Systematic review and meta-analysis of prophylactic mesh placement for prevention of incisional hernia following midline laparotomy

REVIEW

Systematic review and meta-analysis of prophylactic meshplacement for prevention of incisional hernia following midlinelaparotomy

A. Bhangu • J. E. Fitzgerald • P. Singh •

N. Battersby • P. Marriott • T. Pinkney

Received: 10 February 2013 / Accepted: 18 May 2013 / Published online: 28 May 2013

� Springer-Verlag France 2013

Abstract

Purpose Incisional hernia is a common long-term com-

plication after laparotomy. This study investigated whether

prophylactic mesh reinforcement of laparotomy reduced

the rate of incisional hernia, with emphasis on trial design

and quality.

Methods A systematic review of published literature was

performed for studies comparing incisional hernia presence

following conventional closure or prophylactic mesh rein-

forcement. Studies were assessed using the Cochrane Risk

of Bias Tool, the Jadad score and the Newcastle Ottawa

Scale (NOS). The primary endpoint was incisional hernia,

assessed by meta-analysis.

Results Seven studies [four randomised controlled trials

(RCTs) and three prospective trials] included 588 patients;

262 received mesh reinforcement. All studies included

elective patients at high risk of incisional hernia. Six

incorporated a polypropylene mesh and one a biologic

mesh. Four studies were judged high quality by NOS and

two of four RCTs were at low risk of bias, although overall

outcome assessment from all studies was either poor or

mediocre. Mesh significantly reduced the rate of incisional

hernia [odds ratio (OR) 0.15, p \ 0.001]; the same effect

was seen in RCTs only (OR 0.17, p \ 0.001). A borderline

increase of seroma seen with a fixed effect model (OR

1.82, p = 0.050) was not seen with a random effect model

(OR 1.86, p = 0.210, I2 = 45 %).

Conclusion Mesh reinforcement of laparotomy signifi-

cantly reduced the rate of incisional hernia in high-risk

patients. However, poor assessment of secondary outcomes

limits applicability; routine placement in all patients cannot

yet be recommended. More evidence regarding the rates of

adverse events, cost-benefits and quality of life are needed.

Keywords Hernia � Prophylaxis � Mesh � Incisional

hernia

Introduction

Incisional hernia following laparotomy is one of the most

frequent long-term complications, affecting 20 % of

unselected patients and up to 50 % of high-risk (e.g. obese)

patients [1, 2]. This can lead to significant morbidity

including pain, deformity, emergency re-admission and re-

operation, with a resultant financial burden [1]. Incisional

hernia repair is strongly associated with recurrence, leading

to further morbidity and patient dissatisfaction, together

with the associated healthcare and economic costs [3].

Comorbidities that increase the likelihood of incisional

hernia following laparotomy include diabetes, pulmonary

disease and connective tissue disorders, such as aneurysmal

disease [4]. Amongst these higher risk individuals, incidence

of post-operative hernia can reach 50 % [5]. The INLINE

study meta-analysed the optimal method for sutured closure

On behalf of the West Midlands Research Collaborative.

Committee of the West Midlands Research Collaborative—David

Bartlett, Kaori Futaba, Pritesh Mistry, Caroline Richardson, Elizabeth

Hepburn, Abhilasha Patel, Paul Marriott, Andrew Torrance, Dion

Morton.

Electronic supplementary material The online version of thisarticle (doi:10.1007/s10029-013-1119-2) contains supplementarymaterial, which is available to authorized users.

A. Bhangu (&) � J. E. Fitzgerald � P. Singh � N. Battersby �P. Marriott � T. Pinkney

Academic Department of Surgery, Room 29, 4th Floor, Queen

Elizabeth Hospital, Edgbaston, Birmingham B15 2TH, UK

e-mail: [email protected]

123

Hernia (2013) 17:445–455

DOI 10.1007/s10029-013-1119-2

of elective midline laparotomy, and found convincing evi-

dence that a continuous, slowly absorbable suture provided

the optimum risk reduction of incisional hernia [2]. Addi-

tional measures to further reduce this risk, even following

optimum sutured closure, require investigation.

Prophylactic mesh reinforcement of midline wounds has

the potential to be an effective intervention in reducing the

risk of incisional hernia formation. Due to the possibility of

adverse effects and increased costs, several questions

remain. Firstly, specific patient groups who will benefit are

yet to be established. Secondly, the adverse event profile

requires definition, particularly regarding the risk of ser-

oma, haematoma and surgical site infection following mesh

implantation. Finally, the cost-effectiveness requires anal-

ysis, including changes in the duration of the index surgical

procedure. This systematic review aimed to determine

whether prophylactic mesh reinforcement of laparotomy

wounds reduced the rate of incisional hernia, with partic-

ular emphasis on the methodological quality of included

studies.

