Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Is prophylactic haemofiltration during cardiopulmonary bypass of benefit during cardiac surgery?

Three Part Question

In [patients undergoing elective Cardiac Surgery] does [prophylactic Haemofiltration] improve [survival or time to discharge or days in CSU]?

Clinical Scenario

You are performing a difficult aortic valve replacement in an 80-year-old patient that also requires three coronary grafts and has an ejection fraction of only 35%. You know that the bypass time is going to be long. The perfusionist informs you that in the last institution he worked at, every patient was prophylactically haemofiltered if bypass was used and that this reduced inflammatory mediators and improved outcome. You decide to use a haemofilter in this high risk case but resolve to look up the evidence for this after the case.

Search Strategy

Medline 1966-07/03 using the OVID interface.
[(exp hemofiltration/ OR OR OR AND (exp thoracic surgery/ or thoracic or cardiac or OR exp Cardiopulmonary Bypass/ OR Cardiopulmonary] LIMIT to Human and English

Search Outcome

273 papers were found of which 9 were deemed to be relevant.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Despotis et al,
20 patients undergoing CABG Investigated how haemofiltration effects heparin level variability and whether it filters the low molecular weight fraction Heparin measured indirectly using anti-Xa and anti-II.a assaysCase Series (4)Levels of heparin activityPre haemofiltration anti-Xa heparin activity was 3.9 +/- 1.7 U/ml

Post haemofiltration Anti-Xa heparin activity was 5+/-1.8 U/ml P=0.003

Haemofiltration increases heparin concentration and contributes to variability

HF does not filter out low molecular weight heparin into ultrafiltrate
Small study. Main finding is that low molecular weight heparin is NOT filtered out They then report their positive secondary finding which seems to have C.I.s that cross
Babka et al,
60 patients undergoing CBP Ultrafiltration group: n=30, ultrafiltration with CPB control group n=30 standard CPBProspective case – control trial (3b)Physiological parametersNo difference in blood loss, blood transfused, length of stay or cost of patient

Significant difference in post op weight gain (3.5 Vs 4.8 lbs) and mean ultrafiltrate vol was 2510ml
Ultrafiltration was of little clinical value Controls had a significantly longer Xclamp time (32 vs 38mins) No power studies given for null findings
Tassani et al,
43 patients having elective CABG Modified ultrafiltration group: n=21- zero fluid balance maintained Control group : n=22PRCT (1b)Immune mediator levelsIL-6 and IL-8 significantly lower levels in the Ultrafiltration group immediately post CPB

These levels were not significantly different at 2 and 4 hours

No difference in IL-10 and IL-1 levels
There was a short difference in inflammatory mediators that disappeared after 2 hours
Van Norman et al,
2 groups of patients randomised to haemofiltration or no haemofiltration during cardiopulmonary bypassPRCT (1b)Aprotinin levelsNo difference in aprotinin levels in the two groupsHaematocrit improved with haemofiltration
HaematocritSignificantly higher haematocrit levels in the haemofiltration group
Grunenfelder et al,
97 patients undergoing CABG with CPB MUF group, n=60, Modified ultrafiltration for 15mins on CPB Control group, n=37, CPB only Stratified study as 2 groups also subdivided into Normo thermic and Hypothermic CPB. Hypothermic patients had temp 26-28 C on CPBPRCT (2b)Immune mediatorsMUF led to a significantly lower level of cytokines IL-6, IL-8, TNF and IL2R) and adhesion molecules

Normothermia also led to a similar reduction
No clinical differences found Roller pump used for CPB Underpowered to make conclusions regarding mortality or morbidity
Clinical outcomeNo difference in time to discharge or mortality/morbidity between the filtered and unfiltered groups
Boga et al,
40 CABG adult patients Control, n=20, standard CPB Treatment, n=20, modified haemofiltration for 20 mins on rewarming at the end of CPBPRCT (1b)Haemodynamic parametersImmediately postoperatively CI and SVR both higher in filtered group. These differences quickly became non significant

Haematocrit was higher (0.33 vs 0.29 ) and blood transfusion needs also significantly lower 0.83 vs 1.84 in the HF group, P<0.05
Small study but did find significant improvements in post op blood loss and CI post op
Blanchard et al,
2 groups of CABG patients, Control, N=13, standard rewarming and CPB Treatment, N=13, Haemofiltration on rewarming on Cardio Pulmonary Bypass. 15ml/kg filtered Haemodynamic and echogardiographic parameters measured on completion of rewarmingSingle blind PRCT (1b)Haemodynamic parametersTreatment: No change in SVR or CI

