Full text of DOI:10.1016/S0272-6386(00)70050-3

Mild Renal Failure Is Associated With Adverse Outcome
After Cardiac Valve Surgery
Robert J. Anderson, MD, Maureen O’Brien, MS, Samantha MaWhinney, ScD,
Catherine B. VillaNueva, RN, Thomas E. Moritz, MS, Gulshan K. Sethi, MD,
William G. Henderson, PhD, Karl E. Hammermeister, MD,
Frederick L. Grover, MD, and A. Laurie Shroyer, PhD
● The present study was performed to ascertain whether the presence of mild renal failure (defined as a serum
creatinine concentration of 1.5 to 3.0 mg/dL) is an independent risk factor for adverse outcome after cardiac valve
surgery. An extensive set of preoperative and postoperative data was collected in 834 prospectively evaluated
patients undergoing cardiac valve surgery at 14 Veterans Affairs Medical Centers. Univariate and multivariable
analyses were performed to determine whether an independent association of mild renal dysfunction with adverse
outcomes was present. Patients with mild renal failure had significantly greater 30-day mortality rates (P ؍ 0.001;
16% versus 6%) and frequency of postoperative bleeding (P ؍ 0.023; 16% versus 8%), respiratory complications
(P ؍ 0.02, 29% versus 16%), and cardiac complications (P ؍ 0.002; 18% versus 7%) than patients with normal renal
function (serum creatinine F1.5 mg/dL) when controlling for multiple other variables. The presence of a serum
creatinine concentration of 1.5 to 3.0 mg/dL is significantly and independently associated with adverse outcomes
after cardiac valve surgery.
ꢀ 2000 by the National Kidney Foundation, Inc.
INDEX WORDS: Renal failure; surgical outcomes.
ORE THAN 600,000 cardiac surgical pro-
found that an elevated preoperative serum creati-
M cedures are performed annually within nine concentration (median, 1.7 mg/dL) was sig-
nificantly associated with adverse outcome after
coronary artery bypass grafting (CABG) indepen-
dent of multiple other variables.31 The present
analysis was performed to further test our hypoth-
esis that mild renal insufficiency is an indepen-
dent risk factor for adverse outcome after cardiac
valve surgery. To extend our testing of this hy-
pothesis, we selected a surgical population with a
significantly greater mortality than for CABG.
Specifically, we analyzed a prospectively ob-
tained Veterans Affairs (VA) database containing
an extensive set of preoperative and postopera-
tive variables in more than 800 patients undergo-
the United States. Delineation of risk factors for
adverse outcome in patients undergoing cardiac
surgery is of considerable potential value. For
example, if modifiable risk factors are present,
then individualized preoperative management
could potentially improve outcome. Also, preop-
erative risk stratification not only allows for
improved informed consent, but also can result
in improved assessment of quality of care by the
development of risk-adjusted outcome models.1
Most of the models developed to assess risk
for adverse outcome after cardiac surgery focus
on such variables as priority of surgery, previous
cardiac surgery, cardiopulmonary functional sta-
tus, and associated comorbid conditions.2-23
Among the comorbid conditions frequently pre-
sent in patients undergoing cardiac surgery are
heart failure, hypertension, vascular disease, dia-
betes mellitus, and advanced age. These condi-
tions are often associated with some degree of
renal insufficiency.12 Physiological abnormali-
ties resulting from renal insufficiency could con-
tribute to adverse outcome after cardiac surgery.
Although it is clear that severe renal failure is
associated with poor outcome after cardiac sur-
gery,24-30 there is limited information on the
impact of mild renal failure on cardiac surgical
outcome.2,4,6,12,13
From the Department of Veterans Affairs Medical Center;
and the Departments of Medicine and Surgery, University of
Colorado Health Sciences Center, Denver, CO.
For the Participants of the Department of Veterans Affairs
Cooperative Study on Processes, Structures, and Outcomes
of Care in Cardiac Surgery.
Received September 7, 1999; accepted in revised form
January 4, 2000.
Supported in part by the Veterans Affairs Health Services
Research and Development Service and the Cooperative
Studies Program.
Address reprint requests to Robert J. Anderson, MD,
UCHSC, Box B180, 4200 East Ninth Ave, Denver, CO
80262. E-mail: robert.anderson@uchsc.edu
ꢀ 2000 by the National Kidney Foundation, Inc.
0272-6386/00/3506-0016$3.00/0
Recently, using multivariable analyses, we
doi:10.1053/ajkd.2000.7475
American Journal of Kidney Diseases, Vol 35, No 6 (June), 2000: pp 1127-1134
1127

1128
ANDERSON ET AL
nally, whether a procedure other than mitral or aortic valve
replacement was performed. Of note, this model also consid-
ers whether CABG was performed. After the risk model was
built without serum creatinine concentration, the two groups
with elevated serum creatinine concentrations were subse-
quently entered to determine their effect.
