Background
The aim of this study was to evaluate whether diastolic dysfunction at the start of dialysis could influence renal and cardiovascular survival rates in 82 patients undergoing peritoneal dialysis.
Methods
Diastolic dysfunction was determined using left ventricular hypertrophy, the ratio of early peak transmitral inflow velocity to peak diastolic mitral annular velocity (E/E′), and left atrial volume index (LAVI). Residual renal function (RRF) was measured with 24-hour urine collections at baseline (within 1 month of beginning peritoneal dialysis) and thereafter at 6-month intervals for 2 years. To evaluate the long-term prognostic significance of diastolic dysfunction, the 4-year cardiac event–free survival was also evalated.
Results
The median slope of RRF decline was −0.07 mL/min/mo/1.73 m 2 . Forty-five patients (54.9%) with rapid RRF declines (< −0.07 mL/min/mo/1.73 m 2 ) had a higher prevalence of diabetes and eccentric left ventricular hypertrophy, as well as significantly elevated E/E′ ratios and LAVIs. There was a close relationship between baseline E/E′ ratio ( r = −0.221, P = .048), LAVI ( r = −0.276, P = .015), and RRF decline rate, and both E/E′ > 15 (odds ratio, 3.61; 95% confidence interval, 1.07–12.12) and LAVI > 32 mL/m 2 (odds ratio, 3.54; 95% confidence interval, 1.08−11.58) were significant independent predictors of the loss of RRF. Furthermore, E/E′ > 15 also provided additional prognostic value in predicting future cardiac events (hazard ratio, 6.74; 95% confidence interval, 1.07−12.12; P = .023).
Conclusions
Left ventricular diastolic dysfunction may be a significant predictor of rapid decline in RRF and adverse cardiac outcomes in patients starting peritoneal dialysis.
Residual renal function (RRF) is defined as the residual glomerular filtration rate (GFR) in patients with end-stage renal disease. In patients undergoing peritoneal dialysis (PD), RRF is of paramount importance, as it plays a major role in dialysis adequacy, middle molecule clearance, volume control, and nutritional status. Retaining even relatively small RRF may contribute to better cardiovascular survival and quality of life. Therefore, preserving RRF may be a principal goal in managing PD patients. To date, recurrent peritonitis, hypotensive episodes, nonuse of angiotensin-converting enzyme inhibitors, and the presence of diabetes or atherosclerotic disease are well-known risk factors for rapid loss of RRF. However, GFR is not easy to measure in the common clinical setting, especially in patients receiving renal replacement therapy. The most accurate and direct measurements of GFR require timed blood sampling after the administration of a tracer (e.g., iohexol, iothalamate, or inulin). However, this is often impractical for routine use and unrealistic for general practice. The “gold standard” for calculating GFR in advanced renal failure is to measure the mean of urea and creatinine clearance, calculated from 24-hour urine collections and normalized to body surface area (1.73 m 2 ). In the failing kidney, GFR should not be estimated from blood urea or creatinine alone, because creatinine clearance is disproportionately high as a result of tubular creatinine secretion, and the opposite is the case with urea clearance, for which tubular resorption is significant.
Recently, diastolic left ventricular (LV) dysfunction has become a widely recognized cardiac condition. It is acknowledged as a cause of cardinal symptoms and significant predictor of cardiovascular and all-cause mortality in a wide spectrum of populations, ranging from healthy subjects to hemodialysis patients. Its high prevalence (i.e., 25% in the general population and up to 70% in patients with end-stage renal disease ) raises important questions about its pathophysiology and clinical consequences. Increasing aging, diabetes, high blood pressure, cardiac ischemia, severe LV dysfunction, and chronic volume overload are common causes of diastolic dysfunction. Moreover, in a uremic milieu, endothelial dysfunction caused by chronic inflammation, oxidative stress, protein-energy wasting, and vascular calcification might accelerate the process of diastolic dysfunction.
In general, as the decline of RRF contributes to the aggravation of volume overload and LV hypertrophy (LVH), it is expected that patients without RRF will have a higher prevalence of diastolic dysfunction compared with those with preserved RRF. A previous cross-sectional study in maintenance PD patients demonstrated the independent association between LVH and RRF. However, there have been few data to investigate the impact of diastolic dysfunction at the start of dialysis on the RRF decline rate, particularly in PD patients.
