Reduction of hypotensive side effects during online-haemodiafiltration and low temperature haemodialysis

Transcription

1 Nephrol Dial Transplant (2003) 18: DOI: /ndt/gfg206 Original Article Reduction of hypotensive side effects during online-haemodiafiltration and low temperature haemodialysis Johannes Donauer 1, Christoph Schweiger 1, Brigitta Rumberger 1, Bernd Krumme 2 and Joachim Bo hler 2 1 Department ofnephrology, University Hospital Freiburg and 2 Deutsche Klinik fu r Diagnostik, Division ofnephrology, Wiesbaden, Germany Abstract Background. This study compares the effect of onlinehaemodiafiltration (o-hdf, post-dilution mode) with conventional haemodialysis (HD) and temperaturecontrolled HD (Temp-HD) on the haemodynamic stability ofhypotension-prone patients. Methods. Seventeen patients with a history offrequent hypotensive episodes during dialysis sessions were studied, each patient serving as his or her own control. The first 25 HD treatments in comparison with 25 o-hdf sessions were evaluated using identical dialysate temperature. In the second part ofthe study, o-hdf (n ¼ 25) was compared with Temp-HD (n ¼ 25). In the latter method, the temperature ofthe dialysate was adjusted to result in identical energy transfer rates to those in the corresponding o-hdf. The number ofhypotensive episodes, blood temperature and blood volume regulation were assessed. Results. Symptomatic hypotension was much more frequent during HD (40%) than during o-hdf (4%) (P < 0.001). During o-hdf, an enhanced energy loss within the extracorporeal system occurred (o-hdf, 16.6 ± 4.0 W; HD, 5.4 ± 5.1 W; P < ), despite identical temperature settings for dialysate and substitution fluid. As a result, the blood returning to the patient was cooler during o-hdf than during HD (o-hdf 35 ± 0.2 C vs HD 36.5 ± 0.3 C; P < ). In o-hdf, even in the patients circulation, the mean blood temperature was lower (o-hdf 36.7 ± 0.2 C vs HD 36.9 ± 0.3 C; P < ) and blood volume was significantly more reduced (o-hdf, 91.8 ± 3.1%; HD, 94.0 ± 3.2%; P < 0.05). Energy transfer rates and blood temperature did not differ significantly between o-hdf and Temp-HD. The rate ofhypotensive episodes was low and not different between o-hdf Correspondence and offprint requests to: Dr Johannes Donauer, Medizin IV, Universitätsklinik Freiburg, Hugstetter Strasse 55, D Freiburg, Germany. donauer@med1.ukl.uni-freiburg.de (4%) and Temp-HD (4%). Neither was there any significant difference in blood volume reduction. Conclusions. O-HDF showed a significant reduction ofhypotensive episodes compared with HD. Surprisingly, o-hdf resulted in cooling ofthe blood via enhanced thermal energy losses within the extracorporeal system, despite use ofreplacement fluid prepared from pre-warmed dialysate. The incidence of symptomatic hypotension was reduced to that ofo- HDF by using cooler Temp-HD. Thus, unexpected blood cooling appears to be the main blood pressurestabilizing factor in o-hdf. Keywords: blood volume monitoring; dialysis-induced hypotension; dialysate temperature; haemodiafiltration; haemodialysis; relative blood volume Introduction Intradialytic hypotension occurs in up to 20% ofall dialysis treatment sessions. Despite progress in dialysis technology, this complication continues to be the most frequent adverse event during haemodialysis (HD) treatments. The influence ofblood temperature on hypotension during HD treatments has been studied extensively. It is well documented that HD leads to a substantial increase in body temperature [1]. Several factors have been discussed in the literature as contributing to hypotension during HD treatments: blood membrane interaction induces cytokine release, which may act as a pyrogen increasing core temperature; a non-sterile dialysate could enhance this effect [2]. The core temperature ofmany dialysis patients is lower than normal, and inappropriate high dialysate temperature (e.g. 37 C) may cause a transfer of heat from dialysate to the blood [1]. In addition, reduced heat losses via the skin due to peripheral vasoconstriction in response to ß 2003 European Renal Association European Dialysis and Transplant Association

