). Kaplan–Meier curves were used to illustrate survival across the groups under comparison. For survival analysis, the SYNTAX Score was categorized using an optimal cut-off value of ≥6, determined by ROC curve analysis.

Results

Baseline characteristics and clinical outcomes

The baseline characteristics of all subjects in the ARTEMIS cohort are detailed in Table 2. During follow-up, 17.0% (n = 105) of women and 22.8% (n = 302) of men experienced death or SCA (P = 0.002). Of these, 47 deaths in women (including 3 SCAs) and 126 deaths in men (including 14 SCAs) were of cardiac origin (Figure 1). Women showed better survival than men regarding non-cardiac mortality (P = 0.015), while survival related to cardiac mortality did not differ significantly between sexes (P = 0.115) (Figure 2). The mean age at the occurrence of cardiac death was 80.4 ± 7.1 years for women and 76.1 ± 8.3 years for men (P = 0.002). The baseline characteristics of women and men who experienced cardiac death during the follow-up are presented in Table 1.

Table 2.

Baseline characteristics of all subjects in the ARTEMIS cohort

	Women (n = 619)	Men (n = 1327)	P-value
Constant variables
Age (years)	68.9 ± 8.0	65.9 ± 8.7	<0.001
BMI (kg/m2)	28.6 ± 5.0	28.2 ± 4.3	0.098
Waist circumference (cm)	93.9 ± 14.1	101.5 ± 11.8	<0.001
Systolic BP (mmHg)	151.8 ± 25.0	143.2 ± 21.7	<0.001
Diastolic BP (mmHg)	75.6 ± 10.8	78.0 ± 10.4	<0.001
Mean heart rate (b.p.m.)	67.3 ± 8.6	67.7 ± 9.3	0.231
SYNTAX Score (pre)a	8.0 (3.0, 16.0)	11.0 (5.0, 21.0)	<0.001
SYNTAX Score (post)a	0.0 (0.0, 4.0)	1.0 (0.0, 6.0)	<0.001
Laboratory analyses
HbA1c (%)	6.1 (5.7, 6.6)	6.1 (5.7, 6.7)	0.658
Total cholesterol (mmol/L)	4.1 (3.6, 4.6)	3.7 (3.3, 4.3)	<0.001
HDL cholesterol (mmol/L)	1.4 ± 0.3	1.2 ± 0.3	<0.001
LDL cholesterol (mmol/L)	2.2 (1.8, 2.7)	2.1 (1.8, 2.6)	0.005
Triglycerides (mmol/L)	1.3 (1.0, 1.7)	1.2 (0.9, 1.7)	0.007
Creatinine clearance (mL/min)	83.0 ± 29.3	98.6 ± 35.3	<0.001
Urine albumin-creatinine ratio (mg/mmol)	1.1 (0.8, 1.7)	0.8 (0.5, 1.2)	<0.001
hs-CRP (mg/L)	1.0 (0.5, 2.2)	0.9 (0.5, 2.0)	0.072
hs-TnT (ng/L)	7.0 (5.0, 11.0)	9.0 (6.0, 14.0)	<0.001
BNP (ng/L)	59.0 (33.0, 105.0)	43.0 (22.0, 89.0)	<0.001
TSH (mU/L)	2.0 (1.3, 2.9)	1.9 (1.3, 2.6)	0.184
Haemoglobin (g/L)	132.5 ± 12.1	141.0 ± 13.9	<0.001
Echocardiography
LV ejection fraction (%)	66.1 ± 8.3	63.1 ± 9.6	<0.001
LV mass (g)	173.3 ± 46.3	223.7 ± 59.7	<0.001
Categorical variables, % (n)
Comorbidities
Type 2 diabetes	41.5 (257)	43.5 (577)	0.415
Hypercholesterolaemia	86.1 (531)	88.1 (1157)	0.239
Hypertension	76.9 (464)	67.1 (868)	<0.001
Previous stroke or TIA	11.5 (71)	10.4 (137)	0.479
Smoking status
Ex-smoker	21.6 (133)	52.8 (700)	<0.001
Current smoker	6.2 (38)	9.7 (129)	0.009
CAD history
Prior myocardial infarction	43.8 (271)	49.6 (658)	0.017
Prior PCI	64.2 (396)	61.1 (808)	0.192
Prior CABG	16.4 (101)	26.0 (344)	<0.001
CCS ≥ 2	55.9 (346)	37.2 (493)	<0.001
Medication
Oestrogen or testosterone therapy	20.8 (129)	5.2 (69)	<0.001
Statins	90.6 (561)	91.8 (1217)	0.434
Calcium channel blockers	25.2 (156)	24.1 (319)	0.610
Diuretics	42.8 (265)	30.3 (402)	<0.001
ACE inhibitors or ATR blockers	70.1 (434)	67.6 (897)	0.295
Beta-blockers	88.0 (545)	87.5 (1160)	0.767
Diabetes medication	33.5 (207)	35.9 (475)	0.332
Insulin	11.3 (69)	11.7 (153)	0.819
Oral antidiabetic medication + insulin	7.9 (49)	7.9 (104)	1.000

Constant variables are expressed as either mean ± standard deviation or median (interquartile range).

BMI, body mass index; BP, blood pressure; BNP, B-type natriuretic peptide; hs-TnT, high-sensitivity troponin T; hs-CRP, high-sensitivity C-reactive protein; HbA1c, haemoglobin A1c; TSH, thyroid-stimulating hormone; LV, left ventricular; TIA, transient ischaemic attack; CAD, coronary artery disease; PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting; CCS, Canadian Cardiovascular Society; ACE, angiotensin-converting enzyme; ATR, angiotensin receptor blocker.

^aIn the SYNTAX Score, ‘pre’ and ‘post’ specify whether the evaluation occurred before or after revascularization.

Figure 1.

Flow chart of study population by sex and cardiac mortality. SCD, sudden cardiac death; SCA, sudden cardiac arrest; NSCD, non-sudden cardiac death.

Figure 2.

Survival by sex for cardiac and non-cardiac causes of death. Kaplan–Meier survival curves for cardiac and non-cardiac causes of death in women and men. NCD, non-cardiac death.

Risk factors for cardiac death

All variables listed in Table 1 were included in univariate regression analysis. Of these, the variables showing significant univariate associations with cardiac death in women included older age, systolic BP, pre- and post-revascularization SYNTAX Scores, CrCl, BNP, HbA1c, hs-TnT, T2DM, CCS Class 2–3 angina, the presence of Q waves, T wave inversions, and permanent AF (see Supplementary material online, Table S1). When NSCD and SCD were analysed separately, systolic BP, CrCl, HbA1c, hs-TnT, and T wave inversions were associated only with NSCD, whereas age, AF, CCS Class 3, and BNP were linked to both endpoints in women (see Supplementary material online, Table S2). In the sex interaction analysis, higher systolic BP was associated with increased cardiac death risk only in women. Higher CrCl was more protective in women than in men, whereas higher LVEF was protective only in men. Electrocardiogram-defined LVH was similarly associated with increased cardiac death risk only in men (see Supplementary material online, Table S1).

The variables with significant univariate association with cardiac death in women were subsequently entered into the multivariable model, where older age, higher post-revascularization SYNTAX Score, higher systolic BP, HbA1c, hs-TnT, and permanent AF remained associated with cardiac death (Table 3). In a competing risk Cox regression model, these associations were still observed and remained statistically significant after accounting for competing non-cardiac mortality (see Supplementary material online, Table S3). Internal validation using bootstrap resampling (1000 samples) confirmed the stability of the multivariable model, with preserved direction and similar magnitude of the regression coefficients compared with the primary analysis. Kaplan–Meier curves for systolic BP, post-revascularization SYNTAX Score, hs-TnT, and AF are presented in Figure 3. A ROC analysis identified a post-revascularization SYNTAX Score ≥ 6 as the optimal cut-off for survival analyses.

Table 3.

Multivariate analysis of risk factors for cardiac death in women

Risk factor	Adjusted HRa (95% CI)	P-value
Age	1.12 (1.05–1.19)	<0.001
SYNTAX Score (pre)b	NI
SYNTAX Score (post)b	1.05 (1.01–1.08)	0.005
Systolic blood pressure (10-unit increase)	1.14 (1.01–1.28)	0.039
Creatinine clearance	NI
HbA1c	1.44 (1.07–1.94)	0.016
BNP (10-unit increase)	1.02 (1.00–1.04)	0.076
hs-TnT (10-unit increase)	1.47 (1.12–1.94)	0.006
Type 2 diabetes	NI
CCS Grades 2–3	NI
Permanent atrial fibrillation	4.91 (1.61–15.03)	0.005
Q waves	1.96 (0.76–5.05)	0.165
T inversions	NI

CI, confidence interval; HbA1c, haemoglobin A1c; BNP, B-type natriuretic peptide; hs-TnT, high-sensitivity troponin T; CCS, Canadian Cardiovascular Society; ECG, electrocardiogram.

^aThe hazard ratios (HRs) presented are adjusted for the variables that remained in the final model. NI indicates a variable that was initially entered into the backward stepwise multivariable logistic regression model but was excluded from the final model due to a P > 0.20.

^bIn the SYNTAX Score, ‘pre’ and ‘post’ specify whether the evaluation occurred before or after revascularization.

Figure 3.

Survival curves for association of cardiac risk factors with cardiac death in women. Kaplan–Meier survival curves for cardiac death in women with coronary artery disease, stratified by presence of elevated blood pressure (BP, mmHg), high-sensitivity Troponin T (hs-TnT, ng/L), post-revascularization SYNTAX Score, and atrial fibrillation. Cut-off values used for categorization are shown within the figure.

Discussion

Previous studies have highlighted the need for more data on factors affecting prognosis in women with CAD. In this study, we aimed to identify potential signals of factors influencing their outcomes. We found that increasing post-revascularization SYNTAX Score was associated with cardiac mortality, supporting the relevance of residual coronary disease burden in this population. In addition to age, clinical and biochemical factors including systolic BP, hs-TnT, and HbA1c were also associated with cardiac death. Permanent AF likewise demonstrated a potential association with increased risk, although the relatively wide CI suggests that further studies are needed to better define its prognostic significance.

Post-revascularization SYNTAX Score is a significant predictor of cardiac death and other adverse outcomes in CAD patients.^12,13 Our study showed a trend that higher scores may be associated with an increased risk of cardiac death in women, with roughly a 5% increase per one-unit rise. Importantly, post-revascularization SYNTAX Score remained independently associated with cardiac death after adjustment for age and comorbidities. According to previous studies, although the excess risk is most evident at higher residual SYNTAX values, prognostic relevance has also been observed across a broader range of residual disease burden and is not limited to patients exceeding conventional high residual SYNTAX thresholds (e.g. post-revascularization SYNTAX Score ≥ 8).^14,15 In addition, Hayase et al.¹⁶ demonstrated that the prognostic impact of residual SYNTAX Score depends not only on the amount of residual anatomical disease but also on the degree of ischaemic improvement achieved after revascularization. Importantly, post-revascularization SYNTAX Score should not be interpreted solely as a measure of residual coronary anatomy, as it also reflects treatment selection and procedural decision-making factors, such as lesion complexity, technical feasibility, and patient-related considerations influencing revascularization strategies.^14,15 In women, these factors may be particularly relevant, as their older age, higher rates of comorbidities, and higher bleeding risk can influence treatment decisions.⁶

Although sex-based differences in outcomes and responses to revascularization have been reported, the SYNTAX Score II—incorporating female sex and clinical variables to improve individualized risk prediction—remains less commonly used in clinical practice than the original anatomy-based model.^17,18 While women generally present with lower SYNTAX Score,¹² this study found no sex differences in pre- or post-revascularization SYNTAX Scores among subjects with cardiac death. This may be attributable to the relatively high proportion of subjects with T2DM in this population, as in women, T2DM is often linked to a higher burden of comorbidities as well as more pronounced atherosclerotic changes.¹⁹ In addition, women in this population were older, which may have influenced both the extent of their coronary disease and the likelihood of undergoing revascularization procedures.

Among subjects with cardiac death during the 10-year follow-up, women had higher systolic BP level compared to men, and in univariate analysis, systolic BP had stronger association with cardiac death in women. This association persisted in the multivariable model, indicating an estimated ∼10% increase in risk per 10 mmHg rise. Previous studies have shown elevated systolic BP to be more strongly associated with an increased risk of cardiac morbidity and mortality in women compared to men.²⁰ Moreover, women may experience CAD events at lower BP thresholds than men.^20,21 This disparity is likely due to multifactorial causes, including physiological differences in vascular and myocardial responses, hormonal influences, and a higher likelihood of developing heart failure with preserved ejection fraction in hypertensive women.^20,22 Previous research, as seen also in our study, has highlighted that women often have worse BP control and are less likely to receive anti-hypertensive medication, suggesting that improving hypertension management could be a cost-effective strategy to reduce cardiac mortality in women.²⁰

In this study, elevated hs-TnT levels showed a significant association with cardiac death in women with CAD. A previous study²³ showed that elevated hs-TnT levels 1 year after an initial event like PCI or diagnostic angiography significantly predicted cardiac death. Prior studies have also shown that elevated hs-TnT levels are stronger predictors of cardiovascular events in women than in men.²⁴ One proposed explanation is the use of similar hs-TnT thresholds for both sexes, despite women generally having lower hs-TnT levels compared to men, as observed also in our study. As a result, women reaching the threshold may be in a more advanced disease state. Therefore, adopting sex-specific cut-offs for hs-TnT could have potential benefits, as highlighted in previous studies.^24,25 In addition, despite women with CAD often present with less severe angiographic findings, they are more likely to have additional comorbidities such as heart failure, T2DM, and hypertension, which may also affect hs-TnT levels and worsen their overall prognosis.²⁴ The incorporation of hs-TnT levels into secondary prevention strategies for CAD patients has been suggested in order to help identify those at higher risk for adverse outcomes and enabling more targeted follow-up care.²³

In this study, permanent AF appeared to be linked to cardiac death in women with CAD, encompassing both SCD and NSCD. While based on a limited number of events and accompanied by a relatively wide CI, the observed association aligns with previous research showing that the presence of AF attenuates the typical survival advantage observed in women.^26,27 For instance, a meta-analysis of 30 studies found that AF was associated with nearly double the cardiovascular mortality risk in women compared to men.²⁷ Proposed factors contributing to this worse prognosis in women include their higher heart rates during AF episodes, higher prevalence of left atrial fibrosis, and less favourable responses to treatments such as rhythm control and catheter ablation.²⁸ Additionally, women with AF often present with a higher burden of comorbidities, suggesting that improving their outcomes requires a comprehensive approach that addresses the overall disease burden.^29,30

Higher HbA1c has been consistently associated with increased cardiac mortality across multiple cohorts.^31,32 In the PRE-DETERMINE study of CAD patients not meeting criteria for ICD implantation, higher HbA1c was linked to an increased risk of cardiac death, with a higher risk for NSCD than SCD.³² Similarly, in our study of women with CAD, HbA1c remained significantly associated with cardiac death even after adjusting for DM in the multivariate model. In univariate analysis, HbA1c was associated with NSCD but not SCD, suggesting that chronic glycaemic exposure may drive progressive structural or ischaemic changes that are particularly relevant in non-sudden cardiac events.

Strengths and limitations

This study includes a cohort of patients with angiographically verified CAD, with a broad range of data on cardiovascular risk factors at the start of follow-up. However, there are some limitations to acknowledge.

First, the number of cardiac deaths among women was relatively limited, which may reduce statistical power and increase uncertainty in multivariable estimates. For example, the CI for AF was relatively wide, reflecting the limited number of events. Given the exploratory aim of the study, we included a broad set of variables in the multivariate analysis to investigate potential associations, including weaker or less obvious factors, to provide initial insights and generate hypotheses for future studies. Although clinically relevant variables were included in the model, the ratio of events to predictors may introduce a risk of model overfitting. However, when we performed internal validation using bootstrap resampling, it demonstrated consistent direction and magnitude of the observed associations. However, further studies in larger cohorts are required to strengthen and extend these observations.

In the ARTEMIS cohort, approximately 40% of subjects have T2DM, whereas the prevalence of T2DM among CAD patients in the general European population is around 30%.³³ This difference may impact the study results. In addition, at baseline, medical therapy for both DM and coronary atherosclerosis was optimized, but after baseline examinations, patients were managed by their local physicians, who adjusted therapy as clinically indicated. Therefore, long-term medication use was encouraged but could not be systematically monitored. However, medication use and adherence were reassessed after 2 years of follow-up. At the 2-year follow-up, adherence to prescribed therapies remained high, with ≥85% of patients in both sexes continuing their medications as intended.

Conclusions

In this study, increasing post-revascularization SYNTAX Score was associated with cardiac death in women with CAD, underscoring the prognostic relevance of residual coronary disease burden. In addition to age, systolic BP, high-sensitivity troponin T, and HbA1c were also significantly associated with cardiac mortality. Interestingly, permanent AF appeared to be associated with increased risk as well, suggesting a potential prognostic role that warrants further investigation.

Collectively, these findings highlight that multiple clinical and biochemical factors may contribute independently to cardiac death risk in women and may offer tools for future studies further analysing these associations.

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Supplementary material

Supplementary material is available at European Heart Journal Open online.

Ethical approval

The study adhered to the Declaration of Helsinki, and the protocol received approval from the local research ethics committee of the Northern Ostrobothnia Hospital District. All study participants provided written informed consent.

Funding

Sigrid Juselius Foundation, Finnish Foundation for Cardiovascular Research, Academy of Finland, The Finnish Medical Foundation, Aarne Koskelo Foundation, Paavo Nurmi Foundation, and Päivikki and Sakari Sohlberg Foundation.