Logistic regression models were analyzed using a for-
ward, stepwise selection method with the model entry crite-
rion set at P less than 0.25, allowing SAS (SAS Institute Inc,
Cary, NC) to select the best-fit model, and P less than 0.05 as
the model exit criterion, respectively. The models were
evaluated by both the predictive power of the model using
the concordance or C-index and the Hosmer-Lemeshow
goodness-of-fit test statistic.37,38
Five different outcome variables (30-day surgical death
and four groups of postoperative complications) were exam-
ined with multivariable logistic regression analysis. The four
groups of postoperative complications studied included the
presence or absence of: (1) bleeding (defined as dissemi-
nated intravascular coagulopathy, gastrointestinal bleeding,
or return to the operating room for cardiothoracic bleeding),
(2) infections (defined as pneumonia or harvest site wound,
chest wound, mediastinitis, or endocarditis infections), (3)
cardiac complications (defined as cardiac arrest or conges-
tive heart failure symptoms), and (4) respiratory complica-
tions (defined as mechanical ventilation for Ͼ48 hours or the
need for reintubation). These clinical variables were exam-
ined because renal failure results in physiological abnormali-
ties that could potentially predispose to several of these
events.32
Because cost data were not available, length-of-stay vari-
ables were developed and analyzed as surrogates for cost.
Postoperative length of stay was defined as the time from
surgery to discharge. Total length of stay was defined as the
time from admission to discharge. In the VA health care
system, medical care days are not tracked separately from
non–medical care days (for example, days within the hospi-
tal that a patient awaited discharge for transportation rea-
sons). Thus, the length-of-stay data may be biased because
greater than 50% of VA patients travel in excess of 50 miles
for inpatient procedures. The median distance from home to
the center at which the surgical procedure was performed for
patients with serum creatinine levels less than 1.5 mg/dL and
1.5 to 3.0 mg/dL was 90 miles. A stratified Kaplan-Meier
was used to differentiate whether the impact of renal func-
tion was an independent predictor of length of stay.
ing cardiac valve surgery, some of whom con-
comitantly underwent CABG.
METHODS
Data used for this study were extracted from VA Coopera-
tive Study no. 5, Processes, Structures, and Outcomes of
Care in Cardiac Surgery. Patients were entered onto this
study from September 8, 1992, through December 31, 1996,
from 14 of 43 VA hospitals with cardiac surgical programs.
The VA hospitals were chosen to represent a spectrum of
hospitals across the distribution of risk-adjusted 30-day
mortality from cardiac surgery, as well as hospitals represen-
tative of VA facilities east and west of the Mississippi River.
A total of 4,969 patients were entered onto the study; 427
patients underwent cardiac valve surgery without an addi-
tional cardiac surgical procedure and 407 patients underwent
cardiac valve surgery along with CABG. Data obtained from
these 834 cardiac valve surgical patients, who had at least
one baseline serum creatinine determination performed within
2 weeks of surgery, form the basis of this report.
Before data analysis, we categorized patients by three
levels of baseline serum creatinine concentration. We di-
vided patients into clinically relevant categories of serum
creatinine levels less than 1.5 mg/dL (a normal value for
most hospital laboratories), 1.5 to 3.0 mg/dL (a value consid-
ered by many to be mild to moderate renal failure, usually
asymptomatic), and greater than 3.0 mg/dL (considered
indicative of more advanced renal failure32,33). No data on
creatinine clearance were obtained, and a more detailed
analysis (such as use of the Cockcroft-Gault nomogram) of
renal function was not available from the database.
Extensive preoperative and postoperative data were ob-
tained on each patient, described in detail in a previous
publication.34 Statistical comparisons of these variables in
the three groups of patients categorized by serum creatinine
concentration were performed by chi-square test, Fisher’s
exact test, and Mantel-Haenszel test, when appropriate, for
categorical variables and by Kruskal-Wallis tests for continu-
ous variables. Length-of-stay variables were analyzed using
Kaplan-Meier analysis censoring for death before discharge.
Wilcoxon’s rank-sum and log-rank test statistics were used
to compare strata. To ascertain if renal failure was indepen-
dently associated with adverse outcomes that were dichoto-
mous in nature, multivariable analyses were performed
using a logistic regression risk model.
