Human Albumin Use in Adults in U.S.

Academic
Medical Centers
Jose I. Suarez, MD1; Renee H. Martin, PhD2; Samuel F. Hohmann, PhD3; Eusebia Calvillo, RN1;
Eric M. Bershad, MD1; Chethan P. Venkatasubba Rao, MD1; Alexandros Georgiadis, MD1;
Oliver Flower, MD4; David Zygun, MD5; Simon Finfer, MD4,6

Objective: To determine rates and predictors of albumin adminis-
tration, and estimated costs in hospitalized adults in the United
States.
Design: Cohort study of adult patients from the University Health-
System Consortium database from 2009 to 2013.
Setting: One hundred twenty academic medical centers and 299
affiliated hospitals.
Patients: A total of 12,366,264 hospitalization records.
1
Division of Vascular Neurology and Neurocritical Care, Department of
Neurology, Baylor College of Medicine, Houston, TX.
2
Divison of Biostatistics, Data Coordinating Unit, Department of Public
Health Sciences, Medical University of South Carolina, Charleston, SC.
3
Comparative Data and Informatics Research, University HealthSystem
Consortium, Chicago, IL.
4
Malcolm Fisher Department of Intensive Care Medicine, Royal North
Shore Hospital, University of Sydney, Sydney, NSW, Australia.
5
Division of Critical Care Medicine, Faculty of Medicine and Dentistry, Uni-
versity of Alberta, Edmonton, AL, Canada.
6
The George Institute for Global Health, University of Sydney, Sydney,
NSW, Australia.
Supplemental digital content is available for this article. Direct URL cita-
tions appear in the printed text and are provided in the HTML and PDF
versions of this article on the journals website (http://journals.lww.com/
ccmjournal).
Dr. Suarez disclosed other support. Dr. Finfer is a member of the Interna-
tional Sepsis Forum (ISF) council. The ISF receives funding from compa-
nies involved in sepsis diagnostics and treatment. The ISF has contributed
to his travel expenses to attend two ISF council meetings and accompa-
nying symposia each year from the time he joined the council in 2008.
Companies sponsoring ISF during Dr. Finfers membership of ISF coun-
cil: Eisai, Siemens, Agennix, Astra Zeneca, BD diagnostics, bioMerieux,
Interventions: Analysis of rates and predictors of albumin admin-
istration, and estimated costs.
Measurements and Main Results: Overall the proportion of admis-
sions during which albumin was administered increased from
6.2% in 2009 to 7.5% in 2013; absolute difference 1.3% (95%
CI, 1.301.40%; p < 0.0001). The increase was greater in surgi-
cal patients from 11.7% in 2009 to 15.1% in 2013; absolute dif-
ference 3.4% (95% CI, 3.263.46%; p < 0.0001). Albumin use
varied geographically being lowest with no increase in hospitals
in the North Eastern United States (4.9% in 2009 and 5.3% in
2013) and was more common in bigger (> 750 beds; 5.2% in
2009 and 7.3% in 2013) compared to smaller hospitals (< 250
beds; 4.4% in 2009 to 6.2% in 2013). Factors independently
associated with albumin use were appropriate indication for albu-
min use (odds ratio, 65.220; 95% CI, 62.45968.103); surgical
admission (odds ratio, 7.942; 95% CI, 7.8897.995); and high
severity of illness (odds ratio, 8.933; 95% CI, 8.8259.042). Total
estimated albumin cost significantly increased from $325 million
in 2009 to $468 million in 2013; (absolute increase of $233 mil-
lion), p value less than 0.0001.
Conclusions: The proportion of hospitalized adults in the United
States receiving albumin has increased, with marked, and cur-
rently unexplained, geographic variability and variability by hospital
size. (Crit Care Med 2017; 45:e16e22)
Key Words: albumin; critical care; hospital costs; hospital mortality;
treatment outcome

