Albumin A
Supplementation in the Critically
Ill
Prospective, Randomized Trial
Eugene F. Foley, MD; Bradley C. Borlase, MD; Walter H. Dzik, MD; Bruce R. Bistrian, MD, PhD; Peter N. Benotti, MD \s=b\ Albumin replacement to correct hypoalbuminemia in critically ill patients has been controversial. This study was a prospective, randomized trial of 25% albumin administration in 40 hypoalbuminemic (serum albumin, 25 g/L [2.5 g/dL]), critically ill patients. The treatment group (18 patients) received 25% albumin supplementation to achieve and maintain serum albumin levels of 25 g/L (2.5 g/dL) or greater, while the nontreatment group (22 patients) received no concentrated albumin. There was no clinical benefit from albumin therapy when assessing mortality (39% vs 27%, treatment vs control) or major complication rate (89% vs 77% of patients). There were also no significant differences in length of hospital stay, intensive care unit stay, ventilator dependence, or tolerance of enteral feeding, despite significant elevations of albumin in the treatment group. The costly use of exogenous albumin as treatment for hypoalbuminemia in this patient population does not appear to be justified. (Arch Surg. 1990;125:739-742)
Hypoal buminemia patientLundholm1 clinical settings. Warnold
has been shown to be an accurate prog¬ nostic indicator of poor outcome in a wide variety in 1984 docu¬ and of mented an increased complication rate and increased length of stay following surgery in patients with hypoalbuminemia preoperatively. Anderson and Wochos2 confirmed an in¬ creased length of stay among a cohort of medical patients who were hypoalbuminemic on admission when compared with a similar group with normal serum albumin levels. More specif¬ ically, hypoalbuminemia has been associated with a variety of organ system dysfunctions, including pulmonary fluid over¬ load,3 enterai feeding intolerance,4'5 and poor wound healing.6 This correlation is based at least partially on the accuracy of hypoalbuminemia as a marker for serious disease processes, including sepsis' and malnutrition.8 Theoretically, hypoalbu¬ minemia itself may further contribute to morbidity by alter¬ ing albumin's physiologic role in maintaining intravascular oncotic pressure, adversely affecting water balance between the intravascular and interstitial spaces.9 Recent work has also suggested that albumin may have important roles as a scavenger of free radicals, a binding agent for toxic com¬ pounds in disease states, and a carrier for a wide variety of drugs and hormones.10 An increasing awareness of these physiologic functions of albumin and the direct association between hypoalbuminemia and morbidity has resulted in widespread use of albumin replacement therapy to restore normal albumin levels in hy¬ poalbuminemic patients. In 1986, Hardin et al11 showed con¬ clusively that low serum albumin concentrations could be safely and rapidly restored with human albumin infusions. However, there are no conclusive data from prospective, randomized trials that this iatrogenic elevation in serum albu¬ min concentrations reduces the high morbidity associated with hypoalbuminemia. Without this conclusive evidence, the Accepted for publication February 18,1990. From the Departments of Surgery (Drs Foley, Borlase, and Benotti) and Medicine (Drs Dzik and Bistrian), Harvard University Medical School, New England Deaconess Hospital, Boston, Mass. Read before the 70th Annual Meeting of the New England Surgical Society, Bretton Woods, NH, September 23,1989. Reprint requests to Department of Surgery, Suite 3A, 110 Francis St, Boston, MA 02215 (Dr Benotti).
high cost ofhuman albumin administration has made its use as treatment for hypoalbuminemia controversial. This study is a prospective, randomized trial designed to assess the benefits of concentrated human albumin administration in critically ill, hypoalbuminemic patients receiving total parenteral nutri¬ tion.