Using the current method of mortality risk model develop-
ment used by the Continuous Improvement in Cardiac
Surgery Study,2,5,35,36 as well as their patient population, we
built a model to estimate the risk for death and disallowed
baseline serum creatinine concentration from entering the
model. The significant variables in predicting 30-day surgi-
cal death in this model were previous myocardial infarction,
priority of surgery (elective, urgent, or emergent), Canadian
Cardiovascular Society angina classification, age, prior heart
surgery, history of peripheral vascular disease, current
smoker, American Society of Anesthesiology classification,
documented ST depression on an electrocardiogram, pres-
ence of audible rales, body surface area, functional status
(independent, partially dependent, or totally dependent),
performance of mitral valve replacement surgery, and fi-
RESULTS
The mean age of the patient population was
65.7 years, and 97% were men. Eighteen percent
of the patients had undergone prior heart surgery.
Of the 834-patient study population, 76.4% had a
baseline serum creatinine level less than 1.5
mg/dL; 21.9%, between 1.5 and 3.0 mg/dL; and
1.7%, greater than 3.0 mg/dL. Median total (pre-
operative and postoperative) length of stay for
the entire group was 17.0 days, whereas median
postoperative length of stay was 11.0 days. The

RENAL FAILURE AND HEART VALVE SURGERY
1129
30-day surgical mortality rate for the study popu-
lation was 8.6%.
performed in patients with baseline serum creati-
nine levels less than 1.5, 1.5 to 3.0, and greater
than 3.0 mg/dL was 47%, 57%, and 64%, respec-
tively.
The characteristics of patients categorized by
baseline serum creatinine concentrations are listed
in Table 1. Median serum creatinine concentra-
tions in the three study groups were 1.1, 1.7, and
5.5 mg/dL. In general, patients with a serum
creatinine level of 1.5 mg/dL or greater tended to
be older; have a greater incidence of previous
myocardial infarction; more frequently have a
history of diabetes mellitus, peripheral vascular
disease, hypertension, and prior heart surgery;
and have a higher American Society of Anesthe-
siology classification. The frequency of concomi-
tant CABG procedure was 46%, 60%, and 50%
in the groups with serum creatinine concentra-
tions less than 1.5, 1.5 to 3.0, and greater than 3.0
mg/dL, respectively. As noted in the Methods
section, all these variables, including CABG,
were eligible for entry in our model-building
process. Patients with a serum creatinine level of
1.5 to 3.0 mg/dL also more frequently had active
endocarditis and current digoxin and diuretic use
than the group with a serum creatinine level less
than 1.5 mg/dL. The frequency with which a
mechanical valve replacement procedure (versus
biological valve or other type of procedure) was
Several post–valve surgery outcomes in the
three groups of patients, categorized by baseline
serum creatinine concentration, are listed in Table
2. Univariately, the serum creatinine grouping
was significantly associated with increased 30-
day surgical mortality, as well as cardiac arrest,
low cardiac output, heart block requiring pace-
maker, prolonged ventilation, and neurological
(coma for Ͼ24 hours and stroke), gastrointesti-
nal (bleeding), and renal (renal failure requiring
dialysis after discharge) complications. One in-
fectious complication (pneumonia), but not other
infectious complications (mediastinitis, chest
wound infection, and endocarditis), had a signifi-
cant difference among the three groups of pa-
tients.
Using multivariable logistic regression analy-
sis, mild to moderate renal failure (serum creati-
nine, 1.5 to 3.0 mg/dL) compared with normal
renal function (serum creatinine Ͻ1.5 mg/dL)
significantly increased 30-day surgical mortality
(P ϭ 0.001). The C-index was 0.692 for the
mortality model without serum creatinine level;
Table 1. Characteristics of Study Subjects Categorized by Baseline Serum Creatinine Concentration
Serum Creatinine
Ͻ1.5 mg/dL
(n ϭ 637)
Serum Creatinine
1.5-3.0 mg/dL
(n ϭ 183)
Serum Creatinine
Ͼ3.0 mg/dL
(n ϭ 14)
Variable
P
Median serum creatinine (mg/dL)
Age (y)
Preoperative balloon pump
Prior myocardial infarction
1.1
64.7
10 (2)
148 (23)
100 (16)
388 (61)
203 (32)
316 (50)
59 (9)
1.7
69.1
4 (2)
5.5
65.3
1 (7)
Ͻ0.001
Ͻ0.001
0.171
0.001
0.007
0.152
0.449
0.001
0.220
0.045
0.184
0.001
0.001
0.122
0.006
0.273
0.104
0.001
0.012
0.100
70 (38)
48 (26)
127 (70)
67 (36)
123 (67)
25 (14)
12 (7)
44 (24)
73 (40)
115 (63)
40 (22)
39 (21)
36 (20)
49 (27)
125 (68)
78 (43)
86 (47)
8 (57)
2 (14)
12 (86)
4 (29)
9 (64)
2 (14)
0 (0)
4 (29)
4 (29)
11 (79)
3 (21)
4 (29)
1 (7)
Prior heart surgery
New York Heart Association classification III or IV
Angina functional classification III or IV
American Society Anesthesiology classification IV and V
Urgent or emergent surgical priority
Active endocarditis
History of cerebral vascular disease
History of peripheral vascular disease
History of hypertension
History of chronic obstructive lung disease
Presence of diabetes mellitus
Current smoker
Current bronchodilator use
Current diuretic use
16 (3)
118 (19)
122 (19)
301 (47)
99 (16)
89 (14)
145 (23)
110 (17)
296 (46)
188 (29)
364 (57)
0 (0)
5 (36)
3 (21)
9 (64)
Current digoxin use
Mechanical valve
NOTE. Values expressed as number (percent) unless otherwise noted.

