 |
European Dialysis and Transplant Nurses Association/European Renal Care Association |  |
2005/3 EDTNA/ERCA Journal Club Discussion SummaryDiscussion of "Merits and limitations of continuous blood volume monitoring during haemodialysis" Compiled by Elizabeth J Lindley 1 based on contributions from Richard Ward 2, Chris Pearson 3, Maurice Harrington 3, Frankie O’Kane 3, Bertrand Belot 4, Stanley Shaldon 5, Jean-Yves De Vos 6, Franta Lopot 7, Anahita Nikman 8, Jim Curtis 2, Danny Schneditz 9, Martin Gerrish 3, Thomas Roy 10, James Tattersall 3 and Judith Dasselaar 11. 1 Department of Renal Medicine, Leeds Teaching Hospitals NHS Trust, UK; 2USA;3UK; 4Switzerland; 5Monaco; 6Belgium; 7Czech Republic; 8Norway; 9Austria; 10Germany; 11The Netherlands. Abstract The discussion explored and expanded on the issues raised by Dasselaar et al in their review of the measurement of relative blood volume (RBV) changes during dialysis (NDT 2005). Dialysis machines incorporating blood volume monitoring and control are widely available in Europe. The use of continuous blood volume monitoring (CBVM) to help establish dry weight; problems with CBVM due to connection and use of single needle dialysis; the physiological processes that cause RBV changes during eating, exercise and posture changes; and the application of blood volume based biofeedback control were discussed by participants from ten countries. The ‘take-home’ messages from the discussion were that CBVM can assist in setting target weight, but must be used together with traditional measures and experience. Biofeedback control may help to achieve symptom-free dialysis, but staff should be prepared to monitor patients systematically for several weeks to obtain individualised settings. Users of CBVM should be aware of factors that can alter the central haematocrit leading to apparent changes in RBV. Practical guidelines should be developed to help staff interpret CBVM data effectively. Synopsis of the discussion paper The publication selected for discussion by the EDTNA/ERCA Journal Club in November 2005 was a review entitled “Measurement of relative blood volume changes during haemodialysis: merits and limitations” [1]. The authors work in the University Medical Centre in Groningen, The Netherlands and the corresponding author, Judith Dasselaar kindly agreed to take part in the discussion and respond to any queries raised. The authors note that most dialysis machine manufacturers have now incorporated devices that continuously monitor relative blood volume (RBV) changes during treatment. These devices are advocated as a tool to maintain the intravascular volume and avoid intradialytic hypotension (IDH), but evidence-based knowledge on using RBV changes to optimise the dialysis prescription is lacking. The paper describes the physiological mechanisms that compensate for reduction in blood volume during ultrafiltration and the factors that influence the compensatory response and may affect the blood volume reduction at which IDH will occur. Although it is not possible to identify a RBV reduction at which IDH will occur, biofeedback systems, developed to adjust ultrafiltration rate (UFR) and dialysate conductivity in response to RBV changes, have been shown to improve haemodynamic stability. They may achieve this by avoiding rapid fluctuations and prolonged drops in RBV. A number of problems related to the measurement of RBV changes are reviewed. The initial hydration status of the patient strongly influences the course of the RBV change during HD. Blood volume redistribution between the central and microcirculation affects the validity of the measured RBV change because of lower haematocrit (Hct) in capillaries and venules. Postural changes can lead to fluid shifts between the circulation and the interstitial tissue. Amongst the miscellaneous factors that influence RBV, the authors list splenic contraction (which would release red-cell rich blood into the circulation), food intake, exercise during HD, administration of IV fluids and changes in red cell volume due to changes in plasma osmolality. The authors conclude that staff using devices that monitor RBV should be aware of the ‘pitfalls’ and they advocate the development of practical guidelines on how to interpret and use the information generated by these devices. How widely used is continuous blood volume monitoring (CBVM)? One of the first contributions to the discussion was from Dick Ward in the US who explained that, whilst their new Gambro Phoenix machines featured the ‘Hemoscan’ technology for CBVM, it was not enabled. “Apparently”, said Dick, “Gambro’s US management do not think there would be sufficient interest in the device to warrant the cost of obtaining FDA approval. This is a ‘self-fulfilling prophecy’ - if the equipment is not available and people are not aware of it, of course there will be no demand.” The structure of the dialysis industry in the US makes it hard to introduce new technology. Reimbursement for dialysis by Medicare/Medicaid is marginal, so dialysis providers will not usually spend money on new equipment unless it can be shown to reduce operating expenses or, at least, be cost neutral. In turn, the providers of equipment see a limited market and decide not to market the new technology which means that there is no data to demonstrate benefit. Dick explained that the Fresenius CBVM system is approved in the US, but said he felt it was not used widely. To find out how widely used CBVM is in Europe, a short questionnaire was e-mailed to the EDTNA/ERCA active volunteers. 