当前位置: 首页 > 期刊 > 《新英格兰医药杂志》 > 2004年第18期 > 正文
编号:11306608
Fluid Resuscitation in the Intensive Care Unit
http://www.100md.com 《新英格兰医药杂志》
     To the Editor: The Saline versus Albumin Fluid Evaluation (SAFE) Study (May 27 issue)1 debunks a 1998 meta-analysis associating albumin with increased mortality in critical care populations.2 Regrettably, this trial may spawn a new misconception: that albumin and saline are equivalent for resuscitating a heterogeneous mix of patients in intensive care units.

    Patients with trauma are generally young and healthy and should not be grouped for evaluation with older patients without trauma, many of whom have preexisting cardiopulmonary disease. The incidence of pulmonary edema is low in patients with hemorrhagic trauma, and studies have not demonstrated a benefit of albumin over saline.3 Consequently, crystalloids are the accepted standard for volume replacement in patients with trauma.4

    When patients with trauma are excluded, use of albumin in the SAFE Study yielded a relative risk of death of 0.96, as compared with saline use. More specifically, the analysis of the patients with sepsis revealed a stronger mortality trend favoring albumin (relative risk, 0.87).

    The SAFE investigators should enroll additional patients with sepsis to resolve whether 1 of every 10 lives could be saved simply by using albumin for fluid resuscitation. Stratifying patients according to the baseline albumin level could answer another fundamental question: whether, as some suggest,5 supplementation with albumin improves outcomes in patients with severe hypoproteinemia.

    Gary R. Haynes, M.D., Ph.D.

    Medical University of South Carolina

    Charleston, SC 29425

    haynesg@musc.edu

    Keith E. Berman, M.P.H., M.B.A.

    Health Research Associates

    Pasadena, CA 91106

    References

    The SAFE Study Investigators. A comparison of albumin and saline for fluid resuscitation in the intensive care unit. N Engl J Med 2004;350:2247-2256.

    Cochrane Injuries Group Albumin Reviewers. Human albumin administration in critically ill patients: systematic review of randomised controlled trials. BMJ 1998;317:235-240.

    Moss GS, Lowe RJ, Jilek J, Levine HD. Colloid or crystalloid in the resuscitation of hemorrhagic shock: a controlled clinical trial. Surgery 1981;89:434-438.

    Committee on Trauma. Advanced trauma life support manual. Chicago: American College of Surgeons, 1997:103-12.

    Vincent J-L, Dubois MJ, Navickis RJ, et al. Hypoalbuminemia in acute illness: is there a rationale for intervention? A meta-analysis of cohort studies and controlled trials. Ann Surg 2003;237:319-334.

    To the Editor: The SAFE Study investigators provide evidence of an increase in the risk of death among trauma patients with concomitant brain injury who received albumin as compared with those who were treated with saline. These results, obtained from a subgroup analysis, support our own findings in a prospective, randomized, controlled trial of the safety of large doses of medium-molecular-weight hydroxyethyl starch (HES 130/0.4) in patients with traumatic brain injury.1 The control group, which received albumin as an add-on colloid to HES 200/0.5, had a significantly higher incidence of intracranial-pressure peaks above 30 mm Hg and more cumulative hours of elevated intracranial pressure. These differences between groups could be linked neither to the underlying brain trauma nor to intracranial bleeding complications. We hypothesized that albumin, which — unlike HES — is suspected to leak from a regionally disrupted blood–brain barrier,2,3,4 might have caused an increase in intracranial volume and therefore excessive intracranial pressure. Our hypothesis appears to be endorsed by the findings of the SAFE Study and may also explain the finding by the SAFE investigators of higher mortality among patients who had trauma involving brain injury when they were treated with albumin.

    Thomas A. Neff, M.D.

    University of Michigan Medical School

    Ann Arbor, MI 48109

    thneff@umich.edu

    Reto Stocker, M.D.

    University Hospital Zurich

    CH-8001 Zurich, Switzerland

    Donat R. Spahn, M.D.

    University Hospital Lausanne

    CH-1011 Lausanne, Switzerland

    Dr. Spahn reports having served on a data and safety monitoring board for Fresenius Kabi and as a consultant for B. Braun.

    References

    Neff TA, Doelberg M, Jungheinrich C, Sauerland A, Spahn DR, Stocker R. Repetitive large-dose infusion of the novel hydroxyethyl starch 130/0.4 in patients with severe head injury. Anesth Analg 2003;96:1453-1459.