PATIENTS AND METHODS All patients referred to the New England Deaconess Hospital Nutrition Support Service, Boston, Mass, for nutritional assessment and total parenteral nutrition with serum albumin concentrations below 25 g/L (2.5 g/dL) were considered for entry. Potential subjects with Child's class C cirrhosis12 were excluded. Forty patients were identified and entered into the study between August 1988 and April 1989. These patients were randomly assigned to either a treatment or nontreatment group by medical record number. The treatment group (18 patients) received 25 to 50 g/d of 25% albumin in addition to full nutritional support with parenteral nutrition. Albumin administra¬ tion was continued daily until serum albumin levels exceeded 25 g/L (2.5 g/dL), after which patients received additional albumin as needed to keep the albumin level at 25 g/L (2.5 g/dL) or higher. Albumin was administered either continuously by adding the 25% solution directly to the patient's total parenteral nutrition, or by separate intravenous boluses three to four times per day. Intermittent bolus therapy was conducted when the patient was receiving total parenteral nutrition as a 3-in-l admixture, with which human albumin solutions are incom¬ patible. The nontreatment group (22 patients) received no exogenous concentrated albumin. At the time of study entrance, thorough evaluations were con¬ ducted for all patients, which included general demographic data, medical histories, routine laboratory data, and a severity of illness score calculated by the APACHE (acute physiology and chronic feealth evaluation) II scoring system.13 All patients were fully evaluat¬ ed by the Nutrition Support Service, and energy and protein needs were estimated (105 to 147 J/kg per day and 1.5 g of protein per kilogram per day) and provided. All patients initially were fed com¬ pletely intravenously, but transition to enterai feedings was actively promoted by the Nutrition Support Service when clinically indicated. Patients were evaluated daily by the primary investigator, who collected all data. Mortality, major complication rates, hospital stay, intensive care unit (ICU) stay, length of ventilator dependence, and tolerance of enterai feedings were tabulated. Complications were defined at the outset of the study, similar to prior definitions.14'1 Cardiovascular complications included myocardial infarction (persistent electrocardiographic changes and isoenzyme elevations), major arrhythmias causing hemodynamic instability and requiring emergent antiarrhythmia therapy, cardiogenic shock (hypotension and hypoperfusion requiring vasopressor therapy), and deep venous thrombosis docu¬ mented by venography. Pulmonary complications included pneumo¬ nia (chest roentgenographic changes and positive sputum cultures), pulmonary insufficiency requiring mechanical ventilation for greater than 24 hours, pneumothorax requiring chest tube placement, and pulmonary embolism documented by arteriography. Renal complica¬ tions included acute renal failure requiring dialysis. Infectious com¬ plications consisted of all viral, bacterial, or fungal infections docu¬ mented by cultures requiring systemic antimicrobial therapy. Gastrointestinal complications included bowel obstruction requiring surgical exploration, gastrointestinal bleeding documented by endoscopy and requiring transfusion of more than 3 U of packed red blood cells, and acute cholecystitis diagnosed clinically and treated with surgery or antibiotic therapy. Intra-abdominal sepsis was defined as a purulent collection drained either surgically or radiographically. A
Downloaded From: http://archsurg.jamanetwork.com/ by a Penn State Milton S Hershey Med Ctr User on 05/24/2015
Table 1 .—Patient Characteristics at Treatment
Age, y Sex, ICU,
%M %
Study Entry*
2.—Mortality and Complications*
No Treatment
Treatment
63.2 ±4.2
66.4 ±3.4
Mortality, No.
55
68
Major complications
(No.)_83 (15/18)_82 (18/22)_
Ventilator
dependent, % (No.) Entry albumin, g/L (g/dL) Entry APACHE score
Table
72
(13/18)_59 (13/22)_
19±1
(1.9±0.1)
18.4 ±2.3
21 ±1
(2.1 ±0.1)
18.0 ±1.7
*ICU indicates intensive care unit; APACHE, acute physiology and chronic health evaluation scoring system.13 Values for age, albumin level, and APACHE score are means ±SEMs. No difference is significant.
No Treatment
(%)_7/18 (38.9)_6/22
Total No.
40
(27)
33
Patients, No. (%)_16/18 (89)_17/22 (77.3) Complications per 2.2 ±0.6 1.5 ±0.4 patient, mean ± SEM *No difference is
significant.
treatment and nontreatment groups at entrance are outlined in Table 1. There were no significant differences in sex, age,
percentage in the ICU, percentage who were ventilator de¬ pendent, serum albumin concentrations, or APACHE II scores at the time of study entrance. Ninety-three percent of
30(3.0)
patients were on the surgical services.