BRAHMS/ThermoFisher, Lilly, Roche, Spectral, Toray, Philips, Apex, Fer-
ring, BioCritica, PPTA. Dr. Finfer has delivered lectures at industry spon-
sored symposia at ISF meetings with honoraria for such lectures donated
directly to ISF between 2008 and 2012: Eli Lilly (one lecture), PPTA (two
lectures). He has received consulting fee from Edwards to ISF for consult-
ing regarding glucose sensors at ISICEM 2012 also donated to ISF. He
disclosed that his institution (The George Institute for Global Health and
the University of Sydney) received funding from CSL Bioplasma, Frese-
nius Kabi, and Baxter. The remaining authors have disclosed that they do
not have any potential conflicts of interest.
For information regarding this article, E-mail: jisuarez@bcm.edu
Copyright 2016 by the Society of Critical Care Medicine and Wolters
Kluwer Health, Inc. All Rights Reserved.
DOI: 10.1097/CCM.0000000000002010
A
lbumin is the most abundant protein in human plasma
(1), and responsible for 7080% of the colloid osmotic
pressure of normal plasma. Albumin also binds and
transports naturally occurring, therapeutic and toxic materials
in the circulation (1, 2).
Albumin administration in hospitalized patients has been
largely based on the observation that hypoalbuminemia is
associated with increased hospital mortality (3). Albumin ther-
apy also has multiple potential effects (4), leading to albumin
administration in various critical scenarios. However, albumin
therapy has been associated with controversy. Earlier reports

