Original Paper Received: July 5, 2010 Accepted: January 6, 2011 Published online: February 24, 2011
Blood Purif 2011;32:63–68 DOI: 10.1159/000324199
Individualised Dialysate Temperature Improves Intradialytic Haemodynamics and Abrogates Haemodialysis-Induced Myocardial Stunning, without Compromising Tolerability Helen J. Jefferies a James O. Burton a Christopher W. McIntyre a, b
Department of Renal Medicine, Royal Derby Hospital, and b School of Graduate Entry Medicine and Health, University of Nottingham, Nottingham, UK
Key Words Blood pressure ⴢ Cardiovascular disease ⴢ Echocardiography ⴢ Heart failure ⴢ Quality of life
Abstract Background/Aims: Haemodialysis-induced myocardial stunning is associated with intradialytic hypotension, increased likelihood of cardiovascular events and death. Dialysis at 35 ° C reduces stunning, but adverse thermal symptoms limit technique adoption. This study investigated whether individualised body temperature dialysis improves haemodynamic stability and abrogates stunning. Methods: Randomised crossover study of 11 patients compared LV regional wall motion abnormalities (RWMAs) at 37 ° C (HD37) and body temperature (‘individualised’, HDind). Regional systolic function was quantitatively assessed by echocardiography. Haemodynamics were assessed using continuous pulse wave analysis. Thermal symptoms were scored by questionnaire. Results: Mean predialysis body temperature was 36.0 8 0.1 ° C. Mean number of peak stress RWMAs per patient was lower with HDind (3.9 8 1.4 vs. 5.3 8 1.5, p = 0.03). Intradialytic systolic BP was higher during HDind versus HD37 (p ! 0.001). Individualised body temperature dialysis demonstrated symptomatic tolerability comparable to HD37. Conclusions: Individualised-temperature haemodialysis abrogates stunning, providing effective haemodynamic stabilisation at no additional therapy cost. Copyright © 2011 S. Karger AG, Basel
© 2011 S. Karger AG, Basel 0253–5068/11/0321–0063$38.00/0 Fax +41 61 306 12 34 E-Mail
[email protected] www.karger.com
Accessible online at: www.karger.com/bpu
Introduction
Cardiovascular mortality in haemodialysis (HD) patients is grossly elevated [1], and not fully explained by traditional risk factors. The commonest cause is sudden death, followed by heart failure [1]. Vascular calcification [2], microcirculatory dysfunction [3], impaired coronary flow reserve [4], and ineffective vasoregulation during HD and ultrafiltration [5] contribute to cardiovascular morbidity and predispose to demand myocardial ischaemia. The importance of risk factors relating to the dialysis procedure is becoming increasingly appreciated. In the non-renal population, repeated ischaemic episodes causing myocardial stunning contribute to development and progression of ischaemic myocardial injury resulting in heart failure [6, 7]. Standard HD induces acute global and segmental myocardial ischaemia [8, 9]. We have recently reported that repetitive HD-induced myocardial stunning over 1 year is associated with impaired haemodynamic response to dialysis, and observations of an absolute reduction in resting ejection fraction of 10%, and development of hibernated segments of myocardium, have generated the proposal that recurrent dialysis-induced ischaemic cardiac injury may contribute to the development of ischaemic heart failure in this population [10]. Cool dialysis is simple, cost-free, and reduces intradialytic hypotension and the severity of myocardial stunning [5]. However, adverse thermal symptoms and poor Dr. C. McIntyre Department of Renal Medicine Royal Derby Hospital, Uttoxeter Road Derby, DE22 3NE (UK) Tel. +44 1332 789 344, E-Mail chris.mcintyre @ nottingham.ac.uk
Downloaded by: 5.67.89.242 - 7/27/2013 9:12:27 PM
a
Table 1. Patient characteristics
Patient
Age years
1 2 3 4 5 6 7 8 9 10 11 Mean 8 SD
64 73 56 69 50 80 45 66 70 73 80 66812
12 15 24 24 12 50 90 276 7 48 8 52879
Subjects and Methods Subjects Eleven HD patients entered a randomised prospective crossover study from a single hospital centre. All patients were dialysed via native arteriovenous fistulae and were anuric. Patients’ characteristics are shown in table 1. Patients established on HD for 13 months were eligible. Those with pre-existing severe LV systolic dysfunction (NYHA IV) or cardiac transplant were excluded. Patients received 4 h HD thrice-weekly, with bicarbonate-based dialysate (Na+ 138 mmol/l, K+ 1.00 mmol/l, Ca2+ 1.25 mmol/l, Mg2+ 0.5 mmol/l, acetate 3.00 mmol/l, Cl– 107.5 mmol/l, HCO3– 32.0 mmol/l, glucose 1.0 g/l) and synthetic polyethersulphone dialysers. Dialysate flow was 500 ml/min. Blood pump speed varied (250 and 450 ml/min), although it remained unchanged between treatments for individual patients. Studies were conducted after a 2-day interdialytic period. The study was conducted in accordance with the Declaration of Helsinki and informed consent was obtained. Study Protocol Dry weight was determined by clinical examination. Subsequently, medications and dry weight remained unchanged. Predialysis core body temperature was determined by digital tympanic thermometer (First Temp쏐, Sherwood Davis & Geck, St.
