Hypertension, fluid overload and micro inflammation are associated with left ventricular hypertrophy in maintenance hemodialysis patients

Abstract

This cross-sectional study aims to identify the potential risk factors of left ventricular hypertrophy (LVH) in hemodialysis (HD) patients. Echocardiography, anthropometric measurements and biochemical analyses were performed for 112 HD patients. In univariate analysis, body mass index, systolic blood pressure, diastolic blood pressure, glycosylated hemoglobin, glycated albumin, high sensitivity C-reactive protein (hs-CRP), cardiac troponin T (cTnT), amino-terminal pro-B-natriuretic peptide (NT-proBNP) and carotid artery intima-media thickness were positively correlated with left ventricular mass index (LVMI); pre-albumin, serum creatinine, left ventricular ejection fraction (LVEF) and fractional shortening were negatively correlated with LVMI. Linear regression analysis showed systolic blood pressure, NT-proBNP and LVEF were independently associated with LVMI. According to a binary logistic regression model, higher systolic blood pressure, NT-proBNP and hs-CRP levels showed independent correlation with LVH. Receiver operator characteristic curves analysis showed the associations between NT-proBNP and LVH more closely than hs-CRP and cTnT. The area under the curve for NT-proBNP, hs-CRP and cTnT was 0.762 (95% CI: 0.660–0.864, p < 0.001), 0.734 (95% CI: 0.624–0.844, p < 0.001) and 0.677 (95% CI: 0.563–0.790, p = 0.004), respectively. These data support the main conclusions: hypertension, fluid overload and micro inflammation are associated with LVH in maintenance HD patients. It demonstrates traditional and nontraditional risk factors all play important roles in the development of LVH.

• Fluid overload
• hemodialysis
• hypertension
• left ventricular hypertrophy
• micro inflammation

Introduction

Left ventricular hypertrophy (LVH) is frequent in maintenance hemodialysis (HD) patients and associated with a poor outcome. The worsening of preexisting LVH is the strongest predictor of sudden cardiac death in dialysis patients.1–3 Thus, identification of factors that are linked to LVH may lead to new treatment strategies to address cardiovascular risk in HD patients.

Several factors contribute to the development of LVH in HD patients including age, gender, diabetes, fluid overload, anemia, hypertension, disturbed mineral metabolism, arteriovenous fistula, hyperhomocysteinemia, malnutrition, high sympathetic activity, accumulation of the endogenous inhibitor of nitric oxide synthase asymmetrical dimethylarginine, etc.1–4 The objective of this study was to identify the traditional and nontraditional risk factors of LVH for better therapeutic strategies in the future.

Materials and methods

Nonhospitalized patients treated in the blood purification center of Tongji Hospital, Tongji University, were enrolled in January 2013. Plasma and serum were obtained from 112 maintenance HD patients. HD patients received regular HD treatment three times a week (4–5 h per session) using bicarbonate dialysate and synthetic low-flux polysulfone membranes. Inclusion criteria were age more than 18 years and vintage on HD more than 3 months. The exclusion criteria were clinical signs of acute infection, active inflammatory diseases, symptomatic cardiac failure or ischemic heart disease, tumors, chronic pulmonary disease, those treated with immunosuppressive drugs, residual urine volume greater than 200 mL per day and unwillingness to participate in the study. The study was approved by the Ethical Committee on Human Research of Tongji Hospital, Tongji University, Shanghai, China, and performed in accordance with the Declaration of Helsinki. All subjects gave informed consent to the study. Body mass index (BMI) was calculated from the ratio of the prescribed dry weight to the square of the height. Single-pool Kt.V (spKt/V), calculated according to Daugirdas, was used to represent the weekly dialysis dose.5 Blood pressure was estimated by averaging all predialysis blood pressure during the month before this study.

Laboratory methods

Cardiac troponin T and amino-terminal pro-B-natriuretic peptide assay

Peripheral blood was collected from HD patients before undergoing HD in mid-week day. Blood samples were put into ethylenediaminetetraacetic acid pyrogen-free tubes and clot activator tubes for the plasma, and the serum were removed, aliquoted and stored at −70 °C until analyzed. The plasma was used for the evaluation of cardiac troponin T (cTnT) and amino-terminal pro-B-natriuretic peptide (NT-proBNP), the serum for other biochemical parameters. cTnT was measured by a third-generation electrochemoluminescence immunoassay (Elecsys autoanalyzer 2010, Roche Diagnostics, Mannheim, Germany). NT-proBNP was measured using polyclonal antibody recognizing the N-terminal fragment of brain natriuretic peptide (Elecsys autoanalyzer 2010, Roche Diagnostics).

