Beyond glycemia: adropin, asprosin, and irisin as potential biomarkers for cardiovascular risk in diabetes and prediabetes

Abstract

Type 2 diabetes mellitus (T2DM) is a prevalent metabolic disorder. Adropin, asprosin, and irisin, recognized for their metabolic roles, lack clear establishment in cardiovascular disease (CVD) risk and T2DM guidelines. This study aimed to investigate the relationships between serum levels of these biomarkers and CVD risk in healthy, prediabetic, and T2DM individuals. A total of 30 individuals with T2DM, 30 prediabetic subjects, and 29 healthy controls were included in the study. The Framingham Risk Score (FRS) was calculated for each participant. Anthropometric measurements and key biochemical parameters, including fasting blood glucose, HbA1c (glycated hemoglobin), and lipid profile, were recorded. Serum levels of adropin, asprosin, and irisin were quantified using enzyme-linked immunosorbent assay. No statistically significant differences were found in serum adropin, irisin, and asprosin levels across diabetes status groups (p > 0.05). Adropin levels were significantly higher in individuals with lower waist circumference (WC) and body mass index (BMI) (p = 0.029; p = 0.024). Low asprosin levels correlated with greater WC (p = 0.021). FRS correlated significantly with metabolic parameters (age, BMI, WC, blood pressure, glucose, HbA1c, TG (triglycerides), HDL (High-Density Lipoprotein), HOMA-IR (Homeostatic Model Assessment of Insulin Resistance)). While no direct significant relationship was observed between these biomarkers and FRS, a positive significant correlation was found among adropin, asprosin, and irisin levels across all groups. These findings should be interpreted as exploratory and hypothesis generating. To the best of our knowledge, this study is the first to evaluate the associations of adropin, asprosin, and irisin with CVD risk in healthy, prediabetic, and T2DM individuals using FRS. The findings indicate that, although these proteins may not serve as independent predictors of CVD risk, their positive inter-correlations likely reflect underlying subclinical metabolic processes, providing valuable insights for future risk assessment strategies.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-37306-3.

Introduction

Diabetes is an independent risk factor for cardiovascular disease (CVD). Cardiometabolic risks, which lead to adverse outcomes in both renal and cardiovascular terms in patients with diabetes, share a common pathophysiology, and the relationship between them is increasingly being considered. Atherosclerotic heart disease, heart failure, and chronic kidney disease are three comorbidities. These increases are generally caused by increased metabolic risk associated with obesity and obesity-related risk factors, and the prevalence of all these conditions increases with increasing HbA1c levels. In addition to treating hyperglycemia, hypertension, and hyperlipidemia in patients, the use of sodium‒glucose cotransporter-2 (SGLT-2) inhibitors or glucagon‒like peptide-1 (GLP-1) receptor agonists, which have demonstrated benefits, is recognized as a key component of management because it positively influences cardiovascular and renal outcomes¹.

According to the latest 2025 report published by the International Diabetes Federation (IDF), 588.7 million adults (11.1%) worldwide will be living with diabetes in 2024. The 2019 IDF and subsequent reports indicate a significant increase in diabetes incidence, with 2024 data already reaching the 2030 projections made in 2019; this highlights that diabetes is progressing faster than previously estimated^2,3.

Adropin, asprosin, and irisin have recently been identified as molecules with unclear effects on various systems, such as metabolism and the cardiovascular system, and they are not yet included in guidelines. Evaluating whether their blood levels are associated with the risk of T2DM and CVD is important for early diagnosis and risk assessment.

Kumar et al. conducted gene expression analysis in obese mice in 2008 to further investigate the hypothalamic regulation of liver metabolism and discovered that adropin, a previously unknown liver transcript peptide, is downregulated in obese individuals. Adropin is thought to play a role in the development of new therapeutic targets for the treatment of metabolic disorders associated with obesity. In this study, mice fed a high-fat diet secreted higher levels of adropin in the liver than did the control group, whereas fasted mice presented reduced adropin levels⁴.

In 2016, Romere et al. discovered asprosin, a protein hormone involved in this regulation. In humans and mice, asprosin, which pathologically increases with insulin resistance, is synthesized by white adipose tissue and released into the bloodstream during fasting, meaning that plasma levels increase during fasting⁵. Recent research has demonstrated a strong association between elevated circulating asprosin levels and both insulin resistance and obesity, thereby contributing to an increased risk of T2DM^6,7.

Irisin, first identified in 2012 by Boström et al., is a newly discovered myokine. When irisin levels in the bloodstream were increased by feeding obese and insulin-resistant mice a high-fat diet, it was found to increase energy expenditure, reduce body weight, and improve diet-induced insulin resistance. Following exercise, the release of the membrane protein fibronectin type III domain 5 (FNDC5) increases in muscles in response to peroxisome proliferator-activated receptor gamma coactivator-1 alpha (PGC1-α) stimulation, and irisin is composed of a fragmented portion of this protein. A mild increase in irisin levels in mice leads to an increase in energy expenditure without changes in physical activity or food intake⁸.

Materials and methods

This observational, cross-sectional, group-comparison study aims to examine the associations between serum levels of adropin, asprosin, and irisin and the risk of CVD in healthy individuals, those with prediabetes, and patients with T2DM. The current study was approved by the Local Ethics Committee of Selçuk University Faculty of Medicine (2024/252). The study was conducted in accordance with the Declaration of Helsinki issued by the World Medical Association.

A priori sample size calculation was performed using G-Power software (version 3.1). Based on effect sizes reported in previous studies evaluating circulating irisin, adropin and asprosin across groups, a one-way ANOVA model (fixed effects, three groups) was assumed. With an alpha level of 0.05, a statistical power of 95%, and an estimated effect size of f = 0.43, the required total sample size was calculated as 87 participants (29 per group). To minimize potential confounding effects, case and control groups were matched based on age, sex, and BMI. In this context, individuals were selected to be between 30 and 60 years of age, with BMI values ranging from 18.5 to 30 kg/m², and an equal distribution of sex was ensured. Subsequent analyses confirmed that there were no statistically significant differences between the groups in terms of age, sex, or BMI, indicating successful matching and control of these variables in intergroup comparisons.

