Albumin Use Guidelines and Outcome in a Surgical Intensive Care Unit FREE

The administration of intravenous fluids to maintain or increase intravascular volume is a common intervention in the surgical intensive care unit (SICU). The choice of whether to use crystalloid (eg, isotonic sodium chloride solution, Ringer lactate) or colloid (eg, albumin, hetastarch) for fluid resuscitation in the ICU continues to be debated.^1- 3 Costs associated with colloid fluid resuscitation are significantly higher than costs associated with crystalloid fluid resuscitation.

Compared with other colloid and crystalloid solutions, albumin solutions are expensive.⁴ Volume for volume, human albumin solution is more than twice as expensive as hydroxyethyl starch and more than 30 times more expensive than crystalloid solutions, such as sodium chloride or Ringer lactate. Because of the high cost, the limited availability of human albumin, and the potential for associated adverse events,⁵ it is imperative that its use be restricted to indications for which it is efficacious, and hypoalbuminemia and critical illness are controversial indications.^6- 7

The recent publication of the Saline versus Albumin Fluid Evaluations (SAFE) study⁸ in 6997 critically ill patients randomized to receive 4% albumin vs isotonic sodium chloride solution documented no difference in 28-day all-cause mortality, ICU or hospital length of stay, or duration of mechanical ventilation or renal replacement therapy. This study⁹ further documented that albumin has no clear benefit in the fluid resuscitation of most critically ill patients. We therefore sought to assess the effect of implementation of an albumin use guideline in the SICU to reduce albumin use on risk-adjusted ICU mortality and cost savings.

This prospective cohort study was conducted from July 1, 2004, to July 1, 2005. All patients admitted to the University of Michigan SICU (general surgical, transplantation, vascular, specialty surgery, and trauma patients) were included in this study. There was no restriction on albumin use in the first 3 quarters of the year. Before the beginning of the final quarter an organized educational program was initiated by the new intensivist-led critical care team in March 2005, which included the following: journal clubs (review of SAFE trial and other studies), didactic sessions (review of appropriate indications for albumin), discussion during patient care rounds, and an educational program directed toward the SICU resident team, nursing staff, and primary surgical teams. Appropriate albumin use guidelines, adapted from the University Health Consortium albumin guidelines,¹⁰ were instituted in the fourth quarter. Prospective clinical and risk-adjusted outcome data were collected by Acute Physiology and Chronic Health Evaluation (APACHE) III dedicated personnel. Albumin use data and costs were obtained from the pharmacy prospective database. Follow-up data were collected through the end of calendar year 2006.

All patients (n = 1361) admitted to the SICU during the study period were included in this study. The implementation of albumin use guidelines resulted in a statistically significant reduction in albumin use (54%, P = .004, Figure 1). Similarly, a significant cost savings of 56% was found (Figure 2). The cost of albumin was calculated directly based on the unit cost of 12.5 g (25%) of albumin ($43 per 50-mL vial). No difference was seen in mean APACHE III scores on ICU day 1 in all 4 quarters. No difference in ICU mortality was identified with the reduction in albumin use, despite a higher APACHE III score and predicted ICU mortality in that quarter. The ICU mortality was significantly lower than the predicted ICU mortality during all quarters (Table 1). No difference in mean ICU length of stay was noted. Follow-up data for albumin use documented a persistent reduction in albumin use in the SICU during the next 6 quarters (2005-2006, Table 2).

Fluid resuscitation for hypovolemia is integral to the short-term medical management of critically ill patients. When it comes to selecting the resuscitation fluid, physicians are faced with a range of options. At the most basic level, the choice is between a colloid and crystalloid solution. Colloids are widely used, having been recommended in a number of resuscitation guidelines and intensive care management algorithms.¹¹

The American Hospital Consortium Guidelines recommend that colloids be used in hemorrhagic shock until blood products become available and in nonhemorrhagic shock after an initial infusion with crystalloid. A 1995 survey of American academic health centers, however, found that the use of colloids far exceeded these recommendations.¹² There is, however, no scientific evidence for this widespread practice, with the most recent systematic review¹³ of 19 trials and 7576 patients concluding that “there is no evidence from randomized controlled trials that resuscitation with colloids reduces the risk of death compared to resuscitation with crystalloids in patients with trauma, burns or following surgery.”