64
Blood pressure medications
amyloidosis1 amyloidosis1 diabetic nephropathy tubulointerstitial nephritis renal tuberculosis adult polycystic kidney disease unknown diabetic nephropathy chronic lymphocytic leukaemia glomerulonephritis unknown
– – – – bisoprolol 2.5 mg nicorandil 20 mg – lisinopril 5 mg alternate days – – –
There was no evidence of amyloid cardiomyopathy.
tolerability of 35 ° C in some patients (consistent with other studies) presented important limitations to long-term use to reduce intradialytic hypotension and recurrent dialysis-based cardiac injury. Therefore, we aimed to test, using simple and widely available temperature monitoring, whether patient-individualised body temperature dialysate might confer benefits of intradialytic blood pressure stability and reduced myocardial stunning, whilst minimising adverse thermal symptoms.
Cause of end-stage renal failure
Blood Purif 2011;32:63–68
Louis, Mo., USA), prior to each dialysis treatment for 1 month. For each patient, the average of their pre-dialysis temperatures defined the individualised temperature. Patients were then randomised to group A or B. Group A initially underwent thrice-weekly standard HD with dialysate temperature at 37 ° C. Group B initially underwent thrice-weekly HD with dialysate temperature set to the individualised temperature of each individual patient. The third HD treatment at the allocated temperature was studied. Both groups then crossed over to the other arm of the study (hence each patient acted as their own control) and underwent a further three treatments at the alternative temperature; the third HD treatment at the alternative temperature was also studied.
Echocardiography Patients were assessed to evaluate the presence and extent of HD-induced regional wall motion abnormalities (RWMAs). Two-dimensional echocardiography was performed before dialysis (pre-HD), at 2 h, 4 h (peak stress), and 30 min into the recovery period (post-HD) using commercially available equipment (1.5– 3.6 MHz 3S probe, GE Medical Systems, Germany). Images were obtained by a single experienced operator blinded to dialysate temperature, with patients in the left lateral position. Apical twochamber and four-chamber views were digitally recorded for subsequent analysis (Echo-CMS; MEDIS, The Netherlands), as previously described [11]. Three consecutive heartbeats were analysed for each time point and endocardial borders were traced semiautomatically for each frame of the sequence. Anomalies were manually corrected by a single analyst blinded to dialysate temperature. Maximal displacement of the endocardial border from a centre point was measured over each of 100 chords around the LV wall, corrected for end-diastolic LV circumference, and expressed as percentage shortening fraction (%SF). Each apical view was divided into 5 segments, and %SF for the chords in each segment was averaged so that 10 regions of the LV were assessed at each time. New RWMAs were classified as segments that showed a decline in %SF 120% from baseline as previously defined. Ejection fraction was calculated using the biplane disc method. Regions that showed a functional decline of 120%
Jefferies /Burton /McIntyre
Downloaded by: 5.67.89.242 - 7/27/2013 9:12:27 PM
1
Dialysis vintage, months
25 Cardiac output (% change from baseline)
HD37 × HDind
× × × × × × × × × × × × × × × × ×
100 60
× × × × × × × × × × × × × × × × ×
ics. p ! 0.001 for comparison between HD37 and HDind with respect to all variables (by repeated-measures ANOVA).