A total of 89 participants were enrolled between July 1 and December 31, 2024, including 30 with T2DM, 30 with prediabetes, and 29 normoglycemic individuals serving as the control group. All participants with T2DM or prediabetes were under active follow-up and treatment at the time of enrollment. The diagnoses of T2DM, prediabetes, and normoglycemia were made in accordance with the IDF diagnostic criteria (11th edition)⁹.

Participants who refused to participate, insulin-dependent diabetes patients, those with type 1 diabetes, those who had experienced myocardial infarction, cerebrovascular events, peripheral artery disease, ischemic heart disease, angiography, stents, or bypass surgery, those who were currently experiencing at least one complication of diabetes, those with a history of cancer, those with a pacemaker, those with a history of heart valve surgery, those with any thyroid dysfunction, those with chronic inflammatory diseases, those receiving immunosuppressive therapy, pregnant or breastfeeding women, those with malabsorption disorders, those who had undergone gastrointestinal surgery, those with chronic kidney or liver failure, and those with acute infections were excluded from the study.

Anthropometric measurements

All the measurements were performed by the same nurse. Height, weight, and WC measurements were taken, and BMI was calculated. Following a minimum rest period of 10 min, systolic and diastolic blood pressures were measured and documented via a calibrated Omron M6 blood pressure monitor. In accordance with the World Health Organization (WHO) criteria, WC measurements were taken and recorded via a measuring tape while patients stood upright on a flat surface with their feet together, with the midpoint between the anterior superior iliac spine and the costal arch aligned with the midpoint of the measurement.

Blood sample collection and biochemical analysis

Venous blood samples were collected from the antecubital vein between 08:00 and 09:00 AM after a minimum of 8 h of overnight fasting, using gel separator biochemistry tubes. All biochemical analyses were performed in the hospital’s biochemistry laboratory. The parameters measured included complete blood count (CBC), fasting blood glucose, serum insulin, HbA1c, and lipid profile [total cholesterol, low-density lipoprotein (LDL), HDL and TG]. Additionally, HOMA-IR index calculated as fasting insulin (µU/mL) × fasting glucose (mg/dL)/405, for assessing insulin resistance¹⁰.

Serum Adropin, Asprosin and Irisin levels

Blood samples were centrifuged at 1500 × g for 10 min, and then, the serum samples were aliquoted into 1.5 ml Eppendorf tubes and stored at −80 °C until further analysis. After the kits were obtained, they were stored at + 4 °C until analysis according to the manufacturer’s instructions. Serum adropin, asprosin, and irisin levels were measured via an ELISA (human) kit (Bioassay Technology Laboratory, China) and a CLARIOstar ELISA reader following the manufacturer’s guidelines. For adropin, the analytical range was 5–1000 ng/L with a sensitivity of 2.49 ng/L; the intra-assay and inter-assay coefficients of variation (CV) were < 8% and < 10%, respectively. For asprosin, the analytical range was 0.5–100 ng/mL with a sensitivity of 0.23 ng/mL; the intra-assay and inter-assay CVs were < 8% and < 10%, respectively. For irisin, the analytical range was 0.2–60 ng/mL with a sensitivity of 0.095 ng/mL, and the intra-assay and inter-assay CVs were < 8% and < 10%, respectively.

Sociodemographic data

The sociodemographic data of the patients were collected via a 23-question questionnaire prepared by the researcher on the basis of a literature review and administered through face‒to-face interviews^11–13.

Framingham risk score

The participants’ 10-year risk of coronary heart disease was calculated via the Framingham risk score (FRS). Patients with a risk of experiencing a cardiovascular event in the next 10 years of more than 20% were considered high risk, those with a risk between 10% and 20% were considered moderate risk, and those with a risk of less than 10% were considered low risk. The risk calculation was performed via the online calculation tool provided by the Framingham Heart Study^14,15.

Statistical analysis

The data obtained from the study were transferred to a computer and analyzed via the SPSS (Statistical Package for Social Sciences) 21.0 software package. For descriptive analyses, frequency data are presented as numbers (n) and percentages (%), whereas numerical data are presented as arithmetic means ± standard deviations (sd) or medians (1st quartile (Q1) − 3rd quartile (Q3)).

Pearson’s or Fisher’s chi-square (χ2) test was used to compare categorical data. The normality of the numerical data was evaluated via visual (histogram, Q‒Q plot) and analytical (Kolmogorov‒Smirnov) methods. When two or more independent groups were compared, if the data followed a normal distribution, ANOVA was used; if the variances were homogeneous, Tukey’s test was used; and if the variances were not homogeneous, the Tamhane post hoc test was used. For data not following a normal distribution, the Kruskal‒Wallis test was used, and if significant differences were found, the post hoc Mann‒Whitney U test and Bonferroni correction were applied. The Spearman test was used to evaluate the relationship between two nonnormally distributed numerical datasets. The correlation levels were classified as follows: r = 0.00–0.19 for insignificant correlations, r = 0.20–0.39 for low correlations, r = 0.40–0.69 for moderate correlations, r = 0.70–0.89 for high correlations, and r = 0.90–1.00 for very high correlations¹⁶.

According to the FRS, a logistic regression model was used to investigate the predictive power of serum adropin, asprosin, and irisin levels for CVR risk between individuals with a CVR risk of less than 10% and those with a risk of 10% or higher. This analysis was performed using an unadjusted (crude) logistic regression model.

On the other hand, a stepwise logistic regression model was used to evaluate the relationship between adropin, irisin, and asprosin levels and type 2 diabetes. Logistic regression analyses were performed using sequentially adjusted logistic regression models. Model 1 was unadjusted. Model 2 was adjusted for potential confounders, including waist circumference, systolic and diastolic blood pressure, and Framingham risk score. Model 3 additionally included HOMA-IR, and Model 4 was further adjusted for lipid parameters (LDL, TG, and total cholesterol).

To assess the effects of the variables on adropin, irisin, and asprosin, univariate and multivariate linear regression models were created.

The statistical significance level was set at α = 0.05.

Results

In the present study, matching was performed between the control, prediabetes, and diabetes groups for age, BMI, and sex. Statistical analysis revealed no statistically significant differences when age, BMI, or sex were compared.

