Congestive heart failure (CHF) has a high prevalence rate, being present in up to approximately one-third of patients with end stage renal disease (ESRD), and contributes to higher mortality [1-3]. At three-year survival rate of less than 15% has been reported among patients on hemodialysis (HD) with CHF, and the overall mortality rate was reported to be 36 months versus 62 months in patients with ESRD without CHF [4,5].

According to previous studies, the most common cardiac abnormality found in patients undergoing HD is left ventricular (LV) overload, represented by LV hypertrophy (LVH) [6-8]. Excessive arterial flow through venous drainage to the right atrium (RA), as in the case of high-flow arteriovenous fistula (HF-AVF), also contributes to LV overload and needs to be controlled [9-11]. Previous studies have investigated the cardiovascular (CV) risk factors of HD patients using transthoracic echocardiography (TTE) parameters, but AVF-specific heart loading has not been extensively researched [12-16].

LVH, the result of physiological adaptation to chronic workload of the heart, can be diagnosed based on a LV mass index (LVMI) of greater than 131 g/m² for males and 100 g/m² for females [17,18]. More recently, it has been reported that determination of left atrial volume (LAV) in addition to LVMI is important for cardiac risk monitoring in patients with ESRD [19-21].

In this study, we investigated the impact of kidney transplantation (KT) and AVF for HD on the LV workload by evaluating patients undergoing KT during HD and subsequent AVF ligation.

From January 2006 to December 2017, we enrolled patients who undergone KT at our institution, who had undergone HD through AVF prior to KT. The AVF remained functional even after KT, and the original function was maintained until AVF ligation was performed as needed. The time period when dialysis was maintained through the AVF without KT was termed “period A”, the time period after KT when the functional AVF was maintained was termed “period B”, the time period encompassing well-functioning transplanted kidney and after AVF ligation was termed “period C” (Fig. 1).

Figure 1. Study design and time course. (A) Definition of time periods and time course. (B) Flowchart of study design. AVF, arteriovenous fistula.

We analyzed the effects of two components of heart loading; heart loading related to ESRD on HD and heart loading related to flow of the AVF. As a retrospective study, pairing was performed according to each time period in which TTE findings were available for the patient population. To define each period according to heart loading for the purposes of this study, each period can be redefined as follows: period A, heart loading according to both ESRD on HD and flow of the AVF; period B, without heart loading according to ESRD on HD, but with heart loading according to flow of the AVF; period C, without heart loading due to ESRD on HD or flow of AVF. TTE used for analysis was performed at least one month after KT and/or fistula. In the case of multiple TTE results, the result was used for more than one month but closest to the time of one month.

The ratio of early mitral velocity, E, to tissue Doppler velocity, E’, (E/E’ ratio) was used for the estimation of LV filling pressures. The E/E’ ratio is used as a representative index of LV diastolic dysfunction. Normal diastolic function can be considered an E/E’ ratio under 8 (E/E’ <8), and diastolic dysfunction can be considered an E/E’ ratio over 14 (E/E’ >14) among patients without previous underlying heart disorders, including coronary artery disease [22-25].

We reviewed the results of TTE and calculated stroke volume (SV) with the LV outflow tract (LVOT) area and LVOT velocity time integral (LVOT VTI) as follows:

SV=LVOT area ∙ LVOT VTI

LVOT area=π ∙ (LVOT radius)² (Fig. 2)

Figure 2. Calculation of ventricular workload. LLV, workload of left ventricle; TPVR, total peripheral vascular resistance; CO, cardiac output; MAP, mean arterial pressure; SV, stroke volume; PR, pulse rate; LVOT, left ventricular out-track; VTI, velocity time integral.

LV workload (LLV) was obtained by multiplying cardiac output (CO) by mean arterial pressure (MAP). As follows:

LLV=CO ∙ MAP

Arterial blood pressure (BP) and pulse rate (PR) were measured during TTE. MAP and CO were calculated as follows:

MAP=(systolic BP+2 ∙ diastolic BP)/3

CO=SV ∙ PR

The LLV results were obtained for each of the three time periods defined above; LLV_A was the LLV during period A, LLV_B was the LLV during period B, and LLV_C was the LLV during period C. We compared the results of LLV_A versus LLV_B, LLV_B versus LLV_C, and LLV_A versus LLV_C in each patient.

Statistical analyses were performed using SPSS, version 24.0 software (IBM Corp., Armonk, NY, USA). Continuous variables are expressed as mean±standard deviation and categorical variables are expressed as number (percentage). The associations between TTE findings were evaluated using Pearson’s chi-square test and the Student t-test. The paired t-test and Wilcoxon test were used to analyze the change in each TTE variable in each time period. A value of p<0.005 was considered statistically significant.

During the study period, 324 patients underwent HD before KT. Of 324 patients, we were able to obtain data concerning TTE and AVF ligation in 109 patients. Four patients who were underwent TTE for heart disorders were excluded.

