Association between different insulin resistance surrogates and all-cause mortality in peritoneal dialysis patients

doi:10.21203/rs.3.rs-4193924/v1

Background: The insulin resistance (IR) is related with poor outcomes in peritoneal dialysis (PD) patients. This study aimed to explore the association between surrogate IR indexes and all-cause mortality in PD patients.

Methods: 1736 PD patients were enrolled from March 2010 to December 2021 in multiple PD centers. Cox proportional hazards models and restricted cubic splines (RCS) regression were constructed to assess the relationship between TyG index, TyG-BMI index, TG/HDL-C ratio and all-cause mortality. All three IR indexes were integrated into same model to access the predictive stability. Furthermore, forest plot was draw to display the result of subgroup analysis in PD patients.

Results: 378 mortality events were recorded during a median follow-up time of 2098 days. The RCS regression showed a linear association between the TyG index and TyG-BMI index and all-cause mortality (P for non-linearity 0.566, 0.716 respectively), while a non-linear association between TG/HDL-C ratio and all-cause mortality (P for non-linearity 0.022). Multivariable Cox proportional hazards analyses indicated higher levels of TyG index, TyG-BMI index and TG/HDL-C ratio were independent risk factors for all-cause mortality in PD patients (HR 1.587, 95%CI 1.264-1.993; HR 1.449, 95%CI 1.085-1.934; HR 1.424, 95%CI 1.104-1.838, respectively). By integrating three IR indexes into the same model, the TyG index showed the highest predictive stability among them. Forest plot of TyG index showed there were no significant interactions in the subgroups.

Conclusion: The TyG index, TyG-BMI index, and TG/HDL-C ratio were significantly associated with all-cause mortality in PD patients, and the TyG index may be the most stable surrogate marker for IR in PD patients among them, and IR was associated with all-cause mortality in PD patients.

The prevalence of chronic kidney disease (CKD) has reached 8–16% globally[1]. There are approximately 759 per million people globally receiving treatment of end-stage renal disease (ESRD) and 11% of them undergoing peritoneal dialysis (PD)[2, 3]. Due to the convenience and economic advantages of PD, several national health care systems have adopted a “PD first” policy, prioritizing PD unless a medical contraindication exists[4]. Although PD can extend the survival time of ESRD patients, and improve their prognosis[5], in comparison to the age-matched general population, the mortality rate is 6.1 to 7.8 times higher[6]. Thus, it is crucial for PD patients to conduct further research on prognostic factors and identify potential treatment targets.

Insulin resistance (IR) is considered as a decrease in sensitivity or responsiveness to the metabolic actions of insulin including insulin-mediated glucose disposal, usually accompanied by hyperinsulinemia[7]. Along with abnormalities in glucose and lipid metabolism, IR is common feature in PD patients[8]. The hyperinsulinemic-euglycemic clamp (HEC) represents the most reliable method for evaluating IR. The other surrogate measures to estimate levels of IR include the Quantitative Insulin Sensitivity Check Index (QUICKI) and the Homeostasis model assessment of IR (HOMA-IR). Nevertheless, due to the expense and complex calculation of them, their clinically implementation is restricted[9]. Currently, triglyceride-glucose (TyG) index, triglyceride glucose body mass (TyG-BMI) index and triglyceride/high-density lipoprotein cholesterol (TG/HDL-C) ratio have been confirmed to be reliable substitute indicators for IR[10–12]. Numerous studies shown highly consistent association between these indicators and adverse clinical outcomes[13–17]. However, to date, no prior study has investigated the relationship between these IR surrogates and all-cause mortality in PD patients. This study aims to explore the relationship between these indexes and all-cause mortality in PD patients using data from a multi-center cohort investigation, with the objective of identifying the most stable predictive among them.

Study Design and Participants

This was a retrospective, multicenter, cohort study. In total, 1736 PD patients were enrolled from four PD centers during the period spanning from March 2010 to December 2021. Exclusion criteria included: (1) Patients aged below 18 years; (2) PD duration ($<$3 mouths); (3) Patients lacking significant baseline date to calculate TyG index, TyG-BMI index and TG/HDL-C index. 1242 PD patients were eligible for final analysis (Fig. 1). All of them received continuous ambulatory peritoneal dialysis (CAPD) and applied PD as their first form of renal replacement therapy. The optimal cut-off value was determined by the receiver operator characteristic (ROC) curve. All participants signed informed consent before admission to the hospital.

Data Collection

We obtained clinical data by using electronic medical record system, which included information such as sex, age, weight, height, history of cardiovascular disease (CVD), diabetes and stroke, as well as the administration of insulin and statin medications. Biochemical data included fasting blood glucose (FBG), triglyceride (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), albumin (Alb), uric acid (UA) and calcium, were measured by each PD center’s standard laboratory techniques. We used following formula to estimate residual renal function $(mL/min/1.73 {m}^{2})$ (RRF), and then correcting it by body surface area. Gehan and George equation were used to calculate body surface area:$RRF=\frac{1}{2}\left[\frac{Urine Creatinine (\mu mmol/l)}{Serum Creatinine (\mu mmol/l)}\right]+\left[\frac{Urine Urea (mmol/l)}{Serum Urea (mmol/l)}\right]\times \frac{Urine Volum \left(ml\right) }{1440}$[18]. Total Kt/V were calculated by PD Adequest software 2.0 (Baxter, Deerfeld, IL). Patients were required to return each center for a comprehensive medical assessment at least once every quarter, and trained nurses followed up them via telephone monthly to assess their general conditions.

