The American Journal of Surgery (2014) 207, 980-988

Review

Meta-analysis of sublay versus onlay meshrepair in incisional hernia surgery

Lucas Timmermans, M.D.a,*, Barry de Goede, M.Sc.a,Sven M. van Dijk, B.Sc.a, Gert-Jan Kleinrensink, Ph.D.b,Johannes Jeekel, M.D., Ph.D.b, Johan F. Lange, M.D., Ph.D.a

aDepartment of Surgery, bDepartment of Neuroscience, Erasm

us MC, University Medical Center, Z-835,3000 CA Rotterdam, The Netherlands

KEYWORDS:Incisional hernia;Onlay;Sublay;Recurrence;Surgical site infection;Complication

The authors declare no conflict of in

* Corresponding author. Tel.:131-10

E-mail address: lucastimmermans@g

Manuscript received March 25, 2013; r

0002-9610/$ - see front matter � 2014

http://dx.doi.org/10.1016/j.amjsurg.20

AbstractBACKGROUND: Incisional hernia (IH) remains a very frequent postoperative complication. The 2

techniques most frequently used are the onlay repair and sublay repair. However, it remains unclearwhich technique is superior.

DATA SOURCES: A meta-analysis was conducted according to Preferred Reporting Items for Sys-tematic Reviews and Meta-Analyses guidelines. The quality of the nonrandomized studies was assessedusing the Newcastle-Ottawa Scale.

RESULTS: Of 178 articles, 10 articles (2 randomized controlled trials, 1 prospective study, and 7 ret-rospective studies) comprising a total of 1,948 patients (775 onlay operations and 1173 sublay opera-tions) were selected. Two of the studies scored below 5 points on the Newcastle-Ottawa Scale and werenot selected. A trend was observed for IH recurrence in favor of sublay repair (odds ratio 5 2.41; 95%confidence interval, .99 to 5.88; I2 5 70%; P 5 .05). Surgical site infection occurred significantly lessafter sublay repair (odds ratio 5 2.42; 95% confidence interval, 1.02 to 5.74; I2 5 16%; P 5 .05). Nodifference was observed regarding seroma and hematoma.

CONCLUSIONS: Although the majority of the included studies were retrospective studies, sublay re-pair seems the preferred technique for IH repair.� 2014 Elsevier Inc. All rights reserved.

Incisional hernia (IH) remains one of the most frequentpostoperative complications after abdominal surgery, withincidences ranging from 11% to 20%.1,2 The incidence ofIH is even higher in patients with risk factors such as obe-sity and abdominal aortic aneurysms.3,4 Each year around200,000 IHs are treated in the United States. The treatmentof choice for IHs should be mesh repair.5,6 Mesh repair re-sults in lower recurrence rates compared with primary

terest.

-7034519; fax:131-10-2250647.

mail.com

evised manuscript June 25, 2013

Elsevier Inc. All rights reserved.

13.08.030

suture as was shown by Luijendijk et al5 (3-year cumulativerecurrence rate of 24% compared with 43%) and Burgeret al6 (10-year cumulative recurrence rate of 32% comparedwith 63%). In the studies performed by Luijendijk and Bur-ger, the Rives-Stoppa sublay repair (SR) technique wasused.7,8 With this technique, the mesh is placed on the pos-terior rectus fascia after dissection of the fascia from therectus muscle and approximation of the 2 edges of the fas-cia. Another frequently used technique for IH is the Chevrelor onlay repair (OR) technique.9 With this technique, themesh is placed on the anterior rectus fascia after dissectionof the fascia from the subcutis and approximation of the2 edges of the fascia. However, no consensus has been

L. Timmermans et al. Sublay versus onlay incisional hernia repair 981

reached as to which technique is preferable. The anatomicposition of the mesh placement has an impact on tissue in-corporation, tissue reaction, and tensile strength of the ab-dominal wall.10–12 These factors are important regardingIH recurrence and postoperative complications. A system-atic review of the literature and meta-analysis were per-formed in order to discover which of the 2 techniques hasbetter results with regard to IH recurrence, operationtime, and postoperative complications such as surgicalsite infection (SSI), seroma, hematoma, and fistula.