Heart rate (% change from baseline)
60
120 Time (min)
180
× × ×
–25
× × ×
×
×
×
×
×
×
× × ×
240
20 15 10 5 × × × × × × × × 0× × × × × × × × × –5 –10 –15 –20 0 60 120 180 240 Time (min)
during HD compared to rest were classed as stunned segments, and patients with more than 2 affected segments at peak stress were classed as ‘stunners’ (as per previous study of HD patients [10]). Haemodynamic Measurements Haemodynamic response was measured non-invasively by continuous digital artery PWA using Finometer (TNO Instruments, Amsterdam, The Netherlands) described in detail elsewhere [12]. Pulse rate, blood pressure, stroke volume, cardiac output (CO) and total peripheral resistance were calculated as percentage change from baseline. Thermal Symptoms Patients completed a temperature symptom questionnaire [5] at each dialysate temperature. The minimum possible score of 6 signified no significant symptoms of cold; the maximum score of 24 indicated severe cold-related symptoms. Scores were set to give the most severe symptoms additional weighting. Biochemistry Blood samples were collected before and after each monitored session in lithium heparin and EDTA tubes, and biochemical analysis was performed on a multichannel autoanalyser. Cardiac troponin T analysis was performed using a third-generation electrochemiluminescence assay (Roche Diagnostics, Lewes, UK). Pre-dialysis blood tests were drawn immediately after the insertion of dialysis needles, and post-dialysis bloods were taken prior to the removal of dialysis needles. Sample Size Calculation and Statistical Analysis The power calculation was based on detecting a significant difference in the number and severity of regional wall motion (corrected for end-diastolic LV circumference and expressed as percentage shortening fraction). We have previously demonstrated a difference in this variable in 8 patients using an identical protocol when comparing standard and biofeedback dialysis [12]. Using an SD on change in fractional shortening from this study
Individualised Dialysate Temperature and Myocardial Stunning
×
–50 0
Fig. 1. Blood pressure and haemodynam-
0×
0
60
120 Time (min)
180
240
300 200
× × ×
100 × × ×
× × ×
× ×
×
× × ×
0× × –100 0
60
120 Time (min)
180
240
of 0.15, a sample size of 9 would appear to be sufficient to detect a difference of 20%, whilst a sample size of 12 would be sufficient to detect a difference of 17%. In the previous study, a difference of greater than 23% was observed. Results are presented as mean 8 SD, or the median and interquartile range, as appropriate. Biochemical data, RWMAs, ultrafiltration volume (UF) and temperature data were analysed using paired t test or Wilcoxon rank sum test (depending on normality). Blood pressure and haemodynamic data were analysed using repeated-measures ANOVA. An alpha error at 0.05 was judged as significant.
Results
Blood Pressure and Haemodynamic Data These data are displayed in figure 1. Mean intradialytic blood pressure was significantly higher with ‘individualised’ HD (HDind) compared to haemodialysis with dialysate at 37 ° C (HD37; p ! 0.001). At HDind, CO decreased and total peripheral resistance increased during HD, compared with HD37 during which CO and total peripheral resistance did not change. There was no significant difference in UF between the studies at 37 ° C and individualised temperature (1.9 8 0.4 vs. 1.7 8 0.4 litres, p = 0.54).
Echocardiographic Data There was no difference between mean pre-dialysis shortening fractions between temperatures (table 2). Compared to resting function, a total of 58 new RWMAs occurred at peak stress at HD37. At HDind, there were 43 Blood Purif 2011;32:63–68
65
Downloaded by: 5.67.89.242 - 7/27/2013 9:12:27 PM
140
Total peripheral resistance (% change from baseline)
BP (mm Hg)
180
Table 2. Echocardiographic data 10 ×
Parameter
6
*
×
HD37 Pre-dialysis Peak stress Post-dialysis HDind Pre-dialysis Peak stress Post-dialysis
×
× 4
2 0 Pre
HD37 × HDind 2h
Peak
Post
Fig. 2. Mean number of unaffected LV segments. In comparison
to resting function (all regions scored as unaffected), the number of segments unaffected by dialysis-induced RWMAs fell consecutively at 2 h and 4 h (peak stress), and returned towards normal post-dialysis. * p = 0.03.
