Surgery xxx (2020) 1e11

Contents lists available at ScienceDirect

Surgery
journal homepage: www.elsevier.com/locate/surg

Goal-directed hemodynamic therapy versus restrictive

normovolemic therapy in major open abdominal surgery:
A randomized controlled trial
John Diaper, RNa, Eduardo Schiffer, MDa,b, Gleicy Keli Barcelos, MDa, Ste phane Luise, RNa,
a a,b a,b,*
Raoul Schorer, MD , Christoph Ellenberger, MD , Marc Licker, MD
a
Department of Anesthesiology, Pharmacology, Intensive Care, and Emergency Medicine, University Hospital of Geneva, Switzerland
b
Faculty of Medicine, University of Geneva, Switzerland

a r t i c l e i n f o a b s t r a c t

Article history: Background: The aim of this study was to compare the occurrence of postoperative complications in
Accepted 22 September 2020 patients undergoing elective open abdominal surgery and receiving intraoperative goal-directed he-
Available online xxx modynamic therapy or restrictive normovolemic therapy.
Methods: A total of 401 patients were randomized in the goal-directed hemodynamic therapy or
restrictive normovolemic therapy groups. A cardiac output monitor was used in all goal-directed he-
modynamic therapy patients and was left at the discretion of anesthetists in charge of patients in the
restrictive normovolemic therapy group. The primary outcome was a composite morbidity endpoint (30-
day mortality and complications grade 2e4 according to Dindo-Clavien classification). Secondary out-
comes were the hospital duration of stay, the incidence of pulmonary, cardiovascular, and renal com-
plications up to 30 days after surgery, and midterm survival.
Results: Intraoperatively, the goal-directed hemodynamic therapy group received higher intravenous
fluid volumes (mean of 10.8 mL/kg/h and standard deviation of 4.0) compared with the restrictive
normovolemic therapy group (mean of 7.2 mL/kg/h and standard deviation of 2.0; P < .001). On the first
postoperative day, similar fluid volumes were infused in the 2 groups. The primary outcome occurred in
57.7% of goal-directed hemodynamic therapy and 53.0% of restrictive normovolemic therapy (relative
risk, 1.09 [95% confidence interval, 0.91e1.30]), and there was no significant difference between groups
for any secondary outcomes.
Conclusion: Among patients undergoing major open abdominal surgery, the goal-directed hemodynamic
therapy and the restrictive normovolemic therapy were associated with similar incidence of moderate-
to-severe postoperative complications and hospital resource use.
© 2020 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).

Introduction
The traditional “liberal” fluid strategy assumes that anesthesia-
induced vasoplegia causes a reduction of the “stressed” circulatory
Fluid therapy is an important cornerstone in perioperative man-
volume, whereas surgery-induced inflammation generates a
agement, and it may influence clinical outcome and the use of health
capillary leak that allows fluids to move from the intravascular to
care resources, particularly after major surgery.1,2 Fluid overload results
the interstitial compartment. Both mechanisms implicate that
in interstitial edema, increased cardiorespiratory workload, and body
additional intravenous fluids are required to restore the intravas-
weight gain, whereas insufficient fluid loading leads to poor peripheral
cular volume and maintain oxygen transport capacity.5 Neverthe-
blood flow and low tissue oxygen delivery. Both strategies have been
less, the importance of the so-called nonanatomical surgical “third
associated with impaired wound healing and poor outcomes.3,4
space” has been exaggerated, since the disturbances in vascular
permeability are transient and mostly localized around the injured
tissues.6,7
* Reprint requests: Marc Licker, Department of Anesthesiology, Pharmacology,
Nowadays, Enhanced Recovery After Surgery (ERAS) programs
Intensive Care, and Emergency Medicine, University Hospital of Geneva, Rue
Gabrielle-Perret-Gentil 4, CH-1211 Geneva, Switzerland. have been widely implemented to provide cost-efficient perioper-
E-mail address: marc-joseph.licker@hcuge.ch (M. Licker). ative care, and much emphasis has been placed on fluid

