Type 2 diabetes (T2DM) is a highly prevalent disease with a global social and economic burden. Its accompanied vascular complications represent a major cause of morbidity and mortality worldwide. T2DM is an expanding disease with rapidly increasing incidence due to the obesity pandemic which is closely related to T2DM [1].

This link between obesity and T2DM is mediated through insulin resistance (IR). Visceral adipose tissue is considered an endocrinal organ secreting various types of adipokines. These adipokines play a pivotal role in the development of IR and subclinical inflammations which represent the main mechanism of T2DM [2][3].

Early detection of IR could help in improving patients' outcome and delaying the development of chronic complications. Additionally, IR is the most powerful predictor of T2DM development and also a major therapeutic target in treatment of T2DM. However, measurement of IR faces some difficulties [4].

The gold standard for measurement of IR is the hyperinsulinemic euglycemic clamp technique which is technically demanding and time consuming. Numerous surrogate marker of IR have been emerged but none of them proved to be an ideal biomarker. Homeostasis model assessment for insulin resistance (HOMA-IR) is the most commonly used surrogate marker of IR. However, no standard cut off value could be generalized [5][6].

Serum c-peptide level is a gold standard measurement of endogenous insulin secretion. It has been found that serum c-peptide is strongly correlated to urinary c-peptide and urinary c-peptide to creatinine ratio (UCPCR). Thus, UCPCR could be used as a valuable non- invasive marker of endogenous insulin secretion and IR in patients with T2DM [7][8]. Recently, low UCPCR was able to identify people with type 1 diabetes (T1DM) with exceptional sensitivity and specificity making it a cost effective, non-invasive marker of T1DM [9]. UCPCR was also identified as an independent risk factor of coronary artery disease in this cohort [10].

The present work aimed at studying the relation between UCPCR, IR and DVC in subjects with T2DM.

Study site. Included subjects were recruited from diabetes outpatient clinic of Alexandria Main University Hospital, Alexandria, Egypt.

Time of the study. The present study was performed between March 2023 till September 2023.

Population: The present study included 90 subjects with T2DM.

Inclusion criteria: Males or females, diagnosed with T2DM, above the age of 18 years.

Exclusion criteria: Severe renal impairment (eGFR <15 ml/min/1.73), subjects with urinary tract infection, pregnant and lactating females.

Sampling was done randomly.

The present study is a single-center, observational, cross sectional study including one study population.

Based on the study of Wang et al, [8] a minimal required sample size of 36 patients with T2DM is needed to assess UCPCR in patients with T2DM and its relation to IR and vascular complications that detect a difference of .3 between the null hypothesis correlation of 0.5 and the alternative hypothesis correlation of 0.78, using One Correlation Power Analysis that achieves 80% power with a target significance level at 5%. Sample size was calculated using NCSS 2004 and PASS 2000 program.

A detailed history was taken from each participant including age, diabetes duration, smoking and drug history. Physical examination was performed including measuring blood pressure, waist circumference, weight and height from which body mass index (BMI) was calculated [11][12].

Detailed neurological examination was done and diabetic peripheral neuropathy (DPN) was diagnosed as per American Diabetes Association (ADA) recommendations [13].

Peripheral arterial disease (PAD) was diagnosed by calculating ankle brachial index (ABI). Measuring systolic pressure of dorsalis pedis (DP)and posterior tibial (PT) arteries in each lower limb was done with Doppler examination and sphygmomanometer cuff. Calculation of the ABI was done by dividing the higher pressure of the DP and the PT arteries by the higher systolic pressure of the two brachial arteries. The lower value of the two calculated ABI values was used for analysis [14].

Fundus examination was done for each subject using slit lamp biomicroscope plus fundus lens by expert ophthalmologist in the ophthalmology outpatient clinic to diagnose diabetic retinopathy (DR).

Laboratory investigations were done after overnight fasting for measurement of: fasting plasma glucose (FPG), fasting insulin and serum creatinine. A spot urine sample was collected for measurement of urinary albumin to creatinine ratio (UACR) and UCPCR.

IR was calculated from FPG and fasting insulin by measuring HOMA-IR using the equation (HOMA-IR = fasting glucose (mg/dL) X fasting insulin (mU/L) / 405) [15].

Estimated glomerular filtration rate was calculated from serum creatinine using CKD-EPI equation. Diabetic kidney disease (DKD) was diagnosed by the presence of albuminuria (UACR≥30 mg/g) and/or eGFR <60 ml/min/1.73 m² [16].

UCPCR was measured from a spot urine sample collected in aseptic tube, centrifuged for 20 min at 2.000–3.000 rpm and the supernatant was collected. Human C-peptide ELISA kit was used to assess c-peptide level in the processed urine sample [17].

