2007 Spring EDTNA/ERCA Journal Club Discussion Summary
Guarding against hidden hemolysis during dialysis: An overview

A discussion based on "Hemolysis: A Hidden Danger by Elisabeth Harman RN, CNN, Clinical Coordinator, Central Valley Dialysis, Salt Lake City, Utah, USA and Paula Dutka, RN, MSN, CNN, Director Education/Research Nephrology Network, Winthrop University Hospital, New York, USA. The paper was published in The Nephrology Nursing Journal, the official publication of the American Nephrology Nurses Association.

Edited by: Gareth Murcutt1, based on contributions from Jean-Yves De Vos2, Ken Pilley3, Elizabeth Lindley3, Franta Lopot4, Susan K Hansen5, Hans-Dietrich Polaschegg6, Melissa Chamney3, Nic Hoenich3, Narayan Venkatraman7, Stanley Shaldon8, Anna Marti i Monros7, Mukesh Gajaria10, Waltraud Kntzle11, Elisabeth Harman5, Paula Dutka5

1Center for Nephrology, Royal Free and University Hospital Medical Schools, Royal Free Hospital Campus, London, UK; 2Belgium; 3UK; 4Czech Republic; 5USA; 6Austria; 7Singapore; 8Monaco, 9Spain, 10Canada, 11Germany.


The paper discussed during spring 2007 was a case study report entitled "Hemolysis: A Hidden Danger published in The Nephrology Nursing Journal. The authors, Elisabeth Harman and Paula Dutka, agreed to follow the discussion and respond to points raised. Sixteen contributors from ten different countries provided insights into the potential causes, symptoms and effects of both acute and hidden hemolysis during dialysis, as well as discussing some of the safety systems that can be used to try and minimise occurrences. The use of blood volume monitoring as a potential method of seeing hidden hemolysis was explored as well as some reporting mechanisms and organisational safeguards that are used to manage the risks.

back to top


Hemolysis, dialysis, blood volume monitoring

back to top


The paper selected was a case-study paper entitled "Hemolysis: A Hidden Danger by Elisabeth Harman and Paula Dutka [1]. The paper discussed four incidents of post-dialysis hemolysis, adversely affecting several patients at more than one clinic. Hemolysis during haemodialysis has many potential causes so the authors institutions carried out investigations to find any factors common to all patients. Amongst the potential causes ruled out, were water quality, dialysis machines and dialysers (new and re-used). In all cases the most likely cause appeared to be a narrowing of the bloodlines, though it must be mentioned that the manufacturers analysis of these did not offer any evidence of this. The authors thus kept an open mind about the actual root cause of the incidents.

Damage to red blood cells (RBC) occurs in all dialysis sessions to some degree, due to the nature of the treatment. Possibly the most important point made by the authors, was that there were no obvious signs of hemolysis in the extracorporeal circuit, in effect, the hemolysis was hidden! Only towards the end of dialysis did any symptoms become apparent, the most common being abdominal pain. Nothing during the dialysis treatments seemed to predict the severity of the hemolysis and the authors were grateful that no fatalities resulted.

back to top

Hemolysis in Dialysis: A Historical Iinsight

Susan K. Hanson provided a personal description of haemodialysis in the 1960's and pointed out that even with the technology available then, hemolysis was quite rare. My career in dialysis started in 1965 - a very different world. In the US at that time nurses were not allowed to put anything into the bloodstream (i.e., meds, start IVs, etc) so they couldn't do dialysis. We dialyzed two patients at once, a total of ten patients for two treatments each per week. Hemolysis was uncommon even then but it happened. Blood pumps were cruder, shunt flows could abruptly vary during treatment and it was possible to forget to weigh and add one of the dialysate chemicals. There were no dialysis monitoring systems and the operator had to visually monitor the bloodlines, dialyser and dialysate throughout the six-hour treatment. Fortunately, significant hemolysis is very obvious in the venous drip chamber and the large blood volumes of these early dialysers meant that the blood circuit was primed with two units of bank blood before each treatment. Discontinuing the treatment, and discarding the hemolysed extracorporeal circuit, whilst not desirable, had no serious clinical sequellae. In my experience, the first sign of hemolysis was that the blood in the venous line or drip chamber turned strangely transparent - although it was the same red as before - you could see right through it! Another sign was when we did Lee-White clotting times (every half hour) and the blood in the tube was noted to be oddly translucent. We were especially vigilant after dialysate batch changes (we dialyzed with a Travenol 100 L tank unit), even though we double-checked the chemicals pre-bath change.

