Determinants of estimated glomerular filtration rate in patients with type 2 diabetes mellitus: A cross‑sectional study

Introduction

Diabetes mellitus (DM) is a metabolic disorder characterized by chronic hyperglycemia and impaired glucose metabolism, affecting >10% of the global population. Furthermore, >75% of cases occur in low- and middle-income countries, where it poses a notable public health challenge and accounts for ~11.3% of deaths among individuals aged 20-79 years (1). Type 1 DM (T1DM) arises from immune-mediated destruction of pancreatic β cells leading to absolute insulin deficiency, and accounts for 5-10% of all DM cases. On the other hand, type 2 DM (T2DM), which accounts for 90-95% of all DM cases, arises from insulin resistance in peripheral tissue, hyperinsulinemia and later development of β cell failure and relative insulin insufficiency (2). The rise in T2DM prevalence is responsible for the overall increase in DM cases worldwide, with global diabetes prevalence doubling from 7% in 1990 to 14% in 2022 (3,4). Notably, developing countries are experiencing a notable rise in the prevalence of diabetes and its complications, with ~81% of adults with diabetes residing in low- and middle-income countries as of 2024, where prevalence has risen more rapidly than in high-income nations, reaching around 20% in adults across South-East Asia and Eastern Mediterranean regions (3,4). Consequently, it is crucial to implement effective prevention and management strategies in these regions to mitigate the escalating health and economic burdens associated with this disease.

Chronic hyperglycemia, a disordered endocrine profile, systemic low-grade inflammation and development of micro- and macrovascular complications are major hallmarks of T2DM (5). Deterioration of kidney function, and subsequent development of diabetic kidney disease (DKD) and end-stage renal disease, are prevalent and serious repercussions originating from inadequately managed T2DM (6).

The glomerular filtration rate (GFR) is a key indicator of the filtering capacity of the kidney, glomerular health and kidney function, particularly in diabetic patients. Therefore, assessment of the GFR is key for detecting, managing and predicting the prognosis of chronic kidney disease (CKD) in diabetes (7,8). The estimated (e)GFR, based on serum creatinine levels, sex and ethnicity, is routinely used to assess kidney function in clinical practice, since direct measurement of the GFR using renal clearance of different substances, such as inulin, is costly and time-consuming (9).

Numerous factors promote GFR decline in diabetic patients through induction of glomerular injury and tubulointerstitial damage, leading to a cascade of metabolic and hemodynamic disturbances that compromise renal function (10). Key clinical factors implicated in this deterioration include poor glycemic control, insulin resistance, elevated blood pressure, dyslipidemia and aberrant levels of growth factors, such as epidermal growth factor (EGF) (11). In Jordan, where the prevalence of diabetes ranks among the highest globally, data on the clinical predictors of kidney function decline remain scarce, highlighting the need for further investigation (12,13). Therefore, the present study aimed to identify the clinical factors associated with changes in eGFR in diabetic patients, with the goal of informing more effective monitoring and management strategies for this high-risk group.

Materials and methods

Study design and participants

The present study was a retrospective, cross-sectional secondary analysis based on data originally collected for patients with T2DM recruited between December 2018 and December 2019 at the Diabetes and Endocrinology Clinics of King Abdullah University Hospital (Irbid, Jordan), a tertiary care center affiliated with Jordan University of Science and Technology (Irbid, Jordan). The study cohort included adults aged 40-90 years, comprising 47.5% males and 52.5% females. The design, recruitment procedures and primary objectives of the original study were as previously described (14,15). The present investigation focused on determinants of the eGFR and associated hematological indices in this diabetic population. All relevant data, including kidney function tests (KFTs) and complete blood count (CBC) were extracted from the electronic medical records of the hospital. The study protocol was approved by the Institutional Review Board of Jordan University of Science and Technology (approval nos. 7/114/2018 and Sep2025/183-55), and written informed consent was obtained from all participants at the time of enrollment.

Eligibility criteria were adults aged ≥18 years with a confirmed diagnosis of T2DM who were receiving metformin monotherapy. Patients were excluded if they had T1DM, were pregnant, had a diagnosis of malignancy, thyroid dysfunction, Cushing's syndrome or hemoglobinopathy, had received a blood transfusion within the past 4 months or were receiving insulin therapy (due to its confounding effect on insulin resistance assessment).

A total of 240 patients were enrolled. Only participants with complete KFT and CBC data were included, yielding a sample of 236 patients. For subgroup comparisons, patients were categorized into chronic kidney disease (CKD) and non-CKD groups according to the Kidney Disease: Improving Global Outcomes (KDIGO) definition, using an eGFR cutoff of <60 vs. ≥60 ml/min/1.73 m² (8).

Data collection and measurements

Data were drawn from a previously compiled dataset (14,15). Demographic and clinical information, including age, sex, height, weight and BMI, was recorded at the time of original enrollment. Laboratory parameters were originally extracted from the electronic medical records of the hospital and included kidney function test (serum creatinine and urea), complete blood count (CBC), serum electrolytes (potassium and sodium), glycated hemoglobin (HbA1c), fasting blood glucose, and fasting insulin levels. The eGFR calculated using the Modification of Diet in Renal Disease (MDRD) equation based on serum creatinine levels (16). CBC values included hemoglobin (Hb), hematocrit, red blood cell count (RBC), mean corpuscular volume (MCV), mean corpuscular Hb (MCH) and red cell distribution width (RDW).

