Background Atrial fibrillation (AF) is the most common heart rhythm disorder worldwide. Changes in metabolism and damage to mitochondria affect left atrial structure and electrical function. Few studies have applied multi-omics to plasma from AF patients. Methods We enrolled 37 patients with persistent AF and 36 with supraventricular tachycardia (SVT). Blood samples were drawn from the coronary sinus in all patients. Ten subjects from each group were randomly selected for 4D-DIA proteomics and untargeted metabolomics. ELISA validation was done on 35 patients per group. Four GEO atrial tissue datasets and one single-nucleus RNA-seq dataset were used for external support. Differentially expressed proteins (DEPs) were defined by |fold change| ≥ 1.5 and Benjamini-Hochberg adjusted P < 0.05. Protein-protein interaction networks were built with STRING and Cytoscape. Results We quantified 3268 plasma proteins and identified 217 DEPs in AF. Network analysis showed voltage-dependent anion channel 1 (VDAC1) as a central mitochondrial hub. It connected respiratory chain subunits and autophagy regulators. ELISA confirmed higher plasma VDAC1, citrate synthase(CS) and secreted frizzled-related protein 2(SFRP2) in AF patients (all P < 0.001). Metabolomics identified 505 differentially abundant compounds. Glycerophospholipid and fatty acid pathways were disrupted. Integrated analysis revealed eight shared dysregulated pathways linked to energy metabolism. Four GEO datasets showed higher atrial VDAC1 transcripts in AF (P < 0.05). Single-nucleus RNA-seq localized VDAC1 to cardiomyocytes. Conclusions Plasma protein and metabolite profiles differ between persistent AF and SVT. Mitochondrial and lipid pathways are both affected. VDAC1 is elevated in plasma, atrial tissue. These findings are supported across multiple cohorts.
Background Circulating cellular communication network factor 1 (CCN1) improves risk stratification in patients with acute coronary syndrome. We here investigated the association of CCN1 with all-cause mortality in patients with dilated cardiomyopathy (DCM). Methods Patients with a primary diagnosis of DCM, defined as LVEF <45% and increased LVEDD (LVEDD >117%), were included in a derivation (SFB/TR19 Greifswald) and a validation (IKARUS Marburg) cohort. Exclusion criteria comprised primary valvular diseases, acute myocarditis, active infectious diseases, pulmonary diseases, cancer, chronic alcoholism, and heart failure of other origins. CCN1 levels were determined in serum from study inclusion using an enzyme-linked immunosorbent assay. An adjusted multivariable Cox regression model was used to assess the association (hazard ratios) between tertiles of CCN1 concentration and all-cause mortality. Results In the SFB/TR19 cohort and [IKARUS cohort], respectively a total of 283 [236] predominantly male (78% [75%]) DCM patients with a median age of 56 [51] years with a severely reduced LVEF (31% [30%]), increased LVEDD (67.0 [67.0]), and normal eGFR (90.9 [83.6] ml/min) were analyzed. During a median follow-up of 10.6 [14.9] years, a total of 107 (37%) [100 (42%)] patients died. Patients in the highest CCN1 tertile had a significantly higher mortality risk than those in the lower tertile (p = 0.007 [p = 0.004]). In both cohorts, CCN1 remained associated (1.92; 95% CI: 1.14–3.24; p = 0.014 [1.69; 1.00–2.85; p = 0.049]) in adjusted multivariable Cox regression models. Conclusion CCN1 is associated with all-cause mortality in DCM patients, warranting further research into the underlying pathophysiology.
