Association between serum copper concentration and type 2 diabetes mellitus in elderly people in Northeast China: a cross-sectional study Authors: Xiaonan Li, Li Yang, Dongmei Mu, Yuehui Wang Published: 2025 ABSTRACT AbstractBackgroundNumerous studies have demonstrated a significant association between serum copper (Cu) levels and abnormal glucose metabolism. However, there is a notable lack of research specifically examining the relationship between serum Cu concentration and type 2 diabetes mellitus (T2DM) in the elderly population. Consequently, this study aims to investigate the potential association between serum Cu concentration and T2DM in elderly adults.MethodsA total of 565 participants were recruited from the outpatient and inpatient geriatric departments of the First Hospital of Jilin University between July 2021 and July 2023. Logistic regression models were employed to explore the association between serum Cu concentration and the risk of diabetes mellitus, with odds ratios (ORs) and 95% confidence intervals (95% CIs) calculated. Additionally, both smooth curve fitting and restricted cubic splines were utilized to assess the dose–response relationship between serum Cu concentration and T2DM.ResultsThe results showed that serum Cu concentrations were higher in the T2DM group (17.5 ± 4.6μmol/L) than in the control group (16.6 ± 4.0μmol/L) (P = 0.021). After adjusting for potential covariates, the ORs (95% CI) for higher serum Cu levels were 1.00 (reference value), 1.66 (1.08–2.55), 2.07 (1.32–3.26), and 2.06(1.26–3.38) for T2DM, respectively. Both smooth curve fitting and restricted cubic spline curves showed a potential linear relationship between serum Cu concentration and T2DM (P for non-linear = 0.231 and P = 0.08).ConclusionThis study suggests that T2DM in elderly adults is associated with increased serum Cu concentrations. Further prospective cohort studies in different populations are needed to validate our results.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12877-025-06815-7.Keywords: T2DM, Elderly, Insulin Resistance, Serum copper Introduction Type 2 diabetes mellitus (T2DM) is a chronic condition resulting from complex interactions between genetic and environmental factors. It is characterized by reduced insulin secretion from pancreatic β-cells, insulin resistance, or a combination of both, which ultimately leads to elevated blood glucose levels [1]. According to 2019 data, China has the highest number of elderly diabetics in the world, with approximately 35.5 million individuals affected. This accounts for nearly a quarter of the global elderly diabetic population and reflects a rising trend as the population aging [2]. The etiology of T2DM involves a combination of internal and external risk factors. The diabetes epidemic is the result of the interplay between unhealthy lifestyles—such as poor diet and physical inactivity—and genetic and epigenetic influences. Most studies have primarily focused on macronutrient status [3], emphasizing strategies to prevent T2DM by increasing the intake of complex carbohydrates over monosaccharides [4]. However, an imbalance in micronutrient homeostasis, particularly in trace elements (TEs), has been recognized as a significant factor influencing the risk of developing T2DM [5, 6]. Copper (Cu), a trace element and essential mineral, plays a crucial role in enzymatic reactions and oxidative/antioxidative processes in the body. Additionally, it is significant in the pathogenesis of T2DM [7, 8]. In vivo, Lin et al. and Naka et al. demonstrated a positive correlation between serum Cu concentration and HbA1c levels. In studies involving diabetic db/db mice, treatment with a Cu chelator was found to reduce reactive oxygen species (ROS) levels and improve both insulin resistance and glucose intolerance [9–11]. Several studies have highlighted a close relationship between serum Cu concentration and impaired glycemic control [12, 13]. An imbalance in Cu levels can disrupt antioxidant homeostasis and contribute to the progression of diabetes-related complications [8, 14, 15]. Aging is closely linked to the disruption of trace element homeostasis and a progressive decline in antioxidant defense mechanisms [16]. The elderly population exhibits age-dependent alterations in metallothionein expression patterns, diminished renal clearance capacity, and impaired intestinal absorption efficiency, collectively rendering them particularly vulnerable to copper homeostasis dysregulation [17, 18]. This dyshomeostasis may subsequently exacerbate its deleterious metabolic ramifications. While numerous studies