ARTICLE IN PRESS

The effect of early mobilization

protocols on postoperative outcomes
following abdominal and thoracic
surgery: A systematic review
Tanya Castelino, MD,a Julio F. Fiore, Jr, PhD,a Petru Niculiseanu, MD,a Tara Landry, MLIS,b
Berson Augustin, BSc,c and Liane S. Feldman, MD,a Montreal, Quebec, Canada

Background. Early mobilization is considered an important element of postoperative care; however, how
best to implement this intervention in clinical practice is unknown. This systematic review summarizes
the evidence regarding the impact of specific early mobilization protocols on postoperative outcomes after
abdominal and thoracic surgery.
Method. The review was performed according to PRISMA guidelines. We searched 8 electronic databases
to identify studies comparing patients receiving a specific protocol of early mobilization to a control
group. Methodologic quality was assessed using the Downs and Black tool.
Results. Four studies in abdominal surgery (3 randomized controlled trials [RCTs] and 1 observational
prospective study) and 4 studies in thoracic surgery (3 RCTs and 1 observational retrospective study)
were identified. None of the 5 studies evaluating postoperative complications reported differences between
groups. One of 4 studies evaluating duration of stay reported a significant decrease in the intervention
group. One of 3 studies evaluating gastrointestinal function reported differences in favor of the
intervention group. One of 4 studies evaluating performance-based outcomes reported differences in
favor of the intervention group. One of 5 studies evaluating patient-reported outcomes reported
differences in favor of the intervention group. Overall methodologic quality was poor.
Conclusion. Few comparative studies have evaluated the impact of early mobilization protocols on
outcomes after abdominal and thoracic surgery. The quality of these studies was poor and results were
conflicting. Although bed rest is harmful, there is little available evidence to guide clinicians in effective
early mobilization protocols that increase mobilization and improve outcomes. (Surgery 2016;j:j-j.)

From the Steinberg-Bernstein Centre for Minimally Invasive Surgery and Innovation,a McGill University
Health Centre; Montreal General Hospital Medical Library,b McGill University Health Centre; Division of
Anesthesia,c Montreal General Hospital, Montreal, Quebec, Canada

EARLY MOBILIZATION is regarded as an important after surgery was Dr Emil Ries, a gynecologist in
aspect of postoperative care. One of the first sur- Chicago, in 1899.1 Despite the report by Ries, the
geons to describe the concept of early mobilization practice of early postoperative mobilization was
slow to gain favor in North America and patients
were still commonly kept on strict bed rest for
Funding sources: The primary author is funded by the Surgical several weeks after surgery to minimize pain and
Scientist Program at McGill University and the Fonds de re- ensure adequate healing of wounds.2 It was only
cherche du Qu ebec – Sante (FRQS). by the 1940s that early mobilization became
The Steinberg-Bernstein Centre for Minimally Invasive Surgery accepted among surgeons, because a number of
and Innovation (McGill University Health Centre, Montreal,
observational studies suggested that this practice
QC, Canada) is funded through an unrestricted educational
grant from Covidien Canada. was not harmful to patients.3,4 In addition, evi-
Accepted for publication November 8, 2015.
dence about the negative effects of immobilization
Reprint requests: Liane S. Feldman, MD, McGill University
(ie, risk of thromboembolism, pneumonia, muscle
Health Centre, Montreal General Hospital, 1650 Cedar wasting, and physical deconditioning) also became
Avenue, L9-309, Montreal, QC H3G 1A4. E-mail: liane. available, reinforcing the importance of avoiding
feldman@[Link]. prolonged bed rest after surgery.5
0039-6060/$ - see front matter Within the last 20 years, there has been signif-
Ó 2016 Elsevier Inc. All rights reserved. icant progress in perioperative care with the
[Link] development of standardized enhanced recovery

SURGERY 1
 ARTICLE IN PRESS
2 Castelino et al Surgery
j 2016

pathways (ERPs). ERPs combine many different outcomes in comparison to standard care? Rele-
elements of care in the preoperative, intraoper- vant search and index terms were used to capture
ative and postoperative periods, and aim to reduce the following concepts: thoracic and abdominal
morbidity, decrease hospital duration of stay, and surgery (eg, thoracic, abdomen, abdominal,
improve patients’ recovery after surgery.6 ERPs are gastric, colorectal), early mobilization or exercise
comprised of up to 25 different interventions in (eg, early, accelerated, inpatient, postoperative,
the perioperative period; however, the relative postsurgical, ambulation, walking, exercise ther-
contribution of each of these elements to the over- apy) and relevant outcomes (eg, complications,
all recovery process remains unclear.7 Early mobili- duration of stay, patient-reported outcomes
zation is considered to be a key component of [PROs], pain, quality of life). The MEDLINE
ERPs, consistent with the goals of supporting the search strategy is provided in Appendix 1. The
early reestablishment of normal function.8 reference lists of included studies were searched
Guidelines for perioperative care from the for relevant articles. The MEDLINE strategy was
Enhanced Recovery After Surgery Society8 give rerun before submission (February 6, 2015) and
early mobilization a strong recommendation no relevant studies were found.
grade, despite a very low level of evidence support- Inclusion and exclusion criteria. Studies were
ing its use. Although it is suggested that early mobi- included in the review if they met the following
lization within an ERP is an independent predictor criteria: (1) involved adult patients undergoing
of early recovery after colon cancer surgery,9 abdominal or thoracic surgery, (2) a specific pro-
adherence to this intervention remains quite tocol for early in-hospital mobilization was used as
low.10 There is little evidence in the literature an intervention (with out-of-bed activities starting
regarding strategies to promote compliance to no later than postoperative day 1), (3) a control
early mobilization, and significant differences in group receiving either no structured mobilization
mobilization goals between programs. A potential protocol (ie, patients were allowed to mobilize at
approach to increase compliance is by using a spe- will) or a different mobilization protocol (ie, if a
cific mobilization protocol supported by personnel standardized mobilization/physiotherapy protocol
dedicated to mobilizing patients, like a physiother- was already in place at the institution) was used as
apist; however, the additional benefit of this a comparator, (4) reported $1 of the outcome
resource-intensive approach is unknown. In this measures of interest, and (5) were published in
systematic review, we summarize the evidence English or French. Studies were excluded if (1)
regarding the impact of early in-hospital mobiliza- they involved patients undergoing cardiac or
tion protocols on postoperative outcomes after orthopedic procedures, (2) the early mobilization
abdominal and thoracic surgery in comparison protocol was not described by the authors, and (3)
with standard care. the early mobilization protocol was not tested in
isolation (eg, mobilization protocol within an
METHODS enhanced recovery program versus traditional
This systematic review was registered at PROS- care). We also excluded studies that used addi-
PERO International prospective register of system- tional outpatient mobilization strategies (without
atic reviews (CRD42015014684) and was reporting any in-hospital outcomes) because we
conducted according to the PRISMA Statement felt that we would not be able to separate the
guidelines.11 effects of early versus late mobilization on post-
Search strategy. The database search was per- discharge outcomes. Studies where bed rest was
formed by 1 investigator (T.L.) using the following prescribed for the control group were also
databases: MEDLINE (via OvidSP 1946 to January excluded because this practice is no longer reflec-
19, 2015; via PubMed 1946 to January 19, 2015); tive of standard postoperative care.
Embase Classic + Embase (via OvidSP 1947 to Outcome measures. Outcome measures of in-
January 19, 2015); BIOSIS Previews (via OvidSP terest in this review included postoperative compli-
1969 to 2015 week 7); CINAHL (via Ebsco 1937 to cation rates, hospital duration of stay,
January 19, 2015); Web of Science (via Thomson- postoperative pulmonary function (spirometry),
Reuters 1996 to January 19, 2015); Scopus (via postoperative gastrointestinal (GI) function,
Elsevier 1996 to January 19, 2015); and CENTRAL performance-based functional tests (eg, 6-minute
(via the Cochrane Library–issue 1 of 12, January walk test [6MWT]), PROs (measures of health
2015). The search was conducted to answer the status collected directly from patients through
research question: to what extent do early mobili- questionnaires), and adverse events. Data on the
zation protocols impact upon postoperative explanatory variable of amount of physical activity,
 ARTICLE IN PRESS
Surgery Castelino et al 3
Volume j, Number j

