Linear and non-linear relationships between red blood cell indices and cardiovascular risk factors: findings from the China National Health Survey Authors: Huijing He, Li Pan, Feng Liu, Xiaolan Ren, Ze Cui, Lize Pa, Dingming Wang, Jingbo Zhao, Hailing Wang, Xianghua Wang, Jianwei Du, Xia Peng, Guangliang Shan Published: 2024 ABSTRACT IntroductionCardiovascular diseases (CVDs) are a leading cause of mortality worldwide. Red blood cell indices (RBIs) are associated with CVD risk factors (CRFs) and easy to test, making them useful as a screening tool for early identification of individuals at high risk for CVDs.MethodsData from 31,781 participants in the China National Health Survey conducted from 2012 to 2017 were analyzed. Linear and non-linear relationships between RBIs and CRFs (hyperuricemia, diabetes, dyslipidemia) were assessed using restricted cubic splines. Propensity score weighting was used to balance confounders between RBI groups in the multivariable logistic regression models.ResultsHemoglobin concentration, red blood cell count and hematocrit all showed a significant linear dose-response association with all CRFs (p values < 0.001). Higher RBIs levels were associated with increased risk of hyperuricemia, diabetes, high LDL, high triglycerides, and high total cholesterol, but decreased HDL. For example, compared to the lowest quantile of HGB, the highest quantile had a 26% (13-40%) higher risk for hyperuricemia, a 43% (25-63%) higher risk of diabetes, 87% (61%-1.18 fold) higher risk of high LDL, and 68% (52-85%) higher risk of high triglycerides. Non-linear relationships were revealed between RBIs and most CRFs except uric acid and glucose. Sex differences were observed, with stronger associations between RBIs and hyperuricemia in women but stronger links with high LDL in men.ConclusionsElevated RBIs indicated higher risk of multiple CRFs. These findings suggest incorporating RBIs into CVD screening strategies to facilitate early prevention efforts, with consideration of sex differences.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12889-024-20988-7. Introduction Cardiovascular diseases (CVDs) are the leading cause of mortality and morbidity worldwide. According to the Global Disease Burden study, the number of deaths attributed to CVDs has risen steadily from 12.1 million in 1990 to 18.6 million in 2019 [1]. Early identification of individuals at high risk for CVD can play a crucial role in preventing complications. Diabetes and dyslipidemia are well-established risk factors for cardiovascular issues [2]. In particular, high cholesterol and hypertriglyceridemia are significant risk factors for diabetic patients suffering from cardiovascular disease [3]. Additionally, hyperuricemia (HUA) is recognized as an important risk factor for both CVD and mortality and is closely linked to diabetes and dyslipidemia. Previous studies have indicated a high prevalence of hyperuricemia among the Chinese population [4, 5]. Therefore, it is essential to take elevated serum uric acid (SUA) into account when implementing CVD interventions and identifying high-risk populations. Utilizing readily available, low-cost data from a significant proportion of individuals to assess and predict cardiovascular risk can enhance implementation within primary healthcare settings. Red blood cell indices (RBIs) can be obtained through a straightforward complete blood count (CBC) test, which measures both the quantity and distributional characteristics of red blood cells, alongside other indices commonly utilized in screening environments [6]. Prior research has established connections between RBIs and various health outcomes, including all-cause mortality, CVDs, and metabolic disorders [7–9]. It has been suggested that increased blood viscosity, associated with elevated hematocrit levels, may contribute to a heightened risk of CVDs [10]. Previous studies have indicated that elevated levels of RBIs—such as hemoglobin concentration (HGB) and hematocrit (HCT)—are linked to an increased risk of CVDs and their associated risk factors [11]. However, the nature of these associations has not always been consistent, with some research suggesting that lower HGB levels or anemia could also elevate CVD risk [12–14]. Furthermore, these relationships may exhibit a J-shaped or U-shaped pattern [15, 16]. These findings suggest the presence of potential non-linear relationships between the independent and dependent variables, yet the precise nature of these non-linear associations remains unclear, necessitating further investigation. In this study, we aimed to explore both the linear and non-linear relationships between RBIs and CVD risk factors (CRFs), as well as the impact of RBIs on various CRFs. The data for this analysis was sourced from the China National Health Survey (CNHS), a comprehensive nationwide survey conducted by the Institute of Basic Medical Sciences at the Chinese Academy of Medical Sciences between 2012 and 2017. Materials and methods Study population The data for this study were obtained from the CNHS. Detailed descriptions of the CNHS protocol can be found in our previous work [17]. In brief, a stratified, multistage cluster sampling method was employed to select eleven provinces based on geographic regions, urbanization levels, and economic development status. Inclusion criteria required participants to be individuals aged 20 years or older who had resided in the local area for at least one year. Individuals with severe mental or physical illnesses, pregnant women, or active military personnel were excluded from the study. While the CNHS encompasses a diverse range of ethnic populations (over 30%), this analysis focused exclusively on Han Chinese individuals due to their predominance in the overall Chinese population. The study adhered to the principles outlined in the Declaration