Background:The epidemiology of permanent pacemaker [PPM] implantation in the community setting is unclear. Methods:We examined Framingham Heart Study cohort participants aged ≥45 years, without a history of PPM, who attended ≥1 exam cycle during 1972-2019. Fine-Gray subdistribution hazard models were used to calculate subdistribution hazard ratios [sHR] and 95% confidence intervals [CI] for the association between clinical characteristics (age, sex, body mass index [BMI], systolic and diastolic blood pressure [BP], smoking, hypertension treatment, diabetes, history of atrial fibrillation [AF], myocardial infarction, and heart failure) and incident PPM implantation. For continuous measures, sHRs were expressed per 1 standard deviation [SD]. Multiple imputation was used to account for missing data. Results:A total of 11,993 participants contributed 44,711 exam cycles (mean age 63 ± 11 years, 56% women) to the analysis. During follow-up (mean 4.9 ± 2.2 years), there were 382 PPM implantations and 4298 deaths. In multivariable-adjusted models, the following factors were associated with incident PPM implantation: age (sHR [per 1 SD = 11 years] =1.39), female sex (sHR = 0.49), BMI (sHR [per 1 SD = 5.1 kg/m2] =1.17), higher systolic BP (sHR [per 1 SD = 20 mmHg] =1.46), lower diastolic BP (sHR [per 1 SD = 10 mmHg] = 0.68), hypertension treatment (sHR = 1.40), smoking (sHR = 0.59), and history of AF (sHR = 2.60). Conclusions:Higher age, BMI, systolic BP, hypertension treatment, and history of AF were associated with increased rate of PPM implantation, while female sex, smoking, and diastolic BP showed an inverse association. Further study is needed to confirm these findings.
Background Intracoronary pressure-wire pullback allows differentiation between focal and diffuse disease, improving precision of revascularisation strategies. While focal lesions benefit from percutaneous coronary intervention (PCI), diffuse disease yields suboptimal results and is therefore treated with coronary artery bypass grafting (CABG) or conservatively (i.e., optimal medical therapy [OMT]). Current guidelines emphasise the prognostic relevance of proximal left anterior descending artery (LAD) disease, yet its role in guiding treatment in diffuse disease remains uncertain. Objectives To evaluate whether angiographically severe proximal LAD disease influences revascularisation strategy in functionally significant diffuse LAD disease. Methods This retrospective cohort study included patients undergoing coronary angiography and subsequent pullback measurement in the LAD in the presence of ischemia (FFR ≤0.80 or RFR ≤0.89), between 2022 and 2024. Coronary anatomy was reassessed blinded to treatment. Multinomial logistic regression explored predictors of treatment strategy (OMT, PCI, or CABG) in patients with diffuse LAD disease. Results Among 303 patients with significant LAD disease, 214 (71%) had diffuse and 89 (29%) focal disease. Baseline characteristics were similar. Focal disease was more associated with triple-vessel CAD (34% vs. 19%; p = 0.007). Severe proximal (21% vs. 7.9%; p = 0.004) and distal LAD stenosis were more common in diffuse disease. In diffuse disease only, severe proximal LAD was associated with an increase in CABG (28.3% vs. 10.1%; p = 0.004) and a decrease in OMT rates (54.3% vs. 75.6%; p = 0.009). Multinomial regression confirmed this association (OR 0.26; p = 0.003). Conclusion Absence of angiographically severe proximal LAD stenosis is associated with CABG deferral in diffuse, functionally significant LAD disease.
Background:This study aimed to evaluate the association of peak exercise left atrial (LA) reservoir strain with clinical outcomes in patients with heart failure (HF), and the prognostic value across HF phenotypes. Methods:We enrolled 150 patients with HF (age 67 ± 15 years, 61% males) who underwent exercise stress echocardiography. LA reservoir strain at rest and peak workload were evaluated using the speckle-tracking echocardiography. The endpoint was a composite of cardiovascular death or hospitalization for HF. Results:During a median follow-up of 1.1 years (interquartile range 0.7-2.4 years), 49 patients experienced events. Patients were categorized into two groups on the basis of the median value of LA reservoir strain at peak workload of 16.7%. Kaplan-Meier curves showed that patients with low LA reservoir strain at peak workload had a higher event rate than those with high LA reservoir strain (log-rank test, P < 0.001). LA reservoir strain at peak workload was independently associated with events (hazard ratio: 5.57, 95% confidence intervals: 2.74-11.3, P < 0.001), and provided the incremental prognostic value over clinical factors. When analyzed according to HF phenotype, LA reservoir strain at peak workload was significantly associated with events in both patients with HFpEF and those with HFrEF, with a significant interaction between HF phenotype and LA reservoir strain at peak workload. Conclusion:LA reservoir strain at peak workload has prognostic value for adverse outcomes in both HFpEF and HFrEF.