High sensitivity C-reactive protein and other laboratory parameters assay

The concentrations of high sensitivity C-reactive protein (hs-CRP) and β₂-microglobulin were determined using immunoturbidimetry assay, and concentrations of intact parathyroid hormone (iPTH) were measured using electrochemiluminescence immunoassay. Serum albumin (bromocresol purple method), creatinine, urea, uric acid, calcium, phosphorus, 25(OH)D3, blood glucose, glycosylated hemoglobin, glycated albumin, lipid profile, hemoglobin, white blood cell count, transferrin and ferritin were determined by routine procedures at the Department of Clinical Chemistry, Tongji Hospital, Tongji University. Serum calcium–phosphorus product was calculated by multiplication of the two serum concentrations.

Carotid artery intima-media thickness measurement

Ultrasonography of both carotid arteries was performed with the Philips SD800 ultrasound scanner (Best, the Netherlands) using a 7.5 MHz high-resolution linear transducer by two experienced sonographers who were both blinded to all the clinical and biochemical data of the study patients. All patients were examined in a supine position. Both longitudinal and cross-sectional images were viewed. Carotid artery intima-media thickness (cIMT) was defined as a low-level echo gray band that did not project into the arterial lumen and was the distance from the leading edge of the lumen–intima interface to the leading edge of the media–adventitia interface of the far edge. Three segments were measured bilaterally, namely the 1 cm section of the common carotid artery immediately proximal to the beginning of the dilation of the bifurcation, the 1 cm section of the bifurcation immediately proximal to the tip of the flow divider and the 1 cm section of the internal carotid artery immediately distal to the tip of the flow divider.

Echocardiography

Transthoracic echocardiographic examinations were conducted using a Philips echocardiographic machine (Philips IE33, Eindhoven, the Netherlands) with a 3.5-MHz multiphase array probe by a single experienced cardiologist during a midweek dialysis day, within 2 h after undergoing HD. Left ventricular mass (LVM) was calculated using measurements made according to the recommendations of the American Society of Echocardiography: LVM = 0.8{1.04 [(LVEDD + PWT + IVSDT)³ − (LVEDD)³]} + 0.6 g, where LVEDD is left ventricular diameter in end diastole, PWT is posterior wall thickness in diastole and IVSDT is interventricular septum thickness in end diastole.6 Left ventricular mass index (LVMI) was calculated as LVM divided by height (meters)^2.7. Correcting LVM for height^2.7 minimizes the effect of sex, race, age and obesity. LVH was defined by LVMI of over 47 g/m^2.7 in women or over 50 g/m^2.7 in men.7 The left ventricular ejection fraction (LVEF) were determined by two-dimensional echocardiography.

Statistical analysis

All continuous variables were tested for normal distribution with the Kolmogorov–Smirnov test before further statistical analysis. Normally distributed values were presented as mean ± standard deviation, whereas non-normally distributed values are presented as median (interquartile range). Categorical values are presented as number of patients and percent. Comparisons between two groups were assessed for continuous variables with the Student’s unpaired t-test, Mann–Whitney test or χ²-test, as appropriate. Spearman’s rank correlation and linear regression analysis (continuous variables with skewed distribution were natural log-transformed) were used to determine correlations of LVMI with other variables. Significance tests were two-sided, p < 0.05 were considered significant. Correlation of other parameters with LVH was evaluated by binary logistic regression model. Receiver operator characteristic (ROC) curves analysis was performed to evaluate the correlation of NT-proBNP, hs-CRP, cTnT and LVH. All statistical analyses were performed with SPSS (version 17, SPSS Inc., Chicago, IL).

Results

Patient characteristics

The median age of all 112 patients (57 males and 55 females) was 60.5 (53.8–70.0) years. The causes of end-stage renal disease were chronic glomerulonephritis (n = 57, 50.9%), diabetic nephropathy (n = 21, 18.8%), hypertensive nephrosclerosis (n = 9, 8.0%), polycystic kidney disease (n = 8, 7.1%), lupus nephritis (n = 4, 3.6%), aristolochic acid nephropathy (n = 3, 2.7%), pyelonephritis (n = 3, 2.7%), obstructive nephropathy (n = 2, 1.8%) and other or unknown (n = 5, 4.5%). Patients’ characteristics are given in Table 1.

