BEST EVIDENCE TOPIC – ADULT CARDIAC
Interactive CardioVascular and Thoracic Surgery 20 (2015) 415–421 doi:10.1093/icvts/ivu388 Advance Access publication 25 November 2014
Modified ultrafiltration in adult patients undergoing cardiac surgery Mustafa Zakkar*, Gustavo Guida and Gianni D. Angelini Bristol Heart Institute, Bristol, UK * Corresponding author. Bristol Heart Institute, Bristol Royal infirmary, Bristol BS2 8HW, UK. Tel: +44-117-9230000; e-mail: [email protected] (M. Zakkar). Received 6 August 2014; received in revised form 24 October 2014; accepted 28 October 2014
Summary A best evidence topic in cardiac surgery was written according to a structured protocol. The question addressed was the impact of modified ultrafiltration on adult patients undergoing cardiac surgery in terms of inflammatory and metabolic changes, blood loss and early clinical outcomes. A total of 155 papers were identified using the search as described below. Of these, six papers presented the best evidence to answer the clinical question as they reported data to reach conclusions regarding the issues of interest for this review. The author, date and country of publication, patient group, study type and weaknesses and relevant outcomes were tabulated. Modified ultrafiltration in adult patients undergoing cardiac surgery seems to attenuate the levels of inflammatory molecules associated with surgery, reduces blood loss and blood transfusion and improves cardiac output, index and systemic vascular resistance. However, this was not translated in any reduction in length of stay in intensive care unit or hospital. Most studies were single-centre prospective non-blinded trials that included a small cohort of elective coronary artery bypass grafting patients, which makes it underpowered to provide unbiased evidence regarding clinical outcomes. Properly designed and conducted prospective randomized studies are required to answer whether the beneficial effect of modified ultrafiltration on systemic inflammatory molecules associated with surgery can translate with improvement in clinical outcome.
INTRODUCTION A best evidence topic was constructed according to a structured protocol. This is fully described in the ICVTS [1].
THREE-PART QUESTION In adult patients undergoing [cardiac surgery] does [modified ultrafiltration] reduce [inflammatory response and blood loss] and influence [early outcomes].
OR ‘Ultrafiltration/epidemiology’[Mesh] OR ‘Ultrafiltration/ instrumentation’[Mesh] OR ‘Ultrafiltration/methods’[Mesh] OR ‘Ultrafiltration/mortality’[Mesh] OR ‘Ultrafiltration/organization and administration’[Mesh] OR ‘Ultrafiltration/pharmacology’[Mesh] OR ‘Ultrafiltration/standards’[Mesh] OR ‘Ultrafiltration/therapeutic use’[Mesh] OR ‘Ultrafiltration/therapy’[Mesh] OR ‘Ultrafiltration/ trends’[Mesh] OR ‘Ultrafiltration/utilization’[Mesh])) AND (‘Thoracic Surgery’[Mesh] OR ‘Cardiac Surgical Procedures’[Mesh]). The search was limited to English language articles and human studies only. This search was repeated in the Cochrane Central Register of Controlled Trials. In addition, the reference lists of each publication were searched.
CLINICAL SCENARIO
SEARCH OUTCOME
A 75-year old male patient is undergoing isolated coronary artery bypass on-pump. In theatre, the perfusionist suggests using modified ultrafiltration at the end of the operation. You are not aware of evidence supporting the use of this technique in adult patients so you carry out a literature search.
A total of 155 papers were found using the reported search. Of these, six papers provided the best evidence to answer the question. These are presented in Table 1.
SEARCH STRATEGY Medline from 1950 to June 2014 using the PubMed interface (‘Ultrafiltration/adverse effects’[Mesh] OR ‘Ultrafiltration/ classification’[Mesh] OR ‘Ultrafiltration/contraindications’[Mesh] OR ‘Ultrafiltration/economics’[Mesh] OR ‘Ultrafiltration/education’[Mesh]
RESULTS Papadopoulos et al. [2], in a single-centre prospective randomized clinical trial (RCT), investigated the benefits of normovolemic modified ultrafiltration (MUF) after cardiac surgery in a cohort of high-risk patients (mean logistic EuroSCORE II: 17.5%) undergoing different cardiac operations. This study showed a significant reduction in complement C5b9 (preoperatively: 4.18 ± 2.6 pg/ml, postoperatively: 3.05 ± 2.39 pg/ml; P = 0.006) 6 h after MUF. The
© The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
BEST EVIDENCE TOPIC
Keywords: Cardiac surgery • Modified ultrafiltration • Inflammation • Blood loss • Early outcomes
M. Zakkar et al. / Interactive CardioVascular and Thoracic Surgery
416
Table 1: Summary table Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
Key results
Comments
Papadopoulos et al. (2013), Perfusion, Germany [2]
N = 50 high-risk patients
Group 1 (control, n = 25) Group 2 (MUF, n = 25)
Inflammatory and metabolic changes: Reduction in LBP concentration (preoperatively: 23.8 ± 8.4 pg/ml, postoperatively: 14.2 ± 12.9 pg/ml, P = 0.02) and C5b9 (preoperatively: 4.18 ± 2.6 pg/ml, postoperatively: 3.05 ± 2.39 pg/ml, P = 0.006) were detected 6 h after MUF
Small prospective non-blinded study
Prospective randomized control study (level 1b)
Modified ultrafiltration (MUF) leads to a significant reduction of lipopolysaccharide-binding protein (LBP) and terminal complement complex and was associated with reduced blood loss and postoperative lactate concentrations shortly after surgery
Between April 2009 and January 2010
Mixed complex cardiac operations Blood samples were taken after the induction of anaesthesia, before CPB, before CPB weaning, 30 min after CPB and at 6, 24 and 48 h postoperatively
The MUF group had lower mean concentrations of lactate at 12 h postoperatively (N-MUF group: 12.5 ± 5.3 mmol/l, control group: 20.7 ± 7.3 mmol/l, P = 0.04) No changes noted in IL-6 and tumour necrosis factor alpha (TNF-α) levels 30 min and 6 h Blood loss: Postoperative chest tube blood