Methods

Data sources and search strategy

A pre-specified study protocol for this systematic review

was designed and registered with the PROSPERO database

(International Prospective Register of Systematic Reviews,

www.crd.york.ac.uk/prospero, record number: CRD4201

2003179). A systematic search of the OVID SP version of

Medline, Pubmed version of Medline, Embase, the Coch-

rane Library, ClinicalTrials.gov and Google Scholar was

performed for published studies comparing mesh closure

versus conventional closure following laparotomy for any

reason. To assess contemporary evidence, only studies

published after 1980 were included and no language

restrictions were applied. The search was performed

independently by two researchers. The search strategies

used are presented in supplemental table 1. MeSH terms

were used to search Medline and Embase. A manual search

of reference lists in relevant systematic reviews was

undertaken to further identify trials of potential interest.

Abstracts and conference proceedings were excluded

because of the high probability of incomplete data. Cita-

tions were collated with EndNote Reference Manager

(Version X4, Thomson Reuters) and duplicates removed.

The last search was performed in October 2012.

Inclusion and exclusion criteria

To be included, studies had to satisfy the following pre-

determined criteria: (1) include midline laparotomy only;

(2) report post-operative incisional hernia rate (as either

primary or secondary endpoint); (3) design was a ran-

domised controlled trial, prospective observational or ret-

rospective cohort study; (4) published after 1980; (5)

reporting ten or more patients.

Studies were excluded if they: (1) were designed as case

reports, letters, or with \10 patients; (2) considered par-

astomal hernia prevention; (3) included management of the

open abdomen (i.e. in relation to a planned fascial defect).

Data extraction

Two authors extracted data independently. Discrepancies

in outcome extraction were resolved by re-examination and

discussion of the relevant study to achieve consensus. Data

extracted on study design included: randomisation tech-

nique, intervention arms, method of linea alba closure, type

of mesh, definition of incisional hernia. Details relating to

patients included: number, age, gender, operation indica-

tion, presence of diabetes, cardiovascular disease, con-

nective tissue disease, obesity, smoking, mean length of

incision and previous abdominal surgery/laparotomy. The

primary outcome assessed for meta-analysis was the rate of

incisional hernia. Secondary outcomes recorded were rates

of adverse events.

Assessment of bias and study quality

Risk of bias and study quality was assessed in several

ways, in order to allow a multi-domain comparison and to

allow comparison between randomised and non-random-

ised studies. Two reviewers independently allocated study

quality and disagreement was resolved by re-examining the

relevant paper until consensus was achieved.

Firstly, all studies were evaluated using the Newcastle–

Ottawa Scale (NOS) [6]. This assessment examined three

factors: method of patient selection, comparability of the

study groups and number of outcomes reported. Although

initially intended to quality assess non-randomised studies,

the fields were adapted to suit the study population and

outcome measures in the present meta-analysis, and was

applied to all included studies irrespective of design. A score

of seven and above was used to indicate high quality [7].

Secondly, the Cochrane Risk of Bias Tool was used for

randomised trials [8]. The risk of bias tool covers six

domains of bias: selection bias, performance bias, detection

bias, attrition bias, reporting bias, and other bias. Each

domain was scored as a high or low risk of bias and an

overall high and low risk status was assigned. Studies with

inadequate randomisation, unblinded outcome assessment,

and high ([10 %) or unexplained losses to follow-up were

considered to be factors for high risk of bias. Thirdly, the

Jadad scale was used to score RCTs in the domains of

446 Hernia (2013) 17:445–455

123

randomisation, blinding and losses to follow-up using the

five available points (supplemental table 2) [9]. Fourthly,

details on measures used to determine and record outcomes

were collated. Finally, a pre-specified assessment of pub-

lication bias was performed by means of a funnel plot.

Statistical analysis

Meta-analysis was conducted according to guidelines from

the Preferred Reporting Items for Systematic reviews and

Meta-Analysis group (PRISMA) [10]. The odds ratio (OR)

was used as the statistical measure for dichotomous out-

comes and the weighted mean difference (WMD) for

continuous variables. ORs were calculated from the origi-

nal data and meta-analysed using the Mantel–Haenszel

method. An OR of [1.0 indicated greater risk of an

adverse event occurring in the experimental group. A

p value of \0.05 was considered significant for all analy-

ses. Statistical algorithms were used to calculate the stan-

dard deviation if unavailable.

Between-study heterogeneity was assessed using the I2

and v2 statistic, and funnel plots. Higher values of I2 and

the v2 statistic signified increasing levels of heterogeneity,

with a p value \0.05 or an I2 [ 50 % indicating significant

heterogeneity [11]. To take into account the effects of this

heterogeneity, both fixed and random effect models are

presented [12]. Statistical analysis was performed using

Review Manager 5.0 (Copenhagen: The Nordic Cochrane

Centre, The Cochrane Collaboration, 2008).

Sensitivity analysis

Pre-planned sensitivity analysis was planned for the fol-

lowing groups, if two or more studies were available: RCTs

only; RCTs at low risk of bias; studies with NOS score of

seven and above; abdominal aortic aneurysm (AAA) repair

patients only; obesity-related surgery only.