as well as data on costs and adherence to the RESULTS

mobilization protocol were also extracted if Literature search. The literature search yielded
available. a total of 4,546 citations after the removal of
Study selection and data extraction. Two inde- duplicates. Of these citations, 102 full-text arti-
pendent reviewers (T.C., B.A.) screened through cles were screened and 94 were excluded (Fig).
the titles and abstracts of the articles yielded by the The excluded articles and reasons for exclusion
search strategy. Articles that were clearly irrelevant are provided in Appendix 2. The main reasons
were excluded. The remaining full-text articles for exclusion were intervention was tested as
were then screened independently against the part of an enhanced recovery program
selection criteria by 2 reviewers (T.C., P.N.). Dis- (n = 18), the article was an editorial or review
crepancies were resolved by consensus within the (n = 16), the article was not in English or French
research group. (n = 13), and the study did not involve a control
Data were then extracted independently from group (n = 11). Eight full-text articles met our
the articles by 2 investigators (T.C., P.N.) into a selection criteria and were included in the
standardized data collection form. In addition to review.17-24
the outcome measures of interest, information Characteristics of the included studies. Four
about the study design, number of patients, age, studies involved patients undergoing abdominal
gender, American Society of Anesthesiologists surgery (3 RCTs18-20 and 1 prospective observa-
score, preoperative diagnosis, type of operation, tional study17) and 4 studies involved patients un-
and operative approach were collected. dergoing thoracic surgery (3 RCTs21-23 and 1
Quality assessment. The methodologic quality retrospective observational study24). The charac-
of each study was independently evaluated by 2 teristics of these studies are summarized in
investigators (T.C., J.F.) using the Downs and Table I and their quality assessment scores are
Black tool.12 This tool was chosen because it ap- shown in Table II. The RCTs by Ahn et al18 and
praises the quality of both randomized controlled Granger et al22 had the highest methodologic
trials (RCTs) and nonrandomized comparative quality (a score of 18/28 on the Downs and Black
studies and has been shown to have good internal tool). The RCT by Waldhausen et al20 was the study
consistency, test–retest reliability, interrater with the lowest quality (score of 7/28). Common
reliability, and criterion-related validity.12 A methodologic issues observed in the included
large-scale review assessing 194 tools to evaluate studies were poor reporting (ie, no data on adverse
methodologic quality deemed the Downs and events17-21,23 and losses to follow-up18-24), lack of
Black tool as appropriate for use in systematic re- information on external validity (ie, sampling strat-
views.13 The Downs and Black tool consists of 27 egy not described17-24), lack of blinding of
items divided into 5 subscales: reporting (10 outcome assessors,17-21,24 lack of randomiza-
items), external validity (3 items), bias (7 items), tion,17,20,24 lack of concealment of allocation,17-24
confounding (6 items), and power (1 item). The and lack of information on statistical
original tool generates an overall score with a power.17,20,22,24
maximum of 32 points but, as recommended in There were a total of 508 participants included
previous literature,14,15 we used a modified in the 8 studies, 225 abdominal surgery patients
version with a maximum score of 28 (for and 283 thoracic surgery patients. There was
simplicity, the last item was scored 0 or 1 instead inconsistent reporting of sample characteristics,
of the original range of 0–5). Disagreements with 1 study20 completely omitting this informa-
regarding the quality assessment were resolved tion (Table III). There were no differences be-
by consensus within the research group. tween mobilization and control groups in terms
Data analysis. We intended to conduct a meta- of age and gender, except in 1 RCT,21 where there
analysis if studies were sufficiently homogeneous were more females in the mobilization group (55%
with respect to design, population, interventions, vs 36% in the control group; P = .03). Studies in
and outcome measures; however, the studies abdominal surgery included patients undergoing
identified by the search were considerably het- various GI procedures,17,20 colon resection for can-
erogeneous. Because pooling of data from het- cer,18 or hysterectomy for benign and malignant
erogeneous studies into a metaanalysis can diseases.19 The majority of these studies included
produce misleading results,16 this systematic re- patients undergoing both open and laparoscopic
view is reported using a narrative synthesis surgery.17-19 All studies in thoracic surgery
approach. included patients undergoing lung resection for
 ARTICLE IN PRESS
4 Castelino et al Surgery
j 2016

Records identified through

database search (n = 5853)
• MEDLINE ( n = 1075) Additional records
• Embase (n = 3496) identified through
• Biosis (n = 47) reference lists
• CINAHL (n = 251) (n = 534)
• PubMed (n = 40)
• Cochrane Library (n = 73)
• Web of Science (n = 289)
• Scopus (n = 582)

Records after duplicates removed

(n = 4546)

Excluded based on
titles and abstracts
(n = 4444)

