World Journal of Cardiovascular Surgery, 2011, 1, 18-23
doi:10.4236/wjcs.2011.12004 Published Online December 2011 (http://www.SciRP.org/journal/wjcs)
Fluid Overload after Coronary Artery Bypass Grafting
Surgery Increases the Incidence of Post-Operative
Complications
Jean-François Morin1*, Berguez Mistry2, Yves Langlois1, Felix Ma1, Patrick Chamoun2,
Christina Holcroft3
1Cardiac Surgery, SMBD Jewish General Hospital, McGill, Montreal, Canada
2Anesthesia, SMBD Jewish General Hospital, McGill, Montreal, Canada
3Clinical Epidemio logy and Public Health Research Center, Jewish General Hospital, McGill, Montreal, Canada
Received September 16, 2011; revised November 15, 2011; accepted November 30, 2011
Abstract
This study is a prospective trial comparing the incidence of post-operative complications to fluid status in
patients undergoing coronary artery bypass grafting (CABG) surgery. One hundred and nine subjects under-
going CABG surgery at the Jewish general hospital were recruited over a 5 months period in the year 2006.
All of the patients underwent CABG surgery “on pump”. Post operative fluid overload was measured by
weight gain. Using logistic regression with complications (major vs. minor only/none) as an outcome and
fluid overload as a covariate, the risk of major complications significantly increases for fluid overload ≥5 kg
compared to 1 - 5 kg (p < 0.001), while the risk for ≤1 kg is not significantly different from 1 - 5 kg. Also,
the risk of major complications significantly (p = 0.012) increases for days with fluid overload ≥5 days in
comparison to ≤1 day.
Keywords: Fluid Overload, Complications, Coronary Artery Bypass Grafting
1. Introduction
Open heart surgery requiring cardio-pulmonary bypass
(CPB) has been well known to result in post-operative
fluid overload. Several mechanisms have been postulated
as possible causes for this hypervolemia: excessive fluid
replacement pre and intra-operatively and post-operative
crystalloid administration in an effort to correct hy-
potension and a systemic inflammatory reaction caused by
CPB which ultimately results in increased capillary per-
meability and leakage of fluid in the extra-vascular space.
It is believed that fluid overload participates in the
pathogenesis of important severe clinical complications
[1]. The relation between fluid overload and early
post-operative complications has not been studied ade-
quately yet.
Our present proposed study therefore is a prospective
trial comparing the frequency of post-operative complica-
tions to fluid status in patients undergoing coronary ar-
tery bypass grafting (CABG) surgery.
2. Method
One hundred and nine subjects undergoing CABG sur-
gery at the Jewish General Hospital were recruited over a
5 month period in the year 2006. Exclusion criteria in-
cluded redo CABG surgery, valve surgery, combined
procedures (valve + CABG), patients with pre-operative
renal failure (documented serum creatinine level of more
than 2.0 mg/dl or on dialysis) and pre-operative conges-
tive heart failure. Congestive heart failure diagnosis re-
quired at least 3 of the following: presence of dyspnea,
rales thought to represent pulmonary congestion, periph-
eral edema, cardiomegaly on chest x-ray and chest x-ray
compatible with interstitial edema.
Eighty-eight patients were male with an average age
of 62.1 ± 10.4 years old and 21 patients were female with
a mean age of 64.7 ± 11.5 years old. For the entire group
of patients, the mean Canadian Cardio-vascular Society
Class was 3.8, with a median Parsonnet Score of 7 and a
mean left ventricular ejection fraction of 48.6%. All the

J.-F. MORIN ET AL.19
patients underwent coronary artery bypass grafting sur-
gery “on pump”. The median number of grafts per pa-
tient was 4, with mainly the use of the left internal
mammary artery for the left anterior descending artery
and the saphenous vein grafts for the other coronary ar-
teries. The average aortic cross-clamp time and total
pump time were respectively 58 minutes and 77 minutes.
The average amount of fluids given intra-operatively,
including pump prime, cardioplegia, crystalloids, col-
loids, blood transfusions and blood derived products was
4733.46 mL. The average amount of fluids given
post-operatively up to 24 hours, including crystalloids
and colloids, blood transfusions and blood derived prod-
ucts was 3240.04 mL. The amount of patients requiring
blood transfusions and blood derived products was 64%.