e16 www.ccmjournal.org January 2017 Volume 45 Number 1

Copyright 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

indicated that it was associated with increased in-hospital mor-
tality (48). Subsequent studies found significant small-trial
bias and that there was no evidence that albumin significantly
affected mortality (912). The Saline versus Albumin Fluid
Evaluation (SAFE) study investigators reported the first large-
scale randomized controlled clinical trial to examine the effect
of type of resuscitation fluid on mortality (13). The study found
that administration of either 4% albumin or normal saline for
fluid resuscitation resulted in similar 28-day mortality. The
SAFE investigators performed a post hoc analysis of the sub-
group of traumatic brain injury patients and found that resusci-
tation with 4% albumin was associated with increased mortality
in this population (14). However, albumin administration may
offer some benefit by reducing morbidity in patients with sep-
sis and hospitalized patients in general (15, 16). Assessing the
impact of all those studies on the use of albumin in daily clini-
cal practice has been difficult to determine with certainty. An
international study documented annual use of albumin and
synthetic colloids in several industrialized countries (excluding
the United States) between 1995 and 2006 (17). The investiga-
tors found that reliable information on colloid use was difficult
to obtain and data were not available for all countries or years.
There are no data on the trends and predictors of use, and costs
in hospitalized patients receiving albumin in the United States,
despite the perception that albumin is often administered inap-
propriately (18).
We therefore studied albumin use in hospitalized adults in
the United States among patients treated at academic medi-
cal centers that participated in the University HealthSystem
Consortium (UHC) clinical database between 2009 and 2013.
We calculated trends in rates and modelled predictors of albu-
min use. We also performed a sub analysis of trends in deliv-
ered dose and total cost to purchase albumin.
MATERIALS AND METHODS
Study Sample
Institutional review board approval for this study was obtained
through the Baylor College of Medicine Human Investigation
Committee. We used the UHC database to identify all hospital-
ized adult patients ( 18 yr old) who received albumin between
2009 and 2013. UHC is an alliance of more than 90% of aca-
demic medical centers in the United States (120 academic
medical centers and their 299 affiliated hospitals) (19). Mem-
bers have access to a clinical database with discharge data that
allows hospitals to compare clinical performance (2024). The
UHC database is a collection of patient-level Uniform Bill-04
billing data from all participating hospitals (25). The discharge
abstract data contain information regarding patient demo-
graphics, ICU admission, estimated direct cost of albumin,
and in-hospital morbidity. For further clarification of the des-
ignation of severity of illness (SOI), UHC uses a combination
of the 3M All Payer Refined-Diagnosis Related Group (DRG)
grouper and the UHC Complication Profiler in conjunction
with data on specific characteristics (29 specific comorbidities,
age, sex, race, admission status, and payor class) to construct
Critical Care Medicine www.ccmjournal.org e17
Copyright 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Online Clinical Investigations
risk adjustment regression models that predict direct cost and
inpatient mortality (26). The UHC risk adjustment is accom-
plished with four steps: 1) assignment of Medicare Severity-
DRG as defined by the Centers for Medicare and Medicaid
Services (CMS); 2) identification of a patient population for
model generation; 3) use of multiple regression techniques
to predict direct cost and probability of mortality based on
normative patient population; and 4) assignment of expected
direct cost and mortality to every patient in the database (27).
All this allows the SOI and risk of mortality to be disease specific.
The SOI classification used for risk stratification consists of four
severity categories: minor, moderate, major, and extreme. Flags
for comorbid conditions in the database are based on definitions