Among the participants, 50.6% (n = 45) were male, 89.9% (n = 80) were married, 44.9% (n = 40) were university graduates, 46.1% (n = 41) were civil servants, 92.1% (n = 82) lived in urban areas, 41.6% (n = 37) had income exceeding their expenses, 69.7% (n = 62) had never smoked, 95.5% (n = 85) did not consume alcohol, 84.3% (n = 75) did not take vitamin supplements, 48.3% (n = 46) exercised at least 30 min a day, 71.9% (n = 64) consumed fruits and vegetables daily, 67.4% (n = 60) were overweight, and 70.8% (n = 63) were classified as low risk according to the FRS. The participants had an average age of 48.68 ± 7.42 years, an average weight of 72.24 ± 11.01, an average height of 165.59 ± 10.55, an average BMI of 26.96 ± 2.59, and an average waist circumference (WC) of 93.76 ± 6.23. There were no known medical conditions among the participants in the control group. A total of 86.6% of participants in the prediabetes group had no additional medical conditions. A total of 59.9% of patients in the diabetes group had concomitant hyperlipidemia.

The average duration of diabetes in diabetic patients was 11.16 ± 6.71 years. The median age at diagnosis of diabetes was 39.50 years. A total of 73.3% (n = 22) of the participants had a family history of first-degree diabetes. Among patients taking medication, 33.3% (n = 10) used sulfonylurea drugs, 96.7% (n = 29) used metformin, 43.3% (n = 13) used SGLT-2 inhibitors drugs, 46.7% (n = 14) used dipeptidyl peptidase IV inhibitors drugs, 30.0% (n = 9) used thiazolidinediones, and 10.0% (n = 3) used meglitinides. It was determined that alpha-glucosidase inhibitors and GLP-1 receptor agonists group drugs were not used by patients in the diabetes group. 40% (n = 12) of the patients reported visiting a doctor for diabetes every 3 months, and 93.3% (n = 28) reported taking their medications regularly. The average duration of prediabetes in the prediabetes group was 2.76 ± 2.01 years. The median age at diagnosis for prediabetic patients was 49.50 years. A total of 60.0% of the participants (n = 18) had a first-degree relative with diabetes. It was determined that 16.6% of patients (n = 5) used metformin, while other medication groups were not used by patients in the prediabetes group. A total of 26.7% (n = 8) of patients in the prediabetes group reported visiting a healthcare provider every six months, and 13.3% (n = 4) reported taking their medications regularly.

A statistically significant difference was found between exercise status and diabetes status (p = 0.010), with the control group engaging in more exercise. A significant difference was also found between a family history of diabetes and diabetes status (p = 0.003), with the control group having less family history and the diabetes group having more. A significant difference was found between FRS level and diabetes status (p < 0.001), with the diabetes group having fewer individuals in the low-risk group and more individuals in the medium- and high-risk groups. No statistically significant differences were found between other characteristics and diabetes status.

When the participants’ diabetes status was compared with their exercise habits (at least 30 min of exercise per day during work or leisure time), it was statistically significant that individuals in the control group exercised more (p = 0.010). The participants in the control group had a significantly lower prevalence of a family history of diabetes, whereas those in the diabetes group had a significantly greater prevalence (p = 0.003). The participants in the diabetes group had a significantly lower prevalence of low-risk FRS and a significantly greater prevalence of moderate- and high-risk scores (p < 0.001).

The comparison of matched variables between the control, prediabetes, and diabetes groups is presented in Table 1. No statistically significant differences were found between the diabetes, prediabetes, and healthy groups in terms of age, BMI, or sex. When WC, systolic blood pressure, and WBC (white blood cell) values were compared according to diabetes status, the control group was found to have significantly lower WC (p = 0.019), systolic blood pressure (p = 0.003), and WBC values (p = 0.042) than the diabetes group. When HbA1c and fasting blood glucose levels were compared according to the presence of diabetes in individuals, statistically significant differences were detected between all groups in post hoc analyses (p < 0.001). HOMA-IR values were greater in diabetic patients than in control patients (p = 0.023), HDL values were significantly greater in the control group than in the diabetic group (p = 0.044), and TG values were significantly greater in the control group than in the prediabetic group (p = 0.042). The cholesterol and LDL levels in the diabetes group were significantly lower than those in the prediabetes group (p = 0.013) (Table 1).

Table 1.

Comparison of some measurements and blood parameters of participants with diabetes Status.

Feature	Healthy (n = 29)	Prediabetes (n = 30)	Diabetes (n = 30)	p
WC (cm)	90,00 (83,50–97,00)	95,00 (90,75–99,00)	97,00 (93,50–99,00)	0,0191
Systolic blood pressure (mmHg)	115,10 ± 10,45	124,80 ± 14,73	127,30 ± 16,40	0,0032
Diastolic blood pressure (mmHg)	75,75 ± 9,88	80,06 ± 9,97	79,16 ± 10,69	0,2382
Adropin (ng/L)	135,60 (102,91–251,97)	110,49 (85,30–156,22)	127,95 (100,96–189,14)	0,1551
Irisin (ng/ml)	5,21 (4,01–13,49)	5,39 (3,40 − 8,17)	6,18 (5,07–8,73)	0,2851
Asprosin (ng/ml)	8,28 (5,50 − 19,40)	10,01 (7,61 − 20,07)	12,05 (7,70 − 18,05)	0,4001
HGB (g/dL)	14,30 ± 1,30	14,40 ± 1,58	14,42 ± 1,56	0,9522
HCT (%)	43,54 ± 3,28	43,98 ± 4,52	43,76 ± 3,94	0,9162
WBC (×10⁹/L)	6,29 (5,70 − 7,91)	6,52 (5,69 − 7,96)	7,58 (6,26 − 9,32)	0,0421
PLT (×10⁹/L)	265,24 ± 63,89	271,43 ± 73,43	274,26 ± 72,10	0,8802
Glucose (mg/dL)	89,00 (84,50–93,00)	98,00 (92,50–101,00)	131,50 (112,50–148,75)	< 0,0011
TG (mg/dL)	100,00 (73,00–153,00)	143,50 (114,75–196,75)	150,50 (84,25–247,00)	0,0271
Cholesterol (mg/dL)	196,00 ± 34,33	217,13 ± 37,61	191,93 ± 31,72	0,0132
LDL (mg/dL)	113,92 ± 28,78	132,10 ± 39,81	107,96 ± 26,36	0,0132
HDL (mg/dL)	55,00 (45,00–65,00)	52,00 (45,75 − 56,20)	46,50 (39,25–52,75)	0,0441
Insulin (mIU/L)	8,74 (5,72 − 15,80)	11,85 (7,29 − 17,07)	9,95 (6,12–14,47)	0,2441
HOMA-IR	1,71 (1,31 − 3,24)	2,92 (1,58 − 4,24)	3,03 (2,08 − 4,38)	0,0231
HbA1c (%)	5,40 (5,20 − 5,50)	5,80 (5,70 − 5,90)	6,95 (6,57 − 7,10)	< 0,0011

¹: Kruskal-Wallis test was used. Median (Q1-Q3) was given.