1130
ANDERSON ET AL
Table 2. Postoperative Complication Rates in Three Groups of Patients Undergoing Cardiac Valve Surgery
Classified by Baseline Serum Creatinine Concentration
Serum Creatinine
Ͻ1.5 mg/dL
(n ϭ 637)
Serum Creatinine
1.5-3.0 mg/dL
(n ϭ 183)
Serum Creatinine
Ͼ3.0 mg/dL
(n ϭ 14)
Complication
30-day mortality (%)
Myocardial infarction (%)
Cardiac arrest (%)
Low cardiac output (%)
Heart block requiring pacemaker (%)
Renal failure requiring dialysis at discharge (%)
On ventilator Ͼ24 h (%)
Reintubated (%)
P
6.0
5
3
18
12
2
15
8
2
4
2
6
15.8
5
35.7
21
21
21
36
21
21
14
14
7
Ͻ0.001
0.071
0.001
Ͻ0.001
0.029
0.001
0.001
0.092
0.001
0.007
0.009
0.001
10
32
15
10
27
13
5
9
4
14
Coma Ͼ24 h (%)
Stroke (%)
Gastrointestinal bleeding (%)
Pneumonia (%)
14
14
this increased to 0.711 when the two serum
creatinine variables (serum creatinine levels of
1.5 to 3.0 and Ͼ3.0 mg/dL) were added to the
baseline model. This change in C-index, al-
though in the direction toward improved predic-
tive power, was not statistically significant.
3.0 mg/dL were 2.45 (95% CI, 1.43 to 4.19) for
30-day mortality, 1.81 (95% CI, 1.09 to 3.03) for
bleeding, 1.62 (95% CI, 1.08 to 2.44) for ventila-
tory complications, and 2.25 (95% CI, 1.34 to
3.77) for cardiac complications. Multivariable
analysis with regard to selected infectious com-
plications (pneumonia, chest wound infection,
mediastinitis, and endocarditis) showed no sig-
nificant effect when a serum creatinine level of
1.5 to 3.0 mg/dL was added to the original model
(odds ratio, 1.48; 95% CI, 0.93 to 2.36).
With regard to risk for bleeding complications
and renal function, data listed in Table 3 indicate
a significant increase in average number of units
of transfused red blood cells and fresh frozen
plasma, but not platelets, in the intraoperative
and postoperative periods in the group of pa-
tients with a serum creatinine concentration of
1.5 to 3.0 mg/dL. The percentage of patients with
Also using multivariable logistic regression
analysis, mild to moderate renal failure (serum
creatinine, 1.5 to 3.0 mg/dL) compared with
normal renal function (serum creatinine Ͻ 1.5
mg/dL) significantly increased bleeding compli-
cations (defined as disseminated intravascular
coagulation, gastrointestinal hemorrhage, or tho-
racic hemorrhage sufficient to require reopera-
tion; P ϭ 0.023); cardiac complications (defined
as cardiac arrest and low cardiac output; P ϭ
0.002); and ventilatory complications (defined as
Ͼ48 hours on mechanical ventilation and/or rein-
tubation; P ϭ 0.020) independent of other vari-
ables. Multivariable logistic regression analysis
did not show that mild to moderate renal failure
(serum creatinine, 1.5 to 3.0 mg/dL) was associ-
ated with an increase in selected infectious com-
plications (pneumonia, mediastinitis, chest wound
infection, or endocarditis; P ϭ 0.0972) over the
group with normal renal function (serum creati-
nine Ͻ1.5 mg/dL).