of Helsinki, and ethical approval was granted by the Bioethical Committee of the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences (No. 029–2013). All participants provided written informed consent prior to their involvement in the survey. Measurements A standardized questionnaire was developed specifically for conducting face-to-face personal interviews. To ensure the interviewers’ proficiency with the necessary tools and methods, all staff participated in a comprehensive training program. Demographic and socioeconomic data, along with lifestyle factors such as tobacco use and alcohol consumption, were systematically collected. Additionally, information regarding participants’ medical and family histories related to diabetes was obtained. Participants were measured barefoot and dressed in light clothing during anthropometric assessments. Height was recorded to the nearest 0.1 cm using a fixed stadiometer, while weight was measured with a body composition analyzer (TANITA BC-420, Japan), ensuring decimal-level accuracy [18]. Following an overnight fast of at least 8 h, a fasting blood sample of 9 ml was collected from each participant. This comprised two EDTA vacutainers with 2 ml of blood each and one non-EDTA vacutainer containing 5 ml of blood [17]. The serum levels of uric acid (SUA), fasting plasma glucose (FPG), and serum lipids were assessed using the ROCHE Cobas8000C701 Chemistry Analyzer (USA). The serum lipid profile included total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C). Hemoglobin (HGB), red blood cell count (RBC), and hematocrit (HCT) were analyzed using the automated hematology analyzer (XT-4000i, Japan). Definitions The definitions of cigarette smoking, alcohol consumption, and physical activity have been previously outlined [17]. Body mass index (BMI) was calculated by dividing weight in kilograms by the square of height in meters (kg/m²). HUA was defined as a SUA level exceeding 360 µmol/L (approximately 6 mg/dL) in women and 420 µmol/L (approximately 7 mg/dL) in men, in accordance with prior studies [19, 20]. Abnormal serum lipid levels were determined based on the 2016 Chinese guidelines for the management of dyslipidemia in adults [21]. Diabetes was identified as either a prior diagnosis of the condition or a FPG level greater than 7.0 mmol/L. Statistical analyses Our analyses focused on adults who had complete data on red blood cell indices, including HGB, RBC, and HCT. This resulted in a final analytical sample of 31,781 individuals. The study’s design and flow chart are presented in Supplemental Fig. S1. In the descriptive analysis, continuous variables were reported as mean (standard deviation) or median (interquartile range, IQR) when they did not follow a normal distribution. Categorical variables were expressed as number (percentage). The Wilcoxon rank sum test and Chi-square test were utilized to compare the basic characteristics between sexes. We employed the R package “rms” to fit restricted cubic spline (RCS) models to explore both linear and non-linear relationships between RBIs and cardiometabolic traits (SUA, serum lipids, and FPG). Non-linearity was assessed using the likelihood ratio test, comparing a model with only the linear term against one that included both the linear and cubic spline terms [22]. To balance the distribution of potential confounding factors across different RBI quantity groups (Q1, Q2, Q3, and Q4), we applied propensity score weighting (PSW) using the R package “twang.” The factors included in the PSW calculation were age, sex, study sites, urban or rural residence, smoking status, alcohol consumption, BMI, and, for diabetes estimation, family history of diabetes. We used general linear regression models (GLMs) incorporating the calculated propensity scores as weighting factors and additionally adjusted for BMI and study sites to assess the associations between HGB, RBC, HCT, and cardiometabolic traits. Adjusting for BMI and study sites in the outcome models was done to enhance the precision of the estimates, as they, while balanced by PSW, remained independent associated factors of the outcomes. The propensity scores were also used as weighting factors in logistic regression (LR) models, conducted with the survey-logistic procedure in SAS. Forest plots, adjusted for BMI and study sites, were generated based on the LR model results. In both the GLMs and LR models, HGB, RBC, and HCT were categorized according to sex-specific quantiles, with the lowest values (Q1 group) designated as the reference group. Additionally, we performed the regressions separately by sex to further examine the sex-specific associations between RBIs and CRFs. In both the GLMs and LR models for women, menopause status was adjusted for in the association analysis. A sensitivity analysis was also performed using traditional LR models to evaluate the associations between RBIs and CRFs. The data were analyzed using SAS (version 9.4, SAS Institute, Cary, NC) and R software (version 4.0.4). We did not perform the Bonferroni test, and a two-sided p-value less than 0.05 was considered statistically significant. Results Basic characteristics of the study population Basic demographic and clinical characteristics are summarized in Table 1 and compared by sex. The mean age of the 31,781 individuals was 48.76 ± 13.34 years, with men comprising 40.1% (12,740) of the overall population. Significant differences were observed between the sexes regarding demographic and socioeconomic factors, cardiometabolic traits, and red blood cell indices (HGB, RBC, and HCT). Generally, men exhibited higher BMI, SUA, FPG, and serum lipid levels (with the exception of HDL), as well as elevated HGB, RBC, and HCT values (all p-values < 0.05). Additionally, men had a higher prevalence of HUA, diabetes, and dyslipidemia compared to women. Table 1Basic