Anesthesiology 2005; 103:25–32 © 2005 American Society of Anesthesiologists, Inc. Lippincott Williams & Wilkins, Inc.

Effect of Intraoperative Fluid Management on Outcome

after Intraabdominal Surgery
Vadim Nisanevich, M.D.,* Itamar Felsenstein, M.D.,† Gidon Almogy, M.D.,† Charles Weissman, M.D.,‡
Sharon Einav, M.D.,§ Idit Matot, M.D.储

Background: The debate over the correct perioperative fluid The widespread use of “dry” fluid regimen in pulmonary
management is unresolved. surgery with resulting decrease in pulmonary morbidity
Methods: The impact of two intraoperative fluid regimes on
postoperative outcome was prospectively evaluated in 152 pa- supports the safety of this regimen in high-risk patients
tients with an American Society of Anesthesiologists physical undergoing major surgical procedures.1.3,4 Nevertheless,
status of I–III who were undergoing elective intraabdominal no widely accepted recommendations are currently avail-
surgery. Patients were randomly assigned to receive intraoper- able for the optimal perioperative fluid regimen to be used
atively either liberal (liberal protocol group [LPG], n ⴝ 75; bolus
of 10 ml/kg followed by 12 ml · kgⴚ1 · hⴚ1) or restrictive in nonthoracic surgery. According to textbook recommen-
(restrictive protocol group [RPG], n ⴝ 77; 4 ml · kgⴚ1 · hⴚ1) dations, intraoperative fluid administration in patients un-
amounts of lactated Ringer’s solution. The primary endpoint dergoing intraabdominal procedures should be in the range
was the number of patients who died or experienced compli-
of 10 –15 ml · kg⫺1 · h⫺1.5–7 This regimen, however, is not
cations. The secondary endpoints included time to initial pas-
sage of flatus and feces, duration of hospital stay, and changes evidence based. Recent studies that investigated the effects
in body weight, hematocrit, and albumin serum concentration of different amounts of perioperative fluids on outcome
in the first 3 postoperative days. reported conflicting results depending on the patient pop-
Results: The number of patients with complications was
ulation, the type of surgery, and the regimen. Holte et al.8
lower in the RPG (P ⴝ 0.046). Patients in the LPG passed flatus
and feces significantly later (flatus, median [range]: 4 [3–7] days tried to mimic the perioperative course of minor to mod-
in the LPG vs. 3 [2–7] days in the RPG; P < 0.001; feces: 6 [4 –9] erately sized surgery in healthy volunteers and found that
days in the LPG vs. 4 [3–9] days in the RPG; P < 0.001), and their infusion of 40 ml/kg lactated Ringer’s (RL) solution over 3 h
postoperative hospital stay was significantly longer (9 [7–24]
caused significant increases in body weight and reductions
days in the LPG vs. 8 [6 –21] days in the RPG; P ⴝ 0.01). Signif-
icantly larger increases in body weight were observed in the in pulmonary function compared with infusions of 5 ml/kg.
LPG compared with the RPG (P < 0.01). In the first 3 postoper- In a subsequent study,9 the same investigators reported
ative days, hematocrit and albumin concentrations were signif- that the intraoperative administration of 40 ml/kg rather
icantly higher in the RPG compared with the LPG.
than 15 ml/kg RL solution to patients with an American
Conclusions: In patients undergoing elective intraabdominal
surgery, intraoperative use of restrictive fluid management may Society of Anesthesiologists (ASA) physical status of I or II
be advantageous because it reduces postoperative morbidity who were undergoing laparoscopic cholecystectomy led
and shortens hospital stay. to improved pulmonary function, exercise capacity, and
general well-being and shortened hospital stay. The benefit
PERIOPERATIVE fluid management continues to be a daily of administering a “high” volume of fluids (20 ml/kg) has
challenge in anesthesia practice. Abdominal surgical proce- also been demonstrated in patients undergoing general
dures in particular are associated with dehydration from anesthesia for short ambulatory procedures.10 Recently,
preoperative fasting, bowel preparation, underlying illness, the effect of different fluid regimens on outcome was
and intraoperative and postoperative fluid and electrolyte evaluated in patients undergoing more extensive opera-
loss.1 The exact quantity of this fluid loss is difficult to tions. Two studies in patients undergoing colectomy11,12 or
ascertain, and estimates for replacement with balanced salt colorectal resection12 found that restricted postoperative11
solutions range from 0 to 67 ml · kg⫺1 · h⫺1 of surgery.2 and perioperative12 fluid administration resulted in re-
duced hospital stays, faster return of gastrointestinal func-
This article is featured in “This Month in Anesthesiology.” tion,11 and reduced postoperative complications.12 The
䉫 Please see this issue of ANESTHESIOLOGY, page 5A. objective of the current study was to evaluate whether the
postulated benefits of fluid restriction can be demonstrated
in a more diverse population of surgical patients, i.e., pa-
* Resident, ‡ Professor and Chair, § Lecturer and Staff Anesthesiologist, 储 Senior tients with an ASA physical status of I–III who are under-
Lecturer and Staff Anesthesiologist, Department of Anesthesia and Critical Care
Medicine, † Attending, Department of Surgery, Hadassah Hebrew University and going a variety of extensive intraabdominal surgery. Be-
Medical Center. cause the use of liberal fluid regimens has, as reported
Received from Hadassah Hebrew University Medical Center, Jerusalem, Israel.
Submitted for publication January 18, 2005. Accepted for publication March 15,
before,1 deleterious effects on recovery of gastrointestinal
2005. Supported by a grant from the Administrator General, The State of Israel, motility, wound/anastomotic healing, coagulation, and car-
Jerusalem, Israel.
diac and pulmonary function, we tested the hypothesis that
Address reprint requests to Dr. Matot: Department of Anesthesiology and
Critical Care Medicine, Hadassah Hebrew University Medical Center, P.O. Box restrictive fluid administration for patients undergoing in-
12000, Jerusalem 91120, Israel. Address electronic mail to: idit_matot@ traabdominal surgery is associated with a lower incidence
yahoo.com. Individual article reprints may be purchased through the Journal
Web site, www.anesthesiology.org. of adverse outcomes.