Results

The final analysis included seven studies (Fig. 1), of which

four were RCTs [13–16] and three were prospective trials

[4, 17, 18]. These studies included 588 randomised

References identified after initial search and from reference lists

n=1526

Excluded by abstract/ title n=900

Duplicates records excluded n=581

Excluded n=38: No mesh prophylaxis: 14 Letter/review: 9 Emergent/ planned delayed

closure: 8 Non-comparative mesh

prophylaxis: 4 Non-human: 2 Overlap: 1

Full text reviewed n=45

Included in final review n=7

Abstracts reviewed n=945

Fig. 1 PRISMA flowchart of

included studies

Hernia (2013) 17:445–455 447

123

patients, of whom 262 (44.6 %) received mesh reinforce-

ment of the incision. All studies included adults, with

48.1 % being male (age and BMI distributions are shown

in supplemental table 3).

Inclusion criteria

All studies included patients who were considered to be at

high risk of incisional hernia, including those with

abdominal aortic aneurysms [13], those undergoing open

bariatric procedures for obesity [4, 16–18] and cohorts of

patients with mixed-risk factors [14, 15]. Specific inclusion

criteria are shown in Table 1 and co-morbidities affecting

included patients are shown in supplemental table 4.

Surgical details/mesh details

Six of seven studies considered a polypropylene mesh, with

only one prospective study considering a biologic mesh

(Table 2). Three studies stipulated placement between the

rectus muscle and posterior rectus sheath, two above the

anterior rectus sheath (on-lay method) and two deep to the

posterior rectus sheath (sub-lay method). All meshes were

fixed into position with sutures. For the conventional arm,

six studies stipulated a continuous mass closure and one

interrupted sutures of linea alba. Incision types are shown

in supplemental table 5.

Risk of bias

Two of four RCTs provided adequate detail of randomi-

sation, including specifying pre-operative randomisation

[14, 16]. Only one study stipulated a blinded assessor for

post-operative visits [16]. In four studies, the nature of the

person performing clinical assessment was unstated. One

randomised study [15] and one prospective [4] study were

at high risk of bias due to high or unexplained losses to

follow-up.

When scored using the Newcastle–Ottawa Scale, four of

the seven studies achieved seven or more stars and one

achieved nine stars (Table 1 and supplemental table 6).

When assessing the four RCTs with the Cochrane Risk of

Bias Tool, two were considered to be at high risk and two

at low risk of bias (supplemental table 7). RCTs scored

one, two, three and five points on the Jadad scale (sup-

plemental table 8); those scoring three and five were also

Table 1 Cohort demographics

Study Years Setting Design Operation High-risk inclusion

criteria?

If non-RCT,

selection for mesh

NOS

score

Risk

of

biasAAA Bariatric Unclear

Bevis 2010 3 Southwest

hospitals,

UK

RCT 85 0 0 Patients undergoing

abdominal aneurysm

repair

7 Low

Curro 2011 Messina,

Italy

PNR 0 95 0 Morbidly obese undergoing

biliopancreatic diversion

Two consecutive

time periods

7 NA

El-Khadrawy 2009 Tanta, Egypt RCT 0 0 40 C1 risk factors that may

affect the healing process

7 High

de la Pena 2003 Jerez, Spain RCT 0 0 100 High risk of incisional

hernia or one of:

neoplastic pathology,

age [70, respiratory

failure, malnutrition,

BMI [30, smokes [20

cigarettes daily

6 High

Llaguna 2011 New York,

USA

PNR 0 134 0 Patients undergoing Roux-

en-Y gastric bypass

One surgeon

operating at 2

hospital with

different policies

5 NA

Strzeczyk

RCT

2006 Lodz,

Poland

RCT 0 74 0 Morbid obesity undergoing

Roux-en-Y gastric bypass

9 Low

Strzeczyk

non-RCT

2002 Lodz,

Poland

PNR 0 60 0 Patients undergoing Roux-

en-Y gastric bypass

Patients considered

to be at high risk

of hernia

(morbidly obese

with clinically

significant

morbidity)

6 NA

AAA Abdominal aortic aneurysm, RCT randomised controlled trial, PNR prospective non-randomised, NA non-applicable

448 Hernia (2013) 17:445–455

123

judged to be at low risk of bias using the Cochrane tool.

Funnel plot analysis for all studies and the four RCTs for

the primary outcome revealed low risk of publication bias

(supplemental Fig. 1).

Outcome assessment

Details of outcome assessments are shown in Table 3. Two

studies provided guidance for clinical examination tech-

nique [4, 15]. Six of seven studies allowed for radiological

examination to augment clinical examination, with ultra-

sound as routine in three studies [14, 16, 17] and selective

ultrasound or computed tomography for equivocal cases in

three others [4, 13, 15]. No studies provided a definition of

wound infection.

Planned time points for assessment varied considerably

between studies, with most including assessment at 6 and

12 months. Six studies achieved a minimum mean follow-

up of 1-year. Whilst some studies stipulated a minimum

follow-up time (at 6 months, 1 and 3 years), and one study

allowed a minimum follow-up of 35 days.

Outcomes

Assessment of incisional hernia as the primary outcome

was available in 543 patients from seven studies. Patients

lost to follow-up (n = 48) included 33 unexplained losses

or short follow-up periods, six perioperative deaths, six re-

operations due to reasons other than incisional hernia, three

withdrawals of consent following allocation but prior to

receiving treatment (details for individual studies shown in

supplemental table 3).