characteristics of the study population from the China National Health Survey, 2012–2017Men (n = 12740)Women (n = 19041)Overall (n = 31781)PAge49.2113.5448.4613.2048.7613.34< 0.00120-129610.17201910.60331510.43< 0.00130-202915.93310416.30513316.1540-321725.25509726.77831426.1650-317924.95475424.97793324.9660-219417.22304816.01524216.4970–808256.4810195.3518445.80Residence area Urban797262.5712,33264.7720,30463.89< 0.001 Rural475437.32668035.0811,43435.98Education Illiterate/elementary school232118.22576930.30809025.46< 0.001 High school652551.22874245.9115,26748.04 College or above386630.35448723.56835326.28Ever smoke875668.734702.47922629.03< 0.001Ever drink alcohol951774.70334817.5812,86540.48< 0.001BMI24.413.5223.683.5723.973.57< 0.001Serum uric acid360.9110.927285.6303.2113.9< 0.001FPG5.160.935.060.795.10.84< 0.001TC4.691.294.741.374.721.33< 0.001TG1.531.291.290.971.371.1< 0.001LDL2.741.082.761.122.751.110.017HDL1.20.411.40.441.320.46< 0.001Hyperuricemia313724.62217911.44531616.73< 0.001Diabetes129910.2012346.4825337.97< 0.001Dyslipidemia574545.09541928.4611,16435.13< 0.001HGB161.8712.52139.6511.76148.5916.26< 0.001RBC5.370.494.790.415.020.53< 0.001HCT49.284.1243.473.7045.804.81< 0.001BMI body mass index, FPG fasting plasma glucose, TC total cholesterol, TG triglycerides, LDL low-density lipoprotein cholesterol, HDL high-density lipoprotein cholesterol, HGB hemoglobin concentration, RBC red blood cell count, HCT hematocrit Table 1Basic characteristics of the study population from the China National Health Survey, 2012–2017Men (n = 12740)Women (n = 19041)Overall (n = 31781)PAge49.2113.5448.4613.2048.7613.34< 0.00120-129610.17201910.60331510.43< 0.00130-202915.93310416.30513316.1540-321725.25509726.77831426.1650-317924.95475424.97793324.9660-219417.22304816.01524216.4970–808256.4810195.3518445.80Residence area Urban797262.5712,33264.7720,30463.89< 0.001 Rural475437.32668035.0811,43435.98Education Illiterate/elementary school232118.22576930.30809025.46< 0.001 High school652551.22874245.9115,26748.04 College or above386630.35448723.56835326.28Ever smoke875668.734702.47922629.03< 0.001Ever drink alcohol951774.70334817.5812,86540.48< 0.001BMI24.413.5223.683.5723.973.57< 0.001Serum uric acid360.9110.927285.6303.2113.9< 0.001FPG5.160.935.060.795.10.84< 0.001TC4.691.294.741.374.721.33< 0.001TG1.531.291.290.971.371.1< 0.001LDL2.741.082.761.122.751.110.017HDL1.20.411.40.441.320.46< 0.001Hyperuricemia313724.62217911.44531616.73< 0.001Diabetes129910.2012346.4825337.97< 0.001Dyslipidemia574545.09541928.4611,16435.13< 0.001HGB161.8712.52139.6511.76148.5916.26< 0.001RBC5.370.494.790.415.020.53< 0.001HCT49.284.1243.473.7045.804.81< 0.001BMI body mass index, FPG fasting plasma glucose, TC total cholesterol, TG triglycerides, LDL low-density lipoprotein cholesterol, HDL high-density lipoprotein cholesterol, HGB hemoglobin concentration, RBC red blood cell count, HCT hematocrit Basic characteristics of the study population from the China National Health Survey, 2012–2017 BMI body mass index, FPG fasting plasma glucose, TC total cholesterol, TG triglycerides, LDL low-density lipoprotein cholesterol, HDL high-density lipoprotein cholesterol, HGB hemoglobin concentration, RBC red blood cell count, HCT hematocrit The linear and non-linear relationships between red blood cell indexes and cardiometabolic traits The RCS curves, illustrating the linear and non-linear relationships between independent and dependent variables, are presented in Supplementary Fig. S2 and Fig. 1. Supplementary Fig.S2 demonstrates the dose-response relationships between RBIs and cardiometabolic traits. As shown in the figure, HGB, RBC, and HCT exhibited significant linear dose-response associations with all cardiometabolic traits (p-values < 0.001). With the exception of SUA and FPG, the remaining cardiometabolic traits displayed non-linear associations with RBIs. Overall, increases in HGB, RBC, and HCT were associated with higher levels of SUA, FPG, LDL, TC, and TG. The relationships between RBIs and HDL followed an inverted U-shaped pattern. Fig. 1The linear and non-linear relationship between hemoglobin concentration, red blood cell count, hematocrit and multiple cardiovascular risk factors. HGB: hemoglobin concentration; RBC: red blood cell count; HCT: red blood cell count; HUA: hyperuricemia; TC: total cholesterol; TG: triglycerides; LDL: low-density lipoprotein cholesterol; HDL: high-density lipoprotein cholesterol. A the relationship between red blood cell indices (RBIs) and HUA; B the relationship between RBIs and diabetes; C the relationship between RBIs and high LDL; D the relationship between RBIs and low HDL; E the relationship between RBIs and high TG; F the relationship between RBIs and high TC The linear and non-linear relationship between hemoglobin concentration, red blood cell count, hematocrit and multiple cardiovascular risk factors. HGB: hemoglobin concentration; RBC: red blood cell count; HCT: red blood cell count; HUA: hyperuricemia; TC: total cholesterol; TG: triglycerides; LDL: low-density lipoprotein cholesterol; HDL: high-density lipoprotein cholesterol. A the relationship between red blood cell indices (RBIs) and HUA; B the relationship between RBIs and diabetes; C the relationship between RBIs and high LDL; D the relationship between RBIs and low HDL; E the relationship between RBIs and high TG; F the relationship between RBIs and high TC Figure 1 illustrates both linear and non-linear relationships between RBIs and CRFs, including HUA, diabetes, and abnormal serum lipids. HGB demonstrated a linear association with all CRFs (p-values < 0.001) and a non-linear association with diabetes, low HDL, and high TG. Similarly, HGB exhibited linear associations with all CRFs and non-linear associations with HUA and abnormal serum lipids. However, no association was found between HCT and diabetes (Fig. 1-B; p-values for both linear and non-linear associations > 0.1). For the other CRFs, HCT exhibited associations similar to those of the other two RBIs. Associations between red blood cell indexes and cardiometabolic traits Table 2 showed the association between the RBIs and cardiometabolic traits, and Fig. 2 illustrates the correlations between RBIs and CRFs in the overall population. HGB, RBC, and HCT were positively associated with an increased risk of HUA, diabetes, LDL, high TG, and high TC, while showing a negative association with low HDL. For instance, individuals in the highest quantile (Q4) of HGB had a 26% (13–40%) higher risk of HUA, a 43% (25–63%) higher risk of diabetes, an 87% (61%−1.18-fold) higher risk of high LDL, a 68% (52–85%) higher risk of high TG, and a 95% (70%−1.24-fold) higher risk of high TC compared to those in the lowest quantile (Q1). The exact odds ratios (ORs) and their 95% confidence intervals (CIs) for these associations are presented in Supplementary Table S1. Table 2The association between hemoglobin concentration, hematocrit, red blood cell and cardiometabolic traitsHGB (B and 95% CI)RBC (B and 95% CI)HCT (B and 95% CI)Q2Q3Q4Q2Q3Q4Q2Q3Q4SUA3.15 (0.98, 5.33)9.56 (7.39, 11.73)13.77 (11.58, 15.96)2.92 (0.74, 5.09)6.09 (3.92, 8.27)13.61 (11.42, 15.80)6.26 (4.08, 8.44)9.09 (6.92, 11.27)14.62 (12.42, 16.83)P0.005< 0.001< 0.0010.009< 0.001< 0.001< 0.001< 0.001< 0.001FPGa0.01 (−0.03, 0.04)0.06 (0.03, 0.10)0.17 (0.14, 0.20)0.03 (0.000, 0.07)0.08 (0.04, 0.11)0.14 (0.11, 0.18)0.04 (0.002, 0.07)0.07 (0.03, 0.10)0.12 (0.08, 0.15)P0.609< 0.001< 0.0010.050< 0.001< 0.0010.037< 0.001< 0.001LDL0.12 (0.10, 0.15)0.18 (0.16, 0.21)0.23 (0.20, 0.25)0.11 (0.09, 0.14)0.15 (0.13, 0.18)0.20 (0.18, 0.23)0.13 (0.11, 0.16)0.22 (0.19, 0.24)0.31 (0.29, 0.34)P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001HDL0.02 (0.01, 0.03)0.01 (0.01, 0.02)0.01 (0.002, 0.02)0.002 (−0.01, 0.01)−0.01 (−0.02, −0.001)−0.02 (−0.03, −0.01)0.02 (0.01, 0.03)0.03 (0.02, 0.04)0.03 (0.02, 0.04)P< 0.0010.0020.0150.6210.030< 0.001< 0.001< 0.001< 0.001TG0.09 (0.05, 0.12)0.17 (0.13, 0.21)0.28 (0.24, 0.32)0.05 (0.01, 0.09)0.13 (0.10, 0.17)0.19 (0.15, 0.23)0.12 (0.08, 0.16)0.18 (0.14, 0.22)0.26 (0.23, 0.30)P< 0.001< 0.001< 0.0010.009< 0.001< 0.001< 0.001< 0.001< 0.001TC0.17 (0.14, 0.20)0.26 (0.23, 0.29)0.34 (0.31, 0.37)0.12 (0.09, 0.15)0.19 (0.16, 0.22)0.24 (0.22, 0.27)0.19 (0.16, 0.22)0.30 (0.27, 0.33)0.43 (0.41, 0.46)P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001Men SUA−0.59 (−4.48, 3.30)3.89 (0.01, 7.77)6.85 (2.91, 10.78)0.28 (−3.62, 4.17)0.93 (−2.96, 4.83)7.28 (3.36, 11.20)1.53 (−2.37, 5.43)3.08 (−0.83, 7.00)5.42 (1.45, 9.40) P0.7650.050< 0.0010.8900.638< 0.0010.4420.1220.008 FPGa−0.03 (−0.09, 0.03)0.02 (−0.04, 0.08)0.08 (0.01, 0.14)0.03 (−0.04, 0.09)0.01 (−0.05, 0.08)0.06 (−0.01, 0.12)−0.03 (−0.09, 0.04)−0.05 (−0.11, 0.02)−0.02 (−0.08, 0.05) P0.3190.4710.0150.4120.6450.0760.4330.1370.576 LDL0.12 (0.08, 0.15)0.16 (0.12, 0.19)0.20 (0.16, 0.24)0.11 (0.07, 0.15)0.14 (0.10, 0.17)0.18 (0.15, 0.22)0.13 (0.09, 0.17)0.21 (0.17, 0.25)0.28 (0.25, 0.32) P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001 HDL0.01 (0.000, 0.03)0.01 (−0.001, 0.03)−0.01 (−0.02, 0.01)0.01 (−0.005, 0.02)−0.004 (−0.02, 0.01)−0.04 (−0.05, −0.02)0.01 (−0.004. 0.02)0.02 (0.01, 0.04)0.01 (−0.01, 0.02) P0.0450.0770.2170.1860.597< 0.0010.1700.0010.387 TG0.06 (−0.01, 0.13)0.18 (0.11, 0.25)0.29 (0.21, 0.36)0.04 (−0.03, 0.11)0.10 (0.03, 0.17)0.18 (0.10, 0.25)0.11 (0.04, 0.18)0.14 (0.07, 0.21)0.27 (0.19, 0.34) P0.121< 0.001< 0.0010.2410.007< 0.0010.003< 0.001< 0.001 TC0.14 (0.10, 0.18)0.22 (0.17, 0.26)0.27 (0.23, 0.32)0.11 (0.07, 0.16)0.15 (0.11, 0.20)0.19 (0.15, 0.24)0.17 (0.12, 0.21)0.26 (0.21, 0.30)0.36 (0.32, 0.41) P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001Women SUA5.22 (2.79, 7.66)11.89 (9.46, 14.33)15.96 (13.51, 18.40)4.89 (2.46, 7.32)8.74 (6.31, 11.18)16.64 (14.19, 19.10)8.07 (5.64, 10.50)10.99 (8.56, 13.42)16.90 (14.44, 19.36) P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001 FPGa0.002 (−0.04, 0.04)0.05 (0.01, 0.09)0.17 (0.13, 0.21)0.03 (−0.01, 0.07)0.10 (0.06, 0.14)0.18 (0.14, 0.22)0.03 (−0.01, 0.07)0.08 (0.04, 0.12)0.13 (0.09, 0.17) P0.9090.007< 0.0010.178< 0.001< 0.0010.167< 0.001< 0.001 LDL0.09 (0.06, 0.12)0.16 (0.13, 0.19)0.21 (0.18, 0.24)0.09 (0.06, 0.12)0.14 (0.11, 0.17)0.19 (0.16, 0.22)0.09 (0.06, 0.12)0.17 (0.14, 0.20)0.27 (0.24, 0.30) P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001 HDL0.02 (0.003, 0.03)0.02 (0.003, 0.03)0.03 (0.01, 0.04)−0.003 (−0.02, 0.01)−0.01 (−0.03, −0.002)−0.01 (−0.03, 0.004)0.03 (0.01, 0.04)0.03 (0.02, 0.05)0.05 (0.04, 0.06) P0.0150.017< 0.0010.6130.0240.158< 0.001< 0.001< 0.001 TG0.07 (0.03, 0.11)0.12 (0.08, 0.16)0.22 (0.18, 0.26)0.04 (0.003, 0.09)0.13 (0.09, 0.17)0.18 (0.13, 0.22)0.08 (0.04, 0.12)0.15 (0.11, 0.19)0.19 (0.14, 0.23) P< 0.001< 0.001< 0.0010.036< 0.001< 0.001< 0.001< 0.001< 0.001 TC0.15 (0.11, 0.18)0.23 (0.19, 0.26)0.32 (0.29, 0.36)0.11 (0.07, 0.14)0.18 (0.15, 0.22)0.24 (0.20, 0.28)0.15 (0.11, 0.18)0.26 (0.22, 0.29)0.39 (0.35, 0.43) P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001BMI body mass index, SUA serum uric acids, FPG fasting plasma glucose, TC total cholesterol, TG triglycerides, LDL low-density lipoprotein cholesterol, HDL high-density lipoprotein cholesterol, HGB hemoglobin concentration, RBC red blood cell count, HCT hematocrit, Q1-Q4 the first to fourth quantile of the variable, B regression coefficient, CI confidence intervalThe general linear regression models considered propensity score as a weighted factor, and the regression models were adjusted for sex (only in the overall population), BMI and study sites. In women, menopause status was further adjusted. The Q1 group of the red blood cell indices were set as the reference group in the comparisonaFurther adjusted for family history of diabetes