FEATURE
Allowing Normal Food at Will After Major Upper
Gastrointestinal Surgery Does Not Increase Morbidity
A Randomized Multicenter Trial
Kristoffer Lassen, MD, PhD,*† Jørn Kjæve, MD, PhD,*† Torunn Fetveit, MD,‡ Gerd Tranø, MD,§
Helgi Kjartan Sigurdsson, MD,¶ Arild Horn, MD, PhD,储 and Arthur Revhaug, MD, PhD*†
Objective: The aim of this trial was to investigate whether a routine
of allowing normal food at will increases morbidity after major
upper gastrointestinal (GI) surgery.
Summary Background Data: Nil-by-mouth with enteral tube feed-
ing is widely practiced for several days after major upper GI surgery.
After other abdominal operations, normal food at will has been
shown to be safe and to improve gut function.
Methods: Patients were randomly assigned to a routine of nil-by-
mouth and enteral tube feeding by needle-catheter jejunostomy
(ETF group) or normal food at will from the first day after major
upper GI surgery. Primary end point was rate of major complications
and death. Secondary outcomes were minor complications and
adverse events, bowel function, and length of stay. All patients were
invited to a follow-up at 8 weeks after discharge from the hospital.
Results: Four hundred fifty-three patients who underwent major
open upper GI surgery in 5 centers were enrolled between 2001 and
2006. Four hundred forty-seven patients were correctly randomized.
Of 227 patients 76 (33.5%) had major complications in the ETF
group compared with 62 (28.2%) of 220 patients allowed normal
food at will (P ⫽ 0.26, 95% CI for the difference in rate from ⫺3.3
to 13.9). In the ETF group, 36 (15.9%) patients were reoperated
compared with 29 (13.2%) in the group allowed normal food at will
(P ⫽ 0.50) and 30-day mortality was 10 (4.4%) of 227 and 11
From the *Department of Gastrointestinal Surgery, University Hospital
Northern Norway, Tromsø, Norway; †Institute of Clinical Medicine,
University of Tromsø, Norway; ‡Department of Surgery, Sørlandet
Hospital, Arendal, Norway; §Department of Gastrointestinal Surgery, St.
Olavs Hospital, University Hospital of Trondheim, Norway; ¶Depart-
ment of Surgery, Stavanger University Hospital, Stavanger, Norway; and
㛳Department of Gastrointestinal Surgery, Haukeland University Hospital,
Bergen, Norway.
Supported by the Norwegian Research Council Grant 147339/V50 and
Fresenius Kabi AS, Norway.
Kristoffer Lassen and Arthur Revhaug are members of the Enhanced Re-
covery After Surgery (ERAS) group. From 2006, Fresenius Kabi is a
main sponsor of this group.
As the corresponding author, Kristoffer Lassen declares that he has had full
access to all the data in the study and accepts final responsibility for the
decision to submit for publication.
Corresponding author: Kristoffer Lassen, MD, PhD, Department of Gastro-
intestinal Surgery, University Hospital Northern Norway, Tromso, 9038,
Norway. E-mail: lassen@unn.no.
Copyright © 2008 by Lippincott Williams & Wilkins
ISSN: 0003-4932/08/24705-0721
DOI: 10.1097/SLA.0b013e31815cca68
Annals of Surgery • Volume 247, Number 5, May 2008
Kaliane Cunha - kalianebcunha@gmail.com - IP: 177.37.205.206
(5.0%) of 220 patients, respectively (P ⫽ 0.83). Time to resumed
bowel function was significantly in favor of allowing normal food at
will (P ⫽ 0.01), as were the total number of major complications,
length of stay, and rate of postdischarge complications.
Conclusions: Allowing patients to eat normal food at will from the
first day after major upper GI surgery does not increase morbidity
compared with traditional care with nil-by-mouth and enteral feeding.
(Ann Surg 2008;247: 721–729)
T he literature concludes that patients should be allowed
food without delay (at will) after colorectal surgery and
that the customary withholding of oral intake (nil-by-mouth)
for the first postoperative days is unnecessary.1,2 Robust data
also suggest that we should avoid the nil-by-mouth regimen
after major gynecologic,3,4 urologic,5 and vascular surgery.2,6
In upper gastrointestinal (GI) surgery, however, a strict nil-
by-mouth regimen for several days is widely practiced.7–11
This reluctance to allow food at will is based on the concern
for gastric distension and for anastomotic integrity. This
concern is not supported by randomized trials.12,13
Upper GI surgeons hesitate to allow anything by mouth
postoperatively, and many advocate enteral catheter feeding
postoperatively to a jejunal segment distal to a new anasto-
mosis. The safety and feasibility of postoperative enteral
nutrition by needle-catheter jejunostomies (NCJ) has been
extensively documented,14 –18 but serious complications oc-
cur.19 Although enteral catheter feeding has been shown to be
beneficial in upper GI surgery patients,20 the preference for
this modality also rests on the traditional but undocumented
reluctance to allow food at will. These assumed hazards of
allowing normal food in the immediate postoperative period
have not been scientifically tested and should be viewed
against both the benefits and side effects of any artificial
feeding modality.21
The objective of this trial was to investigate whether the
alleged increase in morbidity associated with allowing food at
will after major upper GI surgery could be sustained in a
randomized trial. We compared a routine of allowing normal
food at will from the first day after surgery with a routine of
nil-by-mouth and enteral nutrition for the first 5 postoperative
days.
721
 Lassen et al Annals of Surgery • Volume 247, Number 5, May 2008