have investigated the relationship between Cu and diabetes in humans, there is a notable scarcity of research focusing on the association between T2DM and serum Cu concentration specifically in elderly adults. This lack of research is particularly concerning given the physiological and nutritional differences in this population, which often lead to frequent homeostatic imbalances of serum Cu [19–22]. Therefore, we conducted a cross-sectional study of elderly adults in the Northeast Region to examine the relationship between serum Cu concentration and T2DM in this demographic. Materials and methods Experimental design and population A total of 951 participants were recruited from the outpatient and inpatient geriatric departments of the First Hospital of Jilin University between July 2021 and July 2023. Based on the established inclusion and exclusion criteria, 565 individuals were ultimately included in the analysis (Fig. 1). Our study was a retrospective study not a clinical study, so clinical trial number were not applicable. This study adhered to medical ethics standards and received approval from the Ethics Committee of the First Hospital of Jilin University (NO.19K028-001). All patients involved in the study provided written informed consent to participate in the research, and it was wavied. This study was conducted in compliance with the Declaration of Helsinki and all applicable ethical guidelines. Flowchart. Flowchart for the inclusion of participants. After screening, a total number of 565 participants were included in the final analysis Inclusion criteria and Exclusion criteria Inclusion criteria: Participants were required to be over 60 years of age, willing to enroll, and able to provide informed consent. Exclusion criteria: These were individuals with a prior confirmed diagnosis of T2DM, those currently taking oral medications or following diets that may affect Cu ion concentrations, individuals under 60 years of age, and those with acute cardiovascular or cerebrovascular complications, unconsciousness, chronic kidney disease, hypertension, or malignant tumors were excluded from the study. Diagnosis of T2DM This study of patients with T2DM was conducted by the 1999 World Health Organization's criteria for T2DM, individuals with a prior, established diagnosis were excluded. Measurement of T2DM All participants first completed a structured questionnaire to exclude individuals with T2DM confirmed by self-report or existing medical records. To mitigate the limitations of self-report (undiagnosed cases or recall bias), all remaining control group individuals underwent standardized biochemical tests, specifically Fasting Plasma Glucose (FPG) and Glycated Hemoglobin (HbA1c%) measurement. For subjects with specific risk factors (impaired FPG or Obesity), a 75g Oral Glucose Tolerance Test (OGTT) was additionally performed. Enhanced Validation for High-Risk Individuals: For subjects with specific risk factors (impaired FPG or Obesity), a 75g Oral Glucose Tolerance Test (OGTT) was additionally performed. Measurement of trace elements Trace element measurements were performed in the clinical laboratory of the Department of Laboratory Medicine of the First Hospital of Jilin University, which complies with the ISO 15189 testing standard. Serum Cu concentrations were measured by inductively coupled plasma mass spectrometry (ICP-MS, Agilent 7700X, USA) in the laboratory of the First Hospital of Jilin University. All subjects were required to fast for at least 12 h prior to venipuncture. Blood was collected using trace element-specific, metal-contamination-free tubes. The polypropylene cryovials used for serum aliquoting were subjected to standardized acid-washing procedures (10% nitric acid immersion followed by rinsing with ultrapure water) to eliminate exogenous metal contamination. Samples were centrifuged at 3000 × g for 10 min at 4℃ within 2 h of collection to separate serum. Aliquoted serum was immediately stored in − 80℃ cryovials. Crucially, no serum samples underwent freeze–thaw cycles prior to analysis, and all analyses were completed within 6 months of collection. All samples underwent serum copper measurement within the same analytical batch to eliminate batch-to-batch variation. For quality assurance, we measured the metal content once in every 20 samples using standard reference materials (ClinChek, No. 8883 and 8884). It was confirmed that the Cu measurements in the reference material were within the recommended range. The measured concentration was 917 ± 67 μg/L (certified: 925 ± 185 μg/L) for No. 8883 and 1314 ± 114 μg/L (certified: 1363 ± 