ABSTRACT
Background
Peritoneal dialysis (PD) patients have a high risk of abnormal glucose and lipids metabolism.
Objective
We investigated the effects of baseline fasting plasma glucose (FPG) as well as its interaction with lipid profiles on all-cause and cardiovascular disease (CVD) cause-specific mortality in PD patients.
Methods
A total of 1995 PD patients were enrolled. Kaplan–Meier survival curves and Cox regression models were performed to assess the association of FPG levels with mortality in PD patients.
Results
During a median (25th–75th quartile) follow-up period of 48.1 (21.8–77.9) months, 567 (28.4%) patients died, including 282 (14.1%) CVD deaths. Kaplan–Meier survival curves showed that all-cause and CVD cause-specific mortality increased significantly with elevated baseline FPG levels (Log-rank tests: both P-values <.001). However, with adjustment for potential confounding factors, baseline FPG levels were not significantly associated with all-cause and CVD cause-specific mortality. Nevertheless, a significant interaction between baseline FPG and low-density lipoprotein cholesterol (LDL-C) on all-cause mortality was found (P for interaction test: .013), and subgroup analyses further showed that all-cause mortality was significantly increased for baseline FPG ≥7.0 mmol/L compared with the normal reference (FPG <5.6 mmol/L) (hazard ratio 1.89, 95% confidence interval 1.11–3.23, P-value = .020) for patients with LDL-C ≥3.37 mmol/L only, but not for those with lower LDL-C levels (<3.37 mmol/L).
Conclusion
The significant interaction effect between baseline FPG and LDL-C on all-cause mortality showed that, for PD patients with LDL-C ≥3.37 mmol/L, higher FPG levels (≥7.0 mmol/L) were significantly associated with an increased risk of all-cause mortality and need more intensive management of their FPG by clinicians in the future.
Keywords: fasting plasma glucose, interaction, low-density lipoprotein cholesterol, mortality, peritoneal dialysis
INTRODUCTION
The number of end-stage renal disease (ESRD) patients is increasing at an alarming rate worldwide. The prevalence of dialysis patients in China is predicted to be 629.67 per million population (874 373 patients) in 2025 [1]. Peritoneal dialysis (PD) has been widely used for the long-term treatment of ESRD, which has a similar survival rate to patients receiving in-center hemodialysis (HD) [2]. PD offers some advantages, including better preservation of residual renal function, minimal cardiac stress, and more gradual and continuous fluid and solute clearance [2].
Glucose-based dialysate is the most frequently used dialysate in PD patients. It is estimated that the amount of glucose absorption from dialysate varies from 89 to 316 g/day, accounting for 20.3% of the total energy intake [3], thus PD patients have a higher risk of abnormal glucose metabolism and developing new-onset diabetes compared with those treated with HD [4, 5]. However, the association between baseline fasting plasma glucose (FPG) and mortality of PD patients is still controversial. Szeto et al. reported that baseline FPG level was an independent factor for survival in PD patients [6], while Chung et al. found no association between baseline FPG and all-cause mortality in the Korean PD population [7].
Dyslipidemia often occurs in PD patients, although the mechanisms have not been fully elucidated. Abundant evidence has demonstrated that dyslipidemia in PD patients, such as elevated very low-density lipoprotein cholesterol (VLDL-C) and elevated non-high-density lipoprotein cholesterol (non-HDL-C), were independently associated with increased risk of all-cause and cardiovascular disease (CVD) cause-specific mortality in PD patients [8, 9].
Glucose and lipid metabolism are related to each other in PD patients—long-term exposure to high glucose solution may contribute to dyslipidemia [10]. A low-glucose PD regimen (combination of the dextrose-based solution, icodextrin and amino acids) has been shown to significantly improve serum triglyceride (TG), VLDL-C and ApoB levels compared with the traditional glucose dialysate regimen (dextrose solutions only) [11]. In addition, hyperlipidemia was associated with a higher risk of all-cause mortality only in PD patients with diabetes mellitus (DM) status, but not in non-DM patients [12]. However, it is unknown whether there is an interaction between FPG and lipids that affects the risk of mortality in PD patients.
In this study, we aim to assess the association of baseline FPG with long-term prognosis and identify the interaction effect between baseline FPG and lipid profiles on all-cause and CVD cause-specific mortality in a large 15-year contemporary cohort of incident PD patients.
MATERIALS AND METHODS
Study subjects
This was a single-center cohort study. All incident PD patients at the First Affiliated Hospital, Sun Yat-sen University were recruited from 1 January 2006 to 31 December 2016. Eligible participants were aged 18 years or older and underwent PD for more than 3 months. Patients who had incomplete FPG and serum lipid profiles at baseline, a history of malignancy, transferred from HD or failed renal transplantation, or were taking long-term systemic corticosteroids were excluded. The study was conducted in accordance with the Declaration of Helsinki and was approved by the Research Ethics Committee of The First Affiliated Hospital of Sun Yat-sen University ([2016]215). All participants provided written informed consent.