²: ANOVA test was used. Mean ± SD was given. Data derived from the first author’s Specialization in Medicine thesis⁹

(WC: Waist Circumference, HGB: Hemoglobin, HCT: Hematocrit, PLT: Platelet Count TG: Triglycerides, LDL: Low-Density Lipoprotein Cholesterol, HDL: High-Density Lipoprotein Cholesterol, HOMA-IR: Homeostatic Model Assessment of Insulin Resistance, HbA1c: Glycated Hemoglobin).

When comparing adropin, irisin, and asprosin tertiles according to patients’ diabetes status, no statistically significant differences were detected (Table 2).

Table 2.

Comparison of Adropin, Irisin, and Asprosin tertiles according to participants’ diabetes Status. Data derived from the first author’s Specialization in Medicine thesis⁹

Feature		Healthy n (%)	Prediabetes n (%)	Diabetes n (%)	p
Adropin (ng/L)	T1 T2 T3	9 (31,0) 7 (24,1) 13 (44,8)	14 (46,7) 9 (30,0) 7 (23,3)	7 (23,3) 13 (43,3) 10 (33,3)	0,178
Irisin (ng/ml)	T1 T2 T3	12 (41,4) 7 (24,1) 10 (34,5)	11 (36,7) 10 (33,3) 9 (30,0)	7 (23,3) 12 (40,0) 11 (36,7)	0,576
Asprosin (ng/ml)	T1 T2 T3	14 (48,3) 6 (20,7) 9 (31,0)	8 (26,7) 14 (46,7) 8 (26,7)	8 (26,7) 9 (30,0) 13 (43,3)	0,127

In terms of adropin tertiles, WC and BMI values were significantly greater in the first tertile than in the third tertile (p = 0.029; p = 0.024). No statistically significant differences were found when irisin tertiles were compared. According to the asprosin tertiles, WC was significantly greater in the first tertile than in the third tertile (p = 0.021) (Table 3).

Table 3.

Comparison of cardiovascular Risk-Related parameters according to Adropin, Irisin, and Asprosin Tertiles.

Feature	T1 (n = 30)	T2 (n = 29)	T3 (n = 30)	p
ADROPIN (ng/mL)	85,28 (59,86–96,26)	125,97 (112,31–132,55)	342,20 (181,93–746,48)
Age (years)	50,00 (46,75 − 56,00)	49,00 (43,50–54,50)	49,00 (39,00–54,25)	0,2691
WC (cm)	97,00 (90,50–99,00)	95,00 (90,00–99,00)	92,00 (86,25–97,00)	0,0291
BMI (kg/m 2 )	28,55 (26,80 − 29,45)	27,80 (25,45 − 29,40)	26,10 (23,95 − 28,52)	0,0241
Systolic blood pressure (mmHg)	123,93 ± 16,20	120,34 ± 14,20	123,10 ± 14,53	0,6332
HbA1c (%)	5,75 (5,50 − 6,20)	5,90 (5,60 − 6,90)	5,70 (5,40 − 6,82)	0,3751
FRS	4,50 (1,75 − 7,00)	7,00 (0,50 − 8,50)	3,00 (−2,00–10,00)	0,4491
IRISIN (ng/ml)	3,89 (1,91 − 4,48)	5,77 (5,22 − 6,23)	13,97 (8,46 − 44,49)
Age (year)	50,00 (48,00–56,00)	48,00 (41,00–55,00)	49,50 (39,00–55,00)	0,2011
WC (cm)	96,00 (88,75–98,25)	98,00 (91,00–99,00)	92,00 (88,75–97,00)	0,0581
BMI (kg/m 2 )	28,15 (26,07–29,05)	28,37 (25,25–29,60)	26,40 (23,95 − 28,70)	0,1591
Systolic blood pressure (mmHg)	122,46 ± 13,19	123,34 ± 15,63	121,66 ± 16,25	0,9132
HbA1c (%)	5,60 (5,50 − 6,20)	6,00 (5,60 − 6,90)	5,70 (5,40 − 6,70)	0,2561
FRS	5,00 (1,00–7,25)	4,00 (1,00–7,50)	4,00 (−2,00–9,25)	0,8771
ASPROSIN (ng/ml)	5,66 (2,96 − 6,57)	9,80 (8,72 − 11,18)	30,68 (17,03–54,22)
Age (year)	49,00 (43,75 − 56,00)	50,00 (46,50–55,00)	49,00 (39,00–56,00)	0,7151
WC (cm)	97,50 (92,75–99,25)	93,00 (88,50–99,00)	93,00 (88,50–97,00)	0,0211
BMI (kg/m 2 )	28,30 (26,07–29,12)	27,60 (25,25–29,45)	26,60 (23,95 − 29,17)	0,4731
Systolic blood pressure (mmHg)	122,46 ± 13,19	123,34 ± 15,63	121,66 ± 16,25	0,9132
HbA1c (%)	5,60 (5,40 − 6,12)	5,90 (5,60 − 6,75)	5,80 (5,47 − 6,90)	0,0951
FRS	3,00 (0,00–7,00)	6,00 (1,50 − 8,50)	4,50 (−1,25 − 10,00)	0,4361

¹: Kruskal-Wallis test was used. Median (Q1-Q3) was given.

²: ANOVA test was used. Mean ± SD was given. Data derived from the first author’s Specialization in Medicine thesis⁹

(WC: Waist Circumference, BMI: Body Mass Index, HbA1c: Glycated Hemoglobin, FRS: Framingham Risk Score).

