Correlation Between Cardiometabolic Index and Microalbuminuria in Type 2 Diabetic Patients

Background: In recent years, cardiometabolic index (CMI) has been introduced to predict the risk of diabetes in the general population and is also strongly associated with cardiovascular disease, hypertension, and ischemic stroke, but the relationship between CMI and diabetic microalbuminuria remains unclear. This study intends to investigate the relationship between CMI and microalbuminuria in patients with type 2 diabetes mellitus(T2DM). Methods: 524 subjects (mean age 57.28 ± 10.52 years, 283 males and 241 females) from the Department of Endocrinology, Southwestern Medical University Hospital, China (data collected from June 2017 to June 2019) participated in this study. CMI was calculated by triglyceride to high-density lipoprotein cholesterol ratio(TG/HDL-C) multiply waist-to-height ratio(WHtR). Microalbuminuria was identied by the urinary albumin ‐ to ‐ creatinine ratio (UACR) when the value was between 30–300 mg/g. Multivariate logistic regression aimed to explore the association between CMI and diabetic microalbuminuria; The receiver operating characteristic (ROC) analysis was employed to evaluate the predictive value of CMI for microalbuminuria in T2DM patients. Results: The prevalence of microalbuminuria in all participants was 37.8% in men and 41.1% in women. Regardless of gender, CMI was signicantly more unfavorable in the microalbuminuria group. The incidence of microalbuminuria increased dose-responsively with increasing CMI quartiles. Modeling CMI as a continuous variable in a multivariate-adjusted model, we observed an independent effect of each 1 SD increase in CMI on the risk of developing microalbuminuria. The relationship was more pronounced in women than in men. In women, the odds ratio(OR) for microalbuminuria was 5.666 (95% CI: 2.247-14.289) in the highest quartile of CMI compared with the lowest quartile; in men, the OR was nearly 4-fold increased(OR: 4.667, 95% CI: 1.910-11.405) in the same situation. The AUC for CMI was 0.681 (0.613 to 0.749) in women and 0.648 (0.582-0.713) in men. Conclusion: High levels of CMI were independently associated with diabetic microalbuminuria, and CMI, a novel index covering lipids and central obesity, explained the gender differences in obesity-related microalbuminuria excretion, an effect that was substantially increased in women. The results of this study provide important insights into the potential usefulness and clinical relevance of CMI for microalbuminuria in Chinese patients with T2DM.


Background
Diabetic nephropathy (DN) is one of the most common microvascular complications of diabetes mellitus, the incidence of which has continued to rise in recent years and has become a signi cant hazard in the development of end-stage renal disease in diabetic patients. It is well known that the presence of microalbuminuria is a marker for the diagnosis of early diabetic nephropathy in T2DM. Considering that early detection of microalbuminuria is an essential factor in avoiding progressive renal function deterioration in diabetic patients. Therefore, early detection of microalbuminuria and timely and effective treatment is key to delaying the progression of diabetic nephropathy, preventing end-stage renal disease, and improving the overall prognosis. It has been widely reported in the literature that central obesity and its associated abnormal lipid metabolism in T2DM patients play an essential role in the process of renal damage, causing nonspeci c low-grade in ammation and insulin resistance through the secretion of various in ammatory factors and hormones, leading to renal vascular endothelial dysfunction and the production of albuminuria [1][2]. Dyslipidemia in patients with T2DM is often manifested by high triglyceride (TG) levels and low high-density lipoprotein cholesterol (HDL-C) levels and elevated TG is positively associated with the risk of developing microalbuminuria in T2DM [3]. Also, BMI and WC have been recognized to have a clear relationship with microalbuminuria, but they are less helpful in diagnosing the disease because they do not distinguish between fat and muscle weight or fat distribution [4]. Waist height ratio (WHtR), another anthropometric indicator of central obesity, modi es the relationship between waist circumference and height, is superior to general obesity indicators in predicting the risk of microalbuminuria [5]. Cardiometabolic index(CMI) represents a combination of TG/ HDL-C and WHtR, taking lipid and abdominal obesity indicators into account. As a new and widely recognized metabolic index in recent years, it has been found to predict the risk of diabetes in the general population and is associated with cardiovascular disease, hypertension, and ischemic stroke [6][7][8][9]. In this paper, we investigated the correlation between CMI and microalbuminuria in type 2 diabetes in China and aimed to assess the diagnostic value of CMI on the occurrence of microalbuminuria in T2DM and provide a reference for clinical work. To the best of our knowledge, this has not been investigated.

