Coronary artery disease often develops silently before myocardial infarction, and current risk-based prevention pathways miss some individuals with clinically important disease, particularly those without standard modifiable cardiovascular risk factors. High-sensitivity cardiac troponin I is a blood biomarker of low-level myocardial injury, but its relationship to silent coronary atherosclerosis remains incompletely understood. Here we show, in 1,613 adults undergoing coronary computed tomography (CT) imaging, that low-level high-sensitivity cardiac troponin I is associated with clinically actionable coronary disease, defined by coronary calcium score ≥100, and with quantitative CT markers of plaque burden. A threshold of 2.3 ng l-1 enriched for actionable disease, especially in individuals without standard risk factors. Normalizing troponin for myocardial mass, age and renal function improved prediction and provided mechanistic insight into baseline troponin elevation in chronic coronary syndromes. A two-stage strategy using troponin before CT imaging increased diagnostic yield and reduced the number of scans required. Prospective validation is warranted before clinical implementation.

BACKGROUND:Serum troponin measurement forms a cornerstone of acute myocardial infarction (AMI) diagnosis. A major challenge is interpretation of an elevated first troponin in patients with impaired renal function. We aimed to (1) evaluate the relationship between estimated glomerular filtration rate (eGFR) and first troponin, (2) characterise the performance of different troponin assays for diagnosing AMI and (3) derive eGFR-specific thresholds for potential clinical use. METHODS:We analysed the distribution of troponin values stratified by eGFR and AMI. Diagnostic performance was analysed using the C-statistic. Test detection rate, false positive rate and positive predictive value were calculated for different cut-offs. RESULTS:We included 221 175 patients between 2010 and 2017 from four acute tertiary care hospitals in London, UK, with a median age of 65 years (IQR 49-79). eGFR was<60 mL/min/1.73 m2 in 20.6% of patients and 6.4% of patients had a diagnosis of AMI. In patients without AMI, we observed an inverse log-linear relationship between eGFR and troponin. Diagnostic performance for AMI was best in patients with eGFR>90 (C-statistic 0.93) and worst in eGFR<15 (C-statistic 0.81). For high-sensitive troponin T, using the conventional cut-off of 14 ng/L, false positive rates ranged from 68-93% for eGFRs between 15 and 60 mL/min/1.73 m2. Restricting the false positive rate to 15% yields eGFR specific cut-offs of 73, 112 and 184 ng/L, with detection rates of 73%, 70% and 68% in patients with an eGFR of 45-60, 30-45 or 15-30 mL/min/1.73 m2. CONCLUSIONS:The diagnostic performance of an unadjusted troponin cut-off for AMI falls with worsening renal function. We propose consideration of eGFR specific cut-offs to support more effective triage and early management of suspected AMI in patients with renal impairment. TRIAL REGISTRATION NUMBER:NCT03507309.

Cardiac disease remains the primary cause of mortality worldwide. Waiting times for standard echocardiography are lengthening due to growing needs of ageing populations and shortage of trained sonographers. Point-of-care ultrasound (POCUS) has emerged as a portable, cost-effective alternative, allowing quicker evaluation while waiting for traditional scans to rule out life-threatening pathology. POCUS is often used by clinicians without full echocardiography training. In addition, POCUS images are of lower quality and only used for qualitative cardiac evaluation, particularly left-ventricular systolic function. Deep learning could enhance its capabilities, but no public POCUS dataset currently exists to support it. Here, we introduce PocketLV, an opensource, multi-view POCUS echocardiography dataset with left-ventricular segmentations, covering patients with varied cardiac health, where image quality reflects real-world acquisition by a minimally-trained sonographer. To demonstrate PocketLV enables deep learning studies and establish baseline models, we benchmark state-of-the-art segmentation algorithms, including discriminative and generative models, finding nnU-Net performs best. The dataset is available at https://datacompass.lshtm.ac.uk/id/eprint/4569.