Methods

A systematic search of MEDLINE, Embase, Web ofScience, Scopus, PubMed Publisher, and the CochraneLibrary was performed. All aspects of the PreferredReporting Items for Systematic Reviews and Meta-Analysis statement were followed.13 Manual referencechecks (B.G.) of accepted articles in recent reviews and in-cluded articles were performed to supplement the elec-tronic searches. Details of the search syntax are listed inAppendix 1. Studies were evaluated for inclusion indepen-dently by 2 reviewers (L.T., B.G.) based on title and ab-stract and finally were evaluated independently based onthe full text. Studies were included if they met the follow-ing criteria: (1) participants: adult patients who underwentelective IH repair; (2) interventions: OR and SR as de-scribed by the European Hernia Society; (3) outcome mea-sures: IH recurrence; and (4) secondary outcome measures:SSI, seroma, hematoma, fistula, operation time, and a pe-riod of hospitalization. A random check was performedby the senior author (J.F.L.).14 Articles in which additionaldissections had been implemented were excluded to reduceheterogeneity by intervention because the current reviewwas strictly interested in comparing OR and SR and nota mixture of techniques. Any discrepancies in inclusionwere resolved by discussion between the reviewers andthe senior author (J.F.L.).

All required data from each study included wereextracted using a standardized form that covered studycharacteristics (study design, year of publication, and studyperiod), type of intervention (OR and SR), perioperativeinformation (operation time and period of hospitalization),and postoperative complications (IH recurrence, SSI, se-roma, hematoma, and fistula).

Assessment of study quality

The methodological quality of the included nonrandom-ized studies was assessed according to the NewcastleOttawa Scale (NOS) criteria.15

Data analysis

It was expected that the majority of articles would benonrandomized studies. Therefore, it was decided to

implement a quality assessment before including studiesfor meta-analysis. Nonrandomized studies were deemedeligible if a NOS score of 5 (out of 9) or higher wasestablished. Articles with a lower score were more at risk ofpresenting biased results, and, thus, were excluded from themeta-analysis. The results of these excluded studies will bepresented after each outcome section.

To pool data and calculate a pooled mean for eachpatient level outcome, a random effects model was used,which takes into account both the variance between studiesand the variance within a study.16 Odds ratios or mean dif-ferences with 95% confidence intervals were calculated toevaluate the statistical difference between outcomes afterOR and SR. Statistical heterogeneity was assessed for IHrecurrence, SSI, seroma, hematoma, and fistula by calculat-ing the Q statistic and the I2 statistic.

Selective dissemination of evidence was assessed byplotting each outcome measure of each study againstprecision (1/standard error) in a plot with P value contours.Funnel plot asymmetry, specifically with an apparent lackof studies in high P value areas of the plot, can be indica-tive of publication bias.17 Individual study effects on the re-sults were examined by removing studies 1 at a time todetermine whether removing a particular study wouldchange the significance of the pooled effect. Effects wereconsidered statistically significant if the 95% confidence in-tervals of the overall effect estimate did not overlap 1.Analyses were performed using Review Manager software(RevMan version 5.0.25; The Nordic Cochrane Centre, Co-penhagen, Denmark).

Results

Search and study characteristics

A total of 178 articles were identified after the removalof duplicates. Of these 178, 153 were excluded on the basisof title and abstract. After full-text assessment, 21 articleswere excluded because the article was an abstract or letterto the editor, it could not be obtained, it had been publishedin 2 different journals with similar results, or it did notfulfill the inclusion criterion that information needed to beavailable for distinguishing results for both techniques.

A total of 10 articles (2 randomized controlled trials[RCTs],18,19 1 prospective study,20 and 7 retrospective stud-ies21–27) compromising a total of 1,948 patients (775 onlayoperations and 1173 sublay operations) did meet the inclu-sion criteria. The Preferred Reporting Items for SystematicReviews and Meta-Analysis flow diagram for systematicreviews is presented in Fig. 1. The nonrandomized studieswere assessed for quality using the NOS criteria. Two stud-ies did not meet the criteria for inclusion in the meta-analysis because they scored lower than 5 points (out of9) (Table 1). In these 2 studies, a total of 265 patientswere included .The majority of patients (240/265) under-went SR.

Figure 1 Preferred Reporting Items for Systematic Reviews and Meta-Analysis study flow diagram.

