Blood urea nitrogen to serum albumin ratio, systemic inflammation response index and diabetic macrovascular complications in Chinese inpatients: a cross-sectional analysis Authors: Chen-Xi Jin, Xiao-Yan Wang, Guo-Yong Xie, Yong-Qi Liang, Jing-Jing Liang, Yi-Ning Xu, Xiao-Meng Wang, Jun-Yan Huang, Xian-Bo Wu, Meng-Chen Zou Published: 2026 ABSTRACT AbstractBackgroundDiabetic patients often experience prolonged hyperglycemia, which leads to sustained harm and malfunction of various body organs. The blood urea nitrogen (BUN) to serum albumin ratio (BAR) and systemic inflammation response index (SIRI) have been proven to be reliable indicators. We aimed to investigate the association of BAR, SIRI and their combination with the risk of diabetic macrovascular diseases.MethodsA total of 2,911 participants from Nanfang Hospital diagnosed with type 2 diabetes mellitus (T2DM) were included. Restricted cubic spline (RCS) and logistic regression modelling were used to assess the relation between different values of BAR, SIRI and combined BAR-SIRI and the risk of diabetic macrovascular complications. Receiver operating curve (ROC) analysis was used to compare the predictive value of BAR, SIRI, and their combination for macrovascular outcomes.ResultsHigh value of BAR, SIRI and their combination were associated with increased risk of macrovascular complications (BAR: odds ratio [OR] = 1.45, 95% confidence interval [CI]: 1.18–1.78; SIRI: OR(95% CI) = 1.25(1.03,1.52); BAR-SIRI: OR(95%CI) = 1.51(1.21,1.88); all p < 0.05). ROC curve analysis demonstrated that combined BAR-SIRI has the best predictive value for risk of macrovascular outcomes (area under the ROC curve [AUC]: BAR-SIRI: 0.678 > SIRI: 0.674 > BAR: 0.673).ConclusionBAR, SIRI and their combination exhibit a notable positive correlation with the risk of diabetic macrovascular disease and may be independent predictors of macrovascular complications.Clinical trial registrationClinicalTrials.gov, TRN: NCT06651983, Registration date: 19 May 2025.Graphical abstract Supplementary InformationThe online version contains supplementary material available at 10.1186/s41043-026-01373-z.Keywords: Blood urea nitrogen to serum albumin ratio, Systemic inflammation response index, Type 2 diabetes, Diabetic macrovascular disease Introduction Diabetic patients often experience prolonged hyperglycemia, which leads to sustained harm, malfunction, and potential failure of various body organs, particularly impacting the eyes, renal system, nerves, cardiovascular tissue, and vascular structures [1]. The macrovascular complications of T2DM encompass cardiovascular disorders (coronary disease, cardiomyopathy, arrhythmias), cerebrovascular events, and peripheral vascular disease, accounting for most diabetes-related deaths [2]. As such, the lifespan of diabetes patients has increased, resulting in a longer duration of managing various complications, thereby placing a significant psychological and financial burden on individuals with diabetes [3–5]. Blood urea nitrogen (BUN) is commonly used as an indicator of renal function, and recently it has been examined as a marker in conditions other than renal function [6]. Study has shown that BUN levels were positively associated with incident T2DM risk [7] as well as higher risk of cardiovascular disease (CVD) and all-cause mortality [8]. Albumin (ALB) constitutes the primary component of plasma proteins, and changes are associated with the prognosis of various diseases [9]. The BUN to serum albumin ratio (BAR) is readily available [10], and it serves as an indicator for predicting the severity of chronic obstructive pulmonary disease (COPD), septic conditions, and intracranial hemorrhages [11–13]. Notably, it has been shown to be a more reliable indicator than ALB or BUN alone [9]. Researches have revealed that elevated BAR levels were notably linked to a higher likelihood of mortality in diabetic patients [9] or in those with heart failure (HF) [14]. Importantly, based on our previous fundamental research, integrated metabolomic analyses of both human subjects and animal models revealed that patients with diabetic peripheral arterial disease (PAD) exhibit decreased thyroid hormone levels alongside elevated N-class metabolites (particularly heightened homocitrulline levels) [15]. To further validate our findings regarding N-class metabolites, we incorporated patients from the UK Biobank (UKB) database and demonstrated that BAR effectively predicts mortality risk across the full spectrum of vascular complications - from prediabetes, uncomplicated diabetes, to diabetes with multiple complications [16]. Building upon these results, we subsequently conducted an exploratory