Introduction

Systolic hypertension is highly prevalent in the haemodialysis (HD) population. As diastolic blood pressure (BP) is usually normal, this results in an elevated pulse pressure, which is an independent cardiovascular risk factor.1, 2, 3 While total peripheral resistance reflects the nonpulsatile cardiac load, an elevated pulse pressure reflects a decrease in arterial cushioning function and, consequently, an increase in pulsatile load. This results from arterial stiffening, which normally occurs with increasing age.4 Patients with renal failure therefore seem to suffer from premature arterial aging. Structural vascular changes include increased (intimal) atherosclerosis, which is enhanced by the disadvantageous cardiovascular risk profile in end-stage renal disease (ESRD), and increased (medial) arteriosclerosis, which is related to calcification and chronic volume overload.5, 6 Functional changes include an increased arterial tone induced by angiotensin II,7 increased plasma endothelin 1 levels,8 and impaired nitric oxide synthesis resulting from accumulation of endogenous asymmetrical dimethylarginine (ADMA) in ESRD.9 Both structural and functional abnormalities of the arterial system affect the arterial pressure wave contour. The forward travelling pressure wave results in an initial systolic rise. Decreased compliance of central arteries, such as the aorta and common carotid artery, leads to a decreased Windkessel effect with a higher incident systolic peak pressure. It also leads to an increase in aortic pulse wave velocity (PWV) of both propagated and backward travelling (reflected) pressure waves. Earlier wave reflection changes the phase relationship between the propagated and reflected wave. The resulting central arterial wave contour increasingly resembles a peripheral arterial wave contour, with a second systolic BP (SBP) rise augmenting peak systolic pressure instead of diastolic pressure.

Several markers associated with increased central arterial stiffness have been shown to be predictors of mortality in the dialysis population, including carotid pulse pressure,10 carotid incremental elastic modulus,11 carotid–femoral PWV12, 13 and carotid augmentation index.14 In HD patients, factors such as anuria, anaemia and the presence of an arteriovenous fistula lead to chronic volume overload. In addition, reduced aortic compliance results in chronic pressure overload. This promotes the development of LV hypertrophy, and may lead to both diastolic dysfunction and a reduced coronary perfusion. Coronary perfusion may be further worsened by a reduced coronary perfusion pressure due to the lower diastolic BP (DBP).

The aim of our study was to test whether a reduction in volume overload by ultrafiltration (UF) during HD leads to an improvement of aortic compliance. We estimated aortic compliance noninvasively before and after HD with the use of a pulse pressure method based on the Windkessel model.15 This technique has not been applied before in a dialysis population. With the use of echocardiography, Doppler-derived aortic outflow volume measurement is combined with an estimation of aortic pulse pressure obtained by carotid pulse contour analysis using applanation tonometry. Changes in pressure–volume relations are measured close to the LV outflow tract, reflecting changes in the cushioning function of the aorta, and hence, of the pulsatile cardiac load, an important determinant of LV afterload.

Materials and methods

Study population

A total of 18 HD patients (nine men and nine women, median age 54 years, range 30–85 years) were included in this study. Median time on HD was 31 months (range 4–106 months). Hypertensive medication, including calcium-antagonists in four patients, angiotensin-converting enzyme inhibitor,1 and angiotensin II receptor blockers,4 had been withheld since the previous dialysis session, 2 days before the measurements. Four patients used beta blockers, which were continued. On the day of the measurements, patients refrained from smoking and did not use caffeine-containing beverages. The study protocol was approved by the institutional review committee of our hospital and written informed consent was obtained from all patients.

Haemodialysis

All patients were dialysed three times a week according to a standard dialysis prescription, which had been unchanged for several weeks. Dry weight was considered optimal when patients remained without symptoms of dyspnoea, orthopnoea or oedema during the interdialytic period. Inferior vena cava diameter (IVCD) was measured with patients in supine position during quiet expiration. Indexed by body surface area, overhydration was defined as an IVCD of >11.4 mm/m2 and underhydration as an IVCD of <8.0 mm/m2.16, 17 Hypotension was defined as a drop in SBP of >30% or below 100 mmHg. All dialysis treatments used Fresenius 4008 machines (Fresenius Medical Care, Bad Homburg, Germany), biocompatible membranes (Hemophane or Polysulphone) and bicarbonate-buffered dialysate (Fresenius Medical Care SK-F213).

BP measurement

Sphygmomanometric brachial BP was measured with the use of the Datascope Accutorr Plus™ (Datascope Corp., Paramus, NJ, USA). After a rest period of 5 min in recumbent position, three consecutive readings were made, and the average value for SBP, DBP and mean BP were taken. A regular cuff size was adequate in all patients, and was positioned just above the elbow in the nonfistula arm. Patients were in a recumbent position.

Vascular parameters

Measurements were made 1 h before and 1 h after HD. In a subset of eight patients, measurements were repeated 24 h after the initial measurements before HD. Pressure measurements and echocardiographic measurements were performed by two separate investigators. Pressure curves were obtained with the use of a pen-like transducer for applanation tonometry (Millar SPT-301, Millar Instruments Inc., Houston, TX, USA) of the carotid and femoral artery. Calibration used mean and diastolic sphygmomanometric brachial BP. Carotid BP was assumed to be similar to aortic BP, as invasive recordings have shown a high correlation between carotid and aortic recordings.18 Aortic compliance was calculated with a pulse pressure method combining carotid tonometry signals with echo Doppler-derived aortic outflow measurements. This method is based on an iterative search of the best fit between the measured pulse pressure and the pulse pressure predicted by a 2-element Windkessel. Inputs of the Windkessel were measured pulse pressure and aortic flow (Figure 1). Aortic outflow volume was measured by echocardiography in the LV outflow tract. This is the product of the aortic annulus area (cm2) from a parasternal long axis view and the Doppler outflow velocity−ejection time integral (VTI: (cm/s) s) from an apical four-chamber view.

Figure 1
figure 1

Combined measurements of carotid pressure (P) wave by applanation tonometry and aortic outflow velocity (Ao) by Doppler echocardiography for the calculation of aortic compliance with the use of a pulse pressure method. Aortic outflow volume (ml) is the product of the left ventricular outflow tract area (cm2) and the velocity-time integral (VTI) of the aortic outflow (cm). Aug: augmentation.

Full size image

Echocardiographic measurements

An experienced echocardiographer performed all echocardiographic measurements using a Hewlett Packard Sonos 5500 machine (Hewlett Packard, Andover, MA, USA). At baseline, LV dimensions were measured from the apical four-chamber or parasternal long-axis view. LV mass is reported as LV mass index (LVMI) by height according to the Framingham Study.19 As described above, the aortic outflow volume was measured in the LV outflow tract, which in the absence of significant valvular regurgitation equals stroke volume (SV). Fractional shortening (FS) and ejection fraction (EF) were measured as additional parameters of LV systolic function. Transmitral pulsed-wave Doppler flow velocities were recorded in the apical four-chamber or apical long-axis view. The sample volume was located at the tips of the mitral valve leaflets. Peak E and A transmitral flow velocities were measured to obtain an E/A ratio as measure of LV diastolic function.

Analysis of wave contours

The echo Doppler and tonometric signals were simultaneously acquired with an analogue/digital board and a laptop running customized software for spectral recomputation and automatic detection of maximal velocity. For analysis of wave contours, at least seven consecutive cardiac cycles were measured. Aortic PWV was computed from realigned carotid and femoral tonometry signals, simultaneous ECG recording and sternal notch-femoral distance. Augmentation of carotid systolic pressure was expressed as carotid augmentation index (AIC), which is the height of the second BP rise above the inclination point in the carotid pressure wave contour divided by carotid pulse pressure.

Statistical analysis

The change in vascular and haemodynamic parameters before and after HD was analysed by the Wilcoxon matched pairs test. For assessing correlation the Spearman correlation coefficient was used. A P-value of less than 0.05 was considered to indicate statistical significance. Data are presented as median values and range.

Results

Echocardiographic parameters at baseline were as follows: LVMI in men 141 g/m (117–196 g/m) and in women 141 g/m (86–195 g/m). EF was 71% (56–81%). FS was 33% (24–43%). E/A was 0.8 (0.6–1.3). IVC diameter at baseline was 10.5 mm/m2 (6.4–11.6 mm/m2), so there was no excessive overhydration before the start of HD. Median IVC diameter after HD was 9.8 mm/m2 (4.2–11.0 mm/m2), and all patients reached dryweight. Three patients developed hypotension during the dialysis treatment. In these patients, UF was temporarily halted, patients were placed in supine position, but no fluid had to be infused in order to continue HD. Median UF volume was 2450 ml (1000–4000 ml). SBP after HD (128 mmHg; 98–184 mmHg) was lower (P=0.01) than before HD (142 mmHg; 105–184 mmHg) (Table 1). DBP after HD (80 mmHg; 60–93 mmHg) was not different from DBP before HD (85 mmHg; 69–94 mmHg). Heart rate (HR) after HD (79 min−1; 57–100 min−1) was not different from HR before HD (75 min−1; 54–87 min−1).

Table 1 Haemodynamic and vascular parameters before and after HD
Full size table

Aortic outflow volume after HD (39 ml; 32–53 ml) was lower (P=0.01) than before (46 ml; 29–60 ml). At baseline, all patients showed carotid wave contours with augmentation. Carotid pulse pressure after HD (42 mmHg; 25–85 mmHg) was lower (P=0.01) than before (46 mmHg; 35–93 mmHg). AIC after HD (22%; 3–30%) was significantly lower (P=0.001) than before (31%; 7–53%). Carotid–femoral PWV was not significantly different after HD (8.7 m/s; 5.6–28.9 m/s vs 7.7 m/s; 4.7–36.8 m/s). Aortic compliance was negatively correlated with PWV before HD (r=−0.6; P=0.02) and after HD (r=−0.6; P=0.02). Aortic compliance after HD (1.10 ml/mmHg; 0.60–2.43 ml/mmHg) was significantly higher (P=0.02) than before HD (1.05 ml/mmHg; 0.45–1.69 ml/mmHg).

In the subgroup in which measurements were repeated the day after HD, values 1 h after HD were: AIC 19% (11–30%), PWV 7.3 m/s (4.7–23.3 m/s), and aortic compliance 1.66 ml/mmHg (0.86–2.43 ml/mmHg). At 24 h after the initial measurements before HD, values were: AIC (20%; 12–29%), PWV (7.6 m/s; 4.9–24.3 m/s), and aortic compliance (1.49 ml/mmHg; 0.99–1.95 ml/mmHg). None of the tonometry parameters showed a significant difference between measurements taken after 1 h and 1 day after HD.

Discussion

The results of this study show improvement in arterial wall pressure–volume relations after reduction of volume overload by UF. There was a significant increase in aortic compliance after HD and a significant decline in SBP, but no change in DBP. These results demonstrate that in HD patients the increase in aortic stiffness is partly reversible.

Studies testing the effect of rapid fluid withdrawal on arterial wall characteristics are few and show variable results (Table 2). This partly depends on the parameter used. As seen in our study, AIC has been shown to decrease after HD with UF.7, 20 When carotid–femoral PWV was used as a marker of aortic stiffness, no direct effect of fluid withdrawal was found in previous studies.7, 21 This is in agreement with our study in which there was no significant change in carotid–femoral PWV after HD with UF. The variability of our results for PWV may be partly responsible for this, although there was a trend toward a decrease after HD. In a previous study, UF combined with angiotensin-converting enzyme inhibition did show a significant decrease in carotid–femoral PWV, which was explained by an acute effect on angiotensin II activity.7

Table 2 Changes in pulse pressure and in different vascular parameters after volume withdrawal by HD
Full size table

Both increased PWV and augmentation have been shown to be important predictors of cardiovascular morbidity and mortality. Although they result from altered mechanical properties of arteries, neither provides direct information on those properties.22 AIC is a manifestation of arterial stiffness, and is influenced by confounding factors such as HR, age, gender and height,23 although in our study HR did not change significantly. The function of small muscular arteries and arterioles can also affect AIC by changing the aortic wave reflection pattern independent of aortic PWV.24 Aortic PWV is usually measured between the carotid and femoral artery as a parameter of segmental arterial stiffness. However, PWV is proportional to the square root of the product of wall elastance and thickness. As a result of this square root relationship, PWV is not a sensitive measure of a change in physical arterial properties.25, 26

In the dialysis patients in our study, changes in pressure–volume relations close to the LV outflow tract are estimated with the use of a pulse pressure method. Compliance is a direct measure of arterial elastic properties. It describes arterial stiffness as an absolute volume (or diameter) change for a given pressure increment. The improved aortic compliance following volume withdrawal in our study signifies a decrease in LV afterload resulting from a reduction in pulsatile load.

An inherent characteristic of a method for measuring compliance is that it depends on both changes in volume and changes in pressure. In this study population, as a consequence of UF, there is primarily a change in aortic outflow volume. Subsequently, there is a change in aortic pulse pressure. Since the arterial pressure–volume relationship is nonlinear, compliance increases as pressure decreases. This is because at higher pressure, wall tension is mainly generated by collagen fibres, but at lower pressure by elastin fibres, which are more distensible.3, 6 Therefore, overhydration per se can lead to increased aortic stiffness. In the subgroup that was measured the following day, the improvement of arterial wall characteristics was sustained. This finding suggests that relative overhydration leading to increased arterial stiffness occurs shortly before the start of HD.

The UF-induced increase in compliance may have resulted from changes independent of pressure, which we were unable to detect in the present study design. In two previous studies, the parameter carotid distensibility, another direct measure of arterial stiffness, was used to monitor changes in arterial wall characteristics after HD. It describes arterial stiffness as relative, rather than absolute, diameter change for a pressure increment. In one study, pulse pressure increased rather than decreased, although not significantly, and carotid distensibility decreased after HD.21 This was despite a mean UF volume of 1460 ml. The authors suggested this might be due to sympathetic activation, as there was no change in distensibility when corrected for pressure. In the other study, there was no significant change in carotid distensibility nor in BP and pulse pressure, despite a mean UF volume of 4000 ml.27 It would therefore appear that, when volume reduction by UF does not lead to a decrease in pressure, other factors leading to functional changes of the arterial wall offset the beneficial effect of volume withdrawal.21

Although volume withdrawal in our study had an acute, albeit small, impact on arterial wall characteristics, high PWV values and marked augmentation continued to be present after HD. This indicates that the acute large swing in volume status before and after HD affects arterial stiffness on top of a more permanent alteration in arterial structure and function, which is likely to be caused by chronic volume overload.28

We conclude, that the increase in aortic stiffness in HD patients is partly caused by a reversible reduction in aortic compliance resulting from volume expansion. Volume withdrawal by HD moves the arterial wall characteristics back to a more favourable position on the nonlinear pressure–volume curve, reflected in a concomitant decrease in arterial pressure and improved aortic compliance.