Table 3. Transfusions in Patients Undergoing Cardiac
Valve Surgery Categorized by Baseline Serum
Creatinine Concentration
Serum
Serum
Creatinine
Creatinine
Ͻ1.5 mg/dL 1.5-3.0 mg/dL
P
Red blood cells
Fresh frozen plasma
Platelets
3.27
(n ϭ 459)
1.90
(n ϭ 261)
5.18
6.01
(n ϭ 156)
3.30
(n ϭ 97)
7.69
Ͻ0.001
Ͻ0.002
0.069
Multivariable comparisons were performed us-
ing the group with baseline serum creatinine
level less than 1.5 mg/dL as the reference for
comparison. The odds ratios with 95% confi-
dence intervals (CIs) when comparing patients
with a serum creatinine level less than 1.5 mg/dL
with those with a serum creatinine level of 1.5 to
(n ϭ 293)
(n ϭ 93)
NOTE. Values expressed as average number of units
transfused intraoperatively or postoperatively.

RENAL FAILURE AND HEART VALVE SURGERY
1131
baseline serum creatinine concentrations less than
1.5, 1.5 to 3.0, and greater than 3.0 mg/dL
administered red blood cells (72%, 85%, and
79%, respectively), fresh frozen plasma (41%,
53%, and 50%, respectively), but not platelet
transfusions (46%, 51%, and 64%, respectively)
increased significantly (P ϭ 0.001; P ϭ 0.004;
P ϭ 0.186, respectively) in parallel with the
magnitude of increase in serum creatinine level.
With regard to length of stay, stratified Kaplan-
Meier analysis showed univariately that the group
with a baseline serum creatinine level of 1.5 to
3.0 mg/dL also had a significantly longer total
length of stay than the group with a baseline
serum creatinine level less than 1.5 mg/dL (16
versus 20 days, respectively; P Ͻ 0.001).
preoperative renal dysfunction was a significant
determinant of surgical mortality in patients un-
dergoing valve replacement for aortic stenosis.
Subsequently, Grover et al,2 using a large VA
database, found that the mean serum creatinine
concentration was significantly less (1.3 mg/dL)
in survivors than nonsurvivors (1.7 mg/dL; P Ͻ
0.001) of cardiac valve surgery.
The present analyses confirm and extend some
of these previous observations. We found that
several adverse outcomes occurred with signifi-
cantly greater frequency in patients with preop-
erative serum creatinine levels of 1.5 to 3.0
mg/dL compared with patients with serum creati-
nine concentrations less than 1.5 mg/dL. Multi-
variable logistic regression analysis found that
serum creatinine levels of 1.5 to 3.0 mg/dL were
independently associated with significantly in-
creased 30-day surgical mortality. Stratified
Kaplan-Meier analysis found that preoperative
serum creatinine levels of 1.5 to 3.0 mg/dL were
also associated with increased length of stay.
Collectively, these observations clearly docu-
ment that preoperative mild to moderate renal
failure is significantly associated with adverse
outcome and enhanced resource use after cardiac
valve surgery in the studied population.
The present study also delineates some of the
specific complications encountered with in-
creased frequency in patients with mild renal
failure undergoing heart valve surgery. Our mul-
tivariable analysis found that mild renal failure
was significantly associated with selected bleed-
ing, respiratory, and cardiac complications inde-
pendent of other variables. Our results with re-
gard to a significant association of mild renal
failure with increased bleeding complications
after cardiac valve surgery is corroborated by our
finding of a significant increase in red blood cell
and fresh frozen plasma transfusions adminis-
tered to the group with renal failure.
DISCUSSION
Preoperative assessment before cardiac and
other types of surgery has traditionally focused
on the cardiovascular and pulmonary systems.39-41
In 1992, Browner et al4 suggested that impaired
renal function is an important predictor of surgi-
cal mortality. In this study, including 474 male
veterans with either known coronary artery dis-
ease or at high risk for coronary artery disease, a
reduced estimated creatinine clearance was a
significant (odds ratio, 6.8; 95% CI, 2.8 to 16.0)
independent predictor of postoperative mortality
after noncardiac surgery.4 There is, however,
limited information on the impact of renal failure
on cardiovascular surgical outcomes.