Full-text articles assessed for

eligibility (n = 102)

Full-text articles excluded (n = 94)

-Abstracts (n = 11)
-Editorials/Reviews (n = 16)
-Early mobilization not tested in isolation (n = 18)
Studies included in the review -Intervention did not involve specific protocol of
(n = 8) early mobilization starting no later than POD1
(n = 7)
-Early mobilization not described by authors (n =
2)
-Mobilization after discharge (n = 6)
-No control group (n = 11)
-Did not involve abdominal or thoracic surgery (n
= 4)
-Control group prescribed bed rest (n = 3)
-Not in English or French (n = 13)
-None of outcome measures (n = 1)
-Ambulatory surgery (n = 1)
-Did not involve adult patients (n = 1)

Fig. PRISMA flowchart showing included and excluded articles.

cancer, with the majority including both open sur- thoracic surgery, the intervention comprised a
gery and video-assisted thoracoscopic surgery.21,23 protocol of early mobilization supervised by a
There was a considerable amount of variation health care professional. In 1 study, the protocol
in the protocols of early mobilization received by involved only sitting and walking24 and 3 studies
patients (Table IV). Three studies in abdominal included aerobic (walking or cycling) and
surgery involved early mobilization protocols su- strengthening exercises.21-23 Mobilization proto-
pervised by a health professional17,18,20 and in 1 cols in abdominal surgery were compared with a
study the protocol was unsupervised (involved control group where patients did not receive a
clear goal setting with encouragement and educa- specific early mobilization intervention, but were
tion).19 In 3 studies, the protocol involved only not restricted to bed rest, whereas all studies in
sitting and walking17,19,20 and 1 study included thoracic surgery involved a control group
more complex exercises (stretching, straight- receiving some form of less intensive early mobili-
ening, and balance).18 In all studies involving zation intervention supervised by a health care
 ARTICLE IN PRESS
Surgery Castelino et al 5
Volume j, Number j

Table I. Summary of characteristics of included studies

Total no. of
Reference Type of surgery Study design patients Primary outcome Follow-up*
Le, 2014 Abdominal Prospective 30 PRO 1 month
observational
Ahn, 2013 Abdominal RCT 31 Duration of stay 30 days
Liebermann, 2013 Abdominal RCT 129 No. of steps 24 hours In-hospital
before discharge
Waldhausen, 1990 Abdominal RCT 35 Myoelectric activity of 1 month
bowel wall
Arbane, 2014 Thoracic RCT 131 Physical activity 4 weeks
Granger, 2013 Thoracic RCT 15 Safety, Feasibility 12 weeks
Arbane, 2011 Thoracic RCT 51 PRO 12 weeks
Kaneda, 2007 Thoracic Retrospective 86 Safety Not specified
Observational
*Duration of maximum follow-up.
PRO, Patient-reported outcome; RCT, randomized controlled trial.

Table II. Quality assessment scores of included studies, using modified Downs and Black Checklist
Internal
External Internal validity –
Study Reporting (11) validity (3) validity – Bias (7) Confounding (6) Power (1) Total (28)
Abdominal
Le, 2014 8 0 1 3 0 12
Ahn, 2013 9 0 4 4 1 18
Liebermann, 2013 7 1 4 3 1 16
Waldhausen, 1990 6 0 1 0 0 7
Thoracic
Arbane, 2014 6 1 2 4 1 14
Granger, 2013 8 1 5 4 0 18
Arbane, 2011 6 0 2 1 1 10
Kaneda, 2007 6 1 1 2 0 10

professional, which was already the standard of complications predefined by the surgical team,
care at the institution. and divided them into respiratory, cardiac and
Postoperative complications. Of the studies other, and also included mortality and transfer to
involving abdominal surgery, only 217,18 reported critical care units >72 hours after surgery. They
on postoperative complications. Ahn et al18 found 31% of participants in the mobilization
described 1 wound infection (5.9%) in the mobili- group (16% respiratory, 8% cardiac, 8% other)
zation group and 1 case of postoperative ileus and 33% of participants in the control group
(7.1%) in the control group, and there were no re- (24% respiratory, 0% cardiac, 9% other) suffered
operations or readmissions within 30 days of hospi- postoperative complications, with no inferential
tal discharge. Le et al17 reported the same rate of statistics reported (P values). The 2011 study by Ar-
overall complications in both groups (26.7%), bane et al23 defined postoperative complications as
including pancreatic fistula, abscess, dehiscence, “x-ray changes reported by radiologist as pneu-
difficulty weaning from total parenteral nutrition, monia, respiratory complications requiring addi-
acute anemia secondary to blood loss, and wound tional ventilatory support and/or necessitating a
infection. Neither study reported an inferential return to high dependency care.” There were 2 pa-
comparison between groups (ie, P values). tients in the mobilization group and 3 patients in
Of the thoracic studies,21,23,24 3 reported on the control group who had predefined complica-
postoperative complications; however, these com- tions; however, there were no differences between
plications were not uniformly defined across all the groups. Kaneda et al24 reported only specific
studies. The 2014 study by Arbane et al21 had postoperative complications, including respiratory
 6 Castelino et al
Table III. Sample characteristics
Sample size
(n) Age (y) Male (%) BMI (kg/m2) Diagnosis Surgical approach* Duration of stay
Reference Int Control Int Control Int Control Int Control Int Control Int Control Int Control
Le, 2014 15 15 48.9 ± 9.8 51.4 ± 8.7 40 53 NR NR NR NR“Similar” “Similar” 5.2 ± 5.5 5.1 ± 6
Ahn, 2013 17 14 55.6 ± 7.1 57.4 ± 6.1 71 36 24.27 ± 3.39 22.59 ± 2.01 Colon Colon Lap 14 Lap 11 7.82 ± 1.07 9.86 ± 2.66
cancer cancer
Open 2 Open 2