27 volunteers in 14 countries responded. All of their units had CBVM devices, either the integrated Hemoscan (Gambro-Hospal) or BVM (Fresenius), or the free-standing CritLine (Hemametrics). The majority of the volunteers’ units (18/27, 66%) use CBVM routinely. Eight of these (44%) use the ‘biofeedback’ option routinely. Five (19%) units have only used CBVM for research projects and four (15%) have never used their CBVM devices. Whilst this small survey is likely to over-represent units that interested in this technology, it does suggest relatively widespread use as the units using CBVM were located in all parts of Europe. In most of the units (83%) with machine-integrated devices, some or all of them were provided free of charge. Chris Pearson from the UK felt that blood volume monitors were being ‘given away’ in the hope of future income from specialised disposables which are sold at a premium to recoup investment in research and development and certification costs. The cost of lines was one reason given in the survey for not using CBVM. Maurice Harrington from the UK said that while the majority of the machines in his unit (Gambro AK200/AK200s) have CBVM and the staff do use it, the funding is not available to pay for the special blood lines for all treatments. As a result it is used occasionally as an aid to assessing dry weight, rather than during each session for preventing IDH. The situation is different where the dialysis unit has a special contract with the company. Frankie O'Kane, also from the UK, works in a 43 station unit which will soon have CBVM on all machines. Under their agreement with Gambro, CBVM is provided with no extra charge under their contract so they use it frequently as a guide to fluid status (alongside the usual parameters such as oedema, blood pressure and breathlessness). Dry weight determination Like Maurice and Frankie, the majority (89%) of the volunteers whose units were using CBVM routinely said they used it to help set the patients’ target weight. Bertrand Belot from Switzerland described how he had seen patients with a RBV that is constant, or even rising, during dialysis. “There were no other signs, but these patients were overhydrated and we had to reduce their target weight for several sessions, or in some cases, dialyse them every day for a week.” Bertrand felt that CBVM is not as widely used as it should be to detect fluid overload as many doctors don't accept it as a diagnostic tool. One reason for this may be the lack of published evidence linking CBVM to improved outcome. Stanley Shaldon from Monaco brought the club’s attention to the extraordinary results of the recent CLIMB study [2]. This was a prospective randomised controlled trial (RCT), which should give reliable results, but it appeared to show that patients who had an occasional session with CBVM were more likely to die or be hospitalised than those who didn’t. In fact patients in both the CBVM and control arm had lower mortality rates than expected (77% and 26% of the US mortality rate respectively). One of the authors of the CLIMB paper (Dr Ed Lowrie) criticised the analysis of the study on the RenalWeb site because the conclusion that doing nothing instead of using CBVM dramatically improves clinical outcome is illogic and unreasonable. He called for RCTs to be tested for external validity (how the findings fit into the real clinical world). Prof Shaldon does not advocate basing practice exclusively on RCTs – he pointed out papers suggesting that stifles new ideas and impedes progress that can be based on deduction from existing knowledge – but he wondered if Dr Lowrie’s call for validity tests would have been made if the CLIMB study had demonstrated a benefit of CBVM. None of the participants in the discussion thought that CBVM was harmful. Lizzi Lindley from the UK agreed with Bertrand that flat or rising RBV is a reliable indicator of fluid overload (provided fluid is being removed). Judith Dasselaar said that several publications she had reviewed reported that a patient is overhydrated when the RBV is constant or only mildly decreases. “But a patient still can be overhydrated when the RBV declines severely, simply because cardiovascular compliance is poor and the plasma refilling rate is very slow,” she said, “In using RBV for assessing dry weight we have to keep to whole patient in