Additional biochemical markers were glycemic parameters, including fasting blood glucose (FBG) levels measured by an enzymatic colorimetric assay, glycated Hb (HbA1c) levels assessed by high-performance liquid chromatography and serum insulin levels measured using an immunoassay. Insulin resistance was estimated using the Homeostatic Model Assessment for Insulin Resistance (HOMA-IR), and calculated as follows: Fasting serum insulin (µU/l) x fasting serum glucose (nmol/l)/22.5. The lipid profile included triglyceride and total cholesterol levels, which were measured using standard automated techniques. Serum EGF levels were quantified using ELISA (14,15). Outliers among continuous variables were identified using z-score and boxplot visualization; values >±3 standard deviations were cross-verified against source records and excluded if confirmed as data-entry errors.

Statistical analysis

Statistical analysis was performed using IBM SPSS Statistics v26 (IBM Corp.). Continuous variables were screened for normality with histograms, Q-Q plots and the Kolmogorov-Smirnov test. Data with skewed distributions were either log-transformed or analyzed with non-parametric tests. All data are presented as the mean ± SD or number (%). To explore univariable associations, Pearson's correlation coefficient was used for normally distributed continuous variables, after confirming linearity and homoscedasticity. Multivariable modeling was performed to identify independent predictors of eGFR. A linear regression framework was employed with eGFR as the dependent variable. Model 1 employed a forced-entry approach in which all covariates were included a priori based on clinical relevance and prior literature (age, sex, metabolic indices, hematological parameters and biochemical markers) (17,18). This strategy was selected over data-driven selection methods to reduce omitted-variable bias, ensure estimates were adjusted for known determinants of renal function and avoid the instability and bias associated with stepwise procedure. Multicollinearity between predictors was assessed using variance inflation factors (VIFs). Variables with VIFs >10 were iteratively removed to minimize collinearity, yielding a reduced and statistically stable Model 2. Multivariable models were adjusted for key demographic, glycemic, lipid and hematological variables to minimize confounding. For comparisons between categorical groups, independent-samples unpaired t-tests were applied for normally distributed variables, while the Mann-Whitney U test was used for non-normally distributed variables. P<0.05 (two-tailed) was considered to indicate a statistically significant difference.

Results

Patient characteristics

The study included 236 patients with T2DM, comprising 52.96% females (n=125) and 47.03% males (n=111). The mean age was 60.67±10.42 years, with a median of 60 years (range, 40-90 years), and the mean BMI was 30.22±5.35 kg/m2, indicating a predominantly obese cohort (Table I). Glycemic control was suboptimal, with a mean FBG of 171.67±64.73 mg/dl and HbA1c of 7.59±1.70%. The mean HOMA-IR was 15.50±16.67, and mean insulin levels were 212.42±196.18 pmol/l.

Table I Baseline demographic, clinical and laboratory characteristics of the study population.

Table I

Baseline demographic, clinical and laboratory characteristics of the study population.

Variable	Value
Sex (%)
Female	125 (52.96)
Male	111 (47.03)
Mean age, years	60.67±10.42
Mean BMI, kg/m²	30.22±5.35
Mean FBG, mg/dl	171.67±64.73
Mean HbA1c, %	7.59±1.70
HOMA-IR	15.50±16.67
Mean insulin, pmol/l	212.42±196.18
Mean Hb, g/dl	12.99±1.99
Mean Hematocrit, %	38.88±5.49
Mean MCV, fl	83.63±7.87
Mean MCH, pg	28.15±3.27
Mean RDW, %	15.19±8.42
Mean triglycerides, mg/dl	166.83±117.14
Mean cholesterol, mg/dl	210.35±56.73
Mean sodium, mmol/l	139.64±3.68
Mean potassium, mmol/l	4.69±0.49
Mean creatinine, µmol/l	99.31±83.03
Mean eGFR, ml/min/1.73m²	80.20±29.49
Mean EGF, pg/ml	130.70±105.06

[i] FBG, fasting blood glucose; HbA₁c, glycated hemoglobin A₁c; HOMA-IR, Homeostatic Model Assessment of Insulin Resistance; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; RDW, red cell distribution width; eGFR, estimated glomerular filtration rate; EGF, epidermal growth factor.

Renal function, as assessed by mean eGFR was 80.20±29.49 ml/min/1.73 m2. The mean serum EGF concentration was 130.70±105.06 pg/ml. Mean hemoglobin level was 12.99±1.99 g/dl and hematocrit 38.88±5.49%, while MCV and MCH averaged 83.63±7.87 fl and 28.15±3.27 pg, respectively. The mean RDW was 15.19±8.42%. Lipid parameters showed a mean triglyceride concentration of 166.83±117.14 mg/dl and total cholesterol of 210.35±56.73 mg/dl. Electrolyte levels were within normal limits, with mean serum sodium of 139.64±3.68 mmol/l and potassium of 4.69±0.49 mmol/l (Table I).

Correlation analysis of factors associated with the eGFR

Correlation analysis revealed significant associations between the eGFR and multiple demographic, clinical and laboratory parameters (Table II). Age exhibited the strongest negative correlation with the eGFR (r, -0.393), indicating a significant decline in kidney function with advancing age. Among electrolytes, potassium levels exhibited a significant negative correlation with the eGFR (r, -0.229), while sodium levels showed no significant association (r, 0.003).

Table II Pearson correlation between clinical variables and eGFR.

Table II

Pearson correlation between clinical variables and eGFR.