The results of fixed and random effect meta-analysis are

shown in Tables 4 and 5, and in supplemental Fig. 2. The

rate of incisional hernia was significantly reduced by

placement of mesh (3.9 %, 9/238) versus no mesh (22.0 %,

67/305, Table 4; Fig. 2). In all studies, mesh placement led

to a slightly but significantly longer operation by a mean of

14.9 min and a significantly shorter length of stay by

1 day. A fixed effect model showed a borderline increase in

the rate of seroma or wound leakage (six studies, 12.9 %

with mesh, 26/201 versus 6.9 % without mesh, 18/262, OR

1.82, p = 0.050, Fig 3), although this effect was non-sig-

nificant with a random effect model (OR 1.86, p = 0.210,

Table 4).

There were also no differences in rates of surgical site

infection (SSI, five studies, 5.3 % with mesh [10/190]

versus 5.3 % without mesh [11/209]) or haematoma (two

studies, 3.4 % with mesh [3/89] versus 3.2 % without mesh

[3/94]). There was a trend towards increased chronic pain

with mesh (two studies, 7.8 % with mesh [5/64] versus 0 %

without mesh [0/64]), although the confidence intervals for

the OR from both fixed and random effect models were

wide (Table 4). Mesh led to a significantly reduced risk of

incisional hernia repair compared to conventional closure

(fixed effect OR 0.14, p = 0.002, Table 4). No data were

found for comparative cost-effectiveness or quality of life

assessment. Single studies presented comparative data on

length of incision and wound dehiscence which are thus not

included in this meta-analysis.

Table 2 Surgical details of comparative arms

Arm 1 Arm 2

Study Mesh type Mesh position Mesh fixation n Standard closure n

Bevis Half 15 9 15 cm

polypropylene mesh

Between rectus muscle and

posterior rectus sheath or

posterior peritoneum

Four tacking sutures

to posterior rectus

sheath

40 Mass closure with surgeon’s

preferred non-absorbable suture

45

Curro Polypropylene mesh,

10 cm wider and 8 cm

longer than the defect

Between rectus muscle and

posterior sheath

10–12 sutures 45 Peritoneum with a running 0

Vicryl suture, linea alba with

interrupted 1 Vicryl suture

50

El-Khadrawy Polypropylene mesh 2 cm

each side of linea alba

Between peritoneum and

posterior rectus sheath

4 sutures 20 1–0 polypropylene mass closure

of linea alba

20

de la Pena Polypropylene mesh on

the aponeurosis

3 cm past the incision onto

the aponeurotic surface

Sutures 1 cm apart 50 Continuous non-absorbable

monofilament

50

Llaguna 16 cm 9 6 cm biologic

mesh (AlloDerm)

Below fascia Several interrupted

polydioxanone

sulphate sutures

59 1 PDS continuous suture 75

Strzeczyk

RCT

Polypropylene mesh, 8 cm

wide and 2 cm wider

than defect

Between rectus muscle and

posterior sheath

Interrupted sutures 36 2 PDS mass closure 38

Strzeczyk

non-RCT

Polypropylene mesh 8 cm

wide overlapping length

by 2 cm

On-lay to anterior rectus

sheath

Continuous

polypropylene

suture

12 1 PDS continuous suture 48

RCT Randomised controlled trial

Hernia (2013) 17:445–455 449

123

Sensitivity analysis

Considering only the four RCTs, the OR and heterogeneity

for reduction of incisional hernia were similar to the pooled

results from all studies (Table 5). Seroma and surgical site

infection remained non-significant between groups,

although heterogeneity was greatly reduced (0 and 0 %

respectively). Only incisional hernia assessment was meta-

analysable when considering the two RCTs at low risk of

bias. The effect size was similar (154 randomised patients,

OR 0.18, 95 % CI 0.07, 0.50, p = 0.001) but the hetero-

geneity slightly increased (18 %, p = 0.270).

Four studies scored seven or more NOS points. These

showed a significant reduction of incisional hernia with

mesh placement (OR 0.16, 95 % CI 0.07, 0.36, p \ 0.001)

with low heterogeneity (0 %); there were no significant

differences for any of the secondary outcome measures.

Considering only patients undergoing open bariatric pro-

cedures, the effect size for incisional hernia reduction was

slightly improved (OR 0.10, 95 % CI 0.03, 0.29,

p \ 0.001) and heterogeneity remained low (0 %). Effect

sizes and heterogeneity for seroma, SSI and length of stay

were similar to the overall findings. Sub-group analysis of

those undergoing only AAA repair was not possible.

Table 3 Outcome assessment

Study Follow-up Hernia assessment Radiological

examination

Time point (s) Definition

of wound

infectionMesh Non-mesh Who

performed

clinical

examination

Examination

technique

provided?