Clinical and laboratory data collection
Baseline data were collected in the first 3 months of PD treatment. The comorbidity score was determined according to the Charlson comorbidity index (CCI) [13]. The presence of diabetes was defined as follows: (i) FPG ≥7.0 mmol/L, (ii) 2 h plasma glucose ≥11.1 mmol/L during an oral glucose tolerance test or (iii) glycated hemoglobin (HbA1c) ≥6.5%, (iv) diabetes symptoms plus random plasma glucose ≥11.1 mmol/L or (v) the use of diabetes medication, or (vi) a self-reported history of physician diagnosis. Diabetes symptoms are polydipsia, polyuria, polyphagia and unexplained weight loss. If patients had no symptoms of diabetes and only once had hyperglycemia, criteria (i) to (iii) were confirmed by repeated testing [14]. CVD was defined as ischemic heart disease, congestive heart failure, cerebrovascular disease and peripheral vascular disease [15]. Residual renal function (RRF) and the total Kt/V were calculated using PD Adequest 2.0 for Windows software (Baxter Healthcare Corporation, Chicago, IL, USA). RRF was estimated as an average of 24 h creatinine and urea clearance, corrected for body surface area [16]. Peritoneal glucose load was calculated by the product of glucose concentration and the volume of each exchange, as previously described [17]. Lipid-lowering and hypoglycemic drugs prescribed during the study period were collected.
Study outcomes
The primary outcomes were all-cause and CVD cause-specific mortality. CVD death was determined by the diagnosis codes according to International Classification of Disease, Tenth Revision: acute myocardial ischemia or infarction (I20–I21), congestive heart failure (I50), arrhythmia (I44–I45, I47–I49), cardiac arrest (I46), cardiomyopathy (I42), cerebrovascular accident (I60–I64) and peripheral vascular disease (I70–I71) [18]. To determine the cause of death, three nephrologists at our PD center thoroughly reviewed medical records and/or directly communicated with the referring physician. All patients were censored until death, transfer to HD, kidney transplantation, transfer to other dialysis centers, loss to follow-up or end of the study period on 31 December 2020.
Statistical analysis
Data are presented as numbers (percentages) for categorical variables and mean ± standard deviation (SD) for normally distributed continuous variables as well as median (25th–75th quartiles) for skewed continuous variables. Patients were divided into three categories according to their baseline FPG levels (<5.6 mmol/L, 5.6 ≤ FPG < 7.0 mmol/L, and ≥7.0 mmol/L), per the American Diabetes Association 2021 criteria [14]. Baseline characteristics were compared across FPG categories, using one-way ANOVA, Kruskal–Wallis tests or chi-squared tests, as appropriate.
The cumulative probabilities of survival were analyzed by Kaplan–Meier survival curves, and their differences were assessed by the log-rank tests. Cox proportional hazards models were conducted to examine associations of baseline FPG levels with all-cause and CVD cause-specific mortality. Baseline FPG levels were shown as both categorical and continuous values separately. The proportional hazards assumptions were verified by Schoenfeld residual tests. For the Cox regression analysis, candidate covariates were thought to be related to FPG levels, including age, body mass index (BMI), CCI, albumin, calcium, phosphorus, TC, TG, uric acid, RRF, high-sensitivity C-reactive protein (hs-CRP) level, glucose load and treatment of hypoglycemic drugs; those considered as risk factors of patients’ survival, including gender and hemoglobin, were chosen for multivariate analysis. In the subgroup analysis, lipid levels were classified according to the National Cholesterol Education Program Adult Treatment Panel guidelines [19], and the interaction effect between baseline FPG and lipid profiles was tested further. Additionally, we performed a sensitivity analysis that stratified LDL-C at different cutoff levels to investigate the interaction effect between FPG and LDL-C. Results were expressed as the hazard ratio (HR) and 95% confidence interval (CI). A two-tailed P-value <.05 was considered statistically significant in all analyses. Statistical analyses were performed by using SPSS, version 22.0 for Windows (SPSS Inc., Chicago, IL, USA).
RESULTS
Baseline individual characteristics
A total of 2384 patients undergoing PD in our center were prescreened for the study, of which 389 were excluded. Finally, 1995 patients were eligible for this study (Fig. 1). The average age was 47.1 ± 15.1 years, 1216 (60.9%) were male and the mean BMI was 21.7 ± 3.1 kg/m2. The percentages of diabetes and CVD were 26.2% and 34.4%, respectively (Table 1).
Figure 1:
Flow chart of the participants in the study.
Table 1:
Baseline characteristics of individuals stratified by baseline FPG levels.