METHODS mouth postoperatively.16 Patients with severe extra-abdomi-

Patients and Hospitals
The patients were enrolled in 4 tertiary referral centers
(University Hospitals of Stavanger, Bergen, Trondheim, and
Tromsø) and in 1 secondary referral hospital (Arendal) in
Norway. The patients were informed verbally and in writing,
and written consent was obtained. Adults undergoing both
scheduled and emergency major upper GI surgery and able to
give written informed consent were eligible. The operations
were hepatic, pancreatic, esophageal, and gastric resections,
bilioenteric and gastroenteric bypass procedures, and miscel-
laneous procedures (eg, ileus with gross small bowel disten-
sion and severe contamination of the upper abdominal cavity
after gut perforation) where tradition would indicate nil-by-
nal disease or trauma, life expectancy less than 3 months,
short bowel or other undisputed indication for parenteral
nutrition were not eligible. Preoperative assessment included
weight and albumin concentration for stratification by Nutri-
tional Risk Index (NRI: 1.519 ⫻ serum albumin level (g/L) ⫹
0.417 ⫻ (current weight/usual weight) ⫻ 100).22
Perioperative Care and Interventions
Perioperative routines were largely uniform in all par-
ticipating centers. Bowel cleansing was not used, food was
allowed until midnight before surgery in scheduled cases, and
antithrombotic prophylaxis given as low molecular frag-
mented heparin. Single-dose perioperative antibiotic cover-
age was used, and thoracic epidurals were the mainstay of

TABLE 1. Definitions of Complications

Complication Criteria
Major: Infectious/surgical
SIRS Two or more of the following: temperature ⬎38°C, or ⬍36°C, heart rate ⬎90 beats/min, respiratory rate
⬎20/min or PaCO2 ⬍4.3 kPa, white blood count ⬎12,000 cells/mL, ⬍4000 cells/mL or ⬎10% immature
forms.
Bacteremia At least one positive blood culture of pathogenic organisms
Sepsis SIRS ⫹ bacteremia
Pneumonia X-ray confirmed and necessitating antibiotic treatment
Anastomotic leak Necessitating reoperation or demonstrated on autopsy
Bowel obstruction, necrosis, or Necessitating reoperation or demonstrated on autopsy
perforation
Intra-abdominal abscess Necessitating reoperation or percutaneous drainage or demonstrated on autopsy
Intra-abdominal hemorrhage Necessitating reoperation, transfusion of 6 or more units of packed red blood cells within first 48 h p.o., or
demonstrated on autopsy
Wound rupture Necessitating resuturing in general anesthesia or demonstrated on autopsy
Pancreatitis Serum enzyme level twice upper normal value, absence of known preexisting pancreatitis, ERCP, or
mechanical trauma to pancreas during operation providing satisfactory explanation for elevated enzymes
Cholecystitis Confirmed by histological examination of specimen or by cultured content from percutaneous drainage
Major: Other
Myocardial infarction Diagnostic enzyme pattern and either typical pain or ECG changes, or demonstrated on autopsy
Myocardial arrhythmia ECG-confirmed, with hypotension or symptomatic angina, necessitating stabilizing drugs or electroconversion
Cardiac arrest Confirmed by ECG or cardiac rhythm monitoring, and necessitating resuscitation
Acute cardiac failure Confirmed by echocardiography or necessitating pressure agents
Cerebrovascular hemorrhage or New and persistent (⬎48 h) central neurological deficit, and confirmed by CT scan or demonstrated on
cerebrovascular infarction autopsy
Pulmonary embolism Confirmed by unequivocal nuclear isotope scan, echocardiography, pulmonary angiography, spiral-technique
CT scan, or demonstrated on autopsy
Pulmonary insufficiency Necessitating postoperative ventilation support more than 24 h
Minor
Atelectasis X-ray confirmed and necessitating lung physiotherapy
Wound infection Necessitating removal of stitches and topical treatment, antibiotics or surgical debridement
Incisional hernia As judged clinically
Adverse events
Need for parenteral nutrition, or All intravenous or catheter-delivered substances other than clear crystalloids or colloids will be regarded as
therapeutic enteral feeding (reverting nutrients
to enteral feeding after withdrawal)
Recurrent need of ICU treatment Patients taken from the operating theater to either the ICU department or a postoperative surveillance ward,
depending on clinical status and type of operation. After discharged to the bed ward, any readmission to
these wards will be regarded as failure of this parameter
Use of therapeutic decompression tube Any nasogastric tube inserted after the patient has been primarily extubated and transferred to bed ward
ECG indicates electrocardiogram; ERCP, endoscopic retrograde cholangiography; PaCO2, arterial pressure of carbon dioxide; kPa, kilo pascal; SIRS, systemic inflammatory
response syndrome.