https://doi.org/10.1016/j.surg.2020.09.035
0039-6060/© 2020 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
2 J. Diaper et al. / Surgery xxx (2020) 1e11
management to facilitate physiological recovery.8,9 Vivid scientific
debates are focusing on the amount of fluids to be given, the in-
dications of vasopressors and/or inotropes, and the physiologic
targets to be achieved.10 The goal-directed hemodynamic therapy
(GDHT) requires a close monitoring of the stroke volume for
guiding fluids and cardiovascular drugs. On the other hand,
restrictive normovolemic therapy (RNT) simply requires a fixed-
rate infusion of crystalloids coupled with the administration of
vasopressors to mitigate the vasodilatory effects of anesthetic
agents and/or neuraxial block.7
From more than 45 randomized controlled studies (RCT),11e13
there is some evidence that the GDHT is associated with less
postoperative morbidity compared with the liberal strategy,
although recent well-controlled trials have failed to demonstrate
favorable clinical effects.14,15 Fewer studies using RNT have re-
ported beneficial effects,16,17 although concerns have been raised
regarding the risk of acute kidney injury.18
So far, 4 small studies comparing GDHT to RNT did not report
any difference in the occurrence of complications after intra-
abdominal surgery.19e22 Given uncertainties related to the benefi-
cial impact of GDHT and RNT along with the recent improvements
in overall perioperative care, we carried out an RCT to compare the
clinical impact of intraoperative GDHT versus RNT in patients un-
dergoing major abdominal surgery.
Material and methods
The trial was approved by the Ethics Committee of the Univer-
sity Hospital of Geneva (NAC 09e022) and registered at
ClinicalTrials.gov (NCT02625701), and it was conducted between
2010 and 2018 in a single academic center. All patients provided
written informed consent at least 1 day before surgery.
Study population
Eligible adult patients were scheduled for major abdominal,
urological, or vascular surgery via open laparotomy under general
anesthesia (
2 hours). Patients undergoing emergency procedures,
those with end-stage organ failure (with Child-Pugh class C or
model for end-stage liver disease score >22, on hemodialysis or
hemofiltration, with forced expiratory lung volume in 1 second
<30% of predicted values), those with psychiatric disorders, and
those unable to give an independent consent were all excluded.
Study design
This was a randomized, controlled, parallel-arm, and superiority
trial. Patients were randomly allocated to 1 of 2 groups (GDHT
versus RNT) in a 1:1 ratio using a computer-generated list with
blocks of 4. Allocation details were concealed in sequentially
numbered, sealed, and stapled envelopes. The randomization
sequence was developed before initiation of the trial and concealed
until after enrollment. The investigator opened envelopes on the
day of surgery. Patients, health care professionals in intensive care
(ICU), intermediate care, and postanesthesia care units (PACU), as
well as the assessors who recorded data during the postoperative
period, were all blinded to the treatment allocation. Investigators
who collected the intraoperative data had knowledge of the group
assignment.
Perioperative anesthesia conduct
All included patients were allowed solid foods up to 6 hours
before surgery and fluids up to 2 hours before surgery. Bowel
preparation was not performed in most patients, and an enhanced
recovery after surgery program was not implemented during the
study period. Prophylactic antibiotics were administered to all
patients.
A multicomponent monitoring system (IntelliVue MP70; Philips
Medical Systems, Philips Healthcare, Amsterdam, The Netherlands)
was applied and included a 5-lead electrocardiogram, an oxygen
pulse oximeter, invasive radial arterial pressure, a nasopharyngeal
temperature, as well as processed electroencephalography (BIS
Monitor; Aspect Medical Systems, Norwood, MA) and electro-
myographic response to nerve stimulations. Pulse pressure varia-
tion (PPV) was obtained online from the analysis of the pressure
waveform.
An epidural catheter was placed in most patients to provide
continuous intra and postoperative analgesia, and neuraxial anal-
gesia was performed in some patients with a single intrathecal
injection of morphine (5e10 mcg/kg), particularly in vascular sur-
gery.23 Anesthesia was induced intravenously with sufentanil
(10e15 mg) and propofol (2e3 mg$kg-1). After completion of
neuromuscular blockade, the trachea was intubated, and the lungs
were mechanically ventilated with low tidal volume (6e8 mL$kg-1
of predicted body weight), a positive end-expiratory pressure of 5
to 10 cm of water, hourly alveolar recruitment maneuvers, and
respiratory rate adjusted to keep the end-tidal carbon dioxide close
to 4.5% to 5%. The fractional inspiratory oxygen concentration was
adjusted to maintain SpO2 above 94%. The administration of
neuromuscular blocking agents was titrated to maintain 1 to 2
mechanical twitches in response to supramaximal stimulation of
the ulnar nerve at the wrist. Anesthesia was maintained with
inhaled sevoflurane to target bispectral index values between 40
and 60, whereas analgesia was ensured with neuraxial analgesia. In
the absence of central neuraxial analgesic block, lidocaine, keta-
mine, or dexmedetomidine were infused intravenously to mini-
mize the administration of intravenous opiate and facilitate
weaning from the ventilator at the end of surgery. Normothermia
was maintained with forced-air warming blankets, a heated infu-
sion system, and rewarming intra-abdominal irrigating fluids.
During surgery, arterial blood gas samples were obtained hourly. A
bladder catheter was inserted in all patients.
Postoperatively, patients were transferred to the PACU, inter-
mediate care unit, or ICU, where local practice guidelines entailed
multimodal analgesia, intravenous hypotonic solutions (5%
dextrose/0.40% sodium chloride between 0.8e1.2 mL/kg/h) as well
as early resumption of oral intake, mobilization, and removal of
indwelling catheters and drains.
Hemodynamic management
During anesthesia induction, all patients received 5 to 8 mL$kg-1
of a balanced crystalloid solution (Ringerfundin; B. Braun Medical
AG, Sempach, Switzerland) to compensate for preoperative fasting
and vasodilation associated with general anesthesia and central
neuraxial block. Throughout surgery, balanced crystalloids were
given at a rate of 2 to 4 mL$kg-1$h-1. Surgery-related blood losses
were compensated by infusing balanced crystalloids in a 2:1 ratio
or by infusing colloids (hydroxyethyl starch [HES balanced]; Fre-
senius Kabi AG, Stans, Switzerland) in a 1:1 ratio. Packed red cells
were transfused when the hemoglobin level dropped below 90 to
100 g$L-1 in the elderly and in subjects with cardiac comorbidities
or below 80 g$L-1 in those without cardiac comorbidities.
In the GDHT group, the arterial line was connected to a dedi-
cated monitor (LiDCO Ltd, Cambridge, United Kingdom) to estimate
stroke volume index and cardiac index (non-calibrated) based on
