- Email: [email protected]
Cardiac and hemodynamic effects of hemodialysis and ultrafiltration
Cardiac and Hemodynamic Effects of Hemodialysis and Ultrafiltration Willem Jan W. Bos, MD, Sjoerd Bruin, MSc, Rudolf W. van Olden, MD, Ingrid Keur, MD, Karel H. Wesseling, MSc, Nico Westerhof, MD, Raymond T. Krediet, MD, and Lambertus A. Arisz, MD ● Imbalance between cardiac oxygen supply and demand may trigger cardiac events in already vulnerable hemodialysis (HD) patients. We studied the effect of ultrafiltration (UF) and HD in nine chronic HD patients by continuously measuring blood volume (BV; by Critline), blood pressure (BP; by Portapres), and changes in hemodynamics (Modelflow) during isolated UF (iUF) of 500 mL in 30 minutes and subsequent HD combined with UF (HD ⴙ UF). Aortic pressure was reconstructed from finger pressure. Changes in cardiac oxygen supply were assessed by calculating the area under the aortic pressure curve during diastole (diastolic pressure time index [DPTI]). Changes in cardiac oxygen demand were assessed by calculating systolic pressure time index (SPTI). BV decreased 4.0% ⴞ 1.8% during UF and 7.3% ⴞ 3.3% during HD ⴙ UF (both P F 0.01). Systolic BP did not change; diastolic and mean BP increased 11 ⴞ 7.4 and 11 ⴞ 8.4 mm Hg during iUF, respectively (both P F 0.01), and stabilized during HD ⴙ UF. Overall pulse pressure decreased 19 ⴞ 11.1 mm Hg (P F 0.01). Heart rate increased 13 ⴞ 11 beats/min (P F 0.01) and systemic vascular resistance increased 59% ⴞ 51% (P F 0.01), whereas stroke volume and cardiac output (CO) decreased by 40% ⴞ 17% and 30% ⴞ 13%, respectively (both P F 0.01). Both cardiac oxygen supply (DPTI) and demand (SPTI) increased during iUF, and both decreased during HD ⴙ UF. By the end of the procedure, DPTI/SPTI ratio had increased 9% ⴞ 8% (P F 0.05). Changes in CO correlated closely to changes in BV. Despite large changes in hemodynamics during uncomplicated UF and HD, the balance between cardiac oxygen supply and demand (DPTI/SPTI ratio) did not decrease, but improved slightly. r 2000 by the National Kidney Foundation, Inc. INDEX WORDS: Hemodialysis (HD); hemodynamics; blood pressure (BP); myocardial ischemia; blood volume (BV).
T
HE INCIDENCE OF cardiac morbidity and mortality is high in patients treated with hemodialysis (HD).1 Left ventricular hypertrophy, coronary artery disease, and decreased capillary density in the myocardium make these patients vulnerable to ischemic events.1,2 Interventions causing changes in hemodynamics, such as the HD procedure itself,3-10 might result in a further imbalance between cardiac oxygen supply and demand and thus trigger cardiac events. However, little is known about the effect of these hemodynamic changes during HD on cardiac oxygen supply and demand. Cardiac oxygen demand might decrease during HD because of a decrease in preload, but it also might increase because of increasing heart rate and peripheral resistance. The net effect on the balance between cardiac oxygen supply and demand has only been studied during HD sessions in which no fluid was withdrawn from the patients.11,12 Because ultrafiltration (UF)-induced hypovolemia is the main cause of hemodynamic changes during HD,5 the balance between cardiac oxygen supply and demand should be investigated in HD sessions during which fluid is removed. The objective of this study is to investigate the net effect of UF and HD on cardiac oxygen supply and demand. We estimated both cardiac
oxygen supply and cardiac oxygen demand during isolated UF (iUF) and HD combined with UF (HD ⫹ UF). PATIENTS AND METHODS
Patients Measurements were performed in nine random chronic HD patients (7 men, 2 women); age, 44 ⫾ 17 years; range, 19 to 70 years). Six subjects in whom finger pressure could not be measured properly during HD were not entered into the study. The patients had renal failure for 6.9 ⫾ 4.7 years (range, 1 to 17 years). Seven patients were treated with erythropoietin (average dose, 5,111 U/wk). Four patients
From the Department of Nephrology, General Internal Medicine, and TNO-BMI, Academic Medical Center; Dialysis Center Dianet; and the Laboratory for Physiology, ICaR-VU, Free University, Amsterdam, The Netherlands. Received June 22, 1999; accepted in revised form November 10, 1999. Supported in part by grant no. NWO 902-18-307 from the Netherlands Foundation for Scientific Research (W.J.W.B). Address reprint requests to Willem Jan W. Bos, MD, Department of Internal Medicine, Sint Antonius Hospital, PO Box 2500, 3430 EM Nieuwegein, The Netherlands. E-mail: [email protected]
r 2000 by the National Kidney Foundation, Inc. 0272-6286/00/3505-0005$3.00/0 doi:10.1053/kd.2000.6377
American Journal of Kidney Diseases, Vol 35, No 5 (May), 2000: pp 819-826
819
820
were treated with vasoactive substances (three patients, -blocking agents on nondialysis days; three patients, calcium antagonists; one patient, long-acting nitrates; one patient, angiotension-converting enzyme inhibitor; and one patient, ␣-blocking agent). None of the patients had experienced a myocardial infarction, had been treated for heart failure, or had clinical signs of autonomic failure. The protocol was approved by the Medical Ethics Committee of the Academic Medical Center of the University of Amsterdam (The Netherlands). All participants gave informed consent.
Procedure Patients were dialyzed with an AK-100 (Gambro, Stockholm, Sweden) or Fresenius 4000 (Fresenius Medical Care, Bad Homburg, Germany) using polysulfone or cellulose triacetate membranes. The first 10 minutes after starting the extracorporeal circulation was used as a control period, during which no fluid was withdrawn and dialysate flow was zero (t ⫽ ⫺10 to t ⫽ 0). From 0 to 30 minutes, 500 mL was ultrafiltrated (iUF). The next 20 minutes was used to study the effect of vascular refill; again, no fluid was withdrawn and dialysate flow remained zero. During the next 3 hours, patients were treated with HD ⫹ UF. UF was performed as needed to achieve dry weight. Dialysate temperature was set at 36°C. This low temperature was used to prevent hypotensive episodes. Dialysate contained 32 or 35 mmol/L of bicarbonate, 1.5 mmol/L of calcium, 138 mmol/L of sodium, and 2 mmol/L of potassium. All subjects were in a semirecumbent position during the entire procedure. No additional fluid infusions were needed because no patient developed hypotension.
Measurements Blood volume (BV) change and arterial oxygen saturation (SaO2 ) were measured continuously with the Critline system (In-Line Diagnostics, Riverdale, UT).13 Briefly, the instrument uses optical techniques to measure the absorption and scattering properties of red blood cells passing through the extracorporeal circuit to determine hematocrit and SaO2. Changes in BV are calculated from changes in hematocrit.13 These indirect hematocrit measurements are in close agreement with direct hematocrit measurements.13 The instrument provides 20-second averages of BV and SaO2. Blood pressure (BP) was measured continuously with Portapres, a portable version of Finapres (TNO-BMI, Amsterdam, The Netherlands),14,15 using the middle finger of the nonfistula arm. Finger-pressure measurements have been validated against invasive BP measurements in many studies,14 including studies of subjects with signs of atherosclerotic vascular disease.16 All procedures were marked with an electronic time marker.