by the Agency for Healthcare Research and Quality (28).
We also classified hospitals according to number of staffed
beds into four categories: < 250, 250499, 500749, and > 750.
Furthermore, we grouped hospitals into four regions accord-
ing to the U.S. Census Bureau: Northeast, South, Midwest, and
West (29).
Patient Characteristics
We extracted data on patients age, sex, ethnicity, race, comor-
bidities, and type of hospital admission. Categories of race and
ethnicity follow the CMS methodology (30). We identified
primary and secondary diagnoses from International Classifi-
cation of Diseases, 9th Edition (ICD-9) codes reported in the
database of all patient hospitalizations for any cause. We also
gathered all comorbidities listed for each patient and hospi-
talization. In addition, we extracted SOI category at admis-
sion, which is based on the initial diagnosis. Type of hospital
admission was classified as medical and surgical. The latter
was stratified according to DRGs. We abstracted data on in-
hospital complications defined as any medical condition that
occurred during hospitalization and which was not present or
documented at the time of admission.
Outcomes
The methodology used by UHC for defining and collecting
albumin administration and cost data has been similar, and the
number of participating hospitals has remained approximately
the same since 2009.
Albumin Administration. We located the type and amount
of albumin administered using the hospital charge codes cor-
responding to parenteral administration of albumin. In addi-
tion, we abstracted information on whether albumin was
administered while patients were in the ICU or elsewhere in
the hospital. Furthermore, we considered appropriateness of
albumin administration and classified it into three categories
(indications for appropriate use, indications for occasionally
appropriate use, and inappropriate indications) according to
published consensus (31). Although there is an ICD-9 diagno-
sis code for abnormal albuminemia, directionality of the con-
dition is not specified. Therefore, we were unable to determine
whether albumin was administered for hypoalbuminemia. In
addition, we were unable to determine whether albumin was
administered for intravascular volume expansion.
Suarez et al
Albumin Cost. We estimated direct cost of albumin from
the purchasing cost of albumin available in the database as
reported by each hospital.
Statistical Analysis
We present the overall rate of albumin administration per
annum as percentages. We tried to identify predictors of
albumin administration using logistic regression models and
adjusting for age, sex, comorbidities, underlying SOI, medi-
cal or surgical care, year of albumin administration, albumin
administration in the ICU, any hospital complications, appro-
priateness of albumin use, number of staffed hospital beds,
and U.S. geographic location. We chose these covariates once
we found strong evidence of an association, judged according
to the p values (< 0.05) from the Wald test. Changes over time
in human albumin use were studied by means of a linear mixed
model. In these models, we used interaction tests to determine
differences in the covariate of interest and slope of trajectories
between the patients that received albumin and those who did
not. We tested performance of logistic regression models in
terms of discrimination by using a concordance (C) statistic.
We constructed two logistic regression models. The first model
included the entire dataset. The other model was used to eval-
uate internal validity and included data that were randomly
selected from 1% of the total study sample (32, 33).
Statistical analyses were performed with the use of SAS soft-
ware, version 9.2 (SAS Institute, Cary, NC). p values are two-
sided, and p values of 0.05 or less were considered to indicate
statistical significance.
RESULTS
Patient Characteristics and Comorbidities
The total sample consisted of 12,366,264 hospitalizations; albu-
min was administered during 849,161 hospitalizations (6.9%).
Patient age, sex, and racial/ethnic distribution were different
between the group that received albumin and the group that
did not (p < 0.0001 for all comparisons; Table S1, Supplemen-
tal Digital Content 1, http://links.lww.com/CCM/C219); how-