Bevis mean 939

(134–1,510) days

Mean 609 (range

35–1,333) days

Unblinded

surgeon

None Ultrasound

when doubt

over hernia

1 month,

6 months,

1/2/3 years

Not stated

Curro 25 reached 2 years;

20 reached 1 year

Minimum 2 years Surgeon,

blinding

unclear

None Ultrasound at

day 15, 6/12/

24 months;

CT

augmentation

in mesh group

3, 6, 12,

24 months

Not stated

El-Khadrawy Mean 37.2 (SD 10.0)

months

Mean 36.3 (SD 10.8)

months

Unclear Unclear Ultrasound in

all on day 10

to look for

seroma or

haematoma

2 weeks,

monthly for

6 months,

then 3

monthly

Not stated

de la Pena Minimum 3 years Minimum 3 years Unclear Abdominal

palpation to

detect possible

existence of

incisional

hernia

Computed

tomography

for

inconclusive

examination

3 years Not stated

Llaguna Mean

16.6 ± 7.1 months

Mean

17.7 ± 9.4 months

Unclear Palpable fascial

defects or

visible

protrusions at

or near the

surgical

incision at rest

or with

valsalva

Imaging only

when

clinically

warranted

1 week,

monthly for

3 months, 6

monthly for

2 years

Not stated

Strzeczyk

RCT

Minimum 6 months Minimum 6 months Blinded

registrar

None Ultrasound in

all patients

Reviewed for

at least

6 months

mean 28

(range

6-38 months)

Not stated

Strzeczyk

non-RCT

Minimum 1 year Minimum 1 year Unclear None Unstated 12–14 months Not stated

SD Standard deviation, RCT randomised controlled trial

450 Hernia (2013) 17:445–455

123

Discussion

The main finding of this systematic review is the evidence

for reduced occurrence of incisional hernia in high-risk

patients following prophylactic mesh implantation. How-

ever, when making recommendations for widespread use,

this has to be balanced against the mediocre study design,

inconsistent outcome assessment and uncertainty of true

secondary outcome rates.

Assessing trial quality, risk of bias and outcome

reporting is controversial, can be difficult and is often

inadequately performed [19, 20]. This meta-analysis used

three methods of assessing studies. Four studies scored

seven or above NOS points and two of the included RCTs

were deemed to be at low risk of bias. The assessment

systems did not have perfect agreement, although only one

RCT was deemed to be at low risk of bias and scored

highly on the NOS and Jadad scales. Whilst general het-

erogeneity was low and performing sub-group analyses

compensated for the quality differences, the overall

assessment detected the mixed nature of quality and thus

overall high risk of bias; this was an advantage of assessing

quality using several metrics rather than single tools. The

RCTs were not without limitations, particularly relating to

randomisation reporting and blinding of assessors. Their

generalisability is further limited by their application to

specific populations (e.g. AAA repair patients) in specific

geographical locations. More evidence to support the pri-

mary outcome in a wider range of clinical settings is still

required.

Assessment of secondary outcomes related to adverse

events showed no significant differences between arms, but

studies were inadequate to completely assess these. How-

ever, a trend towards increased seroma formation with

mesh carried borderline significance with a fixed effect

model, but non-significance with a random effect model,

which reflects clinical heterogeneity. Since pooled assess-

ment of seroma formation from the high-quality RCTs was

not possible due to a lack of data, more high-quality evi-

dence to definitively identify these rates is thus required.

Two RCTs which assessed chronic pain found a possible

trend towards an increase with mesh (OR 6.63, 95 % CI

0.77, 56.86, p = 0.080); however, no definition of chronic

pain was given and the assessment methods were unclear

with neither study using validated measurement tools [14,

15]. A definition of surgical site infection was not provided

by any of the included studies. When assessed prospec-

tively, with an accepted definition and the intention of

identifying all events, the risk of SSI has recently been

proven to be higher than expected in elective colorectal

surgery and can approach 25–45 % [21]. Without active,

independent and standardised assessment for adverse

events, it may be that the rates seen within the included

studies of the present meta-analysis are under-estimated.

Table 4 Meta-analysis of outcome measures from all studies

Outcome Studies

(patients)

Fixed effect p Random effect p Heterogeneity p

OR or WMD OR or WMD I2 (%) v2

Incisional hernia 7 (543) 0.15 (0.07, 0.29) \0.001 0.16 (0.08, 0.33) \0.001 0 2.3 0.89

Seroma or wound leakage 6 (463) 1.82 (0.99, 3.37) 0.050 1.86 (0.70, 4.97) 0.210 45 9.1 0.11

Surgical site infection 5 (409) 1.04 (0.43, 2.49) 0.940 0.97 (0.36, 2.59) 0.950 2 4.1 0.39

Haematoma 2 (183) 1.03 (0.23, 4.72) 0.970 1.08 (0.22, 5.33) 0.930 0 0.6 0.44

Duration of surgery 2 (180) 14.9 (12.1, 17.7) \0.001 14.9 (12.1, 17.7) \0.001 0 0.3 0.6

Length of stay 2 (169) -1.02 (-1.37, -0.67) \0.001 -1.02 (-1.37, -0.67) \0.001 0 0.7 0.42

Chronic wound pain 2 (128) 6.63 (0.77, 56.86) 0.080 6.57 (0.76, 56.69) 0.090 0 0.04 0.84

Surgical repair of hernia (all patients) 2 (175) 0.14 (0.04, 0.50) 0.002 0.15 (0.04, 0.52) 0.003 0 0.44 0.51