The group presenting data on high-risk patients (mean logistic EuroSCORE II: 17.5-%) The group report that these patients underwent complex cardiac operations but no further details about the nature of these operations Normovolemic MUF of 3000 ml was performed after weaning from CPB Comprehensive data presented, however, the primary outcomes were plasma levels of LBP, terminal complement complex (C5b9) and cytokines so data regarding secondary outcomes may be affected by the small numbers in the study (underpowered)
Lower in the MUF group vs the control Group at 24 h [mean: 890 (range: 500/1100) ml vs mean: 1075 (range: 800/1413) ml] (P = 0 0.039) and 48 h [mean: 900 (range: 550/1350) ml vs mean: 1400 (range: 900/1750) ml; P = 0 0.026] postoperatively Outcomes: Overall hospital mortality was 10% (n = 5) and was lower, But not significantly in the MUF group (MUF: 8%, n = 2 vs control group: 12%, n = 3, P = 0.1) No differences were observed in terms of the incidence of surgical re-exploration for bleeding (P = 1.0) or the incidence of wound infection (P = 0.15) or the incidence of acute renal failure (P = 1.0) The duration of ICU and hospital stays did not differ (ICU stay: 5.7 ± 9.21 days in the control group vs 5.1 ± 6.45 in the MUF group, P = 0.25, hospital stay: 12.2 ± 6.5 days in the control group vs 13.4 ± 9.27 days in the MUF group, P = 0.18)
Continued
M. Zakkar et al. / Interactive CardioVascular and Thoracic Surgery
417
Table 1: Continued Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
Key results
Comments
No differences noted regarding cardiac index (no P-value) or SVRI at the end of operation (P = 0.15) or 6 h later (P = 0.26) Torina et al. (2012), J Thorac Cardiovasc Surg, Brazil [3] Randomized control study (level 1b)
N = 60 patients No dates given Randomized to the MUF or control (n = 30 each)
The use of MUF was associated with increased inflammatory response, reduced blood loss and less blood transfusions in adults undergoing CABG
Coronary artery bypass grafting (CABG) cases only Data were recorded at the beginning of anaesthesia, end of bypass, end of experimental treatment and 24 and 48 h after surgery
Inflammatory and metabolic changes: No changes were noted for interleukin-6 and P-selectin Plasma levels of intercellular adhesion molecule were higher in the MUF group in the first sampling after MUF (P = 0.01) Similar trend was noted for soluble tumour necrosis factor receptor at 48 h (P < 0.001) Haematocrit level was higher in the MUF group at the end of bypass (37.8% ± 1.1 vs 34.1% ± 1.1%; P < 0.05), but the levels became comparable at 48 h
Small prospective non-blinded study. The surgical team were not blinded for group selection, but all measurements (i.e. chest tube drainage, inflammatory markers, haematocrit, transfusion decisions) were performed in a blinded manner MUF was performed for 15 min at the end of bypass with no data about volumes The study was not designed to evaluate mortality, ICU stay or hospital length of stay (underpowered)
Outcomes: No differences noted in SVR (P = 0.79) and cardiac index (P = 0.93) There were no differences in ICU (P = 0.9) and hospital stay (P = 0.8) Torina et al. (2010), Rev Bras Cir Cardiovasc, Brasil [4] Prospective randomized study (level 1b)
N = 37 patients No dates given Randomized to no MUF (n = 17) or MUF (n = 20) Only CABG surgery Data collected at beginning of anaesthesia, termination of bypass, end of MUF, 24 h and 48 h after surgery
MUF reduces postoperative bleeding and red blood cell transfusion with no differences in clinical outcomes The routine MUF employment was not associated with haemodynamic instability
Inflammatory and metabolic changes: No measurement of inflammatory mediators. Haematocrit changes noted between the two groups favouring the MUF (P < 0.05) Blood loss: Lower drain loss in the MUF group after 48 h (598 ± 123 ml vs 848 ± 455 ml; P = 0.04) The MUF group required less blood transfusion (0.6 ± 0.6 units/patient vs 1.6 ± 1.1 units/ patient; P = 0.03) Outcomes: The MUF group showed lower airway resistance (9.3 ± 0.4 cmH2O.L-1s-1 vs 12.1 ± 0.8
Small non-blinded study over an unspecified period of time No measurement of inflammatory mediators The measurements of respiratory mechanics were only performed in the intraoperative period The respiratory mechanics data are not comparable between the intraoperative period and the intensive care unit period, for the patients requiring prolonged mechanical ventilation
Continued
BEST EVIDENCE TOPIC
Blood loss: The MUF group had less chest tube drainage after 48 h (598 ± 123 vs 848.0 ± 455 ml; P = 0.04) and less red blood cell transfusions (0.6 ± 0.6 units/patient vs 1.6 ± 1.1 units/patient; P = 0.03)
M. Zakkar et al. / Interactive CardioVascular and Thoracic Surgery
418
Table 1: Continued Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
Key results
Comments
cmH2O.L-1s-1; P = 0.04). No differences noted in ICU and hospital stay between the two groups Boga et al. (2000), Perfusion, Turkey [5]
N = 40 patients No dates given
Randomized control study (level 1b)
Randomized to No MUF (n = 20) or MUF (n = 20)
MUF is associated with improved cardiac performance in post-CABG
Only CABG surgery
Inflammatory and metabolic changes: There were no differences in the immune mediators such as IL-6 and IL-8 (no P-value) The mean haematocrit level of the MUF group was higher (33.0 ± 0.33 vs 29.0 ± 0.45%; P = 0.008)
Data were recorded before and after MUF
Blood loss: There was no difference in drainage (no further data) Difference in blood transfusion was noted between the two groups favouring the MUF (0.83 ± 0.27 vs 1.84 ± 0.31; P < 0.05)