analysis to investigate the association between BAR and diabetic macrovascular complications using our institutional patient cohort. Systemic inflammation response index (SIRI) was developed based on lymphocyte, peripheral neutrophil and monocyte counts [17]. As a novel composite index incorporating three distinct white blood cell subsets, it has demonstrated significant associations with both cardiovascular and all-cause mortality [18] and in population with chronic kidney disease [19]. Besides, evidence from a retrospective cohort study confirms that SIRI serves as a significant predictor of clinical outcomes in stroke patients [20]. Additionally, analysis of 6,412 U.S. adults from the National Health and Nutrition Examination Survey (NHANES) demonstrated significant associations between SIRI and both all-cause mortality and cardio-cerebrovascular disease (CCD) mortality in diabetic patients [21]. Although the BAR and SIRI has been examined in diabetes, CVD, diabetic nephropathy, and mortality for diabetes, limited studies have examined the association of the combined effects of BAR and SIRI with diabetic macrovascular complications. Moreover, the majority of existing studies have relied on public databases such as NHANES, whereas our data were derived from hospitalized patients in our own institution, which will provide unique clinical significance. In this study, we aimed to evaluate the associations between the incidence of macrovascular complications in diabetic persons with BAR and SIRI values. The valuable insights gained from this analysis will provide crucial evidence for enhancing the strategies used in preventing and managing diabetic macrovascular disease. Methods Study population In Nanfang Hospital, 3,262 patients were hospitalized in the endocrinology department with a diagnosis of T2DM from 2020 to 2023. Patients with missing SIRI (n = 11), BUN or ALB (n = 255) data, or with missing data on macrovascular disease outcome (n = 77) were excluded. Thus, in the final analysis, a total of 2,911 participants were taken into consideration (Figure S1). Exposure and experimental assessment Demographic characteristics, medical history, and clinical data of patients hospitalized at Nanfang Hospital were collected from the patient admission records. Patient weight, height and blood pressure were measured using standard procedures. The calculation for body mass index (BMI) involves dividing an individual’s weight, measured in kilograms, by the square of their height, measured in meters. Medication information included pre-admission medications and medications administered during hospitalization. Fasting venous blood samples were collected for measurements of HbA1c, serum ALB, BUN and creatinine. The estimated glomerular filtration rate(eGFR) was calculated by the 2021 CKD-EPI equation. Ascertainment of outcomes T2DM is defined as (1) a previously reported diagnosis by a medical professional, or (2) fasting plasma glucose ≥ 7.0 mmol/L, or (3) HbA1c ≥ 6.5 mmol/L, or (4) taking diabetes drugs. Diabetic macrovascular complications include CVD, HF, PAD, stroke, coronary heart disease (CHD) and myocardial infarction. For the data collected from Nanfang Hospital, the diagnosis of diabetes and macrovascular complications were extracted from the patient medical records. Covariates The following covariates were used in the analyses: age < 60 years or ≥ 60 years, sex (male or female), smoking status (yes or no), drinking status (yes or no), history of hypertension (yes or no), history of hyperlipidemia (yes or no), use of insulin (yes or no), and family history of diabetes (yes or no); these were analyzed as categorical variables. Body mass index (BMI, kg/m2), course of diabetes (years), and hemoglobin A1c (HbA1c, %) were analyzed as continuous variables. Statistical analysis Continuous data were reported as mean (SD; standard deviation), or median (IQR; interquartile range) based on the evaluation of a normal distribution assessed by Shapiro-Wilk test, and categorical data were reported as count (n) and percentage (%). Statistical analyses were conducted to detect variations among multiple groups of continuous and categorical variables, utilizing the one-way ANOVA, Kruskal-Wallis H test, and chi-squared tests, respectively. Categorical variables in these and all subsequent regression analyses were implemented using dummy coding. BAR and SIRI values were categorized into 3 distinct groups according to their numerical values, and the relations with macrovascular complications were examined. The reference category was the low BAR and SIRI value group. Odds ratios (ORs) accompanied by 95% confidence intervals (CIs) were