Definitions

Age equal to or older than 65 was considered as the elderly[19]. Body mass index (BMI) was calculated by following formula: $BMI (Kg/{m}^{2})= \left[\frac{Weight \left(Kg\right)}{{Height}^{2} \left({m}^{2}\right)}\right]$. And once patients’ BMI surpasses 24.9 $Kg/{m}^{2}$, they were regarded as overweight or obesity. CVD was defined as a history of conditions including coronary heart disease, myocardial infarction, or heart failure, or a history of undergoing vascular intervention procedures or coronary artery bypass surgery[20]. Diabetes was defined by meeting any of the following criteria: (1) the usage of insulin; (2) Glycated hemoglobin A1c (HbA1c) levels ≥ 6.5%; (3) FBG levels ≥ 7.0 $mmol/l$; (4) Patients with a history of diabetes. Other comorbidities were defined according to the International Classification of Diseases, Ninth Revision (ICD-9). TyG index, TyG-BMI index and TG/HDL-C ratio were calculated by formulas as follows:

$$TyG = ln \left[\frac{\left(TG \right(mg/dl)\times FBG (mg/dl)}{2}\right]$$

$$TyG-BMI = ln \left[\frac{\left(TG \right(mg/dl)\times FBG (mg/dl)}{2}\right]\times BMI (Kg/{m}^{2})$$

$$TG/HDL-C = \frac{TG (mg/dl)}{HDL-C (mg/dl)}$$

Study outcomes

The primary endpoint was all-cause mortality during the PD period. Besides, Patients who transferred to hemodialysis (HD), underwent renal transplantation, transferred to another center and were lost to follow-up were also recorded. The follow-up of our study ended on December 8, 2021.

Statistical analysis

All statistical analysis was performed applying SPSS statistical software (version 22.0), R Project (version 4.3.2) and GraphPad Prism 8 statistical packages. Missing data were imputed utilizing multiple imputation. P < 0.05 was considered as statistically significant. Normality was determined by Kolmogorov–Smirnov normality test. Continuous variables without normal distribution were expressed as medians (25th-75th percentile) while categorical variables were given as number (%). The chi-square test or nonparametric Mann–Whitney test was exploited to examine whether there were differences between two groups. We employed univariable Cox proportional hazards regression to examine the association between significant risk factors and all-cause mortality, and a multivariable Cox proportional hazards regression was performed to assess the association between diverse levels of IR indexes and all-cause mortality. Model 1 was unadjusted for any factors, Model 2 was adjusted for laboratory indices, and Model 3 included adjustments for laboratory indices, elderly, sex, BMI, comorbid conditions and medication. Additionally, we incorporated three IR indexes into Model 3 to form Model 4 in order to identify the most stable predictive indicator among them. Spearman’s correlation analysis was conducted to explore the correlation of IR indexes with other significant mortality risk factors. The adjusted Kaplan-Meier survival curves showed the contrast in survival outcomes between higher and lower IR index groups. We also drew forest plot to present outcomes of subgroup analysis. Furthermore, restricted cubic splines (RCS) were used to elucidate the relationship between various IR indexes and all-cause mortality.

Baseline Characteristics of the Participants

A total of 1242 PD patients were finally enrolled in our study. At the end of follow-up, 378 participants had died, 111 participants transferred to HD, 61 participants underwent renal transplantation, 18 participants transferred to another centers, 18 participants lost to follow-up, 656 participants reached the final follow-up date (Fig. 1).

Table 1 shows the baseline features between death group and non-death group. Compared with non-death group, individuals in death group were more likely to be older, had history of CVD, stroke and diabetes, and were more likely to be on insulin therapy. Besides, patients in daeth group had higher levels of FBG, TG, TC, LDL-C, calcium, TyG index, TyG-BMI index, TG/HDL-C ratio. Furthermore, participants who encountered mortality had lower levels of UA and worse RRF.

