Article Text

Original research
Association between the triglyceride to high-density lipoprotein cholesterol ratio and mortality in Chinese maintenance haemodialysis patients: a retrospective cohort study
  1. Lemuge Qi1,
  2. Aihong Zhang2,
  3. Yanping Zhang1,
  4. Zhuo Ren1,
  5. Chen Zhao1,
  6. Qian Wang1,
  7. Kaiming Ren1,
  8. Jiuxu Bai1,
  9. Ning Cao1
  1. 1Department of Blood Purification, General Hospital of Northern Theatre Command, Shenyang, Liaoning, China
  2. 2Department of Nephrology, Xi’an People’s Hospital (Xi’an Fourth Hospital), Xi'an, Shaanxi, China
  1. Correspondence to Dr Ning Cao; bzxyjhk{at}; Dr Jiuxu Bai; 107034054{at}


Objective To investigate the relationship between the triglyceride to high-density lipoprotein cholesterol (TG/HDL-C) ratio and all-cause and cardiovascular (CV) mortality in Chinese haemodialysis (HD) patients.

Design Retrospective cohort study.

Setting Patients from June 2015 to September 2016 and followed through September 2021 were categorised into quartiles according to the follow-up averaged TG/HDL-C ratio. The association between TG/HDL-C and mortality was examined by univariate and multivariate time-varying Cox regression analyses. The C-index was used to assess the predictive accuracy of the Cox regression models.

Participants A total of 534 maintenance HD patients were enrolled.

Primary and secondary outcome measures The outcomes were all-cause death and CV mortality.

Results During the median follow-up of 61 months, 207 patients died, with 94 (45.4%) classified as CV death. After adjusting for confounders, multivariate time-varying Cox regression analysis showed that the quartile 4 group (TG/HDL-C ≥2.64) was associated with decreased all-cause mortality (adjusted HR 0.51, 95% CI 0.33–0.77, p=0.001) and CV mortality (adjusted HR 0.31; 95% CI 0.16 to 0.62; p=0.001) in maintenance HD patients. Model 1 of all-cause mortality achieved a C-index of 0.72 (95% CI 0.68 to 0.75), and model 2 achieved a C-index of 0.77 (95% CI 0.73 to 0.82). The C-index for model 1 in CV mortality was 0.74 (95% CI 0.70 to 0.77), and the C-index for model 2 was 0.80 (95% CI 0.75 to 0.84).

Conclusions High TG/HDL-C was associated with decreased all-cause and CV mortality in HD patients.

  • Chronic Disease
  • Lipid disorders
  • Nephrology
  • Dialysis
  • End stage renal failure

Data availability statement

Data are available upon reasonable request.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Strengths and limitations of this study

  • Time-varying Cox regression analyses were used to identify the time-varying effect of triglyceride to high-density lipoprotein cholesterol (TG/HDL-C) levels on the mortality of haemodialysis (HD) patients.

  • Restricted cubic spline curves were built.

  • The C-index was applied to assess the predictive ability of the Cox regression models.

  • Retrospective cohort study that limits causal inferences.

  • Limited size and data of this study.


Chronic kidney disease (CKD) is considered a major threat to human health, with a prevalence of approximately 9.1%.1 Patients with CKD treated with haemodialysis (HD) are prone to cardiovascular disease (CVD) and have a significant risk of mortality from CVD.2 Dyslipidaemia has traditionally been considered as the risk factor of atherosclerosis and CVD. However, it is increasingly recognised that the underlying mechanisms are more complicated, and the adverse effects of dyslipidaemia on CVD are in doubt. According to two review studies,3 4 the traditional cholesterol hypothesis that hyperlipidaemia contributes to CVD and mortality is questionable and that statin therapy does not achieve beneficial effects in the general population. Similarly, the effect of lipids on mortality in HD patients remains controversial.5

Compared with each of those individual lipid parameters alone, the ratio of triglyceride to high-density lipoprotein cholesterol (TG/HDL-C), a novel lipid predictor, has recently been found to be more predictive.6 The ratio that reflects the complicated interplay of lipoprotein metabolism overall7 has great predictive power and ability for mortality in the general population8–10 and in patients with a history of CVD.11 However, there has been controversy about the usefulness of the ratio in HD patients, and studies are limited. Given that the amount of research is limited, the goal of this study was to estimate the link between the TG/HDL-C ratio and all-cause and cardiovascular (CV) death in HD patients.