Anesthesiology, V 103, No 1, Jul 2005 25

Downloaded From: http://anesthesiology.pubs.asahq.org/pdfaccess.ashx?url=/data/journals/jasa/931183/ on 11/05/2017

 26 NISANEVICH ET AL.

Materials and Methods

Patients
After institutional review board (Hadassah University
Medical Center, Jerusalem, Israel) approval and written,
informed patient consent, 156 adult patients with an
ASA physical status of I–III who were presenting for
major elective intraabdominal surgery were prospec-
tively studied. Surgical procedures included all types of
colon/rectum procedures, small bowel resections, gas-
tric resections, and pancreaticoduodenectomy/partial
pancreas resections. Patients undergoing hepatectomy
were not included in the study because relative fluid
restriction and low central venous pressures during cer-
tain stages of the operation have been shown to be
beneficial.13,14 Also excluded from the study were pa- Fig. 1. Algorithm for intraoperative fluid administration. a Indi-
tients aged younger than 18 yr; pregnant patients; and cations for blood transfusion were acute massive hemorrhage,
when the hematocrit was less than 24% in patients with no
those with coagulopathy, significant hepatic (liver en- history of coronary artery disease or no evidence of myocardial
zymes ⬎ 50% upper limit of normal value) or renal ischemia, when the hematocrit was less than 30% in patients
(creatinine ⬎ 50% upper limit of normal value) dysfunc- with a history of coronary artery disease or evidence of myo-
cardial ischemia, and when the hematocrit was less than 30%
tion, and congestive heart failure. and greater than 24% but with ongoing bleeding. b Central
venous pressure (CVP) less than 15 mmHg. c CVP greater than
15 mmHg. * Fluid bolus may also be administered in patients
Intraoperative Management with systolic blood pressure (BP) less than 90 mmHg and heart
Patients were randomized into one of two groups, a rate (HR) less than 90 beats/min when the patient’s HR cannot
be increased (␤-blocker treatments, pacemaker). ** CVP may be
liberal protocol group (LPG) or a restricted protocol introduced earlier; however, bolus administration of fluid be-
group (RPG), by using a random number generator in fore this stage should be based on hemodynamic parameters
sealed envelopes. Study investigators and research per- and urine output. IV ⴝ intravenous; RL ⴝ lactated Ringer’s
solution.
sonnel were not directly involved in the care of these
patients and hence were blinded to the treatment assign- Patients in the RPG received 4 ml · kg⫺1 · h⫺1 RL
ments. Diuretics were discontinued the day before sur- solution throughout the intraoperative period, whereas
gery. All patients received identical bowel preparation, patients in the LPG received an initial bolus of 10 ml/kg
which consisted of 3 l Precolonoscopic Solution (poly- RL solution before skin incision followed by 12 ml ·
ethylene glycol). One liter of 5% dextrose– 0.45% NaCl kg⫺1 · h⫺1. No additional boluses of fluid were adminis-
was administered during the night intravenously. All tered before skin incision, and all hemodynamic changes
patients fasted after midnight and received 10 mg diaz- during this period were treated pharmacologically. Intra-
epam orally as premedication 1 h before surgery. Anes- operative treatment of tachycardia (heart rate ⬎ 90
thesia was induced using thiopental (4 –5 mg/kg), fent- beats/min or ⬎ 20% above baseline) accompanied by
anyl (2 ␮g/kg), and vecuronium (0.1 mg/kg) and was low blood pressure (⬍ 90 mmHg or ⬎ 20% below
maintained with a balanced technique involving isoflu- baseline) was guided by a fluid algorithm (fig. 1). Fluid
rane, nitrous oxide, and oxygen. Neuromuscular block- boluses (250 ml RL solution) were also provided if urine
ade was performed with intravenous vecuronium. Addi- output decreased below 0.5 ml · kg⫺1 · h⫺1 for 2 h.
tional doses of 1.5 ␮g/kg intravenous fentanyl were Patients were reassessed after each fluid challenge to
given when the mean arterial blood pressure or heart determine whether the target hemodynamic/urine out-
rate increased 25% above baseline value. Ventilation was put goals were achieved. If hemodynamics or urine out-
adjusted to maintain an arterial carbon dioxide tension of put were not improved with the first bolus of fluid,
35– 40 mmHg, and temperature was maintained at additional boluses of RL solution were administered in
greater than 35.5°C throughout surgery. Patients re- accordance with the fluid algorithm to a maximum of
ceived epidural analgesia for postoperative pain relief. 1,500 ml. A central venous catheter was introduced
No drugs (local anesthetic or narcotic) were adminis- thereafter if no response was observed in urine output or
tered via the epidural catheter during surgery. Postop- hemodynamics. In a hemodynamically unstable patient,
eratively, all patients received continuous epidural ad- a central venous pressure of less than 15 mmHg could
ministration of bupivacaine (0.5%) and methadone trigger one of two treatment options: intravenous admin-
(0.2%) (9 ml bupivacaine and 7 ml methadone at a rate of istration of colloid (6% hydroxyethyl starch) or pharma-
60 – 80 mm/24 h) until postoperative day 3. Thereafter, cologic circulatory support. A central venous pressure of
patients’ pain treatment consisted of nonsteroidal anti- greater than 15 mmHg could trigger the intravenous
inflammatory drugs. administration of furosemide or pharmacologic circula-