20
(2.0) U
18(1.8) U 4
6
Days in Study Mean ( ± SEM) serum albumin concentrations in the treatment (solid line) and nontreatment (broken line) groups over the first 10 days of
study.
wound dehiscence required surgical closure. Fistulas were demon¬ strated radiographically, and wound infections required wound open¬ ing and drainage of purulent material. All patients were followed up until discharge or death, with the exception of two patients (one in each group) who remained hospital¬ ized at the study closure. Both of these patients had been hospitalized and actively enrolled in the study for greater than 2 months at the study closure. Data were analyzed by the unpaired Student's t test, x2 analysis, or Wilcoxon tests where appropriate.16 We considered P
Supplementation in the Critically
Ill
Prospective, Randomized Trial
Eugene F. Foley, MD; Bradley C. Borlase, MD; Walter H. Dzik, MD; Bruce R. Bistrian, MD, PhD; Peter N. Benotti, MD \s=b\ Albumin replacement to correct hypoalbuminemia in critically ill patients has been controversial. This study was a prospective, randomized trial of 25% albumin administration in 40 hypoalbuminemic (serum albumin, 25 g/L [2.5 g/dL]), critically ill patients. The treatment group (18 patients) received 25% albumin supplementation to achieve and maintain serum albumin levels of 25 g/L (2.5 g/dL) or greater, while the nontreatment group (22 patients) received no concentrated albumin. There was no clinical benefit from albumin therapy when assessing mortality (39% vs 27%, treatment vs control) or major complication rate (89% vs 77% of patients). There were also no significant differences in length of hospital stay, intensive care unit stay, ventilator dependence, or tolerance of enteral feeding, despite significant elevations of albumin in the treatment group. The costly use of exogenous albumin as treatment for hypoalbuminemia in this patient population does not appear to be justified. (Arch Surg. 1990;125:739-742)
Hypoal buminemia patientLundholm1 clinical settings. Warnold
has been shown to be an accurate prog¬ nostic indicator of poor outcome in a wide variety in 1984 docu¬ and of mented an increased complication rate and increased length of stay following surgery in patients with hypoalbuminemia preoperatively. Anderson and Wochos2 confirmed an in¬ creased length of stay among a cohort of medical patients who were hypoalbuminemic on admission when compared with a similar group with normal serum albumin levels. More specif¬ ically, hypoalbuminemia has been associated with a variety of organ system dysfunctions, including pulmonary fluid over¬ load,3 enterai feeding intolerance,4'5 and poor wound healing.6 This correlation is based at least partially on the accuracy of hypoalbuminemia as a marker for serious disease processes, including sepsis' and malnutrition.8 Theoretically, hypoalbu¬ minemia itself may further contribute to morbidity by alter¬ ing albumin's physiologic role in maintaining intravascular oncotic pressure, adversely affecting water balance between the intravascular and interstitial spaces.9 Recent work has also suggested that albumin may have important roles as a scavenger of free radicals, a binding agent for toxic com¬ pounds in disease states, and a carrier for a wide variety of drugs and hormones.10 An increasing awareness of these physiologic functions of albumin and the direct association between hypoalbuminemia and morbidity has resulted in widespread use of albumin replacement therapy to restore normal albumin levels in hy¬ poalbuminemic patients. In 1986, Hardin et al11 showed con¬ clusively that low serum albumin concentrations could be safely and rapidly restored with human albumin infusions. However, there are no conclusive data from prospective, randomized trials that this iatrogenic elevation in serum albu¬ min concentrations reduces the high morbidity associated with hypoalbuminemia. Without this conclusive evidence, the Accepted for publication February 18,1990. From the Departments of Surgery (Drs Foley, Borlase, and Benotti) and Medicine (Drs Dzik and Bistrian), Harvard University Medical School, New England Deaconess Hospital, Boston, Mass. Read before the 70th Annual Meeting of the New England Surgical Society, Bretton Woods, NH, September 23,1989. Reprint requests to Department of Surgery, Suite 3A, 110 Francis St, Boston, MA 02215 (Dr Benotti).
high cost ofhuman albumin administration has made its use as treatment for hypoalbuminemia controversial. This study is a prospective, randomized trial designed to assess the benefits of concentrated human albumin administration in critically ill, hypoalbuminemic patients receiving total parenteral nutri¬ tion.