Demographics, clinical data and factors related ESRD of the 105 patients are presented in Table 1 and 2. For comparison of LLV_A and LLV_B, 32 patients were enrolled and the mean LLV values were 5.95 watts and 5.92 watts (p=0.942). For comparison of LLV_B and LLV_C, 25 patients were enrolled and the mean LLV values were 5.65 watts and 5.45 watts (p=0.494). For comparison of LLV_A and LLV_C, 58 patients were enrolled and the mean LLV values were 6.3 watts and 5.51 watts (p=0.011) (Fig. 3).

Table 1 . Demographic data and clinical features of patients.

Clinical features (n=105)
Median age (years, range)	49 (12−67)
Gender, male (%)	56 (53.3)
Body mass index (kg/m2)	22.3±3.18
Systolic blood pressure (mmHg)	140.3±23.0
Diastolic blood pressure (mmHg)	82.2±14.7
Coexisting medical conditions
Smoking (total/active/ex-smoker)	29 (27.6)/19 (18.1)/10 (9.5)
Hypertension	63 (60.0)
Diabetes mellitus	27 (25.7)
Chronic lung disease	4 (3.8)
Chronic liver disease	9 (8.6)
History of cancer	8 (7.4)
Donor type
Living donor	19 (18.1)
ABO-incompatible	3 (2.9)

Continuous data are expressed as mean±standard deviation and categorical data as number (%). The data were based on clinical information just before transplantation..

Table 2 . Factors and laboratory findings related with renal disease.

Cause of renal disease (n=105)
Diabetes mellitus	20 (19.0)
Hypertension	22 (21.0)
Glomerulonephritis	27 (25.7)
Other*	8 (7.6)
Unknown	28 (26.7)
Baseline laboratory findings
Hemoglobin (g/dL)	11.5±1.32
Blood urea nitrogen (mg/dL)	48.6±18.2
Serum creatinine (mg/dL)	8.1±2.4
Serum albumin (g/dL)	4.2±0.6
Total cholesterol (mg/dL)	153.5±35.4

Continuous data are expressed as mean±standard deviation and categorical data as numbers (%). The data were based on clinical information just before transplantation..

*Polycystic kidney disease, n=5; vesicoureteral reflux, n=1; chronic urinary tract infection, n=1; congenital nephrotic syndrome, n=1..

Figure 3. (A) LLV before kidney transplantation (LLV_A) and after kidney transplantation with AVF (LLV_B) (n=32). (B) LLV after kidney transplantation (LLV_B) and after AVF ligation (LLV_C) (n=25). (C) LLV before kidney transplantation (LLV_A) and after AVF ligation (LLV_C) (n=58). LLV, workload of left ventricle; AVF, arteriovenous fistula.

In order to evaluate the change in the degree of LVH, LVMI was compared on confirmatory TTE. The LVMI results showed statistically significant differences between all paired time periods (LVMI_A vs. LVMI_B, p=0.011; LVMI_B vs. LVMI_C, p=0.003; LVMI_A vs. LVMI_C, p<0.001). According to the previously mentioned criteria, 46 patients were diagnosed with LVH at the time of initial TTE (43.8%).

The results of the comparison of LAV values were statistically significant in the comparison of LAV_A and LAV_B and LAV_A and LAV_C, but there was no statistically significant difference in the comparison of LAV_B and LAV_C (LAV_A vs. LAV_B, p=0.048; LAV_B vs. LAV_C, p=0.591; LAV_A vs. LAV_C, p=0.023).

In the analysis of the E/E’ ratio, which reflects LV diastolic dysfunction, statistical significance was found between all time periods except E/E’_A and E/E’_B (E/E’_A vs. E/E’_B, p=0.164; E/E’_B vs. E/E’_C, p=0.002; E/E’_A vs. E/E’_C, p<0.001).

The first issue of this study was to identify the best parameter to determine heart workload. Since LLV is expressed in units of force (watts), including data concerning peripheral vascular resistance (PVR) and volume flow (MAP and CO), it is considered to more realistically reflect LV workload [26-28].

Cardiac hypertrophy, which is associated with changes such as LVH, is a major process in myocardial remodeling associated with heart workload [29]. Such remodeling is associated with unfavorable prognosis in the development of arrhythmia, heart failure, and sudden death. This remodeling process results in mismatch between physiologic supply and demand. This process begins with hypertrophy, leading to increasing output and compensatory hypertrophy to maintain the workload/mass ratio. If this vicious cycle persists, it will progress to heart failure [30,31].

In this study, each parameter (LVMI, LAV, and LLV) showed different results. Changes in LVMI, representing changes in LVH, were statistically significant in reducing heart workload at all time points when performed as in separate procedures and together. However, studies of LAV have reported statistically significant results in reducing heart workload at each stage, including KT. Studies evaluating LLV showed a reduction in heart workload when KT and AVF ligation were involved, but each procedure alone did not reduce heart workload.