982 The American Journal of Surgery, Vol 207, No 6, June 2014

Incisional hernia recurrence

Eight studies18,20–22,24,26–28 comprising a total of 1,359patients reported data regarding IH recurrence and were in-cluded into the meta-analysis (Figs. 2 and 3).20–22,24,26–29 Atrend in favor of SR was observed (odds ratio 5 2.41; 95%confidence interval [CI], .99 to 5.88; I2 5 70%; P 5 .05).During sensitivity analysis, results proved to be unstable,and heterogeneity remained high. However, after exclusionof the study by Weber,28 heterogeneity was reduced to 0,and results became statistically significant (odds ratio 53.35; 95% CI, 1.93 to 5.82; I2 5 0%; P , .001). Bridgingof the defect was not mentioned in any of the includedstudies.

The 2 studies not included in the meta-analysis alsoreported on IH recurrence.23,25 In 1 study, the recurrencerate for OR was 33% (3/9) compared with .48% (1/207)in the SR group. In the other study, the recurrence rate inthe OR group was 12.5% (2/16) compared with 3% (1/33) in the SR group.

Surgical site infection

Six studies comprising a total of 747 patients reporteddata regarding SSI and were included in the meta-analysis

(Fig. 4).18,20,22,24,26,27 SSI occurred significantly less in theSR group (odds ratio 5 2.42; 95% CI, 1.02 to 5.74; I2 516%; P 5 .05). During sensitivity analysis, the resultsproved to be unstable with acceptable heterogeneity scores.Little to no information was present regarding the intrao-perative degree of wound contamination and previousSSIs. In most studies, a standard regimen of antibiotic treat-ment during and after hernia repair was implemented(Table 1). One study not included in the meta-analysis re-ported data regarding SSI.23 The SSI rate in the OR groupwas 11.1% (1/9) compared with 3.4% (7/207) in the SRgroup.

Seroma

Six studies comprising a total of 715 patients reporteddata regarding seroma and were included in the meta-analysis (Fig. 5).18,20,22,24,26,27 No statistical significant re-sults were achieved (odds ratio 5 1.06; 95% CI, .38 to2.95; I2 5 48%; P 5 .89). During sensitivity analysis, theresults proved to be stable with fluctuating heterogeneityscores. One study not included in the meta-analysis re-ported data regarding seroma.23 The seroma rate in theOR group was 11.1% (1/9) compared with 1.9% (4/207)in the SR group.

Table 1 Study characteristics

Reference NOSStudyType Year Hernia size

No. ofORs

No. ofSRs Mesh Antibiotics Outcome measures Follow-up

Kumar et al20 7 Pros 2012 3 cm to 12 cm* 45 18 ? ? IH recurrence, SSI,seroma, postoperativepain

60 months†

Forte et al23 4 Retro 2011 ? 9 207 PP Yes (cephazoline) IH recurrence, SSI, seroma 12 months†

Abdollahi et al24 6 Retro 2010 ? 312 32 PP and PE Yes (cephazoline) IH recurrence, SSI,seroma, intestinalfistula,

98 months†

Venclauskas et al18 (2b‡) RCT 2010 11.5 cm (OR)†

–11 cm (SR)†57 50 PP comp Yes (oxacylinum) IH recurrence, SSI,

seroma, hematoma,ligature fistula,operative time,postoperative pain

12 months†

Weber andHorvath19

(2b‡) RCT 2010 .25 cm2† 235 224 PP Yes (cephazoline) IH recurrence, operativetime

60 months†

Coskun et al27 6 Retro 2009 d 22 23 PP ? IH recurrence, SSI,seroma, hematoma,fistula

54 months†

Gleysteen22 7 Retro 2009 10 cm† 75 50 ? Yes IH recurrence, SSI,seroma, hematoma,

64 months†

Israelsson et al21 5 Retro 2006 ? 281 228 PP Yes (75%) IH recurrence, operativetime

12 months†

de Vries Reilinghet al26

6 Retro 2004 ‘‘Large’’ PP Yes IH recurrence, SSI,seroma,enterocutaneous fistula,hematoma

30 months†

Kingsnorth et al25 3 Retro 2004 ‘‘Loss of domain’’ 16 33 PP Yes IH recurrence d

NOS 5 Newcastle-Ottowa Score, OR 5 onlay repair, SR 5 sublay repair.

*Range.†Mean.‡Oxford level of evidence: ? 5 unknown; IH 5 incisional hernia; PE 5 polyester; PP 5 polypropylene; Pros 5 prospective; RCT 5 randomized controlled trial, Retro 5 retrospective, SSI 5 surgical site

infection.

L.Tim

merm

anset

al.Sublay

versusonlay

incisio

nal

hern

iarep

air983

Figure 2 IH recurrence.

