laparoscopic versus open appendectomy in adults with complicated appendicitis- systematic review and...
TRANSCRIPT
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
1/15
Laparoscopic Versus Open Appendectomy in Adultswith Complicated Appendicitis: Systematic Review
and Meta-analysis
Georgios Markides Daren Subar
Kallingal Riyad
Societe Internationale de Chirurgie 2010
Abstract
Background The goal of the present study was to criti-cally review and identify the strength of available evidence
in the literature on the use of laparoscopic appendectomy
(LA) in complicated appendicitis (CA).
Methods The Cochrane Library and Controlled Trials
Registry, MEDLINE (Ovid), PubMed, Web of knowledge,
and SCOPUS databases were electronically searched, using
the keywords appendectomy, laparoscopy, appen-
dicitis. complicated appendicitis. gangrenous appen-
dicitis, perforated appendicitis, with English language
as a limit. Backward chaining was also employed. The
NHS Public Health Resource Unit Critical Appraisal Skills
Programme Tools were used for critical appraisal.
Results Twelve retrospective case-control studies were
included in the review. Overall methodological quality was
moderate to poor, with heterogeneity, absence of random-
ization and blinding, and presence of important methodo-
logical flaws. Meta-analysis showed that LA in CA has
reduced surgical site infection (SSI) rates compared to
open appendectomy (OA), odds ratio (OR) 0.23, 95%
confidence intervals (CI): 0.140.37 (level 3a evidence),
and no difference with regard to intra-abdominal abscess
(IAA) complication rates OR: 1.02, 95% CI 0.561.86
(level 3a evidence).Conclusions When compared to OA, laparoscopic
appendectomy is advantageous in CA with regard to SSIs,
with no significant additional risk of IAA (level 3a
evidence).
Introduction
Evidence-based practice is the cornerstone for best medical
practice. It is the process of systematically finding,
appraising and using contemporaneous research findings as
the basis of clinical decisions [1]. In the face of a clinical
problem such as the choice between laparoscopic appen-
dectomy (LA) and open appendectomy (OA) best evidence
derived from level 1 quantitative research like well-
designed randomized double-blinded control trials should
be used to decide on the best management. In the absence
of such trials, evidence from lower down in the hierarchical
evidence ladder should be carefully considered.
Acute appendicitis is a surgical condition with incidence
of 1.17 per 1,000 [2] and lifetime risk of 8.6% in males and
6.7% in females [3]. Complicated appendicitis (CA) occurs
once the appendix has become gangrenous and/or has
perforated with various degrees of peritonitis. The rate of
CA is slowly increasing [4] and has been reported at an
incidence of 1230% [5]. It is associated with higher
morbidity rateswound infection rates 20% versus 5% in
noncomplicated appendicitis (nCA) [6]and mortality
rates5% versus 0.8% in nCA [7].
Since 1894, after the description of a series of case
studies by McBurney, OA via the McBurney approach has
been the gold standard procedure for appendectomy [8]. In
the early 1980s Semm [9] described the first laparoscopic
G. Markides (&)
Department of General Surgery, Royal Oldham Hospital,
Oldham, UK
e-mail: [email protected]
D. Subar
Department of General Surgery, Stepping Hill Hospital,
Stockport, UK
K. Riyad
Department of General Surgery, Royal Blackburn Hospital,
Blackburn, UK
123
World J Surg
DOI 10.1007/s00268-010-0669-z
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
2/15
appendectomy. Subsequent technological advances and
improvement of surgical laparoscopic techniques and
expertise have given ground for trying to introduce this
procedure as the mainstay of treatment for appendicitis. A
recent systematic review and meta-analysis has shown that
the advantages of LA in nCA appear to focus on the
reduction of postoperative pain, wound infection rates, and
in-hospital stay compared to OA, and that it has distinctdiagnostic advantages in females of reproductive age [10].
The use of LA in CA though has been controversial and
associated with increased rates of intra-abdominal abscess
(IAA) complications [10, 11]. More recent studies have
refuted the above finding, suggesting reduced postoperative
wound infection rates of LA compared to OA in CA, and
are actually recommending LA as the mainstay of treat-
ment for these patients [12, 13]. The aim of the present
study was therefore to systematically review the literature
on the effectiveness of LA in relation to conventional OA
in the management of adult patients with complicated acute
appendicitis, with a subsequent meta-analysis.
Methods
Types of studies
Our search included both randomized controlled trials and
observational studies. Eligibility criteria included all stud-
ies comparing LA to OA in adult patients with CA, with (a)
postoperative surgical site infection (SSI) rates, and/or (b)
postoperative IAA rates, and/or (c) postoperative analgesia,
and/or time to oral intake, and/or length of stay in hospital,
as end points. Studies with insufficient data to form 2 9 2
tables for use in odds ratio (OR) analysis and 95% confi-
dence interval (CI) or studies not providing adequate data
to obtain mean values and standard deviation for continu-
ous variables were excluded from the relevant calculations.
Search and selection strategies
The Cochrane Library and Controlled Trials Registry,
MEDLINE (Ovid) (1966November 2008), PubMed,
EMBASE (1966November 2008), Web of knowledge,
and SCOPUS databases were electronically searched for
appendectomy, laparoscopy, laparoscopic appendec-
tomy, and appendicitis, complicated appendicitis,
gangrenous appendicitis with perforated appendicitis.
The search was limited to studies published in the English
language, and was completed with reference follow-up.
The citations and/or abstracts or full text of all potentially
relevant studies were independently reviewed and selected
by two reviewers, and any disagreements were resolved
with discussion.
Data abstraction and validity assessment
Data on study methodology, participant characteristics,
intervention characteristics, and primary outcomes were
independently extracted and tabulated by two reviewers
with a predefined data extraction form. Attempts to contact
authors were made if there were missing data or unclear
information in the studies, and data were adjustedaccordingly. The studies were subsequently assessed
independently by two reviewers for methodological quality
using the Critical Appraisal Skills Tools for randomized
controlled trials and case-control studies [14, 15], with
criteria marked as met, unclear, and not met. Overall risk of
bias was assessed with established methods and graded as
A (low risk), B (moderate risk), and C (high risk) [16].