Post-operative fluid overload was measured by weight
gain, which was measured on a ScaleTronix electronic
scale on each day in the hospital after surgery for eight
days and compared to pre-operative weight. The maximum
weight gain over eight days was the value used for fluid
overload. Arbitrarily, this weight gain was divided into 3
categories: less than 1 kg, 1 to 5 kg and more than 5 kg.
The number of days in the hospital until pre-operative
weight was also calculated and defined as number of
days in the hospital with fluid overload. This variable
was categorized arbitrarily into ≤1, 2 to 4, and ≥5 days.
The complications were divided into 2 groups: major
(death, myocardial infarction, cardiac arrest, low cardiac
output syndrome, cardiac tamponade, mediastinal explo-
ration for bleeding, cerebral vascular accident, respira-
tory failure requiring prolonged intubation, renal failure
and deep sternal wound infection) and minor (atrial fib-
rillation, supra-ventricular tachycardia, new heart block,
transient ischemic attack, delirium, pneumonia, leg
wound infection, arm wound infection and superficial
sternal wound infection). The median hospital length of
stay was 5 days.
Statistical analysis to examine the association between
type of complication (major, minor only, none) and fluid
overload or days with fluid overload groups was per-
formed using Fisher’s exact test (because some table cell
counts were <5). The association between type of com-
plication and fluid overload was subsequently simplified
by dichotomizing complications group into major vs.
minor only/none. Two logistic regression models with
complications group as the outcome were run: one with
fluid overload categories and another with days with
fluid overload categories as covariates. Odds ratios,
which represent the odds of major complications in one
fluid overload group compared to the odds of major
complications in the fluid overload reference group, are
reported with corresponding 95% confidence intervals
and p-values. Statistical calculations were carried out
using Stata statistical software (Intercooled Stata v8.2,
College Station, TX, StataCorp).
3. Results
There was no operative mortality (up to one month
post-operatively). The complications and their frequen-
cies are listed in Table 1. Twenty-one patients (19%)
had a major complication, 36 (33%) had only minor
complications, and 52 (48%) had no complications.
Looking at the relationship between demographic fac-
tors and complications, we found that the median age of
patients with major, minor or no complications was
similar at 62 years old. The median left ventricular
ejection fraction was better (55%) in patients without
complications than patients that presented minor or
major complications (50%). The body mass index of
our population did not influence the frequency or sever-
ity of the complications. The gender distribution of
major, minor or no complications was respectively
15.91%, 31.82% and 52.27% for males and 33.33%,
38.10%, and 28.57% for females. For patients with
major complications, 43.36% were in NYHA class III
or IV compared to 0% in NYHA class I or II.
Type of complication (major, minor only, none) was
examined in relation to groups of fluid overload post-
operative weight gain (≤1 kg, 1-5 kg, ≥5 kg) (Figure 1).
Among the 20 patients that presented with a post-opera-
tive weight gain ≤1 kg, the counts of major, minor and
no complications were respectively 1, 7 and 12. Out of
62 patients with a weight gain of 1 to 5 kg, the counts of
major, minor and no complications were 7, 21 and 34. In
the group of 27 patients with a weight gain of ≥5 kg, the
counts of major, minor and no complications were 13, 8
and 6. The association between type of complication and
fluid overload weight gain shown in Figure 1 was statis-
tically significant (Fisher’s exact p = 0.001).
Using logistic regression with complications (major vs.
minor only/none) as an outcome and fluid overload as a
covariate (with 1 - 5 kg as the reference group because of
small numbers in the ≤1 kg category), the risk of major
complications significantly (p < 0.001) differed for fluid
overload ≥5 kg compared to 1 - 5 kg, while the risk for
≤1 kg was not significantly different from 1 - 5 kg (Ta-
ble 2).