the pulse contour analysis of the pressure wave and to guide the
hemodynamic interventions. A fluid challenge of 250 mL (colloids
or crystalloids) given over 10 minutes was initiated before incision
 J. Diaper et al. / Surgery xxx (2020) 1e11
Fig 1. Intraoperative anesthetic and hemodynamic management.
and repeated until the gain in stroke volume index was less than
10% and whenever mean arterial pressure (MAP) decreased (>20%
or < 65e70 mmHg) and PPV was higher than 10% (Fig 1).
In the RNT group, besides the basal infusion of crystalloids,
supplemental fluids were given in a 1:2 ratio to compensate any
fluid losses (blood, urinary output, effusions from the gastrointes-
tinal tract) and to support circulatory homeostasis in cases of
metabolic acidosis or hypotension that showed a poor response to
vasopressors (Fig 1). A cardiac output monitor (LiDCO Ltd) was
available at the discretion of the anesthetists in charge of the
patient.
In both groups, MAP was kept above 65 to 70 mmHg with either
ephedrine and/or phenylephrine boluses. Norepinephrine was used
as a continuous infusion in patients requiring repeated doses of
ephedrine or phenylephrine and when local anesthetics were
administered through the epidural catheter. Moderate oliguria
(0.3e0.8 mL$kg-1$h-1) was not considered an indication for fluid
loading.
Data collected and study endpoints
On the day of enrolment, collected data included demographic
and clinical information as well as the results of blood laboratory
tests and other investigations. The revised cardiac risk index was
computed for each patient. Surgical and anesthetic data were
directly recorded during the procedure. Postoperatively, patients
were followed until hospital discharge to report any adverse events.
Blood samples were taken on the day before surgery and daily
within the first 3 days after surgery to measure plasma concen-
trations of N terminal pro-B type natriuretic peptide (NT-proBNP)
and Troponin I (from 2010 to 2014) or high sensitive Troponin T
(from 2015 to 2018; Elecys immunoassay, Roche Diagnostics, Basel,
Switzerland) to detect myocardial injury after noncardiac
3
surgery.24,25 After hospital discharge, follow-up was performed by
phone calls after surgery. Data were collected by 2 assessors with
more than 2 years of work experience in clinical research.
The primary endpoint was a composite morbidity endpoint (30-
day mortality and complications of grade 2e4, according to Dindo-
Clavien classification26).
Secondary endpoints included the incidence of specific com-
plications such as atelectasis, pneumonia, acute respiratory distress
failure, myocardial infarction, acute heart failure, arrhythmias,
myocardial injury after noncardiac surgery, pulmonary thrombo-
embolism, stroke, wound infection, acute kidney injury, delirium as
well as unplanned ICU admission, ICU and hospital durations of
stay, and midterm survival. We also analyzed the amount of fluids
(crystalloids, colloids, blood products) and cardiovascular drugs
administered as well as hemodynamic parameters (MAP, PPV, heart
rate, cardiac index, and blood lactate levels).
Statistical analysis
According to previous published data,11e18 we assumed that the
primary outcome would occur in about 35% of patients in the RNT
group and that the GDHT would reduce the proportion of patients
experiencing this primary outcome by one-third (from 35% to 22%).
Therefore, a sample size of 200 patients in each group was needed
for a 0.05 difference (2-sided) with a power of 80%. After recruiting
the first 100 and then 200 and 300 patients, interim analyses were
planned.
Continuous variables were reported as mean with standard
deviation (SD) or median with interquartile range and categorical
variables as frequencies (%). Continuous variables were compared
using a Student’s t test or Mann-Whitney U test and categorical
variables using c2 test or Fisher exact test. Standardized differences
were used to assess imbalances between baseline characteristics
4 J. Diaper et al. / Surgery xxx (2020) 1e11
Fig 2. Consolidated Standards of Reporting Trials flow diagram.
between the 2 groups. Repeated-measures 2-way analysis of vari-
ance with Greenhouse-Geisser correction was used to estimate
between group differences of intraoperative hemodynamic and
respiratory parameters. Kaplan-Meier curves were built for survival
probability up to 6 years after surgery.
As the trial included 3 types of surgery and evolved over a 9-year
period of time, a post hoc sensitivity analysis using generalized
linear models was carried out to assess the effect of fluid therapy
after adjusting for type of surgery and time period.
All analyses were performed using STATA 14 software (Stata
Corp, College Station, TX).
Results
A total of 558 subjects were assessed, and 401 were randomized
between 2010 and 2018, with 200 allocated to GDHT and 201 to RNT
algorithm; data from 198 and 196 patients were analyzed in the GDHT
and RNT groups, respectively (Fig 2). Since 2014, given concerns raised
by the Committee on Pharmacovigilance Risk Assessment of the Eu-
ropean Medicines Agency regarding the use of 6% HES in septic,
burned, and critically ill patients, cardiac output (CO) optimization
and compensation of fluid losses were preferably performed with
crystalloids, although the administration of HES was not formerly
forbidden.
Baseline demographic and clinical characteristics were similar
between the 2 groups (Table I). Patients in the GDHT group included
a higher proportion of visceral operations and lower proportion of
urologic procedures than the RNT group, although the Physiologic
and Operative Severity Score for the Enumeration of Mortality and
Morbidity scores were comparable. As reported in Table II, the
duration of surgery and the conduct of anesthesia were comparable
in both groups. A cardiac output monitor was used in 185 of patients
(93%) in the RNT group. A fluid challenge was given in all GDHT
patients (median of 4, range 2 to 10). The mean (SD) volumes of
fluids infused intraoperatively were higher in the GDHT than in the
RNT group (10.8 [4.0] vs 7.2 [2.0] mL$kg-1$h-1; P < .001), whereas the
intraoperative doses of ephedrine, phenylephrine, and norepineph-
rine were lower. Intraoperatively, the time course of MAP and cardiac
index was comparable in both groups whereas toward the end of
surgery, heart rate and PPV increased in the RNT group and remained
unchanged in the GDHT group (Fig 3). The control of ventilatory
parameters and oxygenation indices did not differ between the 2
groups (Fig 4).
Postoperatively, the GDHT group presented higher peak values of
NT-proBNP and larger weight gain compared with the RNT group.
As shown in Table III, a total of 210 patients reached the com-
posite postoperative mortality-morbidity end point: 113 patients
(57.7%) in the GDHT group and 105 (53.0%) in the RNT group.
Regarding the primary outcome, there was no difference among
subgroups stratified according to sex, age, revised cardiac risk in-
dex, Physiologic and Operative Severity Score for the Enumeration
of Mortality and Morbidity score, as well as study time, types of
 J. Diaper et al. / Surgery xxx (2020) 1e11 5

