The Haemodynamic Response to Submaximal Exercise During Isovolaemic Haemodialysis

Nephrol Dial Transplant (2004) 19: 1528–1532 DOI: 10.1093/ndt/gfh237

Original Article

The haemodynamic response to submaximal exercise during isovolaemic haemodialysis

Anindya Banerjee, Chiew H. Kong and Ken Farrington Downloaded from https://academic.oup.com/ndt/article/19/6/1528/1857546 by guest on 29 September 2021

Abstract Keywords: cardiac output; exercise; haemodialysis; Introduction. Exercise during haemodialysis has haemoglobin; relative blood volume; serum albumin potential benefits but may compromise cardiovascular stability. We studied its acute effects on relative blood volume (RBV) and other haemodynamic parameters. Methods. Two groups of 10 patients were exercised Introduction submaximally using a stationary cycle during isovolaemic dialysis whilst RBV was monitored continu- Exercise during haemodialysis has many potential ously. In study 1, patients exercised for two 10 min benefits, including increased peak oxygen uptake [1], periods separated by 10 min rest. Cardiac output (CO), improved self-reported physical functioning [1] and peripheral vascular resistance (PVR), central blood enhanced urea clearances [2]. It has been suggested that volume (CBV) and stroke volume were measured using the cardiovascular response to exercise during haemo- ultrasound dilution immediately before and after each dialysis is superimposed onto that of the dialysis exercise session. In study 2, haemoglobin, serum total process. The response has been shown to remain protein and albumin levels were measured before and ‘adequately stable’ during the first 2 h of treatment, immediately after the exercise session and at the nadir though thereafter cardiovascular decompensation of the RBV trace. may occur which could preclude further exercise [3]. Results. RBV fell immediately on exercise initiation, In previous pilot studies, we observed a rapid fall in the maximum reduction being 2.0±1.1% (after 5.9± relative blood volume (RBV) immediately after 1.4 min of exercise 1: P<0.001) and 2.0±1.2% (after commencement of exercise and before the onset of 4.7±2.3 min of exercise 2: P<0.001). CO increased ultrafiltration (UF). This might have an important significantly after both periods of exercise (4.5±0.96 effect on the capacity of patients to tolerate exercise and 5.1±1.1 to 7.2±2.1 and 7.9±2.4 l/min, P<0.001 during haemodialysis. in both). Stroke volume increased significantly and Blood volume monitors (BVMs) infer RBV from PVR fell significantly during exercise. CBV increased ultrasonically measured changes in the density of in absolute terms but fell as a proportion of CO. Mean blood, which are determined mainly by changes in haemoglobin level at the RBV nadir was significantly haematocrit (Hct) and plasma protein concentration. higher than baseline (12.3±1.8 vs 11.8±1.7 g/dl: A constant mass of red cells and plasma proteins, P<0.05: mean change 4.4±2.3%), as was mean total and uniform mixing throughout the circulation are protein concentration (66.0±6.9 vs 62.0±8.1 g/l: assumed. A number of mechanisms may account for P ¼ 0.001: mean change 6.8±5.9%) and mean serum a fall in RBV during exercise on isovolaemic haemo- albumin concentration (36.0±3.9 vs 34.1±3.9 g/l: dialysis. Exercise-induced redistribution of blood P<0.001: mean change 5.8±3.5%). volume from the central to the microcirculation Conclusion The haemodynamic response to exercise would increase whole body Hct (interpreted by the during haemodialysis is comparable with that in normal BVM as a fall in RBV), since the prevailing Hct is less individuals. The rapid reduction in RBV on exercise in the microcirculation [4]. If this were the sole effect occurs in spite of a significant increase in CO, mainly of exercise, plasma protein concentrations would as a consequence of fluid shifts from the microvascu- not change. Any increased fluid translocation from lature to the interstitium. the microcirculation to the interstitium would cause comparable increases in whole body Hct, plasma total protein and albumin concentrations. Any loss of albumin-rich fluid from the microcirculation to the Correspondence and offprint requests to: Dr K. Farrington, Renal Unit, Lister Hospital, Stevenage, Herts SG1 4AB, UK. Email: interstitium would abrogate the rise in serum albumin [email protected] concentration.