Three-dimensional (3D) mesh reconstruction of the cardiac anatomy from medical images is useful for shape and motion measurements and biophysics simulations. However, 3D medical images are often acquired as 2D slices that are sparsely sampled (e.g., large slice spacing) and noisy, and 3D mesh reconstruction on such data is a challenging task. Traditional voxel-based approaches utilize non-differentiable pre- and post-processing that compromises fidelity to images, while mesh-level deep learning approaches require large 3D mesh annotations that are difficult to obtain. Differentiable cross-domain supervision from 2D images to 3D meshes is therefore crucial for enabling end-to-end optimization in medical imaging. While there have been attempts to approximate the voxelization and slicing of meshes that are being optimized, there has not yet been a method for directly using 2D slices to supervise 3D mesh reconstruction in a differentiable manner. Here, we propose a novel explicit differentiable voxelization and slicing (DVS) algorithm allowing gradient backpropagation to a 3D mesh from its slices, which facilitates refined mesh optimization directly supervised by the losses defined on 2D images. Further, we propose an innovative framework for extracting patient-specific left ventricle (LV) meshes from medical images by coupling DVS with a graph harmonic deformation (GHD) mesh morphing descriptor of cardiac shape that naturally preserves mesh quality and smoothness during optimization. The proposed framework achieves state-of-the-art performance in cardiac mesh reconstruction tasks from densely sampled (CT) as well as sparsely sampled (MRI stack with few slices) images, outperforming alternatives, including Marching Cubes, statistical shape models, algorithms with vertex-based mesh morphing algorithms and alternative methods for image-supervision of mesh reconstruction. Experimental results demonstrate that our method achieves an overall Dice score of 90% during a sparse fitting on multi-datasets. The proposed method can further quantify clinically useful parameters such as ejection fraction and global myocardial strains, closely matching the ground truth and outperforming the traditional voxel-based approach in sparse images.

Current Acute Coronary Syndromes (ACS) rule-out algorithms rely on a combination of clinical assessment and measuring troponin levels. It can take several hours for troponin levels to rise after a myocardial infarction, so initial testing may not show detectable levels of troponin. In order to rule out a false negative result, troponin levels are typically tested again several hours later to look for rising values meaning patients are admitted for observation which has a large resource implication. We developed a machine learning model aimed at improving early discharge at initial assessment. The study was conducted using data from the National Institute for Health Research Health Informatics Collaborative Cardiovascular dataset.(1,2) We trained and tuned a machine learning model (Rapid-RO) using patient data from two separate hospitals to rule-out ACS with simple routine demographic or clinical measurements. The model was then tested for its predictive accuracy in cohorts of patients at four different hospitals from separate time periods. The model was assessed against troponin threshold guided management as recommended by the European Society of Cardiology clinical guidelines. The patient cohorts of the six derived datasets are presented in Figure 1. On the left side are the training and tuning cohorts and the right side are the cohorts from which the derived model was tested. The Rapid-RO machine learning model included input from 11 inputs that had the highest feature importance, including troponin, age, C-reactive protein, urea, platelet count, eGFR, white cell count, haemoglobin, heart failure, diabetes, and hypertension. The Rapid-RO model identified 12037 (35.69%) very low risk patients on top of standard clinical assessment who could have been discharged early, compared with 8967 (26.58%) identified by a troponin threshold approach alone (Figure 2), with significantly fewer missed ACS cases (27 (0.22%) vs. 108 (1.20%)) and similar mortality rates (2 (0.02%) vs. 4 (0.04%) at 30 days). The Rapid-RO model demonstrated a consistently higher rule-out rate for ACS with a lower missed ACS rate across patient subsets, including patients with and without chest pain or COVID-19. The Rapid-RO machine learning model, which uses patient history and initial blood tests, offers a significant advancement in the risk stratification process, presenting a reliable tool for clinicians to rapidly rule out ACS and potentially reduce unnecessary hospital admissions. Its robust performance in diverse patient groups across different time periods, underscores its potential utility in a real-world clinical setting.Figure 1 Figure 2

