1. Albumin: Should It Be Used In Clinical Practice? Presented By: Paul Hebert

2. What is Albumin? Description Human plasma protein Human plasma protein Molecular weight of 66 Kd Molecular weight of 66 Kd Most common plasma protein Most common plasma protein Synthesized in the liver Synthesized in the liver Negatively charged Negatively chargedFunction Responsible for oncotic Responsible for oncoticpressure Binds drugs and other Binds drugs and othersubstances Free radical scavenger Free radical scavenger

3. What do we use Albumin for? Treatment of Hypovolemia Treatment of Hypovolemia Burns Burns Nutritional replacement with TPN Nutritional replacement with TPN Hypoalbuminemia Hypoalbuminemia Hyperoncotic therapy Hyperoncotic therapy

4. Indicated Following large volume paracentesis Nephrotic syndrome resistant to potent diuretics Volume/Fluid replacement in plasmapheresis Possibly indicated Adult respiratory distress syndrome Ovarian hyperstimulation syndrome Cardiopulmonary bypass pump priming Fluid resuscitation in shock/sepsis/burns Neonatal kernicterus To improve enteral feeding intolerance Not indicated Correction of measured hypoalbuminemia or hypoproteinemia Nutritional deficiency, total parenteral nutrition Pre-eclampsia Red blood cell suspension Simple volume expansion (surgery, burns) Wound healing Bucur et al. Hematology:Basic Principles and Practice. 2000; 2266 What do we use Albumin for?

5. Investigational Cadaveric renal transplantation Cerebral ischemia Stroke Common Usages Serum albumin <2.0 g/dl Nephrotic syndrome, proteinuria and hypoalbuminemia Labile pulmonary, cardiovascular status Cardiopulmonary bypass, pump priming Extensive burns Plasma exchange Hypotension Liver disease, hypoalbuminemia, diuresis Protein losing enteropathy, hypoalbuminemia Resuscitation Intraoperative fluid requirement > 5-6 L in adults Premature infant undergoing major surgery Bucur et al. Hematology:Basic Principles and Practice. 2000; 2266 What do we use Albumin for?

6. Back to Basics

7. Intracellular Extracellular Compartment Plasma 40% 60% General Schema Interstitial

8. Intracellular EC Plasma What happens when you infuse H 2 O? EC Plasma Interstitial Intracellular

9. EC Plasma EC Plasma What happens when you infuse 0.9% Saline in health? Interstitial Intracellular

10. EC Plasma What happens when you infuse 0.9% saline to edematous patients? Marked edema EC Plasma Intracellular Interstitial

11. EC Plasma EC Plasma What happens when you infuse 5% Albumin (Iso-Oncotic Colloid)? Interstitial Intracellular

12. EC Plasma EC Plasma What is the effect of hypertonic saline ? Intracellular Interstitial

13. EC Plasma EC Plasma Effect of “Hyper”-Oncotic Colloid ie 25% Albumin Interstitial Intracellular

14. 30 RCTs in systematic review 30 RCTs in systematic review 1419 critically ill patients 1419 critically ill patients Indications included hypovolemia, burns and hypoalbuminemia Indications included hypovolemia, burns and hypoalbuminemia All doses and concentrations of albumin (2.5, 4%, 5% and 25%) All doses and concentrations of albumin (2.5, 4%, 5% and 25%) Any control group (nothing, saline, Ringers, dextrose/Ringers) Any control group (nothing, saline, Ringers, dextrose/Ringers) No protocols of care No protocols of care Limited assessment of quality Limited assessment of quality Cochrane Injuries Group Albumin Reviewers, BMJ 1998;317:235-240 The controversy?...Albumin revisited

15. Copyright ©1998 BMJ Publishing Group Ltd. Cochrane Injuries Group Albumin Reviewers, BMJ 1998;317:235-240 The controversy?...Albumin revisited Favors control Favors Albumin