Table 2The association between hemoglobin concentration, hematocrit, red blood cell and cardiometabolic traitsHGB (B and 95% CI)RBC (B and 95% CI)HCT (B and 95% CI)Q2Q3Q4Q2Q3Q4Q2Q3Q4SUA3.15 (0.98, 5.33)9.56 (7.39, 11.73)13.77 (11.58, 15.96)2.92 (0.74, 5.09)6.09 (3.92, 8.27)13.61 (11.42, 15.80)6.26 (4.08, 8.44)9.09 (6.92, 11.27)14.62 (12.42, 16.83)P0.005< 0.001< 0.0010.009< 0.001< 0.001< 0.001< 0.001< 0.001FPGa0.01 (−0.03, 0.04)0.06 (0.03, 0.10)0.17 (0.14, 0.20)0.03 (0.000, 0.07)0.08 (0.04, 0.11)0.14 (0.11, 0.18)0.04 (0.002, 0.07)0.07 (0.03, 0.10)0.12 (0.08, 0.15)P0.609< 0.001< 0.0010.050< 0.001< 0.0010.037< 0.001< 0.001LDL0.12 (0.10, 0.15)0.18 (0.16, 0.21)0.23 (0.20, 0.25)0.11 (0.09, 0.14)0.15 (0.13, 0.18)0.20 (0.18, 0.23)0.13 (0.11, 0.16)0.22 (0.19, 0.24)0.31 (0.29, 0.34)P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001HDL0.02 (0.01, 0.03)0.01 (0.01, 0.02)0.01 (0.002, 0.02)0.002 (−0.01, 0.01)−0.01 (−0.02, −0.001)−0.02 (−0.03, −0.01)0.02 (0.01, 0.03)0.03 (0.02, 0.04)0.03 (0.02, 0.04)P< 0.0010.0020.0150.6210.030< 0.001< 0.001< 0.001< 0.001TG0.09 (0.05, 0.12)0.17 (0.13, 0.21)0.28 (0.24, 0.32)0.05 (0.01, 0.09)0.13 (0.10, 0.17)0.19 (0.15, 0.23)0.12 (0.08, 0.16)0.18 (0.14, 0.22)0.26 (0.23, 0.30)P< 0.001< 0.001< 0.0010.009< 0.001< 0.001< 0.001< 0.001< 0.001TC0.17 (0.14, 0.20)0.26 (0.23, 0.29)0.34 (0.31, 0.37)0.12 (0.09, 0.15)0.19 (0.16, 0.22)0.24 (0.22, 0.27)0.19 (0.16, 0.22)0.30 (0.27, 0.33)0.43 (0.41, 0.46)P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001Men SUA−0.59 (−4.48, 3.30)3.89 (0.01, 7.77)6.85 (2.91, 10.78)0.28 (−3.62, 4.17)0.93 (−2.96, 4.83)7.28 (3.36, 11.20)1.53 (−2.37, 5.43)3.08 (−0.83, 7.00)5.42 (1.45, 9.40) P0.7650.050< 0.0010.8900.638< 0.0010.4420.1220.008 FPGa−0.03 (−0.09, 0.03)0.02 (−0.04, 0.08)0.08 (0.01, 0.14)0.03 (−0.04, 0.09)0.01 (−0.05, 0.08)0.06 (−0.01, 0.12)−0.03 (−0.09, 0.04)−0.05 (−0.11, 0.02)−0.02 (−0.08, 0.05) P0.3190.4710.0150.4120.6450.0760.4330.1370.576 LDL0.12 (0.08, 0.15)0.16 (0.12, 0.19)0.20 (0.16, 0.24)0.11 (0.07, 0.15)0.14 (0.10, 0.17)0.18 (0.15, 0.22)0.13 (0.09, 0.17)0.21 (0.17, 0.25)0.28 (0.25, 0.32) P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001 HDL0.01 (0.000, 0.03)0.01 (−0.001, 0.03)−0.01 (−0.02, 0.01)0.01 (−0.005, 0.02)−0.004 (−0.02, 0.01)−0.04 (−0.05, −0.02)0.01 (−0.004. 0.02)0.02 (0.01, 0.04)0.01 (−0.01, 0.02) P0.0450.0770.2170.1860.597< 0.0010.1700.0010.387 TG0.06 (−0.01, 0.13)0.18 (0.11, 0.25)0.29 (0.21, 0.36)0.04 (−0.03, 0.11)0.10 (0.03, 0.17)0.18 (0.10, 0.25)0.11 (0.04, 0.18)0.14 (0.07, 0.21)0.27 (0.19, 0.34) P0.121< 0.001< 0.0010.2410.007< 0.0010.003< 0.001< 0.001 TC0.14 (0.10, 0.18)0.22 (0.17, 0.26)0.27 (0.23, 0.32)0.11 (0.07, 0.16)0.15 (0.11, 0.20)0.19 (0.15, 0.24)0.17 (0.12, 0.21)0.26 (0.21, 0.30)0.36 (0.32, 0.41) P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001Women SUA5.22 (2.79, 7.66)11.89 (9.46, 14.33)15.96 (13.51, 18.40)4.89 (2.46, 7.32)8.74 (6.31, 11.18)16.64 (14.19, 19.10)8.07 (5.64, 10.50)10.99 (8.56, 13.42)16.90 (14.44, 19.36) P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001 FPGa0.002 (−0.04, 0.04)0.05 (0.01, 0.09)0.17 (0.13, 0.21)0.03 (−0.01, 0.07)0.10 (0.06, 0.14)0.18 (0.14, 0.22)0.03 (−0.01, 0.07)0.08 (0.04, 0.12)0.13 (0.09, 0.17) P0.9090.007< 0.0010.178< 0.001< 0.0010.167< 0.001< 0.001 LDL0.09 (0.06, 0.12)0.16 (0.13, 0.19)0.21 (0.18, 0.24)0.09 (0.06, 0.12)0.14 (0.11, 0.17)0.19 (0.16, 0.22)0.09 (0.06, 0.12)0.17 (0.14, 0.20)0.27 (0.24, 0.30) P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001 HDL0.02 (0.003, 0.03)0.02 (0.003, 0.03)0.03 (0.01, 0.04)−0.003 (−0.02, 0.01)−0.01 (−0.03, −0.002)−0.01 (−0.03, 0.004)0.03 (0.01, 0.04)0.03 (0.02, 0.05)0.05 (0.04, 0.06) P0.0150.017< 0.0010.6130.0240.158< 0.001< 0.001< 0.001 TG0.07 (0.03, 0.11)0.12 (0.08, 0.16)0.22 (0.18, 0.26)0.04 (0.003, 0.09)0.13 (0.09, 0.17)0.18 (0.13, 0.22)0.08 (0.04, 0.12)0.15 (0.11, 0.19)0.19 (0.14, 0.23) P< 0.001< 0.001< 0.0010.036< 0.001< 0.001< 0.001< 0.001< 0.001 TC0.15 (0.11, 0.18)0.23 (0.19, 0.26)0.32 (0.29, 0.36)0.11 (0.07, 0.14)0.18 (0.15, 0.22)0.24 (0.20, 0.28)0.15 (0.11, 0.18)0.26 (0.22, 0.29)0.39 (0.35, 0.43) P< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001< 0.001BMI body mass index, SUA serum uric acids, FPG fasting plasma glucose, TC total cholesterol, TG triglycerides, LDL low-density lipoprotein cholesterol, HDL high-density lipoprotein cholesterol, HGB hemoglobin concentration, RBC red blood cell count, HCT hematocrit, Q1-Q4 the first to fourth quantile of the variable, B regression coefficient, CI confidence intervalThe general linear regression models considered propensity score as a weighted factor, and the regression models were adjusted for sex (only in the overall population), BMI and study sites. In women, menopause status was further adjusted. The Q1 group of the red blood cell indices were set as the reference group in the comparisonaFurther adjusted for family history of diabetes The association between hemoglobin concentration, hematocrit, red blood cell and cardiometabolic traits BMI body mass index, SUA serum uric acids, FPG fasting plasma glucose, TC total cholesterol, TG triglycerides, LDL low-density lipoprotein cholesterol, HDL high-density lipoprotein cholesterol, HGB hemoglobin concentration, RBC red blood cell count, HCT hematocrit, Q1-Q4 the first to fourth quantile of the variable, B regression coefficient, CI confidence interval The general linear regression models considered propensity score as a weighted factor, and the regression models were adjusted for sex (only in the overall population), BMI and study sites. In women, menopause status was further adjusted. The Q1 group of the red blood cell indices were set as the reference group in the comparison aFurther adjusted for family history of diabetes Fig. 2The associations between hemoglobin concentration (HGB), red blood cell count (RBC), hematocrit (HCT), and cardiovascular risk factors demonstrated