The Utility of Perirenal Fat in Determining the Risk of Onset and Progression of Diabetic Kidney Disease

Background Perirenal fat (PRF) has multiple effects on the kidney through its physical structure and adipocytokine-secreting ability. The present study explored the relationship between PRF thickness and the onset and progression of albuminuria in patients with diabetes. Methods In the cross-sectional analysis, we screened 959 patients from 8764 subjects with type 2 diabetes mellitus (T2DM) who met the inclusion criteria and measured their perirenal fat thickness (PFT) using color Doppler ultrasound. A group of laboratory indexes were included in the analysis models. In a longitudinal study, a total of 218 patients with a baseline UACR <30 mg/g were included in the follow-up study. Results In a cross-sectional analysis, patients with diabetes and higher PFT presented with higher albuminuria. Multiple logistic regression analysis indicated that PFT was an independent risk factor for the degree of albuminuria in patients with T2DM (odds ratio = 4.186, 95%CI: 2.290–7.653, P < 0.001). In a longitudinal study, 218 albuminuria-free patients with T2DM at the baseline were followed up for a mean of 12.3 months. Based on the cutoff value from the ROC diagnostic test in the cross-sectional study, patients were divided into two groups: higher PFT (H-PFT) and lower PFT (L-PFT). Kaplan–Meier survival curve analysis showed that H-PFT was associated with a higher incidence of albuminuria than L-PFT (log-rank test, χ2 = 4.522, P = 0.033). Cox regression analysis showed that PFT was a risk factor for the earlier onset of albuminuria (hazard ratio 2.83, 95% CI: 1.34–4.88, P < 0.001). Conclusions PRF evaluated by color Doppler ultrasound is an easy and reliable tool for predicting the onset and progression of albuminuria in patients with T2DM.


Introduction
Diabetic kidney disease (DKD) is the most frequent and severe complication of diabetes mellitus (DM). Approximately 40% of patients with diabetes present with proteinuria and renal dysfunction and DKD reportedly account for 50% of newly diagnosed end-stage renal disease (ESRD) [1,2]. Metabolic disorders caused by hyperglycemia in DKD can lead to both glucose and lipid metabolism disorders. Adipose tissue, an organ that stores energy and releases endocrine factors, participates in various physiological processes by secreting adipokines and cytokines, thereby maintaining lipid metabolic homeostasis and energy balance [3]. Adipose tissue consists of visceral adipose tissue (VAT) and subcutaneous adipose tissue (SAT) [4]. Accumulating evidence suggests that the abnormal distribution and ectopic accumulation of VAT in humans are more closely associated with cardiovascular risk factors, such as insulin resistance, dyslipidemia, hypertension, and atherosclerosis [5].
Perirenal fat (PRF) is a subtype of VAT located in the retroperitoneal space encapsulating the kidneys and flling the space between the kidney and the adjacent retroperitoneal tissue [6]. PRF has been reported to be associated with renal disease in various ways [7,8]. Previous studies showed that PRF was associated with a loss in renal function in patients with diabetes [9][10][11]. In addition, PRF was positively associated with the degree of albuminuria in a crosssectional study with a small sample size of patients with type 2 diabetes mellitus (T2DM) [12]. However, the value of PRF in predicting the onset of albuminuria in patients with diabetes warrants further longitudinal studies.
Te present study investigated the relationship between PRF and albuminuria in a large sample of patients with T2DM at both cross-sectional and longitudinal levels. A threshold value for perirenal fat thickness (PFT) in patients with diabetes was set to render a noninvasive and reliable indicator for predicting the occurrence of early stage DKD.