Data analysis
The RevMan 5 statistical package [17] was employed to
perform odds ratio (OR) analysis and to assess statisticalheterogeneity via I2, with statistical significance at p\ 0.05
and I2\ 30%, respectively, and to assess publication bias
via funnel plot graphical representation. Meta-analysis was
conducted with the Mantel-Haenszel statistical method for
dichotomous variables, and Inverse Variance was used for
continuous variables. Separate analyses were performed for
each bias risk group and for all groups combined together.
Final recommendations were graded based on the Oxford
Centre of Evidence Based Medicine level of evidence and
grades of recommendation guidelines [18].
Results
The initial literature search identified 381 studies. Based on
the inclusion criteria, 362 studies were excluded, with a
selection of 19 studies for more detailed review. Seven of
those studies were subsequently excluded [1925] (Fig. 1),
including one conference abstract with limited study meth-
odology information, leaving 12 retrospective observational/
case-controlled studies for systematic review. A summary of
the studies methodology, intervention characteristics, and
measured outcomes is provided later in Tables 1, 2 and 3.
Discussion
Summary of studies
Randomization, concealment of allocation, and blinding
are inherently absent from the methodological design of the
included studies. In the absence of a standardized care
protocol, patients were allocated to either operative group
World J Surg
123
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
3/15
based on the surgeons discretion. This is an important bias
because it reflects the surgeons experience with LA, the
presence of co-morbidities, and the patients clinical con-
dition, all of which may have influenced placement into
either of the patient group [2628].
Another of the problems encountered in the appraisal of
the studies is the case definition for CA. Most studies
define this as a perforated appendix with or without abscess
formation, with or without peritonitis. Four of the studies
[13, 2931] include gangrenous appendix in their case
definition, presumably because it is difficult macroscopi-
cally to distinguish between a perforated and a nonperfo-
rated gangrenous appendix, allowing for variation in the
interpretation of results between these groups of studies.
Initial diagnosis of CA in all but two cases [12, 30] is made
subjectively based on clinical opinion, and macroscopic
histological confirmation is provided in only four of the
studies [13, 3133]. Validity of the case definition in the
remaining studies is therefore compromised. In an attempt
to avoid this problem, Fukami et al. [12] used CT (com-
puted tomography) scanning to confirm CA prior to group
allocation. A drawback of this definition though would be
that patients with perforation but minimal abdominal fluid
that cannot be diagnosed by CT scanning were excluded
from the study thus creating a potential selection bias and
less general applicability of the studys results (reduced
external validity). An alternative is provided by Yau et al.
[30], who diagnose CA laparoscopically first and then
proceed to separate the two groups. This appears to be a
reasonable way to reach a good case definition, as the
diagnostic effects of an initial laparoscopy in suspected
appendicitis have been well established [10]. It does though
introduce the bias of an initial diagnostic laparoscopy.
Two of the studies [34, 35] do not report any patient
characteristics and should be viewed with great caution
with regard to generalizing their results to a normal pop-
ulation. In the remaining studies statistical testing of these
independent variables varies significantly (Table 1) with
external validity implications. Yau et al. [30] are the only
authors who used exclusion criteria regarding age. It is
difficult to judge from the reported ages how big the
pediatric contribution is to these studies, as some infor-
mation is presented as median age standard deviation
and some as mean age, rather than medians with the
interquartile ranges. From a cautious review, it appears that
the majority of patients included are adults as the range of
reported median and mean ages is 3648 and 3040 years,
respectively, between groups. The exception to this is the
study by Pokala et al. [29], which reports the inclusion of
24 pediatric cases in a total study population of 104. In
recognition of that, the authors later test the two popula-
tions separately, but they potentially compromise the sig-
nificance of the studys findings because of a change in the
patient number group ratio (from 43:61 to 28:52).
Intervention characteristics
The level of surgical experience and the learning curve are
two important independent variables requiring consider-
ation. A higher surgical level of experience in LA could
mean a reduced number of converted cases to OA and
potentially lower complication rates. Even though the
Potentially relevant studies identifiedand screened for retrieval
n = 381
Studies retrieved for more detailedevaluation
n= 19
Studies excluded n = 362Not satisfying eligibility criteria
Potentially appropriate studies to beincluded in the meta-analysis
n= 17
Studies included in meta-analysisn = 12(retrospective case-control
studies)
Studies with usable information, byoutcome,
n= 12
Studies excluded n = 2
Retrospective case-series n = 2
Studies withdrawn, by outcome, n = 0
Studies excluded from meta-analysis n = 5Serious methodological flaw with high
bias risk directly affecting outcomen=4
Inadequate data for analysis n=1
Fig. 1 Article selection flow
chart
World J Surg
123
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
4/15
Table1
Summaryofstudies
Study
Exposure
measurement
Samplesize
Inclusioncriteria
Independentvariables
Conversionrate&ITT
Findings
Katsunoetal.[31]
Retrospective
case
notereview
LAn=
146
OAn=
84
May1995May
2007
Allpatientswith