| FPG (mmol/L) | |||||
|---|---|---|---|---|---|
| Total | <5.6 | 5.6 ≤ FPG < 7.0 | ≥7.0 | ||
| Variable | (n = 1955) | (n = 1449) | (n = 265) | (n = 281) | P-valuea |
| Demographics | |||||
| Age (years) | 47.1 ± 15.1 | 43.5 ± 14.3 | 53.6 ± 13.8 | 59.7 ± 10.4 | <.001 |
| Men (n, %) | 1216 (60.9) | 886 (61.1) | 164 (61.9) | 166 (59.1) | .765 |
| BMI (kg/m2) | 21.7 ± 3.1 | 21.4 ± 3.07 | 22.4 ± 3.12 | 22.6 ± 2.96 | <.001 |
| Comorbidities | |||||
| History of diabetes (n, %) | 523 (26.2) | 141 (9.7) | 101 (38.1) | 281 (100.0) | <.001 |
| History of hypertension (n, %) | 509 (25.5) | 322 (20.8) | 61 (37.2) | 126 (44.8) | <.001 |
| History of CVD (n, %) | 687 (34.4) | 460 (29.7) | 74 (45.1) | 153 (54.4) | <.001 |
| CCI | 3.5 ± 1.8 | 2.9 ± 1.46 | 4.1 ± 1.83 | 5.8 ± 1.43 | <.001 |
| Laboratory parameters | |||||
| Hemoglobin (g/L) | 101 ± 22 | 101 ± 22 | 98 ± 22 | 102 ± 21 | .067 |
| Albumin (g/L) | 36.9 ± 5.04 | 37.6 ± 4.89 | 35.8 ± 4.98 | 34.1 ± 4.74 | <.001 |
| Hs-CRP (mg/L) | 1.67 (0.60–4.90) | 1.42 (0.53–4.21) | 2.78 (1.06–9.04) | 2.18 (0.86–6.21) | <.001 |
| Corrected calcium (mmol/L) | 2.63 ± 1.46 | 2.58 ± 1.38 | 2.67 ± 1.51 | 2.82 ± 1.76 | .043 |
| Phosphorus (mmol/L) | 1.51 ± 0.47 | 1.53 ± 0.47 | 1.51 ± 0.51 | 1.40 ± 0.44 | <.001 |
| TC (mmol/L) | 4.98 ± 1.30 | 4.89 ± 1.23 | 4.97 ± 1.28 | 5.46 ± 1.54 | <.001 |
| TG (mmol/L) | 1.38 (0.99–1.97) | 1.34 (0.98–1.89) | 1.58 (1.12–2.20) | 1.68 (1.16–2.53) | <.001 |
| HDL-C (mmol/L) | 1.21 ± 0.42 | 1.22 ± 0.40 | 1.20 ± 0.49 | 1.21 ± 0.45 | .555 |
| LDL-C (mmol/L) | 2.92 ± 0.98 | 2.88 ± 0.93 | 2.86 ± 0.99 | 3.22 ± 1.19 | <.001 |
| Uric acid (μmol/L) | 428 ± 95 | 435 ± 94 | 423 ± 105 | 395 ± 83 | <.001 |
| Creatinine (μmol/L) | 722 (583–922) | 754 (614–956) | 683 (569–840) | 571 (467–682) | <.001 |
| Total Kt/v | 2.47 ± 0.71 | 2.45 ± 0.72 | 2.50 ± 0.68 | 2.61 ± 0.70 | .004 |
| RRF (mL/min/1.73 m2) | 3.35 (1.97–5.29) | 3.27 (1.92–5.15) | 3.22 (1.90–5.45) | 4.10 (2.75–5.88) | <.001 |
| 24 h urine output (mL) | 1000 (600–1500) | 1000 (600–1500) | 1000 (525–1500) | 900 (600–1400) | .014 |
| Glucose load (g/day) | 130 ± 23 | 129 ± 22 | 132 ± 25 | 135 ± 27 | .001 |
| Medications | |||||
| Lipid-lowering treatment (n, %) | 227 (11.4) | 130 (9.2) | 29 (11.1) | 68 (24.6) | <.001 |
| Hypoglycemic drugs (n, %) | 500 (25.1) | 139 (9.6) | 93 (35.1) | 268 (95.4) | <.001 |
Values expressed as mean ± SD or median (interquartile range) for continuous variables and number (%) for categorical variables.
Kt/V urea, urea clearance (Kt) normalized to total body water (V).
P < .05 is considered statistically significant. Data shown in bold-type means P < 0.05.
Baseline demographic and clinical characteristics of patients categorized according to baseline FPG levels are shown in Table 1. Patients with higher baseline FPG had significantly higher age, BMI, proportion of diabetes, CCI, hs-CRP, corrected-calcium, TC, TG, LDL-C, total Kt/V, RRF and glucose load, and higher percentages of lipid-lowing and glycemic treatment; however, they had lower levels of albumin, phosphorus, uric acid, creatinine and 24 h urine output (P < .05). No significant differences in gender, hemoglobin or HDL-C were found across the groups. We further classify patients by the low and high levels of LDL-C within each category of FPG, the demographic and clinical characteristics of which can be seen in Supplementary data, Table S1.
Association of baseline FPG with all-cause and CVD cause-specific mortality
During a median follow-up period of 48.1 (21.8–77.9) months, 567 (28.4%) patients died, including 282 (14.1%) CVD deaths. Kaplan–Meier survival curves showed that the cumulative rates of overall survival (Fig. 2A) and cardiovascular death-free survival (Fig. 2B) decreased significantly with increased baseline FPG levels. The association between baseline FPG and all-cause and CVD cause-specific mortality was determined by multivariate Cox proportional hazards regression analysis. As shown in Table 2, in the unadjusted model (Model 1), when FPG was examined as a continuous variable, the baseline FPG level was significantly associated with increased risks of all-cause and CVD cause-specific mortality [HR (95% CI) 1.13 (1.11–1.16), P < .001 and 1.13 (1.10–1.17), P < .001, respectively]. Compared with the reference group (FPG <5.6 mmol/L), higher FPG levels (both 5.6 ≤ FPG < 7.0 mmol/L and FPG ≥7.0 mmol/L) had significantly increased risks of all-cause [HR (95% CI) 1.76 (1.35–2.30), P < .001 and 3.23 (2.66–3.91), P < .001, respectively] and CVD cause-specific mortality [HR (95% CI) 1.69 (1.21–2.34), P = .002 and 3.32 (2.53–4.37), P < .001, respectively]. There were significantly positive trends between increasing categories of FPG levels and risks of all-cause and CVD cause-specific mortality (both trend tests: P-values <.001). When adjusted for age, gender and BMI (Model 2), baseline FPG was still significantly associated with increased risks of all-cause and CVD cause-specific mortality, and the results did not change much. Nevertheless, with additional adjustment for TC, TG, corrected calcium, phosphorus, CCI, hemoglobin, albumin, uric acid, RRF, hs-CRP and glucose load in Model 3, baseline FPG level (either continuous or categorical) was not significantly associated with all-cause or CVD cause-specific mortality. The significantly positive trends between increasing categories of FPG levels and risks of all-cause and CVD cause-specific mortality disappeared. The same trend results as in Model 3 were obtained after further adjustment for hypoglycemic treatment (Model 4). In the sensitivity analysis, the associations between baseline FPG and clinical outcomes were similar among patients with or without diabetes (Supplementary data, Table S2 and S3).