Small prospective non-blinded study MUF was carried out at termination of bypass for 10–15 min The study does not give a correlation between significant elevation of SVR and PVR levels in the MUF group and the elimination of inflammatory mediators from blood The data suggest the changes in SVR, cardiac performance is mainly haemo-concentration related
Clinical outcomes: Cardiac output, cardiac index values were greater in the MUF group (5.32 ± 0.91/min vs 4.42 ± 0.83/ min, P = 0.004, and 3.01 ± 0.47/ min/m2 vs 2.30 ± 0.46/min/m2, P < 0.0001, respectively) in the early postoperative period The SVR and PVR values of the MUF group were greater in the early postoperative period (1635.05 ± 619.9 dyn s/cm5 vs 1010.70 ± 245.63 dyn s/cm5 and 215.40 ± 65.89 dyn s/cm5 vs 154.15 ± 45.26 dyn s/cm5, respectively; P < 0.001) There was a difference in PVR favouring the MUF group at the third postoperative hour (200.60 ± 63.53 dyn s/cm5 vs 143.75 ± 43.39 dyn s/cm5 Grunenfelder et al. (2000), Eur J Cardiothorac Surg, Switzerland [6] Prospective randomized study (level 1b)
N = 97 patients July 1997 to May 1999
MUF has minimal effect in normothermia
Inflammatory and metabolic changes:
50 patients were operated on using normothermic CPB and 47 patients using hypothermic CPB
MUF can be effective in removing cytokines and adhesion molecules. However, there is no significant impact of MUF on outcomes after elective CABG
- After normothermic CPB, soluble E-selectin peaked after 24 h (mean 193 ± 93 ng/ml) in the patient group not undergoing MUF. The difference after 24 h in the patients with MUF was 35% (P < 0.007). No changes were seen after 48 h and 6 days. Soluble ICAM-1 levels in patients with MUF and without MUF differed between 5 and 10%
The normothermic group was subdivided into a group with MUF (n = 30) and a group without MUF (n = 20). The hypothermic group was subdivided into a group with MUF (N = 30) and a group without MUF (n = 17)
The study included MUF groups as a subanalysis of two groups of patients who underwent either normothermic or hypothermic CPB The study looked in detail at different inflammatory mediators The study is underpowered to detect differences in outcomes between subgroups
Continued
M. Zakkar et al. / Interactive CardioVascular and Thoracic Surgery
419
Table 1: Continued Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
Key results
Only coronary artery bypass graft patients
No changes noted in IL-6 or IL-8
MUF was instituted after CPB for 15 min
- After hypothermic CPB, MUF was associated with lowered soluble E-selectin, which decreased by 37% 24 h postoperatively from mean 302.1 ± 123.9 to 191.7 ± 184.4 ng/ml (P = 0.0063), by 40% 48 h postoperatively from mean 241.8 ± 98.9 to 150.3 ± 137.9 ng/ml (P < 0.0027)
Data were collected preoperatively, pre-MUF, in the ultrafiltrate, 24 h, 48 h and 6 days
Comments
After 6 days, E-selectin levels were almost reduced to baseline from mean 146.2 ± 62.5 to 98 ± 73.5 ng/ml (P < 0.0084) Soluble ICAM-1 reduced by MUF after 24 h with a reduction of 43% from mean 575.9 ± 212.9 to 333.5 ± 143.8 ng/ml (P < 0.0001), by 44% after 48 h from 675.7 ± 269.7 to 383.6 ± 123.9 (P < 0.0001) and by 60% from 553.5 ± 26.9 to 232.2 ± 153.4 ng/ml (P < 0.0001) 6 days postoperatively
BEST EVIDENCE TOPIC
Interleukin-6 levels decreased by 60% from mean 792.9 ± 580.4 to 318.4 ± 393.4 ng/l (P < 0.0018). After 48 h and 6 days changes in IL-6 after MUF were noticeable, but not statistically significant Similar results were noted with IL-8 after 24 h (P < 0.0004) and in TNF-α after 24 h (P < 0.0003) without significance after 48 h or 6 days Blood loss: No differences between the two groups (no P-value) Outcomes: There were no differences in the clinical variables such as inotropic support, length of intubation and ICU and hospital stay between groups (no P-values) Onoe et al. (2001), Perfusion, Japan [7] Case control (level 4)
N = 18 patients June 1996 to June 1997 9 patients who underwent different cardiac surgery with CPB and MUF were compared
MUF has beneficial effects on postoperative haemodynamics, and can reduce serum IL-8 levels in adult cardiac surgery
Inflammatory and metabolic changes: In the MUF group, serum IL-8 was reduced from 69.5 ± 33.5 to 58.9 ± 32.4 pg/ml after MUF (P = 0.0029), whereas it was not reduced in
Small study that compared with groups of patients Results presented can be biased considering that the groups compared did not
Continued
M. Zakkar et al. / Interactive CardioVascular and Thoracic Surgery
420
Table 1: Continued Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
with 9 control patients who underwent surgery without MUF in the same period MUF was performed, after weaning off CPB for 15 min Parameters recorded at the start of CPB, immediately after CPB, immediately after MUF and 3 h after MUF
Key results
Comments
the control group
undergo the same operation with significant difference in CPB time
In the MUF group, haematocrit increased significantly from 21.2 ± 2.0 to 24.9 ± 3.3% (P = 0.0008), no such changes in the control group Blood loss: No data Outcomes: There were no differences in time to extubation or in doses of inotropes administered postoperatively Systolic blood pressure increased from 97.5 ± 16.7 to 116.5 ± 23.9 mmHg (P = 0.0024) after MUF. No changes in blood pressure in the control group