Table 1. Univariate associations of LVMI with relevant clinical and laboratory parameters in HD patients.

Parameter	Mean ± SD or median (interquartile range)	Spearman correlation, ρ	p-Value
Age (years)	60.5 (53.8–70.0)	0.128	0.180
BMI (kg/m^2)	21.8 ± 3.0	0.267	0.004
Dialysis vintage (months)	53.0 (21.0–91.3)	−0.254	0.107
spKt/V	1.31 (1.17–1.58)	−0.144	0.042
Ultrafiltration volume (mL)	2500 (2000–3100)	0.117	0.233
erythropoietin dose (U/week)	10,000 (5000–10,000)	0.162	0.093
Systolic blood pressure (mmHg)	136.2 ± 22.2	0.468	<0.001
Diastolic blood pressure (mmHg)	79.3 ± 12.8	0.264	0.007
Serum albumin (g/L)	40.0 ± 3.0	−0.102	0.283
Pre-albumin (g/L)	0.34 ± 0.09	−0.243	0.009
Serum creatinine (μmol/L)	968.7 ± 250.0	−0.151	0.042
BUN (mmol/L)	24.1 (21.7–30.2)	0.065	0.495
Serum uric acid (μmol/L)	436.3 ± 93.2	−0.086	0.369
Calcium (mmol/L)	2.24 (2.12–2.40)	−0.133	0.161
Phosphorus (mmol/L)	1.99 (1.60–2.57)	0.051	0.593
Ca × P product (mmol^2/L^2)	4.42 (3.45–6.04)	0.027	0.778
iPTH (pg/ml)	338.6 (172.4–575.6)	0.142	0.135
25(OH)D3 (nmol/mL)	31.9 (19.3–40.4)	−0.137	0.154
Total cholesterol (mmol/L)	4.46 ± 1.03	−0.017	0.862
Triglyceride (mmol/L)	1.38 (1.04–1.96)	−0.046	0.628
HDL-C (mmol/L)	1.12 ± 0.35	−0.023	0.811
LDL-C (mmol/L)	2.54 ± 0.94	−0.018	0.848
non-HDL (mmol/L)	3.30 ± 1.01	0.027	0.789
β2-Microglobulin (mg/L)	30.2 (18.6–36.2)	0.031	0.746
Blood glucose (mmol/L)	6.1 (4.9–8.2)	0.142	0.135
Glycosylated hemoglobin (%)	5.2 (4.9–5.8)	0.214	0.025
Glycated albumin (%)	14.5 (13.2–16.2)	0.206	0.031
Hemoglobin (g/L)	110.1 ± 14.1	−0.008	0.929
White blood cell count (10^9/L)	5.91 (5.00–7.00)	0.034	0.722
Transferrin (g/L)	1.85 (1.69–2.09)	−0.112	0.248
Ferritin (ng/mL)	168.1 (86.8–321.0)	0.017	0.857
Homocysteine (μmol/L)	30.9 (26.0–38.3)	0.076	0.440
Hs-CRP (mg/L)	3.5 (1.3–10.4)	0.323	0.001
cTnT (ng/mL)	0.046 (0.031–0.065)	0.301	0.002
NT-proBNP (pg/mL)	4168.0 (1975.5–12,570.5)	0.476	<0.001
cIMT (cm)	0.8 (0.7–0.8)	0.267	0.015
LVEF (%)	64 (61–69)	−0.285	0.002
Fractional shortening (%)	36 (32–39)	−0.211	0.026

Note: BMI, body mass index; spKt/V, single pool urea clearance index; BUN, blood urea nitrogen; iPTH, intact-parathyroid hormone; HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol; cTnT, cardiac troponin T; NT-proBNP, amino-terminal pro-B-natriuretic peptide; hs-CRP, high-sensitive C-reactive protein; cIMT, carotid artery intima-media thickness and LVEF, left ventricular ejection fraction.

