Abstract
BACKGROUND: Lipid metabolism-related inflammatory indices (LMIIs) have been recognized as potential biomarkers for the risks of atherosclerosis and major adverse cardiovascular events. This study aims to explore the associations of LMIIs, including neutrophil to high-density lipoprotein cholesterol ratio (NHR), monocyte to high-density lipoprotein cholesterol ratio (MHR), platelet to high-density lipoprotein cholesterol ratio (PHR), and lymphocyte to high-density lipoprotein cholesterol ratio (LHR), with the risk of incident peripheral artery disease (PAD) in elderly T2DM patients. METHODS: A total of 2837 participants aged ≥ 60 years with T2DM from the Jinshan Cohort (China) and 13,542 participants from the UK Biobank (UK) were included in the primary analyses. According to the type of outcome variables, logistic regression models and Cox proportional-hazards models were used to estimate the risks of incident PAD associated with LMIIs. The methods of generalized propensity score (GPS), E-value and negative control exposure (NCE) were applied to control the potential confounding. Additionally, stratified analyses were performed across various populations to examine potential heterogeneity in the effects of LMIIs on PAD risk. Mediation effects of liver and kidney function-related indicators on associations between LMIIs and incident PAD were also explored. RESULTS: The results from traditional regression models suggested positive associations of incident PAD with all four LMIIs. In the Jinshan Cohort, the odds ratios (ORs) and 95% confidence intervals (CIs) of PAD for one-unit increase in NHR (10(9) mmol), MHR (10(8) mmol), PHR (10(11) mmol) and LHR (10(9) mmol) were 1.24 (1.11-1.39), 1.22 (1.08-1.37), 1.69 (1.32-2.15) and 1.57 (1.22-2.01), respectively. While in the UK Biobank, the corresponding hazard ratios (HRs) and 95%CI were 1.14 (1.10-1.18), 1.03 (1.01-1.04), 1.19 (1.09-1.29) and 1.11 (1.07-1.14), respectively, These associations remained robust in regression models weighted by GPS methods. The E-values for the effects of each LMII on PAD were consistently found to be significantly larger than their corresponding observed effects in both cohorts. NCE analyses revealed no statistically significant associations between any selected NCE and PAD in either cohort. On calibration using NCEs, the calibrated P values confirmed significant effect sizes for associations between LMIIs and incident PAD (all for P < 0.05). Additionally, subgroup analyses in the Jinshan Cohort showed different associations varied across sex, residential area, smoking status and HbA1c level, with pronounced HRs in females, urban residents, smokers and individuals with HbA1c ≥ 7%. Findings from the UK Biobank further suggested that aspirin use and HbA1c level may modify the effects of LMIIs on PAD risk (P-interaction < 0.05). Mediation analyses indicated that estimated glomerular filtration rate (eGFR) mediated the relationships between LMIIs and PAD. The proportion of the mediating effects ranged from 10.03% to 19.95% in the Jinshan Cohort and from 10.20 to 20.80% in the UK Biobank. CONCLUSIONS: Elevated levels of LMIIs, such as NHR, MHR, PHR and LHR, were associated with an increased risk of incident PAD among elderly diabetic patients. These associations exhibited notable population heterogeneity, with stronger effects observed in females and individuals with higher HbA1c level. eGFR may serve as a mediating factor in the associations between LMIIs and the occurrence of PAD. These findings provide novel evidence for immune inflammation-based risk assessment and personalized preventive strategies for PAD.