984 The American Journal of Surgery, Vol 207, No 6, June 2014

Hematoma

Four studies comprising a total of 307 patients reporteddata regarding hematoma and were included in the meta-analysis (Fig. 6).18,22,26,27 No statistically significant resultswere achieved (odds ratio 5 .54; 95% CI, .21 to 1.38; I2 50%; P 5 .19). During sensitivity analysis, the results andheterogeneity proved to be stable.

Fistula

Four studies18,24,26,27 reported data regarding fis-tula.17,23,25,26 However, the definition of fistula (enterocuta-neous, ligature, or intestinal) differed in 3 studies and wasnot reported in 1. In 2 studies, no cases of postoperative fis-tula were reported.26,27 In the RCT by Venclauskas et al,18

ligature fistula was reported in 14% (8/57) after OR and in4% (2/50) after SR; this was not statistically significant. Inthe study by Abdollahi et al,24 intestinal fistula was re-ported in 3% (1/33) after OR and in .3% (1/312) afterSR; no statistics were performed.

Operation time

Three studies reported data regarding operatingtime.18,21,28 However, not all studies reported standard

Figure 3 Funnel plot of IH results.

deviations, making pooling of the results impossible. All3 studies reported lower mean operating times for OR.The study by Israelsson et al21 reported a mean OR oper-ation time of 92 minutes (95% CI, 88 to 97) comparedwith 102 minutes (95% CI, 96 to 108) in the SR group;no statistical analysis was performed.21 The RCT byVenclauskas et al18 reported a significantly lower (P 5.001) operation time for the OR group (135 minutes;95% CI, 87.5 to 182.5 minutes) compared with the SRgroup (168.4 minutes; 95% CI, 114.4 to 222.4 minutes).The RCT performed by Weber et al28 reported a mean op-eration time of 75 minutes (range 30 to 210 minutes) inthe OR group compared with 77 minutes (range 25 to220 minutes) in the SR group; this was not statisticallysignificant.28

Postoperative pain

Two studies18,20 reported data regarding postoperativepain. However, the 2 studies reported pain in a differentmanner and thus could not be pooled for meta-analysis.In the RCT by Venclauskas et al,18 a visual analog scale(VAS) was used to asses pain during the hospital stay ofthe patient, and an average was presented in their data.The mean VAS score for OR was not statistically differentcompared with SR during rest (3.96 [95% CI, 2.40 to 4.52]vs 3.78 [95% CI, 1.82 to 5.75]; P 5 .607) and during activ-ity (5.48 [95% CI, 3.82 to 7.14] vs 5.2 [95% CI 2.89 to 7.5];P 5 .607). In the prospective study by Kumar, postopera-tive pain was also measured by means of a VAS.20 How-ever, they opted to group their results into significant pain(VAS .5) and nonsignificant pain (VAS ,5) groups. Pa-tients in the SR group had significantly more pain (33.3%vs 61.1%); however, no P value was provided.

Comments

SR seems to be the preferred technique compared withOR for IH repair because it results in lower SSI rates. Inaddition, a trend toward lower recurrence rates was ob-served when using SR. The 2 techniques did not differ withregard to seroma formation or hematoma. Although results

Figure 4 SSI.

L. Timmermans et al. Sublay versus onlay incisional hernia repair 985

regarding operation time could not be pooled, OR seemedto take up less time than SR. No definitive conclusionscould be drawn from results regarding postoperative painand fistula.

Recurrence

Worldwide, IH repair is a frequently performed surgicalprocedure. OR and SR are well-known techniques for IHrepair; they both have advantages and disadvantages. OR isassumed to be easier to perform and less time-consuming.The dissection of the posterior rectus fascia from the rectusmuscle during SR can be challenging, especially in cases ofprevious infection and adhesions. However, SR is assumedby many to reduce IH recurrence although this was notdirectly confirmed in this study. This reduction of IHrecurrence might be caused by a higher tensile strength ofthe abdominal wall after SR. The intra-abdominal pressurewould fix the mesh between the posterior fascia and theabdominal muscle. However, experimental studies focusingon abdominal wall strength have reported inconsistentresults. Binnebosel et al29 described less stability of themesh with regard to OR in their hernia model, but Koet al12 discovered no significant differences in tensilestrength between OR and SR in their study.

Furthermore, the SR mesh position seems favorable withregard to tissue incorporation. Binnebosel et al10 describedsignificant in growth and lower collagen type 1:3 ratio afterOR in an animal model. These factors are known to pro-mote IH formation. In addition, in an experimental studyby Garcia-Urena et al,30 an increase of mesh shrinkage after

Figure 5 S

OR was described. Mesh shrinkage may reduce this overlapand promote IH recurrence.