postoperative chest tube drainage was significantly lower in the MUF group at 24 h [mean: 890 (range: 500/1100) ml vs mean: 1075 (range: 800/1413) ml; P = 0.039] and 48 h [mean: 900 (range: 550/1350) ml vs mean: 1400 (range: 900/1750) ml; P = 0.026] postoperatively. The study showed no significant changes with regard to cardiac index or systemic vascular resistance index (SVRI) at end of operation or 6 h later. The incidence of surgical re-exploration for bleeding or acute renal failure did not differ among groups. Similarly, no differences noted in intensive care unit (ICU) or hospital stay. Torina et al. [3] randomized 60 patients undergoing coronary artery bypass grafting (CABG) to either MUF (n = 30) or control (n = 30). The study showed that the MUF group had reduced chest tube drainage after 48 h (598 ± 123 ml vs 848.0 ± 455 ml; P < 0.04) and required less blood transfusion (0.6 ± 0.6 units/patient vs 1.6 ± 1.1 units/patient; P < 0.03). Haematocrit levels were higher in the MUF group than in the control group at the end of bypass (37.8% ± 1.1 vs 34.1% ± 1.1%; P < 0.05). Plasma levels of intercellular adhesion molecule-1 (ICAM-1) were higher in the MUF group than in the control group, particularly in the first sampling (P < 0.01). Plasma levels of soluble tumour necrosis factor receptor were higher in the MUF group than in the control group at 48 h. In this RCT, haemodynamic and oxygen transport parameters were similar in both groups throughout the observation period with no differences in clinical outcomes. These results expanded and confirmed previous observations by the same group in an earlier prospective RCT of 37 patients who underwent CABG [4]. The design and primary end point investigated in the studies above, however, meant that they are underpowered to detect differences in clinical outcomes. Boga et al. [5] in an RCT of 40 patients undergoing CABG divided equally into two groups found that haemodynamic parameters and immune mediator levels did not differ between the
two groups during the course of the study, except in the immediate postoperative periods, where cardiac output and cardiac index values were significantly greater in the MUF group (5.32 ± 0.91/ min vs 4.42 ± 0.83/min, P = 0.004, and 3.0 ± 0.47/min/m 2 vs 2.30 ± 0.46/min/m2, P < 0.0001, respectively) in the early postoperative period. Systemic vascular resistance (SVR) and pulmonary vascular resistance (PVR) were significantly greater in the MUF group in the early postoperative period (1635.05 ± 619.9 dyn s/cm5 vs 1010.70 ± 245.63 dyn s/cm5 and 215.40 ± 65.89 dyn s/cm5 vs 154.15 ± 45.26 dyn s/cm5, respectively; P < 0.001). The result of this study suggests however, that such early improvement of haemodynamic is related to haemoconcentration only after MUF. Grünenfelder et al. [6] in an RCT that compared MUF under normo- or hypothermic cardiopulmonary bypass (CPB), showed that MUF has minimum impact under normothermic conditions. In this group, MUF was associated with a 35% reduction in E-selectin levels after 24 h (P < 0.007) with no significant changes 48 h or 6 days postoperatively. There were mild changes in soluble ICAM-1 levels (5–10%) favouring MUF with no significant changes in IL-6 or IL-8. More pronounced modifications were noted when using hypothermic CPB. In this group, MUF was associated with a 37% reduction in soluble E-selectin 24 h postoperatively favouring MUF (P = 0.0063). These changes continued to be significant 48 h (P < 0.0027) and even 6 days postoperatively (P < 0.0084). Soluble ICAM-1 levels were altered significantly by MUF with a reduction of 43% after 24 h (P < 0.0001), 44% after 48 h (P < 0.0001) and 60% (P < 0.0001) 6 days postoperatively. IL-6 levels decreased by 60% 24 h post-surgery in the MUF group (P < 0.0018). Similar results were noted with IL-8 (P < 0.0004) and in tumour necrosis factor alpha (P < 0.0003) after 24 h. However, no significant differences in the clinical variables such as inotropic support, length of intubation and ICU and hospital stay were noted.
M. Zakkar et al. / Interactive CardioVascular and Thoracic Surgery
The impact of MUF on adult patients undergoing cardiac surgery remains unclear. There is some evidence suggesting that it can lead to significant reduction in circulating inflammatory mediators and reduces blood loss and transfusion requirements. However, the studies tend to originate from a single centre including small number of elective CABG patients, do not reflect the real-world situation as patients undergoing more complex operations with long CPB time are usually the ones to suffer the adverse effects of CPB and may in theory benefit most from MUF. A well-powered multi-centre study will be required to address the impact of MUF on clinical outcomes in all-comers’ real-world cardiac surgical cohort. Conflict of interest: none declared.
REFERENCES [1] Dunning J, Prendergast B, kway-Jones K. Towards evidence-based medicine in cardiothoracic surgery: best BETS. Interact CardioVasc Thorac Surg 2003; 2:405–9.
[2] Papadopoulos N, Bakhtiary F, Grün V, Weber CF, Strasser C, Moritz A. The effect of normovolemic modified ultrafiltration on inflammatory mediators, endotoxins, terminal complement complexes and clinical outcome in high-risk cardiacsurgery patients. Perfusion 2013;28: 306–14. [3] Torina AG, Silveira-Filho LM, Vilarinho KA, Eghtesady P, Oliveira PP, Sposito AC et al. Use of modified ultrafiltration in adults undergoing coronary artery bypass grafting is associated with inflammatory modulation and less postoperative blood loss: a randomized and controlled study. J Thorac Cardiovasc Surg 2012;144:663–70. [4] Torina AG, Petrucci O, Oliveira PP, Severino ES, Vilarinho KA, Lavagnoli CF et al. The effects of modified ultrafiltration on pulmonary function and transfusion requirements in patients underwent coronary artery bypass graft surgery. Rev Bras Cir Cardiovasc 2010;25: 59–65. [5] Bogă M, Islamoglu BI, Cikirikçioglu M, Bakalim T, Ya gdi T, Büket S et al. The effects of modified hemofiltration on inflammatory mediators and cardiacperformance in coronary artery bypass grafting. Perfusion 2000; 15:143–50. [6] Grünenfelder J, Zünd G, Schoeberlein A, Maly FE, Schurr U, Guntli S et al. Modified ultrafiltration lowers adhesion molecule and cytokine levels after cardiopulmonary bypass without clinical relevance in adults. Eur J Cardiothorac Surg 2000;17:77–83. [7] Onoe M, Magara T, Yamamoto Y, Nojima T. Modified ultrafiltration removes serum interleukin-8 in adult cardiac surgery. Perfusion 2001;16: 37–42.