Prevalence of LVH and related risk factors

The median LVMI value of all patients was 47.5 (38.2–57.3) g/m^2.7. The prevalence of LVH in our study group was 46.4%. In univariate analysis, BMI, systolic blood pressure, diastolic blood pressure, glycosylated hemoglobin, glycated albumin, hs-CRP, cTnT, NT-proBNP and cIMT were positively correlated with LVMI; pre-albumin, serum creatinine, LVEF and fractional shortening were negatively correlated with LVMI; no correlation was found between LVMI with calcium, phosphorus, iPTH, 25(OH)D3, hemoglobin, homocysteine, blood glucose and lipid profile. Linear regression analysis showed systolic blood pressure, NT-proBNP and LVEF were independently associated with LVMI (Table 2).

Table 2. Linear regression analysis of associations between LVMI and laboratory parameters.

Independent variables	Unstandardized β-coefficient	Standardized β-coefficient	p-Value
Systolic blood pressure	0.007	0.432	<0.001
NT-proBNP	0.219	0.316	0.001
LVEF	−0.009	−0.175	0.038

Compared to non-LVH patients, LVH patients were more likely to be older and had also higher BMI, systolic blood pressure, cTnT, hs-CRP, NT-proBNP and cIMT and had lower serum creatinine. Table 3 lists baseline characteristics of LVH and non-LVH patients. According to a binary logistic regression model (Using variables which were different between LVH and non-LVH patients as independent variables. NT-proBNP, hs-CRP, and cTnT, were divided into two groups according to the median value.), higher systolic blood pressure, NT-proBNP and hs-CRP levels showed independent correlations with LVH (Table 4).

Table 3. Patients characteristics and laboratory parameters in LVH and non-LVH patients.

	Non-LVH (N = 60)	LVH (N = 52)	p-Value
Gender (male/female)	36/24	21/31	0.058
Age (years)	60.0 (53.0–66.0)	66.0 (57.3–73.8)	0.033
Smoker (yes/no)	20/40	9/43	0.083
Diabetes (yes/no)	10/50	11/41	0.629
Dialysis vintage (months)	61 (28–95)	37 (11–82)	0.094
BMI (kg/m^2)	21.5 ± 2.9	22.7 ± 3.0	0.039
SpKt/V	1.32 (1.21–1.56)	1.31 (1.16–1.55)	0.102
Ultrafiltration volume (mL)	2500 (2000–3000)	2350 (1800–3200)	0.535
Erythropoietin dose (U/week)	10,000 (5000–10,000)	10,000 (5500–10,000)	0.654
Systolic blood pressure (mmHg)	127.8 ± 18.7	145.8 ± 23.0	<0.001
Diastolic blood pressure (mmHg)	76.5 ± 10.5	81.2 ± 14.6	0.072
Serum albumin (g/L)	40.3 ± 2.8	39.7 ± 3.1	0.315
Pre-albumin (g/L)	0.35 ± 0.09	0.34 ± 0.09	0.406
Serum creatinine (μmol/L)	1011.6 ± 273.0	919.9 ± 189.5	0.035
BUN (mmol/L)	25.5 ± 7.0	25.1 ± 5.0	0.753
Serum uric acid (μmol/L)	447.1 ± 103.3	424.2 ± 79.6	0.198
Calcium (mmol/L)	2.23 ± 0.29	2.27 ± 0.25	0.391
Phosphorus (mmol/L)	1.97 ± 0.74	2.15 ± 0.68	0.173
Ca × P product (mmol^2/L^2)	4.68 ± 1.92	4.79 ± 1.63	0.757
iPTH (pg/mL)	256.4 (153.1–566.2)	367.6 (207.5–617.2)	0.124
25(OH)D3 (nmol/mL)	34.4 (20.6–46.3)	29.5 (18.2–36.8)	0.065
Triglyceride (mmol/L)	1.37 (0.99–1.99)	1.41 (1.07–2.01)	0.521
Total cholesterol (mmol/L)	4.35 ± 0.97	4.56 ± 1.06	0.291
LDL (mmol/L)	2.50 ± 0.95	2.55 ± 0.89	0.778
HDL (mmol/L)	1.13 ± 0.33	1.17 ± 0.36	0.557
non-HDL (mmol/L)	3.23 ± 0.98	3.39 ± 1.05	0.409
β2-microglobulin (mg/L)	28.4 (18.6–36.4)	30.0 (18.7–35.9)	0.935
Blood glucose (mmol/L)	5.9 (5.1–7.4)	6.6 (4.7–9.5)	0.156
Glycosylated hemoglobin (%)	5.2 (4.9–5.6)	5.4 (5.0–6.0)	0.182
Glycated albumin (%)	14.2 (13.0–16.0)	15.0 (13.6–16.7)	0.093
Hemoglobin (g/L)	110.0 ± 13.9	110.2 ± 14.2	0.926
White blood cell count (10^9/L)	5.89 ± 1.87	6.20 ± 2.03	0.419
Transferrin (g/L)	1.95 ± 0.37	1.88 ± 0.38	0.310
Ferritin (ng/mL)	172.6 (88.2–310.5)	138.2 (84.9–321.1)	0.584
Homocysteine	30.9 (26.6–36.9)	31.2 (26.4–41.0)	0.473
Hs-CRP (mg/L)	2.2 (0.8–7.8)	4.6 (1.8–16.6)	0.004
cTnT (ng/mL)	0.044 (0.024–0.064)	0.051 (0.031–0.071)	0.117
NT-proBNP (pg/mL)	2850.0 (1446.5–5212.0)	7579.0 (2933.0–13,970.5)	0.001
cIMT (cm)	0.7 (0.6–0.8)	0.8 (0.7–0.9)	0.013
LVEF (%)	67 (62–70)	65 (61–69)	0.053
Fractional shortening (%)	37 (32–40)	35 (32–39)	0.154
LVMI (g/m^2.7)	39.3 ± 7.1	59.2 (53.0–66.1)	<0.001