CAfrom
hospitalrecord
s
NSbetweengroups:age,
gender,co-morbiditiesnot
tested
3.4%ITTnotused
SSbetwee
ngroups:lowerrateof
SSIs,lo
ngerOT,analgesicuse,
TTOI,pLOSinLAgroup,
NSbetweengroups:IAA,other
post-op
complications
Fukamietal.[12]
Retrospective
case
notereview
LAn=
34
OAn=
39
Jan1999Dec2004
AllCApatients
undergoing
OAinfirst3yearsandallCA
patientsunderg
oingLAin
last3years
NSbetweengroups:age,
gender,obesity,co-
morbidities
0%N/A
SSbetwee
ngroups:lowerrateof
SSIs,analgesicuse,TTOI,
duration
ofdrainage,pLOSin
LAgroup
NSbetweengroups:OT,IAA,
hernia,fistulaformation
Kirshteinetal.[13]Retrospective
case
notereview
LAn=
50
OAn=
98
Jan2005Aug2005
Allpatientswith
CAfrom
hospitalrecord
s
Sbetweengroups:more
femalesofreproductiveage,
higherco-morbidity,obese
andunknownabdominal
pain
ptsinLAgroup
16%ITTnotused
(conversionstreatedas
separategroup)
NSbetweengroups:OT,TTOI
(solids),pLOS,SSI,IAA
Pokalaetal.[29]
Retrospective
case
notereview
LAn=
43
OAn=
61
Jan2003Feb2006
Allpatientswith
CAfrom
hospitalrecord
s
Sbetweengroups:more
pediatriccasesinLAgro
up
(separateanalysiscontacted),
NSbetweengroups:gende
r,
ASAgroup
18.6%ITTused
SSbetwee
ngroups:longerOT,
higherIAArateinLAgroup
NSbetweengroups:overall
complic
ationsandSSI
Yauetal.[30]
Retrospective
case
notereview
LAn=
175
OAn=
244
Jan1999Jan20
04
Allpatientswith
CAfrom
hospitalrecord
s
NSbetweengroups:age,
gender
13.7%ITTused
SSbetwee
ngroups:shorterOT,
LOSan
dSSIinLAgroup
NSbetweengroups:IAArates
Linetal.[37]
Retrospective
case
notereview
LAn=
99
OAn=
130
Jan2001Dec2003
Allpatientswith
CAfrom
hospitalrecord
s
NSbetweengroups:age,
gender
8%ITTused
SSbetwee
ngroups:longerOT,
shorter
antibioticrequirements,
shorter
TTOIandLOS,lower
SSIinLAgroup
NSbetweengroups:postoperative
analgesia,reoperationrates,IAA
Gulleretal.[35]
NISdatabase
LAn=
1,763
OAn=
12,644
1997
Allpatientswith
appendicitis
fromU.S.nationaldatabase
stratificationfo
rCA
Notreported
Notgiven
SSbetwee
ngroups:shorterLOSin
LAgroup
NSbetweengroups:SSI,post-
operativ
ecomplications
World J Surg
123
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
5/15
Table1
continued
Study
Exposure
measurement
Samplesize
Inclusioncriteria
Independentvariables
Conversionrate&ITT
Findings
Soetal.[36]
Retrospective
case
notereview
LAn=
85
OAn=
146
Jan1992Jun19
99
Allpatientswith
perforated
appendicitis
NSbetweengroups:age,
gender
40%ITTused
SSbetwee
ngroups:less
postoperativeanalgesiaand
complic
ations,reducedTTOI
andLOSinLAgroup
NSbetweengroups:OT
Piksunetal.[32]
Prospectivecase
notecollection
LAn=
28
OAn=
24
52prospectivepatients
Nottested
36%ITTnotused,
groupstested
individually
NSbetweengroups:OT,IAA,
SSI,pLOS
Wullsteinetal.[34]Retrospective
case
notereview
LAn=
217
OAn=
82
Aug1989Mar1
999forOA
Allpatientswith
CA
19911999LA
Notreported
21%(3560%in1991;
23.4%in19981999)
ITTused
SSbetwee
ngroups:SSIlower
inLA
NSbetweengroups:IAA
Stoltzingetal.[33]Retrospective
case
notereview
LAn=
80
OAn=
45
Jul1991Jun1999
Allpatientswith
perforated
appendix
NSbetweengroups:
age,gender,BMI
45%(16%inlatter
years)ITTnotused
SSbetwee
ngroups:SSI
NSbetweengroups:OT,LOS,
IAA
Khalilietal.[44]
Retrospective
case
notereview
LAn=
77
OAn=
122
Jan1994Aug1997
Allpatientswith
acute
appendicitis
Notreported
ITTnotused
SSbetwee
ngroups:OT
NSbetweengroups:LOS,IAA
LAlaparoscopicappendectomy,ASAAmericanSocietyofAnesthesiologistsscore,BMIbodymassindex,OAopen
appendectomy,CAcomplicatedappendic
itis,uCAuncomplicated
appendicitis,
ITTintentiontotreat,
SSstatisticallysignificant,
NSstatisticallynonsignificant,
OToperatingtime,TTOItim
etooralintake,
LOSlengthofstayinhospital,pLOSpostoperative
lengthofstayinhospital,
SSIsurgicalsiteinfection,IAAintra-abdominalabscess
World J Surg
123
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
6/15
Table2
Summaryofstudiesinterventioncharacteristics
Study
CAdiagnosis
Surgeons
LA
OA
Primarywound
closure
Antibiotic
Proce
dure
Meso-appendix&
appendixresection
Retrieval
bag
Procedure
Katsunoetal.
[31]
Intraoperative
1experienced
surgeon?
surgicaltrainees
3trochars,10-5-
5m
m,
?typeof
endoscope
Coagulation?
suture
loopsorstaples
Use
d
McBurneysor
paramedian
Yes
Preoperativeand
postoperativeuseas
clinicallyindicated
Fukamietal.
[12]
CTscan
6experiencedsurgeons
forLA,atleast2other
surgeonsforOA
3trochars,10-10-
5m
m,flexible
end
oscopeused
Electrocauteryor
ultrasound
dissector?
endolinear
cutter
Use
d
McBurneysor
paramedian
Yes
Preoperativeand
postoperativeuseas
clinicallyindicated
Kirshteinetal.
[13]
?clinical
9surgeons?experience
3trochars,45deg
end
oscope
Coagulationor
clips?
sutureloopsor
staples
Use
din
67%
Gridironor
paramedian
Secondary
closurein
contaminated
cases
Preoperativeand
postoperativeuseas
clinicallyindicated
Pokalaetal.[29]Clinical?
CTsca
n
confirmationofCA
Notreported
Notr
eported
Notreported
Notre
ported
Notreported
Notreported
Notreported
Yauetal.[30]
Laparoscopically
Surgicalresidentsat
least3yearsexperience
3trochars,10-5-
5m
m,
?typeof
endoscope
Electrocauteryor
ultrasound
dissector?
suture
loops
Use
d
Gridiron
Notreported
Preoperativeand
postoperativeuseas
clinicallyindicated
Linetal.[37]
Intra-abdominalpus
4experiencedsurgeons
3trochars,10-3-
3m
m,
?typeof
endoscope
Electrocauteryorclipsor
harmonic
scalpel?
clipsor
endoloop
Use
d
McBurneys,
paramedianor
midline
Yes
Preoperativeand
postoperativeuseas
clinicallyindicated
Gulleretal.[35]?clinical
Notreported
Notr
eported
Notreported
Notre
ported
Notreported
Notreported
Notreported
Soetal.[36]
Clinical
?number,lessexperienced
surgeons(n\
20)
associatedwithhigher
conversionrates
3trochars,12-5-
5m
m,
?typeof
endoscope
Electrocautery,clipsor
stapler?
endoloopor
stapler
Use
d
usually
Gridironor
midline
Yesbutin4
cases
Preoperativeand
postoperativeuseas
clinicallyindicated
Piksunetal.[32]?clinical
12surgeons?experience
3trochars,10-5-
12mm,
?typeof
endoscope
?clips?
stapler
Use
d
Notgiven
Yes
Preoperativeand
postoperative5daysand
thenasclinically
indicated
Wullsteinetal.