Anesthesiology, V 103, No 1, Jul 2005

Downloaded From: http://anesthesiology.pubs.asahq.org/pdfaccess.ashx?url=/data/journals/jasa/931183/ on 11/05/2017

 INTRAOPERATIVE FLUID MANAGEMENT 27

tory support, depending on the patient’s hemodynamics. ery room and ending 24 h later than postoperative day 2
Further insertion of a pulmonary artery catheterization would start.
was left to the discretion of the attending anesthesiolo-
gist. The fluid regimen was continued until admission to Endpoints
the recovery room, where departmental routines en- The primary endpoint of the study combined the num-
sued. For management of surgical hemorrhage, in both ber of patients who died or experienced complications.
groups, lost blood was replaced with RL solution in a 3:1 The secondary endpoints included time to initial passage
volume replacement. Blood was transfused during acute of flatus and feces; duration of hospital stay; differences
massive hemorrhage, when the hematocrit was less than in body weight, hematocrit, creatinine, and albumin se-
24% in patients with no history of coronary artery dis- rum concentrations in the first 3 postoperative days;
ease or no evidence of myocardial ischemia, when the changes in oxygen saturation in the first 3 postoperative
hematocrit was less than 30% in patients with a history days; and number of patients receiving transfusion of
of coronary artery disease or evidence of myocardial blood and blood products.
ischemia, and when the hematocrit was less than 30%
and greater than 24% but with ongoing bleeding. To Definition of Complications
ensure uniformity, transfusion guidelines were also es- Wounds were considered infected when pus could be
tablished for administration of fresh frozen plasma (pro- expressed from the incision or aspirated from a loculated
thrombin time, activated partial thromboplastin time ⬎ mass within the wound and when bacteria were cul-
1.5 times normal), cryoprecipitate (fibrinogen concen- tured from the pus. Wound dehiscence was diagnosed
trations ⬍ 100 mg/dl), and platelets (⬍ 50 ⫻ 103/mm3) clinically and was treated by secondary suturing. Perito-
in the presence of continuous uncontrolled bleeding nitis, gastrointestinal bleeding, anastomotic leak, and in-
with no evidence of clot formation.15 Blood loss (esti- testinal obstruction would not be considered complica-
mated by assessment of the suction bottles, sponges, and tions unless they necessitated surgery. Intraabdominal
the surgical drapes and gowns), urine output, and doses abscess required diagnosis by an ultrasound or comput-
of drugs (fentanyl, furosemide) given during the surgical erized tomography scan. Diagnosis of pneumonia re-
procedure or the need to start vasoactive infusion were quired new infiltrate on chest x-ray combined with two
recorded. of the following: temperature greater than 38°C, leuko-
cytosis, and positive sputum culture. Urinary tract infec-
tion was diagnosed when symptoms consistent with the
Postoperative Management and Monitoring diagnosis, such as dysuria, frequency, fever, or an in-
In the postoperative period, the surgical staff, who creased peripheral leukocyte count, prompted urinary
were unaware of the patient’s group assignment and analysis that showed bacterial counts greater than
were not part of the investigator team, guided fluid 100,000 and positive culture. Diagnosis of sepsis re-
therapy. The routine in our General Surgery department quired bacterial infection and at least two of the follow-
is not to feed patients during the early postoperative ing clinical signs: abnormalities of body temperature
period. The volumes of crystalloids administered in the (hypothermia or hyperthermia), heart rate (tachycardia),
first 3 postoperative days were recorded. The “standard” respiratory rate (tachypnea), and leukocyte count (leu-
fluid treatment of the surgical department consists of 5% kocytopenia or leukocytosis). Diagnosis of myocardial
dextrose– 0.45% NaCl at 1–1.5 ml · kg⫺1 · h⫺1. In addi- infarction required an increase of the creatine kinase MB
tion, the number of units of blood and blood products isoenzyme or troponin T concentration above the hos-
administered until hospital discharge was also recorded. pital laboratory’s myocardial infarction threshold and
Postoperative follow-up included measurements of body either new Q waves (duration ⱖ 0.03 s) or persistent
weight (with standardized hospital uniforms), oxygen changes (4 days) in ST-T segment. Congestive heart fail-
saturation, hematocrit, potassium, sodium, albumin, and ure and pulmonary edema were defined by clinical
creatinine concentrations in the first 3 postoperative (shortness of breath, rales, jugular venous distention,
days and before discharge. All measurements were made peripheral edema, third heart sound) and radiologic (car-
in the morning (between 8:00 AM and 10:00 AM). Addi- diomegaly, interstitial edema, alveolar edema) signs that
tional blood tests, electrocardiography, and measure- required a change in medication involving at least treat-
ments of cardiac enzymes were performed when clini- ment with diuretic drugs. Arrhythmias required 12-lead
cally indicated. Time to first passage of flatus and feces electrocardiographic confirmation. Cerebrovascular ac-
was also recorded. Postoperatively, all patients were cident was diagnosed when a new focal neurologic def-
examined and interviewed daily. Complications that icit of presumed vascular etiology persisted more than
were detected by the examining physician were vali- 24 h with a neurologic imaging study that did not indi-
dated by two investigators who were not aware of the cate a different etiology. A diagnosis of acute respiratory
patient’s group assignment. Postoperative day 1 was distress syndrome was established when there was an
defined as starting from patient admittance to the recov- acute onset of respiratory distress, evidence on chest