A, Demographic
Variable	r	P-value
Age	-0.393	<0.001
BMI	0.043	0.263
B, Metabolic
FBG	-0.097	0.074
HbA1c	-0.073	0.140
Insulin	-0.138	0.020
HOMA-IR	-0.145	0.015
Triglycerides	-0.134	0.023
Total cholesterol	0.057	0.197
EGF	0.172	0.005
C, Hematological
HB	0.297	<0.001
RBC	0.29	<0.001
HCT	0.217	<0.001
MCV	-0.163	0.007
MCH	-0.123	0.033
RDW	-0.046	0.247
D, Electrolytes
Potassium	-0.229	<0.001
Sodium	0.003	0.480

[i] BMI, body mass index; EGF, epidermal growth factor; HOMA-IR, homeostatic model assessment of insulin resistance; FBG, fasting blood glucose; HbA1c, glycated hemoglobin; RBC, red blood cell count; HCT, hematocrit; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; RDW, red cell distribution width.

Insulin resistance measured by HOMA-IR was negatively correlated with the eGFR (r, -0.145), as were insulin levels (r, -0.138). However, neither FBG (r, -0.097) nor HbA1c (r, -0.073) exhibited significant correlations with the eGFR, suggesting that insulin resistance was associated with kidney function decline in this population.

Lipid profile analysis revealed that triglyceride levels were negatively correlated with the eGFR (r, -0.134), while total cholesterol levels exhibited no significant association (r, 0.057). BMI also exhibited no significant correlation with the eGFR (r, 0.043).

Hematological parameters exhibited significant positive correlations with the eGFR, including Hb (r, 0.297), hematocrit (r, 0.217) and RBC (r, 0.29). Conversely, MCH exhibited a significant negative correlation (r, -0.123), as did MCV (r, -0.163).

EGF levels exhibited a significant positive correlation with the eGFR (r, 0.172; P=0.005; Table II), suggesting a potential protective role in maintaining kidney function.

Predictors of the eGFR: Multivariate regression analysis. Model 1 (initial model)

In the initial regression model, several factors were significantly associated with eGFR). Age showed a negative association (β=-0.259), indicating that eGFR decreased with advancing age. Sex was also associated with eGFR (β=-0.189), suggesting that male patients had lower kidney function than female patients after adjustment for other covariates. Mean corpuscular hemoglobin (MCH) demonstrated a negative association with eGFR (β=-0.316). Triglyceride levels were negatively correlated with eGFR (β=-0.157), and serum potassium showed a strong negative association (β=-0.231). In contrast, epidermal growth factor (EGF) was positively associated with eGFR (β=0.123), suggesting a potential protective effect on renal function. However, this model included variables with high VIF values, particularly Hb (VIF, 46.583), hematocrit (VIF, 47.446) and RBC (VIF, 13.612), indicating multicollinearity issues (Table III).

Table III Multiple linear regression analysis (Model 1) identifying determinants of estimated glomerular filtration rate in patients with type 2 diabetes mellitus.

Table III

Multiple linear regression analysis (Model 1) identifying determinants of estimated glomerular filtration rate in patients with type 2 diabetes mellitus.

					95% CI for B
Independent variable	B	SE B	β	P-value	LL	UL	VIF
Constant	329.89	87.572	-	<0.001	157.152	502.467	-
Age	-0.732	0.196	-0.259	<0.001	-1.119	-0.346	1.591
Sex	-11.170	3.814	-0.189	0.004	-18.690	-3.650	1.387
BMI	-0.242	0.326	-0.044	0.458	-0.884	0.400	1.158
FBG	-0.023	0.050	-0.051	0.643	-0.123	0.076	4.047
Hb	8.233	5.556	0.555	0.140	-2.720	19.186	46.583
HbA1c	-0.586	1.400	-0.034	0.676	-3.347	2.175	2.165
HOMA-IR	0.279	0.291	0.158	0.339	-0.295	0.854	8.991
Insulin	-0.041	0.022	-0.272	0.067	-0.084	0.003	7.204
RBC	3.809	9.267	0.083	0.681	-14.462	22.081	13.612
HCT	-1.250	2.032	-0.233	0.539	-5.256	2.756	47.446
MCV	0.270	0.510	0.072	0.598	-0.736	1.276	6.156
RDW	-0.071	0.198	-0.020	0.719	-0.461	0.318	1.056
MCH	-2.844	1.267	-0.316	0.026	-5.341	-0.346	6.564
Triglyceride	-0.040	0.017	-0.157	0.018	-0.072	-0.007	1.436
Cholesterol	-0.003	0.032	-0.005	0.932	-0.065	0.060	1.234
Na+	-0.923	0.486	-0.115	0.059	-1.881	0.036	1.224
K+	-14.033	3.486	-0.231	<0.001	-20.907	-7.160	1.091
EGF	0.035	0.017	0.123	0.039	0.002	0.068	1.171

[i] LL, lower limit; UL, upper limit of the 95% confidence interval; VIF, variance inflation factor; FBG, fasting blood glucose; HbA₁c, glycated hemoglobin A₁c; HOMA-IR, Homeostatic Model Assessment of Insulin Resistance; RBC, red blood cell count; HCT, hematocrit; MCV, mean corpuscular volume; RDW, red cell distribution width; MCH, mean corpuscular hemoglobin; EGF, epidermal growth factor.

Model 2 (refined model). After removing variables with high VIF values, the refined model identified the following significant predictors: Age demonstrated a stronger negative association compared with Model 1 (β=-0.418), indicating a more pronounced decline in eGFR with advancing age. Triglyceride levels continued to show a negative association with eGFR (β=-0.176), consistent with findings from Model 1. Serum potassium levels maintained a strong negative correlation with eGFR (β=-0.236). EGF) levels were positively associated with eGFR (β=0.146), reflecting a consistent potential protective effect on renal function. The refined model eliminated multicollinearity issues while maintaining significant predictors, which aligned with the correlation analysis (Table IV).