Table 1

Baseline clinical characteristics of peritoneal dialysis patients.
	Total (n = 1242)	Non-death (n = 864)	Death (n = 378)	P value
Demographics
Male (%)	686 (55.2)	482 (55.8)	204 (54.0)	0.553
Elderly (%)	239 (19.2)	112 (13.0)	127 (33.6)	< 0.001
BMI (Kg/m²)	21.91 (20.04–24.44)	21.98 (19.92–24.43)	21.77 (20.28–24.50)	0.830
Comorbid
Diabetes (%)	468 (37.7)	264 (30.6)	204 (54.0)	< 0.001
CVD (%)	220 (17.7)	101 (11.7)	119 (31.5)	< 0.001
Stroke (%)	61 (4.9)	28 (3.2)	33 (8.7)	< 0.001
Medication use
Use of insulin (%)	190 (15.3)	117 (13.5)	73 (19.3)	0.009
Use of statins (%)	193 (15.5)	135 (15.6)	58 (15.3)	0.900
Laboratory variables
FBG (mmol/L)	4.81 (4.20-6.00)	4.70 (4.12–5.47)	5.30 (4.36–7.49)	< 0.001
TG (mmol/L)	1.41 (0.97–2.13)	1.35 (0.93–2.02)	1.49 (1.04–2.34)	0.001
TC (mmol/L)	4.75 (4.00-5.61)	4.61 (3.90–5.50)	5.01 (4.32-6.00)	< 0.001
HDL-C (mmol/L)	1.11 (0.91–1.38)	1.11 (0.90–1.36)	1.12 (0.92–1.40)	0.580
LDL-C (mmol/L)	2.78 (2.21–3.42)	2.72 (2.17–3.33)	2.92 (2.27–3.57)	0.003
Alb (g/L)	35.50 (31.70–39.30)	35.55 (31.93–39.50)	35.20 (30.98-39.00)	0.262
UA (µmol/L)	408.50 (353.00-472.00)	413.00 (356.00-481.00)	399.50 (346.75–454.00)	0.002
Ca (mmol/L)	2.20 (2.04–2.37)	2.19 (2.03–2.35)	2.24 (2.08–2.39)	0.001
Kt/V	2.31 (1.95–2.79)	2.30 (1.93–2.78)	2.34 (1.96–2.80)	0.298
RRF (mL/min/1.73m²)	3.81 (2.04–8.62)	4.04 (2.09–11.27)	3.31 (1.93–5.62)	< 0.001
TyG-BMI index	190.59 (168.37-215.91)	188.87 (167.13-213.25)	194.28 (172.17-220.84)	0.008
TyG index	8.63 (8.19–9.16)	8.53 (8.14–9.08)	8.81 (8.36–9.37)	< 0.001
TG-HDL index	1.22 (0.75–2.18)	1.16 (0.72–2.13)	1.37 (0.80–2.32)	0.014
Data are shown as medians (25th-75th percentile) or n (%). Age ≥ 65 was considered as the elderly.
Abbreviations: BMI, body mass index; CVD, cardiovascular disease; FBG, fasting blood glucose; TG, triglyceride; TC, total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; Alb, albumin; UA, uric acid; Ca, calcium; RRF, residual renal function; TyG index, triglyceride-glucose index; TyG-BMI index, triglyceride glucose body mass index; TG/HDL-C ratio, triglyceride/high-density lipoprotein cholesterol ratio

Correlations between IR indexes and traditional all-cause mortality risk factors

Tables S₁ separately presented the relationship between TyG index, TyG-BMI index and TG/HDL-C ratio with other risk factors with all-cause mortality in PD patients, respectively. There was a positive correlation between TyG index and age, BMI, TC, LDL-C, hemoglobin (HGB), Kt/V, while a negative correlation was observed with HDL-C and RRF. Moreover, similar to the TyG index, the TyG-BMI index was also found to be positively correlated with age, TC, LDL-C, and negatively correlated with HDL-C. However, there was a negative correlation between the TyG-BMI index and Alb and Kt/V. The TG/HDL-C ratio displayed a positive correlation with age, BMI, FBG, TC, while there was no significant correlation between TG/HDL-C ratio and LDL-C, Alb Kt/V or RRF.

Relationship between IR indexes and all-cause mortality

RCS revealed a linear association between the TyG index and TyG-BMI index with the risk of all-cause mortality (P for non-linearity 0.566, 0.716 respectively). Nevertheless, a non-linear relationship was found between TG/HDL-C ratio and all-cause mortality (Fig. 2).

We constructed a univariable Cox proportional hazards regression to explore the association between baseline characteristics and all-cause mortality in PD patients. In brief, a significant correlation was found between all-cause mortality and advanced age, history of diabetes, stroke and CVD, use of insulin, as well as levels of UA, FBG, TC, TG, Ca, Alb, Kt/V and RRF. Apart from levels of UA, Alb, and RRF which showed a negative correlation with all-cause mortality, the other baseline characteristics were found to be significant positively correlated with all-cause mortality (Table S₂). The ROC curve displayed the optimal cut-off value of TyG index, TyG-BMI index and TG/HDL-C ratio was 8.59, 191.76, and 1.21, respectively (Table S₃). We separated all PD patients into higher and lower IR index groups utilizing the respective optimal cut-off values. Table 2 showed the result of multivariable Cox proportional hazards regression. As shown in Table 2, it was found that elevated levels of the TyG index and TyG-BMI index were associated with a higher risk of mortality for PD patients, regardless of potential confounders were unadjusted in Model 1 or adjusted in Model 2 and Model 3 (HR 1.238, 95%CI 1.074–1.428 P value 0.003; HR 1.009, 95%CI 1.003–1.015, P value 0.005, respectively). Additionally, for TyG index group, patients in higher TyG index group had 1.587 times higher risk of mortality (95%CI 1.264–1.993, P value < 0.001), which would resemble the performance of higher TyG-BMI index (HR 1.392, 95%CI 1.132–1.713, P value 0.002) and higher TG/HDL-C ratio (HR 1.584, 95%CI 1.239–2.204, P value < 0.001) groups. We also plot adjusted Kaplan–Meier survival curves to display the association between diverse IR indexs group and all-cause mortality. Compared to lower IR indexs group, each higher IR index group showed a remarkable better survival rate (Fig. 3).