BEST EVIDENCE TOPIC
CLINICAL BOTTOM LINE
421
Interactive CardioVascular and Thoracic Surgery 20 (2015) 415–421 doi:10.1093/icvts/ivu388 Advance Access publication 25 November 2014
Modified ultrafiltration in adult patients undergoing cardiac surgery Mustafa Zakkar*, Gustavo Guida and Gianni D. Angelini Bristol Heart Institute, Bristol, UK * Corresponding author. Bristol Heart Institute, Bristol Royal infirmary, Bristol BS2 8HW, UK. Tel: +44-117-9230000; e-mail: [email protected] (M. Zakkar). Received 6 August 2014; received in revised form 24 October 2014; accepted 28 October 2014
Summary A best evidence topic in cardiac surgery was written according to a structured protocol. The question addressed was the impact of modified ultrafiltration on adult patients undergoing cardiac surgery in terms of inflammatory and metabolic changes, blood loss and early clinical outcomes. A total of 155 papers were identified using the search as described below. Of these, six papers presented the best evidence to answer the clinical question as they reported data to reach conclusions regarding the issues of interest for this review. The author, date and country of publication, patient group, study type and weaknesses and relevant outcomes were tabulated. Modified ultrafiltration in adult patients undergoing cardiac surgery seems to attenuate the levels of inflammatory molecules associated with surgery, reduces blood loss and blood transfusion and improves cardiac output, index and systemic vascular resistance. However, this was not translated in any reduction in length of stay in intensive care unit or hospital. Most studies were single-centre prospective non-blinded trials that included a small cohort of elective coronary artery bypass grafting patients, which makes it underpowered to provide unbiased evidence regarding clinical outcomes. Properly designed and conducted prospective randomized studies are required to answer whether the beneficial effect of modified ultrafiltration on systemic inflammatory molecules associated with surgery can translate with improvement in clinical outcome.
INTRODUCTION A best evidence topic was constructed according to a structured protocol. This is fully described in the ICVTS [1].
THREE-PART QUESTION In adult patients undergoing [cardiac surgery] does [modified ultrafiltration] reduce [inflammatory response and blood loss] and influence [early outcomes].
OR ‘Ultrafiltration/epidemiology’[Mesh] OR ‘Ultrafiltration/ instrumentation’[Mesh] OR ‘Ultrafiltration/methods’[Mesh] OR ‘Ultrafiltration/mortality’[Mesh] OR ‘Ultrafiltration/organization and administration’[Mesh] OR ‘Ultrafiltration/pharmacology’[Mesh] OR ‘Ultrafiltration/standards’[Mesh] OR ‘Ultrafiltration/therapeutic use’[Mesh] OR ‘Ultrafiltration/therapy’[Mesh] OR ‘Ultrafiltration/ trends’[Mesh] OR ‘Ultrafiltration/utilization’[Mesh])) AND (‘Thoracic Surgery’[Mesh] OR ‘Cardiac Surgical Procedures’[Mesh]). The search was limited to English language articles and human studies only. This search was repeated in the Cochrane Central Register of Controlled Trials. In addition, the reference lists of each publication were searched.
CLINICAL SCENARIO
SEARCH OUTCOME
A 75-year old male patient is undergoing isolated coronary artery bypass on-pump. In theatre, the perfusionist suggests using modified ultrafiltration at the end of the operation. You are not aware of evidence supporting the use of this technique in adult patients so you carry out a literature search.
A total of 155 papers were found using the reported search. Of these, six papers provided the best evidence to answer the question. These are presented in Table 1.
SEARCH STRATEGY Medline from 1950 to June 2014 using the PubMed interface (‘Ultrafiltration/adverse effects’[Mesh] OR ‘Ultrafiltration/ classification’[Mesh] OR ‘Ultrafiltration/contraindications’[Mesh] OR ‘Ultrafiltration/economics’[Mesh] OR ‘Ultrafiltration/education’[Mesh]
RESULTS Papadopoulos et al. [2], in a single-centre prospective randomized clinical trial (RCT), investigated the benefits of normovolemic modified ultrafiltration (MUF) after cardiac surgery in a cohort of high-risk patients (mean logistic EuroSCORE II: 17.5%) undergoing different cardiac operations. This study showed a significant reduction in complement C5b9 (preoperatively: 4.18 ± 2.6 pg/ml, postoperatively: 3.05 ± 2.39 pg/ml; P = 0.006) 6 h after MUF. The
© The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
BEST EVIDENCE TOPIC
Keywords: Cardiac surgery • Modified ultrafiltration • Inflammation • Blood loss • Early outcomes
M. Zakkar et al. / Interactive CardioVascular and Thoracic Surgery
416
Table 1: Summary table Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
Key results
Comments
Papadopoulos et al. (2013), Perfusion, Germany [2]
N = 50 high-risk patients
Group 1 (control, n = 25) Group 2 (MUF, n = 25)
Inflammatory and metabolic changes: Reduction in LBP concentration (preoperatively: 23.8 ± 8.4 pg/ml, postoperatively: 14.2 ± 12.9 pg/ml, P = 0.02) and C5b9 (preoperatively: 4.18 ± 2.6 pg/ml, postoperatively: 3.05 ± 2.39 pg/ml, P = 0.006) were detected 6 h after MUF
Small prospective non-blinded study
Prospective randomized control study (level 1b)
Modified ultrafiltration (MUF) leads to a significant reduction of lipopolysaccharide-binding protein (LBP) and terminal complement complex and was associated with reduced blood loss and postoperative lactate concentrations shortly after surgery
Between April 2009 and January 2010
Mixed complex cardiac operations Blood samples were taken after the induction of anaesthesia, before CPB, before CPB weaning, 30 min after CPB and at 6, 24 and 48 h postoperatively
The MUF group had lower mean concentrations of lactate at 12 h postoperatively (N-MUF group: 12.5 ± 5.3 mmol/l, control group: 20.7 ± 7.3 mmol/l, P = 0.04) No changes noted in IL-6 and tumour necrosis factor alpha (TNF-α) levels 30 min and 6 h Blood loss: Postoperative chest tube blood