Study population
A retrospective cross-sectional study was designed and conducted to describe the prevalence and risk factors for developing microalbuminuria in 524 T2DM patients (age ≥20 years) hospitalized in the Department of Endocrinology at the Southwestern Medical University Hospital between June 2017 and June 2019. Complete details of the study design and rationale are extensively described elsewhere. Exclusion criteria for this study were as follows: 1) other speci c types of diabetes; 2) acute complications of diabetes; 3) renal disease due to primary or secondary nephritis, urinary tract infection, acute and chronic renal failure; 4) severe liver and cardiac dysfunction; 5) acute infection; 6) immunode ciency, tumor, use of lipid-regulating drugs within the past 3 months and/or angiotensin-converting enzyme inhibitors (ACEI )/angiotensin receptor antagonist (ARB) within the past 3 months. A total of 524 patients were recruited for this study. The study protocol followed the Second Declaration of Helsinki, and the Ethics Committee of Southwest Medical University (Sichuan) approved the study protocol. We obtained written informed consent from each participant before enrollment, and the entire data and procedures were under the principles of ethical standards.

Data collection and measurements
Questionnaire: Specially trained and tested doctors and nurses at our hospital were allowed to administer the questionnaire. We used a well-designed questionnaire to collect data from all patients. A central steering committee and a subcommittee performed quality control of the information collection process. The questionnaire covered the following data: age, gender, duration of diabetes, history of hypertension and diabetes, smoking and alcohol consumption, and use of insulin and antihypertensive medication.
Anthropometric measurements: All patients had to rest quietly for 5 minutes, and then blood pressure was measured by a specially trained professional physician and nurse, and three consecutive readings were recorded and averaged. Other than that, they were asked to dress very lightly, barefoot, and measure their standard weight to the nearest 0.1 kg using a calibrated digital scale; maintain a standing position and quantify their standard height to the nearest 0.1 cm using a calibrated telemetry device, and measure their waist circumference with a exible tape measure at the level of 1 cm above the umbilicus. All measurements were performed twice and averaged for analysis.BMI was calculated as weight per height squared (kg/m2).WHtR was de ned as WC divided by height in meters.
Measurement of experimental indicators: Fasting blood was collected by venipuncture after a brief fast (overnight for 8 hours). Fasting blood glucose (FPG) was measured by the glucose oxidase method; glycated hemoglobin (HbA1c) was measured by high-performance liquid chromatography. A fully automated biochemical analyzer (Hitachi, 7600) was used to detect blood uric acid (SUA), blood creatinine (Scr), urea nitrogen (BUN), aspartate aminotransferase (ALT), alanine aminotransferase (AST), alkaline phosphatase (ALP), gamma-glutamine transferase (λ-GGT), TG, total cholesterol (TC), HDL-C, and low-density lipoprotein cholesterol (LDL-C). The 1st-morning urine was retained, and the urinary albumin/creatinine ratio (UACR) was measured by immunoturbidimetric assay. TG/HDL-C ratio was calculated using available TG and HDL-C values. All laboratory equipment was calibrated, and double-blind samples were used.
De nitions CMI was obtained by the following formula [6]: CMI = TG/HDL-C × WHtR. lipid accumulation product (LAP) was calculated according to the gender-speci c formula [10]: female LAP = TG (mmol/L) × [WC (cm) -58] and male LAP = TG (mmol/L) × [WC (cm) -65]. Normoalbuminuria was de ned as UACR < 30 mg/g; microalbuminuria was de ned as 30 mg/g ≤ UACR < 300 mg/g [11]. Hypertension was diagnosed as mean systolic blood pressure (SBP) equal to or greater than 140 mm Hg and/or diastolic blood pressure (DBP) of at least 90 mm Hg and/or the participant was taking antihypertensive medication or self-reported previously diagnosed hypertension [12]. Diagnosis of diabetes according to the American Diabetes Association criteria: fasting plasma glucose (FPG) ≥ 7.0 mmol/L and/or self-reported history of previous diagnosis or treatment with plasma glucose-lowering [13].