Table 2 Association between the IR index and all-cause mortality in the peritoneal dialysis patients
	Events/Total	Model 1		Model 2		Model 3
		HR (95% CI)	P value	HR (95% CI)	P value	HR (95% CI)	P value
		HR (95% CI)	P value	HR (95% CI)	P value	HR (95% CI)	P value
Continuous TyG Index (per unit change)	378/1242 (30.4)	1.420 (1.259-1.600)	<0.001	1.463 (1.287-1.664)	<0.001	1.238 (1.074-1.428)	0.003
Lower TyG index	138/600 (23.0)	Reference	-	Reference	-	Reference	-
Higher TyG index	240/642 (37.4)	1.786 (1.448-2.202)	<0.001	1.845 (1.485-2.293)	<0.001	1.587 (1.264-1.993)	<0.001
Continuous TyG-BMI Index (per unit change)	378/1242 (30.4)	1.004 (1.001-1.006)	0.005	1.004 (1.001-1.007)	0.003	1.009 (1.003-1.015)	0.005
Lower TyG-BMI index	172/642 (26.8)	Reference	-	Reference	-	Reference	-
Higher TyG-BMI index	206/600 (34.3)	1.393 (1.137-1.706)	0.001	1.392 (1.132-1.713)	0.002	1.449 (1.085-1.934)	0.012
Continuous TG/HDL-C ratio (per unit change)*	378/1242 (30.4)	-	-	-	-	-	-
Lower TG-HDL index	164/613 (26.8)	Reference	-	Reference	-	Reference	-
Higher TG-HDL index	214/629 (34.0)	1.350 (1.102-1.655)	0.004	1.584 (1.239-2.204)	<0.001	1.424 (1.104-1.838)	0.007

Model 1 Unadjusted

Model 2 Adjusted by calcium, albumin, uric acid, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, Kt/V

Model 3 Adjusted by model 2 plus sex, elderly, body mass index, diabetes, stroke, cardiovascular disease, use of insulin, use of statins

TyG index, triglyceride-glucose index; TyG-BMI index, triglyceride glucose body mass index; TG/HDL-C ratio, triglyceride/high-density lipoprotein cholesterol ratio

*Since there was not a linear association between TG/HDL-C ratio and all-cause mortality, we didin’t use Cox proportional hazards regression to assess the association between continuous TG/HDL-C ratio and all-cause mortality.

In conclusion, higher TyG index, TyG-BMI index and TG/HDL-C ratio was more likely to undergoing mortality. To determine the most stable index for predicting all-cause mortality in PD patients among them, we integrated three IR indexes into Model 3 to construct Model 4. By adjusting for Model 4, we observed a significant positive correlation between higher TyG index group and all-cause mortality (HR 1.493, 95%CI 1.105–2.017, P value 0.009), while no statistically significant relationship was found between the higher TyG-BMI index and TG/HDL-C ratio group and all-cause mortality (Table 3). This suggests that the TyG index was the most stable predictor.

Table 3

Integrating three IR indexes into Model 3
	Model 4
	HR (95% CI)	P value
Lower TyG index	Reference	-
Higher TyG index	1.493 (1.105–2.017)	0.009
Lower TyG-BMI index	Reference	-
Higher TyG-BMI index	1.156 (0.840–1.590)	0.374
Lower TG-HDL index	Reference	-
Higher TG-HDL index	1.026 (0.740–1.421)	0.879
Model 4 Adjusted by Model 3 plus TyG index group, TyG-BMI index group, TG-HDL index group
TyG index, triglyceride-glucose index; TyG-BMI index, triglyceride glucose body mass index; TG/HDL-C ratio, triglyceride/high-density lipoprotein cholesterol ratio

Subgroup Analyses

We performed subgroup analyses in sex (male or female), elderly (yes or no), diabetes (yes or no), CVD (yes or no), stroke (yes or no), overweight or obesity (yes or no) (Fig. 4). After adjusting for confounders across all subgroups, we didn’t find significant interactions in the subgroups (all P for interaction > 0.05).

In this study, we construct a retrospective, cohort study, and found a significantly positive correlation between IR indexes and all-cause mortality in PD patients. Compared to lower index group, higher TyG index, higher TyG-BMI index and higher TG/HDL-C ratio group had a 158.7%, 144.9% and 142.4% increased risk of occurring mortality, may be considered as a potential target for improving the survival of patients undergoing PD.

Previous studies have found strongly correlation between TyG index, TyG-BMI index and TG/HDL-C ratio and mortality in different populations. In a recent large cross-sectional study, TyG-BMI was found to be associated with CVD mortality among participants from the National Health and Nutrition Examination Survey (NHANES), and ROC analysis showed that the TyG index had better predictive ability for CVD mortality than the TyG-BMI index[21]. Similarly, in a large prospective cohort study, Lopez-Jaramillo et al. reveled the positive correlation between higher TyG index and cardiovascular mortality [22]. Furthermore, in a cohort study aimed at acute ischemic stroke population, a negative correlation between higher TG/HDL-C ratio and 3-month survival was found by Deng and his collaborators, and in the subsequent construction of the clinical prediction mode, model enrolled TG/HDL-C ratio presented a perfect match between predicted and observed values [23]. Additionally, in a cohort study enrolled 6697 patients with chronic heart failure, tertile 2 of the TyG index was associated with a 1.29-times higher risk of all-cause mortality, and tertiles 3 with at 1.84-times risk, compared with the lowest TyG index quartile[24]. Another retrospective cohort study, comprising 8835 critically ill patients, compared the association between the dynamic change of the TyG index and the baseline TyG index with 1-year all-cause mortality. Following propensity score matching, Cheng et al. demonstrated a linear association between TyG variability ratio and 1-year all-cause mortality. And they suggested that the dynamic change of the TyG index may offer a more reliable predictive value for all-cause mortality compared to the baseline TyG index[25]. Additionally, in the field of nephrology, levels of TG/HDL-C ratio were also found to have liner association with all-cause mortality, both in early CKD patients and in advanced CKD patients[26]. However, to date, the relationship between these indexes and all-cause mortality has not been evaluated in the same PD patients, and the linear or non-linear association has not assessed either. Our study filled these gaps.