2 Temperature effects of online-haemodiafiltration 1617 Table 1. Patient and study characteristics ultrafiltration may increase body temperature [3]. As a consequence ofthis heat accumulation, a loss ofvascular tone may take place, resulting in hypotension [4]. It has been reported that patients with cardiovascular instability during conventional HD treatments may experience fewer side effects if treatments based on convective solute and water transport are used [5]. Several mechanisms could be responsible for this phenomenon: differences in convective and diffusive solute clearance [5], differences in sodium removal [6,7], different vascular reactivity [8] or simply cooling ofthe blood [9,10]. Recent data favour the latter explanation, because conventional haemofiltration (HF) as well as haemodiafiltration (HDF) using replacement fluids at room temperature were shown to stabilize blood pressure [10]. However, it is still unclear, how online-hdf (o-hdf) may stabilize blood pressure, since in this treatment, in contrast to conventional HDF, substitution fluid is prepared from pre-warmed dialysate and thus one would not expect thermal energy losses to occur. To clarify the effects of o-hdf on patient stability, blood temperature and blood volume changes, this study compares o-hdf with conventional HD and with temperature-controlled HD (Temp-HD). Subjects and methods Patients Seventeen chronic HD patients ofthe Freiburg University Hospital dialysis unit were included in the study after giving informed consent. For patients characteristics, see Table 1. Only patients known to be hypotension prone defined as three or more dialysis-associated hypotensive episodes during the last 6 weeks were included. Patient monitoring Patients weight loss during the treatment was assessed by use ofa bed scale. Blood pressure and heart rate were monitored at 30 min intervals by automated cuff measurements (Dinamap, Critikon, Jacksonville, FL), or more frequently if indicated. Symptomatic hypotension was defined as a reduction ofthe systolic blood pressure below 100 mmhg associated with reactions ofthe patient prompting nurse intervention, such as placing the patient in Trendelenburg s position, reducing the ultrafiltration rate or infusing intravenous fluids. An asymptomatic fall ofblood pressure is not believed to be ofclinical relevance and therefore was not evaluated. Experimental protocol Patients eligible to participate were studied according to two different protocols: in the first part of the study (study A), 11 patients underwent three o-hdf treatments over 1 week followed by three HD sessions during the next week. In the second part (study B), nine patients were subjected to three o-hdf treatments, followed by three Temp-HD sessions. Three patients participated in both studies. The observation Treatment modalities O-HDF/ HD No. ofpatients 11 9 No. oftreatment pairs observed Treatment pairs excluded according to exclusion criteria: Weight mismatch 4 1 Difference in dialysate 3 O-HDF/ Temp-HD temperature setting Technical problems 1 1 No. oftreatment pairs evaluated Age (years) a 66; ; Sex (M/F) 11/0 6/3 No. ofhypotensive episodes during 5.0 ± ± 1.17 the last 6 weeks b prior to study Non-renal diagnoses (no. of patients/all patients) Diabetes mellitus 4/11 2/9 Congestive heart failure, 1/11 2/9 compensated Coronary heart disease 5/11 3/9 Hypertension 9/11 7/9 Renal diagnoses Polycystic kidney disease 0/11 1/9 Chronic glomerulonephritis 1/11 4/9 Amyloidosis 2/11 1/9 Analgesic nephropathy 0/11 1/9 Tumour nephrectomy 1/11 0/9 Diabetic nephropathy 4/11 2/9 Renal failure of unknown origin 3/11 0/9 No. ofantihypertensive drugs per patient No. ofpatients treated with