Several previous studies reported that the recovery of renal function through KT can be expected to restore LV function [32-34]. When functional change after KT occurs, renal allograft functional problems have been reported to exclusively adversely affect LV remodeling [17]. There is a report that graft function and the possibility of cardiovascular events can be predicted through the evaluation of LV dysfunction after KT [35].

The general mechanism of CHF is equally applicable to patients with ESRD, and includes volume overload, anemia, uncontrolled hypertension, coronary artery disease, and valvular disease. There are other ESRD-specific factors, such as hyper-functioning vascular access, repetitive hemodynamic instability related to intermittent HD, pericardial effusion, and malnutrition [36]. In this study, parameters reflecting these factors are LVMI and LAV. The results of this study using LLV showed no statistically significant effect of KT alone.

LVMI is a parameter that reflects the degree of LVH. Two representative LVHs, eccentric and concentric hypertrophy, appear at a similar rate in patients undergoing HD. Eccentric hypertrophy is the result of volume overload resulting in cardiac myocyte drop-out, and concentric hypertrophy is caused by elevated BP and afterload due to anemia, hyperparathyroidism, and increased angiotensin II concentration [37,38]. LVH reduction requires BP control, regulation of extracellular fluid volume, and reduced volume fluctuation. In order to achieve these, the implementation of appropriate renal replacement therapies affect the status of HF is needed [19,20]. Especially for HF in patients with ESRD associated with eccentric hypertrophy, LVMI may be an easy parameter to follow-up using serial TTE. The results of this study can be interpreted to demonstrate that LVMI is a parameter that is more sensitive to change in heart workload, rather than being a specific indicator of outcomes related to a specific intervention.

Diastolic dysfunction in patients with ESRD on HD is considered an important risk factor for CV-related prognosis. Particularly with studies using the E/E’ ratio, patients with CV events had diastolic dysfunction and a higher baseline E/E’ ratio [13,14]. In current study, 14/105 patients (13.3%) showed findings suggestive of diastolic dysfunction, with a baseline E/E’ ratio greater than 15 (E/E’ ratio >15). During a mean follow-up duration of 155.0±77.5 months, 6 patients (5.7%) experienced CV events (cardiovascular events, n=4; cerebrovascular events, n=2). Although the timing (based on time periods defined above) of TTE was different, there was no statistically significant association between diastolic dysfunction based on the E/E’ ratio at initial TTE and CV events (p=0.138).

For patients with HF-AVF, heart remodeling results in sustained loading to the right heart, but appears to play a role in remodeling of the left heart, especially the LV [39]. However, because of increased right heart loading in patients with ESRD on HD, diastolic dysfunction and pulmonary hypertension are reported to be closely associated with disease progression and mortality. Even if LV function is compensated and preserved, it should not be overlooked [40,41]. Because flow of the AVF can increase not only right heart loading but also left heart loading, AVF flow reduction and/or AVF ligation can be considered for the reduction of heart workload after restoring renal function through KT [42,43].

In this study, most AVF ligations were performed due to cosmetic desires of the patients. In two cases, AVF ligations were performed due to discomfort from aneurysmal change of the native AVF. In one case, AVF ligation was performed because phlebitis could not be completely excluded. Therefore, the results of this study could be applied to patients in whom problems of flow overload, such as HF-AVF, are excluded. The mean duration of HD was 77.59±66.15 months (range, 1 month to 288 months) and the mean duration of AVF was 99.59±69.53 months (range, 3 months to 321 months). The association between occurrence of CV events and the durations of HD and AVF were not statistically significant in this study (HD duration, p=0.365; AVF duration, p=0.470).

This study has some important limitations. First, it was a single center study and has the disadvantage that the number of patients was relatively small. Particularly in the paired tests, there was a difference in the number of available TTE results at different time points, thus analyzation of smaller groups was required. Second, its retrospective design was subject to selection and information biases. This included the possibility of differences in outcomes due to differences in the duration of HD and the timing of TTE, KT, and AVF ligation for each patient. Third, since it was not a comparative study by period targeting the same patient group, bias due to inter-group heterogeneity must be considered. Lastly, enrolled patients comprised only those of Asian descent; because there may be racial differences in the prevalence of HF in patients with ESRD, our findings should be interpreted with caution with respect to different racial groups.

In conclusion, KT and AVF ligation might be able to reduce the workload of heart. KT or AVF ligation alone could be insufficient to regulate the workload of heart resulting from the disease course of ESRD and vascular access flow, which is essential for proper management. If necessary, both two interventions could be applied to reduce the chronic heart workload in patients with ESRD. There may be several different opinions and strategies concerning the maintenance or not of functioning AVF after KT. Based on the results of this study, it may be necessary to perform additional large, prospective studies to establish the appropriate indications for post KT AVF ligation and to apply the appropriate procedures to patients in need.

The authors declare no potential conflict of interest.