mind.” Jean-Yves De Vos from Belgium agreed. “BVM is a helping hand in determining a patient’s dry weight, but it is a tool that is never to be supposed to be used alone. Defining dry weight is still a real art, needing many different tools, experience and continuous adaptation.” Franta Lopot from the Czech Republic described the initial classification for CBVM sessions with a constant UFR in his unit: “Response 1: flat or rising RBV. This indicates overhydration, so we decrease the dry weight by 0.5 to 1.0 kg at the next dialysis. Response 2: RBV flat for some time then a linear decrease. This indicates that the patient is closer to optimal dry weight. It is acceptable in patients with some residual renal function as they could probably tolerate some more ultrafiltation but their urine output the next day would be significantly decreased. This might contribute to accelerated loss of residual kidney function. Response 3: a linear decrease in RBV from the start of dialysis. If the patient can tolerate this we usually keep the dry weight unchanged if the RBV reduction remains within 10-15%.” Franta added that the linear decrease in RBV type 2 and 3 responses has something to do with the applied UFR. “If you use pulsed UFR with rates of 2 to 3 l/hour during the pulses, you see slightly exponential decrease in RBV. With repeated UFR pulses this becomes a more and more steep linear pattern. Sandip Mitra and Paul Chamney have called this ‘linear divergence’ and suggested using it as a marker of being close to optimal dry weight [3].” Anahita Nikman from Norway described how the ICU nurses in her centre use CBVM. “In our ICU, nurses from the chronic dialysis ward provide dialysis using the Fresenius machine with CBVM (without the control method). We are very satisfied with this system and our doctors do believe in it, in fact they try to learn from us! In our experience, ICU patients needing haemodialysis are mostly overhydrated with high systolic blood pressure. CBVM can confirm this information and, with its help, we try to reduce the fluid overload step by step. But it is important to take account of other factors like antihypertensive drugs (which may need to be reduced in time with the reduction of fluid overload) and it is vital to establish if fluid overload is the only reason for hypertension, or if it could be due to other factors.” When should CBVM be switched on? One point made in the discussion paper was that when a person stands up there is shift of fluid from the blood to the interstitial space in the legs giving an increase in Hct and a fall in RBV. The exchange of this haemoconcentrated blood in the legs with the central circulation (which is monitored by CBVM) is rather slow when standing. Lying down facilitates exchange so that there is an initial increase in Hct as blood from the legs mixes into the central circulation. Later the Hct will decrease and the RBV will rise, because of refilling from the tissues. The refilling observed on lying down takes less than 15 minutes in normal controls but more than 30 minutes in dialysis patients. Judith explained that she and her colleagues believe this is because the dialysis patients have a lot more fluid in the interstitium than normal. It is not known how long this refilling process takes as in the study quoted (ref 56 in the paper) the authors waited for 30 minutes and there was still some refilling. This means that when a patient is connected to the machine, their Hct could still be falling due to changing to the recumbent dialysis position. This would give a rise in relative blood volume at the start of dialysis (which is observed sometimes). Should we wait until, say, 15 minutes into dialysis before starting CBVM to avoid an unreliable baseline value due to postural changes? Judith explained that her group has discussed this and they did not think we should delay starting CBVM because we would miss the start of ultrafiltration, but we should be aware that sitting down for dialysis could lead to RBV changes (most likely an increase). Similarly, if we place a patient in the Trendelenburg position during dialysis there is a gradual rise in RBV after a few minutes, even when the patient is almost at their target weight. Jean-Yves raised another ‘start-up’ question. In his unit, the patient is usually connected to circuit filled with saline. Although CBVM is not started until there is blood in the line, the saline ‘infusion’ will have an effect on the RBV. Is this a problem? Franta said that he would not worry about the saline introduced at the start of HD as is quickly diluted and disappears into interstitial space. Judith pointed out that the monitor should not provide any measurements until the Hct is stable. However, as the first stable reading is the referral point for the whole session, an unusual situation at the start will affect all the readings. Postural changes in elderly patients Franta commented that the effect of postural changes