722 © 2008 Lippincott Williams & Wilkins

Kaliane Cunha - kalianebcunha@gmail.com - IP: 177.37.205.206

 Annals of Surgery • Volume 247, Number 5, May 2008 Food at Will After Surgery Does Not Increase Morbidity

perioperative and postoperative analgesia in all hospitals
throughout the trial. Nasogastric tubes were removed at the
end of surgery or at the latest on the first postoperative day
(POD 1) from when mobilization was encouraged.
The patients randomly assigned to be allowed normal
food at will were allowed to take ordinary hospital food as
they wanted from the first day after surgery. They were
encouraged to begin the oral intake carefully and adjust
according to tolerance. The nutritional intake was not mea-
sured and there were no target values.
The patients randomly assigned to enteral tube feeding by
needle-catheter jejunostomy (ETF group) had a feeding catheter
(FloCare Jejunocath, 5 Ch., external diameter 1.7 mm; Nutricia
Norway AS, Oslo, Norway) passed through the abdominal wall
via a split cannula at a 45° angle to the skin and along the axis
of a targeted jejunal loop distal to any anastomosis. Another split
cannula allowed a 3 cm intramural passage to be created before
entering the bowel lumen. The jejunal segment was fixed to the
abdominal wall at the insertion site with absorbable sutures over
a length of 3 to 6 cm to prevent kinking.17 The patients received
isotonic saline by the catheter at 20 mL/h until the morning of
POD 1. Nutrition was then commenced (standard Fresubin, 1
kcal/mL; Fresenius Kabi AS, Halden, Norway) at 20 mL/h. The
rate was increased by 20 mL/h each day if tolerated, up to 80
mL/h. In the event of abdominal pain, severe distension, or
vomiting, infusions were reduced or temporarily halted. A max-
imum of 450 mL of water per day was allowed by mouth. At
POD 6, the patients were allowed food at will and enteral
infusion halted.
SPSS 14.0 for Windows. Standard approximation to binomi-
nal distribution by ␹2 analysis or Fisher exact test was used
when appropriate. Continuous variables were compared by
Student t test. Adjustments were by linear or binary logistic,
or Cox regression as appropriate. Differences were consid-
ered significant at P ⬍ 0.05.
Assignment
Computerized randomization lists were created and the
results were put in sealed opaque envelopes by individuals not
involved in the trial. The envelopes were stored in a central
randomization office in Tromsø, physically removed from wards
and theaters. Randomization was stratified for hospital and for
high (esophageal and pancreatic resections) or medium risk
operations and blocked with a fixed but unknown block size.
After informed consent, the patients were randomized intraop-
eratively if surgery fulfilling the criteria had been performed. A
phone call was made to the central randomization office imme-
diately before the closing of the abdominal incision, after the
intra-abdominal surgical procedure was completed.
Ethics and Consent
The study was approved by the Regional Division of
the National Committee for Research Ethics (P-REK V 52/
2000) and registered at ClinicalTrials.gov on August 23, 2005
(NCT00134407). It was carried out in accordance with the

End Points and Registration

The primary end point was rates of patients with major
complications or death during hospital stay and within 8 weeks.
Complications were predefined in the protocol (Table 1).
Secondary end points were rates of minor complications
and adverse events, gut-recovery indicators (time to bowel
movement and selective use of nasogastric tube), indica-
tors for underfeeding (clinically indicated selective use of
parenteral nutrition and weight loss) and length of stay after
surgery. Blinding of outcome assessors was not attempted.
All patients were invited to a physical follow-up about
8 weeks after discharge to assess trial outcome. Patients
failing to show up had their hospital files examined for
evidence of unscheduled readmissions to other wards and
hence unregistered postdischarge complications. Finally, the
patients lost to follow-up were crosschecked in the Norwe-
gian Central Population Register to exclude unregistered
death within 8 weeks after discharge.

Statistics
We estimated the minimum rate of patients with major
complications in this heterogeneous population at 25%. A
reduction of this by half (to 12.5%) was considered clinically
important. Detecting a difference of this magnitude or greater
at a level of statistical significance of 0.05 and a power of
0.90 with a two-tailed test of proportions required a total of
215 patients in each group. Correcting for the effect of one
interim analysis, a total number of 444 patients was needed.
The data were analyzed on an intention-to-treat basis, with FIGURE 1. Trial flowchart.

© 2008 Lippincott Williams & Wilkins 723

Kaliane Cunha - kalianebcunha@gmail.com - IP: 177.37.205.206

 Lassen et al Annals of Surgery • Volume 247, Number 5, May 2008

Helsinki Declaration. An independent interim analysis in RESULTS

2004 confirmed trial continuation to be ethically sound.
Role of Funding Source
The sponsors did not have any role in the design, data
collection, analysis or interpretation, writing of the paper, or
in the decision to submit for publication.
Trial enrollment was from February 2001 to April 2006
because accrual was slower than expected. Six ineligible
patients were mistakenly randomized and hence excluded. In
3 patients randomized to ETF, catheter insertion was not
attempted but they are included in the analysis (trial flow-

TABLE 2. Baseline Comparison of Treatment Groups

Variable Measure Enteral Tube Feeding Allowed Food at Will
Allocated to intervention n 227 220
Treated at university/referral hospital n (%) 202 (89.0) 195 (88.6)
Male proportion % 64.3 53.2
Age (years) Mean (SD) 65 (13.3) 63 (14.4)
Preoperative weight loss (%)* Mean (SD) 4.5 (7.5) 3.4 (6.9)
Preoperative albumine level (g/L)† Mean (SD) 38.7 (5.9) 39.1 (5.3)
Nutritional Risk Index-score‡ Mean (SD) 99.0 (9.5) 99.9 (9.0)
Nutritional status by NRI‡
Not malnourished, NRI: ⬎100 % 49.5 52.4
Borderline malnourished, NRI: 97.5–100 % 12.4 13.4
Mildly malnourished, NRI: 83.5–97.4 % 33.3 29.4
Severely malnourished, NRI: ⬍83.4 % 4.8 4.8
Emergency (not scheduled) operation % 14.1 19.5
*Data available for 193 patients in the Enteral Tube Feeding group and 195 in the Allowed food at will group.
†
Data available for 206 patients in the Enteral Tube Feeding group and 208 in the Allowed food at will group.
‡
Nutritional Risk Index data for calculation available for 186 patients in the Enteral Tube Feeding group, and 187 patients in the Allowed Food
at Will group.
SD indicates standard deviation.