Anesthesiology, V 103, No 1, Jul 2005

Downloaded From: http://anesthesiology.pubs.asahq.org/pdfaccess.ashx?url=/data/journals/jasa/931183/ on 11/05/2017

 28 NISANEVICH ET AL.

radiographs of airspace changes in all four quadrants, a Table 1. Patient Characteristics

ratio of partial pressure of oxygen to inspired fraction of
Restrictive
oxygen of less than 200, and pulmonary artery wedge Liberal Protocol Protocol Group
pressure less than 18 mmHg or no clinical evidence of Group (n ⫽ 75) (n ⫽ 77)

left atrial hypertension. Pulmonary embolism was diag- Sex, M/F 40/35 38/39
nosed only after evidenced by spiral computerized to- Age, yr 59.4 ⫾ 12.1 62.8 ⫾ 13.4
mography scanning. Renal dysfunction was defined by Weight, kg 68.2 ⫾ 13.5 71.5 ⫾ 14.6
Height, cm 164 ⫾ 8 166 ⫾ 7
creatinine greater than 50% upper limit of normal value. ASA physical status, I/II/III 19/37/19 15/42/20
Ischemic heart disease 14 (19%) 16 (21%)
Hypertension 24 (32%) 29 (38%)
Statistical Analysis
Cholesterol ⱖ 240 mg/dl 11 (15%) 7 (9%)
Categorical data were analyzed using the chi-square Diabetes mellitus* 10 (13%) 13 (17%)
test or Fisher exact test. Differences between the means Smoking 18 (24%) 15 (19%)
of the two groups and the median units of blood trans- Pulmonary disease 4 (5%) 5 (6%)
Cardiac medications
fused were compared using the Student t test and the ␤-Adrenergic blockers 18 (24%) 22 (29%)
Mann–Whitney test, respectively. Data within each Calcium channel 1 (1%) 3 (4%)
group were analyzed using analysis of variance for re- blockers
peated measurements. When appropriate, post hoc anal- Diuretics 7 (9%) 5 (6%)
Nitrates 11 (15%) 15 (19%)
yses were performed with the Newman-Keuls test. Exact ACE inhibitors 15 (20%) 13 (17%)
confidence intervals were computed for the overall rate
of complications. Analysis was performed using Statisti- Values are presented as mean ⫾ SD. There were no significant differences
cal Analysis System software (version 6.12; SAS Institute, between the groups.
Cary, NC). P ⬍ 0.05 was considered to represent statis- * All patients had type II diabetes mellitus.

tical significance. Results are expressed as mean ⫾ SD. ASA ⫽ American Society of Anesthesiologists; ACE ⫽ angiotensin-converting
enzyme.
Analysis was by intention to treat. A power analysis for
postoperative complication rate as an outcome, with
subgroup of patients who received daily cardiac medi-
80% power to detect a 20% reduction in this outcome
cation was evaluated separately for parameters that
and significance of 0.05 or greater, indicated that 75
could reflect hemodynamic instability. There was no
patients were required in each group.
significant difference between patients receiving cardiac
medications in the LPG versus the RPG in the need for
pharmacologic support after induction of anesthesia and
Results before skin incision. In this subgroup of patients, signif-
Demographic and Surgical Data icantly more patients in the RPG compared with the LPG
A total of 156 patients who fulfilled the entry criteria needed intraoperative bolus fluid administration: 11 ver-
were enrolled in the study, 78 in each group; among sus 0, respectively.
them, 4 (3 from the LPG) were excluded because sur-
gery was not extensive. Demographic and surgical data Endpoints
are listed in tables 1 and 2. Randomization was success- None of the patients died during the perioperative
ful in achieving comparable groups for all characteristics period. The number of patients with complications was
listed, including sex, age, weight, height, ASA physical smaller in the RPG compared with the LPG (P ⫽ 0.046;
status, and percentage of patients with concomitant dis- table 4). Significantly greater increases in body weight
eases. The same was true for the type and duration of were observed in patients in the LPG compared with
surgery and estimated blood loss. Significantly more pa- patients in the RPG in the early postoperative period
tients in the RPG received, in accordance with the fluid (1.93 ⫾ 0.52 and 1.85 ⫾ 0.62 kg on the first and third
algorithm (fig. 1), fluid boluses. Despite the administra- postoperative days, respectively, in the LPG vs. 0.51 ⫾
tion of fluid boluses in a third of the patients in the RPG, 0.67 and 0.24 ⫾ 0.61 kg in the RPG; P ⬍ 0.01). Patients
the intraoperative volumes of fluid administered were in the LPG passed flatus and feces significantly later than
significantly lower in the RPG compared with the LPG. RPG patients (flatus, median [range]: 4 [3–7] days in the
Also, in the first 3 postoperative days, the mean amounts LPG vs. 3 [2–7] days in the RPG; P ⬍ 0.001; feces: 6
of fluid infused were similar among the groups (table 3). [4 –9] days in the LPG vs. 4 [3–9] days in the RPG; P ⬍
Compared with the LPG, significantly more patients in 0.001). The duration of hospital stay was 9 days (7–24) in
the RPG experienced episodes of hypotension. Hypoten- the LPG compared with 8 days (6 –21) in the RPG (P ⫽
sion that required the administration of a fluid bolus 0.01). There were no significant differences between the
occurred in 21 patients, 1 from the LPG (1 episode of groups in the number of patients receiving blood or
hypotension) compared with 20 from the RPG (who blood product transfusion or in the median number of
experienced a total of 36 episodes of hypotension). A units of blood transfused.