Statistical analyses
Analyses were performed in a gender-speci c manner using SPSS 25.0 software. Normally distributed measures were expressed as mean±standard deviation (x±s), and comparisons between the 4 groups were made using one-way ANOVA, and 2-pair comparisons between groups were made using the LSD-t-test. Non-normally distributed measures were expressed as M(Q1, Q3), and comparisons between 4 groups were made using Kruskal-Wallis H-test, and comparisons between 2 pairs of groups were made using Nemenyi's method. Categorical variables were expressed as frequencies (%), and comparisons were made using the χ2 test. Multiple logistic regression analysis was used to explore the association between CMI (When included in the analysis as a continuous or a categorical variable, respectively) and diabetic microalbuminuria and to obtain a sex-speci c dominance ratio (OR) for each SD change in CMI to determine the risk of microalbuminuria in T2DM, with results shown as ORs and 95% con dence intervals (95% CI). Finally, the optimal cut-off value for CMI was investigated using subject operating characteristic (ROC) curves to detect the presence of microalbuminuria. The area under the curve (AUC) was used to compare the ability of CMI to discriminate microalbuminuria with other indicators. Differences were considered statistically signi cant at P<0.05.

Results
Demographics, physical measurements values, and laboratory indices of all study participants ( We included a total of 524 patients with con rmed T2DM, with a mean age of 57.28 ± 10.52 years, with a higher mean age in women than in men, 46% were female, and both men and women had microalbuminuria older than their age group. Among men, 37.8% of T2DM patients had microalbuminuria, while the percentage of women with microalbuminuria was higher at 41.1%. Regarding physical measurements, regardless of gender, patients with microalbuminuria had signi cantly higher BMI, WC, SBP, DBP, and VFA than the group without microalbuminuria, and they had a longer duration of diabetes, substantially longer in women than in men. Laboratory index results showed that FPG, HbA1c, ALT, AST, and λGGT were signi cantly higher in the male patients in the group with microalbuminuria. Regardless of gender, WHtR, TG, and TG/HDL-C were considerably higher in the microalbuminuria group, and HDL-C was signi cantly lower than in the regular urine albumin group. Among all subjects, smoking prevalence was signi cantly higher in men than in women, while female patients tended not to smoke and men showed signi cant differences in smoking status. Hypertension prevalence and antihypertensive medication use were more common in the group with microalbuminuria but differed signi cantly in female patients. Also, not surprisingly, CMI and LAP were increased dramatically in patients with microalbuminuria. Dominance ratios (OR) and 95% con dence intervals for patients with microalbuminuria in T2DM according to sex-speci c continuous variables or quartiles of CMI and LAP ( Table 2, Table at the end of the article) Multiple logistic regression analysis was performed to assess the gender-speci c associations of CMI and LAP with microalbuminuria. CMI and LAP were included in the analysis as continuous variables, and the results showed that the risk of microalbuminuria increased by 48% and 59% for each SD increase in CMI level in men and women, respectively. The association between CMI and microalbuminuria was signi cant and robust, more pronounced in women (women, OR, 1.586; 95% CI, 1.181-2.128; men, OR, 1.480; 95% CI, 1.206-1.816). While we took CMI and LAP as categorical variables into the multiple logistic regression analysis and used age, SBP, DBP, BMI, and WC as confounding factors (model 2), then found that in female subjects, the risk of microalbuminuria in the highest quartile of CMI and LAP was 5.83 times (95% CI, 2.38 to 14.32) and 9.16 times (95%CI, 3.04 to 27.54) of those in the lowest quartile, respectively. In model 3, we further included FPG, HbA1C, history of hypertension, antihypertensive medication, and history of insulin use, and the results did not show substantial changes, and the correlation between CMI and LAP and microalbuminuria remained signi cant. Thus, higher CMI and LAP levels in women were independently and positively associated with the prevalence of microalbuminuria in a dose-response pattern (P<0.001). Furthermore, we found a 4.67-fold increase in the OR of microalbuminuria in men with higher CMI quartile levels (95% CI, 1.910 to 11.405; P<0.001), whereas LAP was not signi cantly associated with the prevalence of microalbuminuria after inclusion of confounding factor analysis (P=0.141). Overall, the association of CMI and LAP with microalbuminuria was more prominent in female patients.