[34]
?clinical
?surgeons
3trochars,12-5-
12mm,
?typeof
endoscope
Clipsorstapler?
clips
orendoloop
Use
d
McBurneys
Notreported
Preoperativeand
postoperativeuseas
clinicallyindicated
?experience
Learningcurve
World J Surg
123
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
7/15
finding is not statistically tested, So et al. [36] do actually
report that less experienced laparoscopic surgeons (per-
forming fewer than 20 laparoscopic procedures) had a
higher rate of conversion. Trying to compensate for this
problem and only allowing experienced surgeons to per-
form LA, though, would positively bias the results in favor
of LA and reduce external validity. This is because in the
normal hospital environment of a general hospital there is avariation in the level of laparoscopic experience among
both senior surgeons and trainees. As OA is a procedure
performed both during day and night by all these levels of
surgeons, LA should be assessed under similar
circumstances.
Similarly, a surgeon at the start of his or her learning
curve in performing LA, would be more likely to have a
higher rate of conversion to OA, as well as a higher rate of
complications. Including a large number of these cases in
the analysis sample could potentially bias the results
against LA. Such a learning curve is reported by Katsuno
et al. and Wullstein et al. [31, 34]. The inclusion of thisindividual surgeon learning curve in the studies under
some circumstances may be acceptable and can be
assumed to represent a normal variation around hospitals
as inexperienced surgeons learn LA. This though should
only occur after a plateau has been reached where, theo-
retically, all the senior surgeons in an institution have
mastered the art of laparoscopy. Results from studies near
the introduction of LA would therefore be mostly appli-
cable to clinical practice during that particular time period,
with recent results being more applicable to current prac-
tice. The earlier studies included in the present appraisal,
such as the ones by So et al. [36] and Wullstein et al. [34],
both of which include an early institutional learning curve
from the early 1990s and should therefore be reviewed
carefully with regard to current practice.
Randomization of surgeons laparoscopic experience in
the LA group should therefore be the most appropriate
approach. The number of surgeons performing each pro-
cedure should also be sufficient that individual surgeon
bias is not introduced, affecting the results and external
validity. In the studies included herein there is a wide
variation in the numbers of recruited surgeons and their
laparoscopic experience, both between groups and between
studies (Table 2). The effect of this performance bias can
potentially be seen in the studies by Yau et al. [30] and Lin
et al. [37], groups that only use experienced laparoscopic
surgeons in their studies, reporting relatively low conver-
sion rates compared to the other studies (Table 1). A lower
conversion rate could mean an underestimation of wound
infections in the LA group because the use of intention to
treat (ITT) analysis places patients undergoing LA that
required conversion to OA, which means a bigger
abdominal wall incision, into the LA group. In theseTable2
continued
Study
CAdiagnosis
Surgeons
LA
OA
Primarywound
closure
Antibiotic
Proce
dure
Meso-appendix&
appendixresection
Retrieval
bag
Procedure
Stoltzingetal.
[33]
Clinical
?surgeons
3trochars,10-5-
12mm,
?typeof
endoscope
Electrocautery,clipsor
stapler
Use
d
Pararectalincision
orlowermidline
laparotomy
Notreported
Preoperativeand
postoperativeuseas
clinicallyindicated
?experience
Learningcurve
Khalilietal.[44]Clinical
Notreported
Notr
eported
Notreported
Notre
ported
Notreported
Notreported
Notreported
CTcrosstomographyscan,
CAcomp
licatedappendicitis,uCAuncomplicatedappe
ndicitis
World J Surg
123
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
8/15
Table3
Summaryofstudiesmeasuredoutcomes
Study
OT(min)
Postoperative
analgesia(days)
TTOI(days)
LOS(days)SSI
IAA
Otherpostoperative
complications
Follow-up
Notes
Katsunoetal.[31]
LA:117*
OA:96
impentazosine,
LA:0.8,OA:1.2
POlaxoprofen
sodium,LA:4.2,
OA:6.4*
LA:2.7*,
OA:3.6
LA:8.9
*,
OA:1
6.6
LA:24%*,
OA:6.4%
LA:4.3%,
OA:4.8%
Smallbowelobstruction,
LA:2/146,OA:4/84
Enteritis
,OA:1/84
Pneumonia,OA:2/84
1appt
?timing
Oralintakecommenced
afterbo
welsopened,1
experiencedsurgeonwithteam,
learningcurveincluded
Fukamietal.[12]
(n=73)
LA:97.9,
OA:92
impentazosine,
LA:0.7,OA:0.9
prdiclofenac,LA:
2.1,OA:7.5*
LA:2.6*,
OA:5.1
LA:11.7*,
OA:2
5.8,
pLOS
LA:8.8%*,
OA:
48.6%
LA:5.9%,
OA:5.1%
Hernia,LA:1/34
Fistula,OA:1/39
36patients
?timing
43.6%OAptsperformed
underspinalanaesthesia
Kirshteinetal.[13]
(n=90)
LA:45,
OA:34
Notreported
Tosolids,
LA:1,OA:
1
LA:5,OA:
4,pLOS
LA:2%,
OA:2%
LA:7%,
OA:6%
Abdominalwall
haematoma,LA:1/42
Pulmonaryemboli,OA:
1/48
Wounddehiscence,OA:
1/48
30days
ITTanalysisnotused?