Cardiovascular disease is the leading cause of death worldwide, and among survivors of heart failure, peripheral arterial disease, valvular disease, and atrial fibrillation many develop motor impairment: weakness, slow gait, sarcopenia, frailty. These deficits are usually attributed to deconditioning or coexisting conditions. This narrative review advances a different hypothesis: that reduced arterial blood delivery, from a failing pump or a narrowed artery, itself drives motor impairment at every level of the motor pathway, from cortex to muscle fibre. We group cardiovascular conditions into systemic models of low blood flow (heart failure with reduced or preserved ejection fraction, low-flow aortic stenosis, atrial fibrillation) and regional models of arterial obstruction (carotid, subclavian, and vertebral stenosis; peripheral arterial disease), quantifying the perfusion deficit where haemodynamic data exist. We then examine mechanisms: cerebral hypoperfusion that thins the motor cortex and weakens motor planning; vulnerability of the cerebellum, basal ganglia, and spinal motor neurons, impairing coordination and gait; ischaemic peripheral nerve damage that denervates muscle; loss of muscle capillaries, mitochondrial injury, and failed regeneration; amplified feedback from underperfused muscle that suppresses central motor drive; and gating of corticospinal excitability by the cardiac cycle. Causality is supported by within-patient comparisons, in which a lateralised ischaemic limb or hypoperfused hemisphere weakens while the opposite side, under identical systemic conditions, does not; by graded perfusion-to-outcome relationships; and by recovery after cardiac transplantation, ventricular assist device support, valve replacement, and revascularisation. Deconditioning amplifies these effects rather than causing them, so treatment should restore blood flow and rebuild muscle together.
Background:Stress cardiovascular magnetic resonance (CMR) provides comprehensive assessment of myocardial ischemia and myocardial scar but remains limited by examination duration and accessibility. Whether an ischemia-focused analytical strategy preserves clinically relevant information compared with conventional multiparametric stress CMR remains uncertain. Objectives:To compare the real-world performance of an ischemia-focused analytical strategy (FOCUS-Ischemia) with conventional multiparametric stress CMR for prediction of downstream coronary management and clinical outcomes. Methods:A total of 717 consecutive patients undergoing clinically indicated stress CMR in the Nancy Ischemia Registry were included. Patients were stratified according to inducible ischemia and late gadolinium enhancement (LGE). Endpoints included early coronary revascularization within 30 days, late revascularization >30 days, and a composite outcome including death, unplanned invasive coronary angiography, or coronary revascularization. ROC, Kaplan-Meier, and multivariable Cox analyses were performed. Results:Inducible ischemia was strongly associated with downstream coronary management and adverse outcomes. Early revascularization rates were markedly higher in ischemic versus non-ischemic patients (18.1% and 17.0% vs 0.0% and 0.5%; p < 0.001). The FOCUS-Ischemia and multiparametric stress CMR models demonstrated similar discriminatory performance for early revascularization (AUC 0.880 vs 0.885), late revascularization (0.746 vs 0.748), and the composite clinical endpoint (0.819 vs 0.821). Conclusions:An ischemia-focused analytical strategy derived from a conventional multiparametric stress CMR examination demonstrated similar discriminatory performance for coronary management-related outcomes. Because all patients underwent a conventional multiparametric stress CMR examination, the present study did not evaluate an abbreviated acquisition protocol. These findings should therefore be considered exploratory and hypothesis-generating.