Materials and methods

Study population and data source

Our study was a single-centre retrospective cohort study. Patients who received maintenance HD treatment at the General Hospital of Northern Theater Command from June 2015 to September 2016 were selected and followed up retrospectively until death or censoring in September 2021. The study enrolled patients who were >18 years old and treated with HD three times per week for at least 90 days. We excluded patients who were converted from HD to peritoneal dialysis (PD) or kidney transplantation, transferred to other dialysis centres and patients without available data (especially lipid data). Therefore, the final study population consisted of 534 patients. The baseline demographic data (age, sex, Body Mass Index (BMI)), clinical data (cause of end-stage renal disease (ESRD), dialysis vintage (years), CVD, hypertension, diabetes, use of lipid-lowering agents) and biochemical data (haemoglobin, total cholesterol (TC), TG, HDL-C, low-density lipoprotein cholesterol (LDL-C), serum calcium (Ca), serum phosphorus (P), parathyroid hormone (PTH) and serum albumin (Alb)), which were measured in the fasting state before dialysis in the laboratory of the General Hospital of Northern Theater Command, were gathered from electronic medical records of the dialysis organisation centre. The repetitive data of TG and HDL-C gathered every 6 months during the follow-up period were used to calculate the follow-up averaged TG/HDL-C ratio and used for time-varying analysis.

Patient and public involvement


Definition and outcome

Serum TG levels were divided by HDL-C levels to obtain the TG/HDL-C ratio. BMI was computed by dividing the dry body weight by the height squared. Medication history was defined as patients who took lipid-lowering drugs regularly for more than 6 months at the time of enrolment. The outcomes were all-cause mortality and CV mortality. All-cause mortality was defined as death from any cause. CV mortality was defined as death caused by unstable angina, acute myocardial infarction, cardiac arrhythmia, cardiomyopathy, cardiac arrest, congestive heart failure, cerebrovascular disease or peripheral vascular disease, which was obtained from the electronic medical record system or follow-up and determined by trained physicians or the HD follow-up panel.

Statistical analysis

HD patients were divided into four groups based on the follow-up average TG/HDL-C quartile, and the reference group was the quartile 1 group. Continuous variables are represented as the mean±SD or median (IQR). Between-group comparisons were conducted by applying an analysis of variance or the Kruskal-Wallis test. The SNK (Student-Newman-Keuls) test or Nemenyi method was applied to compare the quartile-to-quartile comparisons within groups. The χ2 test (or Fisher’s exact test) was applied to compare categorical variables that are described as frequencies (proportions). Univariate and multivariate time-varying Cox regression analyses and restricted cubic spline curves were built to assess the relationship between serum TG/HDL-C levels and all-cause and CV mortality rates. The C-index was applied to assess the predictive ability of the Cox regression models. The findings are presented as HRs and 95% CIs. SPSS V.27.0 software (IBM, Armonk, New York, USA) and R V.4.2.2 software were used for all statistical analyses, and a p value of less than 0.05 (two-tailed) was considered statistically significant.


Baseline characteristics

Ultimately, 534 patients on maintenance HD were enrolled in the analysis. The mean age of the patients was 53.7±13.3 years, and 62.7% were male. During the 61-month median follow-up time, 207 HD patients died from any cause, with 94 (45.4%) of those deaths due to CVD. HD patients were divided into four groups based on the quartiles of the follow-up averaged TG/HDL-C ratio: quartile 1 (TG/HDL-C <1.05, n=134), quartile 2 (1.05≤TG/HDL-C<1.64, n=133), quartile 3 (1.64≤TG/HDL-C<2.64, n=133) and quartile 4 (TG/HDL-C ≥2.64, n=134). Table 1 summarises the demographic and laboratory characteristics between the four groups of patients as stratified by the quartiles of the ratio. The differences were statistically significant (p<0.05) between the four groups of patients in the cause of ESRD, diabetes, use of lipid-lowering medications and BMI, TC, TG, HDL-C, LDL-C and Alb levels.