ever, the change over time within each group was small: age
increased from 53.4 to 54.9 years, male proportion increased
from 44.5% to 45.4%, and race/ethnicity proportion for whites
increased from 59.6% to 62.9%. The group that received albu-
min was older, and included a lower proportion of women and
a higher proportion of white individuals (Table S1, Supple-
mental Digital Content 1, http://links.lww.com/CCM/C219).
Several comorbidities were common in both groups including
hypertension, diabetes, deficiency anemias, and fluid and elec-
trolyte disorders. Renal failure was more common in the group
that received albumin, whereas chronic pulmonary disease was
more common in the group that did not receive albumin. The
group of patients that received albumin had a higher admis-
sion SOI category with a greater proportion of patients in the
major and extreme categories (p < 0.0001). The most frequent
diagnoses codes for patients receiving albumin were cardiovas-
cular conditions and sepsis. The most frequent surgical DRGs
e18 www.ccmjournal.org January 2017 Volume 45 Number 1
Copyright 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
for patients receiving albumin were cardiac valve replace-
ment (with or without major complications), extracorporeal
membrane oxygenation or patients undergoing tracheostomy
for prolonged mechanical ventilation, coronary artery bypass
grafting, and major small and large bowel procedures.
Proportion of Admission Where Albumin
Administered
Rates of albumin administration increased from 6.2% in 2009
to 7.5% in 2013; absolute difference 1.3% (95% CI, 1.30
1.40%; p < 0.0001) (Table S1, Supplemental Digital Content 1,
http://links.lww.com/CCM/C219) (Fig. 1). The increase was
more pronounced in surgical patients. In the latter, rates of
albumin administration increased from 11.7% in 2009 to
15.1% in 2013; absolute difference 3.4% (95% CI, 3.263.46%;
p < 0.0001 for trend). In contrast, rates of albumin administra-
tion in medical patients only increased from 2.5% in 2009 to
3.0% in 2013; absolute difference 0.5% (95% CI, 0.460.53%).
Rates of albumin use significantly decreased for patients with
sepsis from 14.3% in 2009 to 12.9% in 2013; absolute differ-
ence 1.4% (95% CI, 0.821.92%; p < 0.0001 for trend).
Most instances of albumin administration occurred while
patients were in the ICU (61%) (Table 1). However, albu-
min administration in ICU decreased from 61.3% in 2009 to
59.5% in 2013; absolute difference 1.8% (95% CI, 1.751.85%;
p < 0.0001 for trend). Albumin administration increased out-
side the ICU environment from 38.7% in 2009 to 40.5% in
2013; absolute difference 1.8% (95% CI, 1.751.85%; p < 0.001
for trend). The rate of albumin administration in patients
with appropriate indication increased from 47.7% in 2009
to 50.5% in 2013; absolute difference 2.8% (95% CI, 2.70
2.90%; p < 0.001 for trend) (Table S2, Supplemental Digital
Content 2, http://links.lww.com/CCM/C220). The rate of
albumin administration in patients with occasionally appro-
priate indication increased from 15.6% in 2009 to 18.1% in
2013; absolute difference 2.5% (95% CI, 2.462.54%; p < 0.001
for trend). In contrast, the rate of albumin administration in
patients with inappropriate indication modestly decreased
from 6.6% in 2009 to 5.8% in 2013; absolute difference 0.8
(95% CI, 0.720.88%).
Albumin administration increased in all the U.S. geographic
regions except in the Northeast where it remained unchanged
from 4.94% in 2009 to 5.28% in 2013; absolute difference 0.34
(95% CI, 0.320.36) (Table1). Overall albumin administra-
tion was lowest in the Northeast compared to all other U.S.
regions (p < 0.001 for trend). The rate of albumin adminis-
tration increased in all the categories of staffed hospital beds
with the highest increment in hospital with greater than 750
beds from 5.25% in 2009 to 7.27% in 2013; absolute difference
2.02% (95% CI, 1.832.21; p < 001 for trend).
Predictors of Albumin Administration
We found that in the logistic regression model, the factors that
were independently associated with albumin use were male sex,
geographic location, year of admission, indications of albu-
min use, any hospital complication, any comorbidity, surgical
 Online Clinical Investigations