The FRS was positively correlated with age (r = 0.560, p < 0.001), BMI (r = 0.301, p = 0.004), WC (r = 0.267, p = 0.011), systolic blood pressure (r = 0.626, p < 0.001), diastolic blood pressure (r = 0.294, p = 0.005), glucose (r = 0.607, p < 0.001), TG (r = 0.371, p < 0.001), HOMA-IR (r = 0.235, p = 0.026), and HbA1c (r = 0.665, p < 0.001). The FRS also was negatively correlated with HDL (r=−0.382, p < 0.001). No statistically significant correlation was found between adropin, irisin, or asprosin levels and the FRS (p > 0.05).

In the control group, adropin was correlated with age (r=−0.460, p = 0.012), WC (r=−0.439, p = 0.017), diastolic blood pressure (r=−0.529, p = 0.003), and the FRS (r=−0.500, p = 0.006). In the prediabetes group, a positive correlation was found between adropin and diastolic blood pressure (r = 0.434, p = 0.017). No significant associations between adropin and any other parameters were found in the diabetes group. In the healthy group, correlations were found between irisin and age (r=−0.427, p = 0.021), systolic blood pressure (r=−0.371, p = 0.048), diastolic blood pressure (r=−0.567, p = 0.001), and FRS (r=−0.509, p = 0.005). No significant relationships were found in the prediabetes and diabetes groups. In the healthy group, asprosin was correlated with age (r=−0.424, p = 0.022), BMI (r=−0.388, p = 0.037), WC (r=−0.586, p = 0.001), systolic blood pressure (r=−0.493, p = 0.007), diastolic blood pressure (r=−0.489, p = 0.007), and the FRS (r=−0.631, p < 0.001). In the prediabetes group, asprosin levels were positively correlated with systolic blood pressure (r = 0.438, p = 0.016) and diastolic blood pressure (r = 0.560, p = 0.001). No significant relationships were found between asprosin and other parameters in the diabetes group (Supplementary Table 1).

When FRS levels were compared with adropin, asprosin, and irisin levels, no statistically significant difference was found (adropin: p = 0.803; irisin: p = 0.793; asprosin: p = 0.545). Additionally, unadjusted logistic regression analysis conducted to investigate the predictive power of serum adropin, asprosin, and irisin levels in individuals with a FRS indicating a CVD risk of less than 10% and those with a risk of 10% or higher demonstrated that these parameters were not significant in predicting high CVD risk. (Adropin: OR = 1.00, p = 0.633; Asprosin: OR = 0.995, p = 0.643; Irisin: OR = 0.998, p = 0.878).

Stepwise logistic regression analyses were performed to evaluate the associations between serum adropin, irisin, and asprosin levels and type 2 diabetes. Model 1 was unadjusted. Model 2 was adjusted for waist circumference, systolic and diastolic blood pressure, and Framingham risk score. Model 3 additionally included HOMA-IR, and Model 4 was further adjusted for lipid parameters (LDL, triglycerides, and total cholesterol). Across all models, no statistically significant associations were observed between serum adropin, irisin, or asprosin levels and type 2 diabetes (p > 0.05) (Table 4).

Table 4.

A Stepwise logistic regression model for evaluating the relationship between Adropin, Irisin, and Asprosin levels and type 2 Diabetes.

Model	B	p	OR	95% Confidence Interval
ADROPIN
Model 1	−0,002	0,173	0,998	0,996-1,001
Model 2	−0,001	0,665	0,999	0,995-1,003
Model 3	−0,001	0,781	0,999	0,996-1,003
Model 4	0,000	0,844	1,000	0,996-1,005
IRISIN
Model 1	−0,007	0,657	0,993	0,962-1,025
Model 2	0,002	0,926	1,003	0,951-1,057
Model 3	0,007	0,775	1,008	0,957-1,061
Model 4	0,015	0,612	1,015	0,958-1,076
ASPROSIN
Model 1	−0,007	0,614	0,993	0,967-1,020
Model 2	0,008	0,744	1,008	0,962-1,056
Model 3	0,010	0,660	1,010	0,965-1,057
Model 4	0,018	0,509	1,018	0,965-1,074

Model 1: Crude model.

Model 2: Model 1 + WC, systolic–diastolic blood pressure, FRS.

Model 3: Model 2 + HOMA-IR.

Model 4: Model 3 + LDL, total cholesterol, TG.

(OR: Odds Ratio). Data derived from the first author’s Specialization in Medicine thesis⁹

Linear regression analyses

In the univariate linear regression model created with factors that may affect adropin levels, a one-unit increase in BMI reduced adropin levels by 30.622 units (p = 0.002), a one-unit increase in WC reduced adropin levels by 8.694 units (p = 0.015), an increase of one unit in the FRS reduced adropin levels by 10.700 units (p = 0.008), and an increase of one unit in the HOMA-IR value reduced adropin levels by 24.897 units (p = 0.001). No significant effect of other variables included in the univariate analysis on adropin levels was detected. In the multivariate linear regression model established with the variables found to be significant in the univariate analysis, no variable was found to have a statistically significant effect on adropin levels (p > 0.05).

These findings suggest that the effects of these factors on adropin are complex or interact with other variables. The univariate linear regression model created with factors that may affect irisin levels revealed that a one-unit increase in age decreased irisin levels by 0.494 units (p = 0.035).

No significant effect of other variables included in the univariate analysis on irisin levels was detected. Multivariate analysis could not be performed because of the lack of sufficient significant variables.

In the single-variable linear regression model created with factors that could affect the asprosin value, a one-unit increase in BMI was determined to reduce the asprosin value by 2.243 units (p = 0.041). The other variables included in the univariate analysis had no significant effect on asprosin levels. Since no sufficient significant variables were found for multivariate analysis, this analysis could not be performed (Tables 5 and 6).

Table 5.

Univariate linear regression model created with factors that may affect Adropin, Irisin, and Asprosin value.