Figure 2:
Kaplan–Meier survival curves for (A) all-cause and (B) CVD cause-specific mortality of all PD patients.
Table 2:
Associations of FPG with all-cause and cardiovascular disease mortality.
| Model 1a | Model 2b | Model 3c | Model 4d | |||||
|---|---|---|---|---|---|---|---|---|
| HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value | |
| All-cause mortality | ||||||||
| Continuous FPG | 1.13 (1.11–1.16) | <.001 | 1.07 (1.04–1.10) | <.001 | 1.01 (0.97–1.05) | .655 | 0.99 (0.95–1.04) | .769 |
| FPG categories | ||||||||
| FPG <5.6 | 1.00 | 1.00 | 1.00 | 1.00 | ||||
| 5.6 ≤ FPG < 7.0 | 1.76 (1.35–2.30) | <.001 | 1.13 (0.89–1.43) | .313 | 0.87 (0.65–1.16) | .342 | 0.87 (0.65–1.16) | .342 |
| FPG ≥7.0 | 3.23 (2.66–3.91) | <.001 | 1.70 (1.39–2.09) | <.001 | 0.93 (0.70–1.25) | .638 | 0.93 (0.70–1.25) | .638 |
| P for trend test | <.001 | <.001 | .629 | .629 | ||||
| Cardiovascular disease mortality | ||||||||
| Continuous FPG | 1.13 (1.10–1.17) | <.001 | 1.06 (1.02–1.10) | .005 | 1.01 (0.96–1.07) | .643 | 1.01 (0.95–1.06) | .985 |
| FPG categories | ||||||||
| FPG <5.6 | 1.00 | 1.00 | 1.00 | 1.00 | ||||
| 5.6 ≤ FPG < 7.0 | 1.69 (1.21–2.34) | .002 | 1.12 (0.80–1.56) | .510 | 0.99 (0.67–1.45) | .984 | 0.99 (0.67–1.48) | .971 |
| FPG ≥7.0 | 3.32 (2.53–4.37) | <.001 | 1.67 (1.25–2.23) | .001 | 1.18 (0.81–1.74) | .375 | 1.09 (0.73–1.63) | .671 |
| P for trend test | <.001 | .002 | .629 | .896 | ||||
Unadjusted model.
Adjusted for age, gender, BMI.
Adjusted for age, gender, BMI, TC, TG, corrected calcium, phosphorus, CCI, hemoglobin, albumin, uric acid, RRF, hs-CRP level and glucose load.
Adjusted for age, gender, BMI, TC, TG, corrected calcium, phosphorus, CCI, hemoglobin, albumin, uric acid, RRF, hs-CRP level, glucose load and hypoglycemic treatment.
P < .05 is considered statistically significant. Data shown in bold-type means P < 0.05.
Interaction of baseline FPG and lipid profiles and their relation to mortality
Interaction tests between baseline FPG and lipid profiles as well as subgroup analyses across lipid profiles to evaluate the effects of FPG on all-cause mortality are shown in Table 3. A significant interaction effect between FPG categories and LDL-C categories on all-cause mortality was found (P = .013). Further subgroup analysis by LDL-C categories showed that, for patients with higher LDL-C (≥3.37 mmol/L) levels, higher baseline FPG levels (≥7.0 mmol/L) were significantly associated with an elevated risk of all-cause mortality [HR (95% CI) 1.89 (1.11–3.23), P = .020]; however, for patients with lower LDL-C levels (<3.37 mmol/L), baseline FPG levels were no longer associated with the risk of all-cause mortality. No other significant interactions between FPG and lipid profiles (TC, TG or HDL-C) were found for all-cause mortality. There is no significant interaction effect between baseline FPG and lipid profiles on CVD cause-specific mortality, and subgroup analyses stratified across lipid profiles show similar results (Table 4). Furthermore, when LDL-C was stratified at 3.0 mmol/L, sensitivity analysis showed significant interactions between baseline FPG and LDL-C on all-cause and CVD cause-specific mortality (P = .008 and P = .021) (Supplementary data, Table S4 and S5); subgroup analyses showed that, for patients with LDL-C ≥3.0 mmol/L, higher baseline FPG levels (≥7.0 mmol/L) were significantly associated with increased risks of all-cause and CVD cause-specific mortalities, with the HRs (95% CI) of 1.59 (1.02–2.46, P = .039) and 1.79 (1.01–3.17, P = .048), respectively. Moreover, the number and rate of deaths in each category of FPG and LDL-C are shown in Supplementary data, Table S6. Patients with higher FPG and LDL-C had a higher rate of all-cause and CVD death.