admission, and severity of

underlying illness (Table2). The
strongest association was seen
for the following factors: appro-
priate indication of albumin
use (odds ratio [OR], 65.220;
95% CI, 62.45968.103), surgi-
cal admission (OR, 7.942; 95%
CI, 7.8897.995), and extreme
underlying SOI (OR, 8.933;
95% CI, 8.8259.042). All the
models performed well (C sta-
tistics, > 0.85) (Table2; and
Table S3, Supplemental Digi-
tal Content 3, http://links.lww.
com/CCM/C221).

Type of Albumin
Administered and
Estimated Direct Cost
The use of each of the avail-
able albumin vials in the
United States from 2009 to
Figure 1. Graphic representation of trends of albumin use in patients receiving albumin as a percent of all
2013 was variable (Table 3).
hospitalized patients by year and type of hospitalization. The most frequently used

Table 1. Outcomes for Hospitalized Patients Between 2009 and 2013 by Albumin Usage
Patient Characteristic 2009 2010 2011 2012 2013

No of patients (%)
Albumin used 120,904 (6.2) 143,847 (6.5) 176,981 (6.8) 195,537 (7.1) 211,895 (7.5)
No albumin used 1,834,096 (93.8) 2,078,335 (93.5) 2,435,419 (93.2) 2,564,847 (92.9) 2,604,403 (92.5)
Patient location at the time of
albumin administration (%)
ICU 74,121 (61.3) 91,958 (63.9) 111,206 (62.8) 117,496 (60.1) 126,114 (59.5)
Other 46,783 (38.7) 51,889 (36.1) 65,775 (37.2) 78,041 (39.9) 85,781 (40.5)
Albumin use according to staffed
hospital beds (%, weighted
average)
< 250 4.4 5.3 5.9 6.0 6.2
250499 6.2 6.5 6.9 6.8 6.9
500749 6.4 6.5 6.7 7.0 8.0
> 750 5.2 5.8 6.4 7.0 7.3
Albumin use according to hospital
geographic location (%,
weighted average)
Northeast 4.9 5.2 5.1 5.3 5.3
South 6.4 6.8 7.6 8.3 8.8
Midwest 5.7 6.1 6.7 7.2 8.3
West 7.0 7.5 8.1 8.4 8.8
Albumin refers to human albumin.

Critical Care Medicine www.ccmjournal.org e19

Copyright 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Suarez et al

Table 2. Predictors of Albumin Administration 2009 to 2013. Most of the albumin was administered when
From Logistic Regression Model Including patients were in the ICU. We found in logistic regression model-
ing that the main factors independently associated with albu-
the Entire Sample Size (C = 0.916)
min administration were appropriate indication of albumin use,
Factors OR 95% Wald CI surgical admission, and extreme underlying SOI at admission.
We also found differences in albumin use according to U.S. geo-
Staffed beds 1.000 1.0001.000
graphic region (lowest in the Northeast), and staffed hospital
Sex (male) 1.246 1.2381.253 beds (highest increase in hospitals with > 750 beds). All this sug-
Geographic location (reference gests that, contrary to a commonly held perception (18), major-
Northeast) ity of albumin administered occurred in those patients with
West 1.744 1.7251.763
appropriate indications. In addition, our analysis suggests that
there is important variability in albumin use depending on U.S.
South 1.559 1.5471.571 geographic region and number of staffed hospital beds.
Midwest 1.400 1.3891.411
Strengths and Weaknesses
Year of admission (reference
2009) The main strength of our study is the analysis of a large dataset
of patients admitted to a majority of academic centers in the
2010 1.069 1.0581.079
United States. In addition, the definitions used for the database
2011 1.197 1.1851.208 have been standardized, and its risk adjustment model is vali-
2012 1.223 1.2111.234 dated and commonly used for comparison of institutions and
estimations of quality of care delivered (2024, 27).
2013 1.284 1.2721.296
This study has several weaknesses. First, our findings are
Indications for albumin use observational and as such a number of factors may have
(reference no albumin use) accounted for the observed changes in albumin use rates over
Appropriate 65.220 62.45968.103 time (e.g., patient selection, healthcare system changes, and
secular changes). Second, we can only infer associations not
Occasionally appropriate 3.413 3.3843.443
causal relationships due to the study design. Third, our study
Inappropriate 1.497 1.4831.511 was limited to the time period between 2009 and 2013. Our
Any hospital complication 1.135 1.1231.147 decision to include data from 2009 to 2013 was based on issues
related to the UHC database design. UHC methodology for
Any comorbidity 1.166 1.1551.176
defining and collecting clinical outcomes has been similar and
Surgical vs medical cases 7.942 7.8897.995 the number of participating hospitals has remained approxi-
Severity of underlying illness mately the same since 2009. Fourth, we were unable to ascer-
(reference minor) tain the temporal relationship between albumin administration
Extreme 8.933 8.8259.042 and reported complications. Fifth, we were unable to determine
with certainty whether albumin was used for intravascular vol-
Major 4.189 4.1454.234 ume expansion or hypoalbuminemia. Finally, our analysis and
Moderate 1.950 1.9301.971 results only represent data from hospitalized patients in the
OR = odds ratio. United States available in the UHC database. However, as the
UHC database contains data from major academic centers in
the United States and their affiliates, which include a substantial
vial per hospital was 25%-50mL with a modest increase
number of community hospitals, it is likely that our results are
from 1,988 in 2009 to 2,038 in 2013; absolute difference of
representative of hospital practice in the United States.
50 (95% CI, 48.5851.42). The total estimated direct cost of
albumin and the estimated direct cost of albumin per hos-
Implications of the Study
pital increased from 2009 to 2013. The total estimated albu-
Even though other studies have reported on albumin use in
min cost across the study cohort was $235 million in 2009
many other countries, ours is the first such report in the United
and $468 million in 2013, (absolute increase of $233 million),
States. Despite great variation across countries, overall albumin
p value less than 0.0001 for trend, and the estimated albu-
use decreased between 1995 and 1999 and then remained stable
min cost per hospital across study cohort hospitals increased
through 2006 in Canada, Europe, Australia, and New Zealand
from $3.9 million in 2009 and $4.9 million in 2013 (absolute
(17). It is not possible to determine if a similar pattern of use
increase of $1.0 million), p value less than 0.0001 for trend.
occurred in the United States due to the absence of reliable
data. However, our findings suggest that there has been a steady
DISCUSSION increase in albumin use in the United States in the past 5 years.
Our study describes an increase in albumin use in the United There are some potential explanations for the increase in
States in patients with data recorded in the UHC database from albumin use. The most common conditions associated with