Signific
antlymorefemalesof
reprodu
ctiveage&higher
co-morbidityinLAgroup
Pokalaetal.[29]
(n=104)
LA:101*,
OA:82
Notreported
Postoperative
ileus,LA:
14%,OA:
11%
LA:6,OA:
6
LA:2.3%,
OA:8.2%
LA:14%*,
OA:0%
Smallbowelobstruction,
LA:1/43,OA:1/61
Medical
complications,
LA:2/43,OA:4/61
Mortality,OA:1/61
30days
24/104w
erepaediatric
patients
Yauetal.[30]
(n=224)
LA:55*,
OA:70
Notreported
NotmeasuredLA:5*,
OA:6
LA:0.6%*,
OA:10%
LA:5.7%,
OA:4.3%
Notreported
2835
days
Linetal.[37]
(n=229)
LA:96*,
OA:68
impethidine,LA:
1.7,OA:1.5
LA:3.2*,
OA:5
LA:6.3
*,
OA:9
.3
LA:15%*,
OA:40%
LA:3%,
OA:4%
Intra-abdominalbleeding,
LA:1/99
Enterocutaneousfistula,
OA:3/130
1follow-
upvisit
Verysignificantselection
bias-lowsocioeconomicstatus
&high
co-morbidityptsforOA
Gulleretal.[35]
(n=8,839)
Not reported
Notreported
Notreported
LA:3.7
*,
OA:5
,
MedV
NS,p=
0.8
NS
Nopvalue
provided
NotreportedforsubgroupsNo
Stratified
analysisused
Ptcharac
teristicsunavailable
Soetal.[36]
(n=231)
LA:73,
OA:71
impethidine,LA:
2.9,OA:3.2
ponaproxen,LA:
4,OA:2.9
LA:3.9*,
OA:4.6
LA:5.2
,
OA:5
.9
LA:14%,
OA:25%
LA:0%,
OA:1.4%
Re-operation,LA:4/85,
OA:14/146
30day
World J Surg
123
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
9/15
Table3
continued
Study
OT(min)P
ostoperative
a
nalgesia(days)
TTOI(days)
LOS(days)SSI
IAA
Otherpo
stoperative
complica
tions
Follow-up
Notes
Piksunetal.[32]
(n=42)
LA:115,
OA:106
N
otreported
LA:22%,
OA:38%
OA:9.2
,
OA:1
0.5
LA:0%,
OA:14%
LA:28%,
OA:29%
Notrepo
rted
No
5perforatedcasesconverteddue
tolackofsurgicalexperience
ITTanaly
sisnotusedfor10
convertedcases
Wullsteinetal.[34]
(n=299)
Not reported
N
otreported
Notreported
Not reported
LA:10.5%*,
OA:34%
LA:4.6%,
OA:6%
Trocharhernia,
LA:2/217
Haemato
ma,
LA:2/217,OA:1/82
Fistula,LA:1/217,
OA:1/82
No
Ptcharacteristicsunavailable
Highinitialconversionrate(27%
total)learningcurve
Stoltzingetal.[33]
(n=125)
LA:75,
OA:70,
Conv:
90,MedV
N
otreported
Ileus,LA:
5%,OA:
2%
Total,L
A:
10,OA:
11,Co
nv:
11,MedV
LA:11%,
OA:18%,
some
laparot-
omies
LA:6%,
OA:4%
Mortality,OA:
5/45
No
Significan
tselectionbiasas
initially
ptswithsuspectedCA
selected
forOA?
longer
symptom
durationinLA
Khalilietal.[44]
(n=276)
LA:86,
OA:70
N
otreported
Notreported
LA:6.3,
OA:6
.4
Not reported
LA:5%,
OA:5%
Notrepo
rted
No
ITTnotu
sed.Ptcharacteristics
unavaila
ble
imintramuscular,prperrectum,po
peroral,
PRNpro-re-nata(asrequired),M
edVmedianvalue
*Statisticallysignificantdifference
(p\
0.05)
World J Surg
123
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
10/15
studies recommendations would strongly favor LA over
OA.
Standardization of the laparoscopic and open interven-
tions within studies can reduce internal validity errors and
improve the studies external validity [38]. All of the nine
studies that give information on the interventions report the
number and size of trochars used with small inter-study
variation (Table 2). The type of endoscope used is onlyreported by Fukami et al. [12]. Even though the rigid
endoscope is the most commonly used instrument for these
operations, failure to disclose the type used may affect the
studys external validity, as other centers may use different
types of endoscopes. Indeed, Fukami et al. [12] describe
LA in CA with a flexible endoscope at very good success
rates and reduced complication rates.
The method of ligation and resection of the meso-
appendix and the appendix is also reported (Table 2) with
some interstudy and intrastudy variation that can affect the
rates of intra-abdominal complications and comparison
within and between studies. A systematic review is beingconducted by the Cochrane group comparing all the sur-
gical techniques used to close the appendix base, and these
findings should provide some evidence of the significance
of variations in these studies.
The use of a retrieval bag for the removal of the per-
forated or gangrenous appendix from the abdominal cavity
can, theoretically, limit the spread of infection within the
abdominal cavity and at the abdominal skin wound site. All
of the studies report use of the retrieval bag. Fukami et al.
[12], however, used this technique in only 28 of their 42
LA cases. This variation could introduce a significant bias
with regard to the findings of the study, as one of the main
assessment areas between LA and OA is postoperative
complications. The authors themselves note that 4 of the 6
LA infective complications in their series occurred in
patients undergoing LA where the retrieval bag was not
used. The inclusion of these cases in the LA versus OA
comparison of a small number of infective complications
might have introduced a significant alteration to the actual
results, with a potential type I statistical error. So et al. [36]
find themselves in a similar situation.
In the studies being considered, for OA the conventional
approach is described by almost all authors, using a
McBurneys, a Gridiron, or a paramedian incision
(Table 2). There is some variation in the approach to either
primary or secondary closure of the skin wounds. A recent
systematic review and meta-analysis of such variation did
not show any difference in the postoperative rates of sur-
gical site infection (SSI) between the two closure tech-
niques [39]. Kirshtein et al. [13] and So et al. [36]
selectively use secondary closure in patients with a con-
taminated abdomen, whereas other authors report per-
forming only primary closure. Similar variability appears
to exist in the use of a drain when an intra-abdominal
abscess (IAA) is encountered.
All of the studies advocated the use of antibiotics in CA,
which involves one preoperative dose followed by regular
intravenous antibiotics until the patients condition clini-
cally improves and antibiotics are not required. In the
absence of a clear care pathway, the decision to stop
antibiotic administration becomes subjective and depen-dent to individual clinicians opinion and may introduce
bias within the studies and affect external validity. One of
the studies [37] does cite a basis for the cessation of anti-
biotic treatment, which is the resolution of pyrexia or
negative microbiology culture results. The use of a protocol
with a fixed duration of antibiotic treatment, as reported in
the study by Piksun et al. [32], could directly affect one of
the study outcomespostoperative length of hospital stay
(pLOS), eliminating any potentially significant pLOS
benefits gained from LA. In addition, a wide range of broad
spectrum antibiotics are used between studies, so it is not
possible to achieve direct absolute comparison betweenstudy findings.