Type of complication was also examined in relation to
number of post-operative days (≤1, 2 to 4, and ≥5 days)
where the patient remains overweight. Overall, 39 pa-
tients (36%) had fluid overload for ≤1 day, 22 (20%) for
2-4 days, and 48 (44%) for ≥5 days. Figure 2 illustrates
the frequency of type of complications in relationship to
the number of days with fluid overload. The association
shown in Figure 2 was statistically significant (Fisher’s

J.-F. MORIN ET AL. 20
Table 1. Subjects in every categor y of complications.
Complications Subjects in Category Results in %
1-Death 0/109 0%
2-MI 2/109 1.83%
3-LCOS 15/109 13.76%
4-A.Fib or SVT 61/109 55.96%
5-Heart block >24 hr 22/109 20.18%
6-Cardiac Arrest 1/109 0.91%
7-Tamponade 1/109 0.91%
8-Re-exploration 3/109 2.75%
9-TIA 0/109 0%
10-RIND 0/109 0%
11-CVA 1/109 0.91%
12-Delirium 4/109 3.66%
13-Pneumonia 0/109 0%
14-Resp. failure + intubation 4/109 3.66%
15-Renal Impairment 8/109 7.33%
16-Leg wound infection 0/109 0%
17-Arm wound infection 0/109 0%
18-Superficial sternal wound 3/109 2.75%
19-Deep sternal wound 1/109 0.91%
exact p = 0.048). Using logistic regression with compli-
cations (major vs. minor only/none) as an outcome and
days with fluid overload as a covariate, the risk of major
complications significantly (p = 0.012) differed for days
with fluid overload ≥5 days in comparison to ≤1 day
(Table 2).
4. Discussion
Open heart surgery is associated with a degree of fluid
overload. The reason being; during cardio pulmonary
bypass (CPB), the patient is exposed to the administra-
tion of a significant amount of exogenous fluids. Such
fluids include the CPB pumps prime (1200 cc), fluids
used for cardioplegia (200 cc) and other fluids given to
deal with hypotensive episodes and/or intra-operative
fluid redistribution. These fluids can often reach the
amount of 3 to 4 litres of combined colloid and crystal-
loid solutions. Often, the patient undergoes cardiac sur-
gery presenting with conditions such as heart or renal
failure and under these circumstances, a degree of fluid
overload already exists and is aggravated by CPB. Irre-
spective of the circumstances, fluid overload often par-
ticipates in the pathogenesis of post-operative hypoxemia,
myocardial oedema and organ oedema. It is also often
responsible for a delay in recovery and several clinically
important complications [1].
In our cohort of patients, the age and body mass index
did not affect the frequency or the severity of the com-
plications. However, as reported in the literature, we
have shown that female gender, patients with lower left
ventricular ejection fraction, and patients with NYHA
class III or IV are at increased risk of complication.
5. Cardiopulmonary Dysfunction
Cardiopulmonary bypass may induce alterations in in-
tra-vascular volume status and extra-vascular body water
content [2,3], leading to oedema and cardiopulmonary
dysfunction [4,5]. With the onset of CPB, the blood is
diluted by the prime volume of the extra-corporeal cir-
cuit, resulting in a decreased colloid osmotic pressure
[3-5]. CPB causes increased capillary permeability by
activating the inflammatory ediators system [6-8]. Fur- m

J.-F. MORIN ET AL.21
010 20 30 40010 20 30 40
MajorMinor OnlyNoneMaj orMino r OnlyNone
MajorMinor OnlyNone
(a) <=1 kg(b) 1-5 kg
(c) >=5 kg
Frequency
G rap hs by M ax weig ht i n 8 days Po st -Op – P re-Op wgt (Gr ouped)
Figure 1. Type of complications by fluid overload group.
05 10 15 20 2505 10 15 20 25
M aj orMinor O nlyNon eMajorMinor OnlyNone
M aj orMinor O nlyNon e
(a) <=1 day(b) 2-4 days
(c) >=5 days
Frequency
Graphs by Groups of days in hospital with fluid overload
Figure 2. Type of complications by days with fluid overload.
Table 2. Logistic regression results with complication group as the outcome (major vs. minor only/none) and fluid overload
and days with fluid overload as covariates (two separate models).