A modest experience with cardiac surgery,
including valve replacement surgery, in patients
with end-stage renal disease undergoing chronic
renal replacement therapy has been published.24-30
These studies show that cardiac surgery can be
successfully performed in dialysis patients with
subsequent improved functional status. How-
ever, a relatively high rate of complications and
mortality was observed.24-30
With regard to lesser degrees of renal failure,
the presence of mild to moderate renal dysfunc-
tion has been reported to significantly increase
surgical mortality after carotid endarterectomy,
but not after elective repair of infrarenal abdomi-
nal aortic aneurysms.42-44 We and others have
found that mild renal failure (serum creatinine,
1.5 to 3.0 mg/dL) significantly increases mortal-
ity after CABG.6,12,13,31 Regarding cardiac valve
replacement, Scott et al16 reported in 1985 that
Two observations regarding our results are
particularly worthy of comment. First, the degree
of renal failure (median serum creatinine, 1.7
mg/dL) present in the group with mild renal
failure that we studied was modest and would be
considered by many clinicians to be unlikely to
adversely affect surgical outcome per se. It is
noteworthy, however, that two other studies found
that a similar, modest degree of renal failure is
associated with adverse outcome after general

1132
ANDERSON ET AL
and heart valve replacement surgery.2,4 More-
over, a recent case-control study showed that
induction of even a very modest degree of radio
contrast–associated renal failure significantly in-
creased mortality and the occurrence of bleeding
and respiratory complications.45
patients that we analyzed with more advanced
renal failure (serum creatinine Ͼ3.0 mg/dL) is
limited. This small sample size prevents ascertain-
ment of whether more advanced degrees of renal
failure are associated with a greater rate of com-
plications.
Second, the findings of the present study are
remarkably consistent with those obtained in a
multivariable analysis of the impact of mild renal
failure on CABG outcome that we recently per-
formed.31 Although the absolute rate of selected
cardiac, bleeding, and respiratory complications
that we observed after valve surgery in patients
with a serum creatinine concentration less than
1.5 mg/dL (7%, 8%, and 16%, respectively) was
approximately twofold greater than that ob-
served after CABG in patients with a serum
creatinine concentration less than 1.5 mg/dL, the
rate of these complications was approximately
doubled in patients with mild renal failure after
both CABG and valve surgery. Together, these
observations suggest that mild renal failure is an
important, independent risk factor in at least two
types of cardiac surgery, and the impact of mild
renal failure to increase morbidity is similar in
the two operations and occurs independently of
the absolute level of risk.
In summary, we find that a preoperative serum
creatinine level of 1.5 to 3.0 mg/dL is signifi-
cantly and independently associated with in-
creased 30-day mortality and selected bleeding,
cardiac, and respiratory complications after car-
diac valvular surgery. Although further study is
required to delineate the mechanism(s) of this
adverse effect, this population of patients should
be considered at enhanced risk for adverse out-
come after valve surgery.
REFERENCES
1. Daley J: Criteria by which to evaluate risk-adjusted
outcome programs in cardiac surgery. Ann Thorac Surg
58:1827-1835, 1994
2. Grover FL, Hammermeister KE, Burchfiel C: Initial
report of the VeteransAdministration preoperative risk assess-
ment study for cardiac surgery. Ann Thorac Surg 50:12-28,
1990
3. Higgins TL, Estafanous FG, Loop FD, Beck GJ, Blum
JM, Parandi L: Stratification of morbidity and mortality
outcome by preoperative risk factors in coronary artery
bypass patients. JAMA 267:2344-2348, 1992
There are several limitations to the present
analysis. First, the observational nature of our
study does not allow us to delineate pathophysi-
ological mechanisms involved to explain our
results. Also, even though we analyzed a very
comprehensive set of variables, it is possible that
we did not identify and analyze all potential
confounding factors. Second, the nature of the