employed, derived through the application of logistic regression models. To evaluate the combined effects of BAR and SIRI, we stratified participants into four groups based on median cutoff values (BAR: 4.028 mg/g; SIRI: 0.936): Group 1 (reference): BAR < 4.028 mg/g & SIRI < 0.936; Group 2: BAR < 4.028 mg/g & SIRI ≥ 0.936; Group 3: BAR ≥ 4.028 mg/g & SIRI < 0.936; Group 4: BAR ≥ 4.028 mg/g & SIRI ≥ 0.936. Three models were developed for logistic regression analysis: Model 1 was unadjusted; Model 2 was adjusted for age and sex; Model 3 was adjusted for age (< 60 or ≥ 60 years old), sex (male or female), BMI (as a continuous variable), course of diabetes (as a continuous variable), drinking status (yes or no), smoking status (yes or no), hypertension (yes or no), hyperlipidemia (yes or no), insulin treatment (yes or no), family history of diabetes (yes or no), and HbA1c (as a continuous variable). In logistic regression models, we employed the Restricted Cubic Spline (RCS) approach to investigate the nonlinear associations of diabetic macrovascular complications with BAR, SIRI and their combination. The knot selection for the RCS was informed by the Akaike Information Criterion (AIC). We systematically tested different knot numbers and positions, and the final adopted configuration was the one that minimized the AIC, thus ensuring an optimal balance between capturing nonlinearity and preventing overfitting. The additive and multiplicative effects of BAR and SIRI on diabetic macrovascular outcomes were assessed. Relative excess risk due to interaction (RERI), attributable proportion (AP), and synergy index (SI) with 95%CI were calculated to assess additive interactions, employing variance estimation methods established in prior studies [22]. Receiver operating characteristic (ROC) analysis was used to compare the predictive value of BAR, SIRI, and their combination for the risk of diabetic macrovascular disease. Sensitivity analysis focused on the association of BAR, SIRI and their combination with the risk of diabetic macrovascular disease was performed, and it excluded patients with extreme BAR or SIRI values (top and bottom 1%), with cancer history or with eGFR < 60 ml/min/1.73m2. Subgroup analysis was performed to examine potential variations in the impact of BAR, SIRI and their integration on the incidence of macrovascular complications. The analysis stratified the final analytical sample by age (< 60 and ≥ 60 years), sex (male or female), drinking status (yes or no), smoking status (yes or no), hypertension (yes or no), hyperlipidemia (yes or no), insulin treatment (yes or no), family history of diabetes (yes or no). Statistical analysis was performed with R (version 4.4.1) and Stata software. All statistical tests were 2-sided, and a value of p < 0.05 was considered statistically significant. Results Baseline characteristics A total of 2,911 patients from Nanfang Hospital were included in the analysis, and 1,043 (36%) were ≥ 60 years old, 1084 (37%) were female, and 1,135 (39%) were diagnosed with diabetic macrovascular disease (Table 1). Compared to participants without macrovascular complications, those with macrovascular disease were more likely to be older, female, smokers, have a longer course of diabetes, be treated with insulin, and have a history of hypertension. Notably, the proportion of patients with diabetic macrovascular complications exhibited a progressive increase with higher values of BAR, SIRI, and their combined measures. Table 1Baseline characteristic of participantsVariablesNon-diabetic macrovascular diseaseDiabetic macrovascular diseaseP value(n = 1,776)(n = 1,135)Age (< 60 years, ≥ 60 years)< 0.001 < 60 years1,294(72.86%)574(50.57%) ≥60 years482(27.14%)561(49.43%)Sex (F, M)0.332 Female649(36.54%)435(38.33%) Male1,127(63.46%)700(61.67%)Hypertension (yes or no)< 0.001 Yes614(34.57%)619(54.54%) No1,162(65.43%)516(45.46%)Hyperlipidemia (yes or no)0.003 Yes797(44.88%)447(39.38%) No979(55.12%)688(60.62%)Insulin treatment (yes or no)0.005 Yes315(17.74%)249(21.94%) No1,461(82.26%)886(78.06%)Family history of diabetes (yes or no)0.058 Yes574(32.32%)329(28.99%) No1,202(67.68%)806(71.01%)Smoke (yes or no)0.197 Yes634(35.70%)432(38.06%) No1,142(64.30%)703(61.94%)Alcohol (yes or no)0.193 Yes536(30.18%)317(27.93%) No1,240(69.82%)818(72.07%)Course of diabetes (years)5(9)10(12)< 0.001BMI23.8(4.8)23.4(4.3)0.003HbA1c9.3(4.0)8.9(3.7)< 0.001BAR (tertiles) (mg/g)< 0.001 Q1688(38.74%)283(24.93%) Q2589(33.16%)381(33.57%) Q3499(28.10%)471(41.50%)SIRI (tertiles)< 0.001 Q1642(36.15%)330(29.07%) Q2620(34.91%)349(30.75%) Q3514(28.94%)456(40.18%)BAR-SIRI< 0.001 Group 1608(34.23%)244(21.50%) Group 2385(21.68%)217(19.12%) Group 