The group presenting data on high-risk patients (mean logistic EuroSCORE II: 17.5-%) The group report that these patients underwent complex cardiac operations but no further details about the nature of these operations Normovolemic MUF of 3000 ml was performed after weaning from CPB Comprehensive data presented, however, the primary outcomes were plasma levels of LBP, terminal complement complex (C5b9) and cytokines so data regarding secondary outcomes may be affected by the small numbers in the study (underpowered)
Lower in the MUF group vs the control Group at 24 h [mean: 890 (range: 500/1100) ml vs mean: 1075 (range: 800/1413) ml] (P = 0 0.039) and 48 h [mean: 900 (range: 550/1350) ml vs mean: 1400 (range: 900/1750) ml; P = 0 0.026] postoperatively Outcomes: Overall hospital mortality was 10% (n = 5) and was lower, But not significantly in the MUF group (MUF: 8%, n = 2 vs control group: 12%, n = 3, P = 0.1) No differences were observed in terms of the incidence of surgical re-exploration for bleeding (P = 1.0) or the incidence of wound infection (P = 0.15) or the incidence of acute renal failure (P = 1.0) The duration of ICU and hospital stays did not differ (ICU stay: 5.7 ± 9.21 days in the control group vs 5.1 ± 6.45 in the MUF group, P = 0.25, hospital stay: 12.2 ± 6.5 days in the control group vs 13.4 ± 9.27 days in the MUF group, P = 0.18)
Continued
M. Zakkar et al. / Interactive CardioVascular and Thoracic Surgery
417
Table 1: Continued Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
Key results
Comments
No differences noted regarding cardiac index (no P-value) or SVRI at the end of operation (P = 0.15) or 6 h later (P = 0.26) Torina et al. (2012), J Thorac Cardiovasc Surg, Brazil [3] Randomized control study (level 1b)
N = 60 patients No dates given Randomized to the MUF or control (n = 30 each)
The use of MUF was associated with increased inflammatory response, reduced blood loss and less blood transfusions in adults undergoing CABG
Coronary artery bypass grafting (CABG) cases only Data were recorded at the beginning of anaesthesia, end of bypass, end of experimental treatment and 24 and 48 h after surgery
Inflammatory and metabolic changes: No changes were noted for interleukin-6 and P-selectin Plasma levels of intercellular adhesion molecule were higher in the MUF group in the first sampling after MUF (P = 0.01) Similar trend was noted for soluble tumour necrosis factor receptor at 48 h (P < 0.001) Haematocrit level was higher in the MUF group at the end of bypass (37.8% ± 1.1 vs 34.1% ± 1.1%; P < 0.05), but the levels became comparable at 48 h
Small prospective non-blinded study. The surgical team were not blinded for group selection, but all measurements (i.e. chest tube drainage, inflammatory markers, haematocrit, transfusion decisions) were performed in a blinded manner MUF was performed for 15 min at the end of bypass with no data about volumes The study was not designed to evaluate mortality, ICU stay or hospital length of stay (underpowered)
Outcomes: No differences noted in SVR (P = 0.79) and cardiac index (P = 0.93) There were no differences in ICU (P = 0.9) and hospital stay (P = 0.8) Torina et al. (2010), Rev Bras Cir Cardiovasc, Brasil [4] Prospective randomized study (level 1b)
N = 37 patients No dates given Randomized to no MUF (n = 17) or MUF (n = 20) Only CABG surgery Data collected at beginning of anaesthesia, termination of bypass, end of MUF, 24 h and 48 h after surgery
MUF reduces postoperative bleeding and red blood cell transfusion with no differences in clinical outcomes The routine MUF employment was not associated with haemodynamic instability
Inflammatory and metabolic changes: No measurement of inflammatory mediators. Haematocrit changes noted between the two groups favouring the MUF (P < 0.05) Blood loss: Lower drain loss in the MUF group after 48 h (598 ± 123 ml vs 848 ± 455 ml; P = 0.04) The MUF group required less blood transfusion (0.6 ± 0.6 units/patient vs 1.6 ± 1.1 units/ patient; P = 0.03) Outcomes: The MUF group showed lower airway resistance (9.3 ± 0.4 cmH2O.L-1s-1 vs 12.1 ± 0.8
Small non-blinded study over an unspecified period of time No measurement of inflammatory mediators The measurements of respiratory mechanics were only performed in the intraoperative period The respiratory mechanics data are not comparable between the intraoperative period and the intensive care unit period, for the patients requiring prolonged mechanical ventilation
Continued
BEST EVIDENCE TOPIC
Blood loss: The MUF group had less chest tube drainage after 48 h (598 ± 123 vs 848.0 ± 455 ml; P = 0.04) and less red blood cell transfusions (0.6 ± 0.6 units/patient vs 1.6 ± 1.1 units/patient; P = 0.03)
M. Zakkar et al. / Interactive CardioVascular and Thoracic Surgery
418
Table 1: Continued Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
Key results
Comments
cmH2O.L-1s-1; P = 0.04). No differences noted in ICU and hospital stay between the two groups Boga et al. (2000), Perfusion, Turkey [5]
N = 40 patients No dates given
Randomized control study (level 1b)
Randomized to No MUF (n = 20) or MUF (n = 20)
MUF is associated with improved cardiac performance in post-CABG
Only CABG surgery
Inflammatory and metabolic changes: There were no differences in the immune mediators such as IL-6 and IL-8 (no P-value) The mean haematocrit level of the MUF group was higher (33.0 ± 0.33 vs 29.0 ± 0.45%; P = 0.008)
Data were recorded before and after MUF
Blood loss: There was no difference in drainage (no further data) Difference in blood transfusion was noted between the two groups favouring the MUF (0.83 ± 0.27 vs 1.84 ± 0.31; P < 0.05)