Table I
Demographic, clinical, biological, and surgical characteristics of patients

Variables GDHT group RNT group STD

(n ¼ 196) (n ¼ 198)

Age, y 65 (14) 64 (12) 0.1127

Female sex 71 (36.2) 71 (35.9) 0.0076
Height, cm 170 (9) 169 (9) 0.0294
Weight, kg 71.9 (15.6) 72.5 (15.0) e0.0427
BMI, kg/m2 24.9 (5.1) 25.2 (4.6) e0.0599
Comorbidities
Hypertension 97 (49.5) 85 (42.9) 0.1319
Hypercholesterolemia 60 (30.6) 43 (21.7) 0.2034
Diabetes mellitus 32 (16.3) 25 (12.6) 0.1053
Vascular disease 28 (14.3) 26 (13.1) 0.0336
Coronary heart disease 23 (11.7) 19 (9.6) 0.0693
Heart insufficiency 16 (8.2) 12 (6.1) 0.0819
Arrhythmia 15 (7.7) 16 (8.1) 0.0159
COPD 38 (19.4) 35 (17.7) 0.0440
Current smoker 74 (37.8) 64 (32.3) 0.1140
Cancer surgery 83 (42.3) 85 (42.9) 0.0118
Renal dysfunction 24 (12.2) 27 (13.6) 0.0415
Revised cardiac risk index
2 77 (39.2) 73 (36.8) 0.0645
ASA-PS classes III & IV 98 (50.0) 85 (42.9) 0.1421
POSSUM physiologic score 15.6 (7.1) 15.5 (6.1) 0.0115
POSSUM operative severity score 15.2 (3.3) 14.7 (3.8) 0.1226
Preoperative medications
Beta-adrenergic blockers 44 (22.4) 40 (20.2) 0.0549
Calcium channel blockers 33 (16.8) 16 (8.1) 0.2675
Statins 52 (26.5) 41 (20.7) 0.1374
ACEI or ARB 67 (34.2) 47 (23.7) 0.2318
Antiplatelets 53 (27.0) 44 (22.2) 0.1120
Anticoagulants 32 (16.3) 21 (10.6) 0.1682
Corticoids 1 (0.5) 5 (2.5) 0.1654
Laboratory
Hemoglobin, g/L 120.7 (18.1) 121.3 (18.3) e0.0292
Creatinine, mmol/L 80.6 (22.2) 85.9 (74.8) e0.0968
NT-pro-BNP, pg/mL 69 (35e159) 56 (29e128) 0.1931
Surgery
Visceral surgery 100 (51.0) 135 (68.2) 0.3552
Urologic surgery 43 (21.7) 20 (10.1) 0.3270
Vascular surgery 53 (27.0) 43 (21.7) 0.1242

Data are presented as mean (SD), number (percentage) or median (interquartile range).
ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin-receptor blocker; ASA-PS, American Society of Anesthesiologists physical status; BMI, body mass index; COPD,
chronic obstructive pulmonary disease; POSSUM, Physiologic and Operative Severity Score for the Enumeration of Mortality and Morbidity; STD, standardized difference.

fluids, surgery, and anesthesia (Fig 5). The incidence of any post-
operative complications, as well as the rate of unplanned admission
in the ICU and the hospital duration of stay, did not differ between
the 2 groups.
Postoperative actuarial survival did not differ according to the
intraoperative fluid strategy (Fig 6). The 30-day and 1-year overall
mortality rates were respectively 1.0% and 11.7% in the GDHT group
and 0.5% and 10.6% in the RNT group.
The post hoc sensitivity analysis (supplemental file) demon-
strated that adjusting for the study period and the type of surgery
had no significant effect on the primary study endpoint (relative
risk [RR] 1.06 [95% confidence interval (CI), 0.89e1.28] vs RR 1.09
[95% CI, 0.91e1.30] unadjusted) and any secondary study endpoints
(pulmonary complications: RR 0.97 [95% CI, 0.67e1.41] vs RR 0.94
[95% CI, 0.65e1.37] unadjusted; cardiovascular complications: RR
1.01 [95% CI, 0.74e1.38] vs RR 1.03 [95% CI, 0.76e1.40] unadjusted;
surgical complications: RR 1.06 [95% CI, 0.69e1.63] vs RR 0.98 [95%
CI, 0.63e1.51] unadjusted; acute kidney injury: RR 1.19 [95% CI,
0.81e1.75] vs RR 1.31 [95% CI, 0.90e1.92] unadjusted).
Discussion
In this single center trial evaluating a 30-day composite
morbidity index among patients undergoing major open
abdominal surgery, we compared the intraoperative application of
GDHT to RNT. Within 30 days after surgery, the composite
morbidity index and the occurrence of any specific complication
were similar in the 2 groups, as was postoperative survival.
Perioperative fluid and cardiovascular drug therapies have been
studied extensively, but the optimal strategy remains controversial
and uncertain given the heterogeneity of patient populations and
different surgical approaches, along with variable tissue injuries
(open and laparoscopic procedures), different algorithms for fluid
therapy, and poorly controlled conditions in the usual-care and
liberal groups.27,28 When compared with a liberal fluid regimen,
both GDHT and RNT have shown favorable effects on early post-
operative morbidity and hospital duration of stay.16,29 The few
small trials directly comparing GDHT with RNT did not demon-
strate any significant differences regarding postoperative clinical
outcome.19e22 Likewise, in the context of ERAS, recent trials
comparing GDHT with a more evidence-based fluid approach also
failed to replicate the positive results reported in earlier studies
where GDHT was compared to a liberal strategy that often resulted
in fluid overload.14,30 A recent systematic review of 112 RCTs on
perioperative GDHT concluded that a meta-analysis was inappro-
priate, given the overall high risk of bias in many trials, large het-
erogeneity in patients populations, different hemodynamic devices
and target physiologic endpoints, and the variable type and timing
6 J. Diaper et al. / Surgery xxx (2020) 1e11