Nephrol Dial Transplant Vol. 19 No. 6 ß ERA–EDTA 2004; all rights reserved Haemodynamic responses to exercise on haemodialysis 1529 We devised this study to quantify the fall in RBV measurements undertaken were pulse, blood pressure and which occurs immediately after commencement of RBV, which were monitored continuously, and cardiac exercise on isovolaemic haemodialysis and to investi- output (CO), peripheral vascular resistance (PVR), central gate its possible causes. blood volume (CBV) and stroke volume, which were measured using ultrasound dilution (Transonic Systems, Ithaca, NY) immediately before and after each exercise Materials and methods session (four times in all for each patient) [5]. In brief, this technique requires the administration of a 30 ml bolus of Patients warm (body temperature) normal saline over 4–7 s into the venous line. After passage of the indicator saline through Two groups of 10 patients (exercise protocols 1 and 2) were the heart and lungs, its passage through the arterial line recruited from our unit. Informed consent was obtained is detected by an ultrasound sensor. CO is calculated from for the study, which was approved by the North Herts the concentration curve obtained with a reproducibility of NHS Trust Ethics Committee. All had been treated by 4.3±3.8%. CBV and PVR can be calculated from the CO Downloaded from https://academic.oup.com/ndt/article/19/6/1528/1857546 by guest on 29 September 2021 thrice-weekly maintenance haemodialysis for at least 3 measurement, with similar reproducibility, using previously months. Dialysis access was by arterio-venous fistula. None described formulae and patient data [5]. had recirculation demonstrable on ultrasound dilution (Transonics). None had current angina or any other clinically Experiment 2. Exercise protocol 2 was employed. Pulse, apparent condition likely to impair their capacity to perform blood pressure and RBV were monitored continuously. In stationary cycling. addition, blood was drawn from the arterial port immediately before and immediately after the exercise session. Dialysis technique Additional samples were taken at the nadir of the RBV trace. All samples were analysed for blood haemoglobin level (LP Dialysis was carried out using a Fresenius Medical Care 20 miniphotometer, Dr Bruno Lange GmbH Medical 4008H machine incorporating a BVM, blood temperature Division), total serum protein concentration (modified monitor and a blood pressure and pulse monitor. The BVM Biuret reaction, Dade Behring Clinical Chemistry System) obtained estimates of relative blood volume by measuring the and serum albumin (by an immunochemical reaction, Dade time taken for ultrasound pulses to travel through a precisely Behring Clinical Chemistry System). defined cuvette located in the arterial blood line. All patients were treated exclusively using high-flux synthetic membranes, predominantly polysulfone. Dialysers were reused with peracetic acid (Renalin, Minntec Inc., USA) as the main Rationale for investigatory protocols processing agent. Bicarbonate was used as buffer. Ultrapure Two identical periods of exercise were performed in protocol water was used for all dialysis-related processes. Stringent 1 in order to allow determination of the reproducibility of bacteriological standards were maintained. Dialysis was the haemodynamic changes and to allow comparison of prescribed and monitored using a two-pool kinetic model to haemodynamic changes in the exercise period with those ensure a Kt/V of 1.2. This was a composite of Kt/V (renal) and in an intervening control period with the patient in the Kt/V (dialysis). Mean dialysis time was 180 min (range same posture. The haemodynamic effects of exercise and 140–225 min). The midweek dialysis session was chosen for its effect on blood haemoglobin, serum total protein and the studies. Blood flow rates and dialysate flow rates were kept albumin levels were assessed on different dialysis sessions to constant over the study period. Dialysis fluid temperature was avoid the dilutional effects on the concentration measure- maintained at 36 C throughout the procedures. No UF was ments of saline boluses used to measure cardiac output. performed during the study period. These