Table 1

Baseline characteristics of individuals according to follow-up averaged TG/HDL-C ratio

TG/HDL-C ratio and mortality

As shown in table 2, the univariate Cox regression indicated that the quartile 4 group with TG/HDL-C ≥2.64 had a substantially reduced unadjusted HR for all-cause death (HR 0.62; 95% CI 0.42 to 0.93; p=0.021). According to the time-varying multivariable Cox proportional hazards models, all-cause mortality was significantly lower in the quartile 4 group (HR 0.50; 95% CI 0.33 to 0.76; p=0.001) after adjusting for age, dialysis vintage, cause of ESRD, hypertension, CVD and diabetes (model 1) and was still significant (HR 0.51; 95% CI 0.33 to 0.77; p=0.001) after further adjustment for Hb, Ca and Alb (model 2).

Table 2

Univariate and multivariate time-varying Cox regression analyses of variables related to all-cause mortality in HD patients

Moreover, compared with the other quartile groups of the TG/HDL-C ratio, the quartile 4 group with TG/HDL-C ≥2.64 was linked to lower CV mortality in univariate Cox regression analysis (HR 0.42; 95% CI 0.22 to 0.80; p=0.009). The association was not affected (HR 0.30; 95% CI 0.15 to 0.60; p=0.001) by age, dialysis vintage, cause of ESRD, hypertension, CVD or diabetes (model 1). In the fully adjusted time-varying multivariable Cox proportional hazards model (model 2), there was significantly decreased CV mortality (HR 0.31; 95% CI 0.16 to 0.62; p=0.001) in the quartile 4 group. The results are presented in table 3.

Table 3

Univariate and multivariate time-varying Cox regression analyses of variables related to cardiovascular mortality in HD patients

The restricted cubic spline curve analysis illustrated the non-linear relationship between time-varying TG/HDL-C and mortality (online supplemental figures 1 and 2). A high TG/HDL-C ratio was associated with reduced all-cause mortality and CV mortality. Figure 1 shows the time-varying all-cause and CV mortality HRs (and 95% CI differences) by the TG/HDL-C ratio.

Figure 1

A high TG/HDL-C ratio was associated with decreased mortality in HD patients. (A) Time-varying all-cause mortality HR (comparison of 95% CI differences); model 1: adjusted for age, dialysis vintage, cause of ESRD, hypertension, CVD and diabetes; model 2: adjusted for model 1 covariates and laboratory parameters including Hb, Ca and Alb. (B) Time-varying cardiovascular mortality HR (comparison of 95% CI differences); model 1: adjusted for age, dialysis vintage, cause of ESRD, hypertension, CVD and diabetes; model 2: adjusted for model 1 covariates and laboratory parameters including PTH, Ca, P and Alb. Alb, albumin; BMI, Body Mass Index; CVD, cardiovascular disease; ESRD, end-stage renal disease; Hb, haemoglobin; HD, haemodialysis; HDL-C, high-density lipoprotein cholesterol; PTH, parathyroid hormone; TG, triglyceride.

As table 4 shows, model 1 of all-cause mortality yielded a C-index of 0.72 (95% CI 0.68 to 0.75), and model 2 yielded a C-index of 0.77 (95% CI 0.73 to 0.82). The C-index for model 1 in CV mortality was 0.74 (95% CI 0.70 to 0.77). Model 2 for CV mortality, which had a C-index of 0.80 (95% CI 0.75 to 0.84), was superior to model 1.

Table 4

C-index of the univariate and multivariate Cox regression analyses


Study summary

In this retrospective cohort study, we estimated the link between the time-varying TG/HDL-C ratio and all-cause death and CV mortality in Chinese patients undergoing maintenance HD. We concluded that in HD patients, the quartile 4 group (TG/HDL-C ≥2.64) had decreased all-cause death and CV mortality. After adjustment for confounders, high TG/HDL-C was linked to a superior overall survival rate.