ARTICLE IN PRESS
Robotic 1 Robotic 1
Liebermann, 61 68 56 53 0 0 30.5 30.6 Gyne-onc 35 Gyne-onc 29 Lap 13 Lap 12 1.54 1.71
2013 Uro-gyne 17 Uro-gyne 24 Open 14 Open 20
Benign 9 Benign 15 Robotic 27 Robotic 22
Vaginal 7 Vaginal 14
Waldhausen, 10 25 NR NR NR NR NR NR NR NR Open Open NR NR
1990
Arbane, 2014 64 67 67 ± 11 68 ± 11 45 64 26 ± 4.6 26 ± 4.7 NSCLC NSCLC VATS 19 VATS 12 7.5 (5–8) 7.1 (6–8)
Open 45 Open 45
Granger, 2013 7 8 57 ± 16.2 72.4 ± 12.4 42.9 62.5 26.9 ± 4.7 28.7 ± 5.9 Suspected or Suspected or NR NR 4 (3–9) 9 (4–17)
confirmed confirmed
cancer cancer
Arbane, 2011 26 25 65.4 62.6 NR NR 25.5 ± 3.6 25.7 ± 4.8 NSCLC NSCLC VATS or VATS or 8.9 ± 3.3 11.0 ± 8.9
open open
Kaneda, 2007 36 50 65 ± 9 66 ± 9 61 54 NR NR NSCLC NSCLC Open Open NR NR
*Open refers to laparotomy for abdominal surgery and thoracotomy for thoracic surgery.
Values are reported as mean ± standard deviation, median (interquartile range).
BMI, Body mass index; gyne-onc, gynecologic-oncology; Int, intervention; lap, laparoscopic; NR, not reported; NSCLC, non-small cell lung cancer; uro-gyne, urologic-gynecology; VATS, video-assisted thoracoscopic
surgery.

j 2016
Surgery
 ARTICLE IN PRESS
Surgery Castelino et al 7
Volume j, Number j

Table IV. Specific mobilization protocols of each study

Study Mobilization group Control group
Le, 2014 Walking with volunteers, a minimum of 1 lap Not walking with volunteers (ie, walking
around the floor independently, ad lib)
Ahn, 2013 POD1:
Supervised exercise (twice/day) Unsupervised sitting or walking in the ward
Stretching (neck, shoulder, wrist, ankle, pelvis)
Core exercise (pelvic tilt)
Resistance exercise
Unsupervised sitting or walking in the ward
POD1–3:
Supervised exercise (twice/day) Unsupervised sitting or walking in the ward
Stretching (whole body, leg, shoulder)
Core exercise (pelvic tilt and thrust, 1 leg raise,
crunch)
Resistance exercise (chest, shoulder, arm,
thigh, calf)
Unsupervised walking in the hallway
POD2–discharge:
One supervised and 1 unsupervised exercise Unsupervised sitting or walking in the ward
Stretching (whole body, leg, shoulder)
Core exercise (pelvic tilt, bridge, 1 leg raise,
crunch)
Resistance exercise (chest, shoulder, arm,
thigh, calf) (12 repetition 33 sets)
Supervised balance exercise (once/day)
One leg standing, 1 leg calf raise, hip adduc-
tion/abduction, hip flexion with knee bent,
hip extension
Unsupervised walking in the hallway

Liebermann, 2013 Specific ambulation goal: $500 steps before No extra encouragement for ambulation
discharge No ambulation goals
Bedside signs, signs on the patient’s hospital
room door
Reminders at every encounter with health care
team members
Waldhausen, 1990 Ambulatory regimen starting 12–24 hours after Did not ambulate outside their hospital rooms
operation until after POD4
Walking $75 yards during each session
Any ad lib walking that patients desired
Arbane, 2014 Standard care, as control group Standard care, including routine
Once-daily cycle (30 min/session) and strength physiotherapy, airway clearance techniques,
training sessions from POD1-5 and upper limb activities
Additional daily mobilization encouraged
Upon discharge, home walking program
Granger, 2013 Standard care, as control group Standard care, including routine physiotherapy
Twice daily structured exercise program and mobilization, respiratory physiotherapy
involving aerobic, resistance, and stretching if developed pulmonary complications, and
exercises from POD1 until discharge thoracic spine and shoulder stretches
Upon discharge, home exercise routine
(continued)
 ARTICLE IN PRESS
8 Castelino et al Surgery
j 2016

Table IV. (continued)

Study Mobilization group Control group
Arbane, 2011 Standard care, as control group Standard care, including routine
Twice daily strength and mobility training from physiotherapy, airway clearance techniques,
POD1-5 mobilization as able, and upper limb
Upon discharge, further 12-week home activities
program of paced exercise
Kaneda, 2007 Sitting position for 30 minutes, 3.5 hours after Same protocol as the mobilization group,
surgery except was performed on POD1
Walking approximately 30 metres total, 4 hours
after surgery
POD, Postoperative day.

complications (bacterial pneumonia, acute exacer- Return of GI function. In abdominal surgery, 2

bation of interstitial pneumonia, acute respiratory studies18,20 included GI function as an outcome of
distress syndrome), deep venous thrombosis, and interest, with varying degrees of detail. Waldhau-
skin ulcers, of which no participants suffered. sen et al20 reported return of GI function as the
Hospital duration of stay. Of the abdominal primary outcome of interest, and the authors
studies, 317-19 reported on hospital duration of stay. report this in terms of myoelectric activity of the
Ahn et al18 reported the mean ± standard devia- gut wall and do not describe this phenomenon
tion (SD) duration of stay in the mobilization in clinical terms. The authors carried out a com-
group to be significantly shorter, at plex study in which they placed seromuscular bipo-
7.82 ± 1.07 days versus 9.86 ± 2.66 days in the con- lar recording electrodes in the stomach, jejunum,
trol group (P = .005). The other 2 studies did not colon, and Roux limb (if present) at the time of
find a significant difference. Liebermann et al.19 laparotomy. They subsequently measured “slow
found a mean duration of stay of 1.54 days in the wave frequency, presence of migrating myoelectric
mobilization group and 1.71 days in the control complexes (MMCs), amount of spike activity in
group (P = .388). Le et al.17 showed a mean ± SD phases II and III and presence of colonic discrete
duration of stay of 5.2 ± 5.5 days in the mobiliza- and continuous electric-response activity patterns”
tion group and 5.1 ± 6 days in the control group at postoperative days (POD) 1–5 and 7, and at 1
(P = .98). month postoperatively, in both the early ambula-
Of the thoracic studies,21-23 3 reported informa- tion and control groups. These recordings were
tion regarding hospital duration of stay and none to determine if early ambulation affects the risk
found differences between the groups. Arbane of postoperative ileus. They found no difference
et al21 found a median duration of stay of 7.5 between the groups in terms of early recovery of
days (interquartile range, 5–8) in the mobilization GI myoelectric activity. Ahn et al18 reported the
group and 7.1 days (interquartile range, 6–8) in mean ± SD time to first flatus, which was demon-
the control group. Arbane et al23 showed a strated to be statistically significant between groups
mean ± SD duration of stay of 8.9 ± 3.3 days in (52.18 ± 21.55 hours after surgery in the interven-
the mobilization group and 11.0 ± 8.9 days in the tion group vs 71.86 ± 29.2 hours in the control
control group. Granger et al22 found a median group; P = .036). The authors also reported the
duration of stay of 4 days (range, 3–9) in the mobi- mean time to liquid diet intake; however, there
lization group and 9 days (4–17) in the control was no difference between the groups
group. (76.91 ± 24.36 hours in the intervention group
Pulmonary function tests. None of the abdom- vs 86.04 ± 20.68 hours in the control group;
inal studies reported pulmonary function test P = .177).
results. Three of the thoracic surgery studies21-23 In thoracic surgery, only 1 study24 reported in-
report baseline pulmonary function test results formation regarding GI function. The Kaneda
(eg, forced expiratory volume in 1 second and et al study quantified the “amount of diet”
forced vital capacity) and found no differences be- consumed on POD1, with increased intake indi-
tween mobilization and control groups; however, cating a faster return of GI function, but found
none of these studies assess pulmonary function no difference between the groups (66.9 ± 35.6%
in the postoperative period as one of their in the intervention group vs 55.8 ± 35.6% in the
outcome measures. control group; P = .16).
 ARTICLE IN PRESS
Surgery Castelino et al 9
Volume j, Number j