It appears that, although the risks of hemolysis were undoubtedly higher in the 1960s, the close scrutiny afforded to all dialysis treatments minimised its occurrence or at least ensured it was quickly detected and acted upon.

back to top

Potential Causes of Hemolysis in Haemodialysis

Haemodialysis (HD) is a technically demanding treatment and there are inherent risks in the process. Though HD machines are designed to detect and safely manage many of these risks the technology is far from perfect. One such risk that HD machines can never totally protect patients against is hemolysis, whose causes can vary from over-occluding bloodpump rollers to a kinked bloodline to inadequate water treatment or incorrect dialysis fluid composition.

Chemically induced hemolysis
Ken Pilley noted, " This is an interesting subject that could be caused by failure of a water treatment system. For example, if a carbon filtration system becomes saturated, then higher chlorine levels could be seen which may contribute to the problem. Other potential causes are the raw materials used during manufacture of the bloodlines and if their construction process has been altered in any way." Meanwhile, Narayan Venkatraman provided a fascinating insight into haemodialysis developments in Asia. "This subject of hemolysis is very pertinent for all renal care professionals and reminds everyone about the complexities involved in this treatment process. In the past 5 years, similar incidents did happen in some Asian countries like Indonesia, Cambodia, Sri Lanka and India, mostly due to inferior water treatment systems, inadequate monitoring protocols and older machines being used. Dialyser reuse is a widely accepted practice in most nations in Asia and some less developed nations also re-use bloodlines. Adaptation of modern technology is improving at a good pace and we are getting to see some state-of-the-art HD facilities established even in remote locations.

Both Mukesh Gajaria and Nic Hoenich pointed out the dangers of incorrect dialysis fluid composition. Mukesh commented, It is well known that hemolysis can also be caused by hypotonic dialysate. I would like to know if anyone has done any osmotic fragility tests on the particular patients affected by hemolysis using the dialysate they were on. Every ESRD patients RBC cell walls could be different, and could be "injured" in different ways. Nic Hoenich brought to our attention the paper by Pendergrast JM et al Hemolysis due to inadvertent hemodialysis against distilled water: Perils of bedside dialysate preparation [2]. The paper analysed four cases of hemolysis, within intensive care settings, over a two-year period and concluded by saying dialysis against distilled water is a potentially serious but preventable complication of bedside dialysate preparation.

Though not covered in depth during the discussion, any article on hemolysis must draw the readers attention to the many-recorded cases of patient exposure to disinfectants, usually chloramines, which have resulted from inadequate water treatment [3, 4]. Chloramines, commonly added by water treatment works to provide safe drinking water, have been the cause of both acute and hidden hemolysis, often characterised by multiple patients being affected [5].

Mechanically induced hemolysis
Jean-Yves De Vos commented that some frequent reasons for hidden hemolysis are neglected kinks) not completely blocking partly the flow and unnoticed by the arterial nor venous pressure measurement ports. Also the extension tubing of the metal needles when using catheter-needles can cause a persistent kink (internally in the vein) if the patient bends their arm, causing hemolysis as a result. Also, we should not forget about the potential hemolysis caused by over occlusion by bloodpump rollers. We have several patients with Polycystic Kidneys with nearly normal Hb levels without EPO and still have no issue of hemolysis. The question is rather what bloodflow is used, with what gauge needle or catheter diameter and how fast we are dialysing ... the slower and longer the better. Susan Hanson noted that Blood flows in excess of 400ml/min are common in the US and this can lead to very high negative pressures (enough to induce mechanical stress and hemolysis) before the blood pump. Not all HD machines monitor this pressure and this can contribute to hidden hemolysis".
Hans-Dietrich Polaschegg pointed out that his paper in the 'Replacement of Renal Function by Dialysis' 5th edition [6] provided some precision as to the causes of mechanical hemolysis. Mechanical hemolysis is caused by shear, not by pressure. The pressure indicated by dialysis machines is a combination of static (height differences) and dynamic pressure. Shear is proportional to dynamic pressure but not identical. The pressure drop in a long catheter can be substantial but may not cause hemolysis while the same pressure drop in a needle may cause it. It is however possible that erythrocytes are damaged but not hemolysed (sub-lethal damage) and are taken out of circulation by the spleen. The result is a drop of hematocrit a few hours after the event. Hans added Surface roughness has an influence as shown over the last few years in the importance of surface finish for rotary blood pumps in heart-assist devices. Dialysis needles are siliconised and without this layer the risk of clotting and hemolysis would be much higher. Another common cause of hemolysis in dialysis is an obstruction, or kink, between the blood pump and the dialyser. It is, however, unclear if the resulting hemolysis is produced in the kink or in the blood pump. Blood pump rollers are spring loaded and give way when the pressure increases above a certain threshold.