Discussion
We found that CMI, a novel metabolic index that expresses both dyslipidemias and accurately re ects central obesity status, was signi cantly and independently associated with the risk of developing microalbuminuria in Chinese patients with type 2 diabetes. The ndings reveal the potential of CMI as a screening marker for microalbuminuria in patients with T2DM, demonstrating that CMI is a crucial independent determinant of early diabetic nephropathy and has important implications for exploring potential areas of research targeting CMI in the future, especially for delaying the progression of diabetic nephropathy and reducing the risk of developing end-stage renal disease. Patients with T2DM are often associated with dyslipidemia and central obesity. Previous studies have shown that individuals with abnormal lipid metabolism and obesity have worse metabolic pro les and poorer renal outcomes [14]. In patients with early-stage chronic kidney disease, dyslipidemia is mainly characterized by atherogenic dyslipidemia: high TG levels, low HDL-C levels, and increased concentrations of small, dense, low-density lipoprotein (sdLDL) particles [15]. Cao et al. [16] showed that high TG levels are an essential risk factor for kidney damage, and the adjusted risk of any renal event associated with increased TG levels remains high. sd-LDL is a vital lipid indicator in diabetic dyslipidemia [17], and TG/HDL-C is thought to re ect the level of "sd-LDL" particles. Because of the combination of TG and HDL-C levels, TG/HDL-C is valuable in differentiating microalbuminuria. Wen et al. [18] found that elevated TG/HDL-C was signi cantly associated with microalbuminuria excretion in Chinese patients with type 2 diabetes. A study of Japanese diabetic patients showed a signi cant decrease in LDL particle size and a considerable increase in the TG/HDL-C ratio in patients with nephropathy than subjects without nephropathy [19]. As our hypothesis in general, TG/HDL-C performed better in predicting proteinuria/chronic kidney disease compared to non-HDL-C/HDL-C, TG, and HDL-C [20].
Obesity is considered to be an essential risk factor for the development of renal damage in T2DM. Data from the Framingham Heart Study, which included more than 2600 patients without chronic kidney disease at baseline, showed that obese patients were at increased risk of developing stage 3 chronic kidney disease ((BMI ≧ 30 kg/m2)) compared to nonobese subjects [21]. Several independent studies have shown that anthropometric measures representing central obesity are superior to general obesity indicators in predicting the development of microalbuminuria in T2DM [22][23][24][25] Cardiometabolic index (CMI) is a new index calculated from TG/HDL-C and WHtR, which was proposed by Ichiro Wakabayashi in 2015 [6]. It can better re ect the status of diabetes mellitus and the progression of atherosclerosis [31][32].
As a new indicator of visceral adipose tissue distribution and dysfunction, CMI has been closely associated with various diseases since its introduction [33][34][35]. In T2DM, considering that insulin resistance, in ammatory response, and abnormal lipid metabolism are inextricably linked to microalbuminuria production, we speculate that CMI is associated with microalbuminuria excretion and may serve as a novel screening index for early diabetic nephropathy. This paper is the rst to explore the predictive value of CMI and other lipid and obesity indicators for the development of microalbuminuria in diabetic patients in a Chinese type 2 diabetic population.
We included CMI and LAP as categorical variables in our analysis and found that both had a more signi cant effect on early kidney damage in women compared to men, but the impact of LAP appeared to be more powerful. After excluding the impact of confounding factors, the risk of developing microalbuminuria in men with T2DM increased signi cantly with increasing quartiles of CMI, independent of LAP. In a cross-sectional study of 5398 healthy Koreans over 20 years of age, in general, the urinary albumin/creatinine ratio increased with increasing CMI levels, and that CMI was associated with increased glucose and urinary albumin/creatinine ratio increased were independently correlated [36]. Wang et al. also concluded that CMI re ects the visceral fat area and the pathological process leading to impaired renal function and can be used as a screening marker for CKD [37]. Our study con rmed that the risk of developing microalbuminuria increased proportionally with changes in CMI quartiles, and CMI showed a strong correlation with microalbuminuria excretion in T2DM. Although the ORs for the risk of developing CMI were weaker than LAP, considering that LAP does not include HDL-C, an essential indicator of dyslipidemia, in its calculations for analysis, and that people of different heights should have other WC criteria, it neglects the role of height. It, therefore, has limited value as a test indicator for combined microalbuminuria in patients with T2DM [38].
To distinguish the predictive value of CMI from lipid and obesity indicators as screening markers for diabetic microalbuminuria, we analyzed T2DM patients of both sexes separately using ROC curves. We found that CMI had the most signi cant AUC value in female patients. Although its discriminatory power was not signi cantly greater than LAP, it had a more comprehensive diagnostic signi cance considering that CMI covered both HDL-C and height-speci c indicators.