Performance-based functional tests. One composite scores between the groups. The study
abdominal study18 evaluated the effect of an early by Liebermann et al19 used a visual analog scale
postoperative mobilization protocol on to rate difficulty with walking, and also developed
performance-based measures. The tests were per- a nonvalidated questionnaire that included 10
formed before surgery and at hospital discharge. questions regarding perception of barriers to
There were no differences detected postopera- ambulation. They found a significant difference
tively for the sit-stand test (15.00 ± 4.75 repetitions among participants between preoperative difficulty
in the intervention group vs 13.00 ± 5.54 repeti- ambulating (1.47/10) and postoperative difficulty
tions in the control group; P = .208), balance abil- ambulating (4.79/10; P < .001). They had an
ity (15.46 ± 15.27 seconds in the intervention 80% response rate for their questionnaire, which
group vs 8.18 ± 6.49 seconds in the control group; showed that the most common obstacles to ambu-
P = .722), and functional capacity as measured by lation included urinary catheters (38.5%), intrave-
the Tecumseh step test (90.21 ± 11.50 beats per nous poles (28%), and pain (12.5%). These
minute in the intervention group vs findings were not different between groups.
100.50 ± 12.00 beats per minute in the control All 4 studies in thoracic surgery evaluated the
group; P = .877). impact of the intervention on PROs. Three of
Three of the studies in thoracic surgery21-23 these studies21-23 included a home exercise pro-
included data on performance-based functional gram and reported only postdischarge data, so
tests; however, these studies involved a home exer- their results were not analyzed. The fourth study
cise program component, so only in-hospital data in thoracic surgery, by Kaneda et al,24 used the
were analyzed for this analysis. The 2014 Arbane modified Borg scale to quantify postoperative
et al21 trial reported results from the incremental pain, and found no difference between the inter-
shuttle walk test and the quadriceps strength test vention and control groups (2.1 ± 1.3 vs
at 5 days after surgery, which showed no differ- 2.0 ± 1.2, respectively; P = .68).
ences between mobilization and control groups. Physical activity. There were 2 studies in abdom-
The 2011 Arbane et al23 trial included results inal surgery18,19 and 1 study in thoracic surgery21
from the 6MWT and the quadriceps strength test. that documented the amount of physical activity.
The authors found no differences in 6MWT results The study by Ahn et al18 stated that the “amount
between the groups at 5 days after surgery; howev- of walking was monitored daily” but the authors
er, they did find a significant difference in quadri- did not describe how this monitoring was accom-
ceps strength between the groups, with the plished, for example, direct observation versus
mobilization group at 37.6 ± 27.1 kg versus the self-reported by patients versus activity monitors.
control group at 21.5 ± 7.7 kg (P = .04). The study The mean ± SD walking distance during hospital
by Granger et al22 did not report any performance- stay was 1,481 ± 651 meters for the exercise group
based outcomes during hospital stay. and 2,187 ± 1,469 meters for the control group;
PROs. In abdominal surgery, 217,19 of the 4 however, these values were not significant
studies included PROs. Le et al17 administered a (P = .12). The primary outcome of the study by Lie-
modified version of the Patient Recovery Profile- bermann et al19 was the number of steps taken in
17 (PRP-17)25 at hospital discharge and the Short the 24 hours immediately before discharge. Pa-
Form 12v2 (SF-12v2)26 at 1 month after discharge. tients wore pedometers to capture these data,
The authors reported an increased PRP-17 com- and the median number of steps taken was 80
posite score in the control group (12.5 vs 9.9 in (range, 0–2,353) for the ambulation group and
the walking group; P = .003), indicating that the 87 (range, 0–3,576) for the control group
control group had a better postoperative recovery. (P = .70).
They also reported indicator sums in the context The only study in thoracic surgery that attemp-
of the PRP-17, which are defined as the “total num- ted to document amount of physical activity was
ber of axes in which an individual reports no symp- the 2014 Arbane study.21 The primary outcome of
toms,” where increasing scores correlate with this study was physical activity, and patients wore
better recovery. The indicator sums were higher activity monitors preoperatively, during the first 5
in the walking group (9.8 vs 8.4 in the control postoperative days or until hospital discharge,
group; P = .04). Finally, the authors described a and for 1 week at 4 weeks postoperatively to cap-
trend toward better scores of the physical compos- ture these data. The authors state that there were
ite score of the SF-12v2 in the walking group changes in hospital admission policy that pre-
(44.4 ± 5.4 vs 41.7 ± 4.3 in the control group; vented the use of these data and that they only ob-
P = .07), but there was no difference in mental tained data for 16% of patients; however, there
 ARTICLE IN PRESS
10 Castelino et al Surgery
j 2016