Small prospective non-blinded study MUF was carried out at termination of bypass for 10–15 min The study does not give a correlation between significant elevation of SVR and PVR levels in the MUF group and the elimination of inflammatory mediators from blood The data suggest the changes in SVR, cardiac performance is mainly haemo-concentration related
Clinical outcomes: Cardiac output, cardiac index values were greater in the MUF group (5.32 ± 0.91/min vs 4.42 ± 0.83/ min, P = 0.004, and 3.01 ± 0.47/ min/m2 vs 2.30 ± 0.46/min/m2, P < 0.0001, respectively) in the early postoperative period The SVR and PVR values of the MUF group were greater in the early postoperative period (1635.05 ± 619.9 dyn s/cm5 vs 1010.70 ± 245.63 dyn s/cm5 and 215.40 ± 65.89 dyn s/cm5 vs 154.15 ± 45.26 dyn s/cm5, respectively; P < 0.001) There was a difference in PVR favouring the MUF group at the third postoperative hour (200.60 ± 63.53 dyn s/cm5 vs 143.75 ± 43.39 dyn s/cm5 Grunenfelder et al. (2000), Eur J Cardiothorac Surg, Switzerland [6] Prospective randomized study (level 1b)
N = 97 patients July 1997 to May 1999
MUF has minimal effect in normothermia
Inflammatory and metabolic changes:
50 patients were operated on using normothermic CPB and 47 patients using hypothermic CPB
MUF can be effective in removing cytokines and adhesion molecules. However, there is no significant impact of MUF on outcomes after elective CABG
- After normothermic CPB, soluble E-selectin peaked after 24 h (mean 193 ± 93 ng/ml) in the patient group not undergoing MUF. The difference after 24 h in the patients with MUF was 35% (P < 0.007). No changes were seen after 48 h and 6 days. Soluble ICAM-1 levels in patients with MUF and without MUF differed between 5 and 10%
The normothermic group was subdivided into a group with MUF (n = 30) and a group without MUF (n = 20). The hypothermic group was subdivided into a group with MUF (N = 30) and a group without MUF (n = 17)
The study included MUF groups as a subanalysis of two groups of patients who underwent either normothermic or hypothermic CPB The study looked in detail at different inflammatory mediators The study is underpowered to detect differences in outcomes between subgroups
Continued
M. Zakkar et al. / Interactive CardioVascular and Thoracic Surgery
419
Table 1: Continued Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
Key results
Only coronary artery bypass graft patients
No changes noted in IL-6 or IL-8
MUF was instituted after CPB for 15 min
- After hypothermic CPB, MUF was associated with lowered soluble E-selectin, which decreased by 37% 24 h postoperatively from mean 302.1 ± 123.9 to 191.7 ± 184.4 ng/ml (P = 0.0063), by 40% 48 h postoperatively from mean 241.8 ± 98.9 to 150.3 ± 137.9 ng/ml (P < 0.0027)
Data were collected preoperatively, pre-MUF, in the ultrafiltrate, 24 h, 48 h and 6 days
Comments
After 6 days, E-selectin levels were almost reduced to baseline from mean 146.2 ± 62.5 to 98 ± 73.5 ng/ml (P < 0.0084) Soluble ICAM-1 reduced by MUF after 24 h with a reduction of 43% from mean 575.9 ± 212.9 to 333.5 ± 143.8 ng/ml (P < 0.0001), by 44% after 48 h from 675.7 ± 269.7 to 383.6 ± 123.9 (P < 0.0001) and by 60% from 553.5 ± 26.9 to 232.2 ± 153.4 ng/ml (P < 0.0001) 6 days postoperatively
BEST EVIDENCE TOPIC
Interleukin-6 levels decreased by 60% from mean 792.9 ± 580.4 to 318.4 ± 393.4 ng/l (P < 0.0018). After 48 h and 6 days changes in IL-6 after MUF were noticeable, but not statistically significant Similar results were noted with IL-8 after 24 h (P < 0.0004) and in TNF-α after 24 h (P < 0.0003) without significance after 48 h or 6 days Blood loss: No differences between the two groups (no P-value) Outcomes: There were no differences in the clinical variables such as inotropic support, length of intubation and ICU and hospital stay between groups (no P-values) Onoe et al. (2001), Perfusion, Japan [7] Case control (level 4)
N = 18 patients June 1996 to June 1997 9 patients who underwent different cardiac surgery with CPB and MUF were compared
MUF has beneficial effects on postoperative haemodynamics, and can reduce serum IL-8 levels in adult cardiac surgery
Inflammatory and metabolic changes: In the MUF group, serum IL-8 was reduced from 69.5 ± 33.5 to 58.9 ± 32.4 pg/ml after MUF (P = 0.0029), whereas it was not reduced in
Small study that compared with groups of patients Results presented can be biased considering that the groups compared did not
Continued
M. Zakkar et al. / Interactive CardioVascular and Thoracic Surgery
420
Table 1: Continued Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
with 9 control patients who underwent surgery without MUF in the same period MUF was performed, after weaning off CPB for 15 min Parameters recorded at the start of CPB, immediately after CPB, immediately after MUF and 3 h after MUF
Key results
Comments
the control group
undergo the same operation with significant difference in CPB time
In the MUF group, haematocrit increased significantly from 21.2 ± 2.0 to 24.9 ± 3.3% (P = 0.0008), no such changes in the control group Blood loss: No data Outcomes: There were no differences in time to extubation or in doses of inotropes administered postoperatively Systolic blood pressure increased from 97.5 ± 16.7 to 116.5 ± 23.9 mmHg (P = 0.0024) after MUF. No changes in blood pressure in the control group