new RWMAs occurring at peak stress (in 11 patients). The proportion of patients classed as ‘stunners’ was lower with individualised dialysis [11/11 (100%) at HD37, 9/11 (82%) at HDind]. The mean number of RWMAs per patient at peak stress was significantly higher during HD37 compared to HDind [5.3 8 1.5 vs. 3.9 8 1.4, p = 0.03; range for new RWMAs per patient was 3–8 (HD37) compared to 2–6 (HDind)]. The mean number of unaffected LV segments was higher with individualised dialysis (fig. 2). The mean pre-dialysis LVEF did not differ between temperatures, and did not change significantly during HD in either temperature group (table 2). Temperature Data and Thermal Symptoms Questionnaire The mean pre-dialysis tympanic temperature was 36.0 8 0.4 ° C. There was no difference in mean pre-dialysis temperatures between studies at different dialysate temperatures (36.1 8 0.6 ° C at HD37; 36.0 8 0.5 ° C at HDind, p = 0.82). There was a trend to increasing body temperature post-dialysis with dialysate temperature of 37 ° C, although it did not reach statistical significance (p = 0.09). There was no statistical difference between pre- and postdialysis body temperatures recorded during HDind, although the magnitude of change in body temperature decreased as the dialysate temperature was lowered (⌬ body temperature +0.6 8 0.3 ° C and +0.2 8 0.2 ° C at HD37 and HDind, respectively). The median temperature-related symptom score for HD37 was 6, which was the lowest possible score, indicating that no symptoms of cold were reported by any of the patients at this temperature. The temperature scores increased as the dialysate temperature decreased. The me
66
Blood Purif 2011;32:63–68
EF, %
SFmean, %
SFWMA, %
64.182.7 61.782.3 58.882.7
2.2480.21 1.7980.25 2.0580.23
2.7380.28 1.4980.13 1.5880.24
66.183.6 67.282.8 67.483.2
1.7380.25 1.8080.25 1.9080.33
2.7380.18 1.4480.10 1.6380.24
EF = Ejection fraction; SFmean = overall mean shortening fraction; SFWMA = mean shortening fraction of regional wall motion abnormalities.
dian score at HDind was 7 (6–10; mean 8 SD temperature scores were 6 8 0 at HD37, 7.4 8 2.4 at HDind). At HDind, 4 patients used blankets or extra clothing to keep warm at least once during the week, and 1 patient reported shivering. At this temperature, 2 patients reported feeling much better than normal overall, 1 felt slightly better, and the others felt no different to usual. Biochemical Data There were no differences in any biochemical measures across the temperature groups (summarised in table 3).
Discussion
This study demonstrates that cooling dialysate from the conventional temperature of 37 ° C to patient-individualised body temperature significantly reduces the number of LV RWMAs occurring during HD. This is consistent with findings from our previous study which compared standard and cool (35 ° C) dialysis. The operationally simple intervention of individualised HD is welltolerated, and is not associated with the adverse symptoms of cold which are experienced at 35 ° C [5]. A number of studies have investigated potential modifications to the dialysis process to ameliorate the severity of myocardial stunning and abrogate the longer-term adverse consequences of recurrent ischaemic injury. Biofeedback dialysis and cool temperature dialysis have demonstrated significant improvements in both intradialytic haemodynamic stability and abrogation of RWMAs, in concert with lower levels of markers of car
Jefferies /Burton /McIntyre
Downloaded by: 5.67.89.242 - 7/27/2013 9:12:27 PM
Number of unaffected regions
8
Table 3. Biochemical data
Parameter
HD37
Haemoglobin, g/dl Bicarbonate, mmol/l Na+, mmol/l Corrected Ca2+, mmol/l Phosphate, mmol/l Albumin, g/l CRP, mg/l cTnT, g/l
HDind
pre-HD
post-HD
pre-HD
post-HD
11.10 (10.7–11.3) 23 (20–25) 136 (136–138) 2.47 (2.37–2.54) 1.53 (1.16–1.73) 35 (33–38) 10 (5–34) 0.05 (0.01–0.07)
11.35 (10.95–11.7) 27 (25–28) 137 (135–137) 2.36 (2.29–2.46) 0.69 (0.60–0.87) 35 (32–38)
11.15 (10.75–12.8) 21.5 (20–24) 136 (135–138) 2.44 (2.4–2.48) 1.35 (1.28–1.80) 34 (33–39) 9.5 (3–17) 0.03 (0.01–0.07)
11.75 (11.05–12.95) 26 (25–27) 136 (135–137) 2.33 (2.28–2.37) 0.65 (0.60–0.78) 35 (33–39)
diac damage [12]. Cool dialysis maintains haemodynamic stability by increasing total peripheral resistance, in the face of limited cardiac ability to increase CO in response to falling intradialytic blood pressure, and improving autonomic reactivity by increasing baroreflex sensitivity [13]. This intervention has the potential for widespread use in dialysis units by virtue of its simplicity and low cost. However, a limiting factor preventing long-term implementation of cool dialysis at 35 ° C is patient discomfort due to adverse thermal symptoms [5]. This is the first study to attempt to optimise the intradialytic response of myocardial stunning and haemodynamics by individualising dialysate temperature. The number of LV RWMAs at peak stress was significantly lower with HDind compared to HD37, demonstrating that dialysate temperature reduction of 1 ° C on average was sufficient to produce detectable abrogation of myocardial stunning. Furthermore, the number of RWMAs increased as HD progressed through the treatment to peak stress, consistent with previous descriptions, and returned towards normal following termination of HD. The magnitude of reduction in RWMAs per patient was comparable but less marked than the previous study comparing dialysate temperatures of 37 and 35 ° C, which might be expected since the mean isothermic temperature was 1 ° C above the lower temperature achieved in that study. Systolic blood pressure was significantly higher during HDind compared to HD37, mediated predominantly through peripheral vasoconstriction as previously described [13]. This is consistent with previous findings that intradialytic fall in systolic blood pressure independently predicts myocardial stunning [10]. Reductions in both heart rate and stroke volume resulted in a reduction in CO. The consequent drop