Postoperative complications

Although OR is thought to be easier and quicker toperform, it has been suggested that the dissection of thesuprafascial space would promote seroma formation andSSI.31 With regard to seroma formation, we could not de-tect any difference in this study. However, of all studiesthat reported data on seroma, only half of them provided in-formation regarding drain placement.18,22

SR was superior to OR with regard to SSI. This could beexplained by the more superficial position of OR making iteasier for bacterial colonization. Additionally, mesh posi-tioning on the posterior rectus fascia would benefit from amore vascularized area compared with the OR position.

Dissection of the SR space could be troubled by thehigher grade of vascularization and the presence of theinferior and superior epigastric arteries, which couldincrease the amount of hematoma formation after SR.However, this was not observed during meta-analysis.

Information regarding postoperative pain and fistulacould not be pooled in this meta-analysis. Definitions ofthe postoperative complications varied and/or were re-ported in such a manner that making assumptions regardingthese topics was not possible. However, it seems plausiblethat the dissection of the space between the posterior rectusfascia and the rectus muscle is a more elaborate procedureand with more possibilities to damage, ligate, or cut nervesand thus induce (chronic) pain. The experience of the

eroma.

Figure 6 Hematoma.

986 The American Journal of Surgery, Vol 207, No 6, June 2014

surgeon in these cases is also of utmost importance andcould make a large difference with regard to postoperativepain.

Limitations

Performing a good meta-analysis can be challenging,and a number of aspects should be kept in mind during theprocess. Ideally, a meta-analysis should consist of a numberof high-quality studies, preferably RCTs, with comparablestudy populations and interventions. The results of theindividual studies should be homogeneous and have acommon dependent variable or end point. Also, the qualityof the data being combined should be similar amongstudies. This limits the possibility of bias and heterogeneitybetween studies.32

In this meta-analysis, only 2 RCTs were included, andthe vast majority of the studies were of retrospective nature.However, all of the included studies had comparable studypopulations, similar interventions, and a common end point(incisional hernia recurrence). Additionally, a quality as-sessment was performed to make sure that nonrandomizedtrials were of decent quality before including them in themeta-analysis. However, excluding trials might also facil-itate publication bias. All data from the excluded studiesregarding IH recurrence, SSI, and seroma displayed com-parable results as calculated in this meta-analysis. Includ-ing these studies would not have changed any outcomesexcept for IH recurrence. Instead of a trend, SR would havebeen significantly better with regard to IH recurrencecompared with OR.

The results regarding IH recurrence, the main outcomeof this study, were subject to a high level of heterogeneity.When looking at the funnel plot for this analysis, the studyby Weber et al28 in particular is the source of asymmetry.Removal of this study reduced the heterogeneity to 0, andresults became statistically significant. Furthermore, theRCT by Weber et al was of mediocre quality and mightbe subject to location bias. Additionally, this study did dif-fer somewhat compared with other studies because Weberet al only included larger hernias, which could contributeto heterogeneity.

Although the included studies were all assessed regardingquality, the retrospective nature of most studies still is alimitation to this meta-analysis. The number of IH recur-rence and postoperative complications are likely to be

underreported especially because no ultrasonography wasperformed in any of the included trials. In addition, a lot ofvariables such as the amount of mesh overlap, experience ofthe surgeon, wound classification, number of stitches, andtime to recurrence remain unclear or differed betweenstudies. These inconsistencies and the instability of thismeta-analysis make it difficult to allow for solid conclusions.

Conclusion

Although the majority of included studies were of aretrospective nature, SR seems to be the preferred tech-nique for IH repair compared with OR.

Acknowledgments

We would like to thank Wichor Bramer for his assistanceon the search strategy.

References

1. Mudge M, Hughes LE. Incisional hernia: a 10 year prospective study

of incidence and attitudes. Br J Surg 1985;72:70–1.

2. Hoer J, Lawong G, Klinge U, et al. Einflussfaktoren der Narbenhernie-

nentstehung. Retrospektive Untersuchung an 2.983 laparotomierten Pa-

tienten uber einen Zeitraum von 10 Jahren [Factors influencing the

development of incisional hernia. A retrospective study of 2,983 laparot-

omy patients over a period of 10 years]. Chirurg 2002;73:474–80.