3356(20.05%)248(21.85%) Group 4427(24.04%)426(37.53%)BAR: Q1 (0.576–3.474), Q2 (3.475–4.805), Q3 (4.810-20.477); SIRI: Q1 (0.004–0.720), Q2 (0.721–1.270), Q3 (1.271–10.192); BAR-SIRI: Group1 (BAR < 4.028 mg/g & SIRI < 0.936), Group2 (BAR < 4.028 mg/g & SIRI ≥ 0.936), Group3 (BAR ≥ 4.028 mg/g & SIRI < 0.936), Group4 (BAR ≥ 4.028 mg/g & SIRI ≥ 0.936)BMI, body mass index; HbA1c, hemoglobin A1c; BAR, blood urea nitrogen to serum albumin ratio; SIRI, system inflammation response index Baseline characteristic of participants BAR: Q1 (0.576–3.474), Q2 (3.475–4.805), Q3 (4.810-20.477); SIRI: Q1 (0.004–0.720), Q2 (0.721–1.270), Q3 (1.271–10.192); BAR-SIRI: Group1 (BAR < 4.028 mg/g & SIRI < 0.936), Group2 (BAR < 4.028 mg/g & SIRI ≥ 0.936), Group3 (BAR ≥ 4.028 mg/g & SIRI < 0.936), Group4 (BAR ≥ 4.028 mg/g & SIRI ≥ 0.936) BMI, body mass index; HbA1c, hemoglobin A1c; BAR, blood urea nitrogen to serum albumin ratio; SIRI, system inflammation response index Association between the BAR and SIRI and risk of diabetic macrovascular disease RCS were employed in logistic regression models and showed that an increased BAR and SIRI value was associated with increased risk of macrovascular complications: BAR, p for overall = 0.0013, p for nonlinear < 0.001; SIRI, p for overall = 0.0245, p for nonlinear = 0.0267 (Fig. 1a&b). Exposure-response associations between the BAR, SIRI and BAR-SIRI and the risk of diabetic macrovascular disease. (a) exposure-response association between BAR and diabetic macrovascular diseases; (b) exposure-response association between SIRI and diabetic macrovascular diseases; (c) exposure-response association between BAR-SIRI and diabetic macrovascular diseases. Model adjusted for age, sex, BMI, drinking status, smoking status, course of diabetes, insulin treatment, family history of diabetes, history of hyperlipidemia, history of hypertension, and HbA1c. BAR, blood urea nitrogen to serum albumin ratio; SIRI, system inflammation response index; BMI, body mass index; HbA1c, hemoglobin A1c Results of the logistic regression analysis of the relations between BAR and SIRI tertile and the risk of macrovascular disease are shown in Table 2. Compared to participants in the lowest BAR and SIRI tertile, those in the highest tertile had a significantly higher risk of the outcomes in fully adjusted model: BAR: OR = 1.45, 95% CI: 1.18–1.78; SIRI: OR = 1.25, 95% CI: 1.03–1.52, all p < 0.05. Table 2Logistic regression analysis of BAR and SIRI and the risk of diabetic macrovascular diseaseVariablesModel 1Model 2Model 3OR(95%CI)P valueOR(95% CI)P valueOR(95%CI)P valueBAR(mg/g)Q1:0.576–3.474Ref-----Q2:3.475–4.8051.57(1.30,1.90) < 0.001 1.41(1.16,1.71) 0.001 1.30(1.06,1.58) 0.010 Q3:4.810-20.4772.29(1.90,2.77) < 0.001 1.81(1.49,2.21) < 0.001 1.45(1.18,1.78) < 0.001 SIRIQ1:0.004–0.720Ref-----Q2:0.721–1.2701.10(0.91,1.32)0.3401.04(0.86,1.27)0.6570.98(0.80,1.19)0.819Q3:1.271–10.1921.73(1.44,2.07) < 0.001 1.47(1.22,1.78) < 0.001 1.25(1.03,1.52) 0.026 Model 1 was not adjustedModel 2 was adjusted for age and sexModel 3 was adjusted for age, sex, course of diabetes, drinking status, smoking status, BMI, HbA1c, insulin treatment, family history of diabetes, history of hyperlipidemia and history of hypertensionBMI, body mass index; HbA1c, hemoglobin A1c; BAR, blood urea nitrogen to serum albumin ratio; SIRI, system inflammation response index; OR odds ratio; CI, confidence intervalBold values represent P < 0.05 (statistically significant) Logistic regression analysis of BAR and SIRI and the risk of diabetic macrovascular disease BAR (mg/g) Model 1 was not adjusted Model 2 was adjusted for age and sex Model 3 was adjusted for age, sex, course of diabetes, drinking status, smoking status, BMI, HbA1c, insulin treatment, family history of diabetes, history of hyperlipidemia and history of hypertension BMI, body mass index; HbA1c, hemoglobin A1c; BAR, blood urea nitrogen to serum albumin ratio; SIRI, system inflammation response index; OR odds ratio; CI, confidence interval Bold values represent P < 0.05 (statistically significant) The joint association of the combined BAR and SIRI incident risk of diabetic macrovascular disease Table 3 shows the combined effects of BAR and SIRI exposure on incident macrovascular complications. Compared with group 1, the estimate of incident complications was highest in group 4 (Model 3: OR, 1.51, 95% CI, 1.21–1.88). The exposure-response associations of combined BAR and SIRI with the complication outcomes is presented in Fig. 1c. The interaction analyses showed that no statistically significant additive or multiplicative interaction was observed between BAR and SIRI (SI: 1.211, 95% CI: 0.101–2.322) (Table S1. Assessment of