Control : Significant drop in SVR and rise in HR and CI
Echocardiographic parametersSignificantly improved kinetic score in treatment group compared to the control group
Onoe et al,
18 patients undergoing cardiac surgery Treatment: 9 patients had CPB and standard ultrafiltration followed by MUF on rewarming Control: 9 controls had CPB only Serum IL-8 measured immediately after CPB and 3h after CPB (IL-8 is a cytokine which strongly promotes neutrophil degranulation and infiltration)PRCT (2b)Serum Levels of IL-8Treatment group: IL-8 reduced from 69.5=/-33 to 58.9+/-32 after ultrafiltration instituted
IL-8 level is reduced by a little and the haematocrit is increased But all findings have v.wide confidence intervals, all of which seem to cross on the graphical images of their results Trend towards higher BP also reported but NS
HaematocritTreatment group: Haematocrit increased from mean 21 to 24 P=0.0008
Control: No change in haematocrit
Leyh et al
48 patients undergoing myocardial revascularisation randomised to: Conventional Ultrafiltration CUF (n=16) Modified ultrafiltration MUF (n=16) Control group (n=16)PRCT (1b)Post-op transfusion volumeMUF 2.0ml/kg bw
CUF 6.9ml/kg bw
Control 7.0ml/kg bw
This study finds that there is less blood loss and need for transfusion with MUF but not CUF Small study Roller pump CPB used
Post-op blood lossMUF 6.4 ml/kg bw in 24hrs
CUF 9.2 ml/kg bw in 24hrs
Control 8.9 ml/kg bw in 24hrs
OtherNo difference in levels of any clotting or fibrinolytic system markers, including ACT, PT, APTT, fibrinogen, platelet count


It is noted that modified haemofiltration (MUF) is the technique of performing haemofiltration for 10-20mins at the end of bypass, during rewarming and in contrast to conventional haemofiltration, which is performed throughout cardiopulmonary bypass. To perform modified haemofiltration, the patient is first weaned off bypass and then the haemofilter pressure is maintained by the difference in pressure between the arterial inflow and the right atrial outflow. Three studies showed a reduction in inflammatory markers including IL-8 and IL-6 with haemofiltration, although one study found no difference. Two studies showed a reduction in post-operative bleeding, although one study found there to be no difference. Three studies report improved haemodynamics after haemofiltration including improved cardiac index and BP. Five studies report an improved haematocrit or reduced patient weight post haemofiltration. Individual studies also find an increased variability in heparin concentration but no change in aprotinin levels. All studies are small and therefore there is no reliable data on the effect of haemofiltration on mortality or morbidity.

Clinical Bottom Line

Haemofiltration will increase the haematocrit, reduce some inflammatory markers and may increase the variability of heparin levels. It may also reduce post-operative blood transfusion and possibly increase BP and cardiac index immediately after haemofiltration, although no differences in morbidity or mortality have ever been shown.


  1. Despotis GJ, Levine V, Filos KS, et al. Hemofiltration during cardiopulmonary bypass: the effect on anti-Xa and anti-IIa heparin activity. Anesth Analg 1997;84(3):479-83.
  2. Babka RM, Petress J, Briggs R, et al. Conventional haemofiltration during routine coronary bypass surgery [erratum appears in Perfusion Nov;12(6):347]. Perfusion 1997;12(3):187-92.
  3. Tassani P, Richter JA, Eising GP, et al. Influence of combined zero-balanced and modified ultrafiltration on the systemic inflammatory response during coronary artery bypass grafting. J Cardiothorac Vasc Anesth 1999;13(3):285-91.
  4. Van Norman GA, Patel MA, Chandler W et al. Effects of hemofiltration on serum aprotinin levels in patients undergoing cardiopulmonary bypass. J Cardiothorac Vasc Anesth 2000;14(3):253-6.
  5. Grunenfelder J, Zund G, Schoeberlein A, et al. Modified ultrafiltration lowers adhesion molecule and cytokine levels after cardiopulmonary bypass without clinical relevance in adults. Eur J Cardiothorac Surg 2000;17(1):77-83.
  6. Boga M, Islamoglu, Badak I, et al. The effects of modified hemofiltration on inflammatory mediators and cardiac performance in coronary artery bypass grafting. Perfusion 2000;15(2):143-50.
  7. Blanchard N, Toque Y, Trojette F, et al. Hemodynamic and echocardiographic effects of hemofiltration performed during cardiopulmonary bypass. J Cardiothorac Vasc Anesth 2000;14:393-8.
  8. Onoe M, Magara T, Yamamoto Y, et al. Modified ultrafiltration removes serum interleukin-8 in adult cardiac surgery. Perfusion 2001;16(1):37-42.
  9. Leyh RG, Bartels C, Joubert-Hubner E, et al. Influence of modified ultrafiltration on coagulation, fibrinolysis and blood loss in adult cardiac surgery. Eur J Cardiothorac Surg 2001;19(2):145-51.