concentration measurements were performed during a single exercise session and recovery period to limit the Experimental protocols volume of blood required. Exercise protocols. Exercise was performed using a stationary cycle (Trailblazer) in the semi-recumbent position. This was placed in front of the patient so pedalling did not disturb the Statistics dialysis procedure or the posture of the patient, which did Analysis of variance for repeated measures was used to test not change throughout the dialysis session. Exercise load was the null hypothesis for differences between repeated values set to achieve a target heart rate at least 20% greater than the of parameters. In experiment 1, the values of RBV used were pre-dialysis baseline pulse rate. Exercise was initiated 15 min those pertaining to before both exercise periods, at the nadir after the start of the dialysis session to allow for stabilization of the RBV trace during both exercise periods and after both of the RBV trace and to minimize any posture-related exercise periods. The values of all other parameters used in changes to RBV. In protocol 1, 10 patients exercised for two experiment 1 were those pertaining to before and after each 10 min periods separated by 10 min rest. In protocol 2, 10 exercise period. In experiment 2, the values of all parameters patients exercised once for 10 min. In each protocol, after used were those pertaining to before the exercise period, at completion of the exercise session(s), the prescribed UF the nadir of the RBV trace during exercise, and at the end volume was removed in a linear fashion during the remainder of exercise. In the event of rejection of the null hypothesis, of the dialysis session. post hoc tests were performed by the Bonferroni multiple comparison test to determine the significance of differences Experiment 1. Exercise protocol 1 was employed during between mean values of individual parameters at these which haemodynamic parameters were measured. The different time points. 1530 A. Banerjee, C. H. Kong and K. Farrington Results 2.5±1.0% (P<0.001) between the end of exercise 1 and the onset of exercise 2. Study groups CO increased significantly after both periods of exercise [4.5±0.96 and 5.1±1.1 to 7.2±2.1 and 7.9± The clinical characteristics of the two exercise protocol 2.4 l/min, P<0.001 in both (Figure 2)]. Stroke volume groups are shown in Table 1. There were no significant increased significantly with exercise from 54.3±19.6 differences between the groups with respect to gender, and 58.6±15.3 to 88.3±37.7 and 92.1±34.9 ml mean weight, dialysis vintage, residual renal function, (P<0.001 in both). CBV increased with exercise interdialytic weight gain, haemoglobin levels, dose of (0.69±0.24 and 0.76±0.26 to 0.93±0.43 and 0.96± erythropoietin and pre-dialysis potassium levels. One 0.40 l: P<0.05 in both). When expressed as a percent- patient in protocol group 2 was diabetic. One patient age of cardiac output, CBV fell significantly during in protocol group 1 had a previous well-recovered exercise (14.5±2.1 and 14.4±2.7% to 11.7±2.2 and cerebrovascular event, and one in protocol group 2 had 11.5±2.6% post, P<0.001 in both). PVR fell sig- undergone previous coronary artery bypass grafting. nificantly (22.4±4.5 and 19.9±4.1 to 14.4±4.0 and Downloaded from https://academic.oup.com/ndt/article/19/6/1528/1857546 by guest on 29 September 2021 Ten patients, five in each group, took antihypertensive 13.5±4.4 mmHg/ml/min, P<0.001 in both). medication. Five patients in study group 1 and one in study group 2 took either angiotensin-converting enzyme inhibitors or angiotensin receptor blockers. Experiment 2 Three patients in study group 2 took low-dose b- Haemodynamic stability was maintained throughout blockers in contrast to none in study group 1. the period of exercise. RBV fell rapidly and significantly in the first 5 min to a nadir of 3.0±0.8% below baseline Experiment 1 (P<0.001) before partial recovery by the end of the Haemodynamic stability was maintained throughout both periods of exercise. The pulse rate rose from Fall in RBV with exercise 87±11 to 112±18 beats per min during the first period of exercise and from 87±14 to 112±18 beats per mine 105 during the second period (P<0.001 in both cases). 