the additive interaction between BAR and SIRI on incident macrovascular disease-Table S2.Assessment of the multiplicative interaction between BAR and SIRI on incident macrovascular disease). Table 3The joint association of BAR with SIRI and incident risk of diabetic macrovascular disease among the participantsVariableModel 1Model 2Model 3OR(95%CI)P valueOR(95% CI)P valueOR(95%CI)P valueGroup 1Ref-----Group 21.40(1.12,1.76) 0.003 1.30(1.03,1.63) 0.025 1.16(0.92,1.47)0.203Group 31.74(1.39,2.16) < 0.001 1.47(1.17,1.84) 0.001 1.25(0.99,1.58)0.057Group 42.49(2.03,3.04) < 0.001 1.96(1.59,2.42) < 0.001 1.51(1.21,1.88) < 0.001 Group1 (BAR < 4.028 mg/g & SIRI < 0.936), Group2 (BAR < 4.028 mg/g & SIRI ≥ 0.936), Group3 (BAR ≥ 4.028 mg/g & SIRI < 0.936), Group4 (BAR ≥ 4.028 mg/g & SIRI ≥ 0.936)Model 1 was not adjustedModel 2 was adjusted for age and sexModel 3 was adjusted for age, sex, course of diabetes, drinking status, smoking status, BMI, HbA1c, insulin treatment, family history of diabetes, history of hyperlipidemia and history of hypertensionAbbreviations: BMI, body mass index; HbA1c, hemoglobin A1c; BAR, blood urea nitrogen to serum albumin ratio; SIRI, system inflammation response index; OR odds ratio; CI, confidence intervalBold values represent P < 0.05 (statistically significant) The joint association of BAR with SIRI and incident risk of diabetic macrovascular disease among the participants Group1 (BAR < 4.028 mg/g & SIRI < 0.936), Group2 (BAR < 4.028 mg/g & SIRI ≥ 0.936), Group3 (BAR ≥ 4.028 mg/g & SIRI < 0.936), Group4 (BAR ≥ 4.028 mg/g & SIRI ≥ 0.936) Model 1 was not adjusted Model 2 was adjusted for age and sex Model 3 was adjusted for age, sex, course of diabetes, drinking status, smoking status, BMI, HbA1c, insulin treatment, family history of diabetes, history of hyperlipidemia and history of hypertension Abbreviations: BMI, body mass index; HbA1c, hemoglobin A1c; BAR, blood urea nitrogen to serum albumin ratio; SIRI, system inflammation response index; OR odds ratio; CI, confidence interval Bold values represent P < 0.05 (statistically significant) Comparison of the predictive value The ROC curves produced using the predictor variables (BAR, SIRI, and their combination) are displayed in Fig. 2. Among all the diabetic patients, combined BAR and SIRI demonstrated a slight improvement in predictive ability for diabetic macrovascular disease over the individual biomarkers (AUC: BAR-SIRI = 0.678 vs. BAR = 0.673 or SIRI = 0.674). The corresponding concordance statistics (c-statistics) for three models are detailed in Table S3 (Table S3.The C-statistic of BAR, SIRI and their combination for diabetic macrovascular disease.). Receiver operating characteristic (ROC) curves for the prediction of diabetic macrovascular disease. (a) ROC curves of BAR and BAR-SIRI for the prediction of diabetic macrovascular disease; (b) ROC curves of SIRI and BAR-SIRI for the prediction of diabetic macrovascular disease. Model adjusted for age, sex, BMI, drinking status, smoking status, course of diabetes, insulin treatment, family history of diabetes, history of hyperlipidemia, history of hypertension, and HbA1c. BAR, blood urea nitrogen to serum albumin ratio; SIRI, system inflammation response index; BMI, body mass index; HbA1c, hemoglobin A1c; AUC, area under curve Sensitivity and subgroup analyses Sensitivity analysis was conducted to assess the potential influence of reclassifying BAR and SIRI values, the presence of extreme BRI and SIRI measurements, reduced renal function (eGFR < 60 mL/min/1.73 m²), or cancer diagnosis on the observed associations between BAR, SIRI, their combined measure, and macrovascular outcomes. As displayed in Fig. 3 and Table S4-7 (Table S4. Results of sensitivity analyses after excluding the maximum (> 99%) or minimum (1%) value of BAR (N = 58), Table S5. Results of sensitivity analyses after excluding the maximum (> 99%) or minimum (1%) value of SIRI (N = 58), Table S6. Results of sensitivity analyses after excluding patients with cancer history (N = 28), Table S7. Results of sensitivity analyses after excluding patients with eGFR < 60 ml/min/1.73m2 (N = 501)), there were still significant and stable associations. Subgroup analyses stratified by sex, age, hypertension, hyperlipidemia, insulin treatment, family history of diabetes, smoking status, and drinking status was subsequently performed, and the results are shown in Figure S2-4. For the BAR, the associations were more pronounced in males and patients with hypertension and SIRI demonstrated stronger associations among smokers (all p < 0.05 for interactions). However, the combined BAR-SIRI index did not exhibit significant effect modification by these