e20 www.ccmjournal.org January 2017 Volume 45 Number 1

Copyright 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
 Online Clinical Investigations

Table 3. Type and Volume of Albumin Used and Purchasing Cost of Albumin for Patients
From 2009 to 2013 (Only Using Encounters With Albumin Usage)
Patient Characteristic 2009 2010 2011 2012 2013

Type of albumin infused, vialsa

5%-50 mL 309 302 312 317 474
5%-250 mL 1,908 1,819 1,745 1,772 1,700
25%-20 mL 142 131 85 76 86
25%-50 mL 1,988 2,047 1,933 1,985 2,038
Albumin volume infused per hospital (mL) b
50,192 49,747 47,192 48,151 48,306
Total estimated albumin cost (US $M) c
235 317 384 440 468
Estimated albumin cost per hospital (US $M) d
3.91 4.29 4.42 4.68 4.92
M = million.
a
Estimates are based on dividing total vials by the number of hospitals.
b
Estimates are based on dividing total volume by the number of hospitals.
c
Estimates are based on multiplying industry unit cost by total volume infused.
d
Estimates are based on dividing total estimated cost by the number of hospitals.
Albumin refers to human albumin.

albumin use were cardiac surgery, cardiovascular diseases, and
sepsis. A meta-analysis of controlled trials of albumin ver-
sus crystalloids for extracorporeal circuit priming in cardiac
surgery published in 2004 suggested that albumin priming
reduced the decline in platelet count compared to crystalloids,
favorably influenced both colloid oncotic pressure, and on-
bypass positive fluid balance (34). Such data may explain the
increase in albumin use in patients undergoing cardiac surgery.
We also found that patients with sepsis received albumin. A
predefined subgroup analysis of patients with severe sepsis in
the SAFE study suggested that fluid resuscitation with albumin
compared to saline might reduce mortality in adult patients
(13). Although a subsequent more detailed analysis supported
these findings, the authors cautioned that confirmatory stud-
ies were still required (35). The latter, added to the safety pro-
file of albumin may have also influenced its administration
in patients with sepsis (15, 16, 36). In addition, in 2008, the
Surviving Sepsis Campaign Guidelines recommended albumin
or crystalloids for initial volume resuscitation in patients with
severe sepsis or septic shock (37).
Future Research
Our study suggests that albumin use is associated with signifi-
cant costs. Further randomized controlled trials or compara-
tive effectiveness studies are needed to determine the efficacy
and effectiveness of albumin in various medical and surgical
conditions; these studies should incorporate cost-effectiveness
analyses. Abandoning albumin use in conditions where efficacy
or effectiveness cannot be proven might contribute to signifi-
cant cost savings to the healthcare system in the United States.
Albumin use was lower in women. The data available in our
study database preclude us from determining whether there is
a true gender bias in albumin administration. Further stud-
ies are needed to investigate this in more detail. However, we
can speculate that similar to what has been reported for other
conditions, women may have been less likely to receive albu-
min (3840). Possible reasons for this differential use of albu-
min may be perceived difference in injury severity, likelihood
of benefiting from albumin therapy, or simply subconscious
gender bias.
CONCLUSIONS
The proportion of hospitalizations in the United States where
albumin was administered and associated cost increased over a
5-year period. The main factors associated with albumin admin-
istration were appropriate indication of albumin use, surgical
admission, and SOI at admission. There was significant, and
currently unexplained, variability in albumin administration
depending on U.S. geographic location and hospital size.
REFERENCES
1. Peters T (Ed): All About Albumin: Biochemistry, Genetics, and Clinical
Applications. San Diego, Academic Press, 1996
2. Baxter: Buminate. Available at: http://www.baxter.com.pr/healthcare_
professionals/products/buminate.html. Accessed September 6, 2016
3. Goldwasser P, Feldman J: Association of serum albumin and mortality
risk. J Clin Epidemiol 1997; 50:693703
4. Foley EF, Borlase BC, Dzik WH, et al: Albumin supplementation in
the critically ill. A prospective, randomized trial. Arch Surg 1990;
125:739742
5. Cochrane Injuries Group Albumin Reviewers: Human albumin admin-
istration in critically ill patients: Systematic review of randomized con-
trolled trials. BMJ 1998; 317:235240
6. Offringa M: Excess mortality after human albumin administration in
critically ill patients. Clinical and pathophysiological evidence sug-
gests albumin is harmful. BMJ 1998; 317:223224
7. Schierhout G, Roberts I: Fluid resuscitation with colloid or crystalloid
solutions in critically ill patients: A systematic review of randomised
trials. BMJ 1998; 316:961964
8. Vermeulen LC Jr, Ratko TA, Erstad BL, et al: A paradigm for con-
sensus. The University Hospital Consortium guidelines for the use of
albumin, nonprotein colloid, and crystalloid solutions. Arch Intern Med
1995; 155:373379