OR odds ratio, WMD weighted Mean Difference

Table 5 Meta-analysis of outcome measures from randomised controlled trials

Outcome Studies (patients) Fixed Effect p Random Effect p Heterogeneity p

OR or WMD OR or WMD I2 v2

Incisional hernia 4 (282) 0.17 (0.07, 0.42) \0.001 0.20 (0.08, 0.50) \0.001 0 1.7 0.63

Seroma or wound leakage 3 (202) 1.03 (0.40, 2.62) 0.950 1.07 (0.41, 2.82) 0.890 0 1.3 0.52

Surgical site infection 3 (208) 0.73 (0.22, 2.42) 0.610 0.74 (0.22, 2.50) 0.630 0 0.54 0.76

Chronic wound pain 2 (128) 6.63 (0.77, 56.86) 0.080 6.57 (0.76, 56.69) 0.090 0 0.04 0.84

OR Odds ratio, WMD weighted mean difference

Hernia (2013) 17:445–455 451

123

All studies in this meta-analysis included only patients

with risk factors for incisional hernia, including connective

tissue disorders, obesity and co-morbidities. This study also

found that mesh placement reduced the risk of future sur-

gery for incisional hernia, and so it may be a well-suited

prophylactic measure for these high-risk patients. There is

inadequate evidence to assess the risks and benefits of

routine use of mesh in un-selected, mixed-risk patients.

Cost-effectiveness data were lacking from the included

studies, which should be addressed in future trials. None of

the included studies assessed quality of life as an endpoint,

which is of increasing concern to both patients and phy-

sicians. Effective assessment of quality of life using vali-

dated tools has been proven feasible from large scale

inguinal hernia trials, and should be applied to future mesh

prophylaxis trials [22].

The risks associated with prophylactic mesh implanta-

tion are potentially serious, including re-operation for

explantation and the potential for chronic pain, which will

both adversely affect quality of life. There was a lack of

data from the included studies regarding these morbidities,

and the outcome from unplanned mesh removal. Other

studies have reported on such morbidities. Herbert et al.

[23] described prosthetic implantation in 16 patients

undergoing Roux-en-Y gastric bypass by a single surgeon,

with the aim to reduce incisional hernia. During follow-up

(mean six-months), four patients required mesh explanta-

tion (three for infection and one for persistent seroma) and

one patient developed incisional hernia despite mesh. The

authors cautioned against the use of prophylactic mesh due

to the unacceptable rate of local complications. However,

three different types of mesh were used which may have

influenced results [Proceed (Ethicon), C-QUR (Atrium

Medical), Supramesh (Pall Corporation)]. This level of

complication was not seen in any included study from the

current meta-analysis, although multi-centre, prospective,

Fig. 2 Meta-analysis of data

considering the outcome of

incisional hernia from all

studies and subgroups. Fixed

effect models are shown when

I2 \ 50 %, and a random effect

when I2 [ 50 %

452 Hernia (2013) 17:445–455

123

independent assessment would be needed to more defini-

tively answer this question.

By analysing study design and risk of bias, this meta-

analysis has implications for future trial design and out-

come reporting. Based on the findings of this study, it is

feasible that future RCTs will find a significant risk

reduction of incisional hernia when mesh is placed. Thus,

careful attention should be paid to the adverse event and

quality of life profile. A future RCT should ideally be

multi-centred, adequately powered, contain a range of

patients at high risk of incisional hernia, and allow for

blinded patient and clinician outcome assessment. Such

outcome assessment should be optimised, to include dis-

crete definitions of outcome measures, a defined clinical

examination technique and supplementation with appro-

priate radiological imaging.

From the present meta-analysis, there was variation in

the assessment times in the included studies, which was

often short. Since only 50 % of incisional hernias become

clinically apparent after 5 months, 75 % after 2 years and

98 % after 5 years [24], correct timing of assessments is

important in hernia detection. As a result, it is feasible that

some of the included studies have not revealed true hernia

rates. However, delaying the findings of trials denies

patients and clinicians early evidence on which to base

treatment decisions. Imaging assessment of hernias carries

two benefits: (1) it allows for standardisation of hernia

detection (especially for computed tomography), and (2) it

may allow for early identification of late clinically relevant

hernias, and thus early delivery of trial results [25]. Finally,

thorough provision for assessment of quality of life

(including pain) and cost-effectiveness should be assessed

and ensured. Recommendations for design and outcome

assessment relevant to future trial design are shown in

Table 6.

A remaining question is whether the findings of this

study can influence the daily practice of the operating

surgeon. Our meta-analysis has limitations, which primar-

ily surround the quality of studies included. It is reasonable

that individually selected patients at high-risk warrant

prophylaxis with mesh. These patients should ideally be

included within prospective, randomised multicentre trials,

which will accurately assess the primary endpoint and

adverse event profile in a standardised way.

Currently, clinical trials registered concerning prophy-

lactic mesh placement include the PRIMAAT trial (pre-

vention of incisional hernia by mesh augmentation after

midline laparotomy for aortic aneurysm treatment, http://

clinicaltrials.gov/ct2/show/NCT00757133), which aims to

randomise 120 patients and complete by December 2013.