BOS ET AL
derived (TNO-BMI) changes in stroke volume (SV), cardiac output (CO), and systemic vascular resistance (SVR).17,18 SV calculations were not individually calibrated with invasive measurements. Therefore, relative changes from control values are given for changes in SV, CO, and SVR. Modelflow-derived changes in CO have been shown to follow thermodilution CO, with a mean deviation of 2% and an SD of 8% during cardiac surgery.17 In patients with septic shock, a bias of ⫺0.1 ⫾ 0.8 L/min was observed during a 2-day experiment.18 These results were in part17 obtained in a patient group with vascular disease. There are no specific pathophysiological reasons to expect problems with the ability of Modelflow to track differences in CO in patients with end-stage renal disease (ESRD). Therefore, we do not consider it necessary and ethically justified to perform validation measurements requiring right-sided cardiac catheterization in HD patients. To estimate left ventricular oxygen supply and demand, the diastolic pressure time index (DPTI) and the systolic pressure time index (SPTI)19,20 were calculated from aortic pressure waves.21 These aortic pressure waves were reconstructed from finger pressure waves using a generalized transfer function.22 The transfer function compensates for the physiological wave distortion of pressure waves traveling toward the periphery by attenuating high-frequency components while slightly amplifying low-frequency components. The present transfer function from finger to aorta was developed in 10 patients by adding a brachial artery– aorta transfer function to the previously published finger-tobrachial-artery transfer function.23,24 As expected, the resulting transfer function closely resembles published transfer functions from finger or radial artery to aorta, with a maximal attenuation at 4 Hz.22,25 To estimate the effect on cardiac oxygen supply, the area under the diastolic part of the aortic pressure curve was calculated (DPTI, measured in millimeters of mercury ⫻ second).20 To estimate the effect on cardiac oxygen demand, the area under the systolic part of the aortic pressure curve was calculated (SPTI, in millimeters of mercury ⫻ second).20 The balance between oxygen supply and demand was expressed as DPTI/SPTI ratio.11,12,20,21 For all parameters, the average value during 2-minute periods at t ⫽ ⫺10, 0, 30, 50, 80, 110, 140, 170, 200, and 230 minutes was calculated.
Statistics All parameters were tested to have a normal distribution. Results are expressed as mean ⫾ SD. After a two-way analysis of variance, paired t-tests were used with a Bonferroni correction for multiple comparisons to test the significance of changes during different procedures. Correlation coefficients were calculated to establish the relationship between changes in different parameters.
RESULTS
Data Analysis
Dialysis
BP and time marker were analogue to digital (A/D) converted with a sampling rate of 100 Hz and stored in a computer. Beat-to-beat values of systolic, mean, and diastolic pressure in millimeters of mercury and heart rate in beats per minute were calculated, together with Modelflow-
During iUF, 500 mL of fluid was withdrawn in all patients. During the subsequent 3-hour HD ⫹ UF, a further 1,544 ⫾ 663 mL of fluid was withdrawn.
CARDIAC EFFECTS OF HD
821
BV BV decreased by 4.0% ⫾ 1.8% during iUF (P ⬍ 0.01). It tended to increase during the refill phase (1.0% ⫾ 2.1%; increased in eight of nine patients; P ⫽ not significant [NS]) and decreased during HD ⫹ UF by 7.3% ⫾ 3.3% (P ⬍ 0.01; Table 1 and Fig 1). Hemodynamics The hemodynamic responses during iUF, refill, and HD ⫹ UF are shown in Table 1 and Fig 2. In short, during iUF, systolic BP did not change and diastolic and mean BP increased by 11 ⫾ 7.4 and 11 ⫾ 8.4 mm Hg, respectively (both P ⬍ 0.01). SV and CO decreased, whereas SVR increased. No significant changes in hemodynamic parameters occurred during the refill phase. During HD ⫹ UF, BP did not change further. Heart rate increased, and SV and CO decreased further. The ejection time decreased from 302 ⫾ 29 to 261 ⫾ 40 ms (P ⬍ 0.01). Diastolic time did not change significantly (522 ⫾ 138 ms at t ⫽ 0 versus 460 ⫾ 135 ms at
t ⫽ 240; P ⫽ NS). Also, the relative contribution of the systolic and the diastolic periods to the cycle length did not change significantly (systolic time, 37.4% ⫾ 4.7% at t ⫽ 0 versus 37.0% ⫾ 4.0% at t ⫽ 240; diastolic time, 62.6% ⫾ 4.7% at t ⫽ 0 versus 63.0% ⫾ 4.0% at t ⫽ 240). Cardiac Oxygen Supply and Demand SaO2 did not change during the entire procedure (Table 1). DPTI increased during iUF, did not change during the refill phase, and decreased during HD ⫹ UF. SPTI increased during both iUF and the refill phase. During HD ⫹ UF, it decreased significantly (Table 1). DPTI/SPTI ratio, representing the balance between cardiac oxygen supply and demand, did not change during the three periods (Table 1). However, when the results of the various phases are combined, an improvement in DPTI/SPTI ratio by 0.14 ⫾ 0.16, or 8.8% ⫾ 7.8%, was found (P ⬍ 0.05; Table 1 and Fig 3).
Table 1. Baseline Values and Changes During Different Procedures Change During Baseline
Blood pressure (mm Hg) Systolic Mean Diastolic Pulse pressure Heart rate (beats/min) Ejection time (ms) Stroke volume (%) Cardiac output (%) SVR (%) Blood volume (% change) Cardiac oxygen supply DPTI (mm Hg · s) Oxygen saturation (%) Cardiac oxygen demand SPTI (mm Hg · s) Supply/demand DPTI/SPTI
iUF
UF ⫹ HD
0 min
0 v 30 min
50 v 230 min
138 ⫾ 22.7 82 ⫾ 19.5 59 ⫾ 17.7 79 ⫾ 18.9 75 ⫾ 13.2 302 ⫾ 30.3 100 100 100 100
3 ⫾ 11.8 11 ⫾ 7.4* 11 ⫾ 6.7* ⫺8 ⫾ 11.7 ⫺1 ⫾ 4.6 3 ⫾ 7.8 ⫺14 ⫾ 11.0* ⫺13 ⫾ 13.7† 33 ⫾ 27.9* ⫺4.0 ⫾ 1.8*
⫺13 ⫾ 23.7 ⫺9 ⫾ 16.2 ⫺3 ⫾ 12.9 ⫺9 ⫾ 17.0 14 ⫾ 11.7* ⫺51 ⫾ 25.5* ⫺23 ⫾ 20.8† ⫺13 ⫾ 16.0† 10 ⫾ 38.7 ⫺7.3 ⫾ 3.3*
38 ⫾ 11.5 98 ⫾ 1.7
7 ⫾ 8.0† 0 ⫾ 0.7
⫺7 ⫾ 8.3† 0 ⫾ 1.3
30 ⫾ 7.3
3 ⫾ 2.3*
⫺8 ⫾ 6.1*
0.08 ⫾ 0.17
0.09 ⫾ 0.15
1.24 ⫾ 0.23
NOTE. Absolute values are given for blood pressure, heart rate, ejection time, DPTI, oxygen saturation, SPTI, and DPTI/SPTI ratio. Stroke volume, cardiac output, systemic vascular resistance, and blood volume are presented as relative changes from baseline values. Values expressed as mean ⫾ SD. *P ⬍ 0.01 in paired t-test (time 30 versus time 0 and time 230 versus time 50). †P ⬍ 0.05 in paired t-test (time 30 versus time 0 and time 230 versus time 50).