Table 3:
Associations of FPG with all-cause mortality stratified by lipid profiles.a
| FPG <5.6 mmol/L | 5.6 ≤ FPG < 7.0 mmol/L | FPG ≥7.0 mmol/L | ||||||
|---|---|---|---|---|---|---|---|---|
| Lipid profiles | HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value | P for trend | P for interactionb |
| TC | .103 | |||||||
| <5.18 mmol/L | Ref. | 0.83 (0.57–1.21) | .333 | 0.69 (0.46–1.03) | .068 | .168 | ||
| ≥5.18 mmol/L | Ref. | 0.99 (0.61–1.61) | .980 | 1.36 (0.87–2.13) | .185 | .336 | ||
| TG | .450 | |||||||
| <1.7 mmol/L | Ref. | 0.98 (0.65–1.47) | .918 | 1.06 (0.69–1.60) | .803 | .949 | ||
| ≥1.7 mmol/L | Ref. | 0.70 (0.43–1.04) | .071 | 0.89 (0.58–1.35) | .572 | .195 | ||
| HDL-C | .244 | |||||||
| <1.04 mmol/L | Ref. | 0.65 (0.40–1.05) | .078 | 0.72 (0.46–1.12) | .144 | .112 | ||
| ≥1.04 mmol/L | Ref. | 1.07 (0.72–1.59) | .729 | 1.25 (0.83–1.88) | .293 | .572 | ||
| LDL-C | .013 | |||||||
| <3.37 mmol/L | Ref. | 0.95 (0.67–1.35) | .785 | 0.73 (0.50–1.05) | .089 | .232 | ||
| ≥3.37 mmol/L | Ref. | 1.20 (0.70–2.15) | .540 | 1.89 (1.11–3.23) | .020 | .059 | ||
Fully adjusted for age, gender, BMI, corrected calcium, phosphorus, CCI, hemoglobin, albumin, uric acid, RRF, hs-CRP level, glucose load, and lipid-lowering and hypoglycemic treatment.
FPG categories × lipid categories.
P < .05 is considered statistically significant. Data shown in bold-type means P < 0.05.
Table 4:
Associations of FPG with cardiovascular disease cause-specific mortality stratified by lipid profiles.a
| FPG <5.6 mmol/L | 5.6 ≤ FPG < 7.0 mmol/L | FPG ≥7.0 mmol/L | ||||||
|---|---|---|---|---|---|---|---|---|
| Lipid profiles | HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value | P for trend | P for interactionb |
| TC | .211 | |||||||
| <5.18 mmol/L | Ref. | 0.79 (0.44–1.41) | .420 | 0.88 (0.50–1.54) | .646 | .700 | ||
| ≥5.18 mmol/L | Ref. | 1.48 (0.83–2.66) | .186 | 1.52 (0.84–2.76) | .167 | .284 | ||
| TG | .379 | |||||||
| <1.7 mmol/L | Ref. | 1.24 (0.71–2.18) | .449 | 1.24 (0.69–2.25) | .474 | .678 | ||
| ≥1.7 mmol/L | Ref. | 0.68 (0.38–1.24) | .209 | 1.08 (0.62–1.89) | .791 | .367 | ||
| HDL-C | .585 | |||||||
| <1.04 mmol/L | Ref. | 0.77 (0.39–1.52) | .449 | 1.01 (0.55–1.87) | .964 | .734 | ||
| ≥1.04 mmol/L | Ref. | 1.17 (0.70–1.98) | .550 | 1.32 (0.75–2.30) | .338 | .621 | ||
| LDL-C | .222 | |||||||
| <3.37 mmol/L | Ref. | 1.09 (0.67–1.79) | .727 | 0.99 (0.60–1.65) | .980 | .933 | ||
| ≥3.37 mmol/L | Ref. | 1.46 (0.71–3.02) | .309 | 1.79 (0.90–3.56) | .100 | .245 | ||
Fully adjusted for age, gender, BMI, corrected calcium, phosphorus, CCI, hemoglobin, albumin, uric acid, RRF, hs-CRP level, glucose load, and lipid-lowering and hypoglycemic treatment.
FPG categories × lipid categories.
P < .05 is considered statistically significant.
DISCUSSION
In the present study, we found that higher baseline FPG was not an independent predictor of all-cause and CVD cause-specific mortality in PD patients. However, a significant interaction effect between baseline FPG and LDL-C on the risk of all-cause mortality was found, and stratified analysis showed that higher baseline FPG levels (≥7.0 mmol/L) were significantly associated with increased risk of all-cause mortality among patients with higher LDL-C levels (≥3.37 mmol/L). These results suggest that high LDL-C levels may modify the association of baseline FPG level with mortality.