were no differences between the groups for preop- Index.28 Most of the studies selectively compared
erative and 4-week postoperative data, using complication rates between mobilization groups
imputed data. The authors do not mention differ- and control groups using descriptive statistics;
ences for postoperative in-hospital activity (ie, however, there was no mention of the significance
POD0-5). of these results, except for 1 study.23 In addition,
Other outcomes. The study by Ahn et al18 re- only 1 study18 reported the duration of follow-up
ported that 84.5% of the patients were adherent in which postoperative complications were consid-
to the mobilization protocol, but no other studies ered. In future research, complications should be
reported on adherence. Two studies in thoracic defined using standardized classification systems,
surgery reported on adverse events during mobili- follow-up periods should be specified, and inferen-
zation. In the study by Granger et al,22 no patients tial statistics (ie, P values and/or 95% CIs)
had abnormal vital signs, new-onset arrhythmias, reported.
chest pain, diaphoresis, or falls during the exercise Duration of stay was a heterogeneous outcome
sessions. Kaneda et al24 also did not report any variable between the studies, which suggests that
adverse events during the intervention (ie, falls there were very different patient populations being
or chest tube issues). studied. However, within each study, whether
None of the studies included in this review involving thoracic or abdominal surgery, the di-
reported a cost analysis. agnoses and types of operations performed were
similar. Duration of stay was reported in the
DISCUSSION majority of studies, and only 118 found a significant
Although convincing evidence suggests that difference between the intervention group and the
patients should not be kept in bed after surgery,5 control group. Although it is a relatively good qual-
there is little guidance on how best to achieve early ity study, it had a small sample size that was even
mobilization, particularly whether adhering to a smaller than anticipated because of early termina-
specific structured mobilization protocol has addi- tion of recruitment owing to implementation of a
tional benefits compared with allowing patients to mandatory ERP (and the resulting violation of
mobilize at will (ie, as tolerated). This systematic the study’s exclusion criteria). These results may
review demonstrates a gap in the body of evidence be suggestive of a decreased duration of stay in pa-
regarding the impact of specific early mobilization tients who ambulate early after surgery. Although
protocols on postoperative outcomes after abdom- some other studies demonstrated a trend toward
inal and thoracic surgery, compared with allowing a shorter duration of stay in the intervention
patients to ambulate but without a specific proto- groups, none of these results were significant.
col. Only 8 relevant articles were identified based This discrepancy of results can be attributed to dif-
on the inclusion and exclusion criteria, and of ferences in statistical power or to the presence of
these studies, only 2 are of relatively good quality. nonclinical factors delaying discharge. Duration
Because results were generally inconsistent, we of stay is not an ideal outcome for studies on inter-
were not able to draw strong conclusions regarding ventions aimed to improve postoperative recovery,
the benefits of early postoperative mobilization because this measure is influenced by several con-
protocols. founders (eg, health care system, surgeon prefer-
One challenge in performing trials of complex ences, patient expectations)29,30 and patients are
interventions is the selection of the primary not necessarily discharged when they are clinically
outcome. Because the negative effects of bed rest ready or “recovered.”
are well-known (eg, thromboembolism, pneu- Functional status is an important outcome of
monia, muscle wasting and physical decondition- recovery for patients31 and performance-based
ing), mobilization could be hypothesized to functional testing is an objective way of measuring
decrease the risk of complications associated with postoperative recovery. The results of
immobilization. Five of 8 studies reported on performance-based functional testing for 3 studies
postoperative complications, 2 in abdominal sur- were considered in this review. In thoracic surgery,
gery and 3 in thoracic surgery, without major only 1 study23 found a significant difference for
differences reported between intervention and one of the tests of functional capacity between
control groups. However, the definition of compli- the groups; however, there were some missing
cations was variable, and none of these studies data and therefore these results should be inter-
used a classification system of postoperative com- preted with caution. The abdominal surgery
plications, such as the Clavien-Dindo classifica- study18 did not find any differences in functional
tion27 or the Comprehensive Complication capacity between the groups. Perhaps these studies
 ARTICLE IN PRESS
Surgery Castelino et al 11
Volume j, Number j

did not find differences in performance testing was a difference in physical activity levels between
simply because it was too early in the process of re- the intervention and control groups. The mea-
covery to find a detectable difference. Alterna- surement of physical activity can be performed in
tively, one could argue that performing these different ways---for example, number of steps, time
functional tests early in recovery may increase the spent out of bed, or intensity of activity (light
likelihood of detecting significant differences be- versus moderate versus high)---and it is not known
tween the groups and that, in fact, early mobiliza- which of these is the best parameter to measure
tion may not increase functional capacity in the physical activity. Future studies should be done to
immediate postoperative period. Because these elucidate the best way to quantify physical activity
studies were RCTs with relatively high methodo- levels in postoperative patients.
logic quality, the somewhat consistent finding Information on cost analysis and adherence was
that early mobilization protocols have little impact not provided by the studies in this review. There is
on performance-based outcomes may be more little known about the costs associated with imple-
meaningful than other results reported in this mentation of specific mobilization strategies that
review. may require additional resources. The types of
The description of PROs was widely variable mobilization protocols included in the studies in
among the studies included in this review and the this review were very variable. Some of them
quality of reporting is questionable. Only 1 study24 included interventions from physiotherapists,
in thoracic surgery used a validated scale, the Borg whereas others simply involved encouragement
scale for pain, and provided a correct interpreta- from health care workers with specific goals of
tion of these results. The other study17 in ambulation, or walking with volunteers. The costs
abdominal surgery using questionnaires for post- associated with each of these different programs
operative recovery (PRP-17) and health-related are likely to be variable, and adherence may differ.
quality of life (SF-12v2) provided conflicting re- One may argue that protocols that are unsuper-
sults. The PRP-17 composite scores and indicator vised and self-motivated may result in lower adher-
sums did not correlate as they should, but this ence in comparison with supervised protocols and,
finding was not explained. Although the authors thus, patients would not experience the theoretical
concluded that these results were both indicative benefits of early mobilization. The results from this
of better recovery in the ambulation group, in review, however, do not provide evidence about the
fact they found conflicting results. The last study19 superiority of supervised protocols compared
in abdominal surgery that included PROs used a with other strategies. Further research in this
questionnaire that was created by the authors area would be helpful to elucidate the best
and not validated by previous research. The au- ways to achieve maximum adherence in a cost-
thors presented information regarding barriers effective way.
specific to ambulation, and did not find a differ- This review was limited by the small number of
ence between groups. This analysis may be helpful studies identified and by the methodologic limita-
to identify potential strategies to improve patients’ tions of these studies. There was substantial het-
ability to walk postoperatively and increase adher- erogeneity in study design, specific mobilization
ence to the mobilization protocol; however, the re- protocols, and outcome reporting, which supports
sults are not necessarily generalizable to other our decision not to conduct a metaanalysis. Also,
centers. language bias cannot be excluded as we targeted
Information regarding physical activity levels only articles in English and French. A large num-
and comparisons between intervention and con- ber of articles (94) were excluded after full-text
trol groups should be documented. Studies consid- review, with the majority excluded because they
ering physical activity levels that were included in did not evaluate mobilization in isolation, did not
this review do not comment on their findings, that involve an early mobilization protocol, or did not
is, the lack of significant differences between the involve a control group. This finding is not a
groups, and how this impacts their ability to draw limitation of the review, but rather demonstrates
conclusions regarding the effect of early mobiliza- the paucity of comparative studies in this field.
tion protocols on outcomes. To demonstrate This may reflect the challenges of conducting
differences in outcome variables such as postoper- RCTs with “behavioral” interventions such as mobi-
ative complications, duration of stay, performance- lization; however, multiple trials involving periop-
based functional testing, and so on, it is first erative exercise (eg, “prehabilitation”32,33) have
necessary to show that the patients actually partic- been successfully conducted, suggesting that
ipated in the mobilization protocol and that there RCTs in this field are feasible. Both abdominal
 ARTICLE IN PRESS
12 Castelino et al Surgery
j 2016