16. Copyright ©1998 BMJ Publishing Group Ltd. The controversy?...Albumin revisited Cochrane Injuries Group Albumin Reviewers, BMJ 1998;317:235-240 Favors AlbuminFavors Control

17. Schierhout and Roberts. BMJ 1998;316:961-964 The controversy?...Colloids versus crystalloids Types of trials: 37 RCTs (n=1622) –Excluded 11 RCTs in systematic review –Mortality information on 1315 patients in 19 RCTs Patients: –All critically ill patients requiring volume replacement –Trauma, surgery, Burn, Sepsis, ARDS, Interventions: Any colloid (2.5% and 5% and 25% albumin, pentaspan, Dextran-70, 6% Dextran, Hydroxyethyl starch, Haemacell,plasma and combination –Colloid in hypertonic (n=10 trials) –Controls included Ringers,.9% and 7.5% saline, 5% dextrose) –No protocols of care Methods: Methods: –Fixed effect models –Limited assessment of quality

18. Copyright ©1998 BMJ Publishing Group Ltd. Schierhout and Roberts. BMJ 1998;316:961-964 The controversy?...Colloids versus crystalloids Favors colloids Favors crystalloids

19. Copyright ©1998 BMJ Publishing Group Ltd. Schierhout and Roberts. BMJ 1998;316:961-964 The controversy?...Colloids versus crystalloids

20. Inferences by Authors “No evidence supporting that albumin administration reduces mortality” “No evidence supporting that albumin administration reduces mortality” “Should not be used outside the context of rigorously conducted RCTs” “Should not be used outside the context of rigorously conducted RCTs” “Resuscitation with colloid solutions was associated with an absolute increase in the risk of mortality of 4%” “Resuscitation with colloid solutions was associated with an absolute increase in the risk of mortality of 4%” Inferences supported by BMJ Editorials Inferences supported by BMJ Editorials

21. But…Significant Limitations with meta-analyses Primary studies were very weak…most neither concealed or blinded Primary studies were very weak…most neither concealed or blinded Significant statistical heterogeneity Significant statistical heterogeneity Use of fixed effect models in analysis Use of fixed effect models in analysis Combined different interventions (2.5%, 5%, 25%) Combined different interventions (2.5%, 5%, 25%) Clinical heterogeneity a major concern Clinical heterogeneity a major concern –Populations (neonates, adults) very different –Many Indications –Different control groups –No protocols for administration –Trials span 2 decades Mortality primarily driven by a few unbalanced studies Mortality primarily driven by a few unbalanced studies

22. AuthorYearPopulationComparator # Studies RR* 95% CI Alderson2002 Critically ill Albumin310.66 0.50 – 0.85 Wilkes2001 No restriction Albumin550.90 0.78 – 1.05 Roberts 1998 Critically ill Albumin300.60 0.45 – 0.79 Alderson2002 Critically ill Colloids380.66 0.49 – 0.93 Choi1999 All adult pts. Colloids170.86 0.63 – 1.17 Schierhout 1998 Critically ill Colloids190.84 0.69 – 1.02 Wade1997Trauma 7.5% Saline/Dextran 81.20** 0.94 – 1.57 *RR<1 favors crystalloids ** Odds ratios

23. AuthorYear Some Sub-group Analyses: Pooled RR (95% CI’s) Wilkes 2001 (A) Surgery/trauma 0.89 (0.69 - 1.18) Ascites 1.08 (0.78 – 1.49) Alderson 2002 (A) Hypovolemia 0.68 (0.90 - 1.03) Burns 0.42 (0.19 - 0.90) Hypoalbuminemia 0.73 (0.49 – 1.06) Schierhout 1998 (C) Trauma 0.77 (.057 – 1.05) Surgery 1.82 (0.61 – 5.56) Burns 0.83 (0.60 – 1.14) Choi 1999 (C) Trauma 0.39 (0.17 - 0.89) Non-Trauma 0.98 (0.70 - 1.36) Alderson 2002 (C) HES 0.86 (0.51 - 1.47) Gelatin 2.0 (0.33 – 12.5) Dextran 0.81 (0.61 – 1.06) *RR < 1 favors crystalloids