on RBV can be quite pronounced, especially in older patients with compromised cardiac condition. Jim Curtis from the US thought the explanation of this might be cardiopulmonary recirculation (CPR). He referred to a paper by Danny Schneditz and colleagues [4] which describes how, as cardiac output (CO) drops, there is more CPR. This is because blood that has passed through the dialyser mixed with blood that has returned from the rest of the body in the heart and lungs. Some of this mixed blood then flows back to the fistula and the dialyser. If the CO is very low, the mixture will contain a high proportion of previously dialysed (and ultrafiltrated) blood. Elderly patients with low CO and high CPR may have inaccurate RBV measurements because the blood in the extracorporeal circuit does not reflect the volume changes in the whole patient. If a patient’s CO changes with the change in posture, the RBV could be affected. Jim noted that CPR does not occur with central catheters, only fistulae or grafts. “Has anyone noticed”, he asked, “if postural RBV changes are less likely to occur in an elderly patient using a catheter?" Danny Schneditz from Austria confirmed that haemoconcentration of the blood in the extracorporeal circuit will increase CPR (and also with access recirculation) due to previous ultrafiltration of the recirculated blood. The sensors might calculate a RBV of 90% while in the whole body it might be only 91%. He explained that “in the absence of access recirculation the size of this error depends on the ratio of UFR to systemic blood flow (CO - access blood flow). We have observed the effect of recirculation on RBV during pulsed ultrafiltration. When the ultrafiltration is stopped after a period of high UFR, there is a very fast (but small) rebound which is due to the dissipation of Hct concentration gradients, similar to the dissipation of urea gradients after stopping dialysis. We have not formally analysed our data for CPR effects - maybe this is an opportunity to take a closer look." Danny asked if Franta could provide more information on the changes seen in the elderly, as he thought differences in venous capacitance would have a greater influence on the size of the RBV change with posture changes than CO. Franta said he had not looked at the effect systematically, but he had noted that older people preferred to be in bed rather than in a chair during dialysis. “For most of the dialysis session they are lying down, but when a meal is being served they sit up and sometimes also hang their leg over the side of the bed,” he explained. “In patients who sit up, the drop in RBV is greater than in those who stay lying down when eating, and the drop is even bigger if they hang their legs down. As soon as they resume the supine position their RBV goes up again.” Franta attributed these changes to the same causes as the higher frequency of oedema in older people - compromised cardiac status and elasticity of vascular system. He also mentioned that in some patients he repeatedly sees an increase in access flow and CO towards the end of dialysis despite significant ultrafiltration. Using CBVM in single needle dialysis Jean-Yves asked for advice on the use of CBVM in single-needle dialysis, which is still used often in Belgium. In this mode there is always a little, often a lot, of recirculation. Lizzi Lindley replied that, with the CritLine at least, CBVM can be used but it gives a noisy trace if the patient has poor fistula outflow. Blood returning to the patient, at perhaps 400 ml/min, in the venous cycle can't leave the fistula so it backs up towards the hand. When the arterial cycle starts, the first blood seen by the monitor is the haemoconcentrated blood that already been untrafiltrated. Later in the arterial cycle, when the backed up blood has been cleared, the monitor will see ‘fresh’ blood with a lower Hct. The timing of the RBV measurements means that the monitor will see different parts of the arterial cycle, giving a noisy trace. The ‘wiggles’ will increase in size with the amount of recirculation and the UFR. Judith suggested that the wiggles during single needle dialysis can be explained by the blood standing still in the measurement chamber. Presumably this would depend on the geometry, but assuming that there is a concentration of red cells at the point of the measurement, this would have a similar effect to recirculation, except that artificially high Hct (low RBV) values would be seen at the end of the arterial cycle, not the start. The two effects could be distinguished by switching the ultrafiltration off which will reduce the noise due to recirculation. Franta agreed that the Critline RBV signal during SN dialysis is indeed unstable but, despite this, the trend is clearly visible. He provided an example showing a patient eating during single needle dialysis (figure 2 When patients have very significant recirculation in dialysis with two needles, switching the UFR off will