variables. Results of sensitivity analyses after by reclassifying BAR and SIRI. (a) results of sensitivity analyses after by reclassifying BAR; (b) results of sensitivity analyses after by reclassifying SIRI. Model adjusted for age, sex, BMI, drinking status, smoking status, course of diabetes, insulin treatment, family history of diabetes, history of hyperlipidemia, history of hypertension, and HbA1c. BAR, blood urea nitrogen to serum albumin ratio; SIRI, system inflammation response index; BMI, body mass index; HbA1c, hemoglobin A1c; OR odds ratio; CI, confidence interval Besides, the distribution of the individual macrovascular complications that constituted the outcome and their association with BAR and SIRI are presented in Table S8 (Table S8. Distribution of individual components of the diabetic macrovascular diseases and their association with BAR and SIRI.). It is noteworthy that the patients were predominantly characterized by CHD and PAD, with a lower observed frequency of HF, stroke and myocardial infarction. As shown in Table S8, the associations of BAR, SIRI, and their combination with the risk of CHD and PAD were basically consistent with the overall macrovascular outcome. Discussion In this cross-sectional study utilizing data from Nanfang Hospital, logistic regression analyses revealed a significant positive association between the BAR, SIRI, and their combined index with diabetic macrovascular disease. Furthermore, subgroup and sensitivity analyses consistently supported the robustness of these associations. According to our research, the RCS curve revealed that elevated BAR and SIRI levels were significantly linked to a heightened risk of diabetic macrovascular disease. Upon adjusting for confounding factors through logistic regression analysis, we discovered that a high BAR and SIRI level as well as their combination was an independent predictor of macrovascular complications. Furthermore, our analysis of subgroups revealed that BAR, SIRI and combined BAR-SIRI served as potent predictors of macrovascular complications in diabetic patients across a range of specific scenarios. Besides, the ROC curve analyses demonstrated that the combined BAR-SIRI index had better predictive value for outcomes compared with BAR and SIRI. However, the overall modest AUC values (0.67–0.68) indicate that while statistically significant, their discriminatory power is limited in a clinical context. It is noteworthy that these values are comparable to those of several established clinical risk scores for cardiovascular disease in T2DM [23, 24]. This supports their potential role not as definitive diagnostic predictors, but as accessible components for initial risk stratification or screening. There are some previous studies that confirm our findings. For example, a study included 7,043 US adults with diabetes from the NHANES revealed that a higher BAR level was significantly related to the incidence of CVD, cardiovascular mortality, and all-cause mortality [25]. Besides, BAR has been found to be a effective predictor for not only the incidence but also the dynamic progress of T2DM (from pre-T2DM to T2DM, diabetic complications, and mortality) [16]. Using the MIMIC-IV database, a retrospective cohort study identified BAR as a novel biomarker reflecting both inflammatory and nutritional status, with significant links to adverse cardiovascular outcomes [26]. As for SIRI, it has been indicated to have a positive correlation with diabetic foot ulcer (OR = 1.24; 95% CI: 1.07, 1.43) according to a cross-sectional study involving 1,246 diabetic patients from NHANES. Moreover, in individuals with diabetes and a high BMI (> 24 kg/m²), elevated SIRI was found to be an independent risk factor for CVD [27]. There are some potential mechanisms that have been described to explain our findings. Some studies have found that elevated circulating levels of urea lead to an increase of islet protein O-GlcNAcylation, and are associated with insulin secretory defects and insulin resistance [28]. Besides, ALB is the other component of the BAR, and as the main component of plasma proteins it has important roles in anti-inflammatory and, antioxidant processes [29]. A previous study has addressed the fact that hypoalbuminemia is associated with insulin deficiency, and insulin is an important regulator of albumin synthesis [30]. Consequently, BAR may contribute to the progression of diabetic macrovascular complications by exacerbating insulin resistance. Therefore, we hypothesize that a high BAR, indicative of uremic toxin accumulation and hypoalbuminemia, may directly creating a substrate of endothelial vulnerability and metabolic derangement, creating a