The ProphMesh trial (Prophylactic Mesh Implantation

for the Prevision of Incisional Hernia, http://clinicaltrials.

gov/ct2/show/NCT01203553) aims to randomise 150

patients and complete by July 2015. The Austrian Hernia

Fig. 3 Meta-analysis of data

considering the outcome of

seroma or wound leakage from

all studies and subgroups. Fixed

effect models are shown when

I2 \ 50 %, and a random effect

when I2 [ 50 %

Hernia (2013) 17:445–455 453

123

Study Group will randomise 300 patients with BMI [27

to the same effect, to complete by January 2016 (http://

clinicaltrials.gov/ct2/show/NCT01507870). In the meantime,

the present meta-analysis provides the most up-to-date

evidence, and perhaps more importantly can inform future

trial design.

Future studies may also provide information to allow for

optimisation of mesh type and position to reduce adverse

events. From studies included in the present meta-analysis,

only one assessed the use of biologic mesh and none

considered emergency patients. Emergency patients, who

often have complex and/or contaminated wounds, are at

high risk of incisional hernia and warrant inclusion in an

RCT. Conventional mesh is typically avoided due to the

perceived risk of infection. It may be that biologic mesh

reduces the risk of SSI in this group and may reduce

chronic pain in a wider group of patients, although

increased risk of seroma formation is a concern; further

evaluation is required [26]. A prospective, single arm study

of biologic mesh closure of 80 patients with contaminated

ventral incisional hernias has revealed infection related

events in 24 patients and seroma in 22, although no com-

parative arm was assessed [27]. The Reinforcement of

Closure of Stoma Site trial will randomise patients

undergoing stoma closure to either suture repair or suture

repair with biologic mesh reinforcement (http://www.rocss.

bham.ac.uk) These wounds act as a frequent and controlled

model for complex and contaminated wounds, where in-

cisional hernia may occur in up to 30 % of patients [25,

28]. Patients undergoing re-laparotomy (e.g. excision of

recurrent cancer or following earlier un-related abdominal

incision) may also be considered high risk and may benefit

from mesh prophylaxis. Temporary bridging of the open

abdomen is a different research question, with the question

of prophylaxis relating to those undergoing primary fascial

closure or at the time of delayed primary closure.

Conclusion

Mesh prophylaxis of midline laparotomy wounds reduced

the rate of incisional hernias in high-risk patients. How-

ever, the secondary outcome measures, including adverse

events, cost-effectiveness and quality of life, have not yet

been adequately assessed to recommend routine mesh

placement in all patients. Well-designed, multicentre RCTs

are required to assess the safety profile more completely.

The findings of this systematic review and meta-analysis

should inform their design, including the need for

Table 6 Recommended core

elements for future trial design

CT Computed tomography,

CDC Centre for Disease

Control, VAS Visual Analogue

Scale

Domain Recommendation Notes

Inclusion criteria Elective (high risk) Obese, AAA, re-laparotomy, other co-

morbidities

Emergency (high risk) Contaminated or dirty, other high risk

Elective (mixed risk) Benefits in low versus high risk

Methodology Mesh position defined Sub-lay, in-lay, on-lay etc.

Mesh type analysis Synthetic versus biologic

Optimum synthetic types

Primary outcome Incisional hernia Blinded outcome assessors

Definition of clinical hernia requires consensus

Standardised clinical examination technique

CT detection for early trial results

Secondary

outcomes

Surgical site infection Defined by CDC criteria

Seroma Definition and classification of size require

consensus

Haematoma Definition and classification of size require

consensus

Intra-operative details/length of

stay

Quality of life Consensus of optimal validated tool required

Cost-effectiveness

Pain Assessed using VAS and hernia specific tools

Blinded outcome assessors

Timing of

assessments

Pre-operative

Early post-operative 30 days

Mid-term 3 months

Long-term 1, 2, 3, 5 years

454 Hernia (2013) 17:445–455

123

improved definition of outcomes, blinded outcome assess-

ment, and assessment of quality of life and cost-effec-

tiveness. The effectiveness in unselected/mixed-risk and

emergency patients is unknown and further evidence is

required from these cohorts.

Conflicts of interest All authors declares no conflicts of interest.

References

1. van’t Riet M, Steyerberg EW, Nellensteyn J, Bonjer HJ, Jeekel J

(2002) Meta-analysis of techniques for closure of midline

abdominal incisions. Br J Surg 89:1350–1356

2. Diener MK, Voss S, Jensen K, Buchler MW, Seiler CM (2010)

Elective midline laparotomy closure: the INLINE systematic

review and meta-analysis. Ann Surg 251:843–856

3. Gecim IE, Kocak S, Ersoz S, Bumin C, Aribal D (1996) Recur-

rence after incisional hernia repair: results and risk factors. Surg

Today 26:607–609

4. Llaguna OH, Avgerinos DV, Nagda P, Elfant D, Leitman IM,

Goodman E (2011) Does prophylactic biologic mesh placement

protect against the development of incisional hernia in high-risk

patients? World J Surg 35:1651–1655

5. Capella RF, Iannace VA, Capella JF (2007) Reducing the inci-

dence of incisional hernias following open gastric bypass surgery.