Feature	B	p	95% Confidence Interval	R 2
ADROPIN
Prediabetes	−62,895	0,397	−210,585 − 84,795	−0,005
Diabetes	−87,630	0,160	−210,905 − 35,645	0,017
Age (year)	−7,016	0,051	−14,072 − 0,039	0,043
BMI (kg/m 2 )	−30,622	0,002	−50,144- −11,100	0,101
WC (cm)	−8,694	0,042	−17,069- −0,320	0,036
FRS	−10,700	0,020	−19,642- −1,758	0,061
TG (mg/dL)	−0,305	0,322	−0,913-0,304	0,011
HbA1c (%)	−35,583	0,291	−102,216 − 31,049	0,001
HOMA-IR	−24,897	0,043	−48,981- −0,813	0,046
IRISIN
Prediabetes	−0,336	0,947	−10,425-9,754	−0,017
Diabetes	−1,893	0,662	−10,517-6,731	−0,014
Age (year)	−0,494	0,047	−0,981-0,006	0,044
WC (cm)	−0,222	0,459	−0,813-0,370	0,006
HbA1c (%)	−1,064	0,649	−5,700-3,571	−0,009
HOMA-IR	−1,131	0,187	−2,820-0,559	0,009
ASPROSIN
Prediabetes	3,048	0,666	−11,021 − 17,117	−0,014
Diabetes	−2,552	0,619	−12,780-7,677	−0,013
Age (year)	−0,377	0,267	−1,047 − 0,293	0,014
BMI (kg/m 2 )	−2,243	0,019	−4,108-0,377	0,051
WC (cm)	−0,553	0,170	−1,347-0,241	0,010
FRS	−0,662	0,126	−1,514-0,190	0,015
TG (mg/dL)	−0,027	0,340	−0,084 − 0,029	−0,001
HbA1c (%)	2,664	0,116	−0,676-6,004	0,028
HOMA-IR	−1,598	0,167	−3,881-0,684	0,011

(BMI: Body Mass Index, WC: Waist Circumference, FRS: Framingham Risk Score, TG: Triglycerides, HbA1c: Glycated Hemoglobin, HOMA-IR: Homeostatic Model Assessment of Insulin Resistance).Data derived from the first author’s Specialization in Medicine thesis ⁹

Table 6.

Multivariate linear regression model with variables found to be significant in univariate linear regression model created with factors that may influence Adropin Levels.

Feature	B	β	p	95% Confidence Interval
BMI (kg/m 2 )	−24,079	−0,249	0,073	−50,452-2,295
WC (cm)	1,648	0,041	0,767	−9,356 − 12,653
FRS	−7,007	−0,162	0,155	−16,712-2,698
HOMA-IR	−9,997	−0,086	0,455	−36,488 − 16,494

B = 770,016 R = 0,359 R² = 0,087 Durbin-Watson = 0,834 F = 3,103 p = 0,020. Data derived from the first author’s Specialization in Medicine thesis ⁹

(BMI: Body Mass Index, WC: Waist Circumference, FRS: Framingham risk score, HOMA-IR: Homeostatic Model Assessment of Insulin Resistance).

Discussion

Several studies have investigated the serum levels of adropin, asprosin, or irisin in diabetic patients and individuals at cardiovascular risk. However, to the best of our knowledge, no previous study has simultaneously evaluated the potential role of these three metabolism-related biomarkers in relation to CVR risk across diabetic, prediabetic, and healthy populations. This study is noteworthy in that it concurrently examines adropin, asprosin, and irisin levels, explores their interrelationships, and evaluates their predictive value for cardiovascular risk using the FRS. By integrating these biomarkers with a validated cardiovascular risk assessment tool, this study provides a novel methodological and analytical contribution to the existing literature.

One of the most notable findings is the positive and statistically significant correlation among these three biomarkers across all patient groups. Given that adropin and irisin are generally linked to protective mechanisms, whereas asprosin is associated with pathological processes, this unexpected correlation suggests a complex balance or compensatory mechanism that warrants further investigation. This finding highlights the dynamic interactions underpinning the pathophysiology of metabolic diseases, extending beyond the current understanding and opening new avenues for research.

Moreover, the observed significant increases in WC and BMI in individuals within the lowest adropin tertiles provide compelling evidence that adropin may serve as an early marker of obesity-related metabolic risk. Although these novel biomarkers did not demonstrate a direct correlation with the FRS, our results suggest that they reflect subclinical metabolic processes beyond classical risk factors and could serve as valuable targets for future cardiovascular risk assessment research. Thus, our study offers important insights that help resolve previous conflicting findings and fill critical gaps in the field.

In the present study, there were no statistically significant differences in adropin, asprosin, or irisin levels between the participant groups. The median adropin level was highest in the control group, while the highest irisin and asprosin levels were observed in the diabetes group.

In the present study, although serum adropin levels were higher in the control group than in the diabetes group, this difference was not statistically significant. Different findings exist in the literature on this subject. Studies reporting significantly higher serum adropin levels in the control group than in the diabetes group are consistent with our results^17,18. In contrast, a study conducted on rats revealed that serum and tissue adropin levels were significantly higher in diabetic rats than in control rats¹⁹. Another study conducted on rats also indicated that hyperglycemia increased adropin levels²⁰. There are also studies suggesting that serum adropin levels are significantly higher in patients with T2DM and that this may be an adaptive response²¹.

In the present study, irisin levels were lowest in the control group and highest in the diabetes group, but these differences were not statistically significant. The literature contains conflicting results regarding irisin levels. Tang et al. reported results consistent with our findings, with higher circulating irisin levels in the diabetes group than in the control group, but the difference was not statistically significant²². Some studies have reported results that are consistent with our findings²³. In contrast, studies have reported that irisin levels are lower in individuals with T2DM, impaired fasting glucose, and impaired glucose tolerance than in those with normal glucose tolerance^24,25. In one study, irisin levels were not associated with weight changes or lipid levels, whereas insulin and HOMA-IR levels were positively correlated²⁶. In a study conducted with football players and sedentary individuals, it was reported that serum irisin values were significantly greater in football players and that irisin may be associated with bone health and physical activity²⁷.

The reasons for these differences may include factors such as diabetes duration, glycemic control, medications used, and lifestyle. Additionally, the reliability of the ELISA kits used to measure irisin is a factor that should be considered. Albrecht et al. noted that commercial irisin kits may cross-react with nonspecific proteins in serum and that the amount of irisin in human serum may be too low to be detected by antibodies, meaning that irisin kits may not actually measure irisin²⁸. In young adults, short-term intense exercise caused a significant and sudden increase in serum irisin levels, which disappeared 30 min after exercise. However, 6-week or 1-year long-term increases in physical activity did not affect irisin levels in children. In adults, a correlation was found between irisin levels and 2-hour glucose and TG levels after an oral glucose tolerance test (OGTT), but this correlation lost its significance after adjusting for BMI, age, and sex. In children, no correlation was detected between serum irisin levels and metabolic or cardiovascular parameters²⁹.