The utility and implications of the TG/HDL-C ratio

The traditional cholesterol hypothesis considers high TC, high TG and high level of LDL-C as risk factors for CV and all-cause mortality. However, it is found that high level of LDL-C is a negative risk factor for mortality in the general population, especially in individuals older than 60 years, contrary to the traditional cholesterol hypothesis.12 And TC is almost unrelated or inversely associated with atherosclerosis and CVD.13 As elevated cholesterol, one of the most important components of organelle membranes, is not the cause of atherosclerosis, but rather one of the changes that occur when the body is exposed to oxidative stress-induced inflammation or other causes that contribute to atherosclerosis.3 4 Moreover, a U-shaped relationship was observed between non-fasting TG levels and fatal CVD events in the Japanese general population.14 The association between hyperlipidaemia and increased mortality in the general population is in doubt.

Similarly, in special populations of HD patients, the impact of lipids on mortality is controversial.15 Studies have shown that reducing LDL-C levels did not improve overall survival in either HD patients or the general population.3 16 17 Moreover, lower TC was discovered to be a risk factor for death and was linked to a poor prognosis in maintenance HD patients.18 19 A previous study showed that hypertriglyceridaemia is paradoxically related to better survival.5 Lamprea-Montealegre et al noticed that TGs and triglyceride-rich lipoproteins were not related to CV events in patients with diabetes on HD.20 And several prior studies noted that high plasma HDL cholesterol concentrations are not linked to a better prognosis in HD patients, but they may be related to a worse prognosis.21–23 Apart from the baseline data study, the time-dependent data study reached similar conclusions, with high TC, non-HDL-C and LDL-C being associated with decreased mortality.24

In addition, some studies have applied combined biomarkers to accurately explore the relationship between lipids with mortality and to potentially predict long-term prognosis in HD patients. The ratio of TG to HDL-C has recently been discovered to be a biomarker for CV events and mortality, with a stronger predictive value than individual lipid parameters in the general population.6 The TG/HDL-C ratio is an uncomplicated, inexpensive, convenient and readily available calculated measure. Additionally, the ratio has an accurate predictive ability of LDL phenotype and HDL subclass distribution.25 It describes a balanced relationship between atherogenic and protective lipoproteins. Compared with traditional lipid profiles, TG/HDL-C has been discovered to be a more accurate risk indicator of all-cause death and CV events,8–10 26 metabolic syndrome,27 insulin resistance28 and incidence of diabetes29 30 in the general population. The ratio is also used in the prediction of mortality in patients with diabetes.31 However, because of the unique characteristics of HD patients, it is not appropriate to extrapolate these findings to HD patients.

The number of studies on the relationship between TG/HDL-C levels and mortality in HD patients is limited, and the results are controversial. A few studies have shown that TG/HDL-C is an independent indicator of death in PD populations, especially in older and female PD patients.32–34 However, PD patients have a more atherogenic lipid profile and a more altered state of dyslipidaemia than HD patients.35 A large retrospective analysis of 50 673 American HD patients found that the TG/HDL-C ratio is a protective factor for mortality.36 Gonzáles-Rubianes et al pointed out that three studies focused on the Asian HD population showed that a higher TG/HDL-C ratio was associated with higher mortality.37 A U-shaped relationship between atherogenic index of plasma (AIP, the logarithmic ratio of TG/HDL-C) and all-cause death in dialysis populations was discovered in a Korean study, with both the greatest and lowest AIP groups being related to high mortality.38 A study focused on Taiwan patients reported that high TG/HDL-C levels were related to worse CV events and death in dialysis populations, but the association did not exist in diabetic dialysis patients.39 This study included both HD patients and PD patients, which may have contributed to the different outcomes. A Japanese study of HD patients concluded that TG/HDL-C provided considerable predictive value for detecting CV events.40 However, in our study, we concluded that the TG/HDL-C ratio is a protective factor for CV and all-cause death, which is not completely consistent with the findings of these three studies. Northeastern China, Korea, Japan and Taiwan differ in their diet, genetics, weather and environment. Therefore, the impact of TG/HDL-C levels on death remains controversial in HD populations from different geographic areas. Our study provided a clinical assessment of the northeast Chinese HD population to evaluate the utility of the TG/HDL-C ratio as a predictor of mortality.