Table IV Refined multiple linear regression model (Model 2) showing independent predictors of estimated glomerular filtration rate.

Table IV

Refined multiple linear regression model (Model 2) showing independent predictors of estimated glomerular filtration rate.

					95% CI for B
Independent variable	B	SE B	β	P-value	LL	UL	VIF
Constant	336.798	80.770	-	<0.001	177.565	496.031	-
Age	-1.183	0.184	-0.418	<0.001	-1.545	-0.820	1.260
Sex	-3.785	3.707	-0.064	0.308	-11.093	3.523	1.178
BMI	-0.171	0.343	-0.031	0.619	-0.846	0.505	1.154
FBG	0.003	0.053	0.007	0.951	-0.100	0.107	3.977
HbA1c	-0.777	1.473	-0.045	0.599	-3.680	2.127	2.153
HOMA-IR	0.133	0.305	0.075	0.664	-0.468	0.733	8.849
Insulin	-0.030	0.023	-0.198	0.200	-0.076	0.016	7.099
Triglyceride	-0.044	0.017	-0.176	0.010	-0.078	-0.011	1.389
Cholesterol	0.011	0.033	0.021	0.742	-0.054	0.076	1.212
MCH	-1.569	0.916	-0.174	0.088	-3.375	0.237	3.086
MCV	0.247	0.388	0.066	0.525	-0.518	1.013	3.206
RDW	-0.047	0.208	-0.013	0.823	-0.456	0.363	1.049
Na++	-0.552	0.504	-0.069	0.275	-1.544	0.441	1.181
K+	-14.356	3.668	-0.236	<0.001	-21.588	-7.124	1.087
EGF	0.041	0.017	0.146	0.020	0.007	0.075	1.148

[i] LL, lower limit of the 95% confidence interval; UL, upper limit of the 95% confidence interval; VIF, variance inflation factor; FBG, fasting blood glucose; HbA₁c, glycated hemoglobin A₁c; HOMA-IR, Homeostatic Model Assessment of Insulin Resistance; RBC, red blood cell count; HCT, hematocrit; MCV, mean corpuscular volume; RDW, red cell distribution width; MCH, mean corpuscular hemoglobin; EGF, epidermal growth factor.

Comparison between the CKD and non-CKD groups

Patients were divided into two groups based on the eGFR: i) CKD group (eGFR <60 ml/min/1.73 m2; n=59); and ii) non-CKD group (eGFR ≥60 ml/min/1.73 m2; n=177). Significant differences between groups were observed in eGFR (39.81±15.23 vs. 94.11±19.03 ml/min/1.73 m2), creatinine (181.77±128.76 vs. 70.36±14.65 µmol/l), age (65.47±10.57 vs. 58.82±9.99 years), HOMA-IR (19.59±19.90 vs. 13.68±14.89), triglyceride levels (199.76±135.25 vs. 153.45±106.81 mg/dl), Hb (11.96±2.11 vs. 13.31±1.84 g/dl), hematocrit (35.87±5.83 vs. 39.84±5.06%), RBC (4.2±0.63 vs. 4.8±0.6x106/mm3), MCV (85.49±6.88 vs. 82.96±8.14 fl) and potassium levels (4.85±0.65 vs. 4.63±0.40 mmol/l; Table V). No significant differences in BMI, glucose, insulin, HbA1c and cholesterol levels, MCH, RDW and sodium or EGF levels were observed between the two groups. These data supported the correlation and regression findings, highlighting the importance of age, insulin resistance, serum triglyceride and potassium levels and hematological parameters in relation to kidney function in patients with T2DM.

Table V Clinical and laboratory parameters between CKD and non-CKD groups.

Table V

Clinical and laboratory parameters between CKD and non-CKD groups.

Parameter	Non-CKD (n=177)	CKD (n=59)	P-value
Age, years	58.82±9.99	65.76±10.75	<0.001
BMI, kg/m²	30.25±5.32	29.88±5.18	0.641
eGFR, ml/min/1.73 m²	94.11±19.03	39.81±15.23	<0.001
Creatinine, µmol/l	181.77±128.76	70.36±14.65	<0.001
FBG, mg/dl	168.80±63.65	184.72±70.06	0.106
HbA1c, %	7.49±1.67	7.93±1.73	0.085
Insulin, pmol/l	194.97±178.88	260.86±230.52	0.026
HOMA-IR	13.68±14.89	20.47±20.02	0.007
Cholesterol, mg/dl	210.18±57.64	207.26±52.36	0.732
Triglycerides, mg/dl	153.45±106.81	198.50±132.92	0.009
Hb, g/dl	13.31±1.84	11.99±2.12	<0.001
Hematocrit, %	39.84±5.06	35.93±5.87	<0.001
RBC, x106/mm³	4.80±0.58	4.20±0.63	<0.001
MCV, fl	82.96±8.14	85.71±6.74	0.020
MCH, pg	27.99±3.38	28.60±2.89	0.212
RDW, %	15.29±9.47	14.91±1.79	0.763
Sodium, mmol/l	139.83±3.45	139.25±4.40	0.299
Potassium, mmol/l	4.63±0.40	4.85±0.65	0.003
EGF, pg/ml	134.15±105.18	110.75±100.19	0.136

[i] Data are presented as the mean ± SD. CKD: chronic kidney disease; eGFR, estimated glomerular filtration rate; FBG, fasting blood glucose; HbA₁c, glycated hemoglobin A₁c; HOMA-IR, Homeostatic Model Assessment of Insulin Resistance; RBC, red blood cell count; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; RDW, red cell distribution width; EGF, epidermal growth factor.

