Modern peptide biomarkers and echocardiography in cardiac healthy haemodialysis patients

Haemodialysis was performed three times per week with a mean time of 4.85 h (five hours n = 17; four hours n = 3), proven arterial fistula blood flow >800 ml/min, machine blood flow 300 ml/min, dialysate flow 500 ml/min, dialysate concentrations SW 285A (Na⁺ 138, K⁺ 3) and SW 393A (Na⁺ 138, K⁺ 4) (Braun, Germany), high flux membrane/dialyzer Polyflux 170H (Nikkiso, Japan), and haemodialysis machine DBB05 (Nikkiso, Japan). All patients had native fistulas. The cut off of the ultrafiltration rate was set by 2000 ml by the median of all patients of the employed session for the dichotomous variable “UF 2000 ml” (median 2000 ml, min 0 ml, max 5000 ml, mean 2055 ml).

Haemoglobin (Hb), haematocrit (Hk), sodium, potassium, phosphate, creatinine, osmolality, albumine, vasopressin (AVP), aldosterone, renin, metanephrine, normetanephrine, MR-proANP, and copeptin [26] have been obtained by serum and plasma samples (EDTA, citrate, and heparin) before and after haemodialysis. Due to intradialytic elimination, NT-proBNP was solely measured before haemodialysis [27]. Additionally, KT/v was calculated using the Daugirdas formula [28, 29].

In all patients standardised transthoracic echocardiography was performed by experienced cardiologists according to national and international recommendations before (after suitable rest of 20 min for each patient) and approximately 20 min after haemodialysis [30]. Echocardiographic investigations were performed with Vivid 7 or Vivid E9 (GE Healthcare, Frankfurt, Germany) and all parameters were analysed offline using the EchoPac-Software (GE Healthcare, Frankfurt, Germany).

Left ventricular (LV) systolic function was assessed by determining LV ejection fraction due to biplane LV volume analysis using the modified Simpson’s rule [31]. Left ventricular diastolic function was assessed by the following echocardiographic parameters. The early-diastolic (Emax) and late-diastolic (Amax) maximum velocity of the mitral valve inflow were assessed by pulsed-wave Doppler measurements in the apical long axis view. The sample volume was positioned 1-2 cm above the mitral ring at the junction of the mitral leaflets to the chord strands and inappropriate Doppler angles were avoided. Both peak velocities permit the calculation of the E/A-ratio. In addition, the deceleration time (EDT) describing the duration from E peak to the end of the E-wave.

The left ventricular end-diastolic pressure was assessed by the determination of E/E’-ratio. E’ describes the tissue velocity of the motion of a specific myocardial segment during diastole and was assessed by tissue Doppler measurements. The sample volume of the pulsed-wave tissue Doppler was positioned at the basal inferoseptal and basal lateral segment in the apical 4-chamber view to obtain E’-inferoseptal and E’-lateral, respectively. Thus, the E/E’-inferoseptal-ratio and E/E’-lateral-ratio could be determined [32].

The diameter of the vena cava inferior (VCID) was assessed in a subcostal short axis view where the vena cava inferior enters the right atrium. Three different diameters were determined via M-Mode during expiration of the patient to obtain the median of the VCID.

The echocardiographic parameters were correlated with the ultrafiltration rate (UR) and the residual excretion (RE).

Pre- and post-HD sampling was performed in our study and biochemistry parameters were determined before and after HD as shown in Table 1.

In patients with an ultrafiltration rate ≥ 2000 ml aldosterone post-HD was significantly higher as in patients with an ultrafiltration rate < 2000 ml (Additional file 1: Table S1). In patients with a residual urinary excretion rate < 500 ml normetanephrine pre-HD was significantly lower as in patients with a rate > 500 ml. Copeptin pre- and post-HD was significantly higher in patients with an urinary excretion rate < 500 ml (Additional file 1: Table S2).