Covariate Odds ratio (95% confidence interval) p-value
Fluid overload
≤1 kg 0.414 (0.048, 3.58) 0.423
1 - 5 kg 1 (reference) -
≥5 kg 7.30 (2.45, 21.70) <0.001
Days with fluid overload
≤1 day 1 (reference) -
2 - 4 days 1.89 (0.35, 10.31) 0.460
≥5 days 5.45 (1.45, 20.55) 0.012

J.-F. MORIN ET AL. 22
thermore, the hydrostatic pressure within the capillaries
is influenced by hypothermia [9] and post-ischemic myo-
cardial dysfunction as a consequence of altered myocar-
dial lymphatic function after cardioplegic arrest [10].
Thus, net filtration of fluid into the interstitium is in-
creased during and after extra-corporeal circulation re-
sulting in elevated extra-vascular lung water content [2,
11] and oedema formation in the myocardial interstitium
[12] which can contribute to ventricular dysfunction,
which may explain the relatively high (13.07%) inci-
dence of low cardiac output syndrome in our series.
6. Other Complications
Acute reduction in cardiac output early after the operation is
the most common and most important risk factor for the
development of acute renal failure. Yet the prevalence of
acute renal failure varies among patients with low car-
diac output in part because of the role played by other
risk factors. Despite fluid overload with weight gain
post-open heart surgery, the intra-vascular compartment
may be depleted leading to renal dysfunction.
Brain embolization and hypo-perfusion with resultant
ischemia, have most generally been considered etiologic
to these organic changes. The causes of both general
neuropsychological subsystem dysfunction and localized
subsystem dysfunction (such as strokes) are the same
except in degree and/or distribution [13]. Could brain
oedema resulting from fluid overload contribute to post-
operative delirium?
In 2003, Brandstrup [14] and al. reported the effects of
intravenous fluid restriction on post-operative complica-
tions after elective colorectal resection surgery. The re-
stricted regimes aimed at maintaining pre-operative body
weight; the standard regimen resembled everyday prac-
tice. The restricted intravenous fluid regimen signifi-
cantly reduced post-operative complications both by in-
tention-to-treat (33% vs. 51%, p = 0.013) and per-pro-
tocol (30% vs. 56%, p = 0.003) analyses. The number of
both cardiopulmonary (7% vs. 24%, p = 0.007) and tis-
sue healing complications (16% vs. 31%, p = 0.04) were
significantly reduced, no patients died in the restricted
group compared to 4 deaths in the standard group (0% vs.
4.7%, p = 0.12). The reduction of cardiopulmonary com-
plications in the restricted group was pronounced. Sub-
clinical oedema in lungs and other tissues may cause
decreased tissue oxygenation [15], and may explain the
findings of significantly increased tissue healing com-
plications in the standard group.
A correlation has been reported between fluid over-
load, thrombosis [16] and prolongation of intestinal pa-
ralysis.
7. Treatment and Prevention of Fluid
Overload
The patients in our series received “standard” amounts of
fluids (mainly crystalloids) during or immediately after
surgery to maintain adequate cardiac output. In the case
of a persistent low cardiac output despite optimum in-
tra-vascular compartment measurements (adequate CVP
and PCWP), the patient would then be started on ino-
tropic treatments.
Most of our patients received Mannitol while on CPB.
Aggressive diuresis is usually started on the first post-
operative day and continued until the patients’ weight
returns to baseline.
Hemofiltration (HF) during CPB in this observational
study was not used. We use it for complex procedures
requiring prolonged CPB. Many advantages of the
hemofiltration during CPB have been mentioned in the
literature. Ultrafiltration is effective in reducing the in-
flammatory response and fluid overload induced by CPB
in paediatric patients [17-19]. In adults, Bellomo et al. [1]
found that post-operative anemia, thrombocytopenia and
hypoalbunemia were significantly decreased by the re-
moval of approximately 3.4 litres of fluid with hemo-
filtration during CPB. One of the possible benefits of such
hemoconcentration is improved post-operative haemostasis
[20-23]. Preservation of colloidal osmotic pressure may
reduce the excessive accumulation of extra-vascular wa-
ter. In their patients, large pleural effusions were signifi-
cantly less common when hemofiltration had been used
during CPB. Hemofiltration during CPB has also been
shown to remove inflammatory mediators [10]. Therefore,
hemofiltration may protect against transcapillary loss of
albumin hence reducing the accumulation of body water
during CPB. Post-operatively, in the intensive care unit,
early use of continuous venovenous hemofiltration
(CVVHF) has shown similar advantages [1].