Table 4. Binary logistic regression analysis of independent risk factors of LVH.

	OR	95% CI	p-Value
Systolic blood pressure (1 mmHg increment)	1.037	1.006–1.069	0.019
NT-proBNP > 4168.0 (pg/mL)	4.698	1.596–13.829	0.005
Hs-CRP > 3.5 (mg/L)	2.496	1.186–11.027	0.043

The correlations between LVH and cardiovascular disease biomarkers

We further performed a comparison analysis using the ROC curves of NT-proBNP, hs-CRP and cTnT with regard to their correlations with LVH. ROC curves analysis showed the correlation between NT-proBNP and LVH was more closely than hs-CRP and cTnT. The area under the curve for NT-proBNP, hs-CRP and cTnT was 0.762 (95%CI: 0.660–0.864, p < 0.001), 0.734 (95%CI: 0.624–0.844, p < 0.001) and 0.677 (95%CI: 0.563–0.790, p = 0.004), respectively (Figure 1).

Figure 1. Receiver operator characteristic curves versus LVH generated with NT-proBNP, hs-CRP and cTnT.

Discussion

LVH is a common finding in maintenance HD patients and it is associated with an increased cardiovascular-related mortality. The data from our study confirmed the high prevalence of LVH in asymptomatic maintenance HD patients, which was well comparable to others,8 although lower than most studies previously reported.9,10 The other finding of our study was that higher systolic blood pressure, NT-proBNP and hs-CRP levels were the independent risk factors of LVH in maintenance HD patients. This finding identified a potential role for these markers to be incorporated into future diagnostic and therapeutic strategies aimed at the earlier detection and management of clinically silent LVH.

Circulating NT-proBNP levels reflect the degree of LV overload.11 Numerous studies demonstrated a close association between NT-proBNP level and left ventricular mass in the end-stage renal disease population; however, only very few studies examined the diagnostic potential of NT-proBNP for LVH.11,12 Sommerer et al.13 showed that NT-proBNP had a high predictive value for hypervolemia in HD patients as defined by a composite score based on clinical assessment of edema, weight change, respiratory collapse of inferior vena cava, and echocardiographic assessment of pulmonary arterial pressure or septal and posterior wall thickness. These findings might strengthen the possible use of NT-proBNP as a marker of variation in LVMI as a result of chronic volume expansion. Circulating cTnT is also linked to LVH in both HD and peritoneal dialysis patients.11 In the study by Mallamaci et al.14, cTnT seemed more strongly associated with LVM than cardiac ischemia or diabetes. In uremic cardiac hypertrophy, myocardial capillary growth did not keep pace with cardiomyocyte hypertrophy. This resulted in cardiomyocyte capillary mismatch, increased oxygen diffusion distance and reduced ischemic tolerance of the heart, which further increased subclinical ischemia of the myocardium and amplified the leakage of cardiac troponins across the plasma membrane of myocardial cells into the circulation. Furthermore, increased mechanical stress altered the permeability of cardiomyocyte plasma membranes, predisposing to leakage of troponins. Thus, the link between elevated cTnT and LVH may partly reflect leakage of this protein from hypertrophic cardiomyocytes.11

