Association of Remnant Cholesterol Inflammation Index with Cardiovascular Risks and All-Cause Mortality in Individuals with Diabetes or Prediabetes
Article information
Abstract
Background
Remnant cholesterol (RC) and low-grade inflammation are established contributors to cardiovascular disease (CVD) risks in diabetes. However, their combined prognostic impact remains unclear in dysglycemia. We evaluated the remnant cholesterol inflammation index (RCII), integrating RC and high-sensitivity C-reactive protein (hsCRP), for predicting mortality and CVD risks in diabetes/prediabetes.
Methods
This study included 2206 United States adults with diabetes/prediabetes from National Health and Nutrition Examination Survey 2015–2018. RCII was calculated as [RC (mg/dL)×hsCRP (mg/L)]/10. All-cause mortality was tracked via National Death Index until 2019; CVD risk was assessed cross-sectionally. Cox proportional hazard regression determined the hazard ratio (HR) and 95% confidence intervals (CIs) of RCII for all-cause mortality. Logistic regression models estimated the odds ratio (OR) and 95% CIs of RCII for CVD risks.
Results
For CVD risks, Q4 vs. Q1 demonstrated increased odds (OR, 2.32; 95% CI, 1.23 to 4.37), though per-standard deviation (SD) increments were non-significant (OR, 1.15; 95% CI, 0.98 to 1.35; P=0.083). During a median of 38 months follow-up, higher RCII quartiles showed graded associations with all-cause mortality (Q4 vs. Q1: HR, 2.45; 95% CI, 1.08 to 5.58; per 1-SD increase: HR, 1.21; 95% CI, 1.08 to 1.35). Restricted cubic splines confirmed dose-dependent relationships for CVD risks and all-cause mortality (all P=0.005 for overall). Subgroup analyses revealed consistent mortality associations but sex-specific CVD interactions (P=0.047 for interaction).
Conclusion
Our study found the RCII as a biomarker for predicting all-cause mortality and CVD risks in individuals with prediabetes or diabetes, highlighting the synergistic effects of RC and low-grade inflammation on adverse outcomes in this population and may facilitate early identification of individuals at heightened risk for CVD.
Highlights
• RC and hsCRP are established, modifiable risk factors for CVD in dysglycemia.
• RCII is a novel indicator combining RC and hsCRP.
• RCII serves as a biomarker for predicting adverse outcomes in prediabetes or DM.
INTRODUCTION
Diabetes mellitus (DM) represents a complex metabolic disorder characterized by multisystem involvement due to chronic dysregulation of glucose homeostasis [1]. The diabetic population demonstrates a 2- to 4-fold elevation in age-adjusted relative risk for cardiovascular disease (CVD) and mortality compared to non-diabetic individuals [2]. Of particular clinical significance, prediabetes defined as the pathophysiological continuum between normal glucose regulation and overt diabetes, serves as a critical window for cardiovascular risk stratification, associating with intermediate-range cardiovascular risks [3]. Therefore, early detection of high-risk patients and identification of more potential risk factors are critical to improving outcomes for patients with diabetes or prediabetes.
In individuals with DM, remnant cholesterol (RC) levels are frequently substantially elevated as a result of the metabolic alterations induced by hyperglycemia and insulin resistance, which are likely to contribute to a heightened risk of CVD and mortality in DM [4]. At the same time, it has been shown that the presence of low-grade inflammation, as indicated by elevated levels of high-sensitivity C-reactive protein (hsCRP), is associated with a worse prognosis in patients with impaired glucose metabolism [5,6]. Therefore, evaluating RC combined with hsCRP seemed to offer a more complete way to assess the risk of CVD and death in individuals with diabetes or prediabetes. Previous studies have established the remnant cholesterol inflammation index (RCII) based on RC and hsCRP to comprehensively assess the combined effects of systemic residual cholesterol and low-grade inflammation on stroke risks, and found that it is significantly related to an increased risk of stroke [7]. Nevertheless, the predictive value of RCII of CVD and mortality risks in patients with diabetes or prediabetes was uncertain.
Therefore, in this study, we estimated the ability of baseline and long-term cumulative RCII on the risks of CVD and all-cause mortality in participants with diabetes and prediabetes, using National Health and Nutrition Examination Survey (NHANES) data 2015 to 2018.