Numerous studies have investigated the association between baseline FPG and the outcome in different populations. In a survey including 820 900 general population from Europe, North America and Japan, baseline FPG level had a J-shaped relationship with the risk of death [20]. Another study conducted on the rural general Chinese population showed similar results [21]. However, studies on the association of baseline FPG level and clinical outcomes in dialysis patients showed conflicting results. In three separate studies of Chinese HD [22] or PD [6, 23] patients, baseline FPG level was found to be an independent risk factor for mortality. Of note, these studies had relatively small sample sizes (693, 405 and 270, respectively). In contrast, our present study showed no significant association between baseline FPG level and all-cause and CVD cause-specific mortality in PD patients after multiple adjustments. The conflicting results may be due to the different FPG classification criteria, sample sizes and population characteristics, and not fully adjusting for potentially relevant confounders, such as glucose load and lipids. Hence, large-scale, multicenter, prospective studies are needed to verify the association between baseline FPG and outcomes in PD patients.
Hyperlipidemia is deemed to be a risk factor for mortality in PD patients [8, 9], but less attention has been paid to the interaction effect between abnormal FPG and dyslipidemia. A large cohort study including 17 902 general population Chinese showed no interaction effect between lipids and FPG on mortality [24]. However, another cohort of 2 939 Chinese PD patients found that PD patients with DM coexisting with hyperlipidemia had the highest risk of all-cause mortality than either DM or hyperlipidemia alone, and hyperlipidemia was associated with poorer long-term outcomes in DM patients, but not in non-DM patients [12]. Notably, by using 2016 Chinese diagnostic criteria for hyperlipidemia as TC ≥4.7 mmol/L, TG ≥2.3 mmol/L or LDL-C ≥4.1 mmol/L, Wei et al. did not find an interaction effect between DM and hyperlipidemia [12]. In the present study, we demonstrated an interaction effect between baseline FPG and LDL-C on all-cause mortality in PD patients. For patients with high LDL-C levels (≥3.37 mmol/L), high FPG levels (≥7.0 mmol/L) were associated with an increased risk of all-cause mortality, but not in patients with LDL-C levels <3.37 mmol/L. Unlike Wei et al. [12], our participants had higher baseline lipids levels and a longer follow-up period. Moreover, we focused more on the interaction effect between lipids and blood glucose per se rather than DM.
Lipid management in dialysis patients remains uncertain due to the relatively limited randomized controlled trials. The 2013 Kidney Disease: Improving Global Outcomes (KDIGO) Clinical Practice Guideline for Lipid Management in CKD did not specify the target values for LDL-C levels and recommended that patients beginning dialysis should not initiate statins or statin/ezetimibe therapy, except for those already receiving such therapy at the time of dialysis initiation [25]. It should be noted that the KDIGO guideline did not distinguish between HD and PD patients and disregarded other factors like comorbidity. Here we show that high LDL-C levels modify the association between baseline FPG and all-cause mortality in PD patients. The sensitivity analysis stratified LDL-C at 3.0 mmol/L demonstrated significant interactions between FPG and LDL-C on CVD cause-specific and all-cause mortality, but not at 1.8 or 2.6 mmol/L. In fact, a recent multicenter cohort study of 2028 Chinese PD patients showed that baseline LDL-C levels were not associated with CVD cause-specific and all-cause mortality after multiple adjustments [26]. These findings suggest that high blood glucose and high LDL-C may act together to increase the risk of death and that clinicians need to lower lipids appropriately in PD patients with hyperglycemia. Lipid-lowering therapy was associated with decreased risk of CVD cause-specific mortality only in diabetic PD patients but not in nondiabetic PD patients, further documenting the importance of lipid management in PD patients with hyperglycemia [27]. A possible mechanism of the interaction is that high FPG and high LDL-C may synergistically facilitate the process of atherogenesis. Namely, insulin resistance owing to peritoneal glucose absorption in PD patients [28] is closely associated with the formation of atherogenic small dense LDL-C [29–31], and glycated LDL particles due to high blood glucose could induce rapid lipid accumulation in macrophage cells [32, 33]. However, we did not observe an interaction effect between FPG and LDL-C on CVD cause-specific mortality when the cutoff was 3.37 mmol/L. This may be due to the insufficient sample size in subgroup analysis and not discriminating between CVD death of atherosclerotic or non-atherosclerotic events when evaluating the interaction effect on CVD cause-specific mortality.
Certain limitations in the present study should be recognized. First, our study was a single-center cohort study, which might limit the generalizability to extrapolate to a larger population. Second, the FPG level was only a single measurement, and we did not evaluate the relationship between longitudinal changes in FPG and the outcome of PD patients. Third, we did not rule out the effect of other unknown confounding factors due to the lack of these kinds of data, such as smoking history, alcohol consumption and adequacy of glycemic control. Finally, we cannot conclude the causal relationship between the interaction of FPG and lipids and mortality because of the observational design.
In conclusion, although the baseline FPG was not an independent predictor for poor prognosis in PD patients, a significant interaction effect between baseline FPG and LDL-C on the risk of all-cause mortality was found in the present study. For PD patients with LDL-C ≥3.37 mmol/L, higher FPG levels (≥7 mmol/L) were significantly associated with an increased risk of all-cause mortality. Therefore, our results implicated the importance of monitoring FPG for PD patients with high LDL-C levels for risk stratification.
Supplementary Material
ACKNOWLEDGEMENTS
We are grateful to all participants, the doctors, and nurses of the PD center for their efforts and contributions to this research.