AVP was in normal range before HD and decreased after HD significantly by 27% (Fig. 1a). Only the percental reduction of copeptin (and not of AVP) was significantly correlated with the percental reduction of the body weight after HD (Spearman test in bivariate correlation analysis, p = 0.028, r = −0.503). In patients with an UF > 2000 ml, serum AVP levels were significantly higher (before HD: median 8.9 ng/l (4.4 - 9.4 ng/l); after HD median 6.9 ng/l (2.9 - 9.1 ng/l) than in patients with an UF < 2000 ml (before HD median 5.1 ng/l (1 - 7.7 ng/l; after HD median 2.7 ng/l (1 - 6.1 ng/l); p < 0.05). There was a trend in higher AVP levels in patients with a RE ≥ 500 ml (Additional file 1: Table S2). No influence of gender, age, or BMI was found. Copeptin levels were 11 fold higher in patients with HD as in the normal population (median 159 pmol/l, 14 - 291 pmol/l, normal median 4.2 pmol/l, 1 - 14 pmol/l [26] and there was a significant reduction of 54% by HD (Fig. 1b). Copeptin levels were significantly higher in patients with an UF ≥ 2000 before HD (median 207 pmol/l, 59 - 291 pmol/l, p < 0.05) and after HD (median 110 pmol/, 27 - 132 pmol/l, p < 0.01) than in patients with an UF < 2000 (median before HD 90 pmol/l, 14 - 196 pmol/l; median after HD 29 pmol/l, 8 - 75 pmol/l). After HD there was in both groups a significant reduction of copeptin and copeptin was significantly correlated with UF (Fig. 2A). Copeptin values were also correlated with the body weight. Thus, in case of increased body weight loss - in case of higher UF - during HD higher copeptin values were obtained (Fig. 2B). Significantly higher copeptin values were revealed in the group with RE < 500 ml (n = 9), either pre-HD (Median 193 pmol/l; 59 - 291 pmol/l) or post-HD (Median 106 pmol/l; 27 - 132 pmol/l) in comparison to the study group with RE ≥ 500 ml (n = 11, median before HD 95 pmol/l, 14 - 218 pmol/l; median after HD 39, 8 - 114 pmol/l). In addition, the statistical analysis showed a negative correlation between copeptin and RE (data not shown). Interestingly, older, male and obese patients tended to higher copeptin concentrations. The relative reduction of copeptin during HD significantly correlated inversely with the Kt/V ratio and the urea reduction indicating that copeptin might be eliminated by HD.

In parallel to AVP and copeptin, MR-proANP was significantly reduced about 26.8% after HD (median before HD 814 pmol/l (285 - 2908 pmol/l), median after HD 665 pmol/l (222 - 2144 pmol/l). Similarly to copeptin, all MR-proANP values were clearly above the normal range as described in the literature in non HD patients. Before HD, MR-proANP values were up to ten fold higher as in the normal population and after HD MR-proANP was still up to eight fold above the normal range in non HD patients (46.1 to 85.2 pmol/l). The percental reduction of MR-proANP was not significantly correlated with the percental reduction of the body weight after HD (Spearman test in bivariate correlation analysis, p = 0.9, r = −0.029). However, there was no significant difference relating to UF, RE, age, gender and BMI. In analogy to copeptin or MR-proANP, the NT-proBNP concentration (median 3020 ng/l, 522 - 72,312 ng/l) was about 18 fold elevated in comparison to the normal range (<169 ng/l).

There was an increase of renin after HD in 13 patients and a decrease of renin in seven patients after HD. In addition, there was a positive correlation of renin to UF after HD showing that a higher UR rate led to higher renin values after HD (r = 0.5 and p < 0.05). The percental reduction of renin was significantly correlated with the percental reduction of the body weight after HD (Spearman test in bivariate correlation analysis, p = 0.046, r = −0.452). Other significant correlations between renin and e.g. RE, age, BMI, blood pressure, antihypertensive drugs or remaining biochemistry parameters as proANP, proBNP, copeptin, AVP, aldosterone, meta- and normetanephrines could not be revealed (data not shown).

According to the different renin reaction patterns, patients with renin decrease (n = 7) and with renin increase (n = 13) after HD showed both significant differences for renin. A comparison of parameters depending from post-HD renin decrease or renin increase is shown in Additional file 1: Table S3.

An inverse correlation between aldosterone and proANP was detected (r = −0.6, p < 0.01). This was similar to proBNP showing a significantly negative correlation to aldosterone (r = −0.5, p < 0.05). The percental reduction of aldosterone was significantly correlated with the percental reduction of the body weight after HD (Spearman test in bivariate correlation analysis, p = 0.00035, r = −0.790). Other significant correlations (e.g. age, BMI, gender, blood pressure, antihypertensive drugs or remaining biochemistry parameters) could not be revealed.