give a rapid rise in apparent blood volume, like the rebound Danny described after a pulse of high UFR. Judith mentioned that recirculation in catheters can be important. “In our unit,” she said, “it is not uncommon to attach the venous line to the arterial side of the catheter and vice versa when the arterial catheter can't provide sufficient blood flow.” Judith’s comment on switching the red and blue catheter connectors to achieve higher flows attracted a number of comments. Jean-Yves reminded the club that this would lead to lower efficiency. This would be offset to some extent if switching the lines reduced the time in alarm states with no flow. Martin Gerrish from the UK said that his unit used the same strategy as Judith’s and their urea reduction ratio measurements suggested that the level of recirculation was not excessive. Franta provided some objective data. The recirculation induced by inverting the double lumen catheter outlets is not usually high because the catheter is inserted in a large vein with a high flow. His unit had measured recirculation in catheters with inverted connections (with the Transonic monitor) and found it was usually below 15-20%. “This is just about double the standard CPR in fistulas and grafts, which is not present in catheters,” he explained. “It is certainly better to invert the connections than to try single needle mode using one lumen.” Jean-Yves took a more holistic view saying that the measured URR (or Kt/V) may be reasonable, but it does not always reflect what is happening in the overall patient. His unit carries out electromyography (nerve conductivity) measurements every 3 months as marker of long term dialysis efficiency, rather like monitoring HbA1c in diabetics. Reduction in nerve conductivity is caused by uraemic neuropathy (inadequate dialysis) and diabetic neuropathy (poor glycaemic control). It is often associated with restless legs. J-Y felt that the trend in EMG over time (6 months and more) should be used to assess the efficacy of using double lumen catheters used as intended and with switched ports. Food intake during blood volume monitoring Figures 1 and 2 Thomas Roy from Germany brought the club’s attention to a paper by Schallenberg, Stiller and H Mann from 1987 [5]. They used off-line monitoring but their data are comparable to with modern systems. In the 7 hour dialysis session shown in figure 3 Prof Shaldon studied the splanchnic blood flow and blood volume control for his MD thesis. "Eating has a profound effect on splanchnic circulation which comprises the vascular bed defined anatomically as the gut and portal (liver and spleen) arterial and venous blood vessels,” he explained. “It is a system that has its own regulatory controls which are largely independent of the systemic circulation. The total capacity of this system can expand to over 30% of total blood volume. This will effectively reduce systemic blood flow (but the reduction should occur without a change in RBV). Judith had a possible explanation for the increase in Hct. "In order to digest food, blood is necessary so it has to be transported to the stomach. During haemodialysis, the patient’s body needs all its capacity to maintain an acceptable blood pressure. It is then possible that eating induces vascular contraction of the spleen (which acts as a reservoir of red blood cells), releasing red cells into the circulation. This would lead to an increase in Hct, which the blood volume monitor would interpret as a fall in blood volume. James Tattersall from the UK described the effect of eating in another way. "Dialysis and ultrafiltration has a similar effect on the body as exercise or stress. The heart rate increases, the arterioles constrict (which causes capillary pressure to fall and allows interstitial fluid to return to the circulation), the arteries supplying gut and skin constrict and gastrointestinal secretion is inhibited. Anticipation of a meal has the opposite effect. Blood is diverted to the gut and gastrointestinal fluid secretion starts. Up to 3 litres of fluid can leave the blood and enter the gut. The body is not ‘designed’ to cope with eating and exercise at the same time so the tachycardia and arteriolar constriction is ‘turned off’ at the same time as the gut secretion increases.” So eating could result in a redistribution of blood volume to the splanchnic circulation, release of red cells from the spleen and/or secretion of fluid from the blood into the gut. Danny added the “F-cell ratio” (described by Robin Fahraeus in the 1920's, see ref 48 in the discussion paper) to the explanation. “The (apparent) haemoconcentration can be explained by a change in the so-called F-cell ratio, i.e. the ratio between whole body and large vessel Hct. The microcirculation is an important fraction of the blood volume, but the Hct in this compartment maybe be as low as 15%. If the large splanchnic microcirculation is opened during eating (as described by Prof