surgery and thoracic surgery were included in this standardized and durations of follow-up specified.
review, because we aimed to evaluate the effects of Examples of well-validated measures for use in
early mobilization protocols on a broad group of future research include postoperative complica-
postoperative patients. We understand that these tions as classified by Clavien-Dindo27 and the
2 patient groups may have different barriers associ- Comprehensive Complication Index28 and recov-
ated with postoperative mobilization; for example, ery of functional walking capacity as measured by
the challenge of having $1 chest tubes may make the 6MWT.36 The use of measures of functional
mobilization more cumbersome in thoracic sur- in-hospital recovery (eg, time to functional recov-
gery. As such, we have chosen to present our re- ery,30 time to readiness for discharge37) can over-
sults and analysis for abdominal and thoracic come the limitations associated with measuring
surgery separately. duration of stay. However, outcomes like complica-
Although it is well-recognized that prolonged tions and hospital duration of stay have many
bed rest is harmful and should not be advocated in other influences in addition to patient mobiliza-
postoperative care,34 the best way to manage post- tion. It is important to report time out of bed
operative mobilization remains unknown. It is and time spent mobilizing as explanatory variables
intuitive that using specific protocols to facilitate that can be measured using Actigraphy or pedom-
(or “enforce”) early mobilization would be benefi- eters. When nonrandomized trials are conducted,
cial; however, several questions regarding the clin- authors should follow specific criteria for reporting
ical effects of this intervention remain their results, for example, using the STROBE
unanswered. At what frequency and intensity Statement.38 Data on the relationship between
should patients mobilize after surgery? What mobi- adherence to mobilization protocols and postoper-
lization targets should be used? Do we need ative outcomes would also provide important
personnel dedicated to facilitate early mobiliza- information.
tion? Do patients treated with an early mobiliza- In conclusion, there remain many unanswered
tion protocol have better postoperative outcomes questions regarding the impact of using a specific
compared with those mobilizing at will (ie, being early mobilization protocol on postoperative out-
counseled not to stay in bed, but mobilizing as comes after abdominal and thoracic surgery, and
much as they feel comfortable)? This review high- whether additional resources should be committed
lights the need for further studies in this field. to achieving specific mobilization goals. No firm
Because RCTs provide the optimal design for conclusions can be drawn from this review as
studies on health interventions, we believe that studies were generally of poor methodologic qual-
this should be the design of choice. In these trials, ity and had conflicting results. Some studies sug-
specific mobilization protocols should be gest that the use of early mobilization protocols
compared with currents standards of care at has the potential to accelerate return of bowel
different institutions (eg, preoperative education function and reduce hospital duration of stay,
regarding early mobilization, daily encouragement which is encouraging for future research.
by surgeons and other health care professionals). Although there is a strong body of literature
Studies should also be performed to determine suggesting that prolonged bed rest is harmful,34
what type of physical activity should be advocated whether a specific protocol ensures early mobiliza-
in postoperative patients, as well as intensity and tion above and beyond what is accomplished by pa-
duration, and whether or not certain thresholds tient education is unknown. Early mobilization
can be targeted in the early postoperative period. protocols, especially when driven by additional
Conducting such studies in institutions using dedicated health professionals, may require addi-
ERPs may help understanding the relative contri- tional resources and should be justified by
bution of early mobilization protocols in this evidence.
context of care. Adherence to early mobilization
within ERPs is low,10 but the need to implement in- SUPPLEMENTARY DATA
terventions to enhance adherence is uncertain. Supplementary data related to this article can be found
Trials should follow the CONSORT Statement35 at [Link]
to optimize study design and reporting. Methods
of randomization and concealment of allocation,
REFERENCES
blinding of outcome assessors, and management
1. Ries E. Some radical changes in the after-treatment of cel-
of missing data should be set carefully. Sample iotomy cases. JAMA 1899;XXXIII:454-6.
size calculations should be conducted to prevent 2. Brieger GH. Early ambulation. A study in the history of sur-
type II error. Outcome measures should be gery. Ann Surg 1983;197:443-9.
 ARTICLE IN PRESS
Surgery Castelino et al 13
Volume j, Number j