822
BOS ET AL
Fig 1. Changes in blood volume during isolated ultrafiltration and hemodialysis combined with ultrafiltration are expressed as percentage of change from baseline. Data are presented as mean ⴞ SD. Comparisons were made against baseline values (t ⴝ 0). *P F 0.05. **P F 0.01 in paired t-test.
Correlations Changes in BV did not correlate with changes in systolic (r 2 ⫽ 0.32; P ⫽ NS), mean (r 2 ⫽ 0.07; P ⫽ NS), or diastolic pressure (r 2 ⫽ 0.34; P ⫽ NS). However, changes in pulse pressure strongly correlated with changes in BV (r 2 ⫽ 0.93; P ⬍ 0.01), as were changes in SV and CO (r 2 ⫽ 0.97 and 0.90; P ⬍ 0.01). The correlation between changes in BV and changes in SV was also very strong in most individual subjects (r 2, 0.38 to 0.96; mean, 0.81; median, 0.85; Fig 4). Changes in DPTI were associated with changes in both diastolic pressure (r 2 ⫽ 0.47; P ⬍ 0.05) and duration of the diastolic period (r 2 ⫽ 0.41; P ⬍ 0.05). Changes in SPTI were equally well associated with changes in systolic pressure (r 2 ⫽ 0.74; P ⬍ 0.01) and ejection time (r 2 ⫽ 0.73; P ⬍ 0.01). Changes in DPTI and SPTI were not related to changes in BV (r 2 ⫽ 0.02 and 0.20, respectively; P ⫽ NS). DISCUSSION
In the present study, a slight improvement of the balance between cardiac oxygen supply and demand was observed. Because DPTI and SPTI represent the area under the aortic pressure curve during diastole and systole, respectively, they are affected by changes in diastolic and systolic BP and the length of the diastolic and systolic period. In the present study, changes in DPTI and SPTI depended on changes in both BP and dia-
stolic and systolic periods. Diastolic BP increased during iUF because of an increase in SVR and remained stable thereafter, despite a significant decrease in BV and CO. The increase in BP during iUF might be caused by a decrease in body temperature.3 Systolic pressure did not change significantly during the procedure, but pulse pressure decreased significantly because of the decrease in SV, which in turn strongly correlated with the decrease in BV. Because systolic and diastolic time decreased at a similar rate, the observed positive effect on DPTI/SPTI ratio can mainly be attributed to the opposing changes in diastolic and pulse pressures. What is the clinical significance of an increase in DPTI/SPTI ratio from 1.24 ⫾ 0.23 to 1.39 ⫾ 0.14? Estimation of DPTI, SPTI, and their ratio yields indirect information about cardiac oxygen supply and demand and has limitations.20 We used these parameters because both cardiac oxygen demand and supply can be estimated in a noninvasive way. Measurement of oxygen saturation in coronary arteries and veins in combination with coronary and ventricular pressure and coronary flow measurements would yield superior information, but such measurements are not feasible during HD. Generally, subendocardial perfusion is considered to be hampered if DPTI/ SPTI ratio is less than 0.6.20 In our studies, the values observed remained much greater than this limit. However, we over-
CARDIAC EFFECTS OF HD
Fig 2. Blood pressure, pulse pressure (Pulse Pr), heart rate (HR) and changes in stroke volume (SV), cardiac output (CO), and systemic vascular resistance (SVR) during isolated ultrafiltration and hemodialysis combined with ultrafiltration. Changes in SV, CO, and SVR are expressed as percentage of change from baseline. Data are presented as mean ⴞ SD. Comparisons were made against baseline values (t ⴝ 0). *P F 0.05. **P F 0.01 in paired t-test.
823
estimated DPTI and thus the DPTI/SPTI ratio because left ventricular diastolic pressure, which might be elevated in patients with ESRD because of left ventricular hypertrophy and fluid overload, was not, as originally described by Buckberg et al,26,27 subtracted from the aortic pressure. Furthermore, the critical DPTI/SPTI ratio is likely to be much greater in subjects with ESRD because the prevalence of left ventricular hypertrophy and coronary artery disease is very high.1 Their hearts are furthermore characterized by extensive interstitial fibrosis.1 Coronary artery stenoses decrease the actual perfusion pressure in the coronary vascular bed,20 whereas left ventricular hypertrophy and interstitial fibrosis increase the resistance to flow while decreasing the vasodilatory reserve.1,2 Thus, the supply of oxygen might be diminished in patients with ESRD, whereas the demand is generally increased because of left ventricular hypertrophy.2 Any change in supply/demand ratio might thus be of clinical significance in this population. We chose to use a dialysate temperature of 36°C to prevent the occurrence of hypotensive episodes.3 It should be realized that the use of dialysate of a higher temperature might compromise coronary perfusion because of hypotensive episodes and temperature-induced vasodilation. Because we did not observe hypotensive episodes in the present study, we were not able to quantify their effect on the balance between coronary oxygen supply and demand. Questions can be raised concerning the reliability of finger-pressure measurements, the reconstruction of the aortic pressure wave, and Modelflow calculations of SV, CO, and SVR. Therefore, these items are discussed here. We had some difficulty finding subjects in whom finger pressure could adequately be measured during HD. This could be caused by the effect of previous arteriovenous fistulas on the nonfistula arm, peripheral vascular disease, or vasoconstriction during HD. That a proper finger-pressure signal could not be obtained during HD in all patients might limit the clinical applicability of fingerpressure measurements in dialysis populations. We consider this technique mainly a noninvasive research tool. Furthermore, it cannot be excluded that a selection bias was introduced by performing the measurements only in selected subjects in whom a proper finger-pressure signal could be
824
BOS ET AL
Fig 3. The balance between oxygen supply and demand during isolated ultrafiltration and hemodialysis combined with ultrafiltration is represented by DPTI/SPTI ratio. Left ventricular oxygen supply is estimated from DPTI and oxygen demand from SPTI. Data are presented as mean ⴞ SD. Comparisons were made against baseline values (t ⴝ 0). *P F 0.05. **P F 0.01 in paired ttest.
obtained. In dialysis patients in whom finger pressures can be measured adequately, this finger pressure might underestimate more centrally measured BPs, as in the elderly and subjects with atherosclerotic disease.16,28 However, group average changes in these central pressures are tracked reliably with finger-pressure measurements, even in populations in which the absolute pressure level at the finger is less than more centrally measured pressures.14,16,28 To estimate the effect of HD and UF on
Fig 4. Group average changes in stroke volume at all measurement points during the experiment are closely related to changes in blood volume (r 2 ⴝ 0.97; P F 0.01).