273 μg/L) for No. 8884. The intra- and inter-assay coefficients of variation for Serum Cu were less than 5%. The limit of detection (LOD) was 0.009 μmol/L and all subjects had serum Cu levels above the LOD. Outlier Handling We employed the widely used Three Standard Deviation Rule (Mean ± 3SD) to identify potential outliers in serum copper concentration, a method suitable for our approximately normally distributed data. For the small number of identified outliers, we applied the Winsorization technique. Measurement of other covariates Epidemiological information was gathered, encompassing sociodemographic factors (age and sex), lifestyle behaviors (smoking and alcohol consumption), and medical history. A trained investigator collected data using an epidemiological questionnaire. Height and weight measurements were taken by qualified staff, with participants dressed in lightweight indoor clothing and without shoes. For laboratory analyses, a Hitachi 7600 automatic biochemical analyzer was utilized, along with reagents and standards from Shanghai Kehua Company and quality control materials from Burroughs Company, USA. The measurements included aspartate aminotransferase (AST), alanine aminotransferase (ALT), serum creatinine (Scr), cholinesterase (CHE), total cholesterol (TC), triglycerides (TG), high-density lipoprotein (HDL), and low-density lipoprotein (LDL) levels. Complete blood counts were performed using a SYSMEX-SE-2100 automatic hematology analyzer, while HbA1c levels were determined using a glycated hemoglobin analyzer (Bio-Rad Variant). All specimens were measured within two hours of collection. Definition of Relevant Indicators BMI = weight (kg)/height (m)2; BMI was classified as ≥ 25 or < 25 (overweight or non-overweight) according to the WHO standard. BMI = weight (kg)/height (m)2; BMI was classified as ≥ 25 or < 25 (overweight or non-overweight) according to the WHO standard. Triglyceride glucose index (TyG) = ln (fasting triglycerides [mg/dL] × FPG [mg/dL]/2). Triglyceride glucose index (TyG) = ln (fasting triglycerides [mg/dL] × FPG [mg/dL]/2). Estimated glomerular filtration rate (eGFR) was calculated according to the Simplified Modification of Diet in Renal Disease equation (MDRD): eGFR = 186 × (Scr)−1.154 × (age)−0.203 × (0.742 female). Estimated glomerular filtration rate (eGFR) was calculated according to the Simplified Modification of Diet in Renal Disease equation (MDRD): eGFR = 186 × (Scr)−1.154 × (age)−0.203 × (0.742 female). HbA1c was used as a reference for glucose control, < 7% for good glycemic control, and ≥ 7% for poor glycemic control. HbA1c was used as a reference for glucose control, < 7% for good glycemic control, and ≥ 7% for poor glycemic control. Smoking and alcohol consumption histories were obtained via a questionnaire. Smoking and alcohol consumption histories were obtained via a questionnaire. Outlier Handling Statistical analysis Continuous variables that followed a normal distribution were expressed as mean ± standard deviation, while those not following a normal distribution were presented as Q50 (Q25, Q75). Categorical variables were reported as counts and percentages. Comparisons between groups were made using the t-test, chi-square test, or Mann–Whitney U test, as appropriate. The serum Cu concentration was divided into tertiles: tertile 1 (Q1) < 33.3%, tertile 2 (Q2) ≥ 33.3% and < 66.7%, and tertile 3 (Q3) ≥ 66.7%. Logistic regression models were employed to investigate the association between serum Cu concentration and the risk of T2DM, calculating odds ratios (ORs) and 95% confidence intervals (95% CIs). Three models were utilized for these analyses: Model 1 was unadjusted for covariates; Model 2 adjusted for age, sex, height, and weight; Model 3 adjusted for age, sex, height, weight, alcohol consumption, smoking, AST, ALT, cholesterol, and triglycerides; Model 4 further adjusted for age, sex, height, weight, alcohol consumption, smoking, AST, ALT, cholesterol, and triglycerides, Zinc, Magnesium, Iron. Model 5 was adjusted for age, sex, height, weight, drinking, smoking, AST, ALT, Cholesterol, triglycerides, Zinc, Magnesium, Iron, Creatinine and BUN. To ensure stability, we performed a formal multicollinearity diagnostic using the Variance Inflation Factor (VIF) for all covariates in Models. In the models, serum Cu concentration was treated as both a continuous and categorical variable, with the lowest tertile serving as the reference group. Additionally, smooth curve fitting and restricted cubic spline methods were applied to explore the dose–response relationship between serum Cu concentration and