antihypertensive drugs AT antagonists 0/11 2/9 Angiotensin-converting 5/11 5/9 enzyme inhibitors -Blocking agents 6/11 5/9 Calcium channel blocking drugs 5/11 3/9 -Blocking agents 4/11 1/9 Diuretics 8/11 9/9 Direct vasodilatating agents 0/11 1/9 Erythropoetin 10/11 9/9 a Median, 25% and 75% interval. b Mean ± SD. period of1 week for each treatment modality was chosen in order to compare treatments with nearly identical weight losses. For each patient, pairs oftreatments were selected which had to fulfil the following matching criteria: (i) the difference of weight losses between matched treatments had to be < 500 g, since it is well known that weight loss during dialysis is a major determinant of side effects and temperature changes [4]; (ii) dry weight had to be reached during all sessions; (iii) treatment time had to be identical in corresponding treatments; and (iv) dialysate temperature was the same for corresponding treatments, except for Temp-HD, where dialysate temperature was changed in order to cool the blood. Patients with significant shunt recirculation [>10%, measured using the blood temperature monitor (BTM)] were excluded, as recirculation may affect blood temperature and blood volume measurements [11]. Each patient served as his or her own control. Based on the above-mentioned criteria, 25 matched treatment pairs were identified for final analysis in both parts of the study.

3 1618 J. Donauer et al. Table 2. Treatment characteristics Study A Study B o-hdf HD o-hdf Temp-HD Mean blood flow (ml/min) ± ± ± ± 27.2 Mean ultrafiltration volume (l) 2.6 ± ± ± ± 0.8 Replacement fluid prescribed per treatment (ml/min) Dialysate flow (ml/min) Mean dialysate temperature 36.8 ± ± ± ± 0.4 Dialysis prescription All treatments were performed using a volume-controlled dialysis machine with optional online-hdf mode (4008H, Fresenius Medical Care, Bad Homburg, Germany). For all treatments, polysulfone hollow-fibre dialysers, F50 or F60 for HD or HF80 for o-hdf (Fresenius Medical Care), were used. Dialysate and substituate composition was 35 mmol/l bicarbonate, mmol/l sodium, 1.75 mmol/l calcium, 2 4 mmol/l potassium, 0.5 mmol/l magnesium, mmol/ l chloride, 1 g/l glucose. Dialysate sodium concentration of an individual patient was not changed and was kept constant throughout the study period. All dialysate was sterile filtered. O-HDF produces sterile replacement fluid from the pre-warmed dialysate. The replacement fluid was infused in post-dilution mode with a substitution rate of50 ml/min. Blood flow rate and treatment time were prescribed individually for each patient aiming at a Kt/V of at least 1.3 per treatment. Dry weight was determined by clinical judgement, chest X-ray or ultrasound ofvena cava collapse, and was not changed during the study period. Detailed data ofthe dialysis prescription are given in Table 2. Dialysate, replacement fluid and room temperature The dialysate temperature for o-hdf and HD was chosen according to the patients previous documented temperature setting (range C). In o-hdf treatments, i.v. replacement fluid and dialysate were taken from the same fluid pool after warming to the operator-set dialysate temperature. The substitution fluid underwent a further step ofsterile filtration to remove all possible bacteria or pyrogens before it was infused into the venous drip chamber. During study A comparing o-hdf and HD, dialysate temperature was set to the same level for all corresponding treatments. In study B (o-hdf vs Temp-HD), dialysate temperature during o-hdf sessions was again set as in study A. During Temp-HD treatments, dialysate temperature was lowered using the E-control option (see below) ofthe BTM (Fresenius Medical Care) which achieved the same blood temperature in the venous line as seen in the preceding o-hdf session. Room temperature was kept constant at 23 C during the whole study period. Blood temperature measurement and control The BTM [12] measures blood temperature in the arterial and venous line and calculates the energy transfer (ET) rate between the extracorporeal system and the blood at 15 s intervals. Based on the temperature ofthe arterial line, the BTM calculates the patient s core temperature by correcting for fistula and cardiopulmonary recirculation. Recirculation is measured by inducing a temporary change in dialysate temperature. The ensuing temperature change ofthe blood passing through the dialyser is detected by the venous temperature sensor head ofthe BTM and later by the sensor head at the arterial line. Recirculation can be calculated from the ratio ofthe amplitudes ofthe temperature transients in the arterial and the venous blood line, respectively [13]. The BTM calculates extracorporeal arterio-venous temperature gradients (T av ) and ET rates. ET is the amount of thermal energy that is transferred from the extracorporeal system to the patient, or vice versa. ET (in kj/h) is calculated using the following formula: c p Qb (T art T ven ), where c ¼ the specific thermal capacity (3.64 kj/kg); Qb ¼ extracorporeal blood flow; and p ¼ density ofthe blood (1052 kg/m 3 )]. For a detailed description, see Rosales et al. [4]. In Temp-HD treatments, dialysate temperature was modified to achieve the same ET rate as seen before in the corresponding o-hdf treatments. ET values are given in Watts (1 W ¼ 3.6 kj/h). Monitoring of relative blood volume Relative blood volume (RBV) was measured with a commercially available blood volume monitor (BVM, Fresenius Medical Care). This monitor measures the transient time of short ultrasonic pulses, which are sent through the blood column in the arterial line. On the basis ofthese measurements, total protein concentration ofthe blood and RBV are calculated consecutively. The changes in RBV during dialysis treatment are assessed every 10 s, starting from 100% at the beginning (for details, see Schneditz et al. [14]). Data collection and graphical depiction ofthe RBV and BTM curves were done with graphical software provided by Fresenius Medical Care. Data analysis and statistics Statistical analysis was performed using Sigmastat statistical software (SPSS Science, Chicago, IL). For comparison of blood temperature and blood volume between treatment modalities within study A or B, the paired Student s t-test or Wilcoxon signed rank test were used as appropriate. Comparison oftemperature and blood volume between HD (study A) and Temp-HD (study B) was performed using unpaired Student s t-test. Comparison ofthe frequencies of dialysis-associated side effects between o-hdf and HD

4 Temperature effects of online-haemodiafiltration 1619 (study A) or o-hdf and Temp- HD (study B) was done using McNemars test. For comparison ofthe rates of hypotensive episodes in HD (study A) and Temp-HD (study B), the z-test was used. Results Data from 17 patients and 120 treatments were obtained in this study. In the 6 week evaluation period prior to the study, the frequency of hypotensive episodes was similar in both groups (study A, 5.0 ± 1.26; study B, 5.1 ± 1.17). Each patient served as his or her own control. Ten pairs oftreatments were excluded from analysis because they did not meet the predefined matching criteria (see above). Finally, 100 treatments representing 25 matched pairs in each part ofthe study (A/B) were analysed. Frequency of hypotensive episodes during o-hdf, HD and Temp-HD In study A, there was a significantly lower rate of hypotensive episodes during o-hdf (4%) compared with HD (40%, P < 0.001). Blood pressure declined significantly during HD, but not during o-hdf treatments (see Table 3). In study B, no difference in the frequency of hypotensive episodes was seen between o-hdf and Temp-HD treatments (4% in both, NS), and decline ofsystolic blood pressure during treatments, although present, did not reach statistical significance (see Figure 1 and Table 3). The rate of