METHODS
Study design
The NHANES, administered by the U.S. Centers for Disease Control and Prevention’s National Center for Health Statistics (NCHS), is a nationally representative surveillance system designed to assess population health dynamics in community-dwelling Americans through biennial data collection cycles. NHANES study employs a stratified multistage probability sampling framework to systematically capture clinical diagnoses, modifiable health risks, and nutritional biomarkers using in-person structured questionnaires and protocol-driven physical evaluations at mobile examination centers. The study protocols of NHANES were conducted in accordance with the Declarations of Helsinki and approved by the Research Ethics Review Board of the NCHS and written informed consent was obtained from all participants involved in the study. The current study utilized data from two survey cycles of the continuous NHANES conducted in 2015–2016 and 2017–2018, as well as mortality data from the National Death Index (NDI).
As present in Supplementary Fig. 1, a total of 19,225 participants were included in the NHANES cohort between 2015 and 2018. DM is diagnosed by self-reported DM, a fasting glucose level greater than or equal to 126 mg/dL, the 2 hours plasma glucose of the oral glucose tolerance test higher than or equal to 199.8 mg/dL, glycosylated hemoglobin (HbA1c) level of ≥6.5%, or use of oral hypoglycemic agents or insulin. Prediabetes is defined as having a fasting plasma glucose value ranging from 100 to 125 mg/dL, the 2 hours of plasma glucose ranging from 140.4 to 198 mg/dL, or HbA1c levels between 5.7% and 6.4% in the absence of the established diagnosis of DM or having hypoglycemic therapies. After excluding participants younger than 20 years old (n=5,136), pregnant women (n=126) and those with missing information on admission fast glucose and HbA1c (n=9,130), hsCRP and lipid-related indicators (including total cholesterol [TC], triglycerides [TG], low-density lipoprotein cholesterol [LDL-C], and high-density lipoprotein cholesterol [HDL-C], n=300), mortality data (n=19), without diabetes or prediabetes (n=1,411), key covariates (education levels, family poverty income ratio [PIR], body mass index [BMI], smoking, and drinking, n=859), medicine uses (lipid-lowering, antidiabetic, and antihypertensive drugs) and weights (n=37), we included 2,206 eligible patients with diabetes or prediabetes in the final analysis.
Exposure and outcome
Blood samples were collected from each participant in the non-fasting state. The concentrations of serum TG and TC were measured using specific enzymatic assays. Serum HDL-C concentrations were determined by immunoassays. The values of serum LDL-C were calculated using the Friedewald equation (LDL-C=TC–HDL-C–TG/5). The NHANES project hsCRP was based on the determination of hsCRP concentration by the highly sensitive near infrared particle immunoassay rate methodology and the quantification of hsCRP using latex enhanced turbidity method (https://wwwn.cdc.gov/nchs/nhanes). RC (mg/dL) was calculated using the formula: RC=TC/(HDL-C+LDL-C) [8], RCII was then computed by multiplying RC by hsCRP levels, normalized as follows: RCII=RC (mg/dL)×hsCRP (mg/L)/10 [7]. All patients were divided into four groups according to RCII quantiles (Q1 to Q4).
CVD diagnosis was determined through structured interviews using a standardized questionnaire that captured self-reported physician diagnoses. Participants were asked if they had ever been diagnosed with coronary heart disease, angina pectoris, myocardial infarction, or stroke by healthcare professionals. A positive response to any of these questions was used to confirm a CVD diagnosis [9].
The mortality status of NHANES participants was ascertained through linkage and probabilistic record matching to the NDI data, with follow-up extending through December 31, 2019. For each individual, the follow-up duration was calculated as the interval between the date of the baseline examination and the last known survival date or the date of removal from the mortality profile. All-cause mortality was defined as death from any cause.
Covariates
Sociodemographic characteristics, family income level ratio, medical history, and laboratory indicators were collected via standardized questionnaires. Education level was categorized as less than high school, high school/equivalent, or college/above. Family PIR was divided into ≤1.3, 1.3–3.5, and >3.5. Smoking status was defined as smoking according to their answers about whether they were current smokers or had smoked at least 100 cigarettes. Alcohol consumption was defined as drinking for participants who reported ≥2 drinks/day for men or ≥1 drink/day for women. Hypertension was diagnosed based on self-report, current antihypertensive treatment, or blood pressure ≥140/90 mm Hg. Baseline concentrations of fasting glucose, and HbA1c were measured using NHANES laboratory protocols.