Background:Atrial fibrillation (AF) frequently requires emergency medical services (EMS), but acute-care pathways may differ by remoteness. We examined ambulance-attended AF care and outcomes across Victoria, Australia. Methods:We conducted a population-based linked cohort study of first ambulance-attended AF events from Jan 1, 2015, to June 30, 2019. Residence was classified using ARIA+ as major cities, inner regional, or outer regional/remote. We estimated incidence rate ratios (IRRs) and used adjusted logistic regression to compare care pathways, rhythm-control procedures, EMS re-attendance, and mortality. Results:Among 16,372 events, 11,914 (72.8%) occurred in major cities, 3691 (22.5%) in inner regional areas, and 767 (4.7%) in outer regional/remote areas. Annualised incidence was higher in inner regional (IRR 1.31, 95% CI 1.27-1.36) and outer regional/remote areas (IRR 1.25, 95% CI 1.16-1.35) than in major cities. Transport to hospital exceeded 99% across groups. Transport longer than 60 min increased with remoteness (0.7%, 6.4%, and 11.2%), whereas offload delay was concentrated in major cities (69.6%, 46.2%, and 35.9%). Rhythm-control procedures, EMS re-attendance for AF, and mortality were similar after adjustment, whereas socioeconomic disadvantage was associated with fewer rhythm-control procedures and higher mortality independently of remoteness. Conclusions:Ambulance-attended AF is more frequent outside major cities and imposes greater transport burden, while metropolitan care is more affected by offload delay. Similar outcomes should not be interpreted as similar care. AF pathways should target ED flow in cities and prehospital triage and follow-up outside cities. Research in context:Evidence before this studyAF is a major contributor to emergency care use, hospitalisation, stroke, and mortality. Prior statewide Victorian studies have described the overall ambulance and ED burden of AF, while international literature suggests that rural patients experience barriers to AF diagnosis, specialist care, and longitudinal management. However, few population-based linked data studies have compared the full prehospital-to-hospital pathway for ambulance-attended AF by remoteness of residence.Added value of this studyThis study links statewide EMS, ED, hospital, and mortality data to show that ambulance-attended AF incidence is higher in regional Victoria and that prolonged transport is substantially more common outside major cities. Conversely, offload delay is concentrated in major cities. Despite these pathway differences, adjusted EMS re-attendance, mortality, and rhythm-control procedure rates were similar across remoteness categories, whereas area-level socioeconomic disadvantage was associated with fewer rhythm-control procedures and higher mortality regardless of where patients lived.Implications of all the available evidenceStatewide AF pathways should not assume that the same bottleneck applies everywhere. Metropolitan strategies should address ED congestion and offload delay, while rural implementation should prioritise transport burden, prehospital risk stratification, and reliable follow-up after ED discharge or non-transport. Socioeconomic disadvantage cuts across both settings and warrants attention in its own right.
Background:The efficacy of intravenous amino acids (AA) in preventing cardiac surgery-associated acute kidney injury (CSA-AKI) remains uncertain. Methods:Three libraries were searched for randomized controlled trials (RCTs) comparing intravenous AA versus placebo or standard care to reduce the risk of CSA-AKI after cardiac surgery with cardiopulmonary bypass (CPB). The primary endpoint was the incidence of CSA-AKI, overall and stratified by KDIGO stage. Secondary endpoints were need for renal replacement therapy (RRT), 30-day mortality, and 90-day mortality. Risk ratios (RRs) with 95% confidence intervals (CIs) were pooled using a random-effects model. Results:Five studies were included, comprising 4499 patients, of whom 2247 (50%) were randomized to intravenous AA infusion. Compared with placebo or standard care, intravenous AA administration significantly reduced the risk of CSA-AKI (RR 0.81, 95% CI 0.70-0.94; p = 0.005), predominantly driven by one trial. The relative risk reduction was more important for stage 2 or 3 AKI (RR 0.69, 95% CI 0.48-0.99; p = 0.046). No significant differences were observed in the need for RRT (p = 0.223), 30-day mortality (p = 0.893), or 90-day mortality (p = 0.581). Conclusion:Intravenous AA administration is associated with a reduced risk of CSA-AKI across the spectrum of AKI severity, without showing an association with other major clinical outcomes.