3. Sugerman HJ, Kellum Jr JM, Reines HD, et al. Greater risk of inci-

sional hernia with morbidly obese than steroid-dependent patients

and low recurrence with prefascial polypropylene mesh. Am J Surg

1996;171:80–4.

4. Bevis PM, Windhaber RA, Lear PA, et al. Randomized clinical trial of

mesh versus sutured wound closure after open abdominal aortic aneu-

rysm surgery. Br J Surg 2010;97:1497–502.

5. Luijendijk RW, Hop WC, van den Tol MP, et al. A comparison of su-

ture repair with mesh repair for incisional hernia. N Engl J Med 2000;

343:392–8.

6. Burger JW, Luijendijk RW, Hop WC, et al. Long-term follow-up of a

randomized controlled trial of suture versus mesh repair of incisional

hernia. Ann Surg 2004;240:578–83; discussion, 83–5.

7. Stoppa R, Louis D, Verhaeghe P, et al. Current surgical treatment of

post-operative eventrations. Int Surg 1987;72:42–4.

8. Rives J, Lardennois B, Pire JC, et al. Les grandes eventrations. Impor-

tance du "volet abdominal" et des troubles respiratoires qui lui sont sec-

ondaires [Large incisional hernias. The importance of flail abdomen and

of subsequent respiratory disorders]. Chirurgie 1973;99:547–63.

9. Chevrel JP. Traitement des grandes eventrations medianes par plastie

en paletot et prothese [The treatment of large midline incisional

L. Timmermans et al. Sublay versus onlay incisional hernia repair 987

hernias by "overcoat" plasty and prothesis (author’s transl)]. Nouv

Presse Med 1979;8:695–6.

10. Binnebosel M, Klink CD, Otto J, et al. Impact of mesh positioning on

foreign body reaction and collagenous ingrowth in a rabbit model of

open incisional hernia repair. Hernia 2010;14:71–7.

11. Johansson M, Gunnarsson U, Strigard K. Different techniques for

mesh application give the same abdominal muscle strength. Hernia

2011;15:65–8.

12. Ko R, Kazacos EA, Snyder S, et al. Tensile strength comparison of

small intestinal submucosa body wall repair. J Surg Res 2006;135:9–17.

13. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for

reporting systematic reviews and meta-analyses of studies that evalu-

ate healthcare interventions: explanation and elaboration. BMJ 2009;

339:b2700.

14. Korenkov M, Paul A, Sauerland S, et al. Classification and surgical

treatment of incisional hernia. Results of an experts’ meeting. Langen-

becks Arch Surg 2001;386:65–73.

15. Wells G, Shea B, O’Connell D, et al. The Newcastle-Ottawa scale

(NOS) for assessing the quality of nonrandomized studies in meta-

analysis. http://www.ohri.ca/programs/clinical_epidemiology/nosgen.

doc. Accessed February 20, 2013.

16. Deeks JJ, Higgins JP, Altman DG. Analysing Data and Undertaking

Meta-Analyses. In: Higgins JP, Green S, eds. Cochrane Handbook

for Systematic Reviews of Interventions: Cochrane Book Series. Chi-

chester, UK: John Wiley & Sons Ltd; 2008. http://dx.doi.org/10.1002/

9780470712184.ch9.

17. Peters JL, Sutton AJ, Jones DR, et al. Contour-enhanced meta-analysis

funnel plots help distinguish publication bias from other causes of

asymmetry. J Clin Epidemiol 2008;61:991–6.

18. Venclauskas L, Maleckas A, Kiudelis M. One-year follow-up after in-

cisional hernia treatment: results of a prospective randomized study.

Hernia 2010;14:575–82.

19. Weber G, Horvath OP. Hasfali servek muteti kezelesenek eredme-

nyei: varrattal, illetve halo beultetessel (onlay vs sublay) torteno

nyitott es laparoszkopos hasfal rekonstrukcio eredmenyeinek ossze-

hasonlitasa (prospektiv, randomizalt, multicentrikus vizsgalat) [Re-

sults of ventral hernia repair: comparison of suture repair with

mesh implantation (onlay vs sublay) using open and laparoscopic

approach–prospective, randomized, multicenter study]. Magy Seb

2002;55:285–9.

20. Kumar V, Rodrigues G, Ravi C, et al. A comparative analysis on var-

ious techniques of incisional hernia repairdexperience from a tertiary

care teaching hospital in South India. Indian J Surg 2012