Study Participants.
Patients with T2DM aged  were included in this study. T2DM was diagnosed according to American Diabetes Association (ADA) Diabetes Diagnostic Criteria [13]. Te exclusion criteria were (1) patients with incomplete clinical data, (2) patients without renal ultrasound records, (3) patients with a history of malignancy and surgical trauma, and (4) patients diagnosed with nondiabetic renal diseases. Based on the inclusion and exclusion criteria, 959 patients with T2DM treated at the Renmin Hospital of Wuhan University from December 2019 to April 2021 were enrolled in this cross-sectional study. Based on the degree of the urine albumin-creatinine ratio (UACR) [14], patients with diabetes were divided into three groups: subjects without albuminuria (DM-1 group, UACR <30 mg/g, 636 cases), subjects with microalbuminuria (DM-2 group, 30 mg/g < UACR < 300 mg/g, 191 cases), and subjects with macroalbuminuria (DM-3 group, UACR > 300 mg/g, 132cases). Moreover, 100 healthy individuals who signed informed consent forms for clinical research from the health examination center in our hospital were included in the control group (CG group).
Additionally, patients with a baseline UACR < 30 mg/g were recruited for the longitudinal study. A total of 218 patients with confrmed consent forms were followed up for a median time of 12.3 months (range: 9.4, 13.1). Te composite endpoints for renal outcomes were UACR >30 mg/g and total 24-hour urinary protein (UTP) > 0.15 g/ 24 h. Based on the cutof value from the receiver operating characteristic (ROC) diagnostic test in the cross-sectional study and according to their median PFT, the subjects in the longitudinal study were assigned to the group with increased PFT (H-PFT group) and the group with decreased PFT (L-PFT group) at the baseline. A fow chart of the included human subjects is presented in Figure 1. Te study was approved by the ethics committee of Renmin Hospital of Wuhan University to access the digital records of subjects who signed informed consent forms to share digital records with local institutions (WDRY2021-KS034). Tis study was approved by the China Clinical Trial Registry (ChiCTR2100052518, https://www.chictr.org.cn).

Ultrasound Measurement of PFT.
Ultrasonography of the kidneys from individual subjects was performed using a duplex Doppler instrument (Acuson Sequoia 512 ultrasound system, Siemens, USA). PFT was measured according to a previously described protocol [11]. Briefy, PRF was the thickness of the fat capsule along the long axis of the kidney, and pararenal fat was the thickness of the fat layer measured between the medial surface of abdominal muscle tissue and the renal fascia. PRF was measured while the patient was in the supine position. Te probe was held vertical to the skin on the transverse aspects of the abdomen and slowly moved  laterally until the optimal position was found. Longitudinal scanning was performed with the kidney surface almost parallel to the skin. Te pressure on the probe should be as small as possible to prevent the fat layer from being compressed. Te thickness of the fat capsule was quantifed by the vertical distance between the renal fascia and the surface of the kidney, also defned as the PFT, which was measured from both sides of the kidney and averaged from individual subjects by three experienced technologists with blind group information. Te correlation between PFT values measured on both sides was 0.45 (P < 0.001) ( Figure S1). Te interoperator agreement between the three experienced technologists is 0.89. Te schematic diagram of PFT measurement is presented in Figure 2.

Statistical
Analysis. Origin (version 2021b; Origin Lab, Inc., Northampton, MA, USA) was used for graphing the data, and data analysis was performed using SPSS (version 22; SPSS, Inc., Chicago, IL, USA). Data were expressed as means ± SD or as medians with quartiles according to the distribution characteristics. T-test, the Mann-Whitney U test, and one-way analysis of variance (ANOVA), and Tukey multiple comparison test were used to compare the diferences between two or more groups. Logistic regression analysis was performed for the univariate analysis, which included variables with a signifcant diference. Kaplan-Meier survival analysis and log-rank tests were performed to compare the incidence of albuminuria onset in subjects with diferent PFTs. Cox univariate and multifactor regression analyses were performed to estimate the risk factors for albuminuria onset and progression to diabetes. Statistical signifcance was set at P < 0.05.