Cardiovascular diseases account for the highest morbidity worldwide. LR11 (also called SorLA), an LDL receptor family member characterized as a sorting receptor, was initially identified in the brain and has a causative role in the development of Alzheimer's disease. However, LR11, and its circulating shed isoform, sLR11, are also associated to cardiovascular diseases and risk factors for atherosclerosis, such as obesity and diabetes. In the current narrative review, we discuss that elevation of plasma sLR11 levels can result from different forms of vascular injury, but also plays a role in subsequent vascular remodeling. We provide an overview of the mechanisms whereby LR11 promotes vascular remodeling and thereby atherosclerosis and how it could be involved in obesity and diabetes. Furthermore, we discuss the possibilities of (s)LR11 as a biomarker and therapeutic target for cardiovascular diseases.
Background:Opioid use is frequently reported as a baseline characteristic in transcatheter aortic valve implantation (TAVI) cohorts, yet its prognostic implications have not been examined. Aim:To examine the association between pre-procedural opioid use and post-TAVI outcomes. Methods:Using Danish nationwide registries, we identified patients undergoing TAVI between 2015 and 2022. Patients were categorized according to opioid use within one year before TAVI. We examined one-year all-cause mortality, cumulative days of hospitalization including the composite outcome of death or > 14 cumulative hospitalization days, and a potential opioid dose-response relationship. Adjusted relative risks were estimated using multivariable Cox and logistic regression. Results:We identified 6505 patients: 1265 (19.4%) with opioid use (48.9% male; median age 80 years) and 5240 (80.6%) without (59.4% male; median age 81 years). The two groups were comparable in cardiac comorbidities and malignancy but differed in musculoskeletal (back disorders: 45.9% vs 20.7%; arthropathies 65.3% vs 47.1%) and psychiatric conditions (antidepressants: 18.8% vs. 9.6%; anxiolytics/hypnotics: 23.7% vs. 12.4%). Baseline opioid use was associated with higher post-TAVI mortality (10.4% vs 6.9%; adjusted HR 1.35, 95% CI 1.11-1.66) and increased risk of the composite outcome (OR 1.47, 95% CI 1.25-1.72). A dose-response relationship was observed, with one-year mortality increasing from 7.0% in non-users to 8.8% in low-dose users and 12.1% in high-dose users. Conclusion:Baseline opioid use was associated with increased mortality and hospitalization burden in the year following TAVI, and this relationship was aggravated among high-dose users. The observed associations were most likely driven by the underlying condition of the opioid therapy.
Background:Sodium-glucose cotransporter 2 inhibitors (SGLT2i) are recommended across the heart failure (HF) spectrum, irrespective of diabetes status. However, prescribing patterns in very old adults with multimorbidity, renal dysfunction, and geriatric vulnerability remain insufficiently characterised. Methods:We conducted a single-centre retrospective cross-sectional study in a specialised acute cardiogeriatric unit between January 1, 2025, and February 13, 2026. Consecutive hospitalisations with confirmed HF and documented SGLT2i discharge status were included. The primary outcome was absence of SGLT2i at discharge. Multivariable logistic regression identified factors independently associated with non-prescription. Results:Among 1235 hospitalisations, median age was 88.6 years [IQR 82-92], 54.5% were women, and 32.7% had diabetes. SGLT2i were prescribed in 1016 hospitalisations (82.3%) and absent in 219 (17.7%). Prescription did not differ by diabetes status (80.4% vs 83.2%, p = 0.276) or across HbA1c tertiles. In the prespecified primary model, non-prescription was associated with CKD stage 4 (aOR 2.79 [95% CI 1.97-3.96]), CKD stage 5 (aOR 20.29 [9.67-42.56]), palliative orientation (aOR 1.62 [1.09-2.41]), and HFrEF versus HFpEF (aOR 1.55 [1.08-2.22]). Higher albumin and medication count were associated with lower odds of non-prescription. Diabetes, age, and Charlson Comorbidity Index were not independently associated with the primary outcome. Conclusion:SGLT2i prescription was frequent and similar according to diabetes status and glycaemic control. Advanced CKD and palliative-oriented care were the most consistent correlates of non-prescription. These findings describe prescribing patterns but do not establish the appropriateness of individual decisions.