Table II
Anesthetic and hemodynamic management

Variable GDHT group RNT group STD

(n ¼ 196) (n ¼ 198)

Anesthesia
GA alone 35 (17.9) 40 (20.2) 0.05976
GA þ epidural analgesia 121 (61.7) 128 (64.6) 0.06040
GA þ spinal analgesia 40 (20.4) 30 (15.2) 0.13781
Inhalational anesthetics 193 (98.5) 196 (99.0) 0.04649
Total iv anesthesia 3 (1.5) 2 (1.0) 0.04649
Duration of anesthesia, min 399.0 (141.7) 388.5 (123.5) 0.70791
Duration of surgery, min 313.1 (135.8) 297.5 (122.7) 0.12098
Intraoperative variables P value
Fluid administration, mL 4,868 (2033) 3,294 (1,384) < .001
Fluid administration, mL/kg/h 10.8 (4.0) 7.2 (2.0) < .001
Crystalloids, mL 4,478 (2157) 3,081 (1,427) < .001
Colloids
Patients, n 73 (37.2) 65 (32.8) < .358*
Volume, mL 1,046 (623) 650 (475) < .001
Diuresis, mL 658 (527) 501 (315) < .001
Ephedrine administration
Patients, n 145 (74) 154 (78) < .378*
Dose, mg 18 (12) 22 (13) < .002
Phenylephrine administration
Patients, n 136 (69) 148 (75) < .236*
Dose, mg 473 (373) 583 (400) < .018
Norepinephrine administration
Patients, n 94 (48.0) 109 (55.1) < .159*
Dose, mg 791 (568) 993 (790) < .041
Blood transfusion
Patients, n 31 (15.8) 29 (14.6) < .747*
Packed red blood cells, unit 2 (1e3) 2 (1e2) < .160y
Hemoglobin end of surgery, g/L 104.8 (14.3) 111.7 (17.2) < .001
Lactate end of surgery, mmol/L 1.4 (0.9) 1.5 (0.9) < .391
Postoperative
Fluids on first postoperative day, mL 2,920 (1,281) 2,780 (1,299) < .289
Peak NT-pro-BNP, pg/L 240 (102e429) 158 (85e313) < .004y
Peak creatinine, mmol/L 91.3 (39.9) 91.8 (41.1) < .892
Lowest hemoglobin, g/L 96.9 (14.2) 98.3 (15.4) < .373
Postoperative blood transfusion
Patients 30 (15.3) 30 (15.2) < .966*
Packed red blood cells, unit 1 (1e2) 2 (1e2) < .226y
Maximal postoperative weight gain, kg 4.9 (2.6e7.1) 3.4 (1.1e5.9) < .001y

Data are presented as mean (SD), number (percentage) or median (interquartile range).
Student’s t tests are used for statistical tests unless otherwise indicated.
GA, general anesthesia; iv, intravenous; STD, standardized difference.
*
c2 tests.
y
Wilcoxon rank sum tests.

of interventions (type of fluids, with or without inotropes/vaso-
pressor).31 Furthermore, merging all data from earlier trials and
more recent trials into one pooled intervention effect estimate is
highly questionable owing to the recent advancements in periop-
erative care (eg, protective lung ventilation) that have led to better
clinical outcomes in both the intervention and control groups.
In contrast with previous trials, the current study is the largest
ever conducted involving exclusively high-risk procedures (open
laparotomy, duration
120 minutes) and comparing 2 individual-
ized active interventions (GDHT versus RNT), one aiming to opti-
mize cardiac output and the other to match fluid administration
with ongoing fluid losses while controlling hemodynamics with
vasopressors.
Although we did not officially adhere to an ERAS program, some
key processes were already implemented, namely preoperative
patient optimization, intraoperative lung protective ventilation,
low-dose opiate anesthesia, and standardized fast-track post-
operative care.
In both groups, treatment goals were flexibly defined to match
with evidence-based scientific knowledge and to avoid local prac-
tice misalignment. Before the start of this study, cardiac output
monitors and arterial pressure wave analysis with PPV were
routinely used in our hospital during high-risk surgery, and anes-
thetists were all trained to identify low-flow conditions and to
correct relative hypovolemia resulting from anesthesia- and
inflammation-induced vasorelaxation. Therefore, hemodynamic
parameters reflecting blood flow and volume responsiveness (car-
diac output and PPV) were available in the majority of RNT patients
and in all GDHT patients for individualized administration of fluids
or/and vasopressors to optimize intravascular volume and blood
flow based on the Franck-Starling relationship.
Restricting fluid administration in the RNT group was associated
with increased vasopressor requirement and resulted in lower
weight gain. In the GDHT group, the higher volume of intra-
operative fluids resulted in lower PPV along with greater post-
operative weight gain and higher blood NT-proBNP values
compared with the RNT. Besides surgery-mediated inflammation,
anesthetic agents and neuraxial analgesia led to reduce vascular
muscular tone within the arterial and venous compartment and, in
turn, to reduce venous return and to lower the arterial pressure.
Consequently, fluids infused in GDHT protocols were partly redis-
tributed within the unstressed venous compartment, the remain-
ing part contributing to maximize cardiac output.32 Alternatively,
administration of vasopressors to stabilize the unstressed blood
 J. Diaper et al. / Surgery xxx (2020) 1e11 7

Fig 3. Intraoperative hemodynamic changes in patients receiving intraoperative GDHT compared to RNT. Repeated-measures 2-way analysis of variance with Greenhouse-Geisser
correction was used to estimate the trend differences between the 2 groups.