susceptible environment within the body [28–30]. SIRI-calculated using neutrophil, monocyte, and lymphocyte counts-represents a valuable prognostic biomarker in disorders characterized by predominant inflammatory pathophysiology [31]. Given that atherosclerosis is fundamentally a chronic inflammatory disease of the arterial wall, characterized by inadequate inflammatory responses and dysregulated lipid metabolism [32], both BAR and SIRI may contribute to the pathogenesis of diabetic macrovascular complications through multiple pathways. These biomarkers potentially promote oxidative stress by disrupting the balance between oxidation and antioxidation systems. The combination of a vulnerable vascular substrate (high BAR) and an active inflammatory driver (high SIRI) may thus synergistically accelerate the disease process, potentially through shared pathways like amplified oxidative stress, ultimately accelerating the development of macrovascular disease in diabetic patients [31, 32]. Notably, diabetic macrovascular disease is comprised of various complications, including CHD, PAD, myocardial infarction, stroke, and HF. This composite nature of our endpoint is critical for interpreting our results. While BAR and SIRI are established predictors of atherosclerotic events [16, 18, 19, 25, 27, 32], their role in heart failure within diabetes remains less defined. The inclusion of heart failure may have contributed to the observed ‘modest’ AUC values, as the underlying mechanisms (e.g., inflammatory and nutritional pathways influencing myocardial remodeling versus plaque instability) likely differ [14, 33–35]. This, however, also represents a novel exploratory aspect of our study, suggesting that these simple biomarkers might capture systemic risk across a broader spectrum of diabetic macrovascular complications. In the future, larger and disease-specific studies hold the potential to clarify the role of BAR and SIRI across different macrovascular diseases. Our subgroup analysis results indicated that the correlation between SIRI and macrovascular outcomes was more evident in smokers. While it is well-established that smoking exacerbates diabetic complications through systemic inflammation, oxidative stress, and endothelial dysfunction [36–39], our finding suggests that the inflammatory state quantified by SIRI and the pro-inflammatory burden induced by smoking may jointly contribute to the elevated vascular risk. This observation warrants further investigation to elucidate the precise nature of their interaction. In addition, an increasing number of studies have confirmed a bidirectional relation between diabetes and cancer. Hyperglycemia has been demonstrated to promote cancer growth and proliferation, and individuals who have survived cancer may face a heightened likelihood of developing diabetes, potentially due to common risk factors shared by both conditions and the impacts of cancer treatment [40]. After excluding patients with diabetic nephropathy and those with cancer, our analysis provided the same results as when it included these patients, which proves the validity of the epidemiologic relationship between BAR, SIRI and their combination and adverse macrovascular disease in patients with nondiabetic kidney disease or noncancer diseases. Our investigation has several advantages. We utilized data from our hospital rather than public databases, which enhances the clinical relevance of our findings as the results are more directly applicable to our local patient population, thereby providing targeted guidance for clinical practice. Additionally, in-house data typically adhere to standardized collection protocols, minimizing heterogeneity often encountered in multi-center datasets. Besides, we took into consideration potential confounding variables, which were selected primarily based on earlier related research, thus ensuring the dependability of our findings. Nonetheless, the current study possesses several limitations. Primarily, our research was constrained by its cross-sectional and retrospective approach, resulting in potential information bias due to incomplete data and presenting difficulties in determining a causal link between the BAR and SIRI and macrovascular complications. Specifically, the cross-sectional nature of the study precludes causal inference as the temporal sequence between biomarker levels and disease outcome cannot be determined. Hence, additional studies utilizing prospective methodologies are required to conclusively demonstrate the effectiveness of the BAR, SIRI as well as their combination. Secondly, given the predominantly Chinese-ancestry composition of our study population, caution should be