24. Why do meta-analyses report discordant results? Clinical QuestionStudy selection and inclusion Populations of patientsSelection criteria InterventionsApplication of the selection criteria Outcome measuresStrategies to search the literature Settings Data ExtractionAssessment of study quality Methods to measure outcomesMethods to assess quality End pointsInterpretations of quality assessments Human error (random or systematic)Methods to incorporate quality assessments in review Assessment of the ability to combine Statistical methods for data synthesis StudiesFixed versus random effects Statistical methods Clinical criteria to judge the ability to combine studies Jadad, Cook, Browman CMAJ 1997:156(10); 1411-1416

25. Types of discordance Jadad, Cook, Browman CMAJ 1997:156(10); 1411-1416 TypeExample ___________________________________________________________________ Results Direction of EffectOne review favors the experimental treatment and another favors the control treatment. Magnitude of EffectOne review suggests that the intervention results in a 30% reduction in mortality and another suggests that it results in a 5% reduction in mortality. Statistical SignificanceOne review shows a statistically significant difference between the experimental and the control treatments and another review shows a non-significant difference between them. Interpretation Authors interpret same results differently

26. Copyright ©1999 BMJ Publishing Group Ltd. Roberts, I. et al. BMJ 1999;318:1214b Has use of Albumin decreased?

27. What type of fluid would you administer in the first 6 hours of resuscitation? (N=210)

28. Does albumin supplementation improve oxygenation? Objective: to determine if 25% albumin added to furosemide improve urine output and pulmonary physiology Objective: to determine if 25% albumin added to furosemide improve urine output and pulmonary physiology Design:Double blind RCT Design:Double blind RCT Patients: 37 mechanically ventilated patients with low total protein and ALI Patients: 37 mechanically ventilated patients with low total protein and ALI Interventions:5 day infusion of 100 mls of Albumin TID plus furosedmide infusion versus furosemide alone Interventions:5 day infusion of 100 mls of Albumin TID plus furosedmide infusion versus furosemide alone Martin et al, CCM,2002; pp2175-2182

29. What did they find? 5.3 kg more weight loss in albumin group (p=0.04) 5.3 kg more weight loss in albumin group (p=0.04) Improved PaO2/FIO2 ratio (171 vs 236, p=0.02) Improved PaO2/FIO2 ratio (171 vs 236, p=0.02) Improved hemodynamics with decreased heart rate and increased blood pressure Improved hemodynamics with decreased heart rate and increased blood pressure No change in other measures of lung mechanics No change in other measures of lung mechanics Martin et al, CCM,2002; pp2175-2182

30. Does albumin supplementation improve outcomes in spontaneous bacterial peritonitis? Objective: to determine whether plasma expansion with 20% albumin prevents renal impairment and reduces mortality Objective: to determine whether plasma expansion with 20% albumin prevents renal impairment and reduces mortality randomized trial involving 7 tertiary centres Design: randomized trial involving 7 tertiary centres Patients: 126 patients with cirrhosis and spontaneous bacterial peritonitis Patients: 126 patients with cirrhosis and spontaneous bacterial peritonitis Interventions: cefotaxime versus cefotaxime and albumin infusion of 1.5 g/kg with cefotaxime. Interventions: cefotaxime versus cefotaxime and albumin infusion of 1.5 g/kg with cefotaxime. No active controls and not blinded No active controls and not blinded Sort et al, NEJM 1999 pp 403-9

31. What did they find? OutcomesAlbuminControlp value (n=63)(n=63) Resolution of infection 98%94%0.36 Renal impairment n(%)21(33%)6(10%)0.002 Hospital mortality n(%)18(29%)6(10%)0.01 3 month mortality26(41%)14(22%)0.03 Sort et al, NEJM 1999 pp 403-9

32. Do protocols of care and timing matter?

33. Evolving Knowledge and Lessons Learned: High risk patients with global tissue hypoxia ? Helpful ? Harmful Treat in early stage of disease