In spite of demonstrating an association between fluid
overload and post-operative complications, this study is
probably underestimating the problem since the patients
underwent only non-complex and short procedures. Fur-
thermore, patients with pre-operative renal dysfunction
or congestive heart failure were excluded. Based on these
results, while performing open heart surgery, our goal
should be to administer a minimal amount of fluids to
maintain adequate cardiac output. As a future research
project we should aim to look at the same fluid overload
and complications on patients undergoing CABG surgery
with the use of the “mini-pump” or “OFF pump”.
8. References
[1] R. Bellomo, J. Raman and C. Ronco, “Intensive Care
Unit Management of the Critically Ill Patient with Fluid

J.-F. MORIN ET AL.23
Overload after Open-Heart Surgery,” Cardiology, Vol. 96,
No. 3-4, 2001, pp. 169-176.
doi:10.1159/000047400
[2] C. G. Olthof, P. G. Jansen, J. P. De Vries, et al., “Intersti-
tial Fluid Volume during Cardiac Surgery Measured by
Means of a Non-Invasive Conductivity Technique,” Acta
Anaesthesiol Scand, Vol. 39, 1995, pp. 508-512.
doi:10.1111/j.1399-6576.1995.tb04109.x
[3] M. E. Koller, J. Bert, L. Segadal and R. K. Reed, “Esti-
mation of Total Body Fluid Shifts between Plasma and
Interstitium in Man during Extracorporeal Circulation,”
Acta Anaesthesiologica Scandinavica, Vol. 36, No. 3,
1992, pp. 255-259.
doi:10.1111/j.1399-6576.1992.tb03460.x
[4] H. J. Geissler and S. J. Allen, “Myocardial Fluid Balance:
Pathophysiology and Clinical Implications,” Thorac Car-
diovascular Surgery, Vol. 46, Suppl 2, 1998, pp. 242-245.
doi:10.1055/s-2007-1013080
[5] T. Hachenberg, A. Tenling, H. Y. Rothen, S. O. Nystrom,
H. Tyden and G. Hedenstierna, “Thoracic Intra Vascular
and Extravascular Fluid Volumes in Cardiac Surgical Pa-
tients,” Anesthesiology, Vol. 79, 1993, pp. 976-984.
doi:10.1097/00000542-199311000-00016
[6] J. H. Levy and K. A. Tanaka, “Inflammatory Response to
Cardiopulmonary Bypass,” The Annals of Thoracic Sur-
gery, Vol. 75, No. 2, 2003, pp. 5715-5720.
doi:10.1016/S0003-4975(02)04701-X
[7] S. Wan, J. L. Leclerc and J. G. Vincent, “Inflammatory
Response to Cardiopulmonary Bypass: Mechanisms In-
volved and Possible Therapeutic Strategies,” Chest, Vol.
112, No. 3, 1997, pp. 676-692.
doi:10.1378/chest.112.3.676
[8] D. E. Chenoweth, S. W. Cooper, T. E. Hugli, R. W.
Stewart, E. H. Blackstone and J. W. Kirklin, “Comple-
ment Activation during Cardiopulmonary Bypass: Evi-
dence for Generation of C3a and C5a Anaphylatoxins,”
The New England Journal of Medicine, Vol. 304, 1981,
pp. 497-503. doi:10.1056/NEJM198102263040901
[9] J. K. Heltne, J. Bert, T. Lund, et al., “Temperature Re-
lated Fluid Extravasation during Cardiopulmonary By-
pass: An Analysis of Filtration Coefficients and Tran-
scapillary Pressures,” Acta Anaesthesiologica Scandi-
navica, Vol. 46, No. 1, 2002, pp. 51-56.
doi:10.1034/j.1399-6576.2002.460109.x
[10] Y. Melhorn, K. L. Davis, E. J. Burke, D. Adams, G. A.