cardiac supply and demand, we calculated DPTI and SPTI from aortic pressure waves reconstructed from finger-pressure measurements using a generalized waveform filter.21,22 On its way to the periphery, especially the systolic part of the pressure wave is amplified. Changes in heart rate cause further changes in pressure amplification,29 making calculations of SPTI in peripheral pressure waves less meaningful. Generalized waveform filters correct for such pulse-wave amplifications.22-25 Modelflow-derived calculations of changes in CO have been validated against changes in thermodilution CO during cardiac surgery and treatment in an intensive care unit (mean error, 2% ⫾ 8%).17,18 In this respect, the close correlation between changes in BV and SV, measured with two independent techniques, was a reassuring finding. A correlation between BV and SV has been predicted in computer models describing hemodynamic changes during HD30 and described in studies in which CO was determined before and after HD.8,31 Our continuous measurements of both BV and SV yield individual BV-SV curves. The individual correlations were strong, even though the slope of these FrankStarling–like curves differed between individuals. Differences in the slope might be attributed to differences in cardiac contractility. However,
CARDIAC EFFECTS OF HD
differences in venous compliance and heart rate response might also contribute to differences in the slope. Thus, the combined use of BV and SV measurements might provide a useful noninvasive tool to study cardiac physiology during HD. In conclusion, uncomplicated UF and HD induce large changes in hemodynamics. These changes cause a slight improvement in the balance between cardiac oxygen supply and demand. REFERENCES 1. Parfrey PS, Foley RN, Harnett JD: Organ and metabolic complications: Cardiac, in Jacobs C, Kjellstrand CM, Koch KM, Winchester JF (eds): Replacement of Renal Function by Dialysis. Dordrecht, The Netherlands, Kluwer, 1996, pp 990-1002 2. London GM, Marchais SJ, Guerin AP, Metivier F, Pannier B: Cardiac hypertrophy and arterial alterations in end-stage renal disease: Hemodynamic factors. Kidney Int 43:S42-S49, 1993 (suppl 41) 3. Van Kuijk WHM, Luik AJ, de Leeuw PW, van Hooff JP, Nieman FHM, Habets HML, Leunissen KML: Vascular reactivity during hemodialysis and isolated ultrafiltration: Thermal influences. Nephrol Dial Transplant 10:1852-1858, 1995 4. Van Kuijk WHM, Leunissen KML: Hemodynamic stability during different forms of dialysis therapy: A pathogenetic analysis. Blood Purif 14:405-420, 1996 5. Daugirdas JT: Dialysis hypotension: A hemodynamic analysis. Kidney Int 39:233-246, 1991 6. Movilli E, Camerini C, Viola BF, Bossini N, Strada A, Maiorca R: Blood volume changes during three different profiles of dialysate sodium variation with similar intradialytic sodium balances in chronic hemodialyzed patients. Am J Kidney Dis 30:58-63, 1997 7. Levy FL, Grayburn PA, Foulks CJ, Brickner ME, Henrich WL: Improved left ventricular contractility with cool temperature hemodialysis. Kidney Int 41:961-965, 1992 8. Chaignon M, Chen WT, Tarazi RC, Nakamoto S, Bravo EL: Blood pressure response to hemodialysis. Hypertension 3:333-339, 1981 9. Mehta BR, Fischer D, Ahmad M, Dubose TD: Effects of acetate and bicarbonate hemodialysis on cardiac function in chronic dialysis patients. Kidney Int 24:782-787, 1983 10. Nixon JV, Mitchell JH, McPhaul JJ, Henrich WL: Effect of hemodialysis on left ventricular function: Dissociation of changes in filling volume and in contractile states. J Clin Invest 71:377-384, 1983 11. Pedersen T, Rasmussen K, Cleemann-Rasmussen K: Effect of hemodialysis on cardiac performance and transmural myocardial perfusion. Clin Nephrol 19:31-36, 1983 12. Wolff J, Pedersen T, Rossen M, Cleemann-Rasmussen K: Effects of acetate and bicarbonate dialysis on cardiac performance, transmural myocardial perfusion and acidbase balance. Int J Artif Organs 9:105-110, 1986
825
13. Steuer RR, Harris DH, Conis JM: A new optical technique for monitoring hematocrit and circulating blood volume: Its application in renal dialysis. Dial Transplant 22:260-265, 1993 14. Imholz BPM, Wieling W, van Montfrans GA, Wesseling KH: Fifteen years of experience with finger arterial pressure monitoring: Assessment of the technology. Cardiovasc Res 38:605-616, 1998 15. Imholz BPM, Langewouters GJ, van Montfrans GA, Parati G, van Goudoever J, Wesseling KH, Wieling W, Mancia G: Feasibility of ambulatory, continuous, 24-hour finger pressure recording. Hypertension 21:65-73, 1993 16. Bos WJW, Imholz BPM, van Goudoever J, Wesseling KH, van Montfrans GA: The reliability of noninvasive continuous finger blood pressure measurement in patients with both hypertension and vascular disease. Am J Hypertens 5:529-535, 1995 17. Wesseling KH, Jansen JRC, Settels JR, Schreuder JJ: Computation of aortic flow from pressure in humans using a nonlinear, three element model. J Appl Physiol 74:25662573, 1993 18. Jellema WT, Wesseling KH, Groeneveld ABJ, Stoutenbeek CP, Thijs LG, van Lieshout JJ: Continuous cardiac output in septic shock by simulating a model of the aortic input impedance. Anesthesiology 90:1317-1328, 1999 19. Sarnoff SJ, Braunwald E, Welch GH, Case RB, Stainsby WN, Macruz R: Hemodynamic determinants of oxygen consumption of the heart with special reference to the tension-time index. Am J Physiol 192:148-156, 1958 20. Hoffman JIE, Buckberg GD: The myocardial supply: demand ratio—A critical review. Am J Cardiol 41:327-332, 1978 21. Bos WJW, Zietse R, Wesseling KH, Westerhof N: Effect of arteriovenous fistulas on cardiac oxygen supply and demand. Kidney Int 55:2049-2053, 1999 22. Karamanoglu M, Feneley MP: On-line synthesis of the human ascending aortic pressure pulse from the finger pulse. Hypertension 30:1416-1424, 1997 23. Bos WJW, van Goudoever J, van Montfrans GA, van den Meiracker AH, Wesseling KH: Reconstruction of brachial artery pressure from noninvasive finger pressure measurements. Circulation 94:1870-1875, 1996 24. Gizdulich P, Prentza A, Wesseling KH: Models of brachial to finger pulse wave distortion and pressure decrements. Cardiovasc Res 33:698-705, 1997 25. Chen CH, Nevo E, Fetics B, Pak PH, Yin FCP, Maugham WL, Kass DA: Estimation of central aortic pressure waveform by mathematical transformation of radial tonometry pressure: Validation of generalized transfer function. Circulation 95:1827-1836, 1997 26. Buckberg GD, Fixler DE, Archie JP, Hoffman JIE: Experimental subendocardial ischemia in dogs with normal coronary arteries. Circ Res 30:67-81, 1972 27. Buckberg GD, Fixler DE, Archie JP, Henney RP, Hoffman JIE: Variable effects of heart rate on phasic and regional left ventricular muscle blood flow in anaesthetized dogs. Cardiovasc Res 9:1-11, 1975 28. Rongen GA, Bos WJW, Lenders JWM, van Montfrans GA, van Lier HJJ, van Goudoever J, Wesseling KH,