Statistical analysis
To ensure the representativeness accuracy of the NHANES data, sample weights, clustering, and stratification were incorporated into all analyses. Sample baseline characteristics were summarized as means (standard deviation [SD]), medians (interquartile range [IQR]), or counts (proportions) according to the distribution of each variable. The distribution pattern of continuous variables was assessed using the Kolmogorov-Smirnov test. Differences in covariates across groups were evaluated using one-way analysis of variance (ANOVA) for normally distributed variables, Kruskal-Wallis tests for non-normally distributed variables, and chi-square (χ²) tests for categorical variables, as appropriate. Survival analysis was performed using the Kaplan-Meier method to estimate death-free survival rates, with comparisons between groups made using the log-rank test. Cox proportional hazards regression was used to calculate hazard ratios (HRs) with 95% confidence intervals (CIs) for all-cause mortality. Logistic regression was applied to estimate odds ratios (ORs) with 95% CIs for CVD. Three models were fitted: model 1 was unadjusted; model 2 adjusted for age, sex, and BMI; and model 3 further adjusted for PIR, hypertension, smoking status, CVD (except in logistic models), lipid-lowering medications, antidiabetic medications, glucose, HbA1c, and TG. Stratified analyses were conducted within subgroups defined by age (<60, ≥60 years), sex (male, female), ethnicity (non-white, white), BMI (<25, ≥25 kg/m²), and DM status (prediabetes, diabetes). Interaction terms between RCII and cumulative RCII levels and stratification variables were tested to assess effect modification. A restricted cubic spline (RCS) model was employed to visualize the dose-response relationships of RCII with CVD risks and all-cause mortality. The added discriminative capacity of the biomarkers (RC, hsCRP, and RCII) beyond conventional risk factors was quantitatively assessed through Harrell’s C-statistic and likelihood ratio (LR) test. All analyses were performed using R version 4.4.3 (R Foundation for Statistical Computing, Vienna, Austria). A two-sided P value <0.05 was considered statistically significant.
Ethics approval and consent to participate
The NCHS and Ethics Review Board approved the protocol for NHANES, and all participants provided written informed consent. The project itself does not provide a single universal ethical approval number.
RESULTS
Baseline characteristics
The study cohort comprised 2,206 individuals with confirmed diabetes or prediabetes. Table 1 detailed the baseline characteristics stratified according to RCII quartiles. Demographically, the population demonstrated a mean age of 51.68±16.01 years with male predominance (54.5%). The median RCII value was 4.55 (IQR, 1.67 to 11.76). Participants with higher RCII levels were more likely to be older, female, have a lower PIR, be smoking, have a higher BMI, and have higher prevalence of hypertension, CVD, and DM compared with those in the lowest quartile (all P<0.05). Additionally, they were more likely to use antidiabetic drugs but less likely to use lipid-lowering drugs (all P<0.05). Meanwhile, significant trends in laboratory parameters across RCII quartiles (Q1 to Q4) were observed. Specifically, the levels of TC, TG, LDL-C, fasting glucose, HbA1c, RC and hsCRP, and RCII itself exhibited a progressive increase from Q1 to Q4 (all P<0.001). Conversely, HDL-C levels showed a descending trend across the quartiles (P<0.001).