symptomatic hypotensive side effects was significantly lower in Temp-HD (study B) than in conventional HD (study A) (P < 0.001, see Table 3). Energy transfer rate The ET rate reached significant negative values in study A, during both o-hdf (P < 0.001) and HD (P < 0.001). Loss ofthermal energy was significantly more pronounced during o-hdf than HD (P < ) Table 3. Thermal and haemodynamic parameters Treatment modality Fig. 1. Number oftreatments with (solid bars) and without (striped bars) symptomatic hypotension. (A) Study A: during o-hdf treatments, significantly fewer hypotensive episodes occurred compared with HD (P < 0.01). (B) Study B: using cooler dialysate during Temp-HD and avoiding most ofthe warming ofthe patients resulted in a similar incidence ofhypotensive episodes compared with o-hdf. Treatments with hypotension/all T art 0 min ( C) T av ( C) Min BV (%) ET (W) Systolic blood pressure (mmhg) treatments (%) Beginning End Study A o-hdf 1/25 (4%) 36.4 ± ± 0.3 c 91.8 ± ± 4.0 e ± ± 17.9 HD 10/25 a (40%) 36.5 ± ± 0.4 c,b 94.0 ± 3.2 d 5.4 ± 5.1 b,e ± ± 17.5 c Study B o-hdf 1/25 (4%) 36.5 ± ± 0.3 c 92.9 ± ± 1.9 e ± ± 25.8 Temp-HD 1/25 f (4%) 36.5 ± ± 0.1 c 93.5 ± ± 4.2 e ± ± 27.8 a P < 0.001, vs o-hdf; b P < , vs o-hdf; c P < , vs beginning; d P < 0.005, vs o-hdf; e P < 0.001, vs beginning; f P < 0.001, vs HD. Values are mean ± SD; T art 0 min, blood temperature ofthe patient at the beginning ofthe treatment session, measured in the arterial blood line; T av, mean difference in blood temperatures before and after passage through the extracorporeal system; Min BV, minimal RBV value reached during treatment sessions.

5 1620 J. Donauer et al. increase during o-hdf, 0.21 ± 0.2 C; or Temp-HD, 0.21 ± 0.16 C; NS; mean arterial blood temperature during o-hdf, 36.7 ± 0.45 C; Temp-HD, ± 0.35 C). Fig. 2. Blood temperature in the venous tube during HD or o-hdf. O-HDF treatments resulted in cooler venous blood returning to the patient despite identical dialysate temperature setting. (Table 3). In study B, there was also a significant loss of thermal energy in both treatment modalities (o-hdf and Temp-HD, P < 0.001); as expected, this did not differ between o-hdf and Temp-HD (Table 3). Blood temperature in the venous outlet of the dialyser In study A, the temperature ofthe dialysate was between 36.5 and 37 C during all treatments. Temperature settings were adopted from previous sessions. For each patient, HD and the corresponding o-hdf treatments were performed at the same dialysate temperature. Despite identical temperature settings for the dialysate, o-hdf treatments resulted in significantly lower blood temperatures in the venous line compared with HD (mean venous blood temperature: o-hdf, ± 0.23 C; HD, ± 0.31 C; P < ; see Figure 2). In study B, blood temperature in the venous line did not differ significantly between Temp-HD and o-hdf (Temp-HD ± 0.4 C vs o-hdf ± 0.43 C, NS). Blood temperature in the arterial line of the dialyser Initial arterial blood temperatures were not significantly different between treatments in both studies (study A: o-hdf, ± 0.39 C; HD, ± 0.47 C; NS; study B: o-hdf, 36.5 ± 0.51 C; Temp- HD, 36.5 ± 0.39 C; NS). During all forms of treatment, arterial blood temperature increased, although to a different degree. In study A, the rise in temperature was more pronounced during HD than during o-hdf (mean arterial temperature increase during o-hdf, 0.26 ± 0.25 C; during HD, 0.39 ± 0.26 C; P < ). Mean arterial blood temperature reached ± 0.23 C in o-hdf but ± 0.29 C in HD (P < ). In contrast, in study B, as expected, the increase in arterial blood temperature was the same in Temp-HD and o-hdf (mean arterial temperature Relative blood volume (RBV) measurements Mean RBV values ofall treatments were significantly higher during HD than during o-hdf (study A: HD ± 3.37%; o-hdf, 95.2 ± 2.86%; P < 0.05). In study B, there was no significant difference in mean RBV between o-hdf