The aims of this study were (1) to evaluate the effect of LV diastolic dysfunction, represented by an increased ratio of early peak transmitral inflow velocity to peak diastolic mitral annular velocity (E/E′) or left atrial (LA) volume index (LAVI) on the changes in RRF, and (2) to determine the role of diastolic dysfunction in predicting renal and cardiac survival in patients starting PD.
Methods
Study Population and Design
This longitudinal, observational cohort study was conducted with 116 patients who initiated PD at two medical centers in Korea between March 2005 and February 2009. Among these patients, 34 were excluded because of a lack of initial echocardiographic data ( n = 9), clinically significant volume overload states ( n = 10), transfer to other hospitals ( n = 6), noncompliance ( n = 3), and anuric status at the start of dialysis (daily urine output < 200 mL; n = 6). To calculate the rate of RRF decline, all patients were evaluated for RRF within 1 month of PD initiation (baseline) and every 6 months thereafter during 24 months or until loss of RRF. We used the average values of renal creatinine and urea clearance from 24-hour urine collections. Loss of RRF was defined as urine volume < 200 mL/day.
Cardiac events were defined as cardiac death, nonfatal acute coronary syndrome, or acute decompensated heart failure requiring hospitalization, and cardiac event–free survival was calculated from the initiation of PD to the date of cardiac event, renal transplantation, switch to hemodialysis, loss of follow-up, or April 30, 2011. The primary goal of this study was to compare the slope of RRF decline in patients who showed diastolic dysfunction at the start of PD versus those who did not. The secondary goal was to evaluate the effects of diastolic dysfunction on renal and long-term cardiac survival rate in PD patients. We did not obtain informed consent, because two-dimensional echocardiography in dialysis patients has been recommended in the Kidney Disease Outcomes Quality Initiative guidelines. This study was performed with the approval of the local institutional review board.
Echocardiographic Data
Comprehensive echocardiographic measurements were performed using an ultrasound machine (Vivid 7; GE Vingmed Ultrasound AS, Horten, Norway) with a 2.5-MHz probe by a single experienced cardiologist blinded to patients’ clinical information. All images were obtained with standard techniques using M-mode, two-dimensional, and Doppler measurements in accordance with the American Society of Echocardiography guide lines. LV ejection fraction (LVEF) was calculated using the modified biplane Simpson’s method in apical two-chamber and four-chamber views. Echocardiographic evidence of LVH was defined as LV mass index (LVMI) divided by body surface area > 115 g/m 2 in men and > 95 g/m 2 in women. LV geometry was stratified into four different patterns according to LVMI and regional wall thickness (0.42 cm). LV diastolic filling pattern was assessed by placing the pulsed Doppler sample volume at the tips of the mitral valve leaflets. From the mitral inflow velocity curve, peak E velocity, its deceleration time, peak A velocity, and E/A ratio were assessed. In addition, the E/E′ ratio was measured using Doppler tissue imaging. Doppler tissue imaging of mitral annular movement was obtained from the apical four-chamber view. A 2.0-mm sample volume was sequentially placed at the septal and lateral annular sites, and the average of the two values was used to evaluate the early (E′) and late (A′) diastolic peak velocities as well as their ratio (E′/A′). Filters and gains were adjusted to minimize background noise and maximize tissue signal ( Figure 1 ). Because E/E′ > 15 and ≤ 8 usually indicates elevated and normal LV filling pressure, respectively, we used these cutoff values. With these parameters, diastolic dysfunction was further categorized into four groups: normal, abnormal relaxation, pseudonormal, and restrictive pattern.

LA volume was also measured from the apical two-chamber and four-chamber views at end-systole by tracing LA borders (the walls of the left atrium and a line drawn across the mitral annulus) using planimetry. LAVI was calculated by dividing LA volume by body surface area and considered moderate or severely increased if it was >32 mL/m 2 , as has been previously reported.