Surprisingly, CMI and LAP showed better predictive effects in women, but both TG/HDL-C and WHtR had larger AUC values than WC. This cause may be related to the smaller number of our subjects. Similarly, the AUC value of CMI was still the largest in male patients; however, the sensitivity of CMI was only 68%, and the speci city was even lower at 54.00%. Compared to men, CMI is much more sensitive than LAP in women, so CMI values are more valuable for diagnosing microalbumin excretion in women with T2DM. In conclusion, CMI has a robust discriminatory ability and high sensitivity and can be an economic screening index for screening people with T2DM combined with microalbuminuria.
Our study found that CMI and LAP are of clinical value as new metabolic indicators in assessing the risk of developing microalbuminuria in type 2 diabetes and differ by gender. The deleterious effect of CMI on early diabetic nephropathy was more signi cant in women compared to men. Our ndings are consistent with previous clinical studies in which abnormal lipid metabolism and central obesity triggered microalbuminuria excretion, particularly in female patients with T2DM. A study based on the Korean National Health and Nutrition Examination Survey found that obese female patients were more likely to develop microalbuminuria and that impaired fasting glucose and high triglycerides were signi cant correlates of microalbuminuria [39]. The interaction of increased glucose exposure with abnormal lipid metabolism associated with central obesity may be one of the leading causes of vascular endothelial dysfunction and microalbuminuria excretion in a 5398 Korean general population [40]. Based on a retrospective observational study of renal injury in elderly participants in a Chinese community, researchers hypothesized that female patients with high TG and high WC levels were signi cantly associated with microalbuminuria excretion [41]. In conjunction with our ndings, we showed that the statistical effect of visceral fat distribution on microalbuminuria excretion in female T2DM patients manifested itself as a signi cant effect of CMI and LAP. The reason for this outcome may be the dysregulation of macrophage and adipocyte secretion in patients with T2DM, which produces a variety of hormones and pro-in ammatory factors, leading to a low in ammatory response, insulin resistance, dyslipidemia, and/or increased synthesis of vasoactive and brogenic substances [42][43][44][45][46]. Their complex interactions may negatively affect the vascular endothelium and impair renal function, thus causing the production of microalbuminuria. In addition, considering that most of the female patients we included were postmenopausal women with decreased sex hormones, the bene cial effects of estrogen on the kidney were inevitably lost. Estrogen is thought to reduce proteinuria, glomerulosclerosis, and tubulointerstitial brosis [47][48][49]. Possible mechanisms are: (i) in diabetic patients, (estradiol) E2 attenuates glomerulosclerosis and tubulointerstitial brosis by reducing type I and type IV collagen synthesis, increasing matrix metalloproteinase expression, and inhibiting apoptosis [50][51][52]; (ii) E2 reduces the angiotensin type II1 receptor expression, transforming growth factor-β and endothelin-1 and regulating vascular dysfunction; (iii) E2 has also been shown to upregulate nitric oxide synthase activity and vascular endothelial growth factor expression in glomeruli and improve vascular permeability, thereby potentially reducing the loss of glomerular function in progressive chronic kidney disease, including diabetic nephropathy [53][54][55][56].
This study still has some limitations: (1) This study is a cross-sectional study, and this study can only provide evidence about the strong correlation between CMI and LAP and microalbuminuria. Still, it cannot further elucidate the causal relationship between CMI, LAP, and microalbuminuria in patients with T2DM, which needs to be further veri ed by a large follow-up study. (2) The sample size was small and limited by the region, and data we only collected from some patients in southwest China with a single ethnicity, so more studies are needed to investigate whether the ndings apply to different regions or ethnic groups. (3) There are too many confounding factors affecting microalbuminuria excretion, and the effects of confounding factors were not eliminated when conducting the analysis, which may produce some data bias.

Conclusion
In summary, our study reveals for the rst time an association between CMI and microalbuminuria in type 2 diabetes that is independent of prevalent CVD, medication use, and traditional cardiovascular risk factors. Thus, these data provide strong evidence for a unique, independent, and economic role of CMI in high-burden nephropathy. These ndings have important implications for guiding original prevention and understanding the mechanisms of VAT-mediated kidney injury.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This study was supported by grants from the "Metabolic Disease Cohort Study" program funds (The NationalKey Research and Development Program of China, Grant#2016YFC0901200, 2016YFC0901205). The founder provided nancial support for our survey, but do not interfere with the design, operation, and analysis of the study.

Authors' contributions
In this study, XY did the study design, statistical analyses, and results in interpretation. YFJ, XLH, and XLC participated in analyzing and resolving di culties of analytic strategies and results in discussion. Finally, QW functioned as a nal reviewer who gave constructional suggestions for the interpretation of data. The corresponding author was QW. All authors have read and approved the manuscript.

Acknowledgments
We would like to express our gratitude to all those who exert their effects in achieving this study.