and Temp-HD (o-hdf, 95.5 ± 3.79%; HD, ± 3.63%; NS). Minimal RBV values (see Table 3) also were higher in conventional HD compared with o-hdf (study A, P < 0.005), whereas Temp-HD and o-hdf in study B showed no significant difference in minimal RBV. During hypotensive episodes, RBV values are influenced by artefacts e.g. increased fistula recirculation or fluid infusion, hence minimal RBVs in Table 3 were compared only for asymptomatic treatments. Discussion Side effects during HD are important determinants of the quality ofchronic dialysis care. They have been linked to a reduction in dialysis efficacy, may influence long-term morbidity and mortality, and substantial resources ofdialysis units have to be allocated to handling hypotension-prone patients. HF and HDF are dialysis procedures using convective solute transport. It has been reported that these treatments provide superior cardiovascular stability compared with standard HD, which uses mainly diffusive transport of solutes [5]. However, the main stabilizing factor in isolated HF seems to be the cooling ofthe blood [15] with use ofrelatively cold substitution fluid. It is also well known that use of cool dialysate during HD can reduce the number of hypotensive episodes [1,4]. In contrast to HF and HD, there is little information about heat balance and occurrence ofsymptomatic hypotension in HDF procedures [15], and no data are available concerning o-hdf. Comparing HDF vs HD, one study over 6 months oftreatment noted a decrease ofhypotensive episodes from 18% (HD) to 14% (HDF) [16]. A second study did not find significant differences between HDF and HD regarding hypotensive episodes [17]. However, the frequency of hypotensive episodes was low, and, therefore, considering the sample size differences between treatment modalities, could hardly be detected. A recently published investigation [10] followed changes in blood pressure and temperature during HDF and HD. A more stable blood pressure was seen in HDF using replacement fluids at room temperature. As this study did not focus on hypotension-prone patients, it did not address the question ofwhether cooling by HDF can prevent symptomatic hypotension. Therefore, our

6 Temperature effects of online-haemodiafiltration 1621 study included only patients known to be hypotension prone, while other treatment variables (e.g. intradialytic weight loss, dialysate and infusate composition, dialysate and room temperature) were kept constant. Focusing on patients at risk, this selected group had an unusually high rate ofhypotensive episodes (40%) with conventional HD at a dialysate temperature set at C. The rate ofhypotensive episodes was significantly lowered with o-hdf (4%) despite the same dialysate temperature settings. In addition, we could show that systolic blood pressure remained stable during o-hdf treatments, whereas it dropped significantly during HD. Since in o-hdf the substitution fluid is prepared from the same pre-warmed fluid that is also used as dialysate, one would expect identical temperature profiles and ET rates during HD and o-hdf. Interestingly, o-hdf reduced blood temperature significantly and increased ET. Loss ofthermal energy in the extracorporeal circuit could cause this effect, as in o-hdf blood temperature in the venous tube was >1 C lower than in the arterial tube. We suggest that substitution ofcold fluid may explain this finding. During o-hdf, the substitution fluid initially has the same temperature as the dialysate. Due to the postdilution mode, it subsequently cools down while passing through an additional tube and sterile filter before it reaches the venous drip chamber. The amount ofthermal energy loss during o-hdf in this study reaches 16.6 W, which is 22% ofthe estimated resting energy expenditure ofan adult person. It should be emphasized, however, that patients were not cooled to temperatures lower than their initial blood temperature, instead