Contributor Information
Yiping Xu, Department of Nephrology, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, China; NHC Key Laboratory of Clinical Nephrology (Sun Yat-Sen University) and Guangdong Provincial Key Laboratory of Nephrology, Guangzhou, China.
Zhong Zhong, Department of Nephrology, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, China; NHC Key Laboratory of Clinical Nephrology (Sun Yat-Sen University) and Guangdong Provincial Key Laboratory of Nephrology, Guangzhou, China.
Yi Li, Department of Nephrology, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, China; NHC Key Laboratory of Clinical Nephrology (Sun Yat-Sen University) and Guangdong Provincial Key Laboratory of Nephrology, Guangzhou, China.
Zhijian Li, Department of Nephrology, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, China; NHC Key Laboratory of Clinical Nephrology (Sun Yat-Sen University) and Guangdong Provincial Key Laboratory of Nephrology, Guangzhou, China.
Yi Zhou, Department of Nephrology, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, China; NHC Key Laboratory of Clinical Nephrology (Sun Yat-Sen University) and Guangdong Provincial Key Laboratory of Nephrology, Guangzhou, China.
Zhibin Li, Epidemiology Research Unit, Translational Medicine Research Center, The First Affiliated Hospital of Xiamen University, School of Medicine, Xiamen University, Xiamen, China.
Haiping Mao, Department of Nephrology, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, China; NHC Key Laboratory of Clinical Nephrology (Sun Yat-Sen University) and Guangdong Provincial Key Laboratory of Nephrology, Guangzhou, China.
FUNDING
This work was supported by the National Natural Science Foundation of China [grant number 82000677]; the NHC Key Laboratory of Clinical Nephrology (Sun Yat-Sen University) and Guangdong Provincial Key Laboratory of Nephrology [grant number 2020B1212060028]; and the Guangdong Basic and Applied Basic Research Foundation [grant number 2019B1515120075, 2022A1515012532].
AUTHORS’ CONTRIBUTIONS
H.M., Zhibin Li, Z.Z. and Y.X. designed the study. Y.L. collected the clinical data of patients. Z.Z., Y.X. and Zhibin Li performed statistical analyses. Y.X. and Z.Z. drafted the manuscript. H.M., Zhibin Li, Zhijian Li and Y.Z. checked and revised the article. All authors read and approved the final manuscript.
DATA AVAILABILITY STATEMENT
The datasets generated and/or analyzed during the current study are not publicly available but are available from the corresponding author upon reasonable request.
CONFLICT OF INTEREST STATEMENT
The authors have no conflicts of interest to declare that are relevant to the content of this article. The results presented in this paper have not been published previously in whole or part, except in abstract format.
REFERENCES
- 1. Yang C, Yang Z, Wang JWet al. Estimation of prevalence of kidney disease treated with dialysis in China: a study of insurance claims data. Am J Kidney Dis 2021;77:889–97.e1. [DOI] [PubMed] [Google Scholar]
- 2. Mehrotra R, Devuyst O, Davies SJet al. The current state of peritoneal dialysis. J Am Soc Nephrol 2016;27:3238–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Grodstein GP, Blumenkrantz MJ, Kopple JDet al. Glucose absorption during continuous ambulatory peritoneal dialysis. Kidney Int 1981;19:564–7. [DOI] [PubMed] [Google Scholar]
- 4. Xue C, Gu YY, Cui CJet al. New-onset glucose disorders in peritoneal dialysis patients: a meta-analysis and systematic review. Nephrol Dial Transplantat 2020;35:1412–9. [DOI] [PubMed] [Google Scholar]
- 5. Wang IK, Lin CL, Chen HCet al. Risk of new-onset diabetes in end-stage renal disease patients undergoing dialysis: analysis from registry data of Taiwan. Nephrol Dial Transplant 2018;33:670–5. [DOI] [PubMed] [Google Scholar]
- 6. Szeto CC, Chow KM, Kwan BCet al. New-onset hyperglycemia in nondiabetic Chinese patients started on peritoneal dialysis. Am J Kidney Dis 2007;49:524–32. [DOI] [PubMed] [Google Scholar]
- 7. Chung SH, Han DC, Noh Het al. High blood glucose independent of pre-existing diabetic status predicts mortality in patients initiating peritoneal dialysis therapy. Int Urol Nephrol 2015;47:1017–24. [DOI] [PubMed] [Google Scholar]
- 8. Xie X, Zhang X, Xiang Set al. Association of very low-density lipoprotein cholesterol with all-cause and cardiovascular mortality in peritoneal dialysis. Kidney Blood Press Res 2017;42:52–61. [DOI] [PubMed] [Google Scholar]
- 9. Yu J, Xia X, Lin Tet al. Non-high-density lipoprotein cholesterol and mortality among peritoneal dialysis patients. J Clin Lipidol 2021;15:732–42. [DOI] [PubMed] [Google Scholar]
- 10. Mehrotra R, de Boer IH, Himmelfarb J.. Adverse effects of systemic glucose absorption with peritoneal dialysis: how good is the evidence? Curr Opin Nephrol Hypertens 2013;22:663–8. [DOI] [PubMed] [Google Scholar]