Statistical Analysis
Statistical analyses were performed using SPSS version 18.0 (SPSS, Inc., Chicago, IL). All variables are expressed as mean ± SD or medians with ranges, unless otherwise indicated. The Kolmogorov-Smirnov test was used to analyze the normality of the distribution, and natural log values were used for skewed data, including RRF. The slope of the RRF decline over time was determined using linear regression analysis of serial urinary urea and creatinine measurements for each patient and expressed as the regression coefficient (mL/min/mo/1.73 m 2 ). Patients were divided into two subgroups according to the median value of RRF decline rate (fast and slow groups), and differences between these two groups were analyzed using independent t tests or Mann-Whitney U tests for continuous variables or χ 2 tests for categorical data. Longitudinal changes in RRF according to the E/E′ ratio (>15) and LAVI (>32 mL/m 2 ) were evaluated using a repeated-measures analysis of variance as a general linear model analysis. Clinical predictors for the loss of RRF were determined using multivariate logistic regression, and Cox regression analysis was used to evaluate cardiovascular mortality. P values < .05 were considered to be statistically significant.
Results
Baseline Clinical and Echocardiographic Characteristics
The baseline characteristics of the study population are shown in Table 1 . A total of 82 patients were included in the study (44 men; mean age, 52.3 ± 12.9 years; median age, 51.0 years; age range, 21–83 years). During the 2-year follow-up of RRF, the median number of RRF measurements was 4 (range, 3–5), except for three patients who showed anuria at 6 months after starting PD, whose RRF was measured only two times (at baseline and at 6 months). Among the 82 patients, 40 (48.8%) lost their RRF, and the mean duration from the start of PD to loss of RRF was 16.4 months (median, 16 months).
Clinical characteristic | Total ( n = 82) | RRF decline rate | P | |
---|---|---|---|---|
Fast ∗ ( n = 45) | Slow † ( n = 37) | |||
Median RRF decline rate (mL/min/mo/1.73 m 2 ) | −0.070 | −0.091 | −0.029 | <.001 |
Age (y) | 52.3 ± 12.9 | 54.1 ± 12.8 | 50.1 ± 12.9 | .173 |
Men | 44 (53.7%) | 29 (64.4%) | 15 (40.5%) | .031 |
BMI (kg/m 2 ) | 23.3 ± 3.5 | 23.9 ± 3.5 | 22.6 ± 3.4 | .065 |
Diabetes mellitus | 42 (51.2%) | 30 (66.7%) | 12 (32.4%) | .002 |
Systolic blood pressure (mm Hg) | 131.5 ± 20.0 | 132.8 ± 22.1 | 130.5 ± 18.8 | .885 |
Diastolic blood pressure (mm Hg) | 71.9 ± 16.7 | 74.6 ± 14.2 | 69.9 ± 20.0 | .754 |
Cause of ESRD | .038 | |||
Diabetic nephropathy | 41 (50%) | 29 (64.4%) | 12 (32.4%) | |
Hypertensive nephrosclerosis | 18 (22%) | 7 (15.6%) | 11 (29.7%) | |
Other | 23 (28%) | 9 (20.0%) | 14 (37.9%) | |
Previous cardiovascular disease | ||||
Coronary artery disease | 14 (17.1%) | 10 (22.2%) | 4 (10.8%) | .437 |
Peripheral artery disease | 3 (3.7%) | 2 (4.4%) | 1 (2.7%) | .676 |
Cerebrovascular accident | 8 (9.8%) | 4 (8.9%) | 4 (10.8%) | .770 |
Baseline laboratory findings | ||||
Hemoglobin (g/dL) | 8.8 ± 1.4 | 8.9 ± 1.3 | 8.7 ± 1.6 | .623 |
Serum albumin (g/dL) | 3.4 ± 0.6 | 3.3 ± 0.6 | 3.6 ± 0.5 | .106 |
Urea nitrogen (mg/dL) | 47.4 ± 16.1 | 48.3 ± 16.7 | 46.4 ± 15.5 | .613 |
Serum creatinine (mg/dL) | 7.2 ± 3.1 | 7.2 ± 3.3 | 7.2 ± 2.7 | .302 |
Serum uric acid | 8.1 ± 2.3 | 7.8 ± 2.3 | 8.3 ± 2.3 | .337 |