Discussion

The present study aimed to assess the influence of key clinical and biochemical factors on the eGFR, an important indicator of kidney function, in patients with T2DM. DM is a pervasive and challenging condition with well-documented implications for renal function as a consequence of prolonged hyperglycemia (2,19,20). Understanding the interplay between these factors is key for tailoring effective interventions and optimizing the overall management of diabetic patients, with a specific focus on preserving renal function. Traditional diagnostic approaches relying solely on serum creatinine levels (for eGFR measurement) and micro- or macroalbuminuria are increasingly recognized as insufficient for detecting early renal dysfunction (21-23). A significant proportion of individuals with diabetes may develop a reduced eGFR without elevated albuminuria, and albuminuria alone exhibits insufficient sensitivity and specificity to serve as a standalone biomarker for CKD progression (24,25). This highlights the multifaceted nature of DKD, and the necessity for more comprehensive screening methodologies that account for metabolic, hematological and hormonal influences on kidney health.

In the present study, ~25% of the studied patients with T2DM had an eGFR <60 ml/min/1.73 m2, indicating CKD according to the KDIGO guidelines (8). Pearson's correlation analysis showed that age, serum triglyceride and K+ levels, HOMA-IR, serum insulin levels, MCV and MCH were negatively correlated with the eGFR. Conversely, serum EGF and Hb levels, hematocrit and RBC exhibited positive correlations with the eGFR. The multivariate linear regression analysis identified four independent predictors of eGFR decline: Advanced age, lower serum EGF and higher serum K+ and triglyceride levels. These findings suggested that these factors may serve key roles in eGFR changes in patients with T2DM, highlighting the need for further investigation into their underlying mechanisms and potential therapeutic interventions.

The prevalence of CKD in the present study aligns closely with regional data from a comprehensive Middle Eastern meta-analysis, which reported a pooled CKD prevalence of 28.96% (95% CI: 19.80-38.11%) among diabetic patients (26). This prevalence is lower than in China (35.5%, 95% CI: 33.7-37.3%) (27) but comparable with that of sub-Saharan Africa (24.7%, 95% CI: 23.6-25.7%) (28). The present study demonstrates consistency with broader regional patterns, where CKD prevalence in diabetic patients ranges from 10.8 to 60.8% across different countries and study methodologies (26).

The present findings align with the established body of literature, confirming the pivotal influence of age on kidney function as estimated by the eGFR (9). The negative unstandardized regression coefficient (B, -1.183) underscores this association, signifying that advancing age is associated with a substantial decrement in the filtering capacity of the kidney. The decline of 1.183 ml/min/1.73 m²/year is consistent with prior population studies worldwide, including both general and diabetic populations (29-32). In Chinese older adults, eGFR declined by 1.06 ml/min/1.73 m²/year in female and 0.91 m;/min/1.73 m²/year in male patients (33). A recent systematic review of 12 longitudinal cohort studies from six countries confirmed that mean decline rates range from -0.24 to -3.60 ml/min/1.73 m²/year across all participants, with rates of -0.37 to -1.07 ml/min/1.73 m²/year in healthy adults (34). However, in diabetic patients, the decline rates are accelerated: The ACCORD study of high-risk diabetic patients demonstrated annual decline rates of 1.52-1.98 ml/min/1.73 m²/year depending on age of diabetes onset and duration (35). Studies in diabetic nephropathy have shown more rapid progression, with decline rates of 2-20 ml/min/1.73 m²/year (median, 12 ml/min/1.73 m²/year) in patients with established KD (19,32). Mechanistically, this decline may be attributed to age-related pathophysiological changes and structural alterations within the glomeruli and tubules, including decreased nephron mass, altered renal blood flow and diminished filtration surface area, with a potential effect of diabetes in accelerating this decline (36). Therefore, vigilant and periodic monitoring is key to promptly detect incipient declines in kidney function, enabling timely therapeutic intervention to mitigate associated risks, particularly in geriatric patients with T2DM. Notably, previous studies identified advanced age as a significant risk factor for DKD and decreased eGFR (37,38). In addition, our previous study revealed a positive association between older age and the prevalence of anemia, a condition associated with CKD (14).

Although male sex was initially associated with a lower eGFR in the unadjusted regression model (model 1; B, -11.170), it was not retained as a significant predictor in the refined model (model 2), suggesting that the association may have been confounded by other associated clinical or biochemical variables. This aligns with epidemiological evidence indicating that male patients with diabetes generally experience a more rapid decline in eGFR compared with female patients (39). Sex-specific differences in CKD progression may stem from a combination of hormonal influences, such as the protective effects of estrogen via nitric oxide modulation, decreased oxidative stress and mitigation of renal fibrosis (40). These complexities underscore the importance of sex-associated factors in the assessment of renal function and highlight the need for further investigation. Clinicians should interpret eGFR values with an awareness of these potential sex-based variations.