postoperative chest tube drainage was significantly lower in the MUF group at 24 h [mean: 890 (range: 500/1100) ml vs mean: 1075 (range: 800/1413) ml; P = 0.039] and 48 h [mean: 900 (range: 550/1350) ml vs mean: 1400 (range: 900/1750) ml; P = 0.026] postoperatively. The study showed no significant changes with regard to cardiac index or systemic vascular resistance index (SVRI) at end of operation or 6 h later. The incidence of surgical re-exploration for bleeding or acute renal failure did not differ among groups. Similarly, no differences noted in intensive care unit (ICU) or hospital stay. Torina et al. [3] randomized 60 patients undergoing coronary artery bypass grafting (CABG) to either MUF (n = 30) or control (n = 30). The study showed that the MUF group had reduced chest tube drainage after 48 h (598 ± 123 ml vs 848.0 ± 455 ml; P < 0.04) and required less blood transfusion (0.6 ± 0.6 units/patient vs 1.6 ± 1.1 units/patient; P < 0.03). Haematocrit levels were higher in the MUF group than in the control group at the end of bypass (37.8% ± 1.1 vs 34.1% ± 1.1%; P < 0.05). Plasma levels of intercellular adhesion molecule-1 (ICAM-1) were higher in the MUF group than in the control group, particularly in the first sampling (P < 0.01). Plasma levels of soluble tumour necrosis factor receptor were higher in the MUF group than in the control group at 48 h. In this RCT, haemodynamic and oxygen transport parameters were similar in both groups throughout the observation period with no differences in clinical outcomes. These results expanded and confirmed previous observations by the same group in an earlier prospective RCT of 37 patients who underwent CABG [4]. The design and primary end point investigated in the studies above, however, meant that they are underpowered to detect differences in clinical outcomes. Boga et al. [5] in an RCT of 40 patients undergoing CABG divided equally into two groups found that haemodynamic parameters and immune mediator levels did not differ between the
two groups during the course of the study, except in the immediate postoperative periods, where cardiac output and cardiac index values were significantly greater in the MUF group (5.32 ± 0.91/ min vs 4.42 ± 0.83/min, P = 0.004, and 3.0 ± 0.47/min/m 2 vs 2.30 ± 0.46/min/m2, P < 0.0001, respectively) in the early postoperative period. Systemic vascular resistance (SVR) and pulmonary vascular resistance (PVR) were significantly greater in the MUF group in the early postoperative period (1635.05 ± 619.9 dyn s/cm5 vs 1010.70 ± 245.63 dyn s/cm5 and 215.40 ± 65.89 dyn s/cm5 vs 154.15 ± 45.26 dyn s/cm5, respectively; P < 0.001). The result of this study suggests however, that such early improvement of haemodynamic is related to haemoconcentration only after MUF. Grünenfelder et al. [6] in an RCT that compared MUF under normo- or hypothermic cardiopulmonary bypass (CPB), showed that MUF has minimum impact under normothermic conditions. In this group, MUF was associated with a 35% reduction in E-selectin levels after 24 h (P < 0.007) with no significant changes 48 h or 6 days postoperatively. There were mild changes in soluble ICAM-1 levels (5–10%) favouring MUF with no significant changes in IL-6 or IL-8. More pronounced modifications were noted when using hypothermic CPB. In this group, MUF was associated with a 37% reduction in soluble E-selectin 24 h postoperatively favouring MUF (P = 0.0063). These changes continued to be significant 48 h (P < 0.0027) and even 6 days postoperatively (P < 0.0084). Soluble ICAM-1 levels were altered significantly by MUF with a reduction of 43% after 24 h (P < 0.0001), 44% after 48 h (P < 0.0001) and 60% (P < 0.0001) 6 days postoperatively. IL-6 levels decreased by 60% 24 h post-surgery in the MUF group (P < 0.0018). Similar results were noted with IL-8 (P < 0.0004) and in tumour necrosis factor alpha (P < 0.0003) after 24 h. However, no significant differences in the clinical variables such as inotropic support, length of intubation and ICU and hospital stay were noted.
M. Zakkar et al. / Interactive CardioVascular and Thoracic Surgery
The impact of MUF on adult patients undergoing cardiac surgery remains unclear. There is some evidence suggesting that it can lead to significant reduction in circulating inflammatory mediators and reduces blood loss and transfusion requirements. However, the studies tend to originate from a single centre including small number of elective CABG patients, do not reflect the real-world situation as patients undergoing more complex operations with long CPB time are usually the ones to suffer the adverse effects of CPB and may in theory benefit most from MUF. A well-powered multi-centre study will be required to address the impact of MUF on clinical outcomes in all-comers’ real-world cardiac surgical cohort. Conflict of interest: none declared.
REFERENCES [1] Dunning J, Prendergast B, kway-Jones K. Towards evidence-based medicine in cardiothoracic surgery: best BETS. Interact CardioVasc Thorac Surg 2003; 2:405–9.
[2] Papadopoulos N, Bakhtiary F, Grün V, Weber CF, Strasser C, Moritz A. The effect of normovolemic modified ultrafiltration on inflammatory mediators, endotoxins, terminal complement complexes and clinical outcome in high-risk cardiacsurgery patients. Perfusion 2013;28: 306–14. [3] Torina AG, Silveira-Filho LM, Vilarinho KA, Eghtesady P, Oliveira PP, Sposito AC et al. Use of modified ultrafiltration in adults undergoing coronary artery bypass grafting is associated with inflammatory modulation and less postoperative blood loss: a randomized and controlled study. J Thorac Cardiovasc Surg 2012;144:663–70. [4] Torina AG, Petrucci O, Oliveira PP, Severino ES, Vilarinho KA, Lavagnoli CF et al. The effects of modified ultrafiltration on pulmonary function and transfusion requirements in patients underwent coronary artery bypass graft surgery. Rev Bras Cir Cardiovasc 2010;25: 59–65. [5] Bogă M, Islamoglu BI, Cikirikçioglu M, Bakalim T, Ya gdi T, Büket S et al. The effects of modified hemofiltration on inflammatory mediators and cardiacperformance in coronary artery bypass grafting. Perfusion 2000; 15:143–50. [6] Grünenfelder J, Zünd G, Schoeberlein A, Maly FE, Schurr U, Guntli S et al. Modified ultrafiltration lowers adhesion molecule and cytokine levels after cardiopulmonary bypass without clinical relevance in adults. Eur J Cardiothorac Surg 2000;17:77–83. [7] Onoe M, Magara T, Yamamoto Y, Nojima T. Modified ultrafiltration removes serum interleukin-8 in adult cardiac surgery. Perfusion 2001;16: 37–42.
BEST EVIDENCE TOPIC
CLINICAL BOTTOM LINE
421