T2DM. Categorical variables such as age, sex, BMI, smoking status, alcohol consumption, and glycemic control were also analyzed. All statistical analyses were conducted using SPSS 22 and Free Statistics 19, with statistical significance set at P < 0.05. Results Baseline characteristics of participants Table 1 presents the baseline characteristics of the participants. A total of 565 eligible participants were included in the study, of whom 202 were diagnosed with diabetes and 363 served as controls. Participants with diabetes exhibited higher body weight, FPG, HbA1c levels, TyG index (P < 0.05). In this study, we further explored the relationship between serum Cu concentration and T2DM. As shown in Table 1, we expressed the tertiles of serum Cu concentration as stratified categorical variables, which also included all relevant clinical characteristics of the subjects. Significant differences were observed in sex distribution, the presence of T2DM, body weight, FPG, HbA1c, blood lipid profiles (cholesterol, HDL, LDL), and liver function markers (AST, GGT, ALP, total protein, albumin, globulin, total bilirubin, direct bilirubin, indirect bilirubin) across different levels of Cu concentration (P < 0.05). Table 1 Baseline characteristics of participantsVariablesTotalNot diabetesDiabetesPserum Cu concentration(μmol/L)Tertile 1(5.90-14.80)Tertlie 2(14.81-17.30)Tertlie 3(17.31-19.30)PSex, n (%).827<.001Female307(54.3)196(54)111(55)73(38.6)96(58.5)138(65.1)Male258(45.7)167(46)91(45)Drinking, n (%).382.596No509(90.1)330(90.9)179(88.6)167(88.4)150(91.5)192(90.6)Yes56(9.9)33(9.1)23(11.4)22(11.6)14(8.5)20(9.4)Smoking, n (%).829No456(80.7)292(80.4)164(81.2).687Yes109(19.3)71(19.6)38(18.8)150(79.4)131(79.9)175(82.5)Age, y, Mean ± SD72.3 ± 7.772.5± 7.971.9 ± 7.5.45271.9 ± 7.671.7 ± 7.873.1 ± 7.8.143BMI, kg/m2, Mean ± SD23.8 ± 3.223.5± 3.324.5 ± 3.0<.00123.9 ± 3.223.5 ± 3.024.0 ± 3.4.285Weight, kg, Mean ±SD64.9 ± 10.363.9±10.366.6 ± 10.0.00266.7 ± 10.963.7 ± 9.164.1 ± 10.5.009Height, cm, Mean ±SD164.8 ± 7.3164.8±7.3164.8 ± 7.4.886166.7 ± 7.3164.6 ± 6.9163.3 ± 7.3<.001FBG, mmol/L, Mean ±SD6.0 ± 2.25.1 ± 0.97.7 ± 2.8<.0015.8 ± 2.15.7 ± 1.96.4 ± 2.4.007HbA1C,%, Mean± SD6.6 ± 1.35.8 ± 0.47.9 ± 1.4<.0016.3 ± 1.16.5 ± 1.26.9 ± 1.5<.001VitD, ng/ml, Median (IQR)16.9(12.1, 22.5)17.3(12.0, 22.7)16.5(12.4,22.4).62919.0 ± 8.618.3 ± 7.40016.8 ± 7.800.019Calcium, mmol/L, Mean ± SD2.21 ± 0.1352.20 ± 0.1292.24 ± 0.141<.0012.22 ± 0.1362.23 ± 0.1192.19 ± 0.141Cholesterol, mmol/L, Median (IQR)4.46(3.63, 5.15)4.46(3.62, 5.17)4.27(3.64, 5.08).4544.01(3.41, 4.73)4.61(3.83, 5.24)4.51(3.70, 5.44)<.001Triglyceride, mmol/L, Mean± SD1.18(0.84, 1.65)1.08(0.78, 1.48)1.34(0.98,2.04)<.0011.0(0.8, 1.6)1.2(0.8, 1.8)1.2(0.9, 1.6).069TyG index, Mean ± SD8.6 ± 0.78.4 ± 0.69.0 ± 0.7<.0018.5 ± 0.78.6 ± 0.78.7 ± 0.7.041Calcium, mmol/L, Mean ± SD2.21 ± 0.1352.20 ± 0.1292.24 ± 0.141<.0012.22 ± 0.1362.23 ± 0.1192.19 ± 0.141.007Cholesterol, mmol/L, Median (IQR)4.46(3.63, 5.15)4.46(3.62, 5.17)4.27(3.64, 5.08).4544.01(3.41, 4.73)4.61(3.83, 5.24)4.51(3.70, 5.44)<.001Triglyceride, mmol/L, Mean± SD1.18(0.84, 1.65)1.08(0.78, 1.48)1.34(0.98,2.04)<.0011.0(0.8, 1.6)1.2(0.8, 1.8)1.2(0.9, 1.6).069HDL, mmol/L, Mean ± SD1.07 ± 0.2871.10 ± 0.2901.02 ± 0.0.277.0031.06 ± 0.2631.12 ± 0.3051.05 ± 0.291.033LDL, mmol/L, Mean ± SD2.8 ± 0.92.8 ± 0.92.8 ± 0.8.4352.5 ± 0.82.9 ± 0.82.9 ± 0.9<.001AST, U/L, Median (IQR)19.5(16.3, 23.8)20.0(17.1, 23.8)18.7(15.3,23.7).00615.4(11.9, 21.6)14.6(11.3, 20.1)13.8(10.1, 20.0).021ALT, U/L, Median (IQR)14.8(11.0, 20.5)14.3(10.8, 20.0)16.4(11.4,21.3).08119.4(16.5, 23.3)19.4(16.5, 23.0)19.6(15.8, 25.0).971GGT, U/L, Median (IQR)24.2(17.4, 39.1)22.8(16.8, 37.6)26.8(18.6,40.4).02421.7(15.0, 30.7)22.6(17.6, 36.1)29.4(19.6, 49.3)<.001Total bilirubin, μmol/L, Median (IQR)12.7(9.9, 16.6)12.8(10.2, 17.0)12.2(9.5, 15.8).04813.8(10.8, 18.7)12.1(9.9, 15.7)12.1(9.3, 15.9)DBIL, μmol/L, Median (IQR)2.5(1.8, 3.3)2.6(1.9, 3.4)2.3(1.8, 3.0).0242.7(2.0, 3.7)2.2(1.7, 2.9)2.3(1.8, 3.3)<.001IBIL, μmol/L, Median (IQR)10.1(8.0, 13.5)10.2(8.2, 13.6)9.9(7.7, 12.7).08011.1(8.5, 14.7)9.9(8.0, 12.7)9.6(7.5, 13.0)<.001Creatinine, μmol/L, Median (IQR)71.1(59.1, 83.8)70.1(59.2, 82.2)72.2(59.1,86.6).30976.4 ± 22.872.8 ± 20.474.7 ± 24.6.345Total bile acid, μmol/L, Median (IQR)3.5(2.2, 5.6)3.5(2.2, 5.6)3.5(2.1, 5.6).7063.8(2.3, 5.6)3.3(2.2, 5.9)3.5(2.1, 5.6).438BUN, mmol/L, Median (IQR)74.3(61.5, 84.8)68.6(59.9, 81.5)75.8 60.9, 83.6).33474.4(61.9, 84.5)69.8(59.9, 81.5)69.3(57.7, 84.3)0.166eGFR, mL/min/1.73m2, Median (IQR)88.2(74.1, 97.0)88.1(74.7, 97.0)88.9(73.3,97.0).52789.7(78.1, 97.2)89.5(74.6, 96.2)86.9(69.7, 97.2).224Continuous