Assessment of bias risk
All the studies appear to be using reasonable statistical tests
with Students t-test measurement for continuous normally
distributed variables such as operating time and 2 9 2 chi
square analysis or Fishers exact test for categorical data
[40, 41]. Significance is tested at 5%. Intention to treat
(ITT) analysis is appropriately used in all but four of the
studies [13, 3133], introducing attrition bias into these
studies [42]. The number of participants in these studies is
relatively small and unequal owing to the lack of true
randomization between the two operative groups, with the
exception of the study by Guller et al., which is based on a
stratification sample from a nationwide database [35].
None of the studies uses a power analysis, and therefore the
possibility of significant type II error is present [43].
The lack of randomization and the presence of selection,
performance, and measurement bias in all of the studies
affect their methodological validity (Table 4). Significant
problems, such as selection bias, absence of patient char-
acteristics, absence of ITT analysis, and absence of any
intervention characteristics, would only allow for classifi-
cation of findings/conclusions into level 3a to 3a evidence
with regard to practice recommendation [18].
Measured outcomes
The main outcomes measured for assessing the efficacy of
LA against OA in CA in the studies appraised, with short-
term follow-up of up to 45 weeks, are shown on Table 3.
All of the studies, although they agreeon the type of
World J Surg
123
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
11/15
Fig. 3 Time to oral intake meta-analysis (all available data)
Fig. 2 Operative time meta-analysis (all available data)
Table 4 Results of studies risk of bias after critical review
Study Center Selection
bias
Performance
bias
Attrition
bias
Selective
reporting bias
Detection
bias
Statistical
bias
External
validity/bias
Overall
bias
BC D ID ITT SM PA
Katsuno et al. [31] One M M U M NM M U M NM U B
Japan
Fukami et al. [12] One M U NM M M M U M NM U B
Japan
Kirshtein et al. [13] One NM M NM M NM M U M NM M C
Israel
Pokala et al. [29] One NM U U M M M U M NM U C
US
Yau et al [30] One U U U M M M U M NM U B
Hong Kong
Lin et al. [37] One NM U M M M M U M NM M C
Taiwan
Guller et al. [35] [1,000 hospitals U U NM M U M U M NM U C
US
So et al. [36] One U U NM M M M U M NM M B
Singapore
Piksun et al. [32] One U M NM M NM M U M NM M C
Italy
Wullstein et al. [34] One U U U M M M U M NM M C
Germany
Stoltzing et al. [33] One M NM NM M NM M U M NM NM C
Germany
Khalili et al [44] One U U NM M NM M U M NM U B
US
BC baseline characteristics, D diagnosis, ID incomplete data, SMstatistical methods, PA power analysis, M met, Uunclear, NMnot met, A low
risk, B moderate risk, C high risk
World J Surg
123
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
12/15
outcomes measured, fail to define these outcomes, allowing
for the widespread introduction of detection bias and
deficiencies in internal and external validity. Important
measurement outcomes such as in-hospital and out-of-
hospital costs, as well as other life quality issues, such as
time to return to normal daily activities and work and
cosmesis have not been assessed. In addition, long-term
follow-up of patients is lacking, and complications such ashernias and intestinal obstruction lack evaluation.
It is important to note the significant mortality rate of
11% in OA in the study by Stoltzing et al., something that
might be attributable to the initial selection bias of the
study, which during its initial period required that patients
with suspected CA were placed in the OA group. At the
same time, the ASA score between LA and OA patients has
not been statistically verified for absence of bias in either
group.
Operating time
The lack of an operating time (OT) definition may
potentially create internal and external validity problems
surrounding this outcome. Most studies suggest that there
is no significant statistical difference in OT between LA
and OA. Four of the studies considered here report sig-
nificantly longer OT in LA compared to OA [2931, 37],
whereas Fukami et al. [12] report a shorter OT for LA, but
with the presence of reduced external validity because only
experienced laparoscopic surgeons participated in this
study. Meta-analysis of the four studies [31, 36, 37, 44] that
report data in mean SD values indicates that OT is
longer in LA by 12.8 min (p\ 0.01, 95% CI 8.417.3), butwith a significant statistical heterogeneity of I2 = 86%
(Fig. 2). Exclusion of the study by So et al. from the
analysis significantly improves heterogeneity to 36%, but
with no significant change in the final result (mean OT
longer in LA: 21.4 min, 95% CI 15.427.3, p\ 0.01).
Conclusions remain the same when only studies with
moderate bias (group B) or all studies are included in the
analysis (level of evidence 3a).
Postoperative analgesia
Regarding postoperative analgesia, one of the importantadvantages of LA, this is measured in four of the studies
[12, 31, 36, 37], as the use of intramuscular and per-rectal
or oral analgesic use. Three of the studies [12, 31, 36]
conclude that there is significantly less use of postoperative
Fig. 5 Surgical site infection rate meta-analysis (all studies included)
Fig. 4 Length of hospital stay meta-analysis (all available data)
World J Surg
123
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
13/15
analgesia by LA patients. All three of these studies state
that analgesia is given as required (PRN), whereas Linet al. [37] do not define analgesia use. The use of PRN
analgesia may have some drawbacks as it depends on
patient understanding and lack of embarrassment with
regard to its use, as well as staff being available to
administer it. A more objective scoring system for the
assessment of postoperative pain would potentially be a
more accurate way of measuring this outcome. No data
meta-analysis was therefore performed for this outcome.
Time to oral intake
Time to oral intake (TTOI) is measured in six of the studies[12, 13, 31, 32, 36, 37], and two of the studies report the
presence of ileus [29, 33] (Table 4). One variation between
studies in this outcome was the measurement by Kirshtein
et al. of time to solid oral intake [13] rather than any oral
intake. Four of the six studies reporting on TTOI individ-
ually report a significantly shorter TTOI in the LA group.
Meta-analysis of data from the three studies [31, 36, 37]
providing mean SD values (Fig. 3) shows a significant
difference between the two procedures, with mean differ-
ence of TTOI of 0.8 days (p\ 0.01, 95% CI: 0.51.15)
shorter for LA and not significant statistical heterogeneity,
both when only group B or all studies are included in theanalysis (level of evidence 3a).