Similarly to renin, different aldosterone reaction patterns could be detected: patients with aldosterone decrease (n = 15) and with aldosterone increase (n = 5) after HD showed both significant differences for aldosterone. A comparison of parameters depending on post-HD aldosterone decrease (e. g. systolic blood pressure 2 h after the begin of HD) or aldosterone increase (e. g. haemoglobin, haematocrit, metanephrines) is shown in Additional file 1: Table S4.

Pre-HD metanephrine values were generally above the normal range (<90 ng/l) for metanephrines and there was a significant decrease after HD (Fig. 3a).

Normetanephrine concentrations were pre-HD and post-HD elevated above the normal range (< 180 ng/l). A significant reduction of normetanephrines was detectable post-HD in comparison to pre-HD normetanephrine concentrations similarly to the metanephrines. The median decrease of normetanephrines with 56.1% was slightly higher compared to the metanephrine decrease (Fig. 3b). The percental reduction of metanephrines and normetanephrines was not significantly correlated with the percental reduction of the body weight after HD (Spearman test in bivariate correlation analysis, p = 0.887, r = −0.034, p = 0.572, r = − 0,134).

The echocardiographic data sets could be analysed in 15 of 20 patients in a standardized professional setting and methods [30, 32]. Three patients were excluded due to insufficient image quality and one patient each due to unknown mitral stenosis and atrial fibrillation. An overview of the echocardiographic parameters before and after HD is shown in Additional file 1: Table S5.

The maximum early-diastolic (Emax, p > 0.05), maximum late-diastolic (Amax; p < 0.05) velocity and the E/A-ratio (p < 0.05) of the mitral valve inflow were lower after HD whereas Amax and the E/A-ratio were significantly reduced after HD (Additional file 1: Table S5 and Fig. 4a). TAPSE (tricuspid annular plane systolic excursion, n = 16) and SPAP (systolic pulmonary arterial pressure, n = 13) were routinely performed for the investigation of the right ventricular function. TAPSE was not significantly different before 2.20 cm (1.5 to 3.0) and after 2.25 cm (1.6 to 3.0) HD (Wilcoxon-test, p = 0.57) but SPAP was significantly reduced from 33.01 mmHg (19.17 to 72.92, SPAP >30, n = 11) to 29.27 mmHg (14.79 to 53.43, SPAP >30, n = 6, Wilcoxon-test, p = 0.0051). Only the SPAP after HD was significantly different in patients with an UF rate less or bigger 2000 ml per session (p = 0.018) and RE less or bigger 500 ml per day (Mann-Whitney-U-test, p = 0.005). TAPSE, SPAP, copeptin, renin and aldosterone before and after HD were not significantly correlated in absolute values and in the difference values (before and after HD) in the correlation analysis (Spearman test, p > 0.05).

The maximum early-diastolic tissue velocity Vmax E’ in region of the basal inferoseptal left ventricular segment was not significantly reduced whereas the Vmax E’ in region of the basal lateral left ventricular segment was significantly reduced after HD (Additional file 1: Table S5). Significant changes in left ventricular end-diastolic pressure (E/E’ inferoseptal, E/E’ lateral) could not be observed after HD. However, values for Vmax E’ inferoseptal were lower than values for Vmax E’ lateral leading to an increased E/E’-ratio when the left ventricular pressure was estimated in region of the basal inferoseptal region of the left ventricle.

Concerning the comparison of the echocardiographic parameters depending on the UF the percentage reduction of the VCID was significantly higher in patients with an UF ≥ 2000 ml (n = 8) than in patients with an UF of <2000 ml (n = 7) after HD. The other parameters did not differ depending on the UF (Additional file 1: Table S6).

The comparison of the echocardiographic parameters depending on the RE showed lower maximum early-diastolic velocity values (Emax) in patients with a RE < 500 ml (n = 8) than in patients with a RE of ≥500 ml (n = 7) before HD. The other parameters did not differ depending on the RE (Additional file 1: Table S7).

In addition, significant correlations between echocardiographic parameters and biochemistry parameters have been revealed before and after haemodialysis, e.g. copeptin and renin with Amax, E/E’ inferoseptal and lateral. These data are shown in Additional file 1: Table S8.

The VCID did show ranges between 13 and 28 mm before and after HD. VCID values of less than 13 mm could be observed in two patients before HD and in seven patients after HD (Fig. 4b). Interestingly, a positive correlation between VCID before HD and patient weights before and after HD was calculated. In addition, the percentage decrease of VCID correlated to UR and other biochemistry parameters (Additional file 1: Table S9). A significant correlation between VCID and echocardiographic parameters or patient body weights has not been revealed.