Obes Surg 17:438–444

6. Wells G SB, O’Connell D et al (2000) The Newcastle–Ottawa

Scale (NOS) for assessing the quality of non-randomised studies

in meta-analyses. In: 3rd Symposium on systematic reviews:

beyond the basics. Improving quality and impact, Oxford, UK

7. Bhangu A, Nepogodiev D, Gupta A, Torrance A, Singh P (2012)

Systematic review and meta-analysis of outcomes following

emergency surgery for Clostridium difficile colitis. Br J Surg

99:1501–1513

8. Higgins JP, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman

AD, Savovic J, Schulz KF, Weeks L, Sterne JA (2011) The

Cochrane Collaboration’s tool for assessing risk of bias in ran-

domised trials. BMJ 343:d5928

9. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ,

Gavaghan DJ, McQuay HJ (1996) Assessing the quality of

reports of randomized clinical trials: is blinding necessary?

Control Clin Trials 17:1–12

10. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC,

Ioannidis JP, Clarke M, Devereaux PJ, Kleijnen J, Moher D

(2009) The PRISMA statement for reporting systematic reviews

and meta-analyses of studies that evaluate healthcare interven-

tions: explanation and elaboration. Brit Med J 339:b2700

11. Lau J, Ioannidis JP, Schmid CH (1997) Quantitative synthesis in

systematic reviews. Ann Intern Med 127:820–826

12. DerSimonian R, Laird N (1986) Meta-analysis in clinical trials.

Control Clin Trials 7:177–188

13. Bevis PM, Windhaber RA, Lear PA, Poskitt KR, Earnshaw JJ,

Mitchell DC (2010) Randomized clinical trial of mesh versus

sutured wound closure after open abdominal aortic aneurysm

surgery. Br J Surg 97:1497–1502

14. El-Khadrawy OH, Moussa G, Mansour O, Hashish MS (2009)

Prophylactic prosthetic reinforcement of midline abdominal

incisions in high-risk patients. Hernia 13:267–274

15. Gutierrez de la Pena C, Medina Achirica C, Dominguez-Adame

E, Medina Diez J (2003) Primary closure of laparotomies with

high risk of incisional hernia using prosthetic material: analysis

of usefulness. Hernia 7:134–136

16. Strzelczyk JM, Szymanski D, Nowicki ME, Wilczynski W,

Gaszynski T, Czupryniak L (2006) Randomized clinical trial of

postoperative hernia prophylaxis in open bariatric surgery. Br J

Surg 93:1347–1350

17. Curro G, Centorrino T, Low V, Sarra G, Navarra G (2012) Long-

term outcome with the prophylactic use of polypropylene mesh in

morbidly obese patients undergoing biliopancreatic diversion.

Obes Surg 22:279–282

18. Strzelczyk J, Czupryniak L, Loba J, Wasiak J (2002) The use of

polypropylene mesh in midline incision closure following gastric

by-pass surgery reduces the risk of postoperative hernia.

Langenbecks Arch Surg 387:294–297

19. Hartling L, Ospina M, Liang Y, Dryden DM, Hooton N, Krebs

Seida J, Klassen TP (2009) Risk of bias versus quality assessment

of randomised controlled trials: cross sectional study. Brit Med J

339:b4012

20. Deeks JJ, Dinnes J, D’Amico R, Sowden AJ, Sakarovitch C, Song

F, Petticrew M, Altman DG (2003) Evaluating non-randomised

intervention studies. Health Technol Assess 7:iii-x, 1–173

21. Anthony T, Murray BW, Sum-Ping JT, Lenkovsky F, Vornik VD,

Parker BJ, McFarlin JE, Hartless K, Huerta S (2011) Evaluating

an evidence-based bundle for preventing surgical site infection: a

randomized trial. Arch Surg 146:263–269

22. Eklund A, Montgomery A, Bergkvist L, Rudberg C (2010)

Chronic pain 5 years after randomized comparison of laparo-

scopic and Lichtenstein inguinal hernia repair. Br J Surg

97:600–608

23. Herbert GS, Tausch TJ, Carter PL (2009) Prophylactic mesh to

prevent incisional hernia: a note of caution. Am J Surg

197:595–598 (discussion 598)

24. Pollock AV, Evans M (1989) Early prediction of late incisional

hernias. Brit J Surg 76:953–954

25. Bhangu A, Fletcher L, Kingdon S, Smith E, Nepogodiev D,

Janjua U (2012) A clinical and radiological assessment of inci-

sional hernias following closure of temporary stomas. Surgeon

10:321–325

26. Lopez-Cano M, Armengol M, Quiles MT, Biel A, Velasco J,

Huguet P, Mestre A, Delgado LM, Gil FX, Arbos MA (2012)

Preventive midline laparotomy closure with a new bioabsorbable

mesh: an experimental study. J Surg Res 181:160–169

27. Itani KM, Rosen M, Vargo D, Awad SS, Denoto G 3rd, Butler CE

(2012) Prospective study of single-stage repair of contaminated

hernias using a biologic porcine tissue matrix: the RICH Study.

Surgery 152:498–505

28. Bhangu A, Nepogodiev D, Futaba K (2012) Systematic review

and meta-analysis of the incidence of incisional hernia at the site

of stoma closure. World J Surg 36:973–983

Hernia (2013) 17:445–455 455

123