Although our study demonstrated the positive relationship between TyG index, TyG-BMI index and TG/HDL-C ratio and all-cause mortality in PD patients, the specific biological mechanisms were unclear. TyG index, TyG-BMI index and TG/HDL-C ratio are simply calculated by TG, FBG, BMI and HDL-C. Earlier studies have demonstrated the strong positive correlation between these indexes and IR. For example, Fernando et al. suggested that the TyG index had a high sensitivity (96.5%) and specificity (85.0%) for diagnosing IR, with an area under the curve (AUC) of 0.858 [27]. The higher TyG-BMI index and TG/HDL-C ratio were also related to higher morbidity of type 2 diabetes[28, 29]. Thus, TyG index, TyG-BMI index and TG-HDL-C ratio were gradually considered as simple and reliable markers for assessing IR. Consequently, our speculation is that IR may play a significant role in linking these indexes with all-cause mortality.

As previously mentioned, IR is characterized by reduced sensitivity or responsiveness to the metabolic actions of insulin. Thus, accompanied by the decreased sensitivity to the insulin, β cells will compensatorily increase insulin secretion, leading hyperinsulinemia. Hyperinsulinemia, as a feature of IR, caused endothelial dysfunction, promoted inflammation, and associated with development of numerous chronic diseases[30–32], especially with CVD[33–35]. A study from Finland also indicated that high fasting serum insulin level was independent risk factor for cardiovascular mortality. Moreover, IR also played a significant role in progressing of atherosclerosis[36, 37]. A recent meta-analysis summarized the association between them and concluded that increased levels of HOMA-IR are associated with a higher risk of developing coronary heart disease [38]. The specific biological mechanisms through which IR contributes to atherosclerosis are not yet clear. However, numerous in vitro and in vivo studies have indicated potential mechanisms were as follows: (1) stimulating lipid synthesis and activating SREBP-C[39, 40]; (2) promoting vascular smooth muscle cell growth and proliferation[41]; (3) decreasing the activation of endothelial nitric oxide synthetase (eNOS) and nitric oxide (NO) production in endothelial cells by inhibiting the palmitoylation of eNOS and promoting the activation of the Endothelial cell Na⁺ channel (EnNaC), ultimately leading to endothelial dysfunction[7, 42]. All these changes may result in higher morbidity of CVD, and contribute to poor prognosis.

Interestingly, our study also indicated that among these three indexes, TyG index showed the highest predictive level and stability for all-cause mortality in PD patients. We assumed it may be related to the best predictive ability of TyG index of IR among these indexes[43, 44]. However, the specific mechanism remains to be studied further. This suggests that clinicians are encouraged to utilize TyG index to evaluate the risk of occurring all-cause mortality in PD patients and take appropriate measures.

There were limitations to the current research. Firstly, since this study was retrospective, we were unable to eliminate the effect of taking different medications on the results. Secondly, due to several relevant data were missed, we were unable to assess the relationship between these indexes and cardiovascular-related mortality. Thirdly, although we hypothesized that the potential mechanism linking these indexes and all-cause mortality may be related to IR, our study didn’t conduct the HEC. Thus, we were unable to evaluate the predictive effects of these indicators on IR in PD patients. Further studies are needed to explore the association between the IR indexes and the gold standard of IR in PD patients. Furthermore, we only analyzed baseline TyG index, TyG-BMI index and TG/HDL-ratio, dynamic changes in indexes were not included in present study. Therefore, future studies are more likely to explore the predictive power of changes in these indexes on all-cause mortality in PD patients.

The TyG index, TyG-BMI index, and TG/HDL-C ratio have clinical significance for predicting all-cause mortality in PD patients.

Conflict of interest:

The authors have no competing interests to declare that are relevant to the content of this article.

Ethical approval:

Consent for participation was not required as the study was a retrospective review. All procedures performed in this study involving human participants was in accordance with the ethical standards of the institution. The study was performed in line with the 1975 Declaration of Helsinki and was approved by the Ethics Committee of the Sixth Affiliated Hospital of Sun Yet-Sen University (No. 2021SLYEC-177).

Funding

This study was supported by the Wu Jieping Medical Foundation (No.320.6750.2022-18-5), the Key Project of Social Science and Technology of Dongguan (No. 201950715046061) and Guangzhou Science and Technology Program (No.2023A04J2244)

Author Contribution

All authors contributed to the study conception and design. Ning Su and Xing Zhang conceived the design of this manuscript. Sijia Shang and Guowen Zhao drafted the original manuscript. The data was provided by all authors. Qian Zhou and Xiaojiang Zhan checked the appraisals of risk of bias and applicability concerns. Data collection ana analysis were performed by Yuanyuan Yang. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Acknowledgements

We would like to thank to The Ever-green Tree Nephrology Group for data support. We thank all the doctors and nurses at the peritoneal dialysis centers.