TABLE 3. Operations Performed

Enteral Tube Feeding, Allowed Food at Will, Total,
Operations n (%) n (%) n (%)
Gastrectomy, total 38 (16.7) 39 (17.7) 77 (17.2)
Gastrectomy, subtotal/distal 47 (20.7) 35 (15.9) 82 (18.4)
Pancreaticoduodenectomy (Whipple) 47 (20.7) 35 (15.9) 82 (18.4)
Distal pancreatic resection or 3 (1.3) 14 (6.4) 17 (3.8)
pancreaticojejunostomy
Biliodigestive anastomoses 9 (4.0) 9 (4.1) 18 (4.0)
Hepatic resections 25 (11.0) 23 (10.5) 48 (10.1)
Gastroenterostomy 9 (4.0) 5 (2.3) 14 (3.1)
Operation for bowel obstruction with gross 12 (5.3) 26 (11.8) 38 (8.5)
small bowel dilatation
Open redo fundoplications 1 (0.4) 4 (1.8) 5 (1.1)
Transhiatal distal esophagectomies/resections 1 (0.4) 1 (0.5) 2 (0.5)
of the cardia
Transthoracic esophagectomies 4 (1.8) 2 (0.9) 6 (1.3)
Repair of perforated gastric/duodenal ulcers 8 (3.5) 11 (5.0) 19 (4.3)
Operation for bowel perforation/anastomotic leak 10 (4.4) 2 (0.9) 12 (2.7)
Duodenal resection 0 (—) 1 (0.5) 1 (0.2)
Open pancreaticocystogastrostomy 3 (1.3) 1 (0.5) 4 (0.9)
Choledochotomy 1 (0.4) 1 (0.5) 2 (0.5)
Resection of central bile duct 3 (1.3) 2 (0.9) 5 (1.1)
Total pancreatectomy 0 (—) 1 (0.5) 1 (0.2)
Other major, upper abdominal operations or 6 (2.7) 8 (3.6) 14 (3.1)
reoperations
Total 227 220 447

724 © 2008 Lippincott Williams & Wilkins

Kaliane Cunha - kalianebcunha@gmail.com - IP: 177.37.205.206

 Annals of Surgery • Volume 247, Number 5, May 2008 Food at Will After Surgery Does Not Increase Morbidity

chart: Fig. 1). The groups were comparable by baseline
characteristics (Table 2), and by procedures performed
(Table 3)—also within each hospital.
Complications and Adverse Events
In the ETF group, 76 (33.5%) of 227 patients suffered
major complications compared with 62 (28.2%) of 220 in the
group allowed normal food at will (P ⫽ 0.26). When adjusted
for main baseline variables (gender, age, emergency surgery,
high-risk operations, and operating hospital) there was still no
significant difference (P ⫽ 0.46). The 227 patients in the ETF
group suffered a total of 165 major complications compared
with 100 major complications in the 220 patients allowed
normal food at will (P ⫽ 0.012). In the ETF group, 36
(15.9%) patients were reoperated compared with 29 (13.2%)
in the group allowed normal food at will (P ⫽ 0.50) (Table
4). Thirty-day mortality was 4.4% in the ETF group and 5.0%
in the group allowed normal food at will (P ⫽ 0.83). Total
mortality within the trial period was 8.4% in the ETF
group versus 5.9% in the group allowed normal food at
will (P ⫽ 0.36).
Subgroup analysis of the different operations showed
no significant difference in rates of major complications