Anesthesiology, V 103, No 1, Jul 2005

Downloaded From: http://anesthesiology.pubs.asahq.org/pdfaccess.ashx?url=/data/journals/jasa/931183/ on 11/05/2017

 INTRAOPERATIVE FLUID MANAGEMENT 29

Table 2. Surgical Data Table 4. Perioperative Complications

Restrictive Restrictive
Liberal Protocol Protocol Group Liberal Protocol Protocol Group
Group (n ⫽ 75) (n ⫽ 77) Complications Group (n ⫽ 75) (n ⫽ 77)

Type of surgery, % Infectious

Colon resection 30 (40) 26 (34) Wound dehiscence/infection 11 7
Resections including the 22 (29) 25 (32) Peritonitis/anastomotic leak/ 3 2
rectum intraabdominal abscess
Small bowel resection 2 (3) 0 Pneumonia 5 3
Gastric resection 8 (11) 10 (13) Urinary tract infection 2 3
Pancreaticoduodenectomy 12 (16) 14 (18) Sepsis 1 0
Pancreas resection 1 (1.3) 2 (3) Cardiovascular
Duration of surgery, min, 251 ⫾ 91 268 ⫾ 112 Myocardial infarction 1 1
mean ⫾ SD Congestive heart 2 0
Total volume of fluid 3,670 1,230 failure/pulmonary edema
administered, ml, (1,880–8,800) (490–7,810)* Arrhythmias (need to start new 3 1
median (range) treatment)
Estimated blood loss, ml, 440 (50–1,800) 400 (50–2,100) Cerebrovascular accident 0 0
median (range) Gastrointestinal
Patients receiving bolus of 1 (1.3) 26 (33)* Bleeding 0 0
fluids (as indicated by Bowel obstruction 2 0
fluid algorithm), Pulmonary
Number of patients 0 3 Acute respiratory distress 2 0
receiving 6% syndrome
hydroxyethyl starch Pulmonary emboli 0 0
Patients receiving blood 19 (25) 12 (15.5) Renal
transfusion, % Renal dysfunction 0 0
Patients receiving blood 3 (4) 2 (3) Death 0 0
products, % Total number of complications 32 17
Number of units of blood 0 (0–3) 0 (0–4) Total number of patients with 23 13*
transfused, median complications
(range)
Dose of fentanyl 345 ⫾ 135 368 ⫾ 124 * P ⬍ 0.05 vs. liberal protocol group.
administered, g, mean
⫾ SD times. Postoperative oxygen saturation decreased signif-
Patients receiving 0 2
icantly in the first 3 postoperative days in both groups;
furosemide
Patients receiving 0 0 however, values were not different between groups.
vasoactive infusion Baseline and postoperative values for sodium and potas-
sium were comparable among the groups (data not
* P ⬍ 0.001. shown). Hemodynamic data at baseline, before skin in-
cision, at skin closure, and 8 and 24 h after the operation
Preoperative hematocrit, creatinine, albumin, and ar-
were not significantly different between the groups
terial oxygen saturation were similar in both groups. In
(table 5).
the immediate postoperative period (first 3 postopera-
tive days), hematocrit and serum albumin concentration
were significantly higher in the RPG compared with the Discussion
LPG; however, at discharge there were no significant
differences between the groups (fig. 2). Mean creatinine The major finding of the current study is that relative
serum concentrations were within the normal range and intraoperative fluid restriction in patients with an ASA
were not significantly different between the groups at all physical status of I–III who are undergoing major intra-
abdominal surgery reduces the number of patients who
Table 3. Total Volume of Fluid Administered experience complications and shortens the time to re-
covery of gastrointestinal function and to hospital dis-
Restrictive
Liberal Protocol Protocol Group charge. Our study extends previous work of Lobo et al.
Group (n ⫽ 75) (n ⫽ 77) and Brandstrup et al., who demonstrated the efficacy of
Intraoperative 3,878 ⫾ 1,170 1,408 ⫾ 946* using postoperative9 and perioperative12 fluid restriction
Postoperative day 1† 2,012 ⫾ 475 2,170 ⫾ 476 in patients undergoing intraabdominal operations. In
Postoperative day 2 1,985 ⫾ 534 2,052 ⫾ 492 contrast to our study, however, these studies were per-
Postoperative day 3 1,870 ⫾ 475 1,955 ⫾ 542
formed in relatively homogenous groups of healthy pa-
Values are presented as mean ⫾ SD (ml).
tients (mainly with an ASA physical status of I or II)
* P ⬍ 0.001. † Postoperative day 1: starting from patient admittance to the undergoing either colectomy or colorectal resection,
recovery room and ending 24 h later than postoperative day 2 would start. whereas the current study included patients undergoing

Anesthesiology, V 103, No 1, Jul 2005

Downloaded From: http://anesthesiology.pubs.asahq.org/pdfaccess.ashx?url=/data/journals/jasa/931183/ on 11/05/2017

 30 NISANEVICH ET AL.

Fig. 2. Hematocrit, serum concentrations

of creatinine and albumin, and oxygen
saturation in the preoperative period
(preop ⴝ day before surgery); postoper-
ative days 1, 2, and 3; and the day of
hospital discharge. Values are presented
as mean ⴞ SD. * P < 0.01 compared with
liberal protocol group (LPG). RPG ⴝ re-
strictive protocol group.