The present correlation analysis identified significant negative associations between the eGFR and insulin resistance markers, specifically HOMA-IR (r, -0.145) and serum insulin levels (r, -0.138), while no significant association was observed with HbA1c or FBG levels. These findings suggested that insulin resistance may contribute directly to DKD through mechanisms independent of glycemic control. These findings align with prior evidence showing that insulin resistance is not only associated with albuminuria and eGFR decline via hemodynamic and inflammatory pathways but also serves as an independent predictor of non-albuminuric DKD phenotypes and adverse outcomes (41). Mechanistically, insulin resistance disrupts the vasodilatory effects of insulin on the renal microvasculature by impairing the PI3K/Akt pathway, leading to decreased nitric oxide bioavailability and endothelial dysfunction (42). Concurrently, unopposed activation of the MAPK pathway fosters pro-inflammatory and vasoconstrictive responses, exacerbating glomerular injury (20,43).

Insulin resistance promotes oxidative stress, stimulates mesangial cell proliferation, enhances renal sodium reabsorption and contributes to lipotoxicity through elevated circulating triglyceride levels (44). These combined effects elevate intraglomerular pressure and promote nephron damage, potentially initiating renal impairment even in the absence of marked hyperglycemia (20). The significant difference in HOMA-IR values between the CKD and non-CKD groups (19.59±19.9 vs. 13.68±14.89) further highlighted insulin resistance as an independent and modifiable risk factor for early renal decline. These insights highlight the potential of insulin-sensitizing therapy for renal protection in diabetes.

While our prior study (14) demonstrated a strong association between poor glycemic control and CKD prevalence, the present analysis suggested that insulin resistance may contribute to renal damage through mechanisms partly independent of hyperglycemia. HbA1c was not a significant correlate or predictor of eGFR. Rather than contradicting the established role of hyperglycemia in DKD, this aligns with emerging evidence that underscores the physiological limitations of HbA1c in KD assessment (45). A recent cross-sectional analysis reported no significant correlation between HbA1c and eGFR (r=-0.444), while glycation gap, an alternative glycemic marker, showed a significant negative correlation with eGFR (r=-0.34), highlighting the differential behavior of glycemic indicators in relation to renal function (45).

Numerous factors may explain the absence of an HbA1c-eGFR association. HbA1c reflects glycemia over the preceding 2-3 months, whereas CKD progression is driven by cumulative long-term exposure that a single cross-sectional measurement cannot capture. In addition, HbA1c accuracy is compromised in CKD due to increased red blood cell turnover, shortened erythrocyte lifespan, and altered hemoglobin glycosylation in the uremic milieu (46,47). Moreover, longitudinal data indicate that HbA1c variability, rather than baseline values, better predicts DKD progression, and early DKD is often detected by albuminuria rather than decreased eGFR (48). The observation that insulin resistance significantly correlates with eGFR suggests that markers of metabolic dysfunction may serve as more clinically relevant early indicators of renal risk than traditional glycemic biomarkers. Collectively, these considerations indicate that the absence of a significant HbA1c-eGFR relationship reflects both physiological constraints of HbA1c and limitations of cross-sectional design, while highlighting insulin resistance as a potentially superior predictor of early renal decline in T2DM. Thus, integrating strategies to reduce insulin resistance alongside traditional glycemic control may be key for effective DKD management. Furthermore, exclusive reliance on HbA1c as a risk indicator may overlook ongoing renal injury in insulin-resistant individuals.

A significant inverse correlation was also observed between triglyceride levels and the eGFR (r, -0.134), with triglyceride levels remaining an independent predictor of renal dysfunction in multivariate analysis (β, -0.176). The 0.044 ml/min/1.73 m² eGFR decline/1 mg/dl triglyceride increase (B, unstandardized regression coefficient) was comparable with a Chinese cohort study that reported 4.93 ml/min/1.73 m² eGFR decrease/1 mmol/l (88.5 mg/dl) triglyceride increase, equivalent to 0.056 ml/min/1.73 m²/1 mg/dl (49). An Italian CKD study demonstrated that each 50 mg/dl triglyceride increase resulted in 6.2% higher risk of eGFR reduction (50). In patients with CKD, the highest triglyceride quartile had 43% higher risk of adverse renal outcomes compared with the lowest quartile (51). The association is particularly pronounced in diabetic patients, where studies showed high triglycerides independently predict KD development with 37-71% increased risk (52,53). This aligns with a previous report linking hypertriglyceridemia to kidney injury via lipotoxicity, inflammation and oxidative stress (44). Lipid accumulation within renal cells provokes inflammation and fibrosis, contributing to functional decline. This association bears particular importance in the Jordanian population given the high prevalence of dyslipidemia. National data show that ~42% of adults have elevated triglyceride levels, and >90% of individuals with T2DM exhibit some form of dyslipidemia (37,54). Here, patients with CKD had significantly higher mean triglyceride levels than those without CKD (199.8±135.3 vs. 153.5±106.8 mg/dl) reinforcing the pathogenic role of hypertriglyceridemia in diabetic nephropathy. These findings support routine assessment and management of lipid profiles, particularly triglycerides, as a modifiable intervention to slow the progression of DKD.

By contrast, total cholesterol levels exhibited no significant association with eGFR (r, 0.057), suggesting limited use as a renal risk marker. This disparity underscores the superior relevance of triglycerides in the context of DKD, a conclusion supported by prior research demonstrating stronger links between hypertriglyceridemia and renal dysfunction than those observed for cholesterol (55). These findings support the growing recognition of triglycerides as a potential modifiable risk factor in the progression of DKD (56,57).