VA patient population that we studied does not
allow us to be able to comfortably generalize our
observations to other patient populations. Third,
we used a discretionary definition to categorize
the presence and degree of renal failure. It is
likely that some of our smaller, older patients
undergoing valve surgery with a serum creati-
nine concentration less than 1.5 mg/dL had mild
to moderate renal failure. Also, we were not able
to clearly ascertain if the mild renal failure pre-
sent was either hemodynamically mediated or
caused by chronic intrinsic renal disease. To-
gether, these factors limit somewhat our ability
to more accurately assess the contribution of
mild renal failure to adverse outcomes after
cardiac valve surgery. Finally, the number of
4. Browner WS, Li J, Mangano DT: In-hospital and
long-term mortality in male veterans following noncardiac
surgery. JAMA 268:228-232, 1992
5. Hammermeister KE, Johnson R, Marshall G, Grover
FL: Continuous assessment and improvement in quality of
care. A model from the Department of Veterans Affairs
cardiac surgery. Ann Surg 219:382-390, 1994
6. O’Connor GT, Plume SK, Olmstead EM, Coffin LN,
Morton JR, Maloney CT, Nowicki ER, Levy DG, Tryzelaar
JF, Hernandez F, Adrian L, Casey KJ, Bandy D, Soule DN,
Marrin CA, Nugent WE, Charlesworth DC, Clough R, Katz
S, Leavitt BJ, Wenneberg JE: Multivariate prediction of
in-hospital mortality associated with coronary artery bypass
graft surgery. Circulation 85:2110-2118, 1992
7. Clark RE: The society of thoracic surgeons national
data base status report. Ann Thorac Surg 57:20-26, 1994
8. Hannan EL, Kilburn H, Rocz M, Sheilds E, Chassin
MR: Improving the outcomes of coronary artery bypass
surgery in New York state. JAMA 271:761-766, 1994
9. Parsonnet V, Dean D, Bernstein AD: A method of
uniform stratification of risk for evaluating the results of
surgery in acquired adult heart disease. Circulation 79:S3-
S12, 1989 (suppl I)
10. Tu JV, Mazer CD, Levington C, Armstrong PW,
Naylor CD: A predictive index for length of stay in the
intensive care unit following cardiac surgery. Can Med
Assoc J 151:177-185, 1994

RENAL FAILURE AND HEART VALVE SURGERY
1133
11. Turner JS, Morgan CJ, Thakrar B, Pepper JR: Difficul-
ties in predicting outcome in cardiac surgery patients. Crit
Care Med 23:1843-1850, 1995
12. Mangano CM, Diamondestone LS, Ramsay JG, Ag-
garwal A, Herskovitz A, Mangano DT: Renal dysfunction
after myocardial revascularization: Risk factors, adverse
outcomes and hospital resource utilization. Ann Intern Med
128:174-180, 1998
13. Shroyer AL, Grover FL, Edwards FH: 1995 Coronary
artery bypass risk model: The Society of Thoracic Surgeons
adult cardiac national data base. Ann Thorac Surg 65:879-
884, 1998
14. Connolly HM, Oh JK, Orszulak TA, Osborn SL,
Roger VL, Hodge DO, Bailey KR, Seward JB, Tajik AJ:
Aortic valve replacement for aortic stenosis with severe left
ventricular dysfunction. Circulation 95:23395-23400, 1997
15. Tseng EE, Lee CA, Cameron DE, Suart RS, Geene
PS, Sussman MS, Watkins L, Gardner TJ, Baumgartner WA:
Aortic valve replacement in the elderly. Risk factors and
long-term results. Ann Surg 225:793-880, 1997
16. Scott WC, Miller DC, Haverich A, Dawkins K,
Mitchell RS, Jamieson SW, Baldwin JC, Shumway NE:
Determinants of operative mortality for patients undergoing
aortic valve replacement. J Thorac Cardiovasc Surg 89:400-
413, 1985
17. Scott WC, Miller DC, Haverich A, Mitchell RC, Oyer
PE, Stinson EB, Jamieson SW, Baldwin JC, Shumway NE:
Operative risk of mitral valve replacement: Discriminant
analysis of 1329 procedures. Circulation 72:S108-S119,
1985 (suppl 2)
18. Christakis GT, Kormos RL, Weisel RD, Fremes SE,
Tong CP, Herst JA, Schwartz L, Mickelborough LL, Scully
HE, Goldman BS: Morbidity and mortality in mitral valve
surgery. Circulation 72:S120-S128, 1985 (suppl 2)
19. Magovern JA, Pennock JL, Campbell DB, Pierce
WS, Waldhausen JA: Risks of mitral valve replacement and
mitral valve replacement with coronary artery bypass. Ann
Thorac Surg 39:346-352, 1985
20. Magovern JA, Pennock JL, Campbell DB, Pae WE,
Bartholomew M, Pierce WS, Waldhaasen JA: Aortic valve
replacement and combined aortic valve replacement and
coronary artery bypass grafting: Predicting high risk groups.