Total cholesterol (mg/dL) | 166.9 ± 51.3 | 169.5 ± 48.1 | 163.6 ± 55.5 | .608 |
C-reactive protein (mg/L) | 4.0 (1.0–29.6) | 4.9 (1.0–29.6) | 2.9 (1.0–23.6) | .274 |
Parathyroid hormone (pmol/L) | 181.9 (88.1–812.1) | 163.9 (88.1–701.6) | 264.3 (121.4–812.1) | .056 |
nPNA (g/kg/day) | 1.0 ± 0.3 | 0.9 ± 0.2 | 1.1 ± 0.3 | .117 |
PD-related parameters | ||||
Urine volume (mL/day) | 868.7 ± 533.3 | 868.5 ± 545.6 | 869.7 ± 533.3 | .993 |
D/P creatinine, 4 hours | 0.71 ± 0.16 | 0.71 ± 0.13 | 0.71 ± 0.18 | .879 |
Total weekly Kt/V urea | 2.4 ± 0.8 | 2.2 ± 0.6 | 2.6 ± 0.8 | .042 |
Peritoneal Kt/V urea | 1.5 ± 0.6 | 1.4 ± 0.5 | 1.8 ± 0.8 | .008 |
Renal Kt/V urea | 0.8 ± 0.5 | 0.9 ± 0.5 | 0.8 ± 0.5 | .684 |
Total weekly creatinine clearance (L/week/1.73 m 2 ) | 86.1 ± 29.7 | 87.3 ± 30.6 | 84.8 ± 29.1 | .712 |
RRF at the start of dialysis (mL/min/1.73 m 2 ) | 2.24 (0.87–10.21) | 2.42 (0.87–6.88) | 2.11 (1.12–10.21) | .353 |
Biocompatible solution | 65 (79.3%) | 38 (84.4%) | 27 (73.0%) | .202 |
∗ RRF decline rate < −0.070 mL/min/mo/1.73 m 2 .
The overall rate of RRF decline was −0.070 mL/min/mo/1.73 m 2 . On the basis of this rate, patients were divided into two groups: those with rapid declines ( n = 45 [54.9%]; median rate, −0.091 mL/min/mo/1.73 m 2 ) and those with slow declines ( n = 37 [45.1%]; median rate, −0.029 mL/min/mo/1.73 m 2 ). RRF decreased more rapidly in men (−0.081 vs −0.051 mL/min/mo/1.73 m 2 , P = .023) and patients with diabetes (−0.085 vs −0.049 mL/min/mo/1.73 m 2 , P = .007) compared with women and patients without diabetes.
The differences in echocardiographic parameters between the fast and slow groups are summarized in Table 2 . Intraobserver and interobserver variability was calculated in a subsample of 10 consecutive patients (11.7% of the total study population): 5.8% and 7.0% for LAVI, 5.2% and 6.3% for E velocity, 5.5% and 6.5% for A velocity, 5.3% and 6.4% for E′ velocity, and 5.8 and 7.1% for A′ velocity, respectively. LV diastolic dysfunction was observed in 66 patients (80.5%): 46 (56.1%) had grade 1 diastolic dysfunction (mild), 18 (22.0%) had grade 2 dysfunction (moderate), and two (2.4%) had grade 3 dysfunction (severe). Of the 82 patients, only 12 (14.6%) showed LV systolic dysfunction (LVEF < 50%). LA diameter ( P = .012) and LAVI ( P = .011) were significantly higher in the fast group, whereas LVMI ( P = .662) and LVEF ( P = .645) were similar between the two groups. LAVIs > 32 mL/m 2 were observed in 29 patients (64.4%) in the fast group and 14 patients (37.8%) in the slow group ( P = .016). Mitral inflow patterns revealed significantly increased E-wave and decreased A-wave velocities in the fast group; therefore, the E/A ratio was much higher in that group. Likewise, Doppler tissue imaging showed that the E/E′ ratio was significantly increased in the fast group (14.5 ± 7.5 vs 11.3 ± 4.4, P = .025). Moreover, echocardiographically detected mitral annular calcification was found in seven patients (8.5%). Although E′ velocity seemed to decrease in patients with mitral annular calcification, it was statistically insignificant compared with those without mitral calcification (5.8 ± 1.2 vs 6.4 ± 2.3 cm/sec, P = .570). Correlation analysis showed a strong positive association between E/E′ ratio and LAVI ( r = 0.688, P < .001; Figure 2 ). Subjects with faster declines of RRF had more elevated E/E′ > 15 and LAVI > 32 mL/m 2 at the start of dialysis ( Figure 2 ).
Echocardiographic parameter | Total ( n = 82) | RRF decline rate | P | |
---|---|---|---|---|
Fast ( n = 45) | Slow ( n = 37) | |||
M-mode and 2D | ||||
LV end-diastolic diameter (mm) | 50.8 ± 6.6 | 52.0 ± 6.6 | 47.5 ± 6.4 | .042 |
LV end-systolic diameter (mm) | 33.5 ± 8.1 | 35.2 ± 8.6 | 30.6 ± 7.1 | .050 |
LV end-diastolic volume index (mL/m 2 ) | 60.8 ± 18.5 | 64.9 ± 20.8 | 53.2 ± 13.8 | .033 |
LV end-systolic volume index (mL/m 2 ) | 27.6 ± 13.4 | 30.2 ± 10.2 | 21.7 ± 10.5 | .047 |
LVMI (g/m 2 ) | 137.9 ± 49.1 | 140.4 ± 45.0 | 128.0 ± 54.9 | .662 |
LVEF (%) | 57.1 ± 5.1 | 57.5 ± 4.3 | 56.5 ± 6.0 | .645 |
LA diameter (mm) | 40.4 ± 6.8 | 42.1 ± 7.3 | 35.1 ± 5.6 | .012 |
LAVI (mL/m 2 ) | 35.9 ± 15.8 | 39.4 ± 15.8 | 30.4 ± 11.6 | .011 |
LVH | .056 | |||
Normal | 30 (36.6%) | 13 (28.9%) | 17 (45.9%) | |
Concentric hypertrophy | 35 (42.7%) | 19 (42.2%) | 16 (43.2%) | |
Eccentric hypertrophy | 14 (17.1%) | 12 (26.7%) | 2 (5.4%) | |
Concentric remodeling | 3 (3.7%) | 1 (2.2%) | 2 (5.4%) | |
Mitral inflow velocities | ||||
E (cm/sec) | 80 ± 27 | 88 ± 22 | 70 ± 17 | .022 |
A (cm/sec) | 85 ± 20 | 80 ± 19 | 95 ± 21 | .016 |
E/A ratio | 0.95 ± 0.48 | 1.07 ± 0.55 | 0.80 ± 0.33 | .007 |
DT (msec) | 208.6 ± 53.0 | 200.4 ± 53.9 | 217.1 ± 52.6 | .128 |
Tissue Doppler parameters | ||||
S′ (cm/sec) | 6.9 ± 1.8 | 6.8 ± 1.9 | 7.1 ± 1.7 | .576 |
E′ (cm/sec) | 6.4 ± 2.2 | 6.1 ± 1.9 | 6.7 ± 1.4 | .140 |
A′ (cm/sec) | 8.5 ± 2.1 | 8.2 ± 2.2 | 8.7 ± 2.1 | .189 |
E′/A′ ratio | 0.7 ± 0.3 | 0.7 ± 0.3 | 0.7 ± 0.3 | .969 |
Pulmonary vein S/D | 1.2 ± 0.4 | 1.0 ± 0.4 | 1.3 ± 0.3 | .229 |
Pulmonary vein Ar velocity (cm/sec) | 28 (16–60) | 30 (16–60) | 25 (18–54) | .507 |
Mitral E/E′ ratio | 13.0 ± 6.4 | 14.5 ± 7.5 | 11.3 ± 4.4 | .025 |
E/E′ > 15 | 26 (31.7%) | 20 (44.4%) | 6 (16.2%) | .001 |
Grading of diastolic dysfunction | .038 | |||
Normal | 16 (19.5%) | 7 (15.6%) | 9 (24.3%) | |
Grade 1 (impaired relaxation pattern) | 46 (56.1%) | 23 (51.1%) | 23 (62.2%) | |
Grade 2 (pseudonormal pattern) | 18 (22.0%) | 13 (28.9%) | 5 (13.5%) | |
Grade 3 (restrictive filling pattern) | 2 (2.4%) | 2 (4.4%) | 0 (0.0%) |
Factors Associated with RRF Decline Rate and Cardiac and Renal Survival
The rate of RRF decline for all subjects is shown in Figure 3 . The decline rate had the largest correlation with diabetes ( r = −0.314, P = .004), followed by LAVI ( r = −0.276, P = .015), aging ( r = −0.256, P = .020), LVH ( r = −0.236, P = .032), and E/E′ ratio ( r = −0.221, P = .048) ( Table 3 ).
Variable | r | P |
---|---|---|
Age | −0.256 | .020 |
Diabetes | −0.314 | .004 |
LA diameter | −0.217 | .053 |
LAVI | −0.276 | .015 |
LVMI | −0.054 | .645 |
LVH | −0.236 | .032 |
LVEF | −0.137 | .219 |
E/A ratio | −0.149 | .182 |
E′ velocity | −0.014 | .217 |
A′ velocity | −0.110 | .406 |
E/E′ ratio | −0.221 | .048 |
Figure 4 shows the effect of increased E/E′ ratio and LAVI on the changes in RRF during the 2 years of follow-up. Patients with E/E′ ratios > 15 or LAVIs > 32 mL/m 2 showed significantly higher rates of RRF decline. A repeated-measures analysis of variance was conducted to compare the effects of those two parameters on RRF decline rates: there was a significant interaction between E/E′, LAVI, and RRF ( P < .001), and the between-subjects effect was F (1, 37) = 7.21 ( P = .011) with E/E′ > 15 and F (1, 37) = 3.26 ( P = .039) with LAVI > 32 mL/m 2 . Table 4 shows the number of patients who lost RRF or had adverse cardiac outcomes during the study period. Loss of RRF and adverse cardiac events occurred more frequently in patients with diabetes, increased E/E′ ratios > 15, LAVIs > 32 mL/m 2 , and LV systolic dysfunction (LVEF < 50%). In a logistic regression model evaluating independent predictors of the loss of RRF, the presence of diabetes, RRF at the start of dialysis, E/E′ > 15, LAVI > 32 mL/m 2 , eccentric hypertrophy (i.e., LV dilation), and serum albumin level emerged as significant predictors in a univariate analysis. After incorporating those factors, as shown in Table 5 , the presence of diabetes and diastolic dysfunction (either E/E′ > 15 or LAVI > 32 mL/m 2 ) were still significant factors (with E/E′: odds ratio, 3.61; 95% confidence interval [CI], 1.07–12.12; P = .038; with LAVI: odds ratio, 3.54; 95% CI, 1.08–11.58; P = .037).
Variable | Loss of RRF | Cardiac events | ||||
---|---|---|---|---|---|---|
Yes | No | P | Yes | No | P | |
Gender | 0.261 | 0.774 | ||||
Male | 24 (54.5%) | 20 (45.5%) | 8 (18.2%) | 36 (81.8%) | ||
Female | 16 (42.1%) | 22 (57.9%) | 6 (15.8%) | 32 (84.2%) | ||
Diabetes | .004 | .001 | ||||
Presence | 27 (64.3%) | 15 (35.7%) | 11 (26.2%) | 31 (73.8%) | ||
Absence | 13 (32.5%) | 27 (67.5%) | 3 (7.5%) | 37 (92.5%) | ||
E/E′ | <.001 | .006 | ||||
>15 | 22 (78.6%) | 6 (21.4%) | 10 (35.7%) | 18 (64.3%) | ||
≤15 | 18 (33.3%) | 36 (66.7%) | 4 (7.4%) | 50 (92.6%) | ||
LAVI (mL/m 2 ) | .008 | .088 | ||||
>32 | 28 (63.6%) | 16 (36.4%) | 10 (22.7%) | 34 (77.3%) | ||
≤32 | 12 (31.6%) | 26 (68.4%) | 4 (10.5%) | 34 (89.5%) | ||
LVEF (%) | .024 | .005 | ||||
≥50 | 27 (38.6%) | 43 (61.4%) | 8 (11.4%) | 62 (88.6%) | ||
<50 | 8 (66.6%) | 4 (33.4%) | 6 (50.0%) | 6 (50.0%) | ||
Loss of RRF | — | <.001 | ||||
Presence | — | — | 13 (32.5%) | 27 (67.5%) | ||
Absence | — | — | 1 (2.4%) | 41 (97.6%) |

Stay updated, free articles. Join our Telegram channel

Full access? Get Clinical Tree