3. Leithauser DJ, Bergo HL. Early rising and ambulatory activ- activity after curative surgery for non-small cell lung cancer:
ity after operation: a means of preventing complications. a multicentre randomised controlled trial. Physiotherapy
Arch Surg 1941;42:1086-93. 2014;100:100-7.
4. D’Ingianni V. Early and late postoperative ambulation: a 22. Granger CL, Chao C, McDonald CF, Berney S, Denehy L.
comparative study of three hundred and three cases. Arch Safety and feasibility of an exercise intervention for patients
Surg 1945;50:214-8. following lung resection: A pilot randomized controlled
5. Bed rest, thrombosis, and embolism. Lancet 1958;1:465-6. trial. Integr Cancer Ther 2013;12:213-24.
6. Kehlet HMDP, Wilmore DWMD. Evidence-based surgical 23. Arbane G, Tropman D, Jackson D, Garrod R. Evaluation of
care and the evolution of fast-track surgery. Ann Surg an early exercise intervention after thoracotomy for non-
2008;248:189-98. small cell lung cancer (NSCLC), effects on quality of life,
7. Varadhan KK, Lobo DN, Ljungqvist O. Enhanced Recovery muscle strength and exercise tolerance: randomised
after surgery: the future of improving surgical care. Crit controlled trial. Lung Cancer 2011;71:229-34.
Care Clin 2010;26:527-47. 24. Kaneda H, Saito Y, Okamoto M, Maniwa T, Minami K,
8. Gustafsson UO, Scott MJ, Schwenk W, Demartines N, Roulin D, Imamura H. Early postoperative mobilization with walking
Francis N, et al. Guidelines for perioperative care in elective at 4 hours after lobectomy in lung cancer patients. Gen
colonic surgery: Enhanced Recovery After Surgery (ERASÒ) Thorac Cardiovasc Surg 2007;55:493-8.
Society recommendations. World J Surg 2013;37:259-84. 25. Allvin R, Svensson E, Rawal N, Ehnfors M, Kling A-M,
9. Vlug MS, Bartels SAL, Wind J, Ubbink DT, Hollmann MW, Idvall E. The Postoperative Recovery Profile (PRP) – a
Bemelman WA, et al. Which fast track elements predict multidimensional questionnaire for evaluation of recovery
early recovery after colon cancer surgery? Colorectal Dis profiles. J Eval Clin Pract 2011;17:236-43.
2012;14:1001-8. 26. Gandek B, Ware JE, Aaronson NK, Apolone G, Bjorner JB,
10. Smart NJ, White P, Allison AS, Ockrim JB, Kennedy RH, Brazier JE, et al. Cross-Validation of Item Selection and
Francis NK. Deviation and failure of enhanced recovery af- Scoring for the SF-12 Health Survey in Nine Countries: Re-
ter surgery following laparoscopic colorectal surgery: early sults from the IQOLA Project. J Clin Epidemiol 1998;51:
prediction model. Colorectal Dis 2012;14:e727-34. 1171-8.
11. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, 27. Dindo D, Demartines N, Clavien P-A. Classification of surgi-
Ioannidis JPA, et al. The PRISMA statement for reporting cal complications: a new proposal with evaluation in a
systematic reviews and meta-analyses of studies that evaluate cohort of 6336 patients and results of a survey. Ann Surg
healthcare interventions: explanation and elaboration. BMJ 2004;240:205-13.
2009;339:b2700. 28. Slankamenac K, Graf R, Barkun J, Puhan MA, Clavien PA.
12. Downs SH, Black N. The feasibility of creating a checklist for The comprehensive complication index: a novel continuous
the assessment of the methodological quality both of rando- scale to measure surgical morbidity. Ann Surg 2013;258:1-7.
mised and non-randomised studies of health care 29. Lee L, Tran T, Mayo NE, Carli F, Feldman LS. What does it
interventions. J Epidemiol Community Health 1998;52:377-84. really mean to “recover” from an operation? Surgery 155:
13. Deeks JJ, Dinnes J, D’Amico R, Sowden AJ, Sakarovitch C, 211–6.
Song F, et al. Evaluating non-randomised intervention 30. Maessen JM, Dejong CH, Kessels AG, von Meyenfeldt MF.
studies. Health Technol Assess 2003;7:iii-x:1-173. Length of stay: an inappropriate readout of the success of
14. Eng JJ, Teasell R, Miller WC, Wolfe DL, Townson AF, enhanced recovery programs. World J Surg 2008;32:971-5.
Aubut J-A, et al. Spinal Cord Injury Rehabilitation Evidence: 31. Lee L, Dumitra T, Fiore JF Jr, Mayo NE, Feldman LS. How
Methods of the SCIRE Systematic Review. Top Spinal Cord well are we measuring postoperative “recovery” after
Inj Rehabil 2007;13:1-10. abdominal surgery? Qual Life Res 2015;24:2583-90.
15. Aubut JA, Marshall S, Bayley M, Teasell RW. A comparison 32. Carli F, Charlebois P, Stein B, Feldman L, Zavorsky G,
of the PEDro and Downs and Black quality assessment tools Kim DJ, et al. Randomized clinical trial of prehabilitation
using the acquired brain injury intervention literature. in colorectal surgery. Br J Surg 2010;97:1187-97.
NeuroRehabilitation 2013;32:95-102. 33. Gillis C, Li C, Lee L, Awasthi R, Augustin B, Gamsa A, et al.
16. Higgins JPT, Green S. Cochrane handbook for systematic Prehabilitation versus rehabilitation: a randomized control
reviews of interventions version 5.1.0 [updated March trial in patients undergoing colorectal resection for cancer.
2011]. The Cochrane Collaboration. Available from: www. Anesthesiology 2014;121:937-47.
[Link].2011. Accessed May 1, 2015. 34. Harper CM, Lyles YM. Physiology and complications of bed
17. Le H, Khankhanian P, Joshi N, Maa J, Crevensten H. Patients rest. J Am Geriatr Soc 1988;36:1047-54.
recovering from abdominal surgery who walked with volun- 35. Schulz KF, Altman DG, Moher D. CONSORT group. CON-
teers had improved postoperative recovery profiles during SORT 2010 statement: updated guidelines for reporting par-
their hospitalization. World J Surg 2014;38:1961-5. allel group randomised trials. BMJ 2010 Mar 23;340:c332.
18. Ahn KY, Hur H, Kim DH, Min J, Jeong DH, Chu SH, et al. 36. Moriello C, Mayo NE, Feldman L, Carli F. Validating the six-
The effects of inpatient exercise therapy on the length of minute walk test as a measure of recovery after elective colon
hospital stay in stages I-III colon cancer patients: random- resection surgery. Arch Phys Med Rehabil 2008;89:1083-9.
ized controlled trial. Int J Colorectal Dis 2013;28:643-51. 37. Fiore JF Jr, Faragher IG, Bialocerkowski A, Browning L,
19. Liebermann M, Awad M, Dejong M, Rivard C, Sinacore J, Denehy L. Time to readiness for discharge is a valid and
Brubaker L. Ambulation of hospitalized gynecologic surgi- reliable measure of short-term recovery after colorectal sur-
cal patients: a randomized controlled trial. Obstet Gynecol gery. World J Surg 2013;37:2927-34.
2013;121:533-7. 38. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC,
20. Waldhausen JH, Schirmer BD. The effect of ambulation on Vandenbroucke JP. The Strengthening the Reporting of
recovery from postoperative ileus. Ann Surg 1990;212:671-7. Observational Studies in Epidemiology (STROBE) state-
21. Arbane G, Douiri A, Hart N, Hopkinson NS, Singh S, ment: guidelines for reporting observational studies. Ann
Speed C, et al. Effect of postoperative physical training on Intern Med 2007;147:573-7.