J Am Coll Cardiol 9:38-43, 1987
cardiac surgery. Scand J Thorac Cardiovasc Surg 29:187-
193, 1995
26. Owen CH, Cummings RG, Sell TL, Schwab SJ, Jones
RH, Glower DO: Coronary artery bypass grafting in patients
with dialysis-dependent renal failure. Ann Thorac Surg 58:
1729-1733, 1994
27. Lucke JC, Samy RN, Atkins BZ, Silvestry SC, Doug-
las JM, Schwab SJ, Wolfe WG, Glower DO: Results of valve
replacement with mechanical and biological prostheses in
chronic renal dialysis patients. Ann Thorac Surg 64:129-
133, 1997
28. Garrido P, Bobadilla JF, Albertos J, Gonzales SJ,
Bastida E, Vallego JL, Arcas R: Cardiac surgery in patients
under chronic hemodialysis. Eur J Cardiothorac Surg 9:6-39,
1995
29. Blum U, Skupin M, Wagner R, Vallbracht C, Opper-
mann F, Satter P: Early and long-term results of cardiac
surgery in dialysis patients. Presse Med 23:28-31, 1994
30. Ko W, Kreiger KH, Isom OW: Cardiopulmonary
bypass procedures in dialysis patients. Ann Thorac Surg
55:677-684, 1993
31. Anderson RJ, O’Brien M, MaWhinney S, VillaNueva
CB, Mortiz TE, Sethi G, Henderson WG, Hammermeister
KE, Grover FL, Shroyer AL: Renal failure predisposes to
adverse outcome after coronary artery bypass surgery. Kid-
ney Int 55:1057-1062, 1999
32. Knochel JP: Biochemical alterations in advanced
uremic failure, in Jacobson HR, Striker GE, Klahr S (eds):
The Principles and Practice of Nephrology. Philadelphia,
PA, Decker, 1991, pp 682-684
33. Luke RG: Chronic renal failure—A vasculopathic
state. N Engl J Med 339:841-843, 1998
34. Shroyer AL, London MJ, VillaNueva C, Sethi G,
Marshall G, Mortiz T, McCarthy MJ, Grover FL, Hammer-
meister KE: The processes, structures and outcomes of care
in cardiac surgery study protocol. Med Care 33:S17-S25,
1995 (suppl)
35. Grover FL, Johnson RR, Shroyer AL, Marshall G,
Hammermeister KE: The Veterans Affairs Continuous Im-
provement in Cardiac Surgery study. Ann Thorac Surg
58:1845-1851, 1994
36. Marshall G, Shroyer AL, Grover FL, Hammermeister
KE: Bayesian-logit model for risk assessment in coronary
artery bypass grafting. Ann Thorac Surg 57:1440-1450,
1994
37. Hosmer DW, Lemeshow S: Applied Logistic Regres-
sion. New York, NY, Wiley, 1989, pp 82-132
38. Ash A, Schwartz M: Evaluating the performance of
risk-adjustment methods: Dichotomous variables, in Iezzoni
L (ed): Risk Adjustment for Measuring Health Care Out-
comes. Ann Arbor, MI, Health Administration Press, 1994,
pp 313-346
39. Palda VA, Detsky AS: Perioperative assessment and
management of risk from coronary artery disease. Ann
Intern Med 127:313-328, 1997
40. Eagle KA, Brumdage BH, Chaitman BR, Evy GA,
Fleisher LA, Hertzer NR, Leppo JA, Ryan T, Schlant RC,
Spencer HW, Spittell JA, Twiss RD, Ritchie JL, Cheitlin,
MD, Gardner TJ, Carson A, Lewis RP, Gibbons RJ, Ryan TJ:
21. Teoh KH, Christakis GT, Weisel RD, Madonik MM,
Ivanov J, Wong PY, Mee AV, Levitt D, Benak A, Reilly P:
The determinants of mortality and morbidity after multiple-
valve operations. Ann Thorac Surg 43:353-358, 1987
22. Christakis GT, Weisel RD, David TE, Salerno TA,
Ivanov J: Predictors of operative survival after valve replace-
ment. Circulation 78:S25-S34, 1988 (suppl 1)
23. Sethi GK, Miller DC, Souchek J: Clinical hemody-
namic and angiographic predictors of operative mortality in
patients undergoing single valve replacement. J Thorac
Cardiovasc Surg 93:884-897, 1987
24. Schreiber S, Korzets A, Parvsner E, Wolloch Y:
Surgery in chronic dialysis patients. Isr J Med Sci 31:479-
483, 1995
25. Ashrof SS, Shaukat N, Kamaly ID, Durrani A, Doran
B, Grotte GJ, Keenan DJ: Determinants of early and late
mortality in patients with end-stage renal disease undergoing

1134
ANDERSON ET AL
Guidelines for perioperative cardiovascular evaluation for
noncardiac surgery. J Am Coll Cardiol 27:910-948, 1996
41. Mangano DT, Goldman L: Preoperative assessment
of patients with known or suspected coronary disease. N
Engl J Med 333:1750-1756, 1995
42. Plecha EJ, King TA, Pitluk HC, Rubin JR: Risk
assessment in patients undergoing carotid endarterectomy.
Cardiovasc Surg 1:30-32, 1993
endarterectomy justified in patients with severe chronic
renal insufficiency? Ann Vasc Surg 11:115-119, 1997
44. Komori K, Kuma S, Eguchi D, Okazaki J, Kawasaki
K, Onchara T, Yamamura S, Itoh H, Sugimachi K: Surgical
strategy of abdominal aortic aneurysm with preoperative
renal failure. Eur J Vasc Endovasc Surg 14:105-108, 1997
45. Levy EM, Visioli CM, Horwitz RI: The effect of
acute renal failure on mortality. A cohort analysis. JAMA
275:1487-1494, 1996
43. Rigdon EE, Monajjem N, Rhodes RS: Is carotid