TABLE 4. Rate of Predefined Complications

Enteral Tube Feeding Allowed Food at will
n ⴝ 227, n (%) n ⴝ 220, n (%) Difference (95% CI) % P
Major, infectious/surgical
SIRS 20 (8.8) 11 (5.0) 3.8 (⫺0.9 to 8.5) 0.14
Bacteremia 6 (2.6) 2 (0.9) 1.7 (⫺0.7 to 4.1) 0.29
Sepsis 12 (5.3) 8 (3.6) 1.7 (⫺2.1 to 5.5) 0.49
Pneumonia 33 (14.5) 26 (11.8) 2.7 (⫺3.6 to 9.0) 0.41
Anastomotic leak 15 (6.6) 10 (4.5) 2.2 (⫺2.1 to 6.5) 0.41
Bowel necrosis 4 (1.8) 2 (0.9) 0.9 (⫺1.2 to 3.0) 0.69
Bowel perforation 1 (0.4) 1 (0.5) ⫺0.1 (⫺1.3 to 1.1) 1.0
Bowel obstruction 6 (2.6) 2 (0.9) 1.7 (⫺0.7 to 4.1) 0.29
Intra-abdominal abscess 16 (7.0) 8 (3.6) 3.4 (⫺0.7 to 7.5) 0.14
Intra-abdominal hemorrhage 8 (3.5) 6 (2.7) 0.8 (⫺2.4 to 4.0) 0.79
Wound rupture 8 (3.5) 6 (2.7) 0.8 (⫺2.4 to 4.0) 0.79
Pancreatitis 1 (0.4) 0 0.4 (⫺0.4 to 1.2) 1.0
Cholecystitis 0 0 — —
Major, other
Myocardial infarction 5 (2.2) 0 2.2 (0.3 to 4.1) 0.06
Myocardial arrhythmia 10 (4.4) 5 (2.3) 2.1 (⫺1.2 to 5.4) 0.29
Cardiac arrest 1 (0.4) 1 (0.5) ⫺0.1 (⫺1.3 to 1.1) 1.0
Acute cardiac failure 5 (2.2) 1 (0.5) 1.7 (⫺0.4 to 3.8) 0.22
Cerebral hemorrhage/infarction 0 1 (0.5) ⫺0.5 (⫺1.4 to 0.4) 0.49
Pulmonary embolism 1 (0.4) 1 (0.5) ⫺0.1 (⫺1.3 to 1.1) 1.0
Pulmonary insufficiency 12 (5.3) 10 (4.5) 0.8 (⫺3.2 to 4.8) 0.83
Patients with major complication 76 (33.5) 62 (28.2) 5.3 (⫺3.3 to 13.9) 0.26
Total number of major complications 165 100 — 0.012
Patients reoperated 36 (15.9) 29 (13.2) 2.7 (⫺3.8 to 9.2) 0.50
Thirty-day mortality 10 (4.4) 11 (5.0) ⫺0.6 (⫺4.5 to 3.3) 0.83
Total mortality within 8 wk (follow-up) 19 (8.4) 13 (5.9) 2.5 (⫺2.3 to 7.3) 0.36
Minor complications and adverse events
Atelectasis 10 (4.4) 11 (5.0) ⫺0.6 (⫺4.5 to 3.3) 0.83
Wound infection 20 (8.8) 11 (5.0) 3.8 (⫺0.9 to 8.5) 0.14
Need for parenteral nutrition 20 (8.8) 25 (11.4) ⫺2.6 (⫺8.2 to 3.0) 0.43
Therapeutic enteral feeding or 6 (2.6) 2 (0.9) 1.7 (⫺0.7 to 4.1) 0.29
reverting to enteral feeding after
transient withdrawal.
Recurrent need of ICU treatment 15 (6.6) 8 (3.6) 3.0 (⫺1.1 to 7.1) 0.20
Use of therapeutic decompression 33 (14.5) 22 (10.0) 4.5 (⫺1.6 to 10.6) 0.15
tube
Patients with any minor complication/ 65 (29.0) 60 (27.3) 1.7 (⫺6.6 to 10.1) 0.75
adverse event
Total number, minor 104 79 — 0.15
complications/adverse events
SIRS indicates systemic inflammatory response syndrome.