"Although arterial and venous pressure readouts on a machine cant always detect too high/low pressures within a specific section of the circuit, sudden changes in pressures (+/- 20% ) should always alert the operator to be cautious. One issue is that, when a pressure alarm occurs, operators often override this, and some machines automatically set a new alarm window. This makes it easy for the operator to loose track of what is going on.
Nic Hoenich offered the following insight into mechanically induced hemolysis during dialysis:

Although it has been long recognised that haemodialysis is associated with some red cell damage, generally, the damage is sub-lethal and not easily detectable. True hemolysis associated with haemodialysis is rare, mainly due to the introduction of [equipment] safety standards. The mechanical injury of erythrocytes from occluded or kinked haemodialysis blood lines or fistula needles has been the subject of a number of articles in the literature. [7, 8, 9]

back to top

Recognising the Symptoms of Hidden Hemolysis

The classical visual signs of acute hemolysis; changes of colour and/or translucence in the venous blood are well known and should be recognised by clinical staff. The problem identified in the paper however, was that the hemolysis was largely hidden. Lizzie Lindley noted the extremely varied symptoms cited in the report, patients who experienced serious hemolysis had one or more of pruritis (itching), increased blood pressure, shivering, high temperature, were slow to respond, appeared flushed, experienced shortness of breath, nausea and back/abdominal pain. All of these symptoms could also be caused by many factors may not be readily associated with hemolysis. This makes the task of recognising hidden hemolysis even more difficult though Paula Dutka added Abdominal pain which was the prevalent symptom that we identified at our facility and was also noted in many of the references that we reviewed when first researching the issue [10] . For the benefit of non-clinical JC members, Jean-Yves explained, Hemolysis can cause back or abdominal pain and jaundice secondary to a rise in unconjugated bilirubin. Jaundice, in the setting of normal liver function, typically does not occur until >50% of the erythrocytes have been hemolysed. Stanley Shaldon also recognised this Hemolysis does produce back pain. I have seen it at a post-mortem following death from acute real failure. The kidneys were full of bilirubin in cast [unconjugated] form. Haemoglobin coming out in urine also causes kidney pain which is a better location descriptor than the more generic back-pain.

back to top

Can Blood Volume Monitoring be Used to See Hidden Hemolysis

The authors noted in their paper that three patients, later found to have experienced hemolysis, were treated on HD machines capable of the on-line monitoring of blood volume (BV). During the treatments of two of these patients, after an initial, expected, decline in relative blood volume, the display unexpectedly trended upwards, indicating probable vascular refilling. Franta Lopot noticed, It seems that hemolysis is manifested by an increasing pattern of the relative blood volume despite significant ultrafiltration. It is very interesting as it gives another dimension to continuous blood volume monitoring (CBVM). Elisabeth Harman one of the authors of the paper provided the following details from the case studies We use CBVM in every treatment (Critline, Hemametrics Inc, Utah, USA). When the hemolysis occurred, the Critline BV became positive instead of the usual negative and the hematocrit decreased. Normally the hematocrit reading increases due to hemo-concentration as ultrafiltration occurs.
This opens the possibility of using CBVM for the early detection of hidden hemolysis. Franta Lopot was asked whether all types of CBVM could be used and replied optical devices will probably exhibit the same false increase in BV with hemolysis but that ultrasonic-based CBVM systems should not be affected. This was further explained by Hans-Dietrich Polaschegg Ultrasound speed is proportional to the mean density of blood; the morphology has a negligible influence. The mean density of blood does not change when the erythrocytes are destroyed so the ultrasonic signal is unchanged when blood is haemolysed. The signal of the optical monitor is influenced, not only by light absorption by haemoglobin, but also by light scattering by erythrocytes. When erythrocytes are destroyed, the scattering is altered which causes a change in the blood volume signal. Hans also added [Mathematical] differential diagnosis would allow the separation of haemolysis from other causes of blood volume increases. Gareth Murcutt questioned whether there was a possibility that if using an optical CBVM system in biofeedback (UF control) mode, any false vascular refilling type signal could lead to incorrect fluid removal rates and Franta Lopot agreed that this was a possibility.