in cardiac workload is consistent with a decreased ischaemic burden and is consistent with the observed reduction in myocardial stunning. An alternative interpretation of these haemodynamic data, which is that a primary increase peripheral resistance resulted in increased afterload and hence reduced CO, is less plausible since this might reasonably be associated with increased cardiac effort and demand ischaemia, and lead to worsening myocardial stunning. Importantly, since UF volume also independently predicts myocardial stunning [10], observed UF volumes were not different between temperatures. It is noteworthy that only 3/11 patients were taking anti-hypertensive medications, explained by the fact that minimisation of UF volume and control of fluid status are actively promoted as key means of controlling blood pressure in this dialysis unit. Aside from the accepted haemodynamic benefits of cool temperature HD [14], recent studies have reported improvements in nocturnal sleep patterns, patient-reported increased energy levels and quality of life [15, 16]. We found no symptomatic differences likely to be significant between HD37 and HDind, suggesting that long-term studies of individualised temperature HD would be well-tolerated. Our study has a number of limitations. Though adequately powered for its stated primary outcome (and statistical assumptions for observed effect were met), patient numbers were too low for any sub-group analyses. Echocardiography is repeatable and quantitative, but was the only technique used. Alternatives may provide additional information but are difficult to perform during dialysis. In any case, a load-independent assessment of regional myocardial function is needed to elucidate the complex mechanisms behind HD-induced RWMAs. We have previously demonstrated that our technique provides clini-
Individualised Dialysate Temperature and Myocardial Stunning
Blood Purif 2011;32:63–68
67
Downloaded by: 5.67.89.242 - 7/27/2013 9:12:27 PM
Data are median (IQR). There were no significant differences across any of the biochemical parameters between temperatures. CRP = C-reactive protein; cTnT = cardiac troponin T.
cally significant insights that relate to hard outcomes, and appear suitable to compare the impact of an intervention. Although several studies support the use of tympanic temperatures [17, 18], some advocate measuring blood line temperatures, but we did not have the necessary equipment. We did not study whether a more sophisticated approach to thermal energy management (blood temperature monitoring) [14] might have produced similar benefits. Such technology is not common to all dialysis machines, and remains unavailable in many dialysis units. Furthermore, insulated dialysis lines were not used since the intention was not to assess energy transfer per se, but to observe differential effects on haemodynamics with changes in programmed dialysate temperature; the stated aim to explore the clinical utility of an individualised temperature prescription that could be applied without specialised dialysis equipment was a deliberate and pragmatic choice to ensure the widest applicability of our findings. Potential confounders to overall thermal balance, including blood pump speed and line length, were controlled. Since this was a short-term study, body temperature was defined with reference to the average of temperatures recorded over the month preceding the intervention phase of the study; setting dialysate temperature on a per-treatment basis according to the body temperature measured before each dialysis is likely to offer better individualisation in the long term [19].
In conclusion, we have confirmed our previous findings that HD-induced acute cardiac dysfunction is abrogated by cool dialysis, and demonstrated that maintenance of intradialytic blood pressure by vasoconstriction, reduction in CO and associated cardiac workload can be achieved by dialysing patients at their body temperature. We have shown that the thermal symptom profile of individualised temperature HD is indistinguishable from standard dialysis at 37 ° C, and well-tolerated by patients. Further work is required to assess long-term effects with respect to myocardial stunning, and clinically important outcomes such as cardiovascular event rate and mortality. Ultimately, individualised temperature HD may potentially reduce progression to heart failure by amelioration of repetitive dialysis-induced cardiac injury and slowing development of ischaemic cardiomyopathy. This potentially cardioprotective approach is being tested in an appropriately powered long-term multi-centre randomised controlled trial in incident patients.