In the present study, the plasma asprosin levels were highest in the diabetes group and lowest in the control group, but these differences were not statistically significant. Studies have reported that plasma asprosin levels are significantly greater in individuals with impaired glucose regulation and/or newly diagnosed T2DM than in those with normal glucose regulation^30–32. Some of these studies were conducted on newly diagnosed and untreated individuals, while others were conducted on individuals who already had a diagnosis and were undergoing treatment. In the present study, however, the diabetes group was composed of individuals with an existing diagnosis and under treatment, aiming to reflect real-world data and the patient type encountered in clinical practice. Thus, our findings may be more consistent with daily practice. The fact that patients in the diabetes group were receiving treatment may have reduced the differences in biomarker levels. While this limits direct comparability with untreated populations, it increases the relevance of our findings to routine clinical practice.

In the present study, a statistically significant positive moderate association was found between the FRS and age, systolic blood pressure, glucose level, and HbA1c level; a statistically significant positive weak association was found between the FRS and BMI, diastolic blood pressure, WC, TG level, and HOMA-IR level; and a statistically significant negative weak association was found between the FRS and HDL level. Since smoking, alcohol consumption, age, sedentary lifestyle, obesity, unhealthy diet, family history, diabetes, hypertension, and hyperlipidemia are known cardiovascular risk factors, our findings are consistent with the literature.

In the present study, there was no statistically significant difference between FRS levels and adropin, asprosin, or irisin levels. In the correlation analysis with the FRS, adropin and irisin were found to be negatively correlated, whereas asprosin was positively correlated, but these correlations were not statistically significant. The fact that most patients in the diabetes group were taking hypolipidemic and antihypertensive medications, the absence of important factors such as BMI, WC and family history in this scoring, and the evaluation of variables such as smoking and diabetes as present/absent only may have influenced the FRS results. Additionally, the fact that these biomarkers reflect subclinical metabolic processes may explain why a direct relationship cannot be established with classical risk scores. Moreover, cardiovascular risk scores such as the Framingham Risk Score may underestimate true risk in patients with diabetes who are receiving lipid lowering or antihypertensive therapy, as treated risk factor levels may not fully reflect the underlying cardiovascular risk. Therefore, the absence of a significant association between these metabolic biomarkers and FRS may partly reflect the limitations of conventional risk models rather than the lack of a biological relationship.

In our study, adropin, asprosin, and irisin levels were divided into three tertiles (T1: lowest, T2: medium, T3: highest) and compared with parameters associated with cardiovascular risk. In the current study, the finding that individuals with low adropin levels in the lowest tertile had significantly greater WCs and BMIs is particularly striking. This finding is consistent with the study by Zang et al., who reported that participants with low adropin levels had higher BMIs and TG, insulin, and HOMA-IR levels¹⁷. In this context, adropin levels may facilitate the early detection of metabolic risks associated with obesity.

In the present study, no significant associations were found between irisin tertiles and cardiovascular risk parameters. However, the trend toward greater WC and BMI in the highest irisin tertile is consistent with studies reporting a positive association between irisin levels and BMI and adiposity^33,34. Stengel et al. reported that circulating irisin levels were significantly higher in obese individuals with a BMI of 40 kg/m² or higher than in other individuals, whereas irisin levels in individuals with a BMI of 30–40 kg/m² were similar to those of normal-weight individuals³⁴. In this case, the relationship between irisin and BMI may be parabolic. In a review investigating the physiology and pathology of irisin, high irisin levels in studies involving obese or cancer patients did not produce this effect. The reason why high irisin levels do not ameliorate obesity may be due to a decrease in the number or sensitivity of irisin receptors, leading to “irisin resistance,” especially under pathological conditions³⁵. As seen from the conflicting findings in the literature, the relationship between irisin and obesity is complex. The present study provides an important contribution to addressing the conflicting results and filling the knowledge gaps in this field. The absence of statistically significant associations may be related to the relatively limited sample size.

In the present study, WC was significantly greater in individuals in the first tertile of asprosin than in those in the third tertile. In contrast, Zhang et al. reported that individuals in the upper tertile of asprosin had higher parameters such as BMI, WC, HbA1c, HOMA-IR, and fasting blood glucose³¹.

The conflicting findings regarding adropin, asprosin, and irisin in studies can be explained by various factors, such as differences in patient populations, patients’ proximity to glycemic targets, and measurement methods. On the other hand, personal factors such as physical activity level, sleep pattern, emotional stress level, and dietary habits can affect the serum levels of these proteins. One study indicated that even short-term nutritional status can affect plasma adropin levels³⁶. Additionally, factors related to study methodology, such as the reliability of ELISA kits and sample size, also play a role in statistical significance. Furthermore, if the relationship between irisin and glucose metabolism is not direct but is mediated by parameters such as obesity, age, or BMI, the fact that age, sex, and BMI were matched between groups in the current study may have influenced the results. On the other hand, exclusion criteria such as smoking and alcohol use or medication use for hypertension or hyperlipidemia were not specified in the current study. While this reflects a more realistic clinical population, it also prevents the exclusion of the effects of regular exercise, dietary habits, sleep patterns, and concomitant treatments. The absence of significant differences in adropin, asprosin, and irisin levels between groups may be related to the biological dynamics of these molecules and the influence of numerous variables. Additionally, further research is needed on topics such as the elimination mechanisms of these proteins from the body and their filtration processes in the kidneys. Furthermore, considering that T2DM can be influenced by both genetic and environmental factors, it is not surprising that a single marker closely related to metabolism may be insufficient for predicting the disease.