by forest plots in the study population. HUA: hyperuricemia; TC: total cholesterol; TG: triglycerides; LDL: low-density lipoprotein cholesterol; HDL: high-density lipoprotein cholesterol. A the associations between red blood cell indices (RBIs) and HUA; B the associations between RBIs and diabetes; C the associations between RBIs and high LDL; D the associations between RBIs and low HDL; E the associations between RBIs and high TG; F the associations between RBIs and high TC The associations between hemoglobin concentration (HGB), red blood cell count (RBC), hematocrit (HCT), and cardiovascular risk factors demonstrated by forest plots in the study population. HUA: hyperuricemia; TC: total cholesterol; TG: triglycerides; LDL: low-density lipoprotein cholesterol; HDL: high-density lipoprotein cholesterol. A the associations between red blood cell indices (RBIs) and HUA; B the associations between RBIs and diabetes; C the associations between RBIs and high LDL; D the associations between RBIs and low HDL; E the associations between RBIs and high TG; F the associations between RBIs and high TC As shown in Fig. 2, the associations between RBIs and abnormal LDL and TC were stronger than those observed for other CRFs. Compared to the Q1 group of HCT, the Q4 group demonstrated an OR (95% CI) of 2.18 (1.84–2.57) for high TC and 2.19 (1.84–2.60) for high LDL. Conversely, the associations between RBIs and diabetes were relatively weak, with only marginal effects observed. Figure 3 illustrates the sex-specific associations between RBIs and CRFs. Differences were observed between males and females regarding the associations between RBIs and CRFs. In women, RBIs had a greater impact on HUA and low HDL levels compared to men. Higher levels of RBIs were negatively associated with low HDL in women, while no such association was identified in men. Similarly, RBIs were positively associated with HUA in women, but no corresponding associations were found in men. Further analysis revealed that RBIs demonstrated linear associations with SUA and HUA in women; however, significant linear or non-linear associations were not observed in men (as shown in Supplementary Fig. S3 and Fig. S4). In contrast, the associations between RBIs and high LDL were stronger in men than in women, with these associations being more pronounced than those of other CRFs. The exact ORs and their 95% CIs are detailed in Supplementary Table S1. Fig. 3The associations between hemoglobin concentration (HGB), red blood cell count (RBC), hematocrit (HCT), and cardiovascular risk factors demonstrated by forest plots, stratified by sex. Males are represented by the blue areas, while females are represented by the yellow areas. RBIs: red blood cell indices. HUA: hyperuricemia; TC: total cholesterol; TG: triglycerides; LDL: low-density lipoprotein cholesterol; HDL: high-density lipoprotein cholesterol. A the associations between red blood cell indices (RBIs) and HUA; B the associations between RBIs and diabetes; C the associations between RBIs and high LDL; D the associations between RBIs and low HDL; E the associations between RBIs and high TG; F the associations between RBIs and high TC The associations between hemoglobin concentration (HGB), red blood cell count (RBC), hematocrit (HCT), and cardiovascular risk factors demonstrated by forest plots, stratified by sex. Males are represented by the blue areas, while females are represented by the yellow areas. RBIs: red blood cell indices. HUA: hyperuricemia; TC: total cholesterol; TG: triglycerides; LDL: low-density lipoprotein cholesterol; HDL: high-density lipoprotein cholesterol. A the associations between red blood cell indices (RBIs) and HUA; B the associations between RBIs and diabetes; C the associations between RBIs and high LDL; D the associations between RBIs and low HDL; E the associations between RBIs and high TG; F the associations between RBIs and high TC The results of the sensitivity analysis are presented in Supplementary Table S2. The main findings from the sensitivity analysis were consistent with those of the primary analysis, indicating the robustness and reliability of the data. Discussion Using data from a large sample from China, we have identified a connection between elevated red blood indices and various cardiovascular risk factors such as hyperuricemia, abnormal serum lipids, and diabetes. Our analysis also revealed non-linear relationships between RBIs and cardiometabolic traits. Furthermore, we found notable sex differences in these associations. These findings suggest that RBIs are significant indicators for CRFs and could be utilized in developing screening strategies within community healthcare settings to facilitate the early identification of high-risk individuals for CVDs. Several studies have explored the associations between HCT and CRFs or CVD. For instance, the Framingham study identified a J- or U-shaped relationship between HCT and CVD events [23]. A study conducted within the Japanese population revealed that both elevated and decreased HCT levels were linked to an increased risk of CVD [9]. Similarly, a study involving 1884 middle-aged Chinese individuals demonstrated a positive association between HCT levels and ischemic cardiovascular diseases [11]. Truslow et al. found that hematologic indices offered better discrimination and calibration in predicting cardiovascular risk [6]. In the Swedish and Danish blood donor population, the risk of myocardial infarction and ischemic stroke rose with higher hemoglobin concentration [8]. Paquette et al. investigated the effect of RBIs on major adverse cardiovascular events (ASCVD) among 482 patients with familial hypercholesterolemia and found that elevated HGB and RBC levels were associated with an increased risk of ASCVD [7]. These studies provide valuable insights into the relationship