Length of hospital stay
Similarly, length of stay in hospital (LOS) is only defined
as postoperative length of stay by Fukami et al. [12] and
Kirshtein et al. [13]. Giving results on total length of stay
would introducesignificant bias, as preoperative time is not
an intervention outcome. Five of the studies [12, 30, 31, 35,
37] individually report LOS as shorter in LA. Four studies
providing mean SD values were included in this out-come meta-analysis (Fig. 4), showing a significantly
shorter mean LOS for LA of 1.1 day (p\ 0.01, 95% CI
0.51.6), but with unacceptably high statistical heteroge-
neity (I2 = 92%), and no change in findings when only
group B studies are included in the analysis (level of evi-
dence 3a).
Surgical site infection and intra-abdominal abscess rates
The important outcomes of SSI and IAA are measured in
all of the identified studies. Unfortunately, the great
majority of studies do not explicitly state how the diagnosiswas reached in each of these cases. A diagnosis of SSI
would more accurately require positive organism culture
confirmation rather than just a clinical diagnosis, and IAA
diagnosis should be reached via imaging studies. None of
the studies include the odds ratio in terms of the rate of
each complication, and all of them only report significance
levels.
The great majority of studies agree that there is reduced
rate of SSI in LA and no significant difference in IAA rates
between LA and CA, whereas So et al. [36] only report a
combined reduced incidence in LA. Guller et al. are the
only authors to suggest no difference in the rate of SSIbetween the two procedures.
Meta-analysis of the SSI rates reported (Fig. 5) indicates
that this complication occurs significantly less frequently in
LA than in OA: OR 0.43 (95% CI 0.340.55) with sig-
nificant statistical heterogeneity (I2 = 73%) when all
studies are used for the analysis, and OR values of 0.23
(95% CI 0.140.37) when only group B studies are used,
with moderate statistical heterogeneity I2 = 61% (level of
evidence 3a).
Fig. 6 Intra-abdominal abscess rate meta-analysis (all studies included)
World J Surg
123
-
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
14/15
Meta-analysis of IAA rates (Fig. 6) shows no significant
difference between the two procedures: OR 1.24 (95% CI
0.841.84), with no statistical heterogeneity when all
studies are used and OR 1.02 (95% CI 0.561.86) when
only group B studies are used, with no statistical hetero-
geneity (level of evidence 3a).
Conclusions
With this systematic review and meta-analysis of reports
from the medical literature we have tried to identify the
best and most up-to-date evidence in regards to LA and OA
in the management of adult patients with CA. Overall the
quality of the studies found was intermediate to poor, with
heterogeneity achieving only low grades of evidence rec-
ommendation. The absence of randomization and blinding
and the presence of important methodological flaws intro-
duce a number of systematic errors and significantly reduce
the studies rigor. There appears to be a genuine attempt to
justify the efficacy of LA in the majority of studies, but
their retrospective nature and lack of care pathway stan-
dardization decreases their internal and external validity,
restricting general interpretation and application of their
results.
A cautious interpretation of the findings, having con-
sidered all their deficiencies, indicates that LA may have a
lower incidence of SSI than OA (level 3a evidence). There
appears to be no difference with regard to IAA complica-
tion rates (level 3a evidence), the main LA drawback, as
previously suggested. An important parameter neglected by
these studies though is the long-term complications. Thereported findings provide the initial framework for higher
level studies on LA versus OA in CA, making it possible to
obtain better evidence on the subject and possibly gain a
significant benefit from the adoption of a new procedure as
the gold standard. Blinded randomized controlled trials
comparing LA and OA in CA would be the next step.
These are feasible as they have been performed in nCA in
the past, comparing the two procedures. Similar studies can
be conducted for CA, taking into consideration the prob-
lems with current evidence as identified through this
review.
In current clinical practice LA can be used in CAaccording to the surgeons discretion and laparoscopic
experience. It should, however, be remembered that lapa-
roscopy is a fairly new intervention requiring significant
resources and surgical expertise to reach maximum effi-
ciency. Industrialized countries appear to be meeting these
technical and surgical criteria, but developing and third-
world countries may find a high level of practice difficult to
achieve. In those instances conventional OA should be the
surgical treatment of choice.
References
1. Rosenberg W, Donald A (1995) Evidence based medicine: an
approach to problem solving. BMJ 310:11221126
2. Williams NMA, Everson NW, Jackson D, Johnstone JM (1998) Is
the incidence of acute appendicitis really falling? Ann R Coll
Surg Engl 80:122124
3. Addiss DG, Shaffer N, Fowler BS, Tauxe RV (1990) The epi-
demiology of appendicitis and appendectomy in the UnitedStates. Am J Epidemiol 132:910925
4. Livingston E, Woodward W, Sarosi GA, Haley RW (2007)
Disconnect between incidence of nonperforated and perforated
appendicitis. Implications for pathophysiology and management.
Ann Surg 245:886892
5. Oliak D, Yamini D, Udani V, Lewis RJ, Vargas H, Arnell T,
Stamos M (2007) Can perforated appendicitis be diagnosed pre-
operatively based on admission factors? J Gastrointest Surg
4:470474
6. Humes DJ, Simpson J (2006)Acuteappendicitis. BMJ333:530534
7. Blomqvist PG, Anderson RE, Granath F, Lambe MP, Ekborn AR
(2001) Mortality after appendectomy in Sweden, 19871996.
Ann Surg 233:455460
8. McBurney C (1894) The incision made in the abdominal wall in
cases of appendicitis, with a description of a new method ofoperating. Ann Surg 20:38
9. Semm K (1983) Endoscopic appendectomy. Endoscopy 15:5964
10. Sauerland S, Lefering R, Neugebauer EAM (2004) Laparoscopic
versus open appendicectomy for suspected appendicitis (review).
Cochrane Database of Systematic Reviews Issue 3. Art. No
CD001546. doi:10.1002/14651858.CD001546.pub2
11. Pedersen AG, Petersen OB, Wara P, Renning H, Qvist N, Laur-
berg S (2001) Randomised control trial of laparoscopic versus
open appendectomy. Br J Surg 88:200205
12. Fukami Y, Hasegawa H, Sakamoto E, Komatsu S, Hiromatsu T
(2007) Value of laparoscopic appendectomy in perforated
appendicitis. World J Surg 31:9397
13. Kirshtein B, Bayme M, Domchik S, Mizrahi S, Lantsberg L
(2007) Complicated appendicitis: laparoscopic or conventional
surgery? World J Surg 31:74474914. NHS P H R U (2008) Critical Appraisal Skills Programme10
questions to help you make sense of randomised control trials.