variables that followed a normal distribution were expressed as mean ± standard deviation (SD), and those that did not follow a normal distribution were expressed as Q50 (Q25, Q75). Categorical variables are expressed as n(%)Fasting glucose(FPG), serum calcium(Ca2+), triglyceride(TG), total cholesterol(TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), blood albumin(ALB), alanine aminotransferase (ALT), aspartate aminotransferase (AST) and BMI(Body mass index), Triglyceride glucose index (TyG) index, Estimated glomerular filtration rate (eGFR), DBIL(Direct Bilirubin), IBIL(Indirect Bilirubin), BUN(Blood urea nitrogen), IQR (Interquartile Range) Baseline characteristics of participants Continuous variables that followed a normal distribution were expressed as mean ± standard deviation (SD), and those that did not follow a normal distribution were expressed as Q50 (Q25, Q75). Categorical variables are expressed as n(%)Fasting glucose(FPG), serum calcium(Ca2+), triglyceride(TG), total cholesterol(TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), blood albumin(ALB), alanine aminotransferase (ALT), aspartate aminotransferase (AST) and BMI(Body mass index), Triglyceride glucose index (TyG) index, Estimated glomerular filtration rate (eGFR), DBIL(Direct Bilirubin), IBIL(Indirect Bilirubin), BUN(Blood urea nitrogen), IQR (Interquartile Range) Association between the association between trace elements and glucose metabolism indicators The involvement of trace elements in metabolic health is well-established, with multiple studies suggesting their association with glucose metabolism and insulin resistance [23].To contribute to this understanding, we investigated the specific relationship of key serum trace elements in diabetic versus non-diabetic individuals, and the results are shown in Table 2. The serum Cu and magnesium concentration significantly differed between those with and without diabetes, with levels being greater in the diabetic group. An association between trace elements and indicators related to glucose metabolism (FPG, HbA1c levels, TyG index) was further analyzed and the results were shown in Table S1. We found that serum copper ions were correlated with all glucose metabolism indicators, while magnesium ions were significantly associated with FPG and HbA1c; Zinc ions were only correlated with the TyG index. Relationship between trace elements and participates Table 2 Relationship between trace elements and participatesVariables Total Not diabetesDiabetesPZinc, μmol/L, Mean± SD12.6 ± 2.812.7 ± 2.912.4 ± 2.8.370Cu, μmol/L, Mean ± SD16.9 ± 4.316.6 ± 4.017.5 ± 4.6.021Magnesium, mmol/L, Mean ± SD0.86 ± 0.1030.87±0.1040.83±0.096<.001Iron, μmol/L, Median (IQR)13.7(10.2, 17.8)13.9(10.2,18.4)13.5(10.1,16.9).164 Relationship between trace elements and participates Association between Cu concentration and T2DM To assess potential multicollinearity among the independent variables included in the multivariable model, we fitted an auxiliary linear regression and calculated the variance inflation factor (VIF) for each covariate. All VIF values were below 5, indicating the absence of significant multicollinearity (Table S2). Subsequently, we investigated the relationship between serum copper and T2DM. Table 3 shows the relationship between serum copper and T2DM. Table 3 details the odds ratios (OR) and 95% confidence intervals (CI) derived from multifactorial logistic regression analysis. When serum Cu was treated as a continuous variable, the ORs and 95% CIs for Models 1, 2, 3, 4 and 5 were (1.05,1.01–1.09), (1.05, 1.01–1.10), (1.08, 1.03–1.12),1.06 (1.01 ~ 1.11) and 1.08 (1.03 ~ 1.14), respectively. In cases where serum Cu concentration was categorized, the ORs and 95% CIs for tertile 2 and 3 in Model 5 were 1.05 (0.63 ~ 1.74)0.866 and 22.03 (1.24 ~ 3.33)0.005 compared to tertile 1. Additionally, we employed smooth curve fitting and restricted cubic spline analyses to explore the dose–response relationship between serum Cu and T2DM. The results, illustrated in Figs. 2 and 3, indicate a linear relationship between serum Cu and T2DM (P for nonlinear = 0.231 and P = 0.08). Table 3Association between Cu and T2DMModel 1Model 2Model 3Model 4Model 5OR(95%) P valueOR(95%) P valueOR(95%) P valueOR(95%) P valueOR(95%) P valueCu1.05 (1.01~1.09).0221.05 (1.01~1.1).0201.08 (1.03~1.12)0.001 1.06 (1.01~1.11).022 1.08 (1.03~1.14) .002Cu TertilesTertile 111111Tertile 20.89 (0.57~1.41).6280.96 (0.6~1.53).8671.07 (0.66~1.75).7821.04 (0.63~1.74).868 1.05 (0.64~1.74).866Tertile 31.61 (1.07~2.42).0221.66 (1.08~2.55).022.07 (1.32~3.26).0022.06(1.26~3.38).004 2.03 (1.24~3.33).005P