Nic Hoenich provided some background information Blood volume changes can be made either on the basis of hematocrit or haemoglobin concentrations, with the latter being more common. A number of corrections are made to take into account effects such as oxygen saturation, scattering induced by the flow, as well as having a pulsatile flow. Red cell damage results in the haemoglobin being released into the plasma, which in-vivo rapidly binds to haptoglobin and the complex is eliminated from the circulating blood by the liver. The liver recycles the iron, heme, and amino acids contained in the haemoglobin and haptoglobin. In the presence of active hemolysis, the rate of haptoglobin destruction will outpace the rate at which new haptoglobin is created. Consequently, changes in the concentration of haptoglobin can be used for the assessment of hemolysis. As the measurement of haemoglobin in vivo is made using an optical method and, as indicated above, free haemoglobin is removed rapidly from the plasma, the real question is "does the release of free haemoglobin result in an appreciable change in the haemoglobin levels, and can these changes be separated from those resulting from hemo-concentration? The answer to this remains unclear at present.

back to top

Guarding Against Hidden Hemolysis

It is very difficult to effectively train someone to recognise symptoms that they may see only once in an entire career, but which nonetheless can have very serious consequences for patient safety. Two factors accentuate this problem: one is that the symptoms may not present until after the patient has left the clinic, the other is that hemolysis often affects more than one patient or one shift. Paula Dutka, one of the authors, pointed out From my nursing perspective, one reason I felt so strongly about publishing this article was to get the word out that the cases we experienced were not presenting as we were all taught - cherry red blood with back pain. Another is that I, and many of my colleagues, have been working in dialysis for more than 25 years and have never experienced hemolysis. We must never forget that we are performing a technically complex procedure every day and must never let our guard down!
Melissa Chamney said I have never seen hemolysis in my 14 years of renal, including six years working in clinics in both the UK and in Australia. The only time I "saw" it was on a video! What is quite scary about this is that there may be many nurses who have never seen hemolysis and therefore will probably not know immediately how to deal with it.
Elisabeth Harman made the following comment In retrospect, we now understand that the CBVM showed us the release of haemoglobin as the RBCs were being lysed. The accompanying increase in blood pressure and decrease in hematocrit as RBCs were destroyed all made sense. The instrument alerted us, but because we had never seen this before we did not interpret it correctly at the time of the events. Now, if we are even the least suspicious we do a quick pink test. We draw a couple of mls of blood into a serum separator tube and spin down. If the serum is pink we have hemolysis. We are now much more alert to our patients' symptoms and also realise that in CBVM, we have another instrument to help in our monitoring.
Lizzie Lindley commented, For me, the lesson learned from the ANNA report is that when a CBVM system gives apparently funny data, it should be taken seriously! A rising BV after the patient is established on dialysis is unusual. Staff need to look at the monitoring chamber, check if there is clotting inside it and then look out for the symptoms described in the report.

As well as technically based solutions there are also organisational procedures we can use to reduce the potential risks. Any signs of hemolysis must be taken seriously, as it is apparent that the full symptoms may not present until long after the patient has left the clinic. To prevent potential injury to other patients, clinics need robust and fast acting adverse incident reporting procedures because it is quite possible that later patient shifts could be exposed to the same factors. Anna Marti i Monros, explained how such a system worked in her clinic I fully agree with the papers title. My experience as a nephrology nurse goes back to 1974 and Ive seen a number of hemolysis episodes. In some of them the cause was never identified but some were caused kinks in bloodlines. In my hospital we have an "Anti-Hemolysis Security Procedure" and it refers to a double check of all the circuit lines during the first 30 minutes of the HD session by a different nurse. We also have the needles lines visible all the time during the session.
Hemolysis incidents may also occur at more than one dialysis facility thus external reporting systems may be critical. Waltraud Kntzle provided the following For Europe we have a medical product law which regulates reporting of adverse incidents. The problem is that this law is not well known or implemented and reports are usually filed by manufacturers. In Germany, the user or provider should also notify a central agency (like a German FDA) called DIMDI [11] who will collate any reports and contact the manufacturer. Gareth Murcutt replied that in the UK, adverse incident reporting is supposedly encouraged by the Medicines and Healthcare products Regulatory Agency (MHRA), but is entirely voluntary [12].