Aim In the UK, pharmacological management of patients with heart failure (HF) occurs predominantly in general practice. Using data from the Clinical Practice Research Datalink, we examined the prevalence and risk factors for medication non-adherence and its association with hospitalisation and mortality over a 9-year period.Methods A retrospective cohort study of 127 927 patients, ≥18 years old in England with incident HF diagnosed during 1 January 2009 to 31 December 2018. We evaluated non-adherence to any ACE inhibitor, angiotensin receptor blocker, β-blocker or mineralocorticoid receptor antagonist, over 24 months. Non-adherence was based on proportion of days covered (PDC) and defined as PDC<80%. Risk factors for non-adherence and all-cause mortality were examined using multiple logistic regression and Cox regression, respectively. Rates of any-cause emergency hospitalisations, cardiovascular disease (CVD) and HF mortality was estimated using Fine-Gray competing risk models. PDC was also assessed as a continuous variable.Results About 43.6% of patients were non-adherent to therapy. Crude rates of emergency admissions, all-cause, CVD and HF mortality overall were 306.8/1000, 119.6/1000, 44.6/1000 and 3.3/1000 person-years, respectively. The strongest predictor of non-adherence was any-cause hospitalisation ≤12 months prior. Non-adherence was associated with a higher rate of all-cause mortality (HR 1.31, 95% CI 1.28 to 1.33) and significantly associated with CVD-related mortality (subdistribution HR (SHR) 1.20, 95% CI 1.16 to 1.23), HF deaths (SHR 1.18, 95% CI 1.05 to 1.32) and any-cause emergency admissions (SHR 1.11, 95% CI 1.10 to 1.13). In the analysis treating PDC as a continuous variable, every 10% decrease in PDC levels was associated with a 6% increased hazard of all-cause mortality (HR 1.06, 95% CI 1.05 to 1.06) and was significantly associated with CVD, but not HF mortality.Conclusion Medication non-adherence over 24 months was relatively high and associated with poorer health outcomes. Interventions to improve adherence among patients with HF are needed.

Background Left ventricular diastolic function as assessed by tissue Doppler echocardiography predicts cardiovascular event rates at 4 years of follow-up in patients with hypertension. Our aim was to evaluate whether this extends to predicting cardiovascular mortality after 20 years of follow-up.Methods Conventional (E) and tissue Doppler (e′) echocardiography was performed on hypertensive participants in the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT) with long-term follow-up ascertained via linkage to the Office of National Statistics. Cardiovascular mortality was defined as death from coronary heart disease, stroke and other cardiovascular aetiology such as heart failure or peripheral vascular disease. Unadjusted and adjusted Cox regression survival models were constructed to investigate the association between tissue Doppler echocardiography measurements and long-term cardiovascular mortality.Results Among 506 hypertensive patients (median age 64, interquartile range (58, 69), 87% male), there were 200 (40%) deaths over a 20-year follow-up period. 60 deaths (12%) were cardiovascular-related.A reduction in e′ was independently associated with increased cardiovascular mortality, after adjusting for the ACC/AHA Atherosclerotic Cardiovascular Disease (ASCVD) risk score, with an inverse HR of 1.22 per 1 cm/s decrease (95% CI 1.04–1.43). A higher E/e′ ratio was independently associated with increased cardiovascular mortality, after adjusting for the ASCVD risk score, with an HR of 1.12 per 1-unit increase (95% CI, 1.02 to 1.23).Conclusions Impaired left ventricular diastolic function, measured using tissue Doppler echocardiography through e′ and E/e′, independently predicts increased cardiovascular mortality over 20 years in hypertensive patients, highlighting its long-term prognostic significance.

This study investigated the sensitivity and specificity of identifying heart failure with reduced ejection fraction (HFrEF) from measurements of the intensity and timing of arterial pulse waves. Previously validated methods combining ultrafast B-mode ultrasound, plane-wave transmission, singular value decomposition (SVD), and speckle tracking were used to characterize the compression and decompression ("S" and "D") waves occurring in early and late systole, respectively, in the carotid arteries of outpatients with left ventricular ejection fraction (LVEF) < 40%, determined by echocardiography, and signs and symptoms of heart failure, or with LVEF >= 50% and no signs or symptoms of heart failure. On average, the HFrEF group had significantly reduced S-wave intensity and energy, a greater interval between the R wave of the ECG and the S wave, a reduced interval between the S and D waves, and an increase in the S-wave shift (SWS), a novel metric that characterizes the shift in timing of the S wave away from the R wave of the ECG and toward the D wave (all P < 0.01). Receiver operating characteristics (ROCs) were used to quantify for the first time how well wave metrics classified individual participants. S-wave intensity and energy gave areas under the ROC of 0.76-0.83, the ECG-S-wave interval gave 0.85-0.88, and the S-wave shift gave 0.88-0.92. Hence the methods, which are simple to use and do not require complex interpretation, provide sensitive and specific identification of HFrEF. If similar results were obtained in primary care, they could form the basis of techniques for heart failure screening.