volume in the RNT group represented a sound physiologically fluid responsiveness under vasopressor treatment, owing to alter-
evidence-based approach to maintain vascular venous tone and ation in venous capacitance and compliance.33
avoid fluid overload while promoting optimal blood flow. Notably, Our standardized protocols for fluids and hemodynamic man-
dynamic indices such as PPV are known to be unreliable to predict agement differed from those used in 3 multicenter trials12,15,18 that

Fig 4. Respiratory parameters in patients receiving intraoperative GDHT compared to RNT. Repeated-measures 2-way analysis of variance with Greenhouse-Geisser correction was
used to estimate the trend differences between the 2 groups.
8 J. Diaper et al. / Surgery xxx (2020) 1e11

Table III
Primary and secondary endpoints

Variable GDHT group RNT group P value

(n ¼ 196) (n ¼ 198)

Composite index of major complications* 113 (57.7) 105 (53.0) .356

Mortality
30-days mortality 2 (1.0) 1 (0.5) .556y
Postoperative pulmonary complications 41 (20.9) 44 (22.2) .753
Acute respiratory distress syndrome 2 (1.0) 6 (3.0) .284y
Reintubation 1 (0.5) 1 (0.5) 1.000y
Pneumonia 7 (3.6) 12 (6.1) .249
Atelectasis 34 17.3) 33 (16.7) .857
Cardiovascular complications 58 (29.6) 57 (28.8) .861
Myocardial infarct 3 (1.5) 3 (1.5) 1.000y
MINS 51 (26.0) 44 (22.2) .378
Arrhythmias 5 (2.6) 4 (2.0) .750y
Heart failure 11 (5.6) 12 (6.1) .849
Thromboembolism 7 (3.6) 9 (4.5) .624
Surgical complications
Reoperation 9 (4.6) 5 (2.5) .268
Bleeding 14 (7.1) 10 (5.1) .385
Wound infection 17 (8.7) 25 (12.6) .204
Systemic inflammatory response syndrome 32 (16.3) 36 (18.2) .626
Acute kidney injury (
25% eGFR decrease) 48 (24.5) 37 (18.7) .161
Postoperative delirium 8 (4.1) 11 (5.6) .495
Admission in ICU 10 (5.1) 13 (6.6) .536
Planned 9 (4.5) 14 (7.1) .298
Unplanned 10 (5.1) 13 (6.6) .536
Hospital duration of stay, d 12 (8e20) 13 (8e19) .779z

Data are presented as number (percentage) or median (interquartile range).

c2 tests were used for statistical tests unless otherwise indicated.
ARDS, acute respiratory distress syndrome; eGFR, estimated glomerular filtration rate; MINS, myocardial injury after noncardiac surgery.
*
Dindo-Clavien classification (grade
2): 30-d mortality, acute respiratory distress syndrome, pneumonia, atelectasis, arrhythmias, MINS, myocardial infarct, acute kidney
injury, wound infection.
y
Fisher exact tests.
z
Wilcoxon rank sum tests.