Serum potassium levels exhibited a significant inverse correlation with the eGFR (r,-0.229), and this association persisted as an independent predictor in multivariate regression analysis (β, -0.236). This association is both clinically relevant and pathophysiologically bidirectional (58,59). Progressive renal impairment compromises the capacity of the kidney to excrete potassium, thereby predisposing patients to hyperkalemia. Conversely, elevated serum potassium levels contribute to renal decline through multiple mechanisms, including altered renal hemodynamics, and direct tubular toxicity (60-62). In a meta-analysis of 27 cohorts from multiple countries including the US, Canada, Japan, Sweden and Netherlands involving 1.2 million participants, both hypo- and hyperkalemia were associated with higher risks of adverse renal outcomes across all eGFR levels (58). In American patients undergoing hemodialysis, hyperkalemia >5.0 mEq/l was associated with the greatest decline in residual kidney function (-0.20 clearance decline, 95% CI: -0.50 to -0.06) (63). In a European multi-country study of 1,714 older patients with CKD (stages 4-5), both low and high potassium levels were independently associated with increased risk of death or kidney replacement therapy initiation, with optimal outcomes at 4.5-5.0 mmol/l reference range (64). The difference in potassium concentrations between patients with CKD and those without (4.85±0.65 vs. 4.63±0.40 mmol/l) further underscored the clinical imperative of vigilant electrolyte monitoring in diabetic populations. Hyperkalemia represents a frequent and potentially hazardous complication in DKD, particularly in the context of pharmacological agents such as renin-angiotensin-aldosterone system inhibitors, which are routinely prescribed, yet may exacerbate potassium retention (59,65).

By contrast, serum sodium levels did not exhibit a significant correlation with the eGFR (r, 0.003), suggesting that the sodium balance may remain relatively stable during the early stages of renal function decline in patients with diabetes. Nonetheless, sodium intake and renal handling are important clinical considerations, as they directly influence blood pressure regulation, a key determinant of kidney function (66).

Anemia is a frequent finding in patients with declining kidney function (67). The present study revealed that a higher eGFR was significantly associated with improved hematological profiles, including higher Hb levels (r, 0.297), hematocrit (r, 0.217) and RBC (r, 0.29). These findings are consistent with the physiological role of the kidneys in supporting erythropoiesis through erythropoietin production. Patients with CKD had markedly lower values for these hematological markers compared with patients in the non-CKD group, reinforcing the link between decreased kidney function and impaired red blood cell production due to diminished erythropoietin synthesis (68). MCH (r, -0.123) and MCV (r, -0.163) exhibited significant negative correlations with the eGFR. These findings suggested that changes in red blood cell morphology may occur with declining kidney function, potentially due to alterations in iron metabolism, inflammation or uremic toxins (69,70).

Our previous analysis demonstrated that anemia was significantly more common among patients with uncontrolled T2DM (40 vs. 27.5%), and that CKD was positively associated with anemia in the multivariate model (14). This reinforces the association between declining kidney function (CKD/low eGFR) and the development of anemia.

A noteworthy finding in the present study was the significant positive correlation between serum EGF levels and the eGFR (r, 0.172), which remained an independent predictor in the multivariate regression model (β, 0.146). This association suggests that EGF may serve a protective role in maintaining kidney function in patients with T2DM. These findings align with our previous study, which demonstrated significantly lower serum EGF levels in patients with poorly controlled diabetes (HbA1c >7.0%) compared with those with improved glycemic control (95.9±82.7 vs. 158.77±111.7 pg/ml), and a significant inverse correlation between EGF and HbA1c levels (r, -0.25) (15). In the aforementioned study, reduced EGF levels were predictive of poor glycemic control. Although the present analysis did not show a significant difference in EGF levels between patients with and without CKD, this may reflect the greater sensitivity of continuous vs. categorical analysis in detecting gradual EGF changes across the physiological range in a cohort with relatively preserved kidney function. Continuous correlation analysis preserves the full spectrum of biological variability and can detect subtle associations that binary CKD classification may obscure (71). Since EGF primarily reflects tubular epithelial health and repair capacity, its levels may decline progressively in parallel with gradual loss of regenerative function, rather than following discrete pathological thresholds. These data suggest a complex relationship where poor glycemic control contributes to decreased EGF expression, which in turn impairs tubular repair mechanisms and potentially accelerates the transition from functional compensation to overt kidney function decline.

EGF is known to support tubular epithelial cell regeneration following injury, a process that may be key in mitigating diabetes-related renal damage (72). Other studies have also shown that urinary EGF levels are lower in patients with DKD and are associated with the severity of renal impairment (11,73). By demonstrating a positive association between serum EGF levels and the eGFR, the present findings underscore the potential of EGF as both a biomarker and a therapeutic target in DKD.

However, it is important to acknowledge that the role of EGF in KD is complex and context-dependent. While the present findings support a protective association between serum EGF and eGFR, chronic or excessive EGF receptor (EGFR) activation can paradoxically promote pro-fibrotic pathways. Sustained EGFR signaling is implicated in mesangial cell proliferation, epithelial-to-mesenchymal transition and progressive interstitial fibrosis in advanced kidney disease (74). The dual nature of EGF signaling suggests that its net renal effect depends on multiple factors including disease stage, duration of exposure, local tissue environment and the balance between regenerative and fibrogenic cellular responses (74). In the present relatively early-stage diabetic cohort with preserved mean eGFR, the positive EGF-eGFR correlation may reflect its beneficial tubular regenerative properties. However, in more advanced KD or different pathological contexts, the same EGF pathways might contribute to maladaptive remodeling and fibrosis (75). This complexity underscores the need for longitudinal studies to understand the temporal dynamics of EGF signaling and its evolving role across different stages of DKD progression.