for trend0.0180.0150.0010.0030.004Model 1, no covariates were adjusted. Model 2 was adjusted for age, sex, height, and weight. Model 3 was adjusted for age, sex, height, weight, drinking, smoking, AST, ALT, Cholesterol, and Triglycerides. Model 4 was adjusted for age, sex, height, weight, drinking, smoking, AST, ALT, Cholesterol, triglycerides, Zinc, Magnesium, and Iron. Model 5 was adjusted for age, sex, height, weight, drinking, smoking, AST, ALT, Cholesterol, triglycerides, Zinc, Magnesium, Iron, Creatinine and BUN.95% CI, 95% confidence interval; OR, odds ratio Association between Cu and T2DM Model 1, no covariates were adjusted. Model 2 was adjusted for age, sex, height, and weight. Model 3 was adjusted for age, sex, height, weight, drinking, smoking, AST, ALT, Cholesterol, and Triglycerides. Model 4 was adjusted for age, sex, height, weight, drinking, smoking, AST, ALT, Cholesterol, triglycerides, Zinc, Magnesium, and Iron. Model 5 was adjusted for age, sex, height, weight, drinking, smoking, AST, ALT, Cholesterol, triglycerides, Zinc, Magnesium, Iron, Creatinine and BUN.95% CI, 95% confidence interval; OR, odds ratio Association between Cu level and T2DM.he smoothed curve of association between Cu level and T2DM. This figure shows the dose–response relationship between serum Cu concentration and the risk of T2DM in elderly individuals. The red line represents the correlation between Cu and T2DM, and the blue line represents the 95% CI The restricted cubic spline (RCS) of association between Cu level and T2DM.The restricted cubic spline (RCS) of association between Cu level and T2DM. RCS shows a linear relationship between Cu and T2DM. The fitted regression line is a solid blue line; the black dashed line indicates the position where the OR is equal to 1; the shaded area indicates the 95% CI The Result of the ROC curve According to the ROC curve (Fig. 4), the commonly used Cu threshold of 17.55μmol/L was the best cutoff to detect T2DM in this population, with a 69.1%sensitivity and 44.9% specificity. The AUC of the association between Cu and T2DM is 0.560 (95% CI 0.510–0.610, P = 0.019).This result strongly indicates that serum copper has limited value as a single, independent clinical diagnostic tool. This is expected given the complex, multifactorial etiology of T2DM.However, the low AUC value does not negate the biological and epidemiological significance of our finding. We conclude that future research must move beyond single markers by focusing on constructing multi-element and multi-omics joint prediction models to improve clinical utility. ROC curve of Cu and the presence of T2DM. AUC: 0.560 (95% CI 0.510–0.610) P = 0.019 We also performed a logistic regression analysis to examine the relationship between serum magnesium ions and diabetes (Table S3). Without considering confounding factors, the results indicated a significant association between diabetes and serum magnesium ions (P < 0.001). After adjusting for confounding factors, the 95% confidence intervals and P-values for models 2, 3, and 4 were 156.48 (0.17–141,203.46) P = 0.146, 953.44 (0.71–1,276,104.72) P = 0.062, and 703.06 (0.43–1,143,928.23) P = 0.082, all showing P-values > 0.05. Subgroup analysis To definitively assess the robustness of our findings across different modeling approaches, we conducted a comprehensive subgroup analysis. The results confirmed that the positive association between T2DM and the variable of interest remained consistent across all predefined subgroups (Fig. 5). Subgroup analysis of the relationship between Cu and T2DM.The association between serum Cu concentration and T2DM in the elderly was stratified according to age, sex, smoking, alcohol consumption, body weight, and glycemic control Discussions Serum Copper Dyshomeostasis and its Association with T2DM in the Elderly In this cross-sectional study of an elderly population in the Northeast Region, our central finding is the significant positive association between elevated serum Cu concentrations and T2DM. This association remained statistically significant and robust even after rigorous adjustment for key metabolic and lifestyle confounders. This result not only substantiates findings from existing literature linking Cu and glucose metabolism disorders [24–26] but also extends this crucial observation to a specific, vulnerable elderly cohort, thereby addressing a gap in the current evidence base. We further demonstrated that serum Cu concentration was positively correlated with multiple indicators of glucose control, including FPG, HbA1c and the TyG index. This differential pattern of