The mechanisms behind the link between TG/HDL-C ≥2.64 and greater HD patient survival are unclear, and the explanations may be as follows. Malnutrition may play an important role in the inverse relationship between lipid levels and mortality.41 Low TGs indicate poor nutritional status, as TGs, which are in preparation for starvation during storage, are vital for the supply and metabolism of energy and fatty acids, and loss of fat stores is likely to increase the risk of CV or cerebrovascular disease and even death.42 Given that TG acts as a fat store in the body, unless its concentration is extremely high, it is negatively correlated with poor outcomes.14 What’s more, a prior study found that the nutritional status of patients with high HDL-C was worse.43 Another explanation for the increased mortality of people with a low TG/HDL-C ratio is the role of inflammation and oxidative stress. HD patients are particularly susceptible to inflammation and oxidative stress owing to the accumulation of uraemic toxins and dialysis treatment itself.44 Under inflammation and oxidative stress, HDL-C particles might be converted to pro-oxidant and pro-inflammatory HDL-C particles, which induce endothelial dysfunction and atherosclerosis by binding to serum amyloid A.45 46 A prospective study found that HD patients with high HDL-C levels had increased levels of oxidised HDL-C.43 It is obvious that higher levels of harmful HDL-C particles are related to a higher mortality rate in HD patients. The existence of these non-traditional risk factors (malnutrition, inflammation and oxidative stress) that correlated with both low TG/HDL-C levels and elevated death rates may contribute to the increased rates of all-cause and CV death in HD populations with low TG/HDL-C, which contrasts with the findings in the general population. Both malnutrition and inflammation, which we usually consider together because of their mutually reinforcing effects and many identical indicators (such as albumin), contribute to the poor prognosis of HD patients. Although after adjusting for albumin, our findings remained significant, albumin was still strongly correlated with mortality in the adjusted model. To better understand the role of inflammation and malnutrition in HD patients, future research should look at more objective nutritional and inflammatory assessments and their impact on the relationship between TG/HDL-C levels and outcomes in HD populations.

Moreover, HD treatment itself may influence the function of lipids. HDL-C may be dysfunctional and have altered function. In HD patients, the composition and activity of HDL are altered,47 and HDL-C has a distinctly impaired capacity to elicit cholesterol efflux and a harmful effect by potentiating inflammation, which is associated with a higher risk of death from CVD.48 A previous study revealed that TG/HDL-C is a good indicator of impaired HDL-C maturation and reduced reverse cholesterol transport ability.49 Our results indicate that a high TG/HDL-C ratio is related to a better prognosis in HD populations, but further studies need to be performed to fully clarify the underlying mechanisms involved.


First, this was an observational study, the nature of which precludes us from drawing conclusions about a causal relationship. Second, because of the limited size of our study, we did not adjust for all potential risk factors that might affect survival. In this way, the interference of residual effects cannot be completely eliminated.


In conclusion, the quartile 4 group (TG/HDL-C ≥2.64) is associated with decreased all-cause and CV death in Chinese patients undergoing HD therapy. The association remained significant after adjustment for relevant confounders. Using this ratio, clinicians would be able to tailor long-term management strategies to improve long-term outcomes in this population.

Data availability statement

Data are available upon reasonable request.

Ethics statements

Patient consent for publication

Ethics approval

The General Hospital of Northern Theater Command’s Ethics Committee (Y (2022) 057) exempted this study. The Ethical Committee waived the requirement for informed consent since this study was retrospective, non-interventional and non-intrusive.


We appreciate the support of colleagues from the Department of Blood Purification, General Hospital of Northern Theater Command.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Contributors JB and NC planned the work. LQ and AZ performed the statistical analyses. YZ, KR, ZR, CZ and QW participated in patient inclusion and demographic data collection. LQ drafted the manuscript. All the authors have read and approved the final manuscript. JB and NC were the guarantors and accepted full responsibility for the work and/or the conduct of the study, had access to the data, and controlled the decision to publish.

  • Funding This work was supported by Liaoning Provincial Science and Technology Plan Joint Fund (2023-MSLH-350).

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.