Region-specific factors may underlie the eGFR predictors. The North Africa and Middle East region has seen a 70.9% rise in CKD incidence since 1990, with kidney dysfunction, hypertension and high BMI as leading risk factors (76). In Jordan, dyslipidemia is highly prevalent: >90% of patients with T2DM are affected, and triglyceride levels in the present patients with CKD were markedly higher than in non-CKD group (77). This aligns with regional patterns shaped by diets rich in saturated fat, sedentary lifestyles and gaps in diabetes care (78). Moreover, as a middle-income country with resource limitations, Jordan faces delayed diagnosis and suboptimal management of complications, which may explain the more advanced metabolic dysfunction compared with high-income settings (79).

The present study has several important limitations. First, it was a secondary analysis of a previously recruited cohort with specific inclusion and exclusion criteria. Only adults with T2DM receiving metformin monotherapy were included, while insulin users and patients with certain comorbidities were excluded. These criteria were intended to reduce heterogeneity and confounding, but limit the generalizability of results to the broader T2DM population, where polytherapy with other oral agents or insulin is common, particularly in patients with longer disease duration or more advanced disease. Thus, the present findings may best represent earlier or less complex stages of T2DM. Second, the duration of diabetes, a critical determinant of microvascular complications such as CKD, was not reliably available. This reflects a common limitation in the Jordanian healthcare context, where delayed diagnosis and incomplete records are frequent (79). The absence of this information restricts the ability to assess the chronicity of disease burden and its interaction with renal outcomes. Kidney function was assessed solely by eGFR estimated using the MDRD equation. While MDRD has been widely applied in Middle Eastern diabetic cohorts and allows comparison with regional studies (80,81), it may underestimate GFR at higher ranges and is less accurate than CKD Epidemiology Collaboration (CKD-EPI) formula, which is recommended by the KDIGO guidelines (82). Nonetheless, MDRD has been shown to perform reasonably well in diabetic populations, and its use maintains comparability with prior literature (83-85). Future studies incorporating CKD-EPI may provide more precise estimates, particularly in patients with preserved function. The lack of albuminuria data (such as urine albumin-creatinine ratio) precludes full KDIGO-based classification and prevents differentiation between albuminuric and non-albuminuric phenotypes. This limits the specificity of findings to renal filtration impairment rather than the broader spectrum of diabetic kidney disease. Other limitations include the cross-sectional design, which precludes causal inference, the use of a single serum EGF measurement, which may not capture temporal variability, and limited information on additional medications, comorbidities or lifestyle factors that could act as residual confounders. While the present findings provide evidence on determinants of eGFR in patients with T2DM, they should be interpreted with caution. Longitudinal studies with broader inclusion criteria, detailed clinical histories and comprehensive KDIGO-based CKD classification are needed to validate and extend the present results.

In conclusion, the present study provides evidence regarding the determinants of kidney function in patients with T2DM in a population with a high burden of both diabetes and CKD. The identification of age, triglycerides, potassium and serum EGF as independent correlates of eGFR decline is consistent with findings from previous studies (34,52,61,73), while also reinforcing their relevance within a Middle Eastern cohort. Clinically, the results emphasize the importance of addressing modifiable metabolic risk factors, particularly insulin resistance and hypertriglyceridemia, through lifestyle interventions and, when appropriate, pharmacological strategies such as omega-3 fatty acids or fenofibrate (86,87). At the same time, vigilant potassium monitoring and age-stratified kidney function surveillance are key in resource-limited settings, where delayed diagnosis and restricted access to nephrology care heighten the risk of CKD progression (88,89).

To the best of our knowledge, the present study is the first to examine serum EGF in relation to renal function in T2DM. Previous research has predominantly focused on urinary EGF, where lower levels consistently predict DKD progression and tubular injury (90,91). While urinary EGF may reflect tubular regenerative capacity, serum EGF may capture systemic growth factor signaling or impaired tubular release. Future mechanistic and longitudinal studies are needed to clarify these pathways, evaluate the complementary prognostic role of serum vs. urinary EGF and determine whether serum EGF could be leveraged as a practical biomarker in resource-limited settings where timed urine collections may not be feasible.

The associations between eGFR and anemia parameters further support the need for systematic hematological evaluation as part of routine diabetes care. Although the association between EGF and eGFR warrants further validation, the present findings complement emerging evidence suggesting that growth factor pathways may influence renal outcomes (91,92).

By demonstrating patterns consistent with those reported in neighboring countries, the present study provides regional data that may inform coordinated healthcare strategies to mitigate the rising burden of DKD in the Middle East, providing a foundation for targeted preventive strategies and integrated nephrology-endocrinology care models in high-risk populations.

Acknowledgements

Not applicable.

Funding

Funding: The present study was supported by the Deanship of Research at Jordan University of Science and Technology (grant nos. 20180162 and 20250455).

Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

Authors' contributions

AAD designed the study, analyzed data and wrote the manuscript. OA and MA designed the study and analyzed data. DGA designed the study and interpreted data. All authors have read and approved the final manuscript. AAD and OA confirm the authenticity of all the raw data.

Ethic approval and consent to participate

Informed written consent was obtained from all participants included in the study. All procedures involving human participants were approved by Jordan University of Science and Technology and King Abdullah University Hospital Institutional Review Board (approval nos. 7/114/2018 and Sep2025/183-55; Irbid Jordan), in accordance with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Use of artificial intelligence tools

During the preparation of this work, artificial intelligence tools were used to improve the readability and language of the manuscript, and subsequently, the authors revised and edited the content produced by the artificial intelligence tools as necessary, taking full responsibility for the ultimate content of the present manuscript.

References