Fig 5. Effect of intraoperative fluid therapy on the primary outcome in prespecified subgroups. GA, general anesthesia; POSSUM, Physiologic and Operative Severity Score for the
Enumeration of Mortality and Morbidity; RCRI, revised cardiac risk index.
 J. Diaper et al. / Surgery xxx (2020) 1e11
Fig 6. Postoperative survival in patients receiving intraoperative GDHT compared to RNT. No significant difference between groups (log-rank test for equality of survivor functions,
P ¼ .523).
included patients with similar preoperative risk profiles but with
less extensive surgical tissue trauma given the shorter duration of
abdominal procedures and the higher rates of laparoscopic
surgeries.
In the OPTIMIZE15 and FEDORA12 trials, patients in the GDHT
group were all equipped with a cardiac output monitor to guide
fluid loading with colloids to maximize stroke volume, whereas
patients in the usual-care group had no cardiac output monitor and
were treated according to undefined local practices (OPTIMIZE) or
received a continuous infusion of balanced crystalloids at 3 to 5 mL/
kg/h for laparoscopic surgery or 5 to 7 mL/kg/h for open procedures
(FEDORA trial). In the RELIEF trial,18 neither a CO monitor nor dy-
namic indices of fluid responsiveness were available, and a
“restrictive” fluid regimen designed to provide “near-zero fluid
balance” was compared to a liberal fluid regimen with crystalloids
infused at a rate of 8 to 11 mL/kg/h. The GDHT was associated with a
nonsignificant decrease in postoperative mortality-morbidity index
in the OPTIMIZE trial15 and with fewer complications only in pa-
tients undergoing laparoscopic surgeries (not laparotomies) in the
FEDORA trial.12 In contrast, the fixed-rate of intraoperative fluid
regimen in the RELIEF trial was associated with comparable
disability-free survival at 1 year, although a higher rate of acute
9
kidney injury was experienced in patients treated with the
restrictive regimen.18
Taken together, these findings indicate that a single algorithm for
fluids and drug administration does not fit in all perioperative situ-
ations. The optimal hemodynamic goals remain unclear, depending
on a patient’s baseline physiological conditions and variable re-
sponses to surgical stress that require individualized adjustments to
fit the metabolic needs. Maximizing intraoperative CO with GDHT
does not represent the sole method of improving postoperative
clinical outcome in most patients with a low-risk profile and un-
dergoing minimally invasive surgery. However, in patients with
limited cardio-respiratory reserve and in unstable hemodynamic
conditions, the GDHT likely represents a suitable approach by tar-
geting CO and oxygen delivery to match the individual patient re-
quirements while maintaining circulatory volume and arterial blood
pressure using inotropes or vasopressors. In line with this proposal,
the 2018 ERAS guidelines in elective colorectal surgery have been
modified by strongly recommending GDHT only in high-risk pa-
tients, not in “all patients” as ascertained in the 2012 guidelines.34
Our trial has certain limitations. First, the generalization of
our findings is limited owing to the single-center settings and
the inclusion of patients undergoing extensive, open intra-
10 J. Diaper et al. / Surgery xxx (2020) 1e11
abdominal procedures. Hence, our results could not be trans-
lated to lesser invasive and laparoscopic procedures that elicit an
attenuated stress response. Second, the trial intervention
dictated the administration of fluids only during the intra-
operative period when most fluids were administered, whereas
the prescription of fluids in the PACU or ICU were left at the
discretion of the attending physicians who were blinded to the
group assignment. Third, the event rate of major adverse events
was higher than expected, and it was likely related to the in-
clusion of higher-risk open laparotomies with prolonged dura-
tion of anesthesia and mechanical ventilation, while patient age
and comorbidity burden were comparable to those reported in
recent trials.12,15,17,18 Fourth, the discharge criteria were not
predefined, which can limit the interpretation of hospital dura-
tion of stay. This secondary endpoint is not only influenced by
objective criteria of functional recovery but also by structural,
social, and logistical aspects inherent to the patient’s environ-
ment and the local health care system. Finally, anesthetists were
allowed to administer either crystalloids or colloids to replace
blood losses and optimize cardiac output despite controversy
surrounding the potential adverse effects of HES in critically ill
patients.35,36 From experimental and recent clinical studies, we
can expect a reduction in the overall amount of fluid infused if
only colloids are administered for volume loading in GDHT
patients.
In conclusion, a goal-directed hemodynamic strategy and a
restrictive normovolemic regimen were associated with compara-
ble effects on the incidence of moderate or severe complications in
patients undergoing high-risk open intra-abdominal surgery.
Conflict of interest/Disclosure
ML received travel funding for lectures from Getinge and Fre-
senius Kabi. JD, SL, PG, CE, RS, GB, and ES claim no conflict of
interest.
Funding/Support
The trial was supported by an institutional grant from the
department of anesthesiology, pharmacology intensive care, and
emergency medicine at the University Hospital of Geneva (Geneva,
Switzerland). The funding bodies had no role in the design and
conduct of the study; data collection, management, analysis, or
interpretation; preparation, review, or approval of the manuscript;
or decision to submit the manuscript for publication.
Acknowledgments
The authors would like to thank all of the administrative staff
who were involved in the patients’ screening and the clinical and
surgical staff who provided care to all the patients included in this
study.
Supplementary materials
Supplementary material associated with this article can be
found, in the online version, at https://doi.org/10.1016/j.surg.2020.
09.035.
References
1. Shin CH, Long DR, McLean D, et al. Effects of intraoperative fluid management
on postoperative outcomes: A hospital registry study. Ann Surg. 2018;267:
1084e1092.
2. Eng OS, Dumitra S, O'Leary M, et al. Association of fluid administration with
morbidity in cytoreductive surgery with hyperthermic intraperitoneal
chemotherapy. JAMA Surg. 2017;152:1156e1160.
3. Chawla LS, Ince C, Chappell D, et al. and the ADQI XII Fluids Workgroup.
Vascular content, tone, integrity, and haemodynamics for guiding fluid ther-
apy: a conceptual approach. Br J Anaesth. 2014;113:748e755.
4. Kulemann B, Timme S, Seifert G, et al. Intraoperative crystalloid overload leads
to substantial inflammatory infiltration of intestinal anastomoses-a histo-
morphological analysis. Surgery. 2013;154:596e603.
5. Strunden MS, Heckel K, Goetz AE, Reuter DA. Perioperative fluid and volume