© 2008 Lippincott Williams & Wilkins 725

Kaliane Cunha - kalianebcunha@gmail.com - IP: 177.37.205.206

 Lassen et al Annals of Surgery • Volume 247, Number 5, May 2008

between groups (data not shown) except for total gastrecto- high-risk operations, and operating hospital) the difference
mies, where 6 (15.8%) of 38 patients in the ETF group had an remained significant (P ⫽ 0.03). The time to first bowel
intra-abdominal abscess compared with none of the 39 pa- movement was not significantly different (mean 4.3 days in
tients in the group allowed normal food at will (P ⫽ 0.012). the ETF group and 4.0 days for those allowed normal food at
Adjusting for the presence or lack of an upper GI anastomosis will, P ⫽ 0.112). The length of stay was longer in the ETF
did not result in any significant differences between treatment group than in the group allowed normal food at will (mean
groups regarding anastomotic leak rate, major infectious 16.7 vs. 13.5 days, P ⫽ 0.046, 95% CI for difference in
complications, or number of patients with any major compli- means from 0.1 to 6.3).
cation.
The rates of minor complications or adverse events Follow-up
were not significantly different (Table 4), nor was the need At follow-up, 40 (19.0%) of 211 patients in the ETF
for parenteral nutrition or selective use of nasogastric tube. group who were discharged alive were found to have suffered
a late complication compared with 24 (11.5%) of 209 in the
Enteral Tube Feeding group allowed normal food at will (P ⫽ 0.04) (Table 6).
Of 224 attempts, 223 NCJ catheters were inserted. Four There were significantly more wound infections in the ETF
patients died within the first 5 postoperative days and another group, 17 of 211 versus 5 of 209 patients in the group allowed
8 patients with functioning catheters requested crossover to normal food at will (P ⫽ 0.01). The patients in the ETF group
food. Thirty-one (13.9%) catheters were prematurely re- had a higher postoperative weight loss (mean 5.2 kg reduc-
moved (Table 5). The remaining 181 patients received a tion in 144 patients) than those in the group allowed normal
median of 4.800 mL enteral feed during the first 5 postoper- food at will (mean 4.0 kg reduction in 152 patients, P ⫽
ative days. Three (1.3%) of 224 patients were reoperated as a 0.06). One of the patients in the ETF group developed a
direct result of catheter complications (two intra-abdominal fistula at the site of the catheter insertion.
leaks and one subcutaneous abscess).
Bowel Function and Recovery DISCUSSION
Time to first flatus was significantly longer in the ETF This randomized trial shows that allowing normal food
group (mean 3.0 days) compared with the group allowed at will immediately after major upper GI surgery does not
normal food at will (2.6 days, P ⫽ 0.01, 95% CI for increase the rate of major complications compared with a
difference in means from ⫺0.7 to ⫺0.2). When adjusted for standard regimen of nil-by-mouth and enteral feeding for 5
main baseline variables (gender, age, emergency surgery, days. Furthermore, both time to resumed bowel function and
postdischarge rate of complications were significantly re-
duced for the group allowed normal food at will.
TABLE 5. NCJ-related Outcome Main Outcome
Enteral Tube Feeding The benefit of allowing normal food at will, as opposed
n ⴝ 227, n (%) to nil-by-mouth, with or without enteral tube feeding, has
Insertion not attempted 3 (1.3) been shown repeatedly for colorectal and major gynecologic
Insertion aborted after difficult 1 (0.4) surgery.1– 4 It does not seem to be dependent either on the
dissection kind or quantity of food taken, and high-caloric sip-feeds
Catheter-related complications Proportion of inserted within the first 5 postoperative days have not been shown to
catheters n ⫽ 223 improve clinical recovery.23 Our aim was to investigate a
Catheter-related infection 3 (1.3) regimen of allowing early food at will in major upper GI
Leakage around insertion site 3 (1.3) surgery as it is practiced in colorectal surgery. The amount of
Intra-abdominal leak verified 2 (0.9) food taken during the first 5 days was not an issue in this trial
Intra-abdominal leak suspected 1 (0.4) as we wanted to investigate the assumed hazards of allowing
Problematic removal 3 (1.3) any food at all, ie, to challenge the tradition of nil-by-mouth.
Other complaints 4 (1.8) Hence, we are not claiming that allowing normal food at will
Total 16 (7.2) confers nutritional benefits in the early postoperative period
Reoperation caused by catheter 3 (1.3) or that selective use of nutritional support is never needed. On
Unscheduled removal (within 31/223 (13.9) the other hand, our results indicate that allowing these pa-
first 5 postoperative days) tients to eat normal food is beneficial and that routine enteral
Clogged 5 feeding by NCJ, although probably providing more calories,
Accidentally pulled or fallen-out 13 is associated with a poorer overall outcome.
Caused by catheter-related 6 Parenteral nutrition support was used selectively in
complaint/complication
both study groups at the discretion of the surgeon based on a
Not caused by catheter 7
clinical judgment of inadequate nutritional intake. The use of
Patient-requested removal 3
parenteral nutrition did not indicate that the group allowed
Removed at reoperation for 3
other reason
food at will was clinically underfed compared with the ETF
Unknown 1
group. Furthermore, we believe that eating food, through the
activation of normal digestive reflexes, has an important

726 © 2008 Lippincott Williams & Wilkins

Kaliane Cunha - kalianebcunha@gmail.com - IP: 177.37.205.206

 Annals of Surgery • Volume 247, Number 5, May 2008 Food at Will After Surgery Does Not Increase Morbidity

TABLE 6. Outcome at Follow-up

Variable (unit) Measure Enteral Tube Feeding Allowed Food at Will Difference (95% CI) % P
Interval for physical follow-up Median (ICR) 74 (63–91) 71 (63–85) — 0.58
(days postsurgery)
Unscheduled readmittance n (%) 30/211 (14.2) 35/209 (16.7) ⫺2.5 (⫺9.4 to 4.4) 0.45
within 8 wk
Unscheduled reoperation within n 5 3 —
8 wk
Unscheduled re-endoscopy n 2 2 —
within 8 wk
Postdischarge complication (not
predefined)
Abscess n 5 4 —
Pneumonia/sepsis n 5 3 —
Bile-/pancreatic fistula n 1 2 —
Anastomotic leakage n 1 0 —
Anastomotic stricture n 3 1 —
Wound infection n (%) 17 (8.1) 5 (2.4) 5.7 (1.5 to 9.9) 0.01
Incisional hernia n 5 9 —
Cardiovascular (myocardial n 4 1 —
infarction, acute
myocardial insufficiency,
pulmonary embolus,
cerebral infarction, venous
thrombosis)
Other significant n 5 5 —
Total n 46 30 —
Patients with postdischarge n (%) 40/211 (19.0) 24/209 (11.5) 7.5 (0.7 to 14.3) 0.04
complication
Weight loss (follow-up ⫺ Mean ⫺5.2 ⫺4.0 1.3 (⫺2.6 to 0.1) 0.06*
preoperative) kg
*Weight change available for 144 patients in the ETF group, and 152 patients in the Allowed Food at Will group.
ICR indicates interquartile range.