Acknowledgements This study was funded by a British Renal Society grant (No. 06-013) and has been adopted onto the UKCRN portfolio (study 5822). We are indebted to the staff and patients at the Royal Derby Hospital dialysis unit for their participation and contributions to this study.
References
68
8
9
10
11
12
Blood Purif 2011;32:63–68
tion after repeated episodes of exerciseinduced ischemia. Am J Physiol 1989; 256:H1462–H1471. Dasselaar JJ, Slart RH, Knip M, et al: Haemodialysis is associated with a pronounced fall in myocardial perfusion. Nephrol Dial Transplant 2009;24:604–610. McIntyre CW, Burton JO, Selby NM, et al: Hemodialysis-induced cardiac dysfunction is associated with an acute reduction in global and segmental myocardial blood flow. Clin J Am Soc Nephrol 2008; 3:19–26. Burton JO, Jefferies HJ, Selby NM, et al: Hemodialysis-induced cardiac injury: determinants and associated outcomes. Clin J Am Soc Nephrol 2009; 4:914–920. Bosch JG, Savalle LH, van Burken G, et al: Evaluation of a semiautomatic contour detection approach in sequences of short-axis twodimensional echocardiographic images. J Am Soc Echocardiogr 1995;8:810–821. Selby NM, Lambie SH, Camici PG, et al: Occurrence of regional left ventricular dysfunction in patients undergoing standard and biofeedback dialysis. Am J Kidney Dis 2006;47:830–841.
13 Chesterton LJ, Selby NM, Burton JO, et al: Cool dialysate reduces asymptomatic intradialytic hypotension and increases baroreflex variability. Hemodial Int 2009;13:189–196. 14 Selby NM, McIntyre CW: A systematic review of the clinical effects of reducing dialysate fluid temperature. Nephrol Dial Transplant 2006;21:1883–1898. 15 Ayoub A, Finlayson M: Effect of cool temperature dialysate on the quality and patients’ perception of haemodialysis. Nephrol Dial Transplant 2004;19:190–194. 16 Parker KP, Bailey JL, Rye DB, et al: Insomnia on dialysis nights: the beneficial effects of cool dialysate. J Nephrol 2008; 21(suppl 13):S71–S77. 17 Edge G, Morgan M: The genius infrared tympanic thermometer. An evaluation for clinical use. Anaesthesia 1993;48:604–607. 18 Robinson J, Charlton J, Seal R, et al: Oesophageal, rectal, axillary, tympanic and pulmonary artery temperatures during cardiac surgery. Can J Anaesth 1998;45:317–323. 19 Pizzarelli F: From cold dialysis to isothermic dialysis: a twenty-five year voyage. Nephrol Dial Transplant 2007;22:1007–1012.
Jefferies /Burton /McIntyre
Downloaded by: 5.67.89.242 - 7/27/2013 9:12:27 PM
1 Foley RN, Parfrey PS: Cardiovascular disease and mortality in ESRD. J Nephrol 1998; 11:239–245. 2 Sigrist MK, Taal MW, Bungay P, et al: Progressive vascular calcification over 2 years is associated with arterial stiffening and increased mortality in patients with stages 4 and 5 chronic kidney disease. Clin J Am Soc Nephrol 2007; 2:1241–1248. 3 Ichimaru K, Horie A: Microangiopathic changes of subepidermal capillaries in endstage renal failure. Nephron 1987;46:144–149. 4 Ragosta M, Samady H, Isaacs RB, et al: Coronary flow reserve abnormalities in patients with diabetes mellitus who have end-stage renal disease and normal epicardial coronary arteries. Am Heart J 2004;147:1017–1023. 5 Selby NM, Burton JO, Chesterton LJ, et al: Dialysis-induced regional left ventricular dysfunction is ameliorated by cooling the dialysate. Clin J Am Soc Nephrol 2006; 1: 1216– 1225. 6 Bolli R: Myocardial ‘stunning’ in man. Circulation 1992;86:1671–1691. 7 Homans DC, Laxson DD, Sublett E, et al: Cumulative deterioration of myocardial func-