correlation supports the hypothesis that Cu dyshomeostasis is implicated in both acute glycemic fluctuations and long-term metabolic dysregulation. Biological Plausibility and Interpretation of Findings The elevated Cu levels observed in the T2DM group compared to controls are biologically plausible and suggest several potential mechanisms. The Copper-Driven Pro-Oxidant/Inflammatory Cascade: Copper is a transition metal with strong redox properties [27]. Even the modest absolute difference observed between diabetic and non-diabetic groups, though small, can be physiologically meaningful. As a catalyst, a minor increase in free or loosely bound Cu ions is sufficient to significantly accelerate the production of ROS via the highly efficient Fenton/Haber–Weiss reactions [28]. Given that oxidative stress is a central mechanism driving insulin resistance and β-cell damage [29], this minor change may actively disrupt the delicate physiological redox balance, especially in the elderly who possess a lower homeostatic reserve and anti-oxidant capacity [30]. Furthermore, the main carrier of serum Cu, Ceruloplasmin (CP), is an acute-phase reactant protein [31]. T2DM is recognized as a state of chronic low-grade inflammation [32]. The elevated total serum Cu may therefore be interpreted not as an etiological agent, but as a sensitive biomarker reflecting the persistent inflammatory environment capable of influencing insulin signaling [32].This interpretation aligns with Mikami's research linking serum Cu to inflammatory cytokines and oxidative stress [33]. Link to Insulin Resistance and Metabolic Syndrome: The positive correlation between serum Cu and the TyG index—a robust surrogate marker for insulin resistance—suggests that elevated Cu may be intrinsically linked to the underlying pathology of insulin resistance [34]. Dysregulation of Cu metabolism is known to potentially impair lipid metabolism and worsen insulin resistance [35, 36]. High Cu levels may promote hepatic lipid synthesis while inhibiting fatty acid oxidation, thereby contributing to the abnormal lipid metabolism and inflammatory states associated with an elevated TyG index [37, 38]. Integrating the Trace Element Network and Confounding Control The biological impact of Cu cannot be studied in isolation; thus, a thorough understanding of the trace element network is critical. Rationale for Covariate Inclusion: Our Model 4 and Model 5 rigorously included Zn, Mg, and Iron Fe as covariates, based on compelling biological rationale: Zn and Cu exhibit competitive intestinal absorption and share homeostatic pathways. Zn is crucial for insulin function, and its deficiency is linked to insulin resistance [39, 40]. Adjusting for Zn controlled for confounding effects from the Cu − Zn metabolic axis. Mg is a key cofactor in glucose metabolism, and its levels are similarly affected by the systemic inflammation central to T2DM [41, 42]. Fe and Cu homeostasis are linked by CP, the copper-containing ferroxidase [43, 44]. Adjusting for Fe helped to control for confounding via this shared CP-mediated pathway, reducing the influence of total body iron stores [45]. The significant, independent association of Cu with T2DM persisting after this rigorous adjustment confirms the robustness of our finding and its independence from the primary metabolic actions of Zn, Mg, and Fe. Furthermore, our dedicated analysis in Sect. 3.2 revealed differential associations: while Cu correlated with all three glycemic indicators (FPG, HbA1c and TyG), Mg was linked to FPG and HbA1c, and Zn was only correlated with the TyG index. This result strongly suggests that these micronutrients play distinct and differential roles within the T2DM pathology. Dietary factors are environmental risk factors for T2DM. Although serum Cu is mainly controlled by endogenous homeostatic regulation [46, 47], our lack of detailed dietary data constitutes a significant source of residual confounding. Furthermore, polymorphisms in key copper metabolism genes (ATP7B) may cause baseline individual differences in Cu levels [48]. The observed association may be partially explained by this underlying genetic background. Clinical Significance and Predictive Limitations The clinical significance of our study is that it provides empirical evidence for monitoring Cu levels in the elderly, suggesting that elevated Cu may indicate an increased risk of metabolic disorders. Regular assessment could aid in the early identification of T2DM risk and inform personalized dietary strategies. However, we must candidly address the low AUC value (0.560) observed in the ROC curve analysis. This