103 Systolic blood pressure also increased during both periods of exercise [129±17 to 142±9 mmHg (P<0.05) 101 and 123±14 to 141±11 mmHg (P<0.001)]. Diastolic blood pressure did not change significantly during 99 RBV % either period of exercise (79±10 to 83±10 and 77±11 97 to 83±10 mmHg). RBV fell rapidly and significantly immediately after 95 the start of exercise (Figure 1). The maximum 01020 30 40 50 reduction of RBV was 2.0±1.1% at 5.9±1.4 min into exercise 1 (P<0.001 compared with onset of Time in minutes exercise 1) and 2.0±1.2% at 4.7±2.3 min into Fig. 1. Rapid fall in relative blood volume at onset of exercise exercise 2 (P<0.001 compared with onset of exercise (black arrows). There is subsequent plateauing and some recovery 2). Some recovery occurred during the remainder of towards end of the exercise period (grey arrows). This trend was the exercise such that the mean reduction in RBV was observed in all 10 patients. 1.3±1.3% (P<0.05) by the end of exercise 1 and 1.0±1.5% by the end of exercise 2 (P ¼ NS). RBV 12 recovery continued in the resting phase, increasing by p < 0.001 p < 0.001 11 10 9 Table 1. Clinical characteristics of patients in study groups 8 7 Study Study P-value group 1 group 2 6 5 Age (years) 37±12 47±24 NS 4 Sex ratio (M:F) 4:6 6:4 NS 3 Cardiac Ouput (l/min) Weight (kg) 61±17 64±20 NS 2 Dialysis vintage (months) 59±50 45±56 NS 1 Residual urea clearance (ml/min) 1.1±1.4 1.2±1.5 NS Interdialytic weight gain (kg) 1.7±0.8 2.0±1.2 NS 0 Haemoglobin (g/dl) 11.6±1.7 12.4±0.7 NS Pre-1 Post-1 Pre-2 Post-2 Pre-dialysis potassium (mmol/l) 5.3±0.5 5.1±0.8 NS Exercise Session No. of antihypertensive agents 0.6±0.7 0.8±1.2 NS Fig. 2. Changes in cardiac out put in 10 patients before and after Values are mean±SD. two periods of exercise during experiment 1. Haemodynamic responses to exercise on haemodialysis 1531 Table 2. The reduction in relative blood volume (RBV), and changes in haemoglobin levels, serum total protein level and serum albumin levels before, during (at nadir of RBV trace) and after exercise (exercise protocol 2)

Start of exercise RBV nadir P-value End of exercise P-value

RBV fall (%) 3.0±0.8 <0.001 2.2±1.5 <0.001 Haemoglobin (g/dl) 11.8±1.7 12.3±1.8 <0.05 12.0±1.8 NS Total protein (g/l) 62.0±8.1 66.0±6.9 <0.001 63.6±5.2 <0.01 Albumin (g/l) 34.1±3.9 36.0±3.9 <0.001 35.5±3.3 0.001

Significance values (shown in comparison with baseline values) were determined by analysis of variance for repeated measures with post hoc testing by the Bonferroni method. Values are mean±SD. exercise session to a value 2.2±1.5% less than baseline contraction and release of red cells has been demon- (P<0.001). Haemoglobin levels increased during the strated following maximal exercise, but has little effect Downloaded from https://academic.oup.com/ndt/article/19/6/1528/1857546 by guest on 29 September 2021 period of exercise (Table 2). The mean haemoglobin on red cell volume measurements [10]. During recovery, level at the nadir of the RBV trace was significantly there is increased protein movement from extravascular higher than baseline (12.3±1.8 vs 11.8±1.7 g/dl: to intravascular compartments [11], such that plasma P<0.05: mean change 4.4±2.3%). The mean level at volume expansion of up to 10% can be seen within 24 h the end of the exercise period remained higher than of intense exercise [12]. baseline (12.0±1.8 g/l) but not significantly so. Serum In haemodialysis patients, cardiac output increased total protein levels also rose significantly during and systemic vascular resistance decreased during the exercise. The mean total protein level was significantly first 2 h of exercise [3]. However, during the following higher at the nadir of the RBV trace than at baseline hour, there was a substantial change in these param- (66.0±6.9 vs 62.0±8.1 g/l: P ¼ 0.001: mean change eters, CO decreasing and systemic vascular resistance 6.8±5.9%). The mean level at the end of exercise increasing, presumably due to a UF-induced reduction remained significantly higher than at baseline in blood volume, though this was not directly assessed. (63.6±5.2 g/l: P ¼ 0.01). Serum albumin likewise A small (1%) exercise-induced reduction in RBV has increased during exercise. The mean serum albumin been observed in the absence of UF, which reversed at level at the RBV nadir was 36.0±3.9 vs 34.1±3.9 g/l at the end of exercise [13]. UF, in the absence of exercise, baseline (P<0.001: mean change 5.8±3.5%). The has been shown to reduce CO in haemodialysis patients mean level at the end of dialysis remained significantly without overt cardiac disease [14]. higher than at baseline (35.5±3.3 g/l: P ¼ 0.001). There Our findings, interpreted in the light of these was no correlation between changes in haemoglobin previous studies, suggest that during submaximal levels, serum albumin concentrations and serum total exercise, though CBV is maintained or increased in protein concentrations. absolute terms with respect to resting values, a greater proportion of the increased CO is shifted from the central circulation to the microcirculation. Within the microcirculation, there is increased hydraulic perme- Discussion ability of the capillary bed resulting in fluid shifts into the interstitium. This fluid shift causes a reduced This study has demonstrated that a significant fall in plasma volume reflected in the observed 6.8% increase RBV occurs immediately after the onset of submaximal in total protein concentration. The changes in serum exercise during isovolaemic haemodialysis. The mean albumin (5.8%) and haemoglobin (4.4%) were similar magnitude of the change was 2–3% of blood volume. but not identical. While it is not possible to exclude The effect may therefore contribute to haemodynamic some loss of albumin-rich fluid from the microcir- instability on haemodialysis especially in the presence culation to the interstitium, or an effect on RBV of concurrent UF. At the same time, there was a of haematocrit redistribution, we have not found any significant increase in CO, and a significant reduction in direct evidence for either of these processes. The PVR. In absolute terms, there was a non-significant changes we have described are maximal after 5 min increase in CBV, but CBV as a proportion of CO fell. of submaximal exercise and tend to diminish as exercise Haemoglobin, total plasma protein and albumin continues. This presumably relates to an increase in the concentration rose abruptly and significantly. circulatory refilling from the interstitium as exercise Exercise in normal individuals, in addition to continues. increasing cardiac output and reducing peripheral resis- There are a number of potential methodological tance, also induces a reduction in total blood volume drawbacks to this study. Changes in haemodynamic and plasma volume up to 8.4 and 15.3, respectively, parameters and blood constituent concentrations were when the exercise is maximal [6]. This reflects fluid not measured simultaneously because the methodology movement from vascular space to the interstitium of used to measure haemodynamic parameters required exercising muscles [7]. The changes are rapid, apparent repeated small volume saline injection which may have by 1 min, maximal by 5 min [8], dependent on posture interfered with concentration measurements. Adoption and on the type of exercise carried out [9]. Splenic of the supine position results in a fall in haematocrit 1532 A. Banerjee, C. H. Kong and K. Farrington and serum albumin levels maximal in 15 min in nor- 2. Kong CH, Tattersall JE, Greenwood RN, Farrington K. The mal individuals, but continuing for at least 30 min effect of exercise during haemodialysis on solute removal. in haemodialysis patients [15]. The phenomenon has Nephrol Dial Transplant 1999; 14: 2927–2931 3. Moore GE, Painter PL, Brinker KR, Stray-Gundersen J, Mitchell been attributed to mobilization of fluid from the JH. Cardiovascular response to submaximal stationary cycling interstitium into the circulation. We commenced our during hemodialysis. Am J Kidney Dis 1998; 31: 631–637 measurements after 15 min of haemodialysis in the 4. Albrecht KH, Gaehtgens P, Pries A, Heuser M. The Fahraeus semi-recumbent position. This was to allow stabiliza- effect in narrow capillaries (i.d. 3.3 to 11.0 micron). Microvasc tion of the RBV trace following the assumption of this Res 1979; 18: 33–47 posture at the commencement of the dialysis session. 