826
Thien T: Comparison of intrabrachial and finger blood pressure in healthy elderly volunteers. Am J Hypertens 8:237-248, 1995 29. Bos WJW, van den Meiracker AH, Wesseling KH, Schalekamp MADH: Effect of regional and systemic changes in vasomotor tone on finger pressure amplification. Hypertension 26:315-320, 1995
BOS ET AL
30. Ursino M, Innocenti M: Mathematical investigation of some physiological factors involved in hemodialysis hypotension. Artif Organs 21:891-902, 1997 31. Kouw PM, Kooman JP, Cheriex EC, Olthof CG, de Vries PMJM, Leunissen KML: Assessment of postdialysis dry weight: A comparison of techniques. J Am Soc Nephrol 4:98-104, 1993
T
HE INCIDENCE OF cardiac morbidity and mortality is high in patients treated with hemodialysis (HD).1 Left ventricular hypertrophy, coronary artery disease, and decreased capillary density in the myocardium make these patients vulnerable to ischemic events.1,2 Interventions causing changes in hemodynamics, such as the HD procedure itself,3-10 might result in a further imbalance between cardiac oxygen supply and demand and thus trigger cardiac events. However, little is known about the effect of these hemodynamic changes during HD on cardiac oxygen supply and demand. Cardiac oxygen demand might decrease during HD because of a decrease in preload, but it also might increase because of increasing heart rate and peripheral resistance. The net effect on the balance between cardiac oxygen supply and demand has only been studied during HD sessions in which no fluid was withdrawn from the patients.11,12 Because ultrafiltration (UF)-induced hypovolemia is the main cause of hemodynamic changes during HD,5 the balance between cardiac oxygen supply and demand should be investigated in HD sessions during which fluid is removed. The objective of this study is to investigate the net effect of UF and HD on cardiac oxygen supply and demand. We estimated both cardiac
oxygen supply and cardiac oxygen demand during isolated UF (iUF) and HD combined with UF (HD ⫹ UF). PATIENTS AND METHODS
Patients Measurements were performed in nine random chronic HD patients (7 men, 2 women); age, 44 ⫾ 17 years; range, 19 to 70 years). Six subjects in whom finger pressure could not be measured properly during HD were not entered into the study. The patients had renal failure for 6.9 ⫾ 4.7 years (range, 1 to 17 years). Seven patients were treated with erythropoietin (average dose, 5,111 U/wk). Four patients
From the Department of Nephrology, General Internal Medicine, and TNO-BMI, Academic Medical Center; Dialysis Center Dianet; and the Laboratory for Physiology, ICaR-VU, Free University, Amsterdam, The Netherlands. Received June 22, 1999; accepted in revised form November 10, 1999. Supported in part by grant no. NWO 902-18-307 from the Netherlands Foundation for Scientific Research (W.J.W.B). Address reprint requests to Willem Jan W. Bos, MD, Department of Internal Medicine, Sint Antonius Hospital, PO Box 2500, 3430 EM Nieuwegein, The Netherlands. E-mail: [email protected]
r 2000 by the National Kidney Foundation, Inc. 0272-6286/00/3505-0005$3.00/0 doi:10.1053/kd.2000.6377
American Journal of Kidney Diseases, Vol 35, No 5 (May), 2000: pp 819-826
819
820
were treated with vasoactive substances (three patients, -blocking agents on nondialysis days; three patients, calcium antagonists; one patient, long-acting nitrates; one patient, angiotension-converting enzyme inhibitor; and one patient, ␣-blocking agent). None of the patients had experienced a myocardial infarction, had been treated for heart failure, or had clinical signs of autonomic failure. The protocol was approved by the Medical Ethics Committee of the Academic Medical Center of the University of Amsterdam (The Netherlands). All participants gave informed consent.
Procedure Patients were dialyzed with an AK-100 (Gambro, Stockholm, Sweden) or Fresenius 4000 (Fresenius Medical Care, Bad Homburg, Germany) using polysulfone or cellulose triacetate membranes. The first 10 minutes after starting the extracorporeal circulation was used as a control period, during which no fluid was withdrawn and dialysate flow was zero (t ⫽ ⫺10 to t ⫽ 0). From 0 to 30 minutes, 500 mL was ultrafiltrated (iUF). The next 20 minutes was used to study the effect of vascular refill; again, no fluid was withdrawn and dialysate flow remained zero. During the next 3 hours, patients were treated with HD ⫹ UF. UF was performed as needed to achieve dry weight. Dialysate temperature was set at 36°C. This low temperature was used to prevent hypotensive episodes. Dialysate contained 32 or 35 mmol/L of bicarbonate, 1.5 mmol/L of calcium, 138 mmol/L of sodium, and 2 mmol/L of potassium. All subjects were in a semirecumbent position during the entire procedure. No additional fluid infusions were needed because no patient developed hypotension.
Measurements Blood volume (BV) change and arterial oxygen saturation (SaO2 ) were measured continuously with the Critline system (In-Line Diagnostics, Riverdale, UT).13 Briefly, the instrument uses optical techniques to measure the absorption and scattering properties of red blood cells passing through the extracorporeal circuit to determine hematocrit and SaO2. Changes in BV are calculated from changes in hematocrit.13 These indirect hematocrit measurements are in close agreement with direct hematocrit measurements.13 The instrument provides 20-second averages of BV and SaO2. Blood pressure (BP) was measured continuously with Portapres, a portable version of Finapres (TNO-BMI, Amsterdam, The Netherlands),14,15 using the middle finger of the nonfistula arm. Finger-pressure measurements have been validated against invasive BP measurements in many studies,14 including studies of subjects with signs of atherosclerotic vascular disease.16 All procedures were marked with an electronic time marker.