a variety of major intraabdominal surgeries, a quarter of volumes are consistent with previous studies, recom-
whom had an ASA physical status of III. In addition, mendations, and hospital routines described in previous
unlike the study by Brandstrup et al.,12 in which oral articles. In a recent study8 designed to mimic minor to
intake was started on the first postoperative day, in the moderate operations, healthy volunteers in the “liberal
current study, patients were treated in the first few group” received 40 ml/kg lactated Ringer’s solution over
postoperative days with intravenous fluid only, as in the 3 h (i.e., approximately 13 ml · kg⫺1 · h⫺1). In laparo-
study by Lobo et al.11 Other major differences between scopic cholecystectomy,9 patients received either 40 or
the study of Brandstrup et al.12 and the current study 15 ml/kg lactated Ringer’s solution infused over 1.5 h
include the higher percentage of alcohol consumers (i.e., 26.7 vs. 10 ml · kg⫺1 · h⫺1). In yet another study8 of
(approximately two third of patients vs. none in our ambulatory surgery lasting approximately 30 min, pa-
study) and the use of a different type of fluids, mostly tients received preoperatively either 20- or 2-ml/kg infu-
normal saline in the standard group, as well as 5% glu- sions of isotonic solution. According to textbook recom-
cose, all of which could have affected outcome.9,16 In mendations, intraoperative fluid administration in
the current study, RL and 5% dextrose– 0.45% NaCl were patients undergoing intraabdominal procedures should
used. range from 10 to 15 ml · kg⫺1 · h⫺1.5–7 In this patient
population, Jenkins et al.17 suggested that the fluid reg-
Fluid Regimen imen should consist of 12–15 ml/kg for the first hour and
The volumes of intraoperative and postoperative fluid
6 –10 ml/kg for the next 2 h. Similarly, Campbell et al.18
administered in the LPG seem high. However, these
observed that cardiovascular stability during major oper-
ations is much better preserved when intraoperative
Table 5. Hemodynamic Measurements and Oxygen Saturation
in the Two Groups crystalloids are given at the rate of 10 –15 ml · kg⫺1 · h⫺1.
The amount of intraoperative fluid used in the LPG in the
Mean current study is therefore consistent with these recom-
Arterial
Blood Oxygen mendations. Two recent prospective studies used different
Pressure, Heart Rate, Saturation, fluid regimens in patients undergoing colectomy. In one
mmHg beats/min %
study in which surgery lasted less than 2 h and involved
LPG baseline* 76 ⫾ 14 75 ⫾ 10 98.1 ⫾ 1.5 minimal bleeding,11 the authors compared postoperative
RPG baseline* 73 ⫾ 11 71 ⫾ 14 98.0 ⫾ 1.4
LPG before skin incision 75 ⫾ 12 68 ⫾ 12 98.2 ⫾ 0.8
administration of 3 l fluid/day (liberal group) with 1.5–2
RPG before skin incision 68 ⫾ 10 73 ⫾ 11 98.3 ⫾ 0.9 l/day. Intraoperatively, all patients received 2.5–2.8 l fluid.
LPG before skin closure 81 ⫾ 14 79 ⫾ 13 97.7 ⫾ 1.0 In the second study,12 the median amounts of fluid admin-
RPG before skin closure 80 ⫾ 12 83 ⫾ 14 97.5 ⫾ 1.6
istered on the day of surgery were 2,740 and 5,388 ml in
LPG 8 h after surgery 86 ⫾ 17 82 ⫾ 12 97 ⫾1.5
RPG 8 h after surgery 82 ⫾ 15 87 ⫾ 11 97.2 ⫾ 1.6 the restrictive and standard groups, respectively. Postoper-
LPG 24 h after surgery 73 ⫾ 16 78 ⫾ 13 97.3 ⫾ 1.2 atively, the total amount of fluid administered (intravenous
RPG 24 h after surgery 75 ⫾ 12 83 ⫾ 12 97.0 ⫾ 1.6 plus oral) was in the range of 2.5 l (restrictive group) versus
3.5 l. In the current study, the volume of fluid administered
Values are presented as mean ⫾ SD. No significant differences were noted
between the groups at the specified time points. in the postoperative period in both groups (1.5–2.5 l) was
* Baseline (before induction of anesthesia). similar to the restrictive groups in both previous stud-
LPG ⫽ liberal protocol group; RPG ⫽ restrictive protocol group. ies.11,12

Anesthesiology, V 103, No 1, Jul 2005

Downloaded From: http://anesthesiology.pubs.asahq.org/pdfaccess.ashx?url=/data/journals/jasa/931183/ on 11/05/2017