34. Does early treatment of global tissue hypoxia improve outcome? The Rivers Hypothesis

35. Oxygen Debt: To Pay or Not to Pay

36. Optimization Trials “Every hemodynamic study is not Shoemaker” Mortality (Boyd, New Horiz, 1996) Early Late (Kern, Crit Care Med, 2002)

37. Goal Directed Therapy in the Critically Ill Goal: to determine if early Goal-directed therapy targeting treatment of venous hypoxia improved clinical outcomes Setting: Single centre study Study Population:263 patients with EARLY sepsis and septic shock Study Design: Open labeled RCT Intervention:Goal-directed vs standard therapy initiated in ER for 6 hours Outcome: In-hospital mortality Rivers et al NEJM 2001;345:1368

38. Early Goal-directed Therapy Supplemental oxygen ± endotracheal intubation and mechanical ventilation Central venous and arterial catheterization CVP Crystalloid Colloid <8 mm Hg MAP 8-12 mm Hg Vasoactive agents <65 mm Hg >90 mm Hg ScvO 2 ≥65 and ≤90 mm Hg Goals achieve d ≥70% Hospital admission Yes No Sedation and/or paralysis (if intubated) Transfusion of red cells to hematocrit ≥30% <70% Inotropic agents <70% ≥70% Rivers et al NEJM 2001;345:1368

39. More Details About the Intervention Goal-directed vs standard therapy initiated in ER for 6 hours: û Both Groups were maintained with: û CVP 8 – 12 û MAP ≥ 65 û Urine output ≥ 0.5 ml/kg/hr n Treatment group had CVP with continuous O 2 û Monitoring with target of mixed saturation ≥ 70%  Targets of therapy were: û Oxygen saturation ≥ 93% û HCT ≥ 30% û Cardiac Index û Systemic oxygen consumption Rivers et al NEJM 2001;345:1368

40. Goal Directed Therapy in the Critically Ill 0 – 6 hours, Goal vs Standard Therapy û Fluids: 4981 ml vs 3499 ml, p <.001 û RBC: 64.1% vs 18.5%, p <.001 û Vasopressor: 27.4% vs 30.3%, p = 0.62 û Inotropes: 13.7% vs 0.8%, p <.001 Rivers et al NEJM 2001;345:1368

41. Early Goal directed therapy (1) DeadAlive%Dead Goal-directed3892A=38/130=30.5% Control5974B=59/133=46.5% Absolute Risk Reduction(ARR)= 16% Relative Risk (RR)= 30.5 /46.5=0.66 (95% CI of 0.38 – 0.87) Relative Risk Reduction(RRR)= (1- 0.66) x 100= 34% Odds Ratio (OR)= a*d /b*c = 0.52 Number needed to treat (NNT)= 1/0.16= 6 (1)Rivers et al, NEJM, 2001,1368-77

42. What can we infer? The type, timing and quantity of fluid resuscitation may impact on mortality The type, timing and quantity of fluid resuscitation may impact on mortality Complex area of care with few high quality trials Complex area of care with few high quality trials Meta-analyses primarily highlight deficiencies in literature Meta-analyses primarily highlight deficiencies in literature Can’t and should not infer treatment choices based upon meta-analyses Can’t and should not infer treatment choices based upon meta-analyses Albumin may be beneficial in improving oxygenation in ALI and supporting patients with cirrhosis who have bacterial peritonitis Albumin may be beneficial in improving oxygenation in ALI and supporting patients with cirrhosis who have bacterial peritonitis Early aggressive fluid resuscitation may save lives Early aggressive fluid resuscitation may save lives Less evidence in support of other colloids Less evidence in support of other colloids

43. What do I recommend? Further clinical trials addressing following questions: Further clinical trials addressing following questions: –Different % albumin versus crystalloid in various settings –Different colloids versus crystalloids in various settings –All crystalloids not created equal either??? –Treatment protocols versus usual care

44. And then we were SAFE’ed

45. Thank You