PATIENTS AND METHODS All patients referred to the New England Deaconess Hospital Nutrition Support Service, Boston, Mass, for nutritional assessment and total parenteral nutrition with serum albumin concentrations below 25 g/L (2.5 g/dL) were considered for entry. Potential subjects with Child's class C cirrhosis12 were excluded. Forty patients were identified and entered into the study between August 1988 and April 1989. These patients were randomly assigned to either a treatment or nontreatment group by medical record number. The treatment group (18 patients) received 25 to 50 g/d of 25% albumin in addition to full nutritional support with parenteral nutrition. Albumin administra¬ tion was continued daily until serum albumin levels exceeded 25 g/L (2.5 g/dL), after which patients received additional albumin as needed to keep the albumin level at 25 g/L (2.5 g/dL) or higher. Albumin was administered either continuously by adding the 25% solution directly to the patient's total parenteral nutrition, or by separate intravenous boluses three to four times per day. Intermittent bolus therapy was conducted when the patient was receiving total parenteral nutrition as a 3-in-l admixture, with which human albumin solutions are incom¬ patible. The nontreatment group (22 patients) received no exogenous concentrated albumin. At the time of study entrance, thorough evaluations were con¬ ducted for all patients, which included general demographic data, medical histories, routine laboratory data, and a severity of illness score calculated by the APACHE (acute physiology and chronic feealth evaluation) II scoring system.13 All patients were fully evaluat¬ ed by the Nutrition Support Service, and energy and protein needs were estimated (105 to 147 J/kg per day and 1.5 g of protein per kilogram per day) and provided. All patients initially were fed com¬ pletely intravenously, but transition to enterai feedings was actively promoted by the Nutrition Support Service when clinically indicated. Patients were evaluated daily by the primary investigator, who collected all data. Mortality, major complication rates, hospital stay, intensive care unit (ICU) stay, length of ventilator dependence, and tolerance of enterai feedings were tabulated. Complications were defined at the outset of the study, similar to prior definitions.14'1 Cardiovascular complications included myocardial infarction (persistent electrocardiographic changes and isoenzyme elevations), major arrhythmias causing hemodynamic instability and requiring emergent antiarrhythmia therapy, cardiogenic shock (hypotension and hypoperfusion requiring vasopressor therapy), and deep venous thrombosis docu¬ mented by venography. Pulmonary complications included pneumo¬ nia (chest roentgenographic changes and positive sputum cultures), pulmonary insufficiency requiring mechanical ventilation for greater than 24 hours, pneumothorax requiring chest tube placement, and pulmonary embolism documented by arteriography. Renal complica¬ tions included acute renal failure requiring dialysis. Infectious com¬ plications consisted of all viral, bacterial, or fungal infections docu¬ mented by cultures requiring systemic antimicrobial therapy. Gastrointestinal complications included bowel obstruction requiring surgical exploration, gastrointestinal bleeding documented by endoscopy and requiring transfusion of more than 3 U of packed red blood cells, and acute cholecystitis diagnosed clinically and treated with surgery or antibiotic therapy. Intra-abdominal sepsis was defined as a purulent collection drained either surgically or radiographically. A
Downloaded From: http://archsurg.jamanetwork.com/ by a Penn State Milton S Hershey Med Ctr User on 05/24/2015
Table 1 .—Patient Characteristics at Treatment
Age, y Sex, ICU,
%M %
Study Entry*
2.—Mortality and Complications*
No Treatment
Treatment
63.2 ±4.2
66.4 ±3.4
Mortality, No.
55
68
Major complications
(No.)_83 (15/18)_82 (18/22)_
Ventilator
dependent, % (No.) Entry albumin, g/L (g/dL) Entry APACHE score
Table
72
(13/18)_59 (13/22)_
19±1
(1.9±0.1)
18.4 ±2.3
21 ±1
(2.1 ±0.1)
18.0 ±1.7
*ICU indicates intensive care unit; APACHE, acute physiology and chronic health evaluation scoring system.13 Values for age, albumin level, and APACHE score are means ±SEMs. No difference is significant.
No Treatment
(%)_7/18 (38.9)_6/22
Total No.
40
(27)
33
Patients, No. (%)_16/18 (89)_17/22 (77.3) Complications per 2.2 ±0.6 1.5 ±0.4 patient, mean ± SEM *No difference is
significant.
treatment and nontreatment groups at entrance are outlined in Table 1. There were no significant differences in sex, age,
percentage in the ICU, percentage who were ventilator de¬ pendent, serum albumin concentrations, or APACHE II scores at the time of study entrance. Ninety-three percent of
30(3.0)
patients were on the surgical services.
20
(2.0) U
18(1.8) U 4
6
Days in Study Mean ( ± SEM) serum albumin concentrations in the treatment (solid line) and nontreatment (broken line) groups over the first 10 days of
study.
wound dehiscence required surgical closure. Fistulas were demon¬ strated radiographically, and wound infections required wound open¬ ing and drainage of purulent material. All patients were followed up until discharge or death, with the exception of two patients (one in each group) who remained hospital¬ ized at the study closure. Both of these patients had been hospitalized and actively enrolled in the study for greater than 2 months at the study closure. Data were analyzed by the unpaired Student's t test, x2 analysis, or Wilcoxon tests where appropriate.16 We considered P