between RBIs and cardiometabolic disorders. Our study revealed both linear and non-linear relationships between RBIs and CRFs, contributing the most recent evidence on these associations. Given that RBIs are easily measurable, the presence of non-linear relationships suggests that more sophisticated methodologies should be employed when developing risk prediction models based on these indices. Additionally, the variability in associations indicates that maintaining RBIs within the normal range is essential for reducing cardiometabolic risk in the general population. Controlling for confounding remains a crucial concern in observational studies. Among the various methods available, propensity score weighting (PSW) is frequently employed in non-randomized studies for confounding control and effect assessment [24]. PSW serves as a balancing score to enable comparisons between treatment groups—in this study, different RBI groups—assuming that the distributions of covariates across these groups are similar. The use of PSW mimics some aspects of a randomized trial study design and allows for an appropriate estimation of causal effects [25, 26]. In our study, the PSW method was implemented to balance the distribution of potential confounders among the different RBI groups, leading to a more reliable assessment of effects. Consistent with prior research, our findings indicate that higher levels of RBIs are associated with an increased risk of CRFs, such as abnormal serum lipids, hyperuricemia, and diabetes. Furthermore, our study contributes new evidence regarding the nonlinear relationships between RBIs and CRFs in the general population. Specifically, RBIs primarily exhibit linear associations with SUA and HUA, while non-linear relationships are observed with other CRFs. Additionally, HGB, RBC, and HCT show varying effect sizes in relation to CRFs, with HGB and HCT demonstrating larger effect sizes than RBC. In clinical practice, simple blood cell tests can serve as a cost-effective method to help identify high-risk populations for CVDs, alongside other modifiable risk factors such as alcohol consumption, cigarette smoking, and excess adiposity considered in this study. Few studies have explored the relationship between RBIs and abnormal SUA levels. HUA is a leading cause of gout, and a significant risk factor for various metabolic disorders, chronic kidney disease, and CVDs [27, 28]. The REACTION study, which focused on individuals living in rural China, identified a positive correlation between HGB and SUA levels. Furthermore, elevated HGB was associated with an increased risk of HUA [29]. However, after adjusting for potential confounders, these associations lost their statistical significance. The findings of this study suggest that, in general, Chinese adults—regardless of whether they reside in urban and rural areas—with higher levels of RBIs are more likely to exhibit elevated SUA and a greater risk of HUA. The mechanisms underlying these associations remain unclear and are rarely studied. Nitric oxide (NO), primarily produced by endothelial cells (ECs) in the vessel wall through the action of endothelial nitric oxide synthase (eNOS), plays a crucial role in regulating blood flow and blood pressure [30]. Research by Leo et al. has demonstrated that RBC eNOS contributes directly to systemic NO bioavailability and blood pressure homeostasis, independent of vascular endothelial eNOS. RBC eNOS is implicated in various disease conditions, including coronary artery disease and chronic kidney disease, where its levels are found to be reduced [31, 32]. Additionally, impaired endothelial function has been linked to the pathophysiology associated with HGB levels [33]. HGB has been associated with soluble CD40 ligand (sCD40L), which can create a pro-inflammatory environment conducive to blood clotting and further exacerbate the development of CRFs and CVDs [29, 34, 35]. Furthermore, increased expression of miR-144 may be triggered by hemoglobin, leading to an inflammatory response in microglia [36]. There are also suggestions that oxidative stress effects may stem from excessive iron accumulation, a component of hemoglobin [14]. Another measure related to red blood cells, HCT, significantly influences blood flow by affecting blood viscosity and the proportion of red blood cells within total blood volume [37, 38]. The elevation in cardiovascular risk associated with high levels of RBIs may be due to increased blood viscosity, platelet activation, and oxidative stress from excessive iron accumulation linked to hemoglobin [39–41]. Moreover, prolonged hypercholesterolemia can result in elevated cholesterol levels within the red blood cell membrane, leading to diminished membrane fluidity and impaired deformability of RBCs [42, 43]. In our analysis, the impact of RBIs and serum lipids appears to be stronger compared to other indicators of CRFs, potentially reflecting these underlying mechanisms. Notably, in females, the relationship between HCT and HDL levels was found to be inverse; higher HCT levels were associated with a lower risk of low HDL. Research exploring the connection between HCT and HDL is limited, and the precise mechanisms remain unclear. Long-term cohort studies and experimental investigations are necessary to elucidate the effects of HCT on HDL and the underlying pathways involved. While higher HCT levels may correlate with elevated HDL, it is important to recognize that increased HCT could also serve as a biomarker for heightened cardiovascular risk when considering overall metabolic profiles, as indicated in previous studies [9]. The relationship between RBIs and CRFs exhibits notable sex differences, particularly concerning HUA. In males, a U-shaped association is observed between HGB levels and HUA, while in females, the relationship is predominantly linear. Moreover, sex