BOS ET AL
derived (TNO-BMI) changes in stroke volume (SV), cardiac output (CO), and systemic vascular resistance (SVR).17,18 SV calculations were not individually calibrated with invasive measurements. Therefore, relative changes from control values are given for changes in SV, CO, and SVR. Modelflow-derived changes in CO have been shown to follow thermodilution CO, with a mean deviation of 2% and an SD of 8% during cardiac surgery.17 In patients with septic shock, a bias of ⫺0.1 ⫾ 0.8 L/min was observed during a 2-day experiment.18 These results were in part17 obtained in a patient group with vascular disease. There are no specific pathophysiological reasons to expect problems with the ability of Modelflow to track differences in CO in patients with end-stage renal disease (ESRD). Therefore, we do not consider it necessary and ethically justified to perform validation measurements requiring right-sided cardiac catheterization in HD patients. To estimate left ventricular oxygen supply and demand, the diastolic pressure time index (DPTI) and the systolic pressure time index (SPTI)19,20 were calculated from aortic pressure waves.21 These aortic pressure waves were reconstructed from finger pressure waves using a generalized transfer function.22 The transfer function compensates for the physiological wave distortion of pressure waves traveling toward the periphery by attenuating high-frequency components while slightly amplifying low-frequency components. The present transfer function from finger to aorta was developed in 10 patients by adding a brachial artery– aorta transfer function to the previously published finger-tobrachial-artery transfer function.23,24 As expected, the resulting transfer function closely resembles published transfer functions from finger or radial artery to aorta, with a maximal attenuation at 4 Hz.22,25 To estimate the effect on cardiac oxygen supply, the area under the diastolic part of the aortic pressure curve was calculated (DPTI, measured in millimeters of mercury ⫻ second).20 To estimate the effect on cardiac oxygen demand, the area under the systolic part of the aortic pressure curve was calculated (SPTI, in millimeters of mercury ⫻ second).20 The balance between oxygen supply and demand was expressed as DPTI/SPTI ratio.11,12,20,21 For all parameters, the average value during 2-minute periods at t ⫽ ⫺10, 0, 30, 50, 80, 110, 140, 170, 200, and 230 minutes was calculated.
Statistics All parameters were tested to have a normal distribution. Results are expressed as mean ⫾ SD. After a two-way analysis of variance, paired t-tests were used with a Bonferroni correction for multiple comparisons to test the significance of changes during different procedures. Correlation coefficients were calculated to establish the relationship between changes in different parameters.
RESULTS
Data Analysis
Dialysis
BP and time marker were analogue to digital (A/D) converted with a sampling rate of 100 Hz and stored in a computer. Beat-to-beat values of systolic, mean, and diastolic pressure in millimeters of mercury and heart rate in beats per minute were calculated, together with Modelflow-
During iUF, 500 mL of fluid was withdrawn in all patients. During the subsequent 3-hour HD ⫹ UF, a further 1,544 ⫾ 663 mL of fluid was withdrawn.
CARDIAC EFFECTS OF HD
821
BV BV decreased by 4.0% ⫾ 1.8% during iUF (P ⬍ 0.01). It tended to increase during the refill phase (1.0% ⫾ 2.1%; increased in eight of nine patients; P ⫽ not significant [NS]) and decreased during HD ⫹ UF by 7.3% ⫾ 3.3% (P ⬍ 0.01; Table 1 and Fig 1). Hemodynamics The hemodynamic responses during iUF, refill, and HD ⫹ UF are shown in Table 1 and Fig 2. In short, during iUF, systolic BP did not change and diastolic and mean BP increased by 11 ⫾ 7.4 and 11 ⫾ 8.4 mm Hg, respectively (both P ⬍ 0.01). SV and CO decreased, whereas SVR increased. No significant changes in hemodynamic parameters occurred during the refill phase. During HD ⫹ UF, BP did not change further. Heart rate increased, and SV and CO decreased further. The ejection time decreased from 302 ⫾ 29 to 261 ⫾ 40 ms (P ⬍ 0.01). Diastolic time did not change significantly (522 ⫾ 138 ms at t ⫽ 0 versus 460 ⫾ 135 ms at
t ⫽ 240; P ⫽ NS). Also, the relative contribution of the systolic and the diastolic periods to the cycle length did not change significantly (systolic time, 37.4% ⫾ 4.7% at t ⫽ 0 versus 37.0% ⫾ 4.0% at t ⫽ 240; diastolic time, 62.6% ⫾ 4.7% at t ⫽ 0 versus 63.0% ⫾ 4.0% at t ⫽ 240). Cardiac Oxygen Supply and Demand SaO2 did not change during the entire procedure (Table 1). DPTI increased during iUF, did not change during the refill phase, and decreased during HD ⫹ UF. SPTI increased during both iUF and the refill phase. During HD ⫹ UF, it decreased significantly (Table 1). DPTI/SPTI ratio, representing the balance between cardiac oxygen supply and demand, did not change during the three periods (Table 1). However, when the results of the various phases are combined, an improvement in DPTI/SPTI ratio by 0.14 ⫾ 0.16, or 8.8% ⫾ 7.8%, was found (P ⬍ 0.05; Table 1 and Fig 3).
Table 1. Baseline Values and Changes During Different Procedures Change During Baseline
Blood pressure (mm Hg) Systolic Mean Diastolic Pulse pressure Heart rate (beats/min) Ejection time (ms) Stroke volume (%) Cardiac output (%) SVR (%) Blood volume (% change) Cardiac oxygen supply DPTI (mm Hg · s) Oxygen saturation (%) Cardiac oxygen demand SPTI (mm Hg · s) Supply/demand DPTI/SPTI
iUF
UF ⫹ HD
0 min
0 v 30 min
50 v 230 min
138 ⫾ 22.7 82 ⫾ 19.5 59 ⫾ 17.7 79 ⫾ 18.9 75 ⫾ 13.2 302 ⫾ 30.3 100 100 100 100
3 ⫾ 11.8 11 ⫾ 7.4* 11 ⫾ 6.7* ⫺8 ⫾ 11.7 ⫺1 ⫾ 4.6 3 ⫾ 7.8 ⫺14 ⫾ 11.0* ⫺13 ⫾ 13.7† 33 ⫾ 27.9* ⫺4.0 ⫾ 1.8*
⫺13 ⫾ 23.7 ⫺9 ⫾ 16.2 ⫺3 ⫾ 12.9 ⫺9 ⫾ 17.0 14 ⫾ 11.7* ⫺51 ⫾ 25.5* ⫺23 ⫾ 20.8† ⫺13 ⫾ 16.0† 10 ⫾ 38.7 ⫺7.3 ⫾ 3.3*
38 ⫾ 11.5 98 ⫾ 1.7
7 ⫾ 8.0† 0 ⫾ 0.7
⫺7 ⫾ 8.3† 0 ⫾ 1.3
30 ⫾ 7.3
3 ⫾ 2.3*
⫺8 ⫾ 6.1*
0.08 ⫾ 0.17
0.09 ⫾ 0.15
1.24 ⫾ 0.23
NOTE. Absolute values are given for blood pressure, heart rate, ejection time, DPTI, oxygen saturation, SPTI, and DPTI/SPTI ratio. Stroke volume, cardiac output, systemic vascular resistance, and blood volume are presented as relative changes from baseline values. Values expressed as mean ⫾ SD. *P ⬍ 0.01 in paired t-test (time 30 versus time 0 and time 230 versus time 50). †P ⬍ 0.05 in paired t-test (time 30 versus time 0 and time 230 versus time 50).
822
BOS ET AL
Fig 1. Changes in blood volume during isolated ultrafiltration and hemodialysis combined with ultrafiltration are expressed as percentage of change from baseline. Data are presented as mean ⴞ SD. Comparisons were made against baseline values (t ⴝ 0). *P F 0.05. **P F 0.01 in paired t-test.