Data were analyzed using IBM SPSS software package version 20.0 (Armonk, NY: IBM Corp, released in 2011). Qualitative data were described using number and percent. Quantitative data were described using mean and standard deviation. Significance of the obtained results was judged at the 5% level. Mann Whitney test was used for analysis of non-normally distributed quantitative variables, to compare between two studied groups. Spearman coefficient was used to correlate between two quantitative variables. Regression analysis was done to detect the most independent factor for IR. Receiver operating characteristic curve (ROC) was generated by plotting sensitivity (TP) on Y axis versus 1-specificity (FP) on X axis at different cut off values. The area under the ROC curve (AUC) denotes the diagnostic performance of the test. Area more than 50% gives acceptable performance and area about 100% is the best performance for the test.

The study was conducted according to the ethical standards of the Helsinki Declaration. Approval of the ethical committee of Faculty of Medicine, Alexandria University was obtained (IRB number 00012089) (date: 16/6/2022). Each participant gave a written informed consent before inclusion in the study.

The present study included 90 subjects with T2DM with mean age of 54.91±8.93 years. Thirty six subjects were males while the rest 54 subjects were females. Baseline characteristics of the studied subjects are shown in table 1. Regarding the distribution of vascular complications among the total sample, 17 patients (18.9%) had DN, 30 patients (33.3%) had DKD and 27 patients (30%) had DR and 44 patients (48.9%) had PAD. There were no statistically significant differences in the mean value of UCPCR between subjects with vascular complications (PVD, DN, DKD, and DR) and those without complications.

There were highly statistically significant positive correlations between UCPCR and HOMA-IR, FPG and HbA1c with P values of <0.001, 0.006 >and 0.005 respectively. However, no significant correlations were detected between UCPCR and the rest of the studied parameters (Table 2) (Figure 1).

On performing the logistic regression analysis for detecting parameters affecting IR (HOMA-IR ≥2.5); FPG, HbA1c, and UCPCR were the independent risk factors for IR in the univariate analysis. However, UCPCR was the only independent risk factor for IR in multivariate analysis (Table 3).

ROC curve was performed to set the cut off value of UCPCR for diagnosis of IR. A cut off value of UCPCR (>0.19) nmol/mmol was able to differentiate significantly (p<0.001) between patients with IR and those without IR with good sensitivity, specificity and AUC (85.11%, 60.47% and 0.716 respectively) (Figure 2).

The study sample is representative to the studied population. Sample size calculation was done.

UCPCR is an emerging promising biomarker for endogenous insulin secretion and IR. The current work uniquely studied this association and also the relation between UCPCR and DVC. Few previous studies investigated these relations and none of them was on Egyptian population.

The present study included 90 subjects with T2DM. The mean UCPCR of the studied subjects was (0.32±0.23 nmol/mmol).

Previous studies [18–20] demonstrated a highly variable range of UCPCR (0.43–2.47 nmol/mmol) in subjects with T2DM and a much lower values in subjects with T1DM. This wide range of UCPCR values across studies could be attributed to several factors including demographic and clinical data. Wang et al,[18] found a higher level of UCPCR in patients with T2DM and monogenic diabetes compared to T1DM. They also divided patients with T2DM into 2 groups regarding their UCPCR and they found that individuals with UCPCR ≤0.20 nmol/mmol had lower BMI, lower serum c-peptide and higher insulin use. Their cohort included higher males (64%) in T2DM group than in our cohort (40%). Besser et al. [19] also differentiated between T2DM, MODY and T1DM using UCPCR. A level of 4.01 (2.84–5.74) nmol/mmol for UCPCR was identified in patients with T2DM. This discrepancy between their results and our results may be due to younger population in Besser et al study (<21 years) than our cohort (54.91±8.93 years). Bhosle et al, [20] concluded that UCPCR can be used to evaluate islets β Cell function in T2DM patients with different renal function status and set a cut-off (UCPCR ≤1.13 nmol/g) for identifying severe insulin deficiency in T2DM patients. Moreover, ethnicity, diabetes duration, age and gender play a crucial role in this variation. Urinary c-peptide excretion declines with increasing age. Furthermore, females have higher values of UCPCR due to lower muscle mass compared to males.

In the current study, the mean value of UCPCR showed a highly statistically significant positive correlation with HOMA-IR in subjects with T2DM. Additionally, UCPCR was the only independent predictor of IR in multivariate regression analysis. HOMA-IR is the most commonly used surrogate marker of IR. Previous studies demonstrated its association with glycemic control, metabolic profile and vascular complications [21].