Baseline demographic and clinical data of RCII quartiles
Association of RCII and CVD risks
A total of 297 CVD were occurred and the prevalence of CVD was higher in the upper quartiles of RCII compared with the lowest quartile (Q1: 7.4% vs. Q2: 10.1% vs. Q3: 11.5% vs. Q4: 14.3%, P<0.001) (Table 1). Univariate and multivariate logistic regression analyses were conducted to assess the association between RCII and CVD risks, with results presented in Table 2. In the unadjusted model, the OR for CVD risks in the highest quartile of RCII was 2.09 (95% CI, 1.18 to 3.68; P=0.013) compared with the lowest quartile. In model 2, the OR for CVD risks in Q4 relative to Q1 was 2.26 (95% CI, 1.25 to 4.07; P=0.009). In the fully adjusted model (model 3), the OR for CVD risks in Q4 versus Q1 was 2.32 (95% CI, 1.23 to 4.37; P=0.013). Furthermore, as a continuous variable, per SD increase in RCII was associated with a 1.21-fold increase in CVD risks in the unadjusted model (OR, 1.21; 95% CI, 1.06 to 1.37; P=0.005) and a 1.27-fold increase in model 2 (OR, 1.27; 95% CI, 1.05 to 1.54; P=0.018). However, this association was not significant in the fully adjusted model 3 (OR, 1.19; 95% CI, 0.96 to 1.47; P=0.111). Supplementary Table 1 presents the correlation between the adjusted covariates included in model 3 and the risk of CVD. RCS analysis identified a significant linear dose-response relationship between RCII and CVD risk (P for overall trend=0.011), with no evidence of a statistically significant non-linear relationship (P for non-linearity=0.244), which indicated that the risk of CVD increases in a monotonic fashion with higher RCII concentrations (Fig. 1).
Logistic regression models for the association between the RCII and CVD risks
Association of the remnant cholesterol inflammatory index (RCII) with cardiovascular disease (CVD) risks (A) and all-cause mortality (B) in patients with diabetes or prediabetes. Logistic regression model was adjusted for age, sex, body mass index (BMI), poverty income ratio, hypertension, smoking, lipid-lowering drugs, antidiabetic drugs, glucose, glycosylated hemoglobin (HbA1c), and triglyceride (TG). Cox regression model was adjusted for age, sex, BMI, poverty income ratio, hypertension, smoking, CVD, lipid-lowering drugs, antidiabetic drugs, glucose, HbA1c, and TG. OR, odds ratio; CI, confidence interval; HR, hazard ratio. P<0.05 suggests significant differences
Association of RCII and all-cause mortality
Over a median follow-up duration of 38 months, a total of 87 all-cause deaths were recorded, corresponding to an all-cause mortality rate of 972 per 1,000 person-years. Kaplan-Meier survival analysis was performed to estimate the cumulative incidence of all-cause mortality across RCII quartiles, which was 1.6%, 3.0%, 4.3%, and 2.7% for Q1 to Q4, respectively. Although the Log-rank test did not reveal a significant difference in all-cause mortality across quartiles (P=0.094), patients with higher RCII had a relatively higher incidence of all-cause death compared with those in the Q1 quartile, the highest all-cause mortality being in Q3 (Fig. 2). Univariate and multivariate Cox regression models used to estimate the association of RCII and all-cause mortality, as shown in Table 3. In the unadjusted model, the HRs and 95% CIs for all-cause mortality across quartiles (Q1 to Q4) were as follows: Q1 (reference; HR, 1.00), Q2 (HR, 1.84; 95% CI, 0.90 to 3.77), Q3 (HR, 2.70; 95% CI, 1.20 to 6.07), and Q4 (HR, 1.56; 95% CI, 0.75 to 3.25). When adjusted for age, sex, and BMI (model 2), the HRs for RCII in Q2, Q3, and Q4 were 2.31 (95% CI, 1.05 to 5.11), 3.42 (95% CI, 1.47 to 7.97), and 3.07 (95% CI, 1.42 to 6.62), respectively, compared with Q1 (all P<0.05). In model 3, which adjusted for age, sex, BMI, PIR, hypertension, smoking, CVD, lipid-lowering drugs, antidiabetic drugs, glucose, HbA1c, and TG, the HRs for RCII across quartiles (Q1 to Q4) were 1.00 (reference), 2.05 (95% CI, 0.96 to 4.39), 2.82 (95% CI, 1.23 to 6.45), and 2.45 (95% CI, 1.08 to 5.58), respectively. In addition, when modeled as a continuous variable, each SD increase in RCII was associated with a 19% increase in all-cause mortality risk in model 1 (HR, 1.19; 95% CI, 1.06 to 1.32; P=0.002), a 29% increase in model 2 (HR, 1.29; 95% CI, 1.18 to 1.41; P<0.001), and a 21% increase in model 3 (HR, 1.21; 95% CI, 1.08 to 1.35; P=0.001). The effects of the adjusted covariates included in model 3 on the risk of all-cause mortality were showed in Supplementary Table 2. Subsequently, RCS analysis revealed a significant linear dose-response relationship between RCII and all-cause mortality (P for overall trend=0.005), with no statistically significant deviation from linearity (P for non-linearity=0.469), suggesting the risk gradient of all-cause mortality increased monotonically across RCII concentrations (Fig. 1).