Abstract Background A raised stress hyperglycaemia ratio (SHR) has been associated with all-cause mortality and may better discriminate than an absolute glucose value. The aim of this meta analysis and systematic review is to synthesise the evidence assessing the relationship between the SHR and all-cause mortality across three common cardiovascular presentations. Methods We undertook a comprehensive search of Medline, Embase, Cochrane CENTRAL and Web of Science from the date of inception to 1st March 2024, and selected articles meeting the following criteria: studies of patients hospitalised for acute myocardial infarction, ischaemic stroke or acute heart failure reporting the risk (odds ratio or hazard ratio) for all-cause mortality associated with the SHR. A random effects model was used for primary analysis. Subgroup analysis by diabetes status and of mortality in the short and long term was undertaken. Risk of bias assessment was performed using the Newcastle Ottawa quality assessment scale. Results A total of 32 studies were included: 26 studies provided 31 estimates for the meta-analysis. The total study population in the meta analysis was 80,010. Six further studies were included in the systematic review. Participants admitted to hospital with cardiovascular disease and an SHR in the highest category had a significantly higher risk ratio of all-cause mortality in both the short and longer term compared with those with a lower SHR (RR = 1.67 [95% CI 1.46–1.91], p < 0.001). This finding was driven by studies in the myocardial infarction (RR = 1.75 [95% CI 1.52–2.01]), and ischaemic stroke cohorts (RR = 1.78 [95% CI 1.26–2.50]). The relationship was present amongst those with and without diabetes (diabetes: RR 1.49 [95% CI 1.14–1.94], p < 0.001, no diabetes: RR 1.85 [95% CI 1.49–2.30], p < 0.001) with p = 0.21 for subgroup differences, and amongst studies that reported mortality as a single outcome (RR of 1.51 ([95% CI 1.29–1.77]; p < 0.001) and those that reported mortality as part of a composite outcome (RR 2.02 [95% CI 1.58–2.59]; p < 0.001). On subgroup analysis by length of follow up, higher SHR values were associated with increased risk of mortality at 90 day, 1 year and > 1year follow up, with risk ratios of 1.84 ([95% CI 1.32–2.56], p < 0.001), 1.69 ([95% CI 1.32–2.16], p < 0.001) and 1.58 ([95% CI 1.34–1.86], p < 0.001) respectively. Conclusions A raised SHR is associated with an increased risk of all-cause mortality following myocardial infarction and ischaemic stroke. Further work is required to define reference values for the SHR, and to investigate the potential effects of relative hypoglycaemia. Interventional trials targeting to the SHR rather than the absolute glucose value should be undertaken. PROSPERO database registration CRD 42023456421 https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023456421

Abstract Background Heart failure (HF) with preserved or mildly reduced ejection fraction includes a heterogenous group of patients. Reclassification into distinct phenogroups to enable targeted interventions is a priority. This study aimed to identify distinct phenogroups, and compare phenogroup characteristics and outcomes, from electronic health record data. Methods 2,187 patients admitted to five UK hospitals with a diagnosis of HF and a left ventricular ejection fraction ≥ 40% were identified from the NIHR Health Informatics Collaborative database. Partition-based, model-based, and density-based machine learning clustering techniques were applied. Cox Proportional Hazards and Fine-Gray competing risks models were used to compare outcomes (all-cause mortality and hospitalisation for HF) across phenogroups. Results Three phenogroups were identified: (1) Younger, predominantly female patients with high prevalence of cardiometabolic and coronary disease; (2) More frail patients, with higher rates of lung disease and atrial fibrillation; (3) Patients characterised by systemic inflammation and high rates of diabetes and renal dysfunction. Survival profiles were distinct, with an increasing risk of all-cause mortality from phenogroups 1 to 3 (p < 0.001). Phenogroup membership significantly improved survival prediction compared to conventional factors. Phenogroups were not predictive of hospitalisation for HF. Conclusions Applying unsupervised machine learning to routinely collected electronic health record data identified phenogroups with distinct clinical characteristics and unique survival profiles.