exercised when extrapolating these results to other ethnic groups, highlighting the need for future multi-ethnic validation studies. Moreover, despite covariate adjustments, residual confounders may influence diabetes outcomes. Additionally, the potential for recall bias in self-reported medical history remains, which could influence the observed associations. Also, we admit that the improvement in AUC for our combined index was marginal, increasing only from 0.673 to 0.678. Such a slight increment raises the question of whether it offers meaningful clinical utility over individual markers. Although statistically significant, this small numerical advantage may not translate into substantial clinical benefit at the general population level. It remains to be determined whether this marginal improvement achieves relevant clinical utility in high-risk subgroups, such as elderly patients or those with a long duration of diabetes. Besides, the outcome of our study includes both atherosclerotic diseases and HF, which are pathophysiologically distinct and the relationship of BAR and SIRI with these different endpoints may not be uniform. Furthermore, the analysis of individual macrovascular complications such as HF and myocardial infarction should be considered exploratory due to the limited number of events and the resulting wide CI, and these findings require validation in larger, prospective cohorts. In addition, the potential for selection bias must be considered, as our population from a specialized tertiary center comprised predominantly of patients with more advanced or complex disease than the general T2DM population, which may limit the generalizability of our findings, particularly to individuals with milder disease managed in primary care settings. The specialized nature of our department also contributed to the cohort being predominantly comprised of patients with advanced atherosclerotic diseases and fewer HF cases, which limits the generalizability of our findings, especially regarding HF. Future studies with larger cohorts are essential to definitively confrim the role of BAR and SIRI across specific macrovascular complications. Conclusion As a cross-sectional study, we indicated that BAR, SIRI and the combined BAR-SIRI exhibit a notable positive association with the risk of diabetic macrovascular disease. Beyond traditional risk indicators, the combined BAR-SIRI may be worthy of consideration. The results offer fresh insights into the detection and prevention of macrovascular disease among individuals with T2DM. Supplementary Information Supplementary Material 1. Acknowledgements We express our gratitude to all the participants of Nanfang Hospital and to all individuals involved in establishing the Nanfang Hospital study. Author contributions CXJ, XYW, and GYX each made significant contributions to the conception and design of the study and are thus recognized as co-first authors. CXJ was responsible for performing statistical analyses and drafting the manuscript. XYW and GYX also made substantial contributions to the study design, data analysis, and interpretation. YQL, JJL, YNX, XMW, and JYH provided important intellectual input by critically revising the manuscript. MCZ oversaw the study conception and design, offered guidance on statistical analysis, and revised the study accordingly. XBW contributed to the development and planning of the study and gave approval for the published version of the manuscript. MCZ acts as the responsible party for this research, overseeing access to the study data and guaranteeing the accuracy and integrity of the analysis. Funding This study was supported by the National Natural Science Foundation of China (Grant No. 82170840). Data availability The relevant data are available from the corresponding author upon reasonable request. Declarations Ethics approval and patient consent statement Hospital study involves human participants and was designed and implemented in accordance with the Declaration of Helsinki (2013) and approved by the Ethics Committee of Nanfang Hospital (NCT06651983). As the waivers of informed consent have been requested and approved for Nanfang Hospital study, patients/participants were not required to provide their written informed consents. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Chen-Xi Jin, Xiao-Yan Wang and Guo-Yong Xie contributed equally to this work and therefore, are considered co-first authors. Contributor Information Xian-Bo Wu, Email: wuxb1010@smu.edu.cn. Meng-Chen Zou, Email: zoumc@smu.edu.cn. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement The relevant data are available from the corresponding author upon reasonable request.