Kaplan-Meier analyses for all-cause mortality among the remnant cholesterol inflammatory index (RCII) quartiles. Q1–Q4 quartiles 1–4. P<0.05 suggests significant differences.
Cox regression models for the association between the RCII and all-cause mortality
Incremental discriminative capacity of RCII
As detailed in Supplementary Table 3, we evaluated the incremental predictive value of RC, hsCRP, and RCII when incorporated into the baseline model for CVD and all-cause mortality. For CVD risk prediction, the addition of RCII yielded the highest discriminative performance (C statistics=0.815, P<0.001), surpassing both RC (C statistics=0.812, P<0.001) and hsCRP (C statistics=0.812, P<0.001). Similarly, in all-cause mortality risk assessment, RCII demonstrated superior predictive accuracy (C statistics=0.839, P<0.001) compared to RC (C statistics=0.833, P<0.001) and hsCRP (C statistics=0.834, P<0.001). The LR tests further supported RCII’s enhanced model fit. For CVD risk, RCII showed significant improvement (LR=14.301, P=0.035), whereas hsCRP (LR=3.568, P=0.132) and RC (LR=1.262, P=0.503) did not reach statistical significance. This advantage persisted in all-cause mortality analysis, with RCII exhibiting the strongest association (LR=36.882, P=0.016) versus hsCRP (LR=25.792, P=0.021) and RC (LR=1.000, P=0.992).
Subgroup analysis
Subgroup analyses were conducted to assess the associations of RCII (per SD increase) with CVD risks and all-cause mortality across different populations stratified by age (<60, ≥60 years), sex (female, male), ethnicity (non-white, white), BMI (<25, ≥25 kg/m²), and diabetes status (diabetes, prediabetes). For CVD risk, a significant interaction was noted between RCII and sex (P for interaction=0.047) (Fig. 3). However, within each sex, the association between RCII and CVD risk was not statistically significant (female: HR, 1.03; 95% CI, 0.88 to 1.22; P=0.667; male: HR, 1.42; 95% CI, 0.94 to 2.13; P=0.087). In other subgroups, including age, ethnicity, BMI, and diabetes status, no significant interactions between RCII and CVD risk were observed (all P for each interaction >0.05). The relationship between RCII and all-cause mortality was consistent across various subgroups, as illustrated in Fig. 3. No significant interaction effects were observed between RCII and the stratified variables (all P for interaction >0.05).
Subgroup analyses of the association of remnant cholesterol inflammatory index (RCII) with cardiovascular disease (CVD) risks (A) and all-cause mortality (B). Logistic regression models were adjusted for sex, body mass index (BMI), poverty income ratio, hypertension, smoking, lipid-lowering drugs, antidiabetic drugs, glucose, glycosylated hemoglobin (HbA1c), and triglyceride (TG). Cox regression models were sex, BMI, poverty income ratio, hypertension, smoking, CVD, lipid-lowering drugs, antidiabetic drugs, glucose, HbA1c, and TG. The strata variable was not included in the model when stratifying by itself. P<0.05 suggests significant differences. OR, odds ratio; CI, confidence interval; SD, standard deviation; DM, diabetes mellitus; HR, hazard ratio.
DISCUSSION
To our knowledge, this prospective study of 2206 individuals with prediabetes or diabetes demonstrates that the RCII-a biomarker combining residual cholesterol (atherogenic lipoprotein remnants) and systemic inflammation-synergistic predicts all-cause mortality and CVD risks. Higher RCII quartiles exhibited progressive metabolic deterioration association. Despite non-significant survival curve divergence by log-rank test (P=0.094), fully adjusted Cox models identified RCII as an independent mortality predictor, with each SD increment conferring 21% excess risk (fully adjusted HR, 1.21; 95% CI, 1.08 to 1.35). Furthermore, higher RCII levels were associated with increased CVD risks after adjusting for multiple covariates (Q4: HR, 2.32; 95% CI, 1.23 to 4.37). Notably, stratified analyses revealed consistent RCII-outcome associations across most subgroups, though we observed potential sex-specific CVD risks modification (interaction P=0.047).