B. Ene-Iordache, N. Perico, B. Bikbov, S. Carminati, A. Remuzzi, A. Perna, N. Islam, R.F. Bravo, M. Aleckovic-Halilovic, H. Zou, L. Zhang, Z. Gouda, I. Tchokhonelidze, G. Abraham, M. Mahdavi-Mazdeh, M. Gallieni, I. Codreanu, A. Togtokh, S.K. Sharma, P. Koirala, S. Uprety, I. Ulasi, G. Remuzzi, Chronic kidney disease and cardiovascular risk in six regions of the world (ISN-KDDC): a cross-sectional study, Lancet Glob Health, 4 (2016) e307-319.
J.S. Thurlow, M. Joshi, G. Yan, K.C. Norris, L.Y. Agodoa, C.M. Yuan, R. Nee, Global Epidemiology of End-Stage Kidney Disease and Disparities in Kidney Replacement Therapy, Am J Nephrol, 52 (2021) 98-107.
Y. Cho, A.K. Bello, A. Levin, M. Lunney, M.A. Osman, F. Ye, G.E. Ashuntantang, E. Bellorin-Font, M.B. Gharbi, S.N. Davison, M. Ghnaimat, P. Harden, H. Htay, V. Jha, K. Kalantar-Zadeh, P.G. Kerr, S. Klarenbach, C.P. Kovesdy, V. Luyckx, B. Neuen, D. O'Donoghue, S. Ossareh, J. Perl, H.U. Rashid, E. Rondeau, E.J. See, S. Saad, L. Sola, I. Tchokhonelidze, V. Tesar, K. Tungsanga, R.T. Kazancioglu, A. Yee-Moon Wang, C.W. Yang, A. Zemchenkov, M.H. Zhao, K.J. Jager, F.J. Caskey, K.K. Jindal, I.G. Okpechi, M. Tonelli, D.C. Harris, D.W. Johnson, Peritoneal Dialysis Use and Practice Patterns: An International Survey Study, Am J Kidney Dis, 77 (2021) 315-325.
P.K. Li, K.M. Chow, M.W. Van de Luijtgaarden, D.W. Johnson, K.J. Jager, R. Mehrotra, S. Naicker, R. Pecoits-Filho, X.Q. Yu, N. Lameire, Changes in the worldwide epidemiology of peritoneal dialysis, Nat Rev Nephrol, 13 (2017) 90-103.
P.K. Li, K.M. Chow, Peritoneal dialysis-first policy made successful: perspectives and actions, Am J Kidney Dis, 62 (2013) 993-1005.
R. Saran, Y. Li, B. Robinson, J. Ayanian, R. Balkrishnan, J. Bragg-Gresham, J.T. Chen, E. Cope, D. Gipson, K. He, W. Herman, M. Heung, R.A. Hirth, S.S. Jacobsen, K. Kalantar-Zadeh, C.P. Kovesdy, A.B. Leichtman, Y. Lu, M.Z. Molnar, H. Morgenstern, B. Nallamothu, A.M. O'Hare, R. Pisoni, B. Plattner, F.K. Port, P. Rao, C.M. Rhee, D.E. Schaubel, D.T. Selewski, V. Shahinian, J.J. Sim, P. Song, E. Streja, M. Kurella Tamura, F. Tentori, P.W. Eggers, L.Y. Agodoa, K.C. Abbott, US Renal Data System 2014 Annual Data Report: Epidemiology of Kidney Disease in the United States 6.1-7.8, Am J Kidney Dis, 66 (2015) Svii, S1-305.
M.A. Hill, Y. Yang, L. Zhang, Z. Sun, G. Jia, A.R. Parrish, J.R. Sowers, Insulin resistance, cardiovascular stiffening and cardiovascular disease, Metabolism, 119 (2021) 154766.
M. Lambie, M. Bonomini, S.J. Davies, D. Accili, A. Arduini, V. Zammit, Insulin resistance in cardiovascular disease, uremia, and peritoneal dialysis, Trends Endocrinol Metab, 32 (2021) 721-730.
A. Gastaldelli, Measuring and estimating insulin resistance in clinical and research settings, Obesity (Silver Spring), 30 (2022) 1549-1563.
T. Du, G. Yuan, M. Zhang, X. Zhou, X. Sun, X. Yu, Clinical usefulness of lipid ratios, visceral adiposity indicators, and the triglycerides and glucose index as risk markers of insulin resistance Cardiovasc Diabetol, 13 (2014) 146.
J.S. Kim, H.T. Kang, J.Y. Shim, H.R. Lee, The association between the triglyceride to high-density lipoprotein cholesterol ratio with insulin resistance (HOMA-IR) in the general Korean population: based on the National Health and Nutrition Examination Survey in 2007-2009, Diabetes Res Clin Pract, 97 (2012) 132-138.
L.K. Er, S. Wu, H.H. Chou, L.A. Hsu, M.S. Teng, Y.C. Sun, Y.L. Ko, Triglyceride Glucose-Body Mass Index Is a Simple and Clinically Useful Surrogate Marker for Insulin Resistance in Nondiabetic Individuals, PLoS One, 11 (2016) e0149731.
B. Che, C. Zhong, R. Zhang, L. Pu, T. Zhao, Y. Zhang, L. Han, Triglyceride-glucose index and triglyceride to high-density lipoprotein cholesterol ratio as potential cardiovascular disease risk factors: an analysis of UK biobank data, Cardiovasc Diabetol, 22 (2023) 34.
C. Zhan, Y. Peng, H. Ye, X. Diao, C. Yi, Q. Guo, W. Chen, X. Yang, Triglyceride glucose-body mass index and cardiovascular mortality in patients undergoing peritoneal dialysis: a retrospective cohort study, Lipids Health Dis, 22 (2023) 143.