- 11. Li PK, Culleton BF, Ariza Aet al. Randomized, controlled trial of glucose-sparing peritoneal dialysis in diabetic patients. J Am Soc Nephrol 2013;24:1889–900. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Wei X, Wen Y, Zhou Qet al. Hyperlipidemia and mortality associated with diabetes mellitus co-existence in Chinese peritoneal dialysis patients. Lipids Health Dis 2020;19:234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Charlson ME, Pompei P, Ales KLet al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83. [DOI] [PubMed] [Google Scholar]
- 14. American Diabetes Association . 2. Classification and diagnosis of diabetes. Diabetes Care 2021;44:S15–33. [DOI] [PubMed] [Google Scholar]
- 15. Weiner DE, Tighiouart H, Vlagopoulos PTet al. Effects of anemia and left ventricular hypertrophy on cardiovascular disease in patients with chronic kidney disease. J Am Soc Nephrol 2005;16:1803–10. [DOI] [PubMed] [Google Scholar]
- 16. Gehan EA, George SL.. Estimation of human body surface area from height and weight. Cancer Chemother Rep 1970;54:225–35. [PubMed] [Google Scholar]
- 17. Jiang N, Qian J, Lin Aet al. Initiation of glucose-based peritoneal dialysis is associated with increased prevalence of metabolic syndrome in non-diabetic patients with end-stage renal disease. Blood Purif 2008;26:423–8. [DOI] [PubMed] [Google Scholar]
- 18. Wang AY, Wang M, Woo Jet al. Cardiac valve calcification as an important predictor for all-cause mortality and cardiovascular mortality in long-term peritoneal dialysis patients: a prospective study. J Am Soc Nephrol 2003;14:159–68. [DOI] [PubMed] [Google Scholar]
- 19. National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) . Third Report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III) final report. Circulation 2002;106:3143–421. [PubMed] [Google Scholar]
- 20. Rao Kondapally Seshasai S, Kaptoge S, Thompson Aet al. Diabetes mellitus, fasting glucose, and risk of cause-specific death. N Engl J Med 2011;364:829–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Liu L, Chen X, Liu Yet al. The association between fasting plasma glucose and all-cause and cause-specific mortality by gender: the rural Chinese cohort study. Diabetes Metab Res Rev 2019;35:e3129. [DOI] [PubMed] [Google Scholar]
- 22. Lin-Tan DT, Lin JL, Wang LHet al. Fasting glucose levels in predicting 1-year all-cause mortality in patients who do not have diabetes and are on maintenance hemodialysis. J Am Soc Nephrol 2007;18:2385–91. [DOI] [PubMed] [Google Scholar]
- 23. Xu L, Hu X, Chen W.. Fibroblast growth factor-23 correlates with advanced disease conditions and predicts high risk of major adverse cardiac and cerebral events in end-stage renal disease patients undergoing continuous ambulatory peritoneal dialysis. J Nephrol 2019;32:307–14. [DOI] [PubMed] [Google Scholar]
- 24. Cheng N, Zhang Y, Yang Jet al. Association between fasting blood glucose and all-cause mortality in a rural Chinese population: 15-year follow-up cohort study. Diabetes Ther 2020;11:2691–701. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Kidney Disease: Improving Global Outcomes (KDIGO) Lipid Working Group . KDIGO clinical practice guideline for lipid management in chronic kidney disease. Kidney Int. Suppl. 2013; 3(3):259, –305. [Google Scholar]
- 26. Xie QH, Shang D, Wang YJet al. The association between baseline serum lipids and mortality in peritoneal dialysis patients. Blood Purif 2022;51:101–10. [DOI] [PubMed] [Google Scholar]
- 27. Goldfarb-Rumyantzev AS, Habib AN, Baird BCet al. The association of lipid-modifying medications with mortality in patients on long-term peritoneal dialysis. Am J Kidney Dis 2007;50:791–802. [DOI] [PubMed] [Google Scholar]
- 28. Tuzcu A, Bahceci M, Yilmaz MEet al. The determination of insulin sensitivity in hemodialysis and continuous ambulatory peritoneal dialysis in nondiabetic patients with end-stage renal disease. Saudi Med J 2005;26:786–91. [PubMed] [Google Scholar]
- 29. Adiels M, Olofsson SO, Taskinen MRet al. Overproduction of very low-density lipoproteins is the hallmark of the dyslipidemia in the metabolic syndrome. Arterioscler Thromb Vasc Biol 2008;28:1225–36. [DOI] [PubMed] [Google Scholar]
- 30. Vergès B. Pathophysiology of diabetic dyslipidaemia: where are we? Diabetologia 2015;58:886–99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Vakkilainen J, Steiner G, Ansquer JCet al. Relationships between low-density lipoprotein particle size, plasma lipoproteins, and progression of coronary artery disease: the Diabetes Atherosclerosis Intervention Study (DAIS). Circulation 2003;107:1733–7. [DOI] [PubMed] [Google Scholar]
- 32. Brown BE, Dean RT, Davies MJ.. Glycation of low-density lipoproteins by methylglyoxal and glycolaldehyde gives rise to the in vitro formation of lipid-laden cells. Diabetologia 2005;48:361–9. [DOI] [PubMed] [Google Scholar]
- 33. Brown BE, Mahroof FM, Cook NLet al. Hydrazine compounds inhibit glycation of low-density lipoproteins and prevent the in vitro formation of model foam cells from glycolaldehyde-modified low-density lipoproteins. Diabetologia 2006;49:775–83. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated and/or analyzed during the current study are not publicly available but are available from the corresponding author upon reasonable request.