Heart failure with preserved ejection fraction (HFpEF) is more prevalent in women, whereas heart failure with reduced ejection fraction (HFrEF) predominates in men. Despite these well-established epidemiological differences, sex-specific alterations in vascular function in heart failure (HF) remain poorly characterised. This systematic review, using a narrative synthesis approach, evaluated sex-related differences in vascular function in individuals with HF. The review was prospectively registered with PROSPERO (CRD42024617745). MEDLINE and CINAHL were searched from inception to 26th November 2024 for studies reporting sex-stratified measures of arterial stiffness among individuals with HF. Nine studies met the eligibility criteria for inclusion (n = 2820; men: n = 1390, women: n = 1430). Of the included studies, 78% were characterised as HFpEF. Compared with men, women exhibited a higher pulsatile arterial load and lower arterial compliance. Representative findings from individual studies showed that women exhibited a higher augmentation index (28.9 ± 13.7% vs 21.7 ± 11.9%, p < 0.001) and augmentation pressure (19.1 ± 12.4 vs 13.7 ± 10.1 mmHg, p = 0.003). Body mass index (BMI) showed variable relationships with arterial stiffness indices, including positive associations with pulse wave velocity (r = 0.24, p < 0.01), central pulse pressure (r = 0.33, p < 0.001), and augmentation index (r = 0.23, p = 0.01), but an inverse relationship was found with cardio-ankle vascular index (r = -0.204, p < 0.001). Importantly, sex differences in HFpEF remained significant after adjustment for BMI. Women with HF exhibit higher pulsatile arterial load and reduced arterial compliance compared with men, which remain after adjustment for BMI. Sex-specific vascular dysfunction contributes to HF pathophysiology and supports the need for sex-informed assessment.
Background:Heart failure (HF) is a common complication in diabetes mellitus (DM), affecting approximately 40% of HF patients. While inflammation plays a crucial role in both conditions, the prognostic value of inflammatory markers in patients with both DM and HF remains unestablished. Objective:To examine the associations of neutrophil-to-lymphocyte ratio (NLR), monocyte-to-lymphocyte ratio (MLR), and (neutrophil + monocyte)/lymphocyte ratio (NMLR) with 30-day mortality in ICU patients with both DM and HF. Methods:This retrospective cohort study analyzed 1318 patients with both HF and DM from the MIMIC-IV database. NLR, MLR, and NMLR were calculated from complete blood counts within 24 h of ICU admission. Multivariate Cox regression, restricted cubic spline analysis, and ROC curves assessed predictive performance. Results:Among 1318 patients, 239 died within 30 days. The deceased group had significantly higher median values of NLR (10.9 vs. 6.5), MLR (0.9 vs. 0.4), and NMLR (12.1 vs. 7.0) compared to survivors (all P < 0.001). After adjustment, each one-unit increase in MLR was associated with 11% increased mortality risk (HR = 1.11, 95% CI: 1.04-1.19, P = 0.002), while NMLR and NLR showed 1% increases, respectively (NMLR: HR = 1.01, 95% CI: 1-1.01; NLR: HR = 1.01, 95% CI: 1-1.01; both P < 0.001). MLR demonstrated the highest discriminative capacity (AUC = 0.672), followed by NMLR (AUC = 0.658) and NLR (AUC = 0.650). Conclusion:NLR, MLR, and NMLR are independent predictors of 30-day mortality in ICU patients with DM and HF, with MLR showing superior discriminative capacity.