management: physiological basis, tools and strategies. Ann Intensive Care.
2011;1:2.
6. Chan ST, Kapadia CR, Johnson AW, Radcliffe AG, Dudley HA. Extracellular fluid
volume expansion and third space sequestration at the site of small bowel
anastomoses. Br J Surg. 1983;70:36e39.
7. Voldby AW, Brandstrup B. Fluid therapy in the perioperative setting-a clinical
review. J Intensive Care. 2016;4:27.
8. Makaryus R, Miller TE, Gan TJ. Current concepts of fluid management in
enhanced recovery pathways. Br J Anaesth. 2018;120:376e383.
9. Carmichael JC, Keller DS, Baldini G, et al. Clinical Practice Guidelines for
Enhanced Recovery After Colon and Rectal Surgery from the American Society
of Colon and Rectal Surgeons and Society of American Gastrointestinal and
Endoscopic Surgeons. Dis Colon Rectum. 2017;60:761e784.
10. Miller TE, Myles PS. Perioperative fluid therapy for major surgery. Anesthesi-
ology. 2019;130:825e832.
11. Sun Y, Chai F, Pan C, Romeiser JL, Gan TJ. Effect of perioperative goal-directed
hemodynamic therapy on postoperative recovery following major abdominal
surgery-a systematic review and meta-analysis of randomized controlled trials.
Crit Care. 2017;21:141.
12. Calvo-Vecino JM, Ripolles-Melchor J, Mythen MG, et al, and the FEDORA Trial
Investigators Group. Effect of goal-directed haemodynamic therapy on
postoperative complications in low-moderate risk surgical patients: a mul-
ticentre randomised controlled trial (FEDORA trial). Br J Anaesth. 2018;120:
734e744.
13. Giglio M, Dalfino L, Puntillo F, Brienza N. Hemodynamic goal-directed therapy
and postoperative kidney injury: an updated meta-analysis with trial
sequential analysis. Crit Care. 2019;23:232.
14. Schmid S, Kapfer B, Heim M, et al. Algorithm-guided goal-directed haemody-
namic therapy does not improve renal function after major abdominal surgery
compared to good standard clinical care: a prospective randomised trial. Crit
Care. 2016;20:50.
15. Pearse RM, Harrison DA, MacDonald N, et al, and the OPTIMISE Study Group.
Effect of a perioperative, cardiac output-guided hemodynamic therapy algo-
rithm on outcomes following major gastrointestinal surgery: a randomized
clinical trial and systematic review. JAMA. 2014;311:2181e2190.
16. Schol PB, Terink IM, Lance MD, Scheepers HC. Liberal or restrictive fluid
management during elective surgery: a systematic review and meta-analysis.
J Clin Anesth. 2016;35:26e39.
17. Wuethrich PY, Burkhard FC, Thalmann GN, Stueber F, Studer UE. Restrictive
deferred hydration combined with preemptive norepinephrine infusion
during radical cystectomy reduces postoperative complications and hospi-
talization time: a randomized clinical trial. Anesthesiology. 2014;120:
365e377.
18. Myles PS, Bellomo R, Corcoran T, et al. and the Australian and New Zealand
College of Anaesthetists Clinical Trials Network and the Austrialian and New
Zealand Intensive Care Society Clinical Trials Group. Restrictive versus liberal
fluid therapy for major abdominal surgery. N Engl J Med. 2018;378:
2263e2274.
19. Zhang J, Qiao H, He Z, Wang Y, Che X, Liang W. Intraoperative fluid manage-
ment in open gastrointestinal surgery: goal-directed versus restrictive. Clinics
(Sao Paulo). 2012;67:1149e1155.
20. Srinivasa S, Taylor MH, Singh PP, Yu TC, Soop M, Hill AG. Randomized clinical
trial of goal-directed fluid therapy within an enhanced recovery protocol for
elective colectomy. Br J Surg. 2013;100:66e74.
21. Phan TD, D'Souza B, Rattray MJ, Johnston MJ, Cowie BS. A randomised
controlled trial of fluid restriction compared to oesophageal Doppler-guided
goal-directed fluid therapy in elective major colorectal surgery within an
Enhanced Recovery After Surgery program. Anaesth Intensive Care. 2014;42:
752e760.
22. Brandstrup B, Svendsen PE, Rasmussen M, et al. Which goal for fluid therapy
during colorectal surgery is followed by the best outcome: near-maximal
stroke volume or zero fluid balance? Br J Anaesth. 2012;109:191e199.
23. Licker M, Ellenberger C, Cartier V, et al. Impact of anesthesia technique on the
incidence of major complications after open aortic abdominal surgery: a cohort
study. J Clin Anesth. 2013;25:296e308.
24. Biccard BM, Scott DJA, Chan MTV, et al. Myocardial injury after noncardiac
surgery (MINS) in vascular surgical patients: A prospective observational
cohort study. Ann Surg. 2018;268:357e363.
25. Writing Committee for the VSI, Devereaux PJ, Biccard BM, et al. Association of
postoperative high-sensitivity troponin levels with myocardial injury and 30-
day mortality among patients undergoing noncardiac surgery. JAMA.
2017;317:1642e1651.
26. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a
new proposal with evaluation in a cohort of 6336 patients and results of a
survey. Ann Surg. 2004;240:205e213.
 J. Diaper et al. / Surgery xxx (2020) 1e11
27. Rahbari NN, Zimmermann JB, Schmidt T, Koch M, Weigand MA, Weitz J. Meta-
analysis of standard, restrictive and supplemental fluid administration in
colorectal surgery. Br J Surg. 2009;96:331e341.
28. Ripolles-Melchor J, Chappell D, Espinosa A, et al. Perioperative fluid therapy
recommendations for major abdominal surgery. Via RICA recommendations
revisited. Part I: Physiological background. Rev Esp Anestesiol Reanim. 2017;64:
328e338.
29. Feng S, Yang S, Xiao W, Wang X, Yang K, Wang T. Effects of perioperative goal-
directed fluid therapy combined with the application of alpha-1 adrenergic
agonists on postoperative outcomes: a systematic review and meta-analysis.
BMC Anesthesiol. 2018;18:113.
30. Corcoran T, Rhodes JE, Clarke S, Myles PS, Ho KM. Perioperative fluid man-
agement strategies in major surgery: a stratified meta-analysis. Anesth Analg.
2012;114:640e651.
31. Kaufmann T, Clement RP, Scheeren TWL, Saugel B, Keus F, van der Horst ICC.
Perioperative goal-directed therapy: A systematic review without meta-anal-
ysis. Acta Anaesthesiol Scand. 2018;62:1340e1355.
11
32. Gelman S, Bigatello L. The physiologic basis for goal-directed hemodynamic
and fluid therapy: the pivotal role of the venous circulation. Can J Anaesth.
2018;65:294e308.
33. Bouchacourt JP, Riva JA, Grignola JC. The increase of vasomotor tone avoids the
ability of the dynamic preload indicators to estimate fluid responsiveness. BMC
Anesthesiol. 2013;13:41.
34. Gustafsson UO, Scott MJ, Hubner M, et al. Guidelines for Perioperative
Care in Elective Colorectal Surgery: Enhanced Recovery After
Surgery (ERAS) Society Recommendations: 2018. World J Surg. 2019;43:
659e695.
35. Zarychanski R, Abou-Setta AM, Turgeon AF, et al. Association of hydroxyethyl
starch administration with mortality and acute kidney injury in critically ill
patients requiring volume resuscitation: a systematic review and meta-anal-
ysis. JAMA. 2013;309:678e688.
36. Raiman M, Mitchell CG, Biccard BM, Rodseth RN. Comparison of hydrox-
yethyl starch colloids with crystalloids for surgical patients: A systematic
review and meta-analysis. Eur J Anaesthesiol. 2016;33:42e48.