Cross-Sectional Analysis in Patients with Diabetes.
A total of 959 patients were enrolled in this cross-sectional analysis. Te clinical characteristics of patients with diabetes (DM group) and healthy subjects without diabetes (CG group) are presented in Table 1. Te HbA1C, FPG, UA, and PFT thickness levels were higher in the DM group than in the CG group, whereas the HDL and ALB levels were lower in the DM group than in the CG group (P < 0.05).
Among the subjects with diabetes, the DM-1, DM-2, and DM-3 groups were divided according to the degree of albuminuria. A signifcant diference was observed in the PFT, HbA1C, UTP, age, urine creatinine, urine microalbumin, UACR, ALB, urea, SCr, UA, TG, LDL, eGFR, Hb, PT, duration of diabetes mellitus, and hypertension levels among the three DM subgroups (P < 0.05) ( Table 2). An increasing trend from the DM-1 group to the DM-3 group was indicated in age, duration of diabetes mellitus, UTP, UACR, urea, SCr, UA, PFT, and PFT/BMI.
According to the Pearson correlation analysis, PFT thickness was positively correlated with the degree of UTP, UACR, urea, SCr, UA, TG, BMI, and the history of hypertension (Table S1). In contrast, a negative correlation ( Figure 3) was found between PFT, HbA1C, and eGFR (P < 0.05). Te heatmap in Figure 3(a) describes the mutual relationship between the two indexes with signifcant differences among the subgroups of DM. Representative distribution patterns with PFT are presented in Figures 3 To compare the ability to predict the diferent stages of DKD, the parameters of ALB, eGFR, and PFT were included in the ROC curve. As shown in Figure 3(f ), PFT can better predict the onset of albuminuria (damaging albuminuria to microalbuminuria). In contrast, the power of PFT to predict the stage of microalbuminuria to macroalbuminuria was lower than that of ALB and eGFR (Figure 3(g)). Moreover, the cutof value [1] of PFT for predicting a new onset of albuminuria was 0.90 cm (sensitivity, 96.3%; specifcity, 30.5%; and Youden index, 1.27). Te cutof value [2] of PFT for predicting the ( Figure 2) worsening of albuminuria was 1.56 cm (sensitivity, 28.0%; specifcity, 88.48%; and Youden index, 1.17).
Using proteinuria as the endpoint, multiple logistic regression showed that increased UTP, increased PFT, increased TG, and age were the risk factors for the onset of proteinuria (P < 0.05), as shown in Table S2, and a scoring system for risk factors for the presence of proteinuria was obtained ( Figure S2A). When the presence of proteinuria was used as the endpoint time, multiple logistic regression International Journal of Endocrinology showed that increased BMI, male sex, and increased UACR were risk factors for thickened PFT (P < 0.05), as shown in Table S3, and a scoring system for risk factors for increased PFT was obtained ( Figure S2B).

Follow-Up Analysis of Patients with Albuminuria-Negative Diabetes.
Based on the cutof value from the ROC diagnostic test in the cross-sectional part, 218 DM patients without albuminuria enrolled in the follow-up study were divided into two groups: L-PFT (n = 41) and H-PFT (n = 177). Te baseline characteristics of all the subjects in the longitudinal study are shown in Table 3. Albuminuria was set as the endpoint for a median followup time of 12.3 months; three patients (7.3%) with L-PFT reached this endpoint, including two patients with UACR > 30 mg/g and one with UACR > 300 mg/g. In contrast, 46 patients (26.0%) with H-PFT reached the endpoint, which included 32 with UACR > 30 mg/g and 14 with UACR > 300 mg/g. A Kaplan-Meier survival curve indicated a signifcantly higher incidence in the H-PFT group than in the L-PFT group (log-rank test, χ2 = 4.522, P = 0.033), as shown in Figure 4(a). Multivariate Cox regression models were used to estimate the risk factors for albuminuria onset. Various models (model 1: demographic parameters + PFT; model 2: demographic parameters + history of disease + concomitant medications + PFT; and model 3: demographic parameters + history of disease + concomitant medications + clinical indexes + PFT) were tested, and PFT was found to be an independent risk factor for the onset of albuminuria in patients with albuminuria-negative DM (model 1: HR 2.91, 95% CI: 1.16-7.30, P � 0.003; model 2: HR 2.89, 95% CI: 1.14-7.50, P � 0.026; and model 3: HR 2.83, 95% CI: 1.34-4.88, P < 0.001), as shown in Table 4. Forest plots of the individual models are shown in Figure 4(b).