Controversy surrounding the efficacy and potential harm of albumin use in a variety of clinical settings is long-standing.¹⁴ A meta-analysis published by the Cochrane Injuries Group¹⁵ in 1998 documented an increased mortality rate in patients randomized to albumin supplementation in 30 controlled trials (1419 patients). They concluded that “albumin should not be used outside the context of a properly conceived and rigorously controlled trial with mortality as the endpoint.” An accompanying editorial suggested that “the administration of albumin should be halted.”¹⁶

Significant criticism of the Cochrane analysis ensued, given that it had excluded a number of trials that had shown reduced mortality with albumin administration. A subsequent, more inclusive (not just critically ill patients) meta-analysis¹⁷ of 55 trials that involved surgery or trauma, burns, hypoalbuminemia, high-risk neonates, ascites, and other indications was published. This analysis concluded that “no effect of albumin on mortality was detected.”³

An update to the Cochrane review of albumin use in critically ill patients was published in 2004, including 32 trials and 8452 patients,¹⁸ and concluded that “there is no evidence that albumin reduces mortality in critically ill patients.” The authors again recommended that “albumin should only be used within the context of a well concealed and adequately powered randomized controlled trial.”

This landmark trial of the safety of albumin in intensive care, performed by the Australian and New Zealand Intensive Care Society Clinical Trials Group, has recently been published. The SAFE study is a prospective trial of 6997 critically ill patients randomized to 4% albumin or isotonic sodium chloride solution for fluid resuscitation during the 28 days after ICU admission. No differences in 28-day all-cause mortality (relative risk [RR], 0.99; 95% confidence interval [CI], 0.91-1.09; P = .87), ICU or hospital length of stay, or duration of mechanical ventilation or renal replacement therapy were identified.

A subgroup analysis of the SAFE study documented that among patients with trauma (n = 1186), albumin was associated with a trend toward increased mortality (RR, 1.36; 95% CI, 0.99-1.86), particularly in patients with traumatic brain injury. Another subgroup analysis of patients with severe sepsis (n = 1218) documented that albumin use was associated with a trend toward reduced mortality (RR, 0.87; 95% CI, 0.74-1.02). These results must be interpreted with caution, particularly in light of emerging data documenting that albumin administration in sepsis is associated with a rapid reduction in serum albumin concentration and decreased volume expansion.¹⁹ The study cohort was also analyzed to determine the effect of baseline serum albumin concentration (≤2.5 vs >2.5 g/dL) on outcome of resuscitation in the SAFE study, and no difference was identified in 28-day all-cause mortality, ICU or hospital length of stay, or duration of renal replacement therapy or mechanical ventilation.²⁰

The substantial differences in the wholesale costs of these 2 therapies is notable ($232 for 1 L of 5% albumin vs $2.32 for 1 L of isotonic sodium chloride solution), with the cost for each patient treated in the SAFE study estimated at $521.30 for albumin compared with $6.90 for isotonic sodium chloride solution, a 75-fold difference.²¹ These cost differences suggested that the routine use of albumin is hard to justify because crystalloid use is associated with similar outcomes at a lower cost. However, some physicians may interpret the SAFE study to have documented an absence of harm associated with albumin and will continue to use it based on presumed potential benefit in select patients.²²

More recently, the results of the Sepsis Occurrence in Acutely Ill Patients²³ study, a prospective, multicenter, observational cohort study (n = 3147) conducted in Europe, examined albumin use and outcomes. Interestingly, 354 patients (11.2%) received albumin, and most (157, 44.4%) received it during the first 24 hours after ICU admission. Albumin administration was independently associated with decreased 30-day survival (hazard ratio, 1.57; 95% CI, 1.11-2.22; P = .01) in this population of critically ill patients, of whom 1388 (44.1%) were surgical patients. Furthermore, surgical patients were the largest cohort to receive albumin, comprising 218 of the 354 patients who received albumin during the study period.