o-hdf merely prevented some ofthe patients temperature increase typically seen with conventional HD. To prove that the observed cooling ofthe venous blood during o-hdf was responsible for better haemodynamic tolerance, the second part ofthe study compared o-hdf treatments with cooler HD treatments. In order to imitate the temperature course of the previous o-hdf, we used the E-control option of the BTM, which automatically reduced the dialysate temperature as needed to ensure identical ET in both treatments. As blood flow was unchanged, this resulted in identical venous and arterial blood temperature profiles during Temp-HD and o-hdf, respectively. Under these conditions, the rate ofhypotensive episodes during Temp-HD was significantly lower than during conventional HD treatments in study A and was no longer different compared with o-hdf sessions. Maggiore et al. showed in a recently published, randomized, crossover trial that the incidence of dialysis-induced symptomatic hypotension during HD could be reduced from 50 to 25% in hypotension-prone patients using isothermic HD [18]. The reduction ofside effects in our study seems to be more pronounced (40 to 4%), but this may be explained by a different definition ofthe term symptomatic hypotension and different patient inclusion criteria. In our study, body temperature rises up 0.26 C in HDF and Temp-HD. Despite this increase, a significant reduction ofhypotension takes place. We conclude that patients tolerate a slight increase ofbody temperature well. However, isothermic dialysis may be the more physiological treatment procedure and therefore should be recommended. The study design (short-term single centre interventional study) may limit the interpretation, but we believe that the highly significant reduction ofside effects with o-hdf and Temp-HD sufficiently validates the data presented here. It should be mentioned that comparison between HD treatment in study A and Temp-HD treatment in study B has to be made with caution, since different patients were followed in both studies. However, our findings prove sufficiently that mainly temperature effects are responsible for the reduced incidence ofhypotensive episodes during o- HDF. Moreover, they also indicate that increases in body temperature should be avoided during HD in order to prevent unnecessary symptomatic hypotensive episodes. As shown for example in Figure 3, the patient s blood temperature at the beginning is clearly lower than the venous temperature, but later continues to rise above venous temperature levels. Thus, warming the blood within the extracorporeal circuit only partly explains the increase ofpatients core temperature at the beginning ofthe treatment. In accordance with other studies [18], we found that the body temperature (measured in the arterial line) continues to rise despite cooler blood returning to the patient (measured in the venous line). Thus, it is obvious that endogenous heat accumulation also contributes to the temperature increase during HD. As shown by Rosales et al. [4], this is most probably due to peripheral vasoconstriction in response to ultrafiltration. Overall, even during conventional HD, the blood is cooled in the extracorporeal circuit as indicated by ET balance. Blood volume reduction correlates with incidence of hypotensive episodes in HD treatments [19]. In this study, however, RBV was significantly lower during treatments with better haemodynamic stability (o-hdf and Temp-HD). Compared with conventional HD, these treatments provided cooler core temperatures. Supposedly, cooling ofthe blood allows for better Fig. 3. Typical temperature curve ofcorresponding HD and o-hdf treatments in a stable dialysis patient. Dialysate temperature was set at 37 C. For both treatment regimes, an increase ofarterial blood temperature could be observed. During HD, the increase in arterial blood temperature was much higher compared with o-hdf.