Wang et al, [8] in concordance with the results of the present study, found a strong positive correlation between UCPCR and HOMA-IR in 1299 hospitalized Chinese subjects with T2DM. Furthermore, Zhou et al, [20] showed a positive correlation between UCPCR and HOMA-IR. Few more studies demonstrated a positive correlation between UCPCR and HOMA-IR in healthy individuals [22][23], prediabetes [22], and metabolic syndrome [24]. However, in Egypt, only two previous studies [25][26] discussed the correlation between UCPCR and HOMA-IR. Both cohorts included children with obesity and both showed positive correlations between UCPCR and HOMA-IR in obese children.

The present study could set a cut-off value of (>0.19) nmol/mmol for UCPCR to predict IR in patients with T2DM. By far, this is the first study to detect a cut-off value of UCPCR for detection of IR in patients with T2DM. Additionally, the present study uniquely studied the relation between UCPCR and IR in Egyptian population with T2DM.

The current work, additionally, found a significant positive correlation between UCPCR and FPG. In addition, FPG was detected as an independent risk factor for IR in univariate but not multivariate analysis. An Indian study by Kulkarni et al [27] showed similar results. They suggested that urinary c-peptide and UCPCR could be used as predictors of endogenous insulin secretion. Thus, they concluded that UCPCR could predict glycemic control in subjects with T2DM. Gedam et al [24] and Reinter et al [22] also reported a significant positive correlation between UCPCR and FPG in studied participants.

Regarding the relation between UCPCR and HbA1c, the present work revealed a significant positive correlation between them. HbA1c also was detected as an independent risk factor for IR in univariate but not multivariate analysis. No previous studies were detected concerning the relation between UCPCR and HbA1c in T2DM. However, Hassan et al [25] found similar results but in obese children with IR. Moreover, Reinter et al [22] reported the same correlation in healthy individuals. However, Elzahar et al [28] showed a significant negative correlation between UCPCR and HbA1c in children with T1DM and T2DM. This discordance may arise from the use of different cohort with different age and type of diabetes.

The present work also studied the relation between UCPCR and DVC. Results revealed that there is no relation between UCPCR and DVC (DKD, DR, DN and PAD). In disagreement with these results Wang et al [8] found that UCPCR is an independent risk factor of DKD in univariate and multivariate analysis. This discordance between their results and ours could be attributed to the small number of subjects in each complication subgroup in the present study. Thus, further studies with larger sample size are encouraged to demonstrate this relation clearly.

UCPCR was highly statistically significantly correlated to IR and glycemic parameters but not DVC in patients with T2DM.

The present study was a single center study. When the studied subjects were divided according to DVC, a small number of subjects was present in each group which resulted in non significant results regarding the relation between UCPCR and DVC.

Further studies are required to assess the relation between UCPCR and DVC. Multicenter studies on larger sample size and different study population are needed to generalize the results of the present study.

The present study demonstrated a strong association between UCPCR and IR in patients with T2DM. This suggests that UCPCR could be used as a simple, noninvasive and available biomarker for IR in T2DM. UCPCR could be a predictor of endogenous insulin secretion and diabetes control as it showed a strong significant correlation with FPG and HbA1c. However, the present study found no relation between UCPCR and VDC which will need further studies with larger sample size to identify this relation.

Funding. No funding.

Conflict of interest. The authors declare no obvious and potential conflicts of interest related to the content of this article.

Contribution of authors. Eman Y Morsy — Substantial contributions to the conception and design of the work; acquisition, analysis, interpretation of data for the work; and Drafting the work and revising it critically for important intellectual content. Noha G Amin — Substantial contributions to the conception and design of the work; acquisition, analysis, interpretation of data for the work; and Drafting the work and revising it critically for important intellectual content. Yasmine A Issa — Substantial contributions to the conception and design of the work; acquisition, analysis, interpretation of data for the work; and Drafting the work and revising it critically for important intellectual content. Rowan KA Khalifa — Substantial contributions to the conception and design of the work; acquisition, analysis, interpretation of data for the work; and Drafting the work and revising it critically for important intellectual content. Heba S Kassab — Substantial contributions to the conception and design of the work; acquisition, analysis, interpretation of data for the work; and Drafting the work and revising it critically for important intellectual content.

All of the authors read and approved the final version of the manuscript before publication, agreed to be responsible for all aspects of the work, implying proper examination and resolution of issues relating to the accuracy or integrity of any part of the work.

Acknowledgment. We would like to acknowledge Diabetes and Metabolism Unit, Internal Medicine Department, Faculty of Medicine, Alexandria University where the work was done.