F. Sato, Y. Nakamura, K. Kayaba, S. Ishikawa, TG/HDL-C ratio as a predictor of stroke in the population with healthy BMI: The Jichi Medical School Cohort Study, Nutr Metab Cardiovasc Dis, 32 (2022) 1872-1879.
R.R. Huo, L. Zhai, Q. Liao, X.M. You, Changes in the triglyceride glucose-body mass index estimate the risk of stroke in middle-aged and older Chinese adults: a nationwide prospective cohort study, Cardiovasc Diabetol, 22 (2023) 254.
Y. Yang, X. Huang, Y. Wang, L. Leng, J. Xu, L. Feng, S. Jiang, J. Wang, Y. Yang, G. Pan, B. Jiang, Y. Wang, L. Chen, The impact of triglyceride-glucose index on ischemic stroke: a systematic review and meta-analysis, Cardiovasc Diabetol, 22 (2023) 2.
X. Zhan, Y. Yang, Y. Chen, X. Wei, J. Xiao, L. Zhang, C. Yan, P. Qiu, S. Liu, Q. Hu, Q. Chen, Y. Wang, Serum alkaline phosphatase levels correlate with long-term mortality solely in peritoneal dialysis patients with residual renal function, Ren Fail, 41 (2019) 718-725.
T. Tong, C. Yang, W. Tian, Z. Liu, B. Liu, J. Cheng, Q. Cheng, B. Zhou, Phenotypes and outcomes in middle-aged patients with diabetic foot ulcers: a retrospective cohort study, J Foot Ankle Res, 13 (2020) 24.
Y. Zhang, X. Lai, Q. Chen, X. Zhan, H. Long, F. Peng, F. Zhang, X. Feng, Q. Zhou, X. Wu, L. Liu, Z. Wang, X. Peng, K. Hu, D. Wang, G. Guo, Y. Zeng, Y. Wen, J. Liang, The relationship between neutrophil-to-lymphocyte ratio and the first occurrence of pneumonia in peritoneal dialysis patients, Clin Exp Nephrol, 24 (2020) 770-778.
K. Dang, X. Wang, J. Hu, Y. Zhang, L. Cheng, X. Qi, L. Liu, Z. Ming, X. Tao, Y. Li, The association between triglyceride-glucose index and its combination with obesity indicators and cardiovascular disease: NHANES 2003-2018, Cardiovasc Diabetol, 23 (2024) 8.
P. Lopez-Jaramillo, D. Gomez-Arbelaez, D. Martinez-Bello, M.E.M. Abat, K.F. Alhabib, Á. Avezum, O. Barbarash, J. Chifamba, M.L. Diaz, S. Gulec, N. Ismail, R. Iqbal, R. Kelishadi, R. Khatib, F. Lanas, N.S. Levitt, Y. Li, V. Mohan, P.K. Mony, P. Poirier, A. Rosengren, B. Soman, C. Wang, Y. Wang, K. Yeates, R. Yusuf, A. Yusufali, K. Zatonska, S. Rangarajan, S. Yusuf, Association of the triglyceride glucose index as a measure of insulin resistance with mortality and cardiovascular disease in populations from five continents (PURE study): a prospective cohort study, Lancet Healthy Longev, 4 (2023) e23-e33.
Q.W. Deng, S. Li, H. Wang, L. Lei, H.Q. Zhang, Z.T. Gu, F.L. Xing, F.L. Yan, The Short-term Prognostic Value of the Triglyceride-to-high-density Lipoprotein Cholesterol Ratio in Acute Ischemic Stroke., Aging Dis, 9 (2018) 498-506.
Y. Zhou, C. Wang, H. Che, L. Cheng, D. Zhu, C. Rao, Q. Zhong, Z. Li, X. Wang, Z. Wu, K. He, Association between the triglyceride-glucose index and the risk of mortality among patients with chronic heart failure: results from a retrospective cohort study in China, Cardiovasc Diabetol, 22 (2023) 171.
L. Cheng, F. Zhang, W. Xue, P. Yu, X. Wang, H. Wang, J. Wang, T. Hu, H. Gong, L. Lin, Association of dynamic change of triglyceride-glucose index during hospital stay with all-cause mortality in critically ill patients: a retrospective cohort study from MIMIC IV2.0, Cardiovasc Diabetol, 22 (2023) 142.
Y. Kim, S. Lee, Y. Lee, M.W. Kang, S. Park, S. Park, K. Han, J.H. Paek, W.Y. Park, K. Jin, S. Han, S.S. Han, H. Lee, J.P. Lee, K.W. Joo, C.S. Lim, Y.S. Kim, D.K. Kim, Predictive value of triglyceride/high-density lipoprotein cholesterol for major clinical outcomes in advanced chronic kidney disease: a nationwide population-based study TG/HDL-C, Clin Kidney J, 14 (2021) 1961-1968.
F. Guerrero-Romero, L.E. Simental-Mendía, M. González-Ortiz, E. Martínez-Abundis, M.G. Ramos-Zavala, S.O. Hernández-González, O. Jacques-Camarena, M. Rodríguez-Morán, The product of triglycerides and glucose, a simple measure of insulin sensitivity. Comparison with the euglycemic-hyperinsulinemic clamp, J Clin Endocrinol Metab, 95 (2010) 3347-3351.
B. Song, X. Zhao, T. Yao, W. Lu, H. Zhang, T. Liu, C. Liu, K. Wang, Triglyceride Glucose-Body Mass Index and Risk of Incident Type 2 Diabetes Mellitus in Japanese People With Normal Glycemic Level: A Population-Based Longitudinal Cohort Study, Front Endocrinol (Lausanne), 13 (2022) 907973.
H. Yuge, H. Okada, M. Hamaguchi, K. Kurogi, H. Murata, M. Ito, M. Fukui, Triglycerides/HDL cholesterol ratio and type 2 diabetes incidence: Panasonic Cohort Study 10, Cardiovasc Diabetol, 22 (2023) 308.
J. Janssen, Hyperinsulinemia and Its Pivotal Role in Aging, Obesity, Type 2 Diabetes, Cardiovascular Disease and Cancer, Int J Mol Sci, 22 (2021).
A.M.Y. Zhang, E.A. Wellberg, J.L. Kopp, J.D. Johnson, Hyperinsulinemia in Obesity, Inflammation, and Cancer, Diabetes Metab J, 45 (2021) 285-311.
R. Liu, S. Guan, Z. Gao, J. Wang, J. Xu, Z. Hao, Y. Zhang, S. Yang, Z. Guo, J. Yang, H. Shao, B. Chang, Pathological Hyperinsulinemia and Hyperglycemia in the Impaired Glucose Tolerance Stage Mediate Endothelial Dysfunction Through miR-21, PTEN/AKT/eNOS, and MARK/ET-1 Pathways, Front Endocrinol (Lausanne), 12 (2021) 644159.
G.M. Reaven, Insulin resistance/compensatory hyperinsulinemia, essential hypertension, and cardiovascular disease, J Clin Endocrinol Metab, 88 (2003) 2399-2403.
M. Pyörälä, H. Miettinen, M. Laakso, K. Pyörälä, Hyperinsulinemia predicts coronary heart disease risk in healthy middle-aged men: the 22-year follow-up results of the Helsinki Policemen Study, Circulation, 98 (1998) 398-404.
R. Muniyappa, M. Montagnani, K.K. Koh, M.J. Quon, Cardiovascular actions of insulin, Endocr Rev, 28 (2007) 463-491.
G. Reaven, Insulin resistance and coronary heart disease in nondiabetic individuals, Arterioscler Thromb Vasc Biol, 32 (2012) 1754-1759.
V. Ormazabal, S. Nair, O. Elfeky, C. Aguayo, C. Salomon, F.A. Zuñiga, Association between insulin resistance and the development of cardiovascular disease, Cardiovasc Diabetol, 17 (2018) 122.
K.B. Gast, N. Tjeerdema, T. Stijnen, J.W. Smit, O.M. Dekkers, Insulin resistance and risk of incident cardiovascular events in adults without diabetes: meta-analysis, PLoS One, 7 (2012) e52036.
S.J. Koopmans, R.S. Kushwaha, R.A. DeFronzo, Chronic physiologic hyperinsulinemia impairs suppression of plasma free fatty acids and increases de novo lipogenesis but does not cause dyslipidemia in conscious normal rats, Metabolism, 48 (1999) 330-337.
D. Azzout-Marniche, D. Bécard, C. Guichard, M. Foretz, P. Ferré, F. Foufelle, Insulin effects on sterol regulatory-element-binding protein-1c (SREBP-1c) transcriptional activity in rat hepatocytes, Biochem J, 350 Pt 2 (2000) 389-393.
I. Golovchenko, M.L. Goalstone, P. Watson, M. Brownlee, B. Draznin, Hyperinsulinemia enhances transcriptional activity of nuclear factor-kappaB induced by angiotensin II, hyperglycemia, and advanced glycosylation end products in vascular smooth muscle cells, Circ Res, 87 (2000) 746-752.
K.E. Bornfeldt, I. Tabas, Insulin resistance, hyperglycemia, and atherosclerosis, Cell Metab, 14 (2011) 575-585.
T. Du, G. Yuan, M. Zhang, X. Zhou, X. Sun, X. Yu, Clinical usefulness of lipid ratios, visceral adiposity indicators, and the triglycerides and glucose index as risk markers of insulin resistance, Cardiovasc Diabetol, 13 (2014) 146.
M. Mirr, D. Skrypnik, P. Bogdański, M. Owecki, Newly proposed insulin resistance indexes called TyG-NC and TyG-NHtR show efficacy in diagnosing the metabolic syndrome, J Endocrinol Invest, 44 (2021) 2831-2843.

No competing interests reported.

Association between different insulin resistance surrogates and all-cause mortality in peritoneal dialysis patients

Status:

Version 1

Abstract

Figures

Introduction

Methods

Study Design and Participants

Data Collection

Definitions

Study outcomes

Statistical analysis

Results

Baseline Characteristics of the Participants

Correlations between IR indexes and traditional all-cause mortality risk factors

Relationship between IR indexes and all-cause mortality

Subgroup Analyses

Discussion

Conclusion

Declarations

Funding

Author Contribution

Acknowledgements

References

Additional Declarations

Supplementary Files

Status:

Version 1