Critical Care Medicine www.ccmjournal.org e21

Copyright 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Suarez et al
9. Wilkes MM, Navickis RJ: Patient survival after human albumin admin-
istration. A meta-analysis of randomized, controlled trials. Ann Intern
Med 2001; 135:149164
10. Roberts I, Blackhall K, Alderson P, et al: Human albumin solution for
resuscitation and volume expansion in critically ill patients. Cochrane
Database Syst Rev 2011; 11:CD001208
11. Bunn F, Trivedi D, Ashraf S: Colloid solutions for fluid resuscitation.
Cochrane Database Syst Rev 2011; 3:CD001319
12. Perel P, Roberts I, Ker K: Colloids versus crystalloids for fluid resus-
citation in critically ill patients. Cochrane Database Syst Rev 2013;
2:CD000567
13. Finfer S, Bellomo R, Boyce N, et al; SAFE Study Investigators: A com-
parison of albumin and saline for fluid resuscitation in the intensive
care unit. N Engl J Med 2004; 350:22472256
14. Myburgh J, Cooper JD, Finfer S, et al: Saline or albumin for fluid resus-
citation in patients with traumatic brain injury. N Engl J Med 2007;
357:874884
15. Delaney AP, Dan A, McCaffrey J, et al: The role of albumin as a resus-
citation fluid for patients with sepsis: A systematic review and meta-
analysis. Crit Care Med 2011; 39:386391
16. Vincent JL, Navickis RJ, Wilkes MM: Morbidity in hospitalized patients
receiving human albumin: A meta-analysis of randomized, controlled
trials. Crit Care Med 2004; 32:20292038
17. Jones D, McEvoy S, Merz TM, et al: International albumin use: 1995
to 2006. Anaesth Intensive Care 2010; 38:266273
18. Tanzi M, Gardner M, Megellas M, et al: Evaluation of the appropri-
ate use of albumin in adult and pediatric patients. Am J Health Syst
Pharm 2003; 60:13301335
19. University HealthSystem Consortium: Available at: https://www.uhc.
edu/cps/rde. Accessed September 6, 2016
20. Varelas PN, Conti MM, Spanaki MV, et al: The impact of a neuroin-
tensivist-led team on a semiclosed neurosciences intensive care unit.
Crit Care Med 2004; 32:21912198
21. Simorov A, Reynoso JF, Dolghi O, et al: Comparison of perioperative
outcomes in patients undergoing laparoscopic versus open abdomi-
noperineal resection. Am J Surg 2011; 202:666670
22. Pallati P, Buettner S, Simorov A, et al: Trends in adolescent bariatric
surgery evaluated by UHC database collection. Surg Endosc 2012;
26:30773081
23. Hinojosa MW, Murrell ZA, Konyalian VR, et al: Comparison of laparo-
scopic vs open sigmoid colectomy for benign and malignant disease at
academic medical centers. J Gastrointest Surg 2007; 11:14231429
24. Jalisi S, Bearelly S, Abdillahi A, et al: Outcomes in head and neck
oncologic surgery at academic medical centers in the United States.
Laryngoscope 2013; 123:689698
25. Clinical Database/Resource Manager User Manual, October 2010.
University HealthSystem Consortium Web site. Available at: https://
www.cms.gov/Medicare/CMS-Forms/CMS-Forms/CMS-Forms-
Items/CMS1196256.html. Accessed September 6, 2016
e22 www.ccmjournal.org January 2017 Volume 45 Number 1
Copyright 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
26. Avril RF, Goldfield N, Hughes JS, et al: All Patient Refined Diagnosis
Related Groups (APR-DRGs) Version 20.0: Methodology Overview.
Wallingford, CT, 3M Health Information Systems, 2003
27. Iezzoni, Lisa I: Risk Adjustment for Measuring Health Care Outcome,
Fourth Edition. Ann Arbor, MI, Health Administration Press, 2012
28. Elixhauser A, Steiner C, Harris DR, et al: Comorbidity measures for
use with administrative data. Med Care 1998; 36:35
29. U.S. Census Bureau. Available at: https://www.census.gov/. Accessed
September 6, 2016
30. Eicheldinger C, Bonito A: More accurate racial and ethnic codes for
Medicare administrative data. Health Care Financ Rev 2008; 29:2742
31. Liumbruno GM, Bennardello F, Lattanzio A, et al; Italian Society of
Transfusion Medicine and Immunohaematology (SIMTI): Recommen-
dations for the use of albumin and immunoglobulins. Blood Transfus
2009; 7:216234
32. Miller ME, Hui SL, Tierney WM: Validation techniques for logistic
regression models. Stat Med 1991; 10:12131226
33. Altman DG, Vergouwe Y, Royston P, et al: Prognosis and prognostic
research: Validating a prognostic model. BMJ 2009; 338:b605
34. Russell JA, Navickis RJ, Wilkes MM: Albumin versus crystalloid for
pump priming in cardiac surgery: Meta-analysis of controlled trials. J
Cardiothorac Vasc Anesth 2004; 18:429437
35. Finfer S, McEvoy S, Bellomo R, et al: Impact of albumin compared to
saline on organ function and mortality of patients with severe sepsis.
Intensive Care Med 2011; 37:8696
36. Vincent JL, Wilkes MM, Navickis RJ: Safety of human albuminseri-
ous adverse events reported worldwide in 1998-2000. Br J Anaesth
2003; 91:625630
37. Dellinger RP, Levy MM, Carlet JM, et al; International Surviving Sep-
sis Campaign Guidelines Committee; American Association of Crit-
ical-Care Nurses; American College of Chest Physicians; American
College of Emergency Physicians; Canadian Critical Care Society;
European Society of Clinical Microbiology and Infectious Diseases;
European Society of Intensive Care Medicine; European Respiratory
Society; International Sepsis Forum; Japanese Association for Acute
Medicine; Japanese Society of Intensive Care Medicine; Society of
Critical Care Medicine; Society of Hospital Medicine; Surgical Infec-
tion Society; World Federation of Societies of Intensive and Critical
Care Medicine: Surviving Sepsis Campaign: International guidelines
for management of severe sepsis and septic shock: 2008. Crit Care
Med 2008; 36:296327
38. Bonita R, Beaglehole R: Women and NCDs: Overcoming the neglect.
Glob Health Action 2014; 7:23742
39. Gomez D, Haas B, de Mestral C, et al: Gender-associated differ-
ences in access to trauma center care: A population-based analysis.
Surgery 2012; 152:179185
40. Dreyer RP, Beltrame JF, Tavella R, et al: Evaluation of gender differ-
ences in Door-to-Balloon time in ST-elevation myocardial infarction.
Heart Lung Circ 2013; 22:861869