Background Cardiac troponin is commonly raised in patients presenting with malignancy. The prognostic significance of raised troponin in these patients is unclear. Objectives We sought to investigate the relation between troponin and mortality in a large, well characterised cohort of patients with a routinely measured troponin and a primary diagnosis of malignancy. Methods We used the National Institute for Health Research (NIHR) Health Informatics Collaborative data of 5571 patients, who had troponin levels measured at 5 UK cardiac centres between 2010 and 2017 and had a primary diagnosis of malignancy. Patients were classified into solid tumour or haematological malignancy subgroups. Peak troponin levels were standardised as a multiple of each laboratory's 99th -percentile upper limit of normal (xULN). Results 4649 patients were diagnosed with solid tumours and 922 patients with haematological malignancies. Raised troponin was an independent predictor of mortality in all patients (Troponin > 10 vs. <1 adjusted HR 2.01, 95% CI 1.73 to 2.34), in solid tumours (HR 1.84, 95% CI 1.55 to 2.19), and in haematological malignancy (HR 2.72, 95% CI 1.99 to 3.72). There was a significant trend in increasing mortality risk across troponin categories in all three subgroups (p < 0.001). Conclusion Raised troponin level is associated with increased mortality in patients with a primary diagnosis of malignancy regardless of cancer subtype. Mortality risk is stable for patients with a troponin level below the ULN but increases as troponin level increases above the ULN in the absence of acute coronary syndrome.

ObjectiveThe COVID-19 pandemic was associated with a reduction in the incidence of myocardial infarction (MI) diagnosis, in part because patients were less likely to present to hospital. Whether changes in clinical decision making with respect to the investigation and management of patients with suspected MI also contributed to this phenomenon is unknown.MethodsMulticentre retrospective cohort study in three UK centres contributing data to the National Institute for Health Research Health Informatics Collaborative. Patients presenting to the Emergency Department (ED) of these centres between 1st January 2020 and 1st September 2020 were included. Three time epochs within this period were defined based on the course of the first wave of the COVID-19 pandemic: pre-pandemic (epoch 1), lockdown (epoch 2), post-lockdown (epoch 3).ResultsDuring the study period, 10,670 unique patients attended the ED with chest pain or dyspnoea, of whom 6,928 were admitted. Despite fewer total ED attendances in epoch 2, patient presentations with dyspnoea were increased (p < 0.001), with greater likelihood of troponin testing in both chest pain (p = 0.001) and dyspnoea (p < 0.001). There was a dramatic reduction in elective and emergency cardiac procedures (both p < 0.001), and greater overall mortality of patients (p < 0.001), compared to the pre-pandemic period. Positive COVID-19 and/or troponin test results were associated with increased mortality (p < 0.001), though the temporal risk profile differed.ConclusionsThe first wave of the COVID-19 pandemic was associated with significant changes not just in presentation, but also the investigation, management, and outcomes of patients presenting with suspected myocardial injury or MI.

IntroductionEjection fraction (EF) is widely used to evaluate heart function during heart failure (HF) due to its simplicity compared but it may misrepresent cardiac function during ventricular hypertrophy, especially in heart failure with preserved EF (HFpEF). To resolve this shortcoming, we evaluate a correction factor to EF, which is equivalent to computing EF at the mid-wall layer (without the need for mid-layer identification) rather than at the endocardial surface, and thus better complements other complex metrics.MethodThe retrospective cohort data was studied, consisting of 2,752 individuals (56.5% male, age 69.3 ± 16.4 years) admitted with a request of a troponin test and undergoing echocardiography as part of their clinical assessment across three centres. Cox-proportional regression models were constructed to compare the adjusted EF (EFa) to EF in evaluating risk of heart failure admissions.ResultComparing HFpEF patients to non-HF cases, there was no significant difference in EF (62.3 ± 7.6% vs. 64.2 ± 6.2%, p = 0.79), but there was a significant difference in EFa (56.6 ± 6.4% vs. 61.8 ± 9.9%, p = 0.0007). Both low EF and low EFa were associated with a high HF readmission risk. However, in the cohort with a normal EF (EF ≥ 50%), models using EFa were significantly more associative with HF readmissions within 3 years, where the leave one out cross validation ROC analysis showed a 18.6% reduction in errors, and Net Classification Index (NRI) analysis showed that risk increment classification of events increased by 12.2%, while risk decrement classification of non-events decreased by 16.6%.ConclusionEFa is associated with HF readmission in patients with a normal EF.