The rationale for giving albumin solutions rather than crystalloid solutions in cases of hypovolemic shock is that fluid reabsorption from the interstitial space is enhanced, and fluid therefore remains in the vascular system longer as predicted by the Starling principle that governs hydrostatic and oncotic pressures within the capillaries. However, the Starling principle may not appropriately reflect the microcirculation in critically ill patients, especially under conditions of capillary leakage, as may happen in sepsis, burns, and other acute illness.²⁴ Additional preclinical studies²⁵ have documented that albumin protects against gut-induced lung injury in vitro and in vivo. Small studies^26- 29 have confirmed some potential benefits, particularly with the more concentrated 25% solution for restoration of colloid osmotic pressure. Additional trials are under way (Early Albumin Resuscitation During Septic Shock,³⁰ Albumin in Acute Ischemic Stroke,³¹ and Treatment of Subarachnoid Hemorrhage With Human Albumin³²) to further investigate the efficacy of albumin.

The lack of demonstrable evidence in improvement of mortality with the use of albumin in fluid resuscitation, coupled with the availability of equally effective lower-cost alternatives, emphasizes the need for a published set of guidelines to ensure the safe and cost-effective use of albumin. Several studies have previously shown discrepancies in albumin use even in the presence of set guidelines, with the variability dependent on the individual ordering physician's preconceived bias, usually based on experience and not on factual evidence of superior efficacy. Fan et al³³ showed that the most common indications for albumin use were hypotension in hemodialysis (18.9%), volume replacement or expansion (15%), and hypoalbuminemia (14.8%). No indication for albumin use was identified in 9.4% of patients. Albumin use was dependent on the specialty of the ordering physician, with surgeons and nephrologists accounting for most use. Additional educational efforts are therefore warranted.

Few data exist regarding the use of albumin and outcome in SICU patients. An administrative database study in patients undergoing coronary artery bypass graft surgery (n = 19 578) who received colloids (albumin vs nonprotein colloids [hetastarch or dextran]) identified that albumin use in 8084 cases (41.3% of the study cohort) was associated with lower all-cause in-hospital mortality (2.47% vs 3.03%, P = .02).³⁴ In the multivariable logistic regression analysis, albumin use was associated with 25% lower odds of mortality compared with nonprotein colloid use (OR, 0.80; 95% CI, 0.67-0.96). Whether this study documented the improved efficacy of albumin or the adverse effects of nonprotein colloids in this patient population is controversial. Most importantly, this study's major limitation was that there was no crystalloid control group. Additional studies in burn shock resuscitation document the lack of efficacy of albumin on the incidence of multiple organ dysfunction³⁵ or other outcome measures.³⁶

Our current study documented that the implementation of albumin use guidelines and educational efforts was effective in significantly reducing albumin use in a SICU. During the first 3 quarters of this study, no attempt was made to assess the appropriateness of albumin use, in terms of either indication or dosage, but it is evident that albumin was used contrary to the established evidence-based guidelines. This is a common occurrence, with an evaluation of albumin use in 53 hospitals in the United States confirming that albumin was inappropriately prescribed for 57.8% of adult patients and 52.2% of pediatric patients.³⁷

The absence of any demonstrable negative impact of reduction in albumin use on risk-adjusted ICU mortality or ICU length of stay in SICU patients further supports the use of crystalloids as the first-line fluid during resuscitation in the critical care setting. Additional efforts are required to further reduce albumin use in SICUs, and ongoing efforts are under way. Institutions should implement guidelines that focus on responsible use of albumin in an increasingly cost-conscious health care environment, and continued educational efforts promoting evidence-based practices in the ICU are warranted.

Correspondence: Lena Napolitano, MD, Department of Surgery, University of Michigan Health System, Room 1C421 University Hospital, 1500 E Medical Center Dr, Ann Arbor, MI 48109-0033 (lenan@umich.edu).

Accepted for Publication: July 5, 2007.

Author Contributions:Study concept and design: Charles, Purtill, Dickinson, Meldrum, and Napolitano. Acquisition of data: Charles, Purtill, Dickinson, Kraft, Pleva, Meldrum, and Napolitano. Analysis and interpretation of data: Charles, Purtill, Kraft, Pleva, Meldrum, and Napolitano. Drafting of the manuscript: Charles, Purtill, and Napolitano. Critical revision of the manuscript for important intellectual content: Dickinson, Kraft, Pleva, Meldrum, and Napolitano. Statistical analysis: Napolitano. Administrative, technical, and material support: Charles, Purtill, Dickinson, Kraft, Pleva, Meldrum, and Napolitano. Study supervision: Charles, Purtill, Dickinson, Kraft, Pleva, and Napolitano.

Financial Disclosure: None reported.