Correlations Changes in BV did not correlate with changes in systolic (r 2 ⫽ 0.32; P ⫽ NS), mean (r 2 ⫽ 0.07; P ⫽ NS), or diastolic pressure (r 2 ⫽ 0.34; P ⫽ NS). However, changes in pulse pressure strongly correlated with changes in BV (r 2 ⫽ 0.93; P ⬍ 0.01), as were changes in SV and CO (r 2 ⫽ 0.97 and 0.90; P ⬍ 0.01). The correlation between changes in BV and changes in SV was also very strong in most individual subjects (r 2, 0.38 to 0.96; mean, 0.81; median, 0.85; Fig 4). Changes in DPTI were associated with changes in both diastolic pressure (r 2 ⫽ 0.47; P ⬍ 0.05) and duration of the diastolic period (r 2 ⫽ 0.41; P ⬍ 0.05). Changes in SPTI were equally well associated with changes in systolic pressure (r 2 ⫽ 0.74; P ⬍ 0.01) and ejection time (r 2 ⫽ 0.73; P ⬍ 0.01). Changes in DPTI and SPTI were not related to changes in BV (r 2 ⫽ 0.02 and 0.20, respectively; P ⫽ NS). DISCUSSION
In the present study, a slight improvement of the balance between cardiac oxygen supply and demand was observed. Because DPTI and SPTI represent the area under the aortic pressure curve during diastole and systole, respectively, they are affected by changes in diastolic and systolic BP and the length of the diastolic and systolic period. In the present study, changes in DPTI and SPTI depended on changes in both BP and dia-
stolic and systolic periods. Diastolic BP increased during iUF because of an increase in SVR and remained stable thereafter, despite a significant decrease in BV and CO. The increase in BP during iUF might be caused by a decrease in body temperature.3 Systolic pressure did not change significantly during the procedure, but pulse pressure decreased significantly because of the decrease in SV, which in turn strongly correlated with the decrease in BV. Because systolic and diastolic time decreased at a similar rate, the observed positive effect on DPTI/SPTI ratio can mainly be attributed to the opposing changes in diastolic and pulse pressures. What is the clinical significance of an increase in DPTI/SPTI ratio from 1.24 ⫾ 0.23 to 1.39 ⫾ 0.14? Estimation of DPTI, SPTI, and their ratio yields indirect information about cardiac oxygen supply and demand and has limitations.20 We used these parameters because both cardiac oxygen demand and supply can be estimated in a noninvasive way. Measurement of oxygen saturation in coronary arteries and veins in combination with coronary and ventricular pressure and coronary flow measurements would yield superior information, but such measurements are not feasible during HD. Generally, subendocardial perfusion is considered to be hampered if DPTI/ SPTI ratio is less than 0.6.20 In our studies, the values observed remained much greater than this limit. However, we over-
CARDIAC EFFECTS OF HD
Fig 2. Blood pressure, pulse pressure (Pulse Pr), heart rate (HR) and changes in stroke volume (SV), cardiac output (CO), and systemic vascular resistance (SVR) during isolated ultrafiltration and hemodialysis combined with ultrafiltration. Changes in SV, CO, and SVR are expressed as percentage of change from baseline. Data are presented as mean ⴞ SD. Comparisons were made against baseline values (t ⴝ 0). *P F 0.05. **P F 0.01 in paired t-test.
823
estimated DPTI and thus the DPTI/SPTI ratio because left ventricular diastolic pressure, which might be elevated in patients with ESRD because of left ventricular hypertrophy and fluid overload, was not, as originally described by Buckberg et al,26,27 subtracted from the aortic pressure. Furthermore, the critical DPTI/SPTI ratio is likely to be much greater in subjects with ESRD because the prevalence of left ventricular hypertrophy and coronary artery disease is very high.1 Their hearts are furthermore characterized by extensive interstitial fibrosis.1 Coronary artery stenoses decrease the actual perfusion pressure in the coronary vascular bed,20 whereas left ventricular hypertrophy and interstitial fibrosis increase the resistance to flow while decreasing the vasodilatory reserve.1,2 Thus, the supply of oxygen might be diminished in patients with ESRD, whereas the demand is generally increased because of left ventricular hypertrophy.2 Any change in supply/demand ratio might thus be of clinical significance in this population. We chose to use a dialysate temperature of 36°C to prevent the occurrence of hypotensive episodes.3 It should be realized that the use of dialysate of a higher temperature might compromise coronary perfusion because of hypotensive episodes and temperature-induced vasodilation. Because we did not observe hypotensive episodes in the present study, we were not able to quantify their effect on the balance between coronary oxygen supply and demand. Questions can be raised concerning the reliability of finger-pressure measurements, the reconstruction of the aortic pressure wave, and Modelflow calculations of SV, CO, and SVR. Therefore, these items are discussed here. We had some difficulty finding subjects in whom finger pressure could adequately be measured during HD. This could be caused by the effect of previous arteriovenous fistulas on the nonfistula arm, peripheral vascular disease, or vasoconstriction during HD. That a proper finger-pressure signal could not be obtained during HD in all patients might limit the clinical applicability of fingerpressure measurements in dialysis populations. We consider this technique mainly a noninvasive research tool. Furthermore, it cannot be excluded that a selection bias was introduced by performing the measurements only in selected subjects in whom a proper finger-pressure signal could be
824
BOS ET AL
Fig 3. The balance between oxygen supply and demand during isolated ultrafiltration and hemodialysis combined with ultrafiltration is represented by DPTI/SPTI ratio. Left ventricular oxygen supply is estimated from DPTI and oxygen demand from SPTI. Data are presented as mean ⴞ SD. Comparisons were made against baseline values (t ⴝ 0). *P F 0.05. **P F 0.01 in paired ttest.
obtained. In dialysis patients in whom finger pressures can be measured adequately, this finger pressure might underestimate more centrally measured BPs, as in the elderly and subjects with atherosclerotic disease.16,28 However, group average changes in these central pressures are tracked reliably with finger-pressure measurements, even in populations in which the absolute pressure level at the finger is less than more centrally measured pressures.14,16,28 To estimate the effect of HD and UF on
Fig 4. Group average changes in stroke volume at all measurement points during the experiment are closely related to changes in blood volume (r 2 ⴝ 0.97; P F 0.01).