Inflammation, as characterized by an increased CRP level, is clearly associated with an increased cardiovascular disease mortality especially in dialysis patients.15 Thus, it would appear rational to use the CRP level as a predictor for cardiovascular disease and a target for therapy. In dialysis patients, there is a combination of an impaired immune response related to the uremic state and persistent immune/inflammatory responses (blood-membrane contact, water quality, bioincompatible membranes, vascular access, etc.), resulting in persistent immune system stimulation, low-grade systemic inflammation and altered cytokine balance. This may characterize the uremic state, which may translate an increased risk of developing vascular disease.15 Inflammation may promote the development of LVH that would change morphology and function of vascular smooth muscle cells, which results in increasing arterial stiffness so the development of LVH would be promoted. In addition, subclinical inflammation can lead to adverse left ventricular geometry by altering the equilibrium that regulates cell growth, apoptosis, phenotype and matrix turnover of cardiac tissue.16 The result of our study that higher hs-CRP level was a independent risk factor of LVH was in line with the above-mentioned point of view and some previous studies.17–19

The other major factor that was associated with LVH in our study was systolic blood pressure, which had been associated with increased afterload. This observation was consistent with many reports of an association between blood pressure and LVH in HD patients.4,8,9 Although measurement of blood pressure before starting and immediately at the end of dialysis is highly variable, dependent on fluid volume status and the hemodynamic effects of dialysis, it remains the clinical standard.

cIMT is a well-established surrogate marker of subclinical atherosclerosis.20 In this study, we found a positive association between cIMT and LVMI, however, which was not confirmed in multivariate analysis. This finding suggested that in addition to atherosclerosis, other factors played an important role in promoting LVH in these patients.

In our study, a positive correlation was observed between LVMI and BMI, the patients with LVH had higher BMI than those without LVH, however, which was also not confirmed in multivariate analysis. This might be attributed to the confounding factors, including preexisting chronic disease, age, malnutrition, inflammation and changes in vitamin D metabolism, etc.9 Furthermore, this result was discrepancy with some other studies,4 probably the impact of BMI on LVH was likely to vary according to race. In this study, spKt/V was negatively correlated with LVMI; this result might be explained by good control for volume overload, anemia and hypertension with a higher spKt/V.9 Our study also observed negative correlations among LVMI, pre-albumin and serum creatinine. This result suggested that malnutrition might play a role in LVH. One explanation of this result maybe malnutrition caused by micro inflammation lower serum creatinine. Inflammation stimulates muscular proteolysis and catabolism and inhibits repairing mechanisms, leading to sarcopenia, muscle mass loss and consequently low serum creatinine levels.21,22 Our data showed that LVMI correlated positively with glycosylated hemoglobin and glycated albumin, this maybe due to LVMI of diabetes patients in our study was significantly higher than non-diabetic patients (data was not shown).

Our study did not demonstrate a predictive role of hemoglobin concentration for LVMI. This contrasted with other published data that demonstrated a strong correlation between LVM and anemia, and the most likely explanation was the higher values and narrower range of hemoglobin in our cohort compared with other studies.3,8 Previous studies demonstrated a potential role of PTH in promoting LVH, but our data showed no correlation between LVMI and PTH. The reason of this result is not clear, but maybe related, in part, to some methodological aspects.8,23

Conclusions

The main findings of our study were that hypertension, fluid overload and micro inflammation were associated with LVH in maintenance HD patients. It demonstrated traditional and nontraditional risk factors all played important roles in the development of LVH. LVH is potentially preventable and reversible.24,25 Thus, our results suggested that strict volume control, long-term active antihypertensive treatment and anti-inflammatory treatment probably could prevent the progression of LVH among these patients. However, this hypothesis cannot be proved from our cross-sectional analysis and merits further longitudinal cohort study. Extensive treatment of risk factors for LVH and monitoring of changes in LVM by echocardiography are required in the follow-up of dialysis patients.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Acknowledgments

The authors are grateful to all the staff in Blood Purification Center, Tongji Hospital, Tongji University.