Background:Despite advances in percutaneous coronary intervention (PCI), patients with obstructive coronary artery disease (CAD) continue to experience recurrent ischemic symptoms and adverse cardiovascular events. Skin post-occlusive reactive hyperemia (PORH) is a noninvasive measure of systemic microvascular function that may add prognostic information beyond epicardial coronary anatomy, but its relation to long-term cardiovascular outcomes after PCI has not been well defined. Methods:We performed a retrospective analysis of a prospectively established CAD registry that included 197 patients with obstructive CAD who underwent PCI and baseline assessment of the skin PORH response. Long-term composite cardiovascular outcomes were evaluated over follow-up of up to 5 years using Kaplan-Meier methods and Cox proportional hazards regression, with propensity score matching used as a sensitivity analysis. Results:The median age of the cohort was 66 years, and 84.8% of patients were male. Compared with those in the higher PORH group, patients with a lower PORH index had more cardiometabolic comorbidities and experienced significantly higher rates of long-term composite cardiovascular events. In multivariable Cox analysis, a higher PORH index was independently associated with a lower risk of composite cardiovascular events (adjusted hazard ratio, 0.422; 95% confidence interval, 0.198-0.898), and this inverse association was preserved in the propensity score-matched cohort. Conclusions:In patients with obstructive CAD undergoing PCI, impaired skin PORH was independently associated with an increased risk of long-term composite cardiovascular events. These findings support a potential role for skin microvascular assessment in cardiovascular risk stratification, and warrant confirmation in larger prospective studies.
Background: Coronary heart disease (CHD) remains a leading cause of mortality worldwide, with individuals with diabetes mellitus (DM) facing markedly elevated risk due to complex inflammatory and metabolic disturbances. Emerging composite inflammatory and metabolic indices have demonstrated promise in enhancing cardiovascular risk stratification, yet research quantifying and comparing their respective predictive performance and assessing their relative contributions remains limited. This study aimed to develop a clinically applicable model for early CHD risk prediction in diabetic patients using novel composite inflammatory and metabolic indices. Methods: This study utilized data from 3379 diabetic participants in the NHANES 1999–2018 survey cycles. Additionally, an independent external set of 902 patients from Qilu Hospital of Shandong University served as the validation cohort. Novel inflammatory and metabolic indices were calculated. Feature selection was performed via LASSO regression, the univariate logistic regression, and the Boruta algorithm. Nine machine learning (ML) models were developed using selected predictors. Model performance was evaluated using the receiver operating characteristic (ROC) curve, calibration curves, and decision curve analysis. SHapley Additive exPlanations (SHAP) values were used to interpret model predictions and identify key contributing features. Results: Five composite indices UHR (uric acid-to-HDL ratio), MHR (monocyte-to-HDL ratio), NPAR (neutrophil-to-albumin ratio), NLR (neutrophil-to-lymphocyte ratio), and AIP (atherogenic index of plasma) were identified as robust predictors of CHD. The random forest (RF) algorithm achieved the highest performance, with an AUC of 0.852 in the internal validation set and 0.713 in the external cohort. Calibration plots, Brier scores, and decision curve analysis further confirmed the RF model’s predictive reliability and clinical utility. SHapley Additive exPlanations (SHAP) value analysis revealed that UHR, MHR, NLR, age, and hypertension were the key features driving CHD prediction. Conclusion: We developed and externally validated an ML model incorporating five composite inflammatory and metabolic indices (UHR, MHR, NLR, NPAR, AIP), which demonstrated promising performance in predicting CHD risk in diabetic patients.