Discussion
Tis study found that PRF was thickened in patients with diabetes. Additionally, PRF thickness measured by ultrasound was associated with albuminuria in diferent stages of DKD in cross-sectional and longitudinal studies, rendering it an easy and practical approach to predict the onset and progression of albuminuria in patients newly diagnosed with diabetes.
Traditionally, PRF has been considered a specifc component of visceral fat. An increasing numbers of studies have reported that PRF is associated with the progression of various chronic diseases, such as cardiovascular disease, DM, and renal disease [15]. Our study showed that PFT levels were higher in patients with diabetes than in healthy controls. In addition, remarkable PRF was also present in diabetic subjects with albuminuria compared to those without albuminuria. After adjustment for confounders, logistic regression analysis showed that increased PFT was an independent risk factor for the development of proteinuria in patients with diabetes. In the longitudinal study, patients with thicker PRF had a poorer prognosis and were more likely to develop proteinuria. Terefore, a close relationship between proteinuria in patients with DKD and thickened PFT has been proposed. A previous study reported the relationship between perirenal fat and proteinuria in DKD using a small sample size [12]. Additionally, a recent study [9] applied computerized tomography (CT) to  International Journal of Endocrinology outline PFT in diabetic patients and indicated that PFT was a more sensitive tool for monitoring renal function loss than other VAT. Compared to the precise measurement of the PFT by CT, ultrasound is less accurate. However, B-mode ultrasound imaging is easy to access and manipulate to evaluate kidney morphology and PFT parameters. Te present study indicates that a PFT cutof value of 0.90 cm is reliable for predicting the onset of albuminuria in subjects with T2DM. Moreover, the cutof value of 1.56 cm for PFT is useful for predicting the progression of DKD in patients. Both BMI and age are well-accepted risk factors for onset of DKD [16]. Intriguingly, the thickness of PRF was also associated with the state of DKD independent with age and BMI in the present study, but whether PRF is a dependent marker of kidney function disorder in a diabetic environment needs further studies. Dyslipidemia is a metabolic disorder in patients with T2DM that contributes to worsening renal function and proteinuria in DKD [17,18]. In our study, TG and PFT levels were independent risk factors for proteinuria in DKD.
However, TG and other serum lipid parameters, such as Tch and LDL, had weak relationships with the thickness of PRF, suggesting that the source of PRF is independent of systemic lipid metabolism in the diabetic environment.
Renal lipid accumulation toxicity (lipotoxicity) is believed to promote kidney disease progression [19,20]. It has been found that patients with DKD have increased cholesterol synthesis and impaired efux from renal cells, increased fatty acid uptake, and decreased fatty acid oxidation [21]. Several studies have reported an association between increased visceral fat, remarkably increased PRF, metabolic and cardiovascular disease determinants, and renal dysfunction. Te mechanisms by which PRF afects renal injury have not been fully elucidated.
It has been suggested that PRF may infuence the development and progression of renal disease by afecting the RAS system and secretion of multiple adipokines and infammatory cytokines [22]. Te proximal location of PRF in the kidney makes its specifc anatomical and morphological features closely related to the development and progression     [23]. However, little is known about the origin of the perirenal adipocytes. In DKD, the origin of PRF thickening needs to be investigated in depth.
Te present study had some limitations as follows: Te sample size of the longitudinal study was relatively small. Since albuminuria is a time-dependent complication of diabetes, a longer follow-up time could strengthen the predictive value of PRF for the onset of albuminuria. Population recruitment from multiple centers in diferent ethnic regions could reduce the selection bias. However, the specifc mechanism of PRF thickening and its efect on renal function have not yet been elucidated.

Conclusion
Tickened PRF is a clinical feature of albuminuria-positive patients with T2DM and is an independent risk factor for the development of proteinuria in DKD. Based on the potential functions of PRF, accumulation of PRF is an alternative therapeutic target to prevent the onset and progression of DKD.

Data Availability
Te resources generated and/or analyzed in the present study are accessible from the corresponding authors on reasonable request.

Ethical 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. Tis study was approved by the ethics committee of the Renmin Hospital of Wuhan University (WDRY2021-KS034) and the China Clinical Trial Registry (ChiCTR2100052518, https://www.chictr.org.cn).

Conflicts of Interest
Te authors declare no conficts of interest.