Shaldon), blood with a low Hct will be sequestered to this compartment. For mass balance, the Hct in the large vessels has to increase. The blood volume monitor, which measures the haemoconcentrated blood in the central vessels, will see an Hct increase and interpret this as a decrease in RBV. In reality, there is an intravascular shift of blood without a real change in blood volume. Of course, one cannot rule out that some plasma water is removed from splanchnic circulation and secreted into the GI system, which also will lead to hemoconcentration. But the changes observed with eating are too fast to be explained by secretion effects alone.” Although there is no real change in blood volume, these apparent RBV drops may be associated with IDH as blood is being diverted from the central circulation. However, the solution is to change the timing of food intake (or the patient’s posture during and after eating), not to increase the patient’s dry weight. Lizzi wondered if the low Hct in the microcirculation might be that there has to be a layer of plasma next to the vessel wall, so the more surface area you have, the more plasma you need to provide this layer. If the F-cell effect is responsible for the rapid (apparent) fall in RBV on eating, it could also be responsible for the rapid falls seen when patients start exercising as the blood flow to the muscle microcirculation (capillaries) increases. This can be seen in Fig 1c of the discussion paper, where a patient is cycling for a few minutes during dialysis. As Judith points out in the discussion paper, during exercise blood is recruited into the central circulation by contraction of other parts of the microcirculation. This may explain the rapid rebound in RBV when the patient stopped cycling. Some research has been published by Sandip Mitra and colleagues [6] showing a steady increase in the F-cell ratio during dialysis, presumably due to constriction of the microcirculation to maintain the central blood volume during ultrafiltration. This increase in the F-cell ratio means that the total blood volume change will be underestimated. Biofeedback systems Judith was asked if her unit was actively using the CBVM. She pointed out that, although the discussion paper described a number of limitations, it also referred to studies that showed positive effects where biofeedback systems were used to control fluid removal. Her own centre had observed a decrease in hypotensive episodes when using the Gambro-Hospal biofeedback system (‘Hemocontrol’), though it can be difficult to work with. Judith felt that innovative research is still needed to improve these devices and make it easier to get an insight into the patient’s condition. “We find it very important to evaluate the RBV together with the other information we have about the patient such as shortness of breath, peripheral oedema, blood pressure variation, antihypertensive medication and presence of diabetes,” said Judith. “The RBV measurements should come in addition to standard physical examination and other physiological measurements (heart size on chest X-ray, measurement of inferior vena cava, bioimpedance analysis etc).” Jean-Yves was uncomfortable about handing over control of fluid removal to the machine. “What if the blood volume device gets faulty information and sends the wrong feed back instructions to the dialysis machine? I would prefer to receive ‘suggestions’ so that I can combine them with my observations of the situation and change parameters by hand.” Martin agreed. “We use CBVM to help set target weights, as a support to our clinical judgement. We don't use feedback control because we don't feel able to rely on RBV to ‘control’ the session. Also, according to the training I have received, if the patient is approaching their critical blood volume the machine reduces the UFR to allow refill. This means there is a need for a higher UFR later in the session which seems more likely to cause volume depletion. Fortunately, the majority of our sessions are 4 hours or longer, so IDH is not a big problem.” Bertrand responded to the point about high UFR. “With blood volume control, it is important to enter the right UF-max for the beginning of the dialysis. If UF-max is too low, you will have the problem Martin describes, but with the right setting most of the fluid is removed early in dialysis. The Fresenius BVM controls only the UFR, while Hemocontrol controls UFR and Na, so it is like having both a UF and Na profile. With Hemocontrol, we find you not only achieve the UF-max even with a high target weight loss, but frequently you can remove another 200 to 500 ml without problems like cramps, headaches and hypotension.” For the benefit of club members who have no experience with this system, Bertrand kindly provided some information on how Hemocontrol works. In his unit in Switzerland, the system is incorporated in Integra machines. “When Hemocontrol reduces UF, the dialysate sodium is increased