 INTRAOPERATIVE FLUID MANAGEMENT 31

Effect on Outcome study, but possible mechanisms have been recently

Studies in minor/ambulatory surgery suggest that high- reviewed.1
dose fluid regimens may improve early recovery mea- There are several limitations to this study. The possi-
sures such as dizziness, drowsiness,9,19 nausea, and bility that the observed differences were due to factors
thirst; improve pulmonary function and exercise capac- other than the amount of resuscitation fluid cannot be
ity; and shorten hospital stay.9 The results of these stud- excluded. However, this is a prospective study with an
ies, however, cannot be extrapolated to major intraab- intent-to-treat design and with well-defined endpoints. In
dominal surgical procedures in which substantially addition, the study was not conducted in a totally
larger third space loss, larger stress response, and altered blinded fashion. Although the anesthesiologist treating
capillary permeability occur. The decreases in hemoglo- the patient in the intraoperative period was not blinded
bin and albumin concentrations that were observed only to the patient’s group assignment, the indications for
in patients from the LPG in the current study are in additional fluid administration were standardized. Dur-
accord with previous studies11,12 and probably reflect ing the postoperative period, adverse outcomes were
the dilutional effect of the larger fluid volumes. Also, the detected by the examining physician, who was not
increase in body weight is most probably the result of aware of the patient’s assignment. Late complications
fluid overload as changes in weight have been shown to could have been missed because we followed patients
reflect fluid balance.20,21 The current study also found only until hospital discharge. In addition, the design of
that the median time to flatus and feces passage was this study does not enable us to compare its results with
significantly longer in the LPG compared with the RPG. studies that used algorithms focused on achieving end-
Positive postoperative fluid balance can result in gut points directed by invasive monitoring. In those studies,
edema, which may contribute to intestinal dysfunc- however, patients did not receive overly large volumes
tion.1,22 In the 1930s, Mecray et al.23 found that modest of fluid but were adequately resuscitated to optimize
positive salt and water balance caused weight gain after oxygen delivery. Finally, the postoperative management
elective colectomy and was associated with delayed re- of patients undergoing intraabdominal surgery is fre-
covery of gastrointestinal function, increased complica- quently institution and department specific. Some surgi-
tion rates, and extended hospital stays. Similar findings cal centers start oral intake of fluid early in the postop-
were demonstrated later in human studies.11,24 –26 Hy- erative period. Nevertheless, this may add up to a
poproteinemia has been associated with extended gas- significant amount as shown in a recent study,12 and the
tric emptying, delayed small bowel transit, and postop- results may therefore be applicable to these patients as
erative ileus.11,23–26 Whether the effect is due to well. Further studies are needed to address this point.
hypoalbuminemia per se or the result of positive fluid Significant healthcare resources are used to provide
balance is unknown because it is difficult to separate care to patients with prolonged postoperative hospital-
these two conditions.20 Others have challenged these ization. Clinicians, hospitals, and healthcare payers are
findings and reported that increased perioperative fluid increasingly focusing on reducing “unnecessary” days of
administration was associated with improved indices of hospitalization after surgery. Gastrointestinal dysfunc-
gut perfusion and reduced intestinal dysfunction.27,28 tion has a substantial effect on resource utilization.29 In
These studies, however, were unblinded and involved a two large studies that included patients undergoing ma-
different patient population (cardiac28 and mostly uro- jor noncardiac surgeries, Bennett-Guerrero et al.30,31
logic and gynecologic patients27). In the study by Gan et demonstrated that gastrointestinal dysfunction was the
al.,27 both groups received “liberal” fluid administration most common morbid event that was associated with
(approximately 4.5 l clear crystalloids for surgeries with prolongation of hospital stay. The current study found
a mean duration of 4 h), with no significant difference in that intraoperative use of “restrictive” fluid regimen
the amount of crystalloids administered between the shortens return of gastrointestinal function and reduces
groups. Moreover, in both studies, the differences in the number of patients experiencing postoperative com-
outcome could have been attributed to the type of fluid, plications with subsequent shortening of hospital stay.
because patients in the protocol group received more The results from this trial demonstrate that some mor-
colloids. The finding that more patients in the high- bidity observed in surgical patients may be preventable
volume group had complications is consistent with pre- by using this fluid strategy. As this study evaluated only
vious reports in similar patient populations.12 Larger two volumes of fluid and focused only on intraoperative
fluid volumes may exert harmful effects on cardiac and management; it is possible to speculate that a different
pulmonary function, tissue oxygenation, coagulation, dose regimen might have further improved outcome.
wound healing, and gastrointestinal function.1 The exact Therefore, additional studies are needed to establish the
mechanism by which liberal intraoperative fluid admin- optimal volume of fluid to be infused during and after
istration increases morbidity was not evaluated in this intraabdominal surgery and other major procedures.

Anesthesiology, V 103, No 1, Jul 2005

Downloaded From: http://anesthesiology.pubs.asahq.org/pdfaccess.ashx?url=/data/journals/jasa/931183/ on 11/05/2017