low value confirms that serum Cu has limited discriminative power when used as a single, independent biomarker. The core contribution of this study is not to provide a new clinical prediction tool, but to advance the understanding of T2DM pathophysiology by demonstrating a strong, biologically plausible association. The clinical utility of Cu will likely be realized through its integration into multi-dimensional prediction models combining genetic markers, inflammatory indicators, and other trace elements. Study Limitations and Future Directions Our study has several limitations that necessarily frame the interpretation of our results:1) Cross-Sectional Design and Causality: The design fundamentally restricts our ability to establish causality or the temporal sequence. The observed hypercupremia could equally be a consequence of T2DM-related inflammation (reverse causality), metabolic dysregulation, or tissue redistribution, rather than an initiating risk factor [7, 8]. Only rigorously designed large-scale prospective cohort studies and Mendelian Randomization (MR) studies can definitively verify the true causal link.2) Lack of Dietary Data: The absence of detailed dietary intake data is a major limitation, potentially introducing residual confounding. Although we mitigated this by emphasizing that serum Cu is under strict hepatic endogenous homeostatic regulation [37, 38] and by performing rigorous multivariate adjustments (including competing trace elements), we cannot rule out the influence of long-term dietary patterns. Future studies must incorporate Food Frequency Questionnaires (FFQ) and animal models to fully differentiate the contributions of exogenous dietary factors from endogenous homeostatic imbalance to glucose metabolism. 3) Generalizability: Despite the geographical limitations and selection bias of this single-center study's sample (where patients with severe comorbidities were excluded to maximize internal validity), which restrict external generalizability, its value is undeniable. Given the unique context of China's large elderly and T2DM populations, this research provides important localized data and, as an exploratory investigation, offers strong preliminary evidence and a scientific hypothesis for the association between serum copper and T2DM. Future studies must be expanded to multi-center, international cohorts to enhance external validity and confirm the universality of these findings.4) Incomplete Pathophysiological Markers: While we adjusted for general liver/kidney enzymes, more detailed pathological changes or inflammation markers (like CRP) could still act as unmeasured confounders. Future research should combine Cu levels with these specific inflammatory markers to provide a more complete perspective on the Cu-inflammation-diabetes axis. In conclusion, this study provides strong empirical evidence of a robust and biologically plausible association between elevated serum copper concentrations and T2DM in the elderly population. Despite the inherent limitations of a cross-sectional design, our findings lay a crucial groundwork for future longitudinal and mechanistic studies, which should focus on the complex interactions within the micronutrient network to offer comprehensive support for clinical practice and personalized metabolic strategies. Supplementary Information Supplementary Material 1. Acknowledgements No. Authors’ contributions Conceptualization: Yuehui Wang. Methodology: Li Yang. Formal analysis: Xiaonan Li. Data curation: Dongmei Mu and Li Yang. Writing—original draft preparation: Xiaonan Li. Writing—review and editing: Yuehui Wang, Dongmei Mu and Li Yang. Funding This work was funded by the Technology Development Plan of Jilin Province (Grant No. 20240304164SF). Data availability Please contact the corresponding author if you want the data. Declarations Ethics approval and consent to participate The study adhered to the medical ethics standards and was approved by the Ethics Committee of the First Hospital of Jilin University (NO.19K028-001). Our study was a retrospective study not a clinical study, so clinical trial number were not applicable. All patients involved in the study provided written informed consent to participate in the research. Consent for publication All patients involved in the study provided written informed consent for the publication of the research materials without mentioning their names. Competing interests. 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. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement Please contact the corresponding author if you want the data.