to help refilling until the RBV shows an adequate response. The dialysate Na can be a high as 160 mmol/l, but only for one or two minutes. The patient’s serum Na, which is detected by the Diascan, either doesn’t change increases transiently, e.g. from 138 to 140 mmol/l for a few minutes before going back to 138.” “Before the Hemocontrol system can be used,” Bertrand explained, “you need to collect data for 12 dialysis sessions with the patient. The UF volume, dialysis time, total change in RBV and, where there were incidents, the RBV when the problem occurred must be recorded. You then need to calculate some input values for the Hemocontrol; the %RBV to UF ratio (e.g. -3:1, which is a 3% fall in RBV for each kg weight lost), the average UFR and the UF-max (average UFR x 1.8). These values are programmed into the patient’s ‘Dialpass’ flash card with their target weight, urea distribution volume (we use 53% of target weight) and the alarm limits. The latter should be as wide as possible, otherwise the system will have not enough freedom to work. It takes another 12 dialysis sessions using the ‘Dialpass’ to refine the initial inputs. After this, you have a good basis to work with but you may still need to adjust the inputs for some sessions depending on the patient.” Bertrand also described how the Hemocontrol feeds back to the operator using a cross on the screen with a moving dot. Ideally the dot stays close to the centre of the cross. If the patient is refilling easily it will move up and right and you may be able to remove more fluid. If it moves down and left (as shown in figure 4 Judith added some warnings about using biofeedback. “Do not start biofeedback during a session where packed red blood cells will be administered or where hemodiafiltration (on-line or other wise) is to be used. Experience has shown us that the RBV changes during dialysis with and without HDF in the same patient are very different. We also consider administration of large (>100 ml) quantities of saline or colloids during a treatment to be a reason to stop using biofeedback treatment for that particular dialysis.” Prof Shaldon was concerned about the potential of the Hemocontrol system for loading patients with sodium. Judith replied that a number of the papers included in the review (refs 32, 34, 35, 38 and 39 in the discussion paper) looked at this issue, either by comparing plasma sodium concentrations sessions with and without Hemocontrol, or by monitoring sodium levels before, during and after a Hemocontrol session. Prof Shaldon was not convinced. “I am afraid that the evidence provided does not fully justify the statement that these systems do not deliver sodium to the patient during dialysis. The patient most likely will have stored the excess sodium in non-osmotically active sites bound to. At the 2005 International Society of Blood Purification meeting in Rotterdam, Dr Chazot from Tassin showed that even with a dialysate sodium of 137mmol/L the patient received salt during dialysis [7]. The commonest post-dialysis symptom used to be thirst, has this changed with all these gadgets now in use?” The osmotically inactive sodium stores were described by Dr Jens Titze in a previous journal club discussion [8]. Anahita provided the club with a presentation on how her unit in Norway uses the Fresenius biofeedback system. Judith explained that this device uses the ‘crash-crit’ principle and varies the UFR based upon RBV changes, while the Hemocontrol described above varies the UFR and dialysate conductivity to guide a patients RBV along a pre-set curve. The patients considered for biofeedback are those who cannot tolerate a constant UFR throughout dialysis and those who need very high UF volumes. The first step is to monitor the patient using CBVM without control for 4 to 6 sessions, systematically recording the UFR, total RBV reduction, minimum RBV and RBV where problems occurred. After evaluating the data collected, the minimum RBV and maximum allowed UFR are programmed into the control system. Like Bertrand, Anahita stressed that the importance of finding the maximum UFR that the patient can tolerate. As an example, Anahita showed a patient who suffered with symptomatic hypotension during dialysis and headaches on non-dialysis days. With blood volume control, combined with a strategy of lying down without food intake for the first hour (while the UFR is high) and lying down again for 30 minutes after eating, the patient had no fall in blood pressure and felt good between dialysis days. This individualisation of dialysis gives a more pleasant and efficient dialysis, but it requires co-operation between staff and good documentation. Combining biofeedback control and on-line clearance monitoring On-line clearance monitoring based on ionic dialysance measurements (the Fresenius ‘OCM’ and Hospal ‘Diascan’) requires the machine to make large changes in the dialysate conductivity for short periods. Judith commented that