 32 NISANEVICH ET AL.

References 15. Practice guidelines for blood component therapy: A report by the Amer-
ican Society of Anesthesiologists Task Force on blood component therapy.
1. Holte K, Sharrock NE, Kehlet H: Pathophysiology and clinical implications ANESTHESIOLOGY 1996; 84:732–47
of perioperative fluid excess. Br J Anaesth 2002; 89:622–632 16. Finney SJ, Zekveld C, Eia A, Evans TW: Glucose control and mortality in
2. Shires T, Williams J, Brown F: Acute change in extracellular fluids associated critically ill patients. JAMA 2003; 209:2041–7
with major surgical procedures. Ann Surg 1961; 154:803–10 17. Jenkins MT, Giesicke AH, Johnson ER: The postoperative patient and his
3. Parquin F, Marchal M, Mehiri S, Herve P, Lescot B: Post-pneumonectomy fluid and electrolyte requirements. Br J Anaesth 1975; 47:143–50
pulmonary edema: Analysis and risk factors. Eur J Cardiothorac Surg 1996; 18. Campbell IT, Bater JN, Tweedie IE, Taylor GT, Keens SJ: IV fluids during
10:929–32 surgery. Br J Anaesth 1990; 65:726–9
4. Slinger PD: Perioperative fluid management for thoracic surgery: The puzzle 19. Holte K, Kehlet H: Compensatory fluid administration for preoperative
of postpneumonectomy pulmonary edema. J Cardiothorac Vasc Anesth 1995; dehydration: Does it improve outcome? Acta Anaesth Scand 2002; 46:1089–93
9:442–51 20. Lobo DN, Stanga Z, Simpson JA, Anderson JA, Rowlands BJ, Allison SP:
5. Hwang G, Marota JA: Anesthesia for abdominal surgery, Clinical Anesthesia Dilution and redistribution effects of rapid 2-litre infusions of 0.9% (w/v) saline
Procedures of the Massachusetts General Hospital. Edited by Hurford WE, Bailin and 5% (w/v) dextrose on haematological parameters and serum biochemistry in
MT, Davison JK, Haspel KL, Rosow C. Philadelphia, Lippincott–Raven, 1997, pp normal subjects: A double-blind crossover study. Clin Sci (Lond) 2001; 101:173–9
330–46 21. Lobo DN, Bjarnason K, Field J, Rownalds BJ, Allison SP: Changes in weight,
6. Sendak M: Monitoring and management of perioperative fluid and electro- fluid balance, and serum albumin in patients referred for nutritional support. Clin
lyte therapy, Principles and Practice of Anesthesiology, 1st edition. Edited by Nutr 1999; 18:197–201
Rogers MC, Longnecker DE, Tinker JH. New York, Mosby-Year Book, 1993, pp 22. Moretti EW, Robertson KM, El Moalem H, Gan TJ: Intraoperative colloid
863–966 administration reduces postoperative nausea and vomiting and improves post-
7. Tonnesen AS: Crystalloids and colloids, Anesthesia, 3rd edition. Edited by operative outcomes compared with crystalloid administration. Anesth Analg
Miller RD. New York, Churchill Livingstone, 1990, pp 1439–65 2003; 96:611–7
8. Holte K, Jensen P, Kehlet H: Physiologic effects of intravenous fluid admin- 23. Mecray PM, Barden RP, Ravdin IS: Nutritional edema: Its effect on the
istration in healthy volunteers. Anesth Analg 2003; 96:1504–9 gastric emptying time before and after gastric operations. Surgery 1937; 1:53–6
9. Holte K, Klarskov B, Christensen DS, Lund C, Nielsen KG, Bie P, Kehlet H: 24. Barden RP, Thompson WD, Ravdin IS, Frank IL: The influence of serum
Liberal versus restrictive fluid administration to improve recovery after laparo- protein on the motility of the small intestine. Surg Gynecol Obstet 1938; 66:
scopic cholecystectomy: A randomized, double-blind study. Ann Surg 2004; 819–21
240:892–9 25. Andrassy RJ, Durr ED: Albumin: Use in nutrition and support. Nutr Clin
10. Yogendran S, Asokumar B, Cheng DC, Chung F: A prospective randomized
Pract 1988; 3:226–9
double-blinded study of the effect of intravenous fluid therapy on adverse
26. Woods MS, Kelley H: Oncotic pressure, albumin and ileus: The effect of
outcomes on outpatient surgery. Anesth Analg 1995; 80:682–6
albumin replacement on postoperative ileus. Ann Surg 1993; 59:758–63
11. Lobo DN, Bostock KA, Neal KR, Perkins AC, Rowlands BJ, Allison SP:
27. Gan TJ, Soppitt A, Maroof M, el-Moalem, H, Robertson KM, Moretti E,
Effect of salt and water balance on recovery of gastrointestinal function after
elective colonic resection: A randomised controlled trial. Lancet 2002; 359: Dwane P, Glass PS: Goal-directed intraoperative fluid administration reduces
1812–8 length of hospital stay after major surgery. ANESTHESIOLOGY 2002; 97:820–6
12. Brandstrup B, Tonnesen H, Beier-Holgersen R, Hjortso E, Ording H, Lin- 28. Mythen MG, Webb AR: Perioperative plasma volume expansion reduces
dorff-Larsen K, Rasmussen MS, Lanng C, Wallin L, Iversen LH, Gramkow CS, the incidence of gut mucosal hypoperfusion during cardiac surgery. Arch Surg
Okholm M, Blemmer T, Svendsen PE, Rottensten HH, Thage B, Riis J, Jeppesen 1995; 130:423–9
IS, Teilum D, Christensen AM, Graungaard B, Pott F, Danish Study Group on 29. Mythen MG: Postoperative gastrointestinal tract dysfunction. Anesth Analg
Perioperative Fluid Therapy:Effects of intravenous fluid restriction on postoper- 2005; 100:196–204
ative complications: Comparison of two perioperative fluid regimens: A random- 30. Bennett-Guerrero E, Welsby I, Dunn TJ, Young LR, Wahl TA, Diers TL,
ized assessor-blinded multicenter trial. Ann Surg 2003; 238:641–8 Phillips-Bute BG, Newman MF, Mythen MG: The use of a postoperative morbidity
13. Mendelez JA, Arslan V, Fischer ME, Wuest D, Jarnagin WR, Fong Y, survey to evaluate patients with prolonged hospitalization after routine, moder-
Blumgart LH: Perioperative outcomes of major hepatic resections under low ate-risk, elective surgery. Anesth Analg 1999; 89:514–9
central venous pressure anesthesia: Blood loss, blood transfusion, and the risk of 31. Bennett-Guerrero E, Panah MH, Barclay GR, Bodian CA, Winfree WJ,
postoperative renal dysfunction. Am Coll Surg 1998; 187:620–25 Andres LA, Reich DL, Mythen MG: Decreased endotoxin immunity is associated
14. Jones RM, Moulton CE, Hardy KJ: Central venous pressure and its effect on with greater mortality and/or prolonged hospitalization after surgery. ANESTHESI-
blood loss during liver resection. Br J Surg 1998; 85:1058–60 OLOGY 2001; 94:992–8

Anesthesiology, V 103, No 1, Jul 2005

Downloaded From: http://anesthesiology.pubs.asahq.org/pdfaccess.ashx?url=/data/journals/jasa/931183/ on 11/05/2017