5. Krivitski N, Depner T. Cardiac output and central blood volume during hemodialysis: methodology. Adv Renal Replace The same posture was maintained during the subse- Ther 1999; 6: 225–232 quent exercise period and during the intervening rest 6. Van Beaumont W, Greenleaf JE, Juhos L. Disproportional period in experiment 1. We are confident, therefore, changes in hematocrit, plasma volume, and proteins during that the haemodynamic and concentration changes exercise and bed rest. J Appl Physiol 1972; 33: 55–61 observed during the study were related to the periods of 7. Ploutz-Snyder LL, Convertino VA, Dudley GA. Resistance Downloaded from https://academic.oup.com/ndt/article/19/6/1528/1857546 by guest on 29 September 2021 exercise rather than to the initiation of haemodialysis exercise-induced fluid shifts: change in active muscle size and or to changes in posture. plasma volume. Am J Physiol 1995; 269: R536–R543 8. Pivarnik JM, Goetting MP, Senay LC Jr. The effects of body We have found that submaximal exercise induces position and exercise on plasma volume dynamics. Eur J Appl haemodynamic changes similar to those seen in normal Physiol Occup Physiol 1986; 55: 450–456 subjects. There is a significant increase in cardiac 9. Gore CJ, Scroop GC, Marker JD, Catcheside PG. Plasma output in spite of a fall in plasma volume, which results volume, osmolarity, total protein and electrolytes during from loss of fluid from the microvasculature to the treadmill running and cycle ergometer exercise. Eur J Appl interstitium. These findings have an important bearing Physiol Occup Physiol 1992; 65: 302–310 10. Stewart IB, Warburton DE, Hodges AN, Lyster DM, on the tolerability of exercise during haemodialysis. McKenzie DC. Cardiovascular and splenic responses to Short periods of intermittent submaximal exercise are exercise in humans. J Appl Physiol 2003; 94: 1619–1626 well tolerated, inducing a haemodynamic response 11. Harrison MH, Edwards RJ, Leitch DR. Effect of exercise and similar to that in normal patients. These findings thermal stress on plasma volume. J Appl Physiol 1975; 39: complement previous work, which has suggested that 925–931 exercise during [3] and immediately after [16] or 12. Gillen CM, Lee R, Mack GW et al. Plasma volume expansion in humans after a single intense exercise protocol. J Appl between [17–19] haemodialysis sessions is well tolerated Physiol 1991; 71: 1914–1920 and has many significant benefits. However, it should 13. Rosales LM, Schneditz D, Chmielnicki H, Shaw K, Levin NW. be noted that our patients were young compared with Exercise and extracorporeal blood cooling during hemodialysis. the dialysis population in general and had well ASAIO J 1998; 44: M574–M578 corrected anaemia, that the exercise was brief and 14. Vandenbogaerde JF, Vanholder RC, Everaert JA et al. Cardiac submaximal, and that it took place in the absence of output-changes during hemodialysis with ultrafiltration. Clin UF. Our findings may not extrapolate to older patients, Nephrol 1988; 29: 88–92 15. Inagaki H, Kuroda M, Watanabe S, Hamazaki T. Changes in subjected to more prolonged or intense exercise in the major blood components after adopting the supine position context of high UF rates. during haemodialysis. Nephrol Dial Transplant 2001; 16: 798–802 16. Ross DL, Grabeau GM, Smith S et al. Efficacy of exercise for end-stage renal disease patients immediately following high- Conflict of interest statement. None declared. efficiency hemodialysis: a pilot study. Am J Nephrol 1989; 9: 376–383 17. Goldberg AP, Geltman EM, Hagberg JM et al. Therapeutic benefits of exercise training for hemodialysis patients. Kidney Int Suppl 1983; 16: S303–S309 References 18. Koufaki P, Mercer TH, Naish PF. Effects of exercise training on aerobic and functional capacity of end-stage renal disease 1. Painter P, Moore G, Carlson L et al. Effects of exercise patients. Clin Physiol Funct Imaging 2002; 22: 115–124 training plus normalization of hematocrit on exercise capacity 19. Deligiannis A, Kouidi E, Tassoulas E et al. Cardiac effects of and health-related quality of life. Am J Kidney Dis 2002; 39: exercise rehabilitation in hemodialysis patients. Int J Cardiol 257–265 1999; 70: 253–266

Received for publication: 5.05.03 Accepted in revised form: 25.02.04