differences are evident in their effects on serum lipids. These findings underscore the need for more tailored prevention and identification strategies for high-risk individuals, considering their sex. The main strength of this study lies in its relatively large sample size and national-level survey design. Additionally, the analysis of multiple cardiometabolic traits and CRFs provides a more comprehensive evaluation of health effects. Furthermore, the exploration of non-linear relationships among indicators reveals diverse associations between different dependent and independent variables, enhancing the clarity of the assessment. However, the study has several limitations. First, establishing causal inference may be challenging due to uncertain temporality in a cross-sectional study design. It should be noted that reverse causality may be a concern, as the directions of these associations are not definitive. For instance, overweight or obese individuals may exhibit higher levels of HGB or HCT [44], thereby increasing their risk of CRFs. Although we adjusted BMI in the regression models, residual confounding may still exist. Nonetheless, as potential biomarkers of CRFs, RBIs can assist in screening at the public health level to identify populations that may benefit from further risk stratification. Secondly, the study did not collect information on nutritional factors, such as dietary intake or the use of dietary supplements like iron, or on potential diseases that could cause hypoxia or red blood cell deformities. Conclusions In summary, our study has revealed both linear and non-linear associations between RBIs and various CRFs in a general Chinese population. The findings indicate that elevated levels of RBIs are associated with an increased risk of hyperuricemia, abnormal serum lipids, and diabetes. Notably, the relationships between RBIs and serum lipids are stronger than those observed with other cardiovascular risk factors, with hemoglobin and hematocrit serving as more effective indicators compared to red blood cell count. Furthermore, we have identified notable sex differences in these associations, emphasizing the importance of developing sex-specific risk prediction models and high-risk identification strategies in future research. Supplementary Information Supplementary Material 1: Table S1. The associations between red blood cell indexes and cardiovascular risk factors using survey-logistic regression models. Table S2. Sensitivity analysis-The associations between red blood cell indexes and cardiometabolic risk factors using traditional logistic regression without PSW. Fig.S1. Flow chart and design of the study. HGB: hemoglobin concentration; RBC: red blood cell count; HCT: hematocrit. HUA: hyperuricemia; SUA: serum uric acid; FPG: fasting plasma glucose; TC: total cholesterol; TG: triglycerides; LDL: low-density lipoprotein cholesterol; HDL: high-density lipoprotein cholesterol; OR: odds ratio; CI: confidence interval. Fig.S2. The linear and non-linear relationship between hemoglobin concentration, red blood cell count, hematocrit and multiple cardiometabolic traits. FPG: fasting plasma glucose; TC: total cholesterol; TG: triglycerides; LDL: low-density lipoprotein cholesterol; HDL: high-density lipoprotein cholesterol; HGB: hemoglobin concentration; RBC: red blood cell count; HCT: hematocrit. A: relationships between RBIs and serum uric acid; B: relationships between RBIs and FPG; C: relationships between RBIs and LDL; D: relationships between RBIs and HDL; E: relationships between RBIs and TG; F: relationships between RBIs and TC. Fig.S3. Sex comparison on the association between red blood cell indices and serum uric acid, demonstrated by the restrict cubic spline regression models. HGB: hemoglobin concentration; RBC: red blood cell count; HCT: hematocrit. Fig.S4. Sex comparison on the association between red blood cell indexes and hyperuricemia, demonstrated by the restrict cubic spline regression models. HUA: hyperuricemia. HGB: hemoglobin concentration; RBC: red blood cell count; HCT: hematocrit. Supplementary Material 1: Table S1. The associations between red blood cell indexes and cardiovascular risk factors using survey-logistic regression models. Table S2. Sensitivity analysis-The associations between red blood cell indexes and cardiometabolic risk factors using traditional logistic regression without PSW. Fig.S1. Flow chart and design of the study. HGB: hemoglobin concentration; RBC: red blood cell count; HCT: hematocrit. HUA: hyperuricemia; SUA: serum uric acid; FPG: fasting plasma glucose; TC: total cholesterol; TG: triglycerides; LDL: low-density lipoprotein cholesterol; HDL: high-density lipoprotein cholesterol; OR: odds ratio; CI: confidence interval. Fig.S2. The linear and non-linear relationship between hemoglobin concentration, red blood cell count, hematocrit and multiple cardiometabolic traits. FPG: fasting plasma glucose; TC: total cholesterol; TG: triglycerides; LDL: low-density lipoprotein cholesterol; HDL: high-density lipoprotein cholesterol; HGB: hemoglobin concentration; RBC: red blood cell count; HCT: hematocrit. A: relationships between RBIs and serum uric acid; B: relationships between RBIs and FPG; C: relationships between RBIs and LDL; D: relationships between RBIs and HDL; E: relationships between RBIs and TG; F: relationships between RBIs and TC. Fig.S3. Sex comparison on the association between red blood cell indices and serum uric acid, demonstrated by the restrict cubic spline regression models. HGB: hemoglobin concentration; RBC: red blood cell count; HCT: hematocrit. Fig.S4. Sex comparison on the association between red blood cell indexes and hyperuricemia, demonstrated by the restrict cubic spline regression models. HUA: hyperuricemia. HGB: hemoglobin concentration; RBC: red blood cell count; HCT: hematocrit.