RC, a pathogenic constituent of TG-rich lipoproteins comprising very low-density lipoproteins (VLDL), intermediate-density lipoproteins, and chylomicron remnants, has emerged as a critical factor of residual cardiovascular risk in lipid management [10-13]. This association is particularly pronounced in individuals with DM, where pathophysiological alterations associated with chronic hyperglycemia and insulin resistance may induce threefold consequences: enhanced hepatic VLDL overproduction via sterol regulatory element-binding protein 1c activation; impaired lipoprotein lipase-mediated RC clearance; and increased glycation of apolipoprotein B-100, collectively driving RC accumulation in diabetic populations [14-17].
The vascular toxicity of RC manifests through infiltrating the arterial wall, accumulating in the intima, causing endothelial dysfunction and initiating vascular inflammation and cholesterol deposition [18-20]. Notably, the causal role of inflammation in atherosclerotic CVD has been definitively established through landmark trials (the Canakinumab Anti-Inflammatory Thrombosis Outcomes Study [CANTOS] and the Colchicine Cardiovascular Outcomes Trial [COLCOT]), which demonstrated 15% to 31% risk reduction with targeted anti-inflammatory therapies [21,22]. Current guidelines accordingly recognize elevated (hsCRP ≥2 mg/L) as a risk-enhancing factor for atherosclerotic CVD, particularly in high-risk subgroups such as diabetic patients [23].
The pathophysiological interplay between RC and systemic inflammation exhibits amplified clinical significance in dysglycemic populations. Mechanistically, chronic low-grade inflammation, quantified by circulating interleukin-6 (IL-6) and C-reactive protein (CRP) elevations, demonstrated significant associations with diabetes progression (IL-6: RR, 1.31; 95% CI, 1.17 to 1.46; CRP: RR, 1.26; 95% CI, 1.16 to 1.37) [24]. Clinically, this RC-inflammation synergy manifests elevation of cardiovascular risk. Analysis of the Danish National Health Registry (n=103,221) revealed that co-elevation of RC and CRP conferred a 1.9-fold increased risk of atherosclerotic CVD (HR, 1.90; 95% CI, 1.70 to 2.12) and 40% excess all-cause mortality (HR, 1.40; 95% CI, 1.30 to 1.50) versus those with the lowest levels of both biomarkers [25]. This synergistic pattern was further validated in a prospective Chinese cohort, where elevated RC and hsCRP levels together constituted the highest risk of new-onset stroke, exceeding the risk associated with each factor alone [26].
Despite these findings, clinical evidence quantifying the combined prognostic impact of RC and systemic inflammation in dysglycemic populations remains limited. Therefore, our study directly filled in this knowledge gap by introducing the RCII, a novel composite biomarker which has been reported to be associated with stroke risks [7]. Critically, our results found that high RCII levels were significantly associated with an increased risk of all-cause death and CVD in individuals with prediabetes and diabetes. Furthermore, the prognostic value of RCII in these patients was greater than the effect of RC and hsCRP alone (Supplementary Table 4). In addition, our analysis demonstrates that RCII exhibited significantly greater incremental predictive value for both CVD and all-cause mortality compared to RC and hsCRP, as evidenced by improved Harrell’s C statistics and LR test results. In individuals with prediabetes or diabetes, RCII captures the interplay between two modifiable risk pathways, remnant lipoprotein accumulation and subclinical inflammation, thereby offering superior risk discrimination compared to conventional biomarkers. The linear risk continuum and the identified treatment gap underscore the need for RCII-guided therapeutic strategies that target both lipid and inflammatory axes. Future prospective trials should evaluate whether RCII-directed escalation of lipid-lowering and/or anti-inflammatory therapies can improve clinical outcomes in this high-risk population.
Subgroup analysis demonstrated significant sex-specific differences in CVD risk associated with the RCII (P for interaction=0.047). Female participants exhibited a stronger correlation between elevated RCII and CVD risks compared to males, although within-sex comparisons showed non-significant trends (both P>0.05). Notably, females in this cohort presented higher baseline cardiometabolic risks, including older age, elevated BMI, atherogenic dyslipidemia (TC and LDL-C), HbA1c, systemic inflammation and RCII (all P<0.05). These observations align with established evidence that older women with metabolic disorders face amplified cardiovascular vulnerability due to estrogen depletion and impaired lipid metabolism [27,28], potentially explaining the sex-specific association of RCII with CVD. While our findings reinforce the need for sex-stratified risk assessment, more confirmatory studies should be conducted.
Additionally, subgroup analysis demonstrated no statistically significant interaction in CVD risk and all-cause mortality between prediabetic and diabetic subgroups (all P interaction >0.05), indicating comparable RCII association patterns across these metabolic states in our cohort. This observation aligns with our lipid profile findings, where prediabetic subjects exhibited significantly higher baseline TC (193.99±38.59 mg/dL vs. 183.48±46.29 mg/dL, P<0.001) and LDL-C levels (117.52±33.20 mg/dL vs. 106.39±40.01 mg/dL, P<0.001) compared to diabetic individuals, potentially suggesting a metabolic state-independent role of RCII that warrants validation through prospective large-scale studies. However, current subgroup analyses might lack sufficient statistical power to detect modest yet clinically meaningful intergroup differences due to inherent dataset limitations. Moreover, the relatively short follow-up duration (median 38 months) may constrain our ability to discern potential divergence in long-term clinical outcomes between these populations.
This study has several limitations that warrant consideration. First, while leveraging the nationally representative NHANES cohort, the observational design inherently precludes causal inference between RCII and clinical outcomes. Second, the lack of diabetes subtype classification in NHANES. Nevertheless, the present study may be more applicable to individuals with type 2 DM as individuals aged <20 years were excluded. Third, given the limited number of hsCRP and lipid index measurements, this study only analyzed data from two cycles spanning 2015 to 2018. The relatively short follow-up period and the small number of deaths that occurred precluded a detailed subgroup analysis of causes of death and resulted in the observation period possibly being insufficient to capture the differences in long-term outcomes. Fourth, the NHANES dataset utilized in this study captured only a single measurement of lipid and inflammatory markers, which should be considered in light of the potential for measurement error. Furthermore, we are unable to evaluate the influence of dynamic changes in RCII during follow-up on the risk of death and CVD. Lastly, although we incorporated a comprehensive array of potential confounders into our analysis and conducted subgroup analyses to ensure the robustness of our findings, the potential impact of residual confounders on our results cannot be entirely ruled out.
In conclusion, as the first study to evaluate the impact of RCII on all-cause mortality and CVD risks in patients with prediabetes and diabetes, our findings suggested that RCII was a significant predictor and may serve as a potential target for preventive interventions in these high-risk populations.
SUPPLEMENTARY MATERIALS
Supplementary materials related to this article can be found online at https://doi.org/10.4093/dmj.2025.0305.
ORs with 95% CIs for the covariates in the fully adjusted logistic regression model for CVD
HRs with 95% CIs for the covariates in the fully adjusted Cox proportional hazards model for all-cause mortality
Incremented predictive value of RC, hsCRP, and RCII for CVD and all-cause mortality risk
Association of RC and hsCRP with CVD risks and all-cause mortality
Flowchart of the study population. DM, diabetes mellitus; HbA1c, glycosylated hemoglobin; hsCRP, highsensitivity C-reactive protein; TC, total cholesterol; TG, triglyceride; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; PIR, poverty income ratio; BMI, body mass index; NHANES, National Health and Nutrition Examination Survey.
Notes
CONFLICTS OF INTEREST
No potential conflict of interest relevant to this article was reported.
AUTHOR CONTRIBUTIONS
Conception or design: Q.L.M., J.P.
Acquisition, analysis, or interpretation of data: all authors.
Drafting the work or revising: Q.L.M., J.P.
Final approval of the manuscript: all authors.
FUNDING
This study was supported by grants from the National Natural Science Foundation of China (No. 82300677) and the Hunan Provincial Natural Science Foundation of China (No.2023JJ 40978) awarded to Jia Peng.
ACKNOWLEDGMENTS
The authors thank U.S. Centers for Disease Control and Prevention’s National Center for Health Statistics (NCHS) of Centers for Disease Control and Prevention (CDC) for sharing the data.
The data for this study were obtained from the public database of the National Center for Health Statistics https://www.cdc.gov/nchs/nhanes/.