her unit does not use biofeedback control in patients with over 20% recirculation because, with this degree of recirculation, the RBV can be seen to react to these changes in dialysate sodium concentration. She provided a chart showing apparent falls of 3-5% in RBV when the dialysate conductivity was increased from 14 to 15 mS/cm to make clearance measurements. Danny replied that this artefact has positive and negative implications. The good thing is that it can help with the identification of some types of access problem. The bad thing is that it interferes with the Hemocontrol concept. “Multiple scattering, which leads to the non-linear relationship between optical density and Hct, is affected by red blood cell shape. The shape and volume of red cells is likely to change when changing the osmolarity and conductivity of blood during an on-line clearance test. For me,” said Danny, “the most plausible explanation is that the red blood cells shrink when exposed to higher osmolarity. The optical density apparently increases and this is (erroneously) interpreted as a drop in RBV. But a volume drop is the opposite effect to what would be expected when increasing dialysate conductivity, as this is like giving an infusion of hypertonic saline. Judith’s data shows that an increase in conductivity appears to decrease RBV. In the control algorithm, dialysate conductivity is automatically increased when blood volume decreases. This should promote vascular refilling and increase RBV so the dialysate conductivity can be reduced. However, the artefact Judith sees will lead the control algorithm to further increase conductivity which will lead to another apparent decrease in blood volume. This is positive feedback which is inherently unstable!” Danny felt that, whilst the artefact was very obvious in the presence of recirculation, the problem of positive feedback when varying dialysate conductivity is always present so the two devices should not be used concurrently. Raphaël Cottancin, the Product Manager for Hospal HD machines in France responded to Danny’s points, though not in time for his answer to be circulated to the club. “In the presence of recirculation, the dialysate conductivity step of the Diascan measurement induces a change in the osmolarity of the recirculating blood. When the Diascan conductivity step is up, the osmolarity of the blood is increased and we see a quick drop of blood volume synchronized with the step. In fact, this change of osmolarity leads to a water shift from the red blood cells to the plasma water. Although the haemoglobin concentration is constant, we see a drop in the measured RBV, which means that the light transmitted to the receiver through the blood line is reduced. There are two possibilities: either the light absorption is increased or the light reflection is decreased. Since amount of haemoglobin present is not changed, there is no reason to think there should be more absorption, so we should consider the second hypothesis. The mechanism leading to reduced reflection could be higher protein adsorption on the surface of the red blood cells due to the higher osmolarity [9].” In Raphaël’s opinion, it should be possible to use the Hemocontrol and Diascan simultaneously on the Integra machine. “In the case of recirculation, the two minute change in conductivity will affect the blood volume curve, but it is not representative of the trend and it does not affect the efficacy of the Hemocontrol system.” Judith feels strongly that the two devices should not be used where there is significant recirculation (over 20%) so there is an interesting project here! Is there a need for practical guidelines? In the discussion paper, Judith and her colleagues advocated the development of practical guidelines on how to interpret and use the information generated by CBVM. The volunteers who responded to the survey all agreed that this would be useful. EDTNA/ERCA will consider the possibility of producing some guidance in the form of case studies and/or a forum where CBVM users can share experiences. Take-home messages
Acknowledgements We are very grateful to Prof Norbert Lamiere, Editor-in-Chief of Nephrology, Dialysis and Transplantation, for giving EDTNA/ERCA permission to circulate the paper freely and to Dr Judith Dasselaar for her enthusiastic participation in the discussion. We would also like to thank Christel Levebre of United Networks in Belgium for providing technical support for the Journal Club mailing system for the last four discussions. And finally, thanks to the EDTNA-ERCA volunteers who responded to the survey and the JC members who shared their thoughts and ideas on the use of CBVM. References
|
 |
 |
 |
 |
 |
 |
 |
 |
 |
EDTNA/ERCA Secretariat - Prague Congress Centre - 5. kvetna 65, 140 21 Praha 4, CZECH REPUBLIC - Tel +420 261 174 318 - Gsm +420 777 777 098 - Fax +420 261 174 318 - Email: queries@edtnaerca.org
EDTNA/ERCA, Pilatusstrasse 35, CH-6003 Lucerne, Switzerland