cardiac supply and demand, we calculated DPTI and SPTI from aortic pressure waves reconstructed from finger-pressure measurements using a generalized waveform filter.21,22 On its way to the periphery, especially the systolic part of the pressure wave is amplified. Changes in heart rate cause further changes in pressure amplification,29 making calculations of SPTI in peripheral pressure waves less meaningful. Generalized waveform filters correct for such pulse-wave amplifications.22-25 Modelflow-derived calculations of changes in CO have been validated against changes in thermodilution CO during cardiac surgery and treatment in an intensive care unit (mean error, 2% ⫾ 8%).17,18 In this respect, the close correlation between changes in BV and SV, measured with two independent techniques, was a reassuring finding. A correlation between BV and SV has been predicted in computer models describing hemodynamic changes during HD30 and described in studies in which CO was determined before and after HD.8,31 Our continuous measurements of both BV and SV yield individual BV-SV curves. The individual correlations were strong, even though the slope of these FrankStarling–like curves differed between individuals. Differences in the slope might be attributed to differences in cardiac contractility. However,
CARDIAC EFFECTS OF HD
differences in venous compliance and heart rate response might also contribute to differences in the slope. Thus, the combined use of BV and SV measurements might provide a useful noninvasive tool to study cardiac physiology during HD. In conclusion, uncomplicated UF and HD induce large changes in hemodynamics. These changes cause a slight improvement in the balance between cardiac oxygen supply and demand. REFERENCES 1. Parfrey PS, Foley RN, Harnett JD: Organ and metabolic complications: Cardiac, in Jacobs C, Kjellstrand CM, Koch KM, Winchester JF (eds): Replacement of Renal Function by Dialysis. Dordrecht, The Netherlands, Kluwer, 1996, pp 990-1002 2. London GM, Marchais SJ, Guerin AP, Metivier F, Pannier B: Cardiac hypertrophy and arterial alterations in end-stage renal disease: Hemodynamic factors. Kidney Int 43:S42-S49, 1993 (suppl 41) 3. Van Kuijk WHM, Luik AJ, de Leeuw PW, van Hooff JP, Nieman FHM, Habets HML, Leunissen KML: Vascular reactivity during hemodialysis and isolated ultrafiltration: Thermal influences. Nephrol Dial Transplant 10:1852-1858, 1995 4. Van Kuijk WHM, Leunissen KML: Hemodynamic stability during different forms of dialysis therapy: A pathogenetic analysis. Blood Purif 14:405-420, 1996 5. Daugirdas JT: Dialysis hypotension: A hemodynamic analysis. Kidney Int 39:233-246, 1991 6. Movilli E, Camerini C, Viola BF, Bossini N, Strada A, Maiorca R: Blood volume changes during three different profiles of dialysate sodium variation with similar intradialytic sodium balances in chronic hemodialyzed patients. Am J Kidney Dis 30:58-63, 1997 7. Levy FL, Grayburn PA, Foulks CJ, Brickner ME, Henrich WL: Improved left ventricular contractility with cool temperature hemodialysis. Kidney Int 41:961-965, 1992 8. Chaignon M, Chen WT, Tarazi RC, Nakamoto S, Bravo EL: Blood pressure response to hemodialysis. Hypertension 3:333-339, 1981 9. Mehta BR, Fischer D, Ahmad M, Dubose TD: Effects of acetate and bicarbonate hemodialysis on cardiac function in chronic dialysis patients. Kidney Int 24:782-787, 1983 10. Nixon JV, Mitchell JH, McPhaul JJ, Henrich WL: Effect of hemodialysis on left ventricular function: Dissociation of changes in filling volume and in contractile states. J Clin Invest 71:377-384, 1983 11. Pedersen T, Rasmussen K, Cleemann-Rasmussen K: Effect of hemodialysis on cardiac performance and transmural myocardial perfusion. Clin Nephrol 19:31-36, 1983 12. Wolff J, Pedersen T, Rossen M, Cleemann-Rasmussen K: Effects of acetate and bicarbonate dialysis on cardiac performance, transmural myocardial perfusion and acidbase balance. Int J Artif Organs 9:105-110, 1986
825
13. Steuer RR, Harris DH, Conis JM: A new optical technique for monitoring hematocrit and circulating blood volume: Its application in renal dialysis. Dial Transplant 22:260-265, 1993 14. Imholz BPM, Wieling W, van Montfrans GA, Wesseling KH: Fifteen years of experience with finger arterial pressure monitoring: Assessment of the technology. Cardiovasc Res 38:605-616, 1998 15. Imholz BPM, Langewouters GJ, van Montfrans GA, Parati G, van Goudoever J, Wesseling KH, Wieling W, Mancia G: Feasibility of ambulatory, continuous, 24-hour finger pressure recording. Hypertension 21:65-73, 1993 16. Bos WJW, Imholz BPM, van Goudoever J, Wesseling KH, van Montfrans GA: The reliability of noninvasive continuous finger blood pressure measurement in patients with both hypertension and vascular disease. Am J Hypertens 5:529-535, 1995 17. Wesseling KH, Jansen JRC, Settels JR, Schreuder JJ: Computation of aortic flow from pressure in humans using a nonlinear, three element model. J Appl Physiol 74:25662573, 1993 18. Jellema WT, Wesseling KH, Groeneveld ABJ, Stoutenbeek CP, Thijs LG, van Lieshout JJ: Continuous cardiac output in septic shock by simulating a model of the aortic input impedance. Anesthesiology 90:1317-1328, 1999 19. Sarnoff SJ, Braunwald E, Welch GH, Case RB, Stainsby WN, Macruz R: Hemodynamic determinants of oxygen consumption of the heart with special reference to the tension-time index. Am J Physiol 192:148-156, 1958 20. Hoffman JIE, Buckberg GD: The myocardial supply: demand ratio—A critical review. Am J Cardiol 41:327-332, 1978 21. Bos WJW, Zietse R, Wesseling KH, Westerhof N: Effect of arteriovenous fistulas on cardiac oxygen supply and demand. Kidney Int 55:2049-2053, 1999 22. Karamanoglu M, Feneley MP: On-line synthesis of the human ascending aortic pressure pulse from the finger pulse. Hypertension 30:1416-1424, 1997 23. Bos WJW, van Goudoever J, van Montfrans GA, van den Meiracker AH, Wesseling KH: Reconstruction of brachial artery pressure from noninvasive finger pressure measurements. Circulation 94:1870-1875, 1996 24. Gizdulich P, Prentza A, Wesseling KH: Models of brachial to finger pulse wave distortion and pressure decrements. Cardiovasc Res 33:698-705, 1997 25. Chen CH, Nevo E, Fetics B, Pak PH, Yin FCP, Maugham WL, Kass DA: Estimation of central aortic pressure waveform by mathematical transformation of radial tonometry pressure: Validation of generalized transfer function. Circulation 95:1827-1836, 1997 26. Buckberg GD, Fixler DE, Archie JP, Hoffman JIE: Experimental subendocardial ischemia in dogs with normal coronary arteries. Circ Res 30:67-81, 1972 27. Buckberg GD, Fixler DE, Archie JP, Henney RP, Hoffman JIE: Variable effects of heart rate on phasic and regional left ventricular muscle blood flow in anaesthetized dogs. Cardiovasc Res 9:1-11, 1975 28. Rongen GA, Bos WJW, Lenders JWM, van Montfrans GA, van Lier HJJ, van Goudoever J, Wesseling KH,
826
Thien T: Comparison of intrabrachial and finger blood pressure in healthy elderly volunteers. Am J Hypertens 8:237-248, 1995 29. Bos WJW, van den Meiracker AH, Wesseling KH, Schalekamp MADH: Effect of regional and systemic changes in vasomotor tone on finger pressure amplification. Hypertension 26:315-320, 1995
BOS ET AL
30. Ursino M, Innocenti M: Mathematical investigation of some physiological factors involved in hemodialysis hypotension. Artif Organs 21:891-902, 1997 31. Kouw PM, Kooman JP, Cheriex EC, Olthof CG, de Vries PMJM, Leunissen KML: Assessment of postdialysis dry weight: A comparison of techniques. J Am Soc Nephrol 4:98-104, 1993