In the present study, the correlations between adropin, irisin, and asprosin levels and various metabolic parameters were examined. In the healthy group, adropin was negatively associated with age, weight, WC, diastolic blood pressure, and FRS and positively associated with irisin and asprosin levels. In the prediabetes group, positive correlations were found between adropin and diastolic blood pressure, irisin, asprosin, and TG. In the diabetes group, a positive correlation was observed between adropin and irisin and asprosin. In the diabetes group, positive correlations were found between adropin and age, weight, WC, diastolic blood pressure, FRS, and HbA1c levels, but these correlations were not statistically significant. Palizban et al. reported findings that were partially consistent with our findings, indicating a positive correlation between adropin levels and HbA1c in patients with T2DM but no correlation with TG, HDL, LDL, cholesterol, or fasting blood glucose²¹. The absence of these correlations in the diabetes group, which were present in the healthy group, may be explained by time-dependent pathophysiological mechanisms influenced by factors such as diabetes duration, lipid levels, fat mass, treatment, and glycemic control.

In the present study, there was a negative correlation between irisin levels and age, weight, systolic-diastolic blood pressure, FRS, and TG levels in the healthy group and a positive correlation between irisin levels and adropin, asprosin, and insulin levels. In the prediabetes group, irisin levels were positively correlated with height, adropin, and asprosin levels, whereas in the diabetes group, adropin and asprosin levels were positively correlated with irisin levels. These findings are consistent with a previous study indicating that irisin levels are negatively correlated with height, weight, WC and hip circumference and positively correlated with cholesterol, LDL, HDL, and diastolic blood pressure percentiles. In this study, after adjusting for age, sex, and BMI, only HDL was found to be positively correlated with irisin³⁷. In Rana and colleagues’ study, plasma irisin levels in patients with T2DM were positively correlated with BMI, total fat percentage, HbA1c, and e-selectin and negatively correlated with visceral fat score, age, and leptin levels²³. Duran et al. reported that irisin was negatively correlated with BMI, fasting blood glucose, LDL, and TG and positively correlated with HDL across all study groups. In this study, after adjusting for age and BMI, the correlations with fasting blood glucose, LDL, and TG levels remained significant²⁵. In another study, the direction of correlations between irisin and metabolic parameters varied between groups of healthy individuals and diabetic individuals³⁸. The findings of the present study are generally consistent with the literature, suggesting that irisin shows more pronounced correlations in healthy individuals, whereas these relationships weaken as the disease progresses. Irisin may exhibit protective and metabolism-regulating effects in the early stages, but as glucose metabolism deteriorates, it may lose this effect due to changes in hormonal regulation.

In our study, correlation analyses with asprosin revealed negative correlations with age, height, weight, BMI, WC, systolic-diastolic blood pressure, FRS, HCT (hematocrit), TG, insulin, and HOMA-IR in the healthy group and positive correlations with adropin and irisin levels. In the prediabetes group, as the asprosin level increased, the systolic-diastolic blood pressure, adropin, and irisin levels also increased, whereas in the diabetes group, as the asprosin level increased, the adropin and irisin levels also increased. In the study by Naiemian et al., serum asprosin concentrations in the control group were positively correlated with BMI and fasting plasma glucose levels, whereas in the diabetes group, they were positively correlated with BMI, fasting blood glucose, HbA1c, HOMA-IR, and TG levels³². In Zhang and colleagues’ study, asprosin levels increased in all participants, along with fasting blood glucose and HbA1c, and after adjusting for age, sex, and BMI, a negative correlation with HOMA-IR was found. However, when the normal glucose tolerance and diabetes groups were analyzed separately, no associations were found in the diabetes group, and only a positive association with HbA1c was detected in the normal glucose tolerance group³⁹.

In our study, the significant positive correlation between asprosin, adropin, and irisin levels in all participant groups was quite striking. This unexpected positive correlation could indicate the presence of a complex balance or compensatory mechanism between these proteins that is not yet fully understood. Adropin, asprosin, and irisin are involved in metabolic processes and glucose regulation. Adropin and irisin are generally believed to have insulin-sensitizing and anti-inflammatory effects, whereas asprosin is associated with inflammation, hyperglycemia, and insulin resistance^40–43. These findings suggest that multiple and potentially opposing pathophysiological mechanisms may be simultaneously activated in our patient population. One possibility is that adropin and irisin may increase together to compensate for impaired glucose homeostasis associated with elevated asprosin levels. Additionally, previous studies have reported that the regulatory roles of these proteins may vary across different patient groups, such as those with severe obesity or newly diagnosed diabetes. Therefore, the relationships among these biomarkers may dynamically differ depending on disease stage and underlying metabolic conditions.

In conclusion, this study is the first to evaluate the associations of adropin, asprosin, and irisin with cardiovascular risk in diabetic, prediabetic, and healthy individuals via the FRS. Our findings revealed that these three biomarkers are positively correlated with each other and that adropin, in particular, may serve as a potential early marker for obesity-related metabolic risk. These findings suggest that next-generation biomarkers may reflect subclinical metabolic processes and could represent potential targets for future cardiovascular risk assessment. However, these conclusions remain hypothesis generating, and their verification will require further mechanistic and longitudinal studies.

Study limitations

Our study is cross-sectional and may not fully capture the dynamic relationships between these proteins over time. The sample size may have limited statistical power, particularly in subgroup analyses; therefore, regression findings should be interpreted as exploratory. Ongoing literature discussions regarding the reliability and specificity of the ELISA kits used for biomarker measurements should be considered when interpreting our findings. Additionally, the absence of certain exclusion criteria, such as smoking, alcohol use, or medication use for hypertension or hyperlipidemia, has prevented us from fully controlling for the potential confounding effects of these factors on biomarker levels. In addition, cardiovascular risk was assessed using the Framingham Risk Score, which has known limitations in individuals with diabetes and in those receiving cardiometabolic therapies.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

E.K.G. and B.Ü.Ü. conceptualized and designed the study. E.K.G. was responsible for clinical data collection and patient recruitment. E.K.G. and B.Ü.Ü. interpreted the clinical findings and laboratory data and performed statistical analyses. B.Ü.Ü. supervised the study. E.K.G. drafted the manuscript. A.Ü. and İ.A. performed the biochemical analyses, including ELISA assays. All authors critically reviewed and approved the final version of the manuscript.

Funding

This research was supported by the Scientific Research Projects Coordination of Selçuk University under project number 24122020.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to privacy restrictions but are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The study was approved by the Local Ethics Committee of Selçuk University Faculty of Medicine (2024/252).

Consent to participate

Informed consent was obtained from all individual participants included in the study.