7 1622 J. Donauer et al. peripheral vasoconstriction ofblood vessels [20] which on one hand stabilizes blood pressure, but on the other hand may interfere with vascular refilling and fluid shifts from tissues to the blood compartment [6]. Comparing warm and cool HD, a similar inverse relationship oftemperature and blood volume has been observed [6]. Therefore, in contrast to HD, RBV measurements appear to be less important during o-hdf and cool HD in order to avoid hypotensive episodes. In summary, the blood pressure-stabilizing effect of o-hdf is due to blood cooling despite the administration ofreplacement fluids which are prepared from warm dialysate. In this case, cooling may be due to heat loss from longer extracorporeal fluid lines, as o-hdf was performed in post-dilution mode. HD has the same low rate of side effects as o-hdf provided that blood temperature measurements and use ofappropriately cooler dialysate prevent warming ofthe patient. Conflict of interest statement. J.B. has repeatedly served as a consultant for Fresenius Medical care. J.D., C.S., B.R. and B.K. declare no conflict ofinterest. References 1. Fine A, Penner B. The protective effect of cool dialysate is dependent on patients predialysis temperature. Am J Kidney Dis 1996; 28: Pereira BJ, Sundaram S, Barrett TW et al. Transfer of cytokine-inducing bacterial products across hemodialyzer membranes in the presence ofplasma or whole blood. Clin Nephrol 1996; 46: Schneditz D, Martin K, Kramer M, Kenner T, Skrabal F. Effect of controlled extracorporeal blood cooling on ultrafiltration-induced blood volume changes during hemodialysis. J Am Soc Nephrol 1997; 8: Rosales LM, Schneditz D, Morris AT, Rahmati S, Levin NW. Isothermic hemodialysis and ultrafiltration. Am J Kidney Dis 2000; 36: Whele B, Asaba H, Castenfors J et al. Hemodynamic changes during sequential ultrafiltration and dialysis. Kidney Int 1979; 15: Gotch FA, Sargent JA. An unnecessarily complex method to achieve hypotonic sodium removal and controlled ultrafiltration. Blood Purif 1983; 1: Pedrini LA, Ponti R, Faranna G, Cozzi G, Locatelli F. Sodium modeling in hemodiafiltration. Kidney Int 2001; 40: van der Sande FM, Mulder AW, Hoorntje SJ et al. The hemodynamic effect of different ultrafiltration rates in patients with cardiac failure and patients without cardiac failure: comparison between isolated ultrafiltration and ultrafiltration with dialysis. Clin Nephrol 1998; 50: Maggiore Q, Pizzarelli F, Zoccali C et al. Effect of extracorporeal blood cooling on dialytic arterial hypotension. Proc Eur Dial Transplant Assoc 1981; 18: van der Sande FM, Kooman JP, Konings CJ, Leunissen KM. Thermal effects and blood pressure response during postdilution hemodiafiltration and hemodialysis: the effect of amount ofreplacement fluid and dialysate temperature. J Am Soc Nephrol 2001; 12: Wang E, Schneditz D, Kaufman AM, Levin NW. Sensitivity and specificity ofthe thermodilution technique in detection of access recirculation. Nephron 2000; 85: Kra mer M, Polaschegg HD. Control ofblood temperature and thermal energy balance during hemodialysis. Proc IEEE EMBS 1992; 14: Schneditz D, Wang E, Levin NW. Validation ofhaemodialysis recirculation and access blood flow measured by thermodilution. Nephrol Dial Transplant 1999; 14: Schneditz D, Pogglitsch H, Horina J, Binswanger U. A blood protein monitor for the continuous measurement of blood volume changes during hemodialysis. Kidney Int 1990; 38: Maggiore Q, Pizzarelli F, Dattolo P, Maggiore U, Cerrai T. Cardiovascular stability during haemodialysis, haemofiltration and haemodiafiltration. Nephrol Dial Transplant 2000; 15 [Suppl 1]: Movilli E, Camerini C, Zein H et al. A prospective comparison ofbicarbonate dialysis, hemodiafiltration, and acetate free biofiltration in the elderly. Kidney Int 1996; 27: Locatelli F, Mastrangelo F, Redaelli B et al. Effects of different membranes and dialysis technologies on patient treatment tolerance and nutrition parameters. Kidney Int 1996; 50: Maggiore Q, Pizzarelli F, Santoro A et al. Effects of control of thermal balance on vascular stability in hemodialysis patients: results ofthe European randomized clinical trial. Am J Kidney Dis 2002; 40: Donauer J, Kolblin D, Bek M, Krause A, Bohler J. Ultrafiltration profiling and measurement ofrelative blood volume as strategies to reduce hemodialysis-related side effects. Am J Kidney Dis 2000; 36: van Kuijk WH, Luik AJ, de Leeuw PW et al. Vascular reactivity during haemodialysis and isolated ultrafiltration: thermal influences. Nephrol Dial Transplant 1995; 10: Received for publication: Accepted in revised form: