Introduction

Diabetes mellitus (DM) and its complications have emerged as a serious public health problem worldwide (Sun et al. 2022). As estimated by the International Diabetes Federation (IDF), diabetes affects 536.6 million (10.5%) globally in 2021, with projections indicating an increase to 783.2 million (12.2%) by 2045 (Sun et al. 2022). Prediabetes, also termed the intermediate stage between normal glucose regulation and diabetes, is far more prevalent than diabetes. Prediabetes affects approximately one-third of adults in the US and 720 million individuals worldwide (Echouffo-Tcheugui et al. 2023). Epidemiological evidence consistently links prediabetes to an elevated risk of diabetes-related complications, including cardiovascular disease (CVD), nephropathy, retinopathy, and neuropathy compared to those with normal glucose regulation (Brannick and Dagogo-Jack 2018; Echouffo-Tcheugui et al. 2023; Schlesinger et al. 2022). Moreover, individuals with diabetes or prediabetes also faced a remarkably heightened risk of both all-cause and cardiovascular mortality (Raghavan et al. 2019; Schlesinger et al. 2022). Therefore, identifying reliable predictors in individuals with diabetes or prediabetes is essential to decrease CVD risk and mortality and improve patient prognosis.

Chronic inflammatory processes are crucial in the pathogenesis of diabetic complications including nephropathy and vascular diseases (Yamamoto and Yamamoto 2013). The neutrophil–lymphocyte ratio (NLR) is an innovative biomarker of immunoinflammatory and is computed by neutrophil count/lymphocyte count (Zahorec 2021). The NLR integrates the contributions of innate (neutrophils) and adaptive (lymphocytes) immune responses (GARCíA-ESCOBAR A et al. 2023). Recently, accumulative evidence suggested that NLR is a strong predictor of CVD risk and related to the outcomes of CVD (Angkananard et al. 2018; JHUANG Y-H et al. 2019; Soylu et al. 2015). A prospective cohort study suggested that compared with lower NLR scores, higher NLR scores independently correlate with the severity of coronary artery disease (CAD) and elevated risk of CVD event (HR = 1.55) in patients undergoing angiography (Arbel et al. 2012). Studies have revealed that NLR is strongly related to higher glycemia and HbA1C levels and elevated risk of mortality in diabetic patients (Adane et al. 2023; Dong et al. 2023). The systemic immune-inflammation index (SII), proposed by Hu et al., reflects immunoinflammatory status through the calculation: platelet count × neutrophil count/lymphocyte count (Hu, et al. 2014). Previous research has demonstrated that SII has a superior prognostic value over other inflammation markers in cancer settings, including NLR and platelet–lymphocyte ratio (PLR) (Aziz et al. 2019; CHEN J-H et al. 2017; Jomrich et al. 2021). Emerging evidence suggested that SII was also a predictor of CVD risk and clinic prognostic (Cao et al. 2023; Xu et al. 2021; YANG Y-L et al. 2020; Ye et al. 2022). A nationwide cohort study demonstrated that SII was remarkably related to CVD risk even after accounting for multiple risk factors (Wang et al. 2023). A prospective cohort study indicated that elevated SII scores were linked to a 65% higher risk of major adverse cardiac events (MACEs) following percutaneous coronary intervention in patients with CAD compared with lower SII scores with SII proving to be a superior predictor of MACEs compared to traditional risk factors (YANG Y-L et al. 2020). Moreover, SII was positively correlated with both all-cause and CVD mortality in diabetic patients (Chen et al. 2023).

Despite these insights, investigations into the relationship between NLR and SII with CVD risk and mortality among prediabetic patients remain scarce. This study aimed to explore whether the NLR and SII have a predictive value for CVD risk, and mortality among patients with diabetes or prediabetes and to compare the predictive power for mortality.

Methods

Study design and populations

The National Health and Nutrition Examination Survey (NHANES) is a biennial cross-sectional research program aimed to evaluate the health and nutritional status of the U.S. population. The study was approved by the NCHS Institutional Review Board, and all participants provided informed consent; therefore, no extra ethical approval was required. Initially, 91,351 individuals were included in this study (NHANES 2001–2018). After excluding those with missing NLR and SII data (N = 16,559), without diabetes or prediabetes (N = 59,602), incomplete follow-up information (N = 1,199), and incomplete covariates (N = 7120), ultimately, 6871 subjects were included in the analysis (Supplementary Fig. 1).

Assessment of neutrophil–lymphocyte ratio and systemic immune-inflammation index

NLR and SII were used as exposure variables. NLR was determined by the ratio of neutrophil count to lymphocyte count. SII was computed by multiplying the platelet count and neutrophil count, then dividing by the lymphocyte count. Neutrophil, platelet, and lymphocyte counts (expressed as × 109/L) were measured using automated hematology analysis equipment.

Diagnosis of diabetes and prediabetes

Diabetes was diagnosed by fasting plasma glucose (FPG) ≥ 126 mg/dL, glycated hemoglobin A1c (HbA1c) value ≥ 6.5%, or 2-h blood glucose ≥ 200 mg/dL from an oral glucose tolerance test (OGTT), in addition to self-reported diagnosis or use of insulin or oral hypoglycemic medication (Zou et al. 2019). Prediabetes is defined by self-reported prediabetes status, FPG levels between 100 and 125 mg/dL, OGTT values ranging from 140 to 199 mg/dL, or HbA1c between 5.7 and 6.5% (Zou et al. 2019).

Diagnosis of cardiovascular disease

Diagnoses of CVD were determined by self-reported physician diagnoses derived from personal interviews. Participants were categorized with CVD if they affirmed having been diagnosed with congestive heart failure (CHF), coronary heart disease (CHD), angina pectoris, myocardial infarction (MI), or stroke by a healthcare professional (Zhang et al. 2023a). In addition, specific events related to these conditions were analyzed as secondary outcomes.

Ascertainment of mortality

Mortality status was determined via linkage to the National Death Index (NDI) records, with follow-up time durations computed from NHANES examination dates to death or December 31, 2019. Cardiovascular deaths were ascertained using ICD-10 codes I00–I09, I11, I13, I20–I51, and I60–I69.

Assessment of covariates

Structure questionnaires acquired information on sociodemographic characteristics, smoking, drinking, disease, and medication use. Education was categorized (less than high school, high school/equivalent, college, or above), income to poverty ratios (< 1, 1–3, > 3), and smoking status (never, former, current) (Li, et al. 1999). Drinking status (heavy, low-moderate drinkers, nondrinker) (Chang et al. 2017), BMI (normal, overweight, obese), hypertension (Yes/No) (Unger et al. 2020), lab analyses included triglyceride (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), serum creatinine, serum uric acid, urinary albumin to creatinine ratio (ACR), and estimated-glomerular filtration rate (eGFR) calculated via CKD-EPI (Levey et al. 2009).

Statistical analysis

All statistical analyses considered NHANES’ complex survey design, integrating sample weights, clustering, and stratification. Continuous variables were reported as means ± standard deviation (SD), while categorical variables were presented as counts and weighted proportions. Group differences were assessed via weighted chi-square and Wilcoxon rank-sum tests. The optimal cut-off for NLR and SII was determined by maximally selected rank statistics (Zhang et al. 2023b). Multivariate logistic regression was employed to evaluate odds ratios (OR) with 95% confidence intervals (CI) for CVD risk in relation to NLR and SII. Survival outcomes were depicted using Kaplan–Meier curves, with between-group comparisons via log-rank tests. The survey-weighted Cox proportional hazards regression was employed to evaluate hazard ratios (HR) and 95% CI for mortality in connection with NLR and SII. The potential nonlinear relationships were explored via restricted cubic spline (RCS) analysis. Stratification and interaction analyses were carried out using variables including gender, age, BMI, smoking status, hypertension, and diabetes. Time-dependent receiver-operator characteristic curve (ROC) analysis was performed to assess the accuracy of the NLR and SII in predicting survival outcomes at multiple time points, utilizing the ‘timeROC’ package (Kamarudin et al. 2017). R software was used for all statistical analyses (version 4.2.1). A two-tailed p value < 0.05 was deemed statistically significant.

Results

Baseline characteristics of the study population

In this study, 6871 participants with diabetes or prediabetes were enrolled (Supplementary Fig. 1). Utilizing maximally selected rank statistics, we determined the optimal NLR threshold to be 3.42, dividing participants into a lower group (NLR ≤ 3.42, n = 6,214) and a higher group (NLR > 3.42, n = 747) (Fig. 1A). Similarly, the optimal SII threshold identified was 853.86 (Fig. 1B).

Fig. 1
figure 1

The cut-off point was calculated using the maximally selected rank statistics based on the ‘maxstat’ package

Full size image

Baseline characteristics of the study participants were stratified according to NLR (Table 1). The mean (SD) age of participants was 51.3 (16.2) years, with 55.0% being male. The higher NLR group was predominantly older and of white ethnicity compared with the lower NLR group. They exhibited higher levels of TC, fasting glucose, serum creatinine, ACR, and neutrophil count, and suffered CVD, CHF, CHD, Angina, MI, stroke, hypertension, and diabetes. Furthermore, they also exhibited a lower proportion of never smokers and lower LDL, eGFR, and lymphocyte levels. The baseline characteristics stratified by SII are detailed in Supplementary Table 1.

Table 1 Baseline characteristics of included participants according to the NLR
Full size table

Relationships between NLR and SII with the risk of CVD

The relationship between NLR and CVD risk was assessed by survey-weighted logistic regression. In the unadjusted model, individuals with higher NLR exhibited a heightened risk of CVD (OR = 2.36, p < 0.001). Following multivariate adjustment, the relationship persisted across Model 2 (OR = 1.48, p < 0.001) and Model 3 (OR = 1.29, p = 0.033) (Table 2). However, NLR as a continuous variable did not exhibit a significant correlation with CVD risk after full adjustment for covariates. For CHF, multivariate logistic regression analysis displayed that the risk of CHF notably elevated in the higher NLR group (OR = 1.80, p = 0.002) (Table 3) than in the lower NLR group.

Table 2 The association between NLR and SII with the risk of CVD
Full size table
Table 3 The association between NLR and SII with the risk of CHF
Full size table

After multivariate adjustment, SII, as a categorical variable, exhibited an association with increased risks of CVD (OR = 1.29, p = 0.034) and CHF (OR = 1.54, p = 0.035) (Table 2–3). Notably, this study did not find significant associations between NLR and SII with the risks of CHD, angina, heart attack, and stroke in this study (Supplementary Table 2–5).

The potential nonlinear correlations between NLR and the risk of CVD and CHF were assessed using RCS. The results revealed a linear correlation between the NLR and the risk of CVD and CHF (both p for nonlinear > 0.05) (Fig. 2 A-B). Conversely, RCS results illustrated a U-shaped association between the SII and the risk of CVD and CHF (both p for nonlinear < 0.05) (Fig. 2 C-D).

Fig. 2
figure 2

Restricted cubic spline analysis between NLR and the risk of CVD (A) and CHF (B) in participants with prediabetes or diabetes. Restricted cubic spline analysis between SII and the risk of CVD (C) and CHF (D). Red lines represent references for odds ratios, and blue areas represent 95% confidence intervals. The model was adjusted for age, gender, race, education level, family income poverty ratio, smoking status, drinking status, BMI, TC TG, HDL, HbA1c, and hypertension

Full size image

We also explored the association between NLR and the risk of CVD and CHF in different population subgroups based on age, gender, BMI, smoking status, hypertension, and diabetes. The CVD risk was elevated only in participants who were female, age (≥ 60), obese, and had hypertension (Fig. 3A). Regarding CHF, the correlation of NLR with CHF was observed in participants who were females, age (≥ 60), normal weight, overweight, never smokers, and individuals with hypertension and prediabetes (Fig. 3B). Interaction tests indicated no statistically significant interactions between NLR and the stratification variables (all interactions p > 0.05) (Fig. 3). Similarly, there were no significant interactions between SII and the stratification variables (all interactions p > 0.05) (Supplementary Fig. 2).

Fig. 3
figure 3

Stratified analyses of the associations between NLR and CVD (A) and CHF (B) among individuals with prediabetes or diabetes

Full size image

Associations of the NLR with all‑cause and cardiovascular mortality

Throughout a mean follow-up period of 191 months, a total of 1146 (16.7%) deaths occurred, including 382 (5.6%) caused by CVD. Kaplan–Meier survival curves suggested that higher NLR was related to increased all-cause mortality and cardiovascular mortality (both log-rank p < 0.001, Fig. 4). Weighted multivariable Cox regression analysis, following comprehensive adjustment for covariates, revealed that each unit increase in NLR was independently associated with a 12% increased risk of both all-cause mortality and cardiovascular mortality (both p < 0.001) (Table 4). Statistical significance persists when NLR is transformed from a continuous to a categorical variable. After adjusting for all covariates, higher NLR was linked to elevated risk of all-cause (HR = 1.82, p < 0.001) and cardiovascular mortality (HR = 2.07, p < 0.001) (Table 4). The core results remained consistent across subgroup analyses based on age, sex, BMI category, smoking status, hypertension, and diabetes (Fig. 5). Furthermore, age-stratified analyses demonstrated interactions between NLR and all-cause mortality, highlighting a particularly pronounced effect among older individuals (Fig. 5).

Fig. 4
figure 4

Kaplan–Meier analysis of all-cause (A) and cardiovascular (B) mortality based on NLR groups among individuals with prediabetes or diabetes

Full size image
Table 4 The association between NLR and SII with all-cause and cardiovascular mortality
Full size table
Fig. 5
figure 5

Stratified analyses of NLR associations with all-cause (A) and cardiovascular (B) mortality among individuals with prediabetes or diabetes

Full size image

Additionally, RCS analysis illustrated a linear relationship between NLR and both all-cause and cardiovascular mortality (both p for nonlinear > 0.05) after full adjustment of covariates, specifically in participants with diabetes or prediabetes (Fig. 6).

Fig. 6
figure 6

Restricted cubic spline analysis between NLR and all-cause (A) and cardiovascular (B) mortality in participants with prediabetes or diabetes. Restricted cubic spline analysis between SII and the risk of all-cause (C) and cardiovascular (D) mortality. Red lines represent references for hazard ratio, and blue areas represent 95% confidence intervals. The model was adjusted for age, gender, race, education level, family income poverty ratio, smoking status, drinking status, BMI, TC TG, HDL, HbA1c, and hypertension

Full size image

Associations of the SII with all‑cause and cardiovascular mortality

Regarding SII, Kaplan–Meier survival curves similarly indicated higher SII levels were linked to elevated risks of all-cause and cardiovascular mortality (Log-rank p < 0.001, Supplementary Fig. 3). Cox regression analysis showed that after full adjustment of covariates, each unit increase in log-SII score was associated with a 67% and 144% elevated risk of all-cause (HR = 1.67, p = 0.006) and cardiovascular mortality (HR = 2.44, p = 0.004) (Table 4), respectively. Statistical significance persists when SII as a categorical variable. After adjusting for all covariates, higher SII levels were found to be associated with a remarkably elevated risk of all-cause (HR = 1.64, p < 0.001) and cardiovascular mortality (HR = 1.98, p < 0.001) compared to lower SII levels (Table 4). (Table 4). Except for overweight individuals who showed SII was not associated with cardiovascular mortality, the correlations of SII with all-cause and cardiovascular mortality were unchanged across subgroups. No significant interaction between SII and the stratification variables was observed (Supplementary Fig. 4).

Interestingly, after full adjustment of covariates, the RCS curves displayed a U-shaped association between log-SII and all-cause mortality in individuals with diabetes or prediabetes (p nonlinear < 0.0001) (Fig. 6). As depicted in Fig. 6, when log-SII was less than 2.65, the risk of all-cause mortality tended to decline as log-SII continued to increase. However, after exceeding 2.65, all-cause mortality rose as log-SII grew. RCS curves exhibited a similar nonlinear connection between SII and cardiovascular mortality (p nonlinear = 0.006), as depicted in Fig. 6. These findings underscore the complex associations between systemic inflammation, as measured by SII, and mortality risks in individuals with diabetes or prediabetes.

The predictive ability of NLR and SII for mortality in patients with prediabetes or diabetes

Time-dependent ROC analysis displayed the area under the curve (AUC) values for NLR in predicting all-cause mortality in patients with prediabetes or diabetes, which were 0.674 at 3 years, 0.659 at 5 years, and 0.639 at 10 years (Fig. 7A). Similarly, NLR demonstrated AUC of 0.674, 0.665, and 0.657 for cardiovascular mortality over the corresponding periods (Fig. 7B). The AUC for log-SII were 0.612, 0.587, and 0.558 for all-cause mortality (Fig. 7C), and 0.612, 0.594, and 0.578 for cardiovascular mortality at 3-year, 5-year, and 10-year, respectively. (Fig. 7D). Additionally, we evaluated the predictive performance of neutrophils, lymphocytes, and platelets individually for all-cause and cardiovascular mortality (Supplementary Fig. 5). These findings underscored that NLR exhibited superior predictive ability over SII, as well as neutrophils, lymphocytes, and platelets, across both short-term and long-term mortality predictions.

Fig. 7
figure 7

Time-dependent ROC curves and time-dependent AUC values of the NLR for predicting all-cause mortality (A) and cardiovascular mortality (B). Time-dependent ROC curves and time-dependent AUC values of the SII for predicting all-cause mortality (C) and cardiovascular mortality (D)

Full size image

Discussion

In this study, among the 6871 individuals diagnosed with diabetes or prediabetes from nine NHANES cycles (2001–2018), we found independent associations of higher NLR and SII with cardiovascular disease (CVD) and mortality. RCS analysis revealed a linear relationship between NLR and the risk of CVD and mortality, while SII showed a nonlinear correlation. Notably, time-dependent ROC analysis underscored the superior predictive performance of NLR over SII and other biomarkers in both short-term and long-term mortality assessments.

Chronic inflammation and immune system activation play pivotal roles in the pathogenesis of diabetes and its complications (Esser et al. 2014). Several key molecular mechanisms have been elucidated in recent years, revealing how inflammatory processes lead to vascular damage, plaque destabilization, and adverse cardiovascular events. Hyperglycemia leads to endothelial and smooth muscle dysfunction by increasing the expression of pro-inflammatory cytokines, leukocyte recruitment and activation, and subendothelial lipid accumulation, which are the main characteristics of diabetic macrovascular lesions (Petrie et al. 2018; SILVEIRA ROSSI J L et al. 2022). Chronic inflammation also leads to changes in platelet function and hemostatic components, such as fibrinogen, thrombin, and fibrin, promoting thrombosis and thereby increasing cardiovascular mortality in diabetic patients (Folsom et al. 2018; Lee et al. 2021). Additionally, neutrophil-derived neutrophil extracellular traps (NETs) have emerged as potent mediators of vascular inflammation, promoting thrombosis and exacerbating endothelial dysfunction through direct cytotoxic effects and activation of coagulation pathways. Several clinical and epidemiological studies have shown that plasma inflammatory markers, including C-reactive protein, interleukin1β, and interleukin-6, are frequently raised in diabetes patients, and correlate with prognosis (Kato et al. 2019; Lowe et al. 2014; Yang et al. 2021). In addition, clinical trials have shown that IL-1β inhibitor canakinumab reduces levels of inflammatory markers and the risk of cardiovascular events in patients with diabetes (Mikkelsen et al. 2022; Ridker et al. 2012, 2017; Rissanen et al. 2012). Hydroxychloroquine, an immunomodulatory drug, is primarily used for treating malaria and autoimmune disorders and has also been shown to reduce the risk of diabetes and CVD by decreasing the production of pro-inflammatory cytokines and blocking Toll-like receptor signaling (Pareek et al. 2014; Sharma et al. 2016; Wondafrash et al. 2020). This implies a potential cumulative effect between inflammation levels with hyperglycemia and CVD.

The white blood cell (WBC) count is a straightforward, cost-effective, and widely accessible marker of nonspecific inflammation. Multiple studies have shown that elevated levels of WBC are associated with both the onset and adverse outcomes of cardiovascular disease (CVD) (Jee et al. 2005; Kabat et al. 2017; Nilsson et al. 2014). For instance, a prospective study from Korea found that increased WBC count was significantly associated with both all-cause mortality and atherosclerotic cardiovascular death (Jee et al. 2005). Nilsson et al. further supported the clinical value of WBC count in predicting long-term survival, showing that higher WBC counts in the elderly were linked to poorer survival outcomes (Nilsson et al. 2014). However, WBC count does not differentiate between various types of white blood cells. The novel inflammatory markers NLR and SII, computed from the counts of neutrophils, lymphocytes, and platelets, better reflect the levels of systemic inflammatory response (Li et al. 2023; Marik and Stephenson 2020). NLR and SII have received attention as potential inflammation biomarkers for various diseases, including cancer, metabolic syndrome, neurodegenerative, and inflammation diseases (HASHEMI MOGHANJOUGHI P et al.. 2022; Sayed et al. 2020; Zhao et al. 2023). NLR and SII are also predictors of composite cardiovascular events (Angkananard et al. 2018; Ye et al. 2022). A large cohort study involving 13,929 middle-aged and older adults demonstrated a significant association between elevated SII and incident CVD among individuals without prior heart disease at baseline (Xu et al. 2021). Our results found that higher levels of NLR and SII were associated with CVD. Interestingly, the subgroup analyses showed that higher SII was strongly related to elevated CVD risk in participants with prediabetes than diabetes. This difference may be attributed to the use of diabetes medications, such as SGLT2 inhibitors and GLP-1R agonists, which have been proven to improve heart health (Packer et al. 2020; Palmer et al. 2021). In addition, contrary to previous studies that showed higher SII and NLR are associated with CHD risk (Ma and Li 2023; Wang et al. 2024), our results suggested no association between them. The possible reason is that smaller scales may blunt the link between NLR and SII with event outcomes.

Most studies have evaluated the association of hematologic parameters with the risk of all-cause and cardiovascular mortality only in diabetic patients. In a recent study, after full adjustment of covariates, neutrophils and platelets did not exhibit associations with either all-cause or cardiovascular mortality in individuals with diabetes, whereas lymphocyte was only related to all-cause mortality (Cardoso et al. 2021). Furthermore, this study showed a robust correlation between the NLR and both types of mortality (Cardoso et al. 2021). Consistent findings were observed in a national cohort study comprising 3,251 diabetes patients (Dong et al. 2023). Few studies assessed association between the NLR with mortality risk in prediabetic patients. Compared to normoglycemia, all-cause mortality also exhibited a higher risk in prediabetes. A meta-analysis demonstrated that the excess absolute risk in prediabetes at baseline for mortality rates was 7.36 per 10,000 person-years (Cai et al. 2020). In our investigation, we found that elevated NLR levels was linked to an elevated the risk of all-cause and cardiovascular mortality both in prediabetes or diabetes at baseline, and the core results were reliable in subgroup analyses. Consistent with the Chen et al.’s study in type 2 diabetes (Chen et al. 2023), we also found that higher SII correlated with increased the risk of all-cause and cardiovascular mortality, either in patients with diabetes or prediabetes. These results suggested that NLR and SII may also be a reliable predictor for managing in prediabetes patients. The RCS curve revealed a linear correlation between NLR with all-cause and cardiovascular mortality, while SII showed a nonlinear correlation. The U-shaped relationship between platelet count and mortality may be the underlying cause of this discrepancy. A cohort study enrolling 21,252 individuals suggested that platelet < 175 × 109/L or > 300 × 109/L was dramatically increased for mortality (Vinholt et al. 2016). Furthermore, Tsai M et al. found that both thrombocytopenia and increased platelet count were associated with cardiovascular mortality (TSAI M-T et al. 2015).

The predictive power of NLR and SII for all-cause and cardiovascular mortality was evaluated. SII, which integrates three common immune cells, may possess a higher potential clinical application. In the field of cancer, SII had a better predictive value for prognosis than NLR (Aziz et al. 2019; CHEN J-H et al. 2017; Jomrich et al. 2021). In contrast, in the present study, the time-dependent ROC results showed that NLR showed a higher predictive ability than SII in predicting mortality in individuals with diabetes or prediabetes.

This study possesses several notable advantages, and its substantial sample size and extended follow-up duration contribute to reliable results and strong statistical power. In addition, all data were collected in a standardized manner, thus avoiding selection bias. Second, the use of sampling weight enhances the applicability of our findings to the U.S. population. However, several limitations should be considered in this study. First, causality could not be established since this study was observational design. Second, NLR and SII was obtained from a single baseline blood sample and may not represent the average levels during the follow-up period, so the effect of changes in the NLR and SII on the risk of cardiovascular events and mortality across the follow-up period cannot be determined. Third, although we endeavor to account for potential confounding covariates, residual confounders cannot be entirely ruled out. The predominant composition of the study cohort from the United States may limit the generalizability of the findings globally.

Conclusion

This cohort study suggested that NLR and SII are valid predictors of the risk of CVD and both all-cause and cardiovascular mortality in populations diagnosed with diabetes or prediabetes, with a linear correlation between NLR and CVD and mortality, whereas there is a nonlinear correlation in SII. The predictive power of NLR was superior to SII. Therefore, measurement of the NLR may be useful in evaluating the risk and predicting the outcome of such patients. Further studies should examine whether interventions targeting NLR improve clinical outcomes and understand the underlying mechanisms.