back to top


Significant hemolysis is a very rare problem in modern day dialysis, however, when it does occur the results are often extremely serious. As well as the classical hemolysis, presenting as strangely coloured or translucent venous blood and often associated with back-pain, there also exists a level of RBC damage that may be visually hidden from clinical staff. Of the factors that can cause hemolysis, such as inadequate water treatment and faulty dialysis machines or defective consumables, most are likely to affect more than one patient. It is therefore vital that clinical staff are aware of the potential risks and use all the tools and information at their disposal. From the evidence and practical experience presented in the report, and the ensuing discussion, it may be that a single nurse, recognising one or two patients with slightly unusual back pain towards the end of dialysis, could save the entire next patient shift from far more serious consequences. That the nurse may only see this situation once in their entire career puts into perspective how difficult a task it is we are asking them to perform. To that end all signs of hemolysis must be acted upon, and organisations should have reporting mechanisms to rapidly decide if any action is required.

back to top


Harman E, Dutka P. Hemolysis: A Hidden Danger. The Nephrology Nursing Journal, March-April 2007, Volume 34, Number 2 nephrologynursing.net/MA2...efault.htm
Pendergrast JM, Hladunewich MA, Richardson RM. Hemolysis due to inadvertent hemodialysis against distilled water: Perils of bedside dialysate preparation. Crit Care Med. 2006 Oct; 34(10): 2666-2673
Fluck S, McKane W, Cairns T, et al. Chloramine-induced haemolysis presenting as erythropoietin resistance. Nephrol Dial Transplant. 1999 Jul; 14(7): 1687-1691.
Richardson D, Bartlett C, Goutcher E, et al. Erythropoietin resistance due to dialysate chloramine: the two-way traffic of solutes in haemodialysis. Nephrol Dial Transplant. 1999 Nov; 14(11): 2625-2627.
Prez-Garca R, Rodrguez-Bentez P, Chloramine, a sneaky contaminant of dialysate. Nephrol Dial Transplant (1999) 14: 2579-2582
Polleschegg H-D, in Replacement of Renal Function by Dialysis. Hrl WH, Koch KM, Lindsay RM, Ronco C, Winchester JF, (Eds.) 5th ed., 2004, XXI, ISBN: 978-1-4020-0083-6
Seukeran D, Fletcher S, Sellars L, et al. Sudden deepening of pigmentation during haemodialysis due to severe hemolysis Br J Dermatol 1997 Dec; 137(6): 997-999.
Sweet SJ, McCarthy S, Steingart R, et al. Hemolytic reactions mechanically induced by kinked hemodialysis lines Am J Kidney Dis 1996 Feb; 27(2): 262-266
Gault MH, Duffett S, Purchase L, et al. Hemodialysis intravascular hemolysis and kinked blood lines Nephron 1992; 62(3): 267-271
Duffy R, Tomashek K, Spangenberg M, et al. Multistate outbreak of hemolysis in hemodialysis patients traced to faulty blood tubing sets Kidney International 2000 Apr,57(4):1668-1674
www.dimdi.de/static/en/mpg/recht/index.htm (Sept 2007)
www.mhra.gov.uk/home/idcplg?IdcService=SS_GET_PAGE&nodeId=291 (Sept 2007)
back to top



EDTNA/ERCA Secretariat - 24 rue Chauchat, F-75009 Paris, France - Tel. + 33 (0)1 53 85 82 69 - Fax + 33 (0)1 53 85 82 83 - Email: info@edtnaerca.org
Head Office - Pilatusstrasse 35, Postfach 3052, CH-6002 Lucerne, Switzerland