Laine and S. J. Allen, “Impact of Cardiopulmonary By-
pass and Cardioplegic Arrest on Myocardial Lymphatic
Function,” American Journal of Physiology, Vol. 268, No.
1, 1995, pp. H178- H183.
[11] A. Hoeft, H. Korb, Y. Mehlhorn, H. Stephan and H.
Sonntag, “Priming of Cardiopulmonary Bypass with
Human Albumin or Ringer Lactate: Effect on Colloid
Osmotic Pressure and Extra-Vascular Lung Water,” Brit-
ish Journal of Anaesthesia, Vol. 66, No. 1, 1991, pp. 73-
80. doi:10.1093/bja/66.1.73
[12] Y. Mehlhorn, S. J. Allen, D. L. Adams, et al., “Nor-
mothermic Continuous Antegrade Cardioplegia Does Not
Prevent Myocardial Edema and Cardiac Dysfunction,”
Circulation, Vol. 92, 1995, pp. 1940-1946.
[13] Kirklin/Barratt-Boyes, “Cardiac Surgery,” 3rd Edition,
Chapter 5, p. 222.
[14] B. Brandstrup, et al., “Effects of Intravenous Fluid Re-
striction on Postoperative Complications: Comparison of
Two Perioperative Fluid Regimens,” Annals of Surgery,
Vol. 238, No. 5, pp. 641-648.
doi:10.1097/01.sla.0000094387.50865.23
[15] K. Lane, J. Boldt, S. Suttner, et al., “Colloids versus
Crystalloids and Tissue Oxygenation Tension in Patients
Undergoing Major Abdominal Surgery,” Anesthesia &
Analgesia, Vol. 93, 2001, pp. 405-409.
[16] S. B. Janvrin, G. Davies and R. M. Greenhalgh, “Postop-
erative Deep Vein Thrombosis Caused by Intravenous
Fluids during Surgery,” British Journal of Surgery, Vol.
67, No. 10, 1980, pp. 690-693.
doi:10.1002/bjs.1800671004
[17] D. N. Lobo, K. A. Bostoch, K. R. Neal, et al., “Effect of
Salt and Water Balance on Recovery of Gastrointestinal
Function after Elective Colonic Resection: A Random-
ized Controlled Trial,” Lancet, Vol. 359, No. 9320, 2002,
pp. 1812-1818. doi:10.1016/S0140-6736(02)08711-1
[18] S. K. Naik, A. Knight and M. Elliott, “A Prospective
Randomized Study of a Modified Technique of Ultrafil-
tration during Pediatric Open-Heart Surgery,” Circulation,
Vol. 84, Suppl 5, 1991, pp. III422-III431.
[19] S. K. Naik and M. Elliott, “Ultrafiltration and Pediatric
Cardiopulmonary Bypass,” Perfusion, Vol. 8, 1993, pp.
101-112. doi:10.1177/026765919300800114
[20] J. D. Cooper, M. Maeda and E. Lowenstein, “Lung Water
Accumulation with Acute Hemodilution in Dogs,” The
Journal of Thoracic and Cardiovascular Surgery, Vol. 69,
No. 6, 1975, pp. 957-965.
[21] A. B. Millar, L. Armstrong, J. Van Der Linden, N. Moat,
R. Ekroth, J. Westwick, M. Scallan and C. Lincoln, “Cy-
tokine Production and Hemofiltration in Children Un-
dergoing Cardio Pulmonary Bypass,” British Journal of
Surgery, Vol. 56, No. 6, 1993, pp. 1499-1502.
doi:10.1016/0003-4975(93)90740-9
[22] R. J. Howard, C. Crain, D. A. Franzini, C. I. Hood and T.
E. Hugli, “Effects of Cardiopulmonary Bypass on Pul-
monary Leukostasis and Complement Activation,” Ar-
chives of Surgery, Vol. 123, No. 12, 1988, pp. 496-501.
doi:10.1001/archsurg.1988.01400360066010
[23] J. Boldt and D. Kline, “Extravascular Lung Water and
Hemofiltration during Complicated Cardiac Surgery,”
Thoracic and Cardiovascular Surgery, Vol. 35, No. 3,
1987, pp. 161-165. doi:10.1055/s-2007-1020221