impact on gut recovery, which is central in overall recovery
after GI surgery.24,25 A multimodal care regimen as used in
this trial (no tubes, thoracic epidural analgesia, avoidance of
parenteral opioids, early mobilization, etc) is probably crucial
to exploit the benefits of allowing early food.24,25 In our
opinion, the literature somewhat erroneously focuses on the
different enteral feeding formulas and unjustly disregards the
benefits of eating food as such.
The total number of major complications was signifi-
cantly lower in the group allowed normal food at will. We
consider such pooling of complications, although widely used
in the medical literature, to be less robust than our primary
end point: the rate of patients suffering one or more major
complications. We therefore refrain from emphasizing this
further. On the other hand, it is important to note that the
catheter-related complications in this trial were similar or
fewer than in most of the recently reported series.14 –16,26 This
indicates that our results are not attributable to errors of
catheter use or technique of insertion.
Recovery
Allowing normal food at will improved recovery of gut
function as indicated by the time to first flatus, which was
significantly shorter for those who were allowed to eat. Also,
significantly fewer patients in the group allowed normal food
had postdischarge complications and they tended to lose less
weight. Length of stay, in an unblinded trial without pre-
defined discharge criteria, is a poor outcome parameter and
may not reflect recovery. However, alongside the data above,
it might indicate that allowing normal food at will enhances
recovery.
Limitations
Separate analysis of the subgroups in this trial did not
reveal any significant increase in morbidity for those allowed
normal food at will. Although the trial was not powered for
such stratification, the results of the largest subgroups
(Whipple procedures, total and subtotal gastrectomies, and
hepatic resections) accorded with this. Because the group of
operations was heterogeneous, our conclusions may still not
be automatically valid for each subgroup. Further trials are
needed for each type of surgery to assess the optimal post-
operative practice and there may be subgroups that would
benefit from nasogastric tubes and nil-by-mouth for some
days. The paucity of esophageal resections in our material
reflects the reluctance of the participating surgeons to include
these patients in the trial. Particular features of these opera-
tions might indicate that they should be treated differently
from the others.13

© 2008 Lippincott Williams & Wilkins 727

Kaliane Cunha - kalianebcunha@gmail.com - IP: 177.37.205.206

 Lassen et al
Approximately 17 percent of the patients were lost to
postdischarge follow-up at 8 weeks. These were evenly
distributed between the treatment groups, and there were no
deaths among them. Major surgical complications would
have prompted immediate readmittance to the operating trial-
hospital. We cannot exclude the possibility of minor nonsur-
gical complications among those lost to follow-up.
Our pretrial estimated rate of major complications was
well below that of the trial population. This reduces the risk
of type-II error as does our pretrial power set at 0.90. It may
be argued that our target reduction (by 50%) is large and that
a smaller difference may well be clinically important. Clearly, a
type-II error for a smaller but clinically important difference
cannot be excluded. The tendency of our results, however, is
uniform. This might indicate that a nonshowing, smaller differ-
ence would be that of a decrease, not an increase, in major
morbidity associated with allowing normal food at will.
Implications and Recommendations
This is the first large randomized trial to challenge the
existing routine of nil-by-mouth in upper GI surgery. Periop-
erative practice for hepatic and pancreatic surgery and other
major upper GI procedures may include early removal of the
nasogastric tube and food at will at some centers, but the
literature in English is scarce.13 Thus, our results may have
important clinical implications for a large group of patients.
We believe that the customary regimen of nil-by-mouth after
major upper GI surgery is unnecessary in most cases. Future
trials are needed to establish the optimal clinical pathway for
each of the main types of operations reported in this study,
but we believe that the majority of these patients should be
allowed normal food at will. There is no evidence that it is
inferior to the current routines, it may enhance recovery, and
there is a potential for reducing costs through less expenditure
on catheters and feeding formulas, and possibly through
shorter lengths of stay. Future trials should focus on the role
of nutrition support as an adjunct to food at will, not as an
alternative.
CONCLUSIONS
This trial has shown that allowing patients to eat normal
food at will from the first day after major upper GI surgery
does not increase morbidity compared with traditional care
with nil-by-mouth and enteral feeding. It is the first large
randomized trial to challenge the existing routine of nil-by-
mouth in these patients.
ACKNOWLEDGMENTS
The authors thank the following individuals who par-
ticipated in various phases of the trial: senior consultant Tore
Gauperaa, MD, PhD, implemented the trial in Arendal and
offered important comments on the manuscript. Implementa-
tion in the other centers was totally dependent on Professors
of Surgery Asgaut Viste (Haukeland University Hospital,
Bergen), Jon Arne Søreide (Stavanger University Hospital,
Stavanger); and Helge Myrvold and Jon-Erik Grønbech (St.
Olavs Hospital, Trondheim University Hospital). For some
time, trial inclusion locally was done by surgeons Glenn
728
Kaliane Cunha - kalianebcunha@gmail.com - IP: 177.37.205.206
Annals of Surgery • Volume 247, Number 5, May 2008
Bjørk and Atle Bernstein (Hospital of Arendal), Jon Sen (now
East Iceland Regional Hospital, Iceland), Jorunn Sandvik
(now Hospital of Ålesund), and Kristian Storli (now St. Olavs
Hospital, Trondheim University Hospital). In addition the
authors thank all the doctors and nurses at the 5 participating
centers that made the trial possible by including patients and
registering outcomes. Bjørn-Odvar Eriksen, MD, PhD and
the staff at the Clinical Research Centre, University Hospital
Northern Norway, Tromsø, provided important assistance in
methodological discussions, the construction of the database,
and in pretrial and post-trial statistical calculations. Profes-
sors Olle Ljungqvist (Ersta Hospital and Karolinska Institu-
tet, Stockholm, Sweden), Barthold Vonen (University Hospi-
tal Northern Norway), and senior consultant Cornelis H.C.
Dejong, MD, PhD (University Hospital Maastricht, The
Netherlands) kindly read the manuscript and offered con-
structive comments. The authors also thank Fresenius Kabi
AS, Norway for providing the enteral nutrition formulas.
REFERENCES
1. Lewis SJ, Egger M, Sylvester PA, et al. Early enteral feeding versus “nil
by mouth” after gastrointestinal surgery: systematic review and meta-
analysis of controlled trials. BMJ. 2001;323:773–776.
2. Han-Geurts IJ, Jeekel J, Tilanus HW, et al. Randomized clinical trial of
patient-controlled versus fixed regimen feeding after elective abdominal
surgery. Br J Surg. 2001;88:1578 –1582.
3. Schilder JM, Hurteau JA, Look KY, et al. A prospective controlled trial
of early postoperative oral intake following major abdominal gyneco-
logic surgery. Gynecol Oncol. 1997;67:235–240. 关See comments兴.
4. Steed HL, Capstick V, Flood C, et al. A randomized controlled trial of
early versus “traditional” postoperative oral intake after major abdomi-
nal gynecologic surgery. Am J Obstet Gynecol. 2002;186:861– 865.
5. Brodner G, Van Aken H, Hertle L, et al. Multimodal perioperative
management– combining thoracic epidural analgesia, forced mobiliza-
tion, and oral nutrition–reduces hormonal and metabolic stress and
improves convalescence after major urologic surgery. Anesth Analg.
2001;92:1594 –1600.
6. Mukherjee D. “Fast-track” abdominal aortic aneurysm repair. Vasc
Endovascular Surg. 2003;37:329 –334.
7. Lassen K, Dejong CH, Ljungqvist O, et al. Nutritional support and oral
intake after gastric resection in five northern European Countries. Dig
Surg. 2005;22:346 –352.
8. Tanaka M. Gastroparesis after a pylorus-preserving pancreatoduodenec-
tomy. Surg Today. 2005;35:345–350.
9. Suehiro T, Matsumata T, Shikada Y, et al. Accelerated rehabilitation
with early postoperative oral feeding following gastrectomy. Hepato-
gastroenterology. 2004;51:1852–1855.
10. Hirao M, Tsujinaka T, Takeno A, et al. Patient-controlled dietary
schedule improves clinical outcome after gastrectomy for gastric cancer.
World J Surg. 2005;29:853– 857.
11. Atkins BZ, Shah AS, Hutcheson KA, et al. Reducing hospital morbidity
and mortality following esophagectomy. Ann Thorac Surg. 2004;78:
1170 –1176.
12. Bisgaard T, Kehlet H. Early oral feeding after elective abdominal
surgery–what are the issues? Nutrition. 2002;18:944 –948.
13. Lassen K, Revhaug A. Early oral nutrition after major upper gastroin-
testinal surgery: why not? Curr Opin Clin Nutr Metab Care. 2006;9:
613– 617.
14. Ryan AM, Rowley SP, Healy LA, et al. Post-oesophagectomy early
enteral nutrition via a needle catheter jejunostomy: 8-year experience at
a specialist unit. Clin Nutr. 2006;25:386 –393.
15. Chin KF, Townsend S, Wong W, et al. A prospective cohort study of
feeding needle catheter jejunostomy in an upper gastrointestinal surgical
unit. Clin Nutr. 2004;23:691– 696.
16. De Gottardi A, Krahenbuhl L, Farhadi J, et al. Clinical experience of
feeding through a needle catheter jejunostomy after major abdominal
operations. Eur J Surg. 1999;165:1055–1060.
© 2008 Lippincott Williams & Wilkins
 Annals of Surgery • Volume 247, Number 5, May 2008
17. Sarr MG. Appropriate use, complications and advantages demonstrated
in 500 consecutive needle catheter jejunostomies. Br J Surg. 1999;86:
557–561.
18. Myers JG, Page CP, Stewart RM, et al. Complications of needle catheter
jejunostomy in 2,022 consecutive applications. Am J Surg. 1995;170:
547–550.
19. Han-Geurts IJ, Verhoef C, Tilanus HW. Relaparotomy following com-
plications of feeding jejunostomy in esophageal surgery. Dig Surg.
2004;21:192–196.
20. Bozzetti F, Braga M, Gianotti L, et al. Postoperative enteral versus
parenteral nutrition in malnourished patients with gastrointestinal can-
cer: a randomised multicentre trial. Lancet. 2001;358:1487–1492.
21. Woodcock N, MacFie J. Optimal nutrition support (and the demise
of the enteral versus parenteral controversy). Nutrition. 2002;18:
523–524.
© 2008 Lippincott Williams & Wilkins
Kaliane Cunha - kalianebcunha@gmail.com - IP: 177.37.205.206
Food at Will After Surgery Does Not Increase Morbidity
22. Perioperative total parenteral nutrition in surgical patients. The Veterans
Affairs Total Parenteral Nutrition Cooperative Study Group. N Engl
J Med 1991;325:525–532. 关see comments兴.
23. MacFie J, Woodcock NP, Palmer MD, et al. Oral dietary supplements in
pre- and postoperative surgical patients: a prospective and randomized
clinical trial. Nutrition. 2000;16:723–728.
24. Kehlet H, Holte K. Review of postoperative ileus. Am J Surg. 2001;
182:3S–10S.
25. Gatt M, Anderson AD, Reddy BS, et al. Randomized clinical trial of
multimodal optimization of surgical care in patients undergoing major
colonic resection. Br J Surg. 2005;92:1354 –1362.
26. Lobo DN, Williams RN, Welch NT, et al. Early postoperative jejunostomy
feeding with an immune modulating diet in patients undergoing resectional
surgery for upper gastrointestinal cancer: a prospective, randomized, con-
trolled, double-blind study. Clin Nutr. 2006;25:716 –726.
729