Abstract Background Percutaneous coronary intervention (PCI) for stable angina is currently recommended for patients with residual symptoms despite maximal anti-anginal medication. Placebo-controlled data demonstrating efficacy of PCI has recently become available. Purpose We evaluated the placebo-controlled efficacy of anti-anginal agents and PCI, as monotherapies for stable angina. Methods We performed a systematic review and meta-analysis of placebo-controlled randomised controlled trials (RCTs) of anti-anginal agents (17 medications across 7 drug classes [beta-blocker, dihydropyridine (DHP) calcium channel blocker (CCB), non-DHP CCB, ivabradine, nicorandil, nitrates and ranolazine]), and PCI, which reported the placebo-controlled treatment effect on standardised exercise duration in patients with stable angina, who were receiving no background anti-anginal medications. We searched MEDLINE, EMBASE, and Cochrane CENTRAL from database inception to 01 December 2023. Relevant systematic reviews and reference lists were inspected to identify publications which had not been identified in the primary searches. Two independent reviewers extracted the data and assessed the risk of bias of included studies. Intention-to-treat data were used. Meta-analyses were carried out using random-effects methods and reported using the PRISMA guidelines. The primary outcome was mean difference in total exercise time with treatment compared with placebo. A pairwise comparison of all anti-anginal medications and PCI was performed to evaluate their relative placebo-controlled efficacy on total exercise time. Results A total of 52 trials were eligible, evaluating 83 treatment arm comparisons of anti-anginal monotherapy (2408 patients) and placebo (2444 patients). The number of trials per medication class ranged from one (ivabradine) to 19 (non-DHP CCBs). All classes improved exercise time compared to placebo, with mean increment ranging from 22.4 seconds (95% confidence interval, 0.8 to 44.0, p=0.04) for ranolazine, to 76.5 seconds (63.2 to 89.8, p<0.0001) for DHP CCBs (Figure 1). There was substantial heterogeneity in exercise time increment between drug classes (I2 98%, p<0.0001). However, across all classes, there was a clear dose-response relationship (b-coefficient 24.8, 15.8 to 33.9, <0.0001). The only trial which compared PCI (151 patients) to placebo (150 patients) showed a placebo-controlled exercise duration increment of 59.5 seconds (16 to 103, P=0.001). This was comparable to the global mean effect across all anti-anginal medications (59, 39 to 79, P<0.0001). Conclusions Each anti-anginal drug class showed a clear increase in placebo-controlled exercise time, with varying effect sizes between classes. The DHP CCB class showed the largest effect size, and ivabradine and ranolazine showed the smallest. The effect of PCI monotherapy was similar to the mean anti-anginal medication monotherapy effect.Figure 1

Background Reduced estimated glomerular filtration rate (eGFR) is associated with lower use of invasive management and increased mortality after acute coronary syndrome (ACS). The reasons for this are unclear.Methods A retrospective clinical cohort study was performed using data from the English National Institute for Health Research Health Informatics Collaborative (2010–2017). Multivariable logistic regression was used to investigate whether eGFR<90 mL/min/1.73 m2 was associated with conservative ACS management and test whether (a) differences in care could be related to frailty and (b) associations between eGFR and mortality could be related to variation in revascularisation rates.Results Among 10 205 people with ACS, an eGFR of <60 mL/min/1.73m2 was found in 25%. Strong inverse linear associations were found between worsening eGFR category and receipt of invasive management, on a relative and absolute scale. People with an eGFR <30 mL compared with ≥90 mL/min/1.73 m2 were half as likely to receive coronary angiography (OR 0.50, 95% CI 0.40 to 0.64) after non-ST-elevation (NSTE)-ACS and one-third as likely after STEMI (OR 0.30, 95% CI 0.19 to 0.46), resulting in 15 and 17 per 100 fewer procedures, respectively. Following multivariable adjustment, the ORs for receipt of angiography following NSTE-ACS were 1.05 (95% CI 0.88 to 1.27), 0.98 (95% CI 0.77 to 1.26), 0.76 (95% CI 0.57 to 1.01) and 0.58 (95% CI 0.44 to 0.77) in eGFR categories 60–89, 45–59, 30–44 and <30, respectively. After STEMI, the respective ORs were 1.20 (95% CI 0.84 to 1.71), 0.77 (95% CI 0.47 to 1.24), 0.33 (95% CI 0.20 to 0.56) and 0.28 (95% CI 0.16 to 0.48) (p<0.001 for linear trends). ORs were unchanged following adjustment for frailty. A positive association between the worse eGFR category and 30-day mortality was found (test for trend p<0.001), which was unaffected by adjustment for frailty.Conclusions In people with ACS, lower eGFR was associated with reduced receipt of invasive coronary management and increased mortality. Adjustment for frailty failed to change these observations. Further research is required to explain these disparities and determine whether treatment variation reflects optimal care for people with low eGFR.Trial registration number NCT03507309.