Accessed 17 January 2009
15. NHS P H R U (2008) Critical Appraisal Skills Programme10
questions to help you make sense of a case control study.
Accessed 17 January 2009
16. Higgins JPT, Green S (eds) (2008) Cochrane Handbook for
Systematic Reviews of Interventions Version 5.0.1 [updated
September 2008] The Cochrane Collaboration. www.cochrane-
handbook.org
17. Centre TNC, Review Manager (RevMan). 2008 The Cochrane
Collaboration: Copenhagen
18. Oxford Centre of Evidence Based Medcine (2009) Levels of
evidence. Accessed April 2009
19. Sileri P, Gentileschi P, Sica GS, Rossi P, Franceschilli L, PerroneF, Gaspari AL (2008) Abstract: laparoscopic appendectomy for
complicated appendicitis. Society for Surgery of the Alimentary
Tract 2008 Abstracts T1715:A883
20. Cueto J, DAllemagne B, Vazquez-Frias JA, Gomez S, Delgado
F, Trullenque L, Fajardo R, Valencia S, Poggi L, Balli J, Diaz J,
Gonzalez R, Mansur JH, Franklin ME (2006) Morbidity of lap-
aroscopic surgery for complicated appendicitis: an international
study. Surg Endosc 20:717720
21. Kiriakopoulos A, Tsakayiannis D, Linos D (2006) Laparoscopic
management of complicated appendicitis. J Soc Lap Surg
10:453456
World J Surg
123
http://dx.doi.org/10.1002/14651858.CD001546.pub2http://www.cochrane-handbook.org/http://www.cochrane-handbook.org/http://www.cochrane-handbook.org/http://www.cochrane-handbook.org/http://dx.doi.org/10.1002/14651858.CD001546.pub2 -
7/30/2019 Laparoscopic Versus Open Appendectomy in Adults With Complicated Appendicitis- Systematic Review and Meta-A
15/15
22. Towfigh S, Chen F, Mason R, Katkhouda N, Chan L, Berne T
(2006) Laparoscopic appendectomy significantly reduces length
of stay for perforated appendicitis. Surg Endosc 20:495499
23. Mancini GJ, Mancini ML, Nelson HS (2005) Efficacy of lapa-
roscopic appendectomy in appendicitis with peritonitis. Am Surg
71:14
24. Ball CG, Kortbeek JB, Kirkpatrick AW, Mitchell P (2004)
Laparoscopic appendectomy for complicated appendicitis: an
evaluation of post-operative factors. Surg Endosc 18:969973
25. Johnston AB, Peetz ME (1998) Laparoscopic appendectomy is an
acceptable alternative for the treatment of perforated appendici-
tis. Surg Endosc 12:940943
26. Lachin JM, Matts JP, Wei LJ (1988) Randomisation in clinical
trials: conclusions and recommendations. Control Clin Trial
9:365374
27. Schulz KF, Grimes DA (2002) Blinding in randomised trials:
hiding who got what. Lancet 359:696700
28. Schulz KF, Grimes DA (2002) Allocation concealment in ran-
domised trials: defending against deciphering. Lancet 359:614
618
29. Pokala N, Sadhasivam S, Kiran RP, Parithivel V (2007) Com-
plicated appendicitisis the laparoscopic approach appropriate?
A comparative study with the open approach: Outcome in a
community hospital setting. Am Surg 73:737741
30. Yau KK, Siu WT, Tang CN, Yang GPC, Li MKW (2007) Lap-
aroscopic versus open appendectomy for complicated appendi-
citis. J Am Coll Surg 205:6065
31. Katsuno G, Nagakari K, Yoshikawa S, Sugiyama K, Fukunaga M
(2009) Laparoscopic appendectomy for complicated appendicitis:
a comparison with open appendectomy. World J Surg 33:208214
32. Piksun G, Kozic D, Rajpal S, Shaftan G, Fogler R (2001)
Comparison of laparoscopic, open and converted appendectomy
for perforated appendicitis. Surg Endosc 15:660662
33. Stoltzing H, Thon K (2000) Perforated appendicitis: is laparo-
scopic operation adviseable? Digest Surg 17:610616
34. Wullstein C, Barkhausen S, Gross E (2001) Results of laparo-
scopic vs conventional appendectomy in complicated appendi-
citis. Dis Colon Rectum 44:17001705
35. Guller U, Hervey S, Purves H, Muhlbaier LH, Peterson ED,
Eubanks E, Pietrobon P (2004) Laparoscopic versus open
appendectomy. Outcomes comparison based on a large admin-
istrative database. Ann Surg 239:4352
36. So JBY, Chiong E-C, Chiong E, Cheah W-K, Lomanto D, Goh P,
Kum C-K (2002) Laparoscopic appendectomy for perforated
appendicitis. World J Surg 26:14851488
37. Lin H-F, Wu J-M, Tseng L-M, Chen K-H, Huang S-H, Lai I-R
(2006) Laparoscopic versus open appendectomy for perforated
appendicitis. J Gastrointest Surg 10:906910
38. Brewer M (2000) Research design and issues of validity. Cam-
bridge University Press, Cambridge
39. Henry MC, Moss RL (2005) Primary versus delayed wound
closure in complicated appendicitis: an international systematic
review and meta-analysis. Ped Surg Intern 21:625630
40. Greenwood PE, Nikulin MS (1996) A Guide to Chi-Squared
Testing. Wiley, New York
41. Fisher RA (1922) On the interpretation of v2 from contingency
tables, and the calculation of P. J R Stat Soc 85:8794
42. Lachin JM (2000) Statistical considerations in the intent-to-treat
principle. Control Clin Trial 21:167189
43. Cohen J (1988) Statistical power analysis for the behavioural
sciences, 2nd edn. Lawrence Erlbaum Associates, Mahwah
44. Khalili TM, Hiatt JR, Savar A, Lau C, Margulies DR (1999)
Perforated appendicitis is not a contraindication to laparoscopy.
Am Surg 65:965967
World J Surg
123