    Fukuda K, Tanno H, Okimura Y, Nakamura M, Yamaura A. The blood-brain barrier disruption to circulating proteins in the early period after fluid percussion brain injury in rats. J Neurotrauma 1995;12:315-324.

    Peters T. All about albumin: biochemistry, genetics, and medical applications. San Diego, Calif.: Academic Press, 1995.

    Dieterich HJ, Reutershan J, Felbinger TW, Eltzschig HK. Penetration of intravenous hydroxyethyl starch into the cerebrospinal fluid in patients with impaired blood-brain barrier function. Anesth Analg 2003;96:1150-1154.

    To the Editor: Although the SAFE Study supports my bias about the lack of superiority of albumin over crystalloid, I have one concern. Were the patients really hypovolemic? Baseline central venous pressures were about 9 mm Hg in both groups and increased to only 11 mm Hg (in the albumin group) and 10 mm Hg (in the saline group) during the four days of the study. Furthermore, the baseline heart rates and mean arterial pressures, as well as responses to the infusions, do not suggest severe hypovolemia. The study may merely show that in mildly unstable, critically ill patients, the choice of fluid does not matter.

    James E. Barone, M.D.

    Lincoln Hospital and Mental Health Center

    Bronx, NY 10451

    To the Editor: The SAFE Study demonstrates similar clinical outcomes whether 4 percent albumin or normal saline is used for fluid resuscitation. The authors and the editorialist1 both describe well the limitations and applicability of this study. However, neither authors nor editorialist comments on the difference in the cost of the two therapies. On the basis of data from the SAFE Study and published information on average wholesale prices2 — $232 for 1 liter of 5 percent albumin and $2.32 for 1 liter of saline — the cost for each patient treated with albumin would be $521.30, as compared with $6.90 for each patient treated with saline, a 75-fold difference. This calculation may overstate the difference, since somewhat less of the 5 percent albumin preparation available in the United States might have been required. Nevertheless, intensivists should be aware of the substantial difference in costs, especially since the cost of fluid resuscitation with 5 percent albumin will often not be reimbursed by insurers.

    William A. Primack, M.D.

    Kristina Estes, Pharm.D.

    University of Massachusetts Medical School

    Worcester, MA 01655

    william.primack@fallon-clinic.com

    References

    Cook D. Is albumin safe? N Engl J Med 2004;350:2294-2296.

    Drug topics red book. Montvale, N.J.: Thomson PDR, 2003.

    To the Editor: The recent SAFE Study provides definitive evidence that the choice of albumin or crystalloid for fluid resuscitation does not affect overall mortality in adult patients in intensive care. Unfortunately, those who work with pediatric patients are left to extrapolate from data on adults once again. In theory, albumin may be more beneficial in pediatric patients, in whom volume management is crucial, yet this issue has not been well studied. Since 1992, there have been five randomized, controlled trials comparing albumin and crystalloid with respect to mortality in pediatric patients in intensive care. Four trials involved neonates, and mortality did not differ statistically.1,2,3,4 One pediatric trial involving patients with burns showed a higher mortality among those treated with albumin.5 All trials were relatively small, and not all compared albumin and crystalloid directly. The SAFE Study has shown that without corroborative results from a well-designed clinical trial, the conclusions drawn from a series of small trials may be incorrect. Until a similar trial is conducted with pediatric patients, intensivists will continue to practice on the basis of dogma and empiricism, rather than evidence-based medicine.

    Eric C. Walter, M.D.

    Richard Wendorf, M.D.

    Youngki Kim, M.D.

    University of Minnesota

    Minneapolis, MN 55455

    walte069@umn.edu

    References

    Kanarek KS, Williams PR, Blair C. Concurrent administration of albumin with total parenteral nutrition in sick newborn infants. JPEN J Parenter Enteral Nutr 1992;16:49-53.

    Greenough A, Emery EF, Hird MF, Gamsu HR. Randomised controlled trial of albumin infusion in ill preterm infants. Eur J Pediatr 1993;152:157-159.

    So KW, Fok TF, Ng PC, Wong WW, Cheung KL. Randomised controlled trial of colloid or crystalloid in hypotensive preterm infants. Arch Dis Child Fetal Neonatal Ed 1997;76:F43-F46.

    Oca MJ, Nelson M, Donn SM. Randomized trial of normal saline versus 5% albumin for the treatment of neonatal hypotension. J Perinatol 2003;23:473-476.

    Greenhalgh DG, Housinger TA, Kagan RJ, et al. Maintenance of serum albumin levels in pediatric burn patients: a prospective, randomized trial. J Trauma 1995;39:67-73.

    The authors reply: Dr. Haynes and Mr. Berman are concerned that the SAFE Study may "spawn misconception . . . that albumin and saline are equivalent for resuscitating a heterogeneous mix of patients in intensive care units." Far from being a misconception, we believe this is the key message of the SAFE Study. Whether one or the other is beneficial in particular subgroups still requires further study. Haynes and Berman may have misinterpreted the results of our subgroup analyses. For patients without trauma, 22.6 percent of those assigned to albumin died (641 of 2831) as compared with 23.5 percent of those assigned to saline (666 of 2830) — a difference of 0.9 percent (relative risk, 0.96; 95 percent confidence interval, 0.88 to 1.06; P=0.52). For patients with sepsis, 30.7 percent of those assigned to albumin died (185 of 603) as compared with 35.3 percent of those assigned to saline (217 of 615); the difference in mortality was 4.6 percent (relative risk, 0.87; 95 percent confidence interval, 0.74 to 1.02; P=0.09). The results do not indicate that albumin reduces mortality in patients without trauma or in patients with severe sepsis. We agree with Dr. Cook that the subgroup with severe sepsis requires further study.

    Dr. Neff and colleagues comment on patients with trauma and brain injury. Our study does not provide definitive evidence that the administration of albumin increased mortality among patients with trauma who had brain injury, and thus it would be unwise to extrapolate our results to their findings in patients who received albumin in addition to other fluids. Appropriately designed randomized, controlled trials are required to determine the safety of albumin and other resuscitation fluids in patients with brain injury.

    We are not surprised that, as noted by Dr. Barone, baseline hemodynamic variables do not suggest severe hypovolemia. The current practice is to administer fluid as soon as intravascular volume depletion becomes apparent, rather than awaiting severe hypovolemia. We believe patients received a sufficient volume of the study fluid to address the issue of albumin's safety, as had been raised by the Cochrane reviewers.1 Although albumin costs more than saline, direct costs of treatment are only one factor in the economic evaluation of new or established therapies.2 We were unable to secure resources for a formal cost-effectiveness assessment in the SAFE Study, and at present the overall cost implications of treatment with albumin or saline remain unknown. In reply to Dr. Walter and colleagues: we explored the idea of including children in the study, but the blinded administration sets were unsuitable for use in small children, and the lower mortality rate for children in intensive care units would have required studying thousands of additional patients.

    We hope that the SAFE Study will encourage others to conduct large randomized trials involving patients in intensive care units and in subgroups of these patients, including children.

    Simon Finfer, M.B., B.S.

    Australian and New Zealand Intensive Care Society Clinical Trials Group

    Carlton, VIC 3053, Australia

    ctg@anzics.com.au

    Neil Boyce, M.B., B.S., Ph.D.

    Australian Red Cross Blood Service

    Melbourne, VIC 3053, Australia

    Robyn Norton, Ph.D., M.P.H.

    George Institute for International Health

    Sydney, NSW 2050, Australia

    References

    Cochrane Injuries Group Albumin Reviewers. Human albumin administration in critically ill patients: systematic review of randomised controlled trials. BMJ 1998;317:235-240.

    Coughlin MT, Angus DC. Economic evaluation of new therapies in critical illness. Crit Care Med 2003;31:Suppl:S7-S16.

    The editorialist replies: Drs. Primack and Estes cite wholesale costs for 5 percent albumin and for normal saline, suggesting that intensivists should be aware of this difference. Although few may be able to cite institution-specific prices, I hope that all intensivists are familiar with this cost differential. An economic analysis requires quantitatively comparing two or more interventions with respect to both resource use and clinical end points. Accordingly, some intensivists will point to equivalent outcomes in the SAFE Study, including the lack of harm overall, and use albumin selectively based on a pathophysiological rationale. As I stated in my editorial, "Others will conclude that without proof of benefit, routine use of albumin is hard to justify; for similar clinical outcomes at a lower cost, crystalloids may suffice in most circumstances."

    Primack and Estes highlight the fact that the costs of albumin may not be reimbursed by insurers. In countries with universal health care, insurance policies do not influence fluid choices, although rationing may occur for interventions without promise of a health advantage. As I noted, the rate of use of albumin will thus reflect interpretation of the SAFE Study's point estimate and confidence limits, as well as "patient-specific conditions, clinicians' preferences, perceptions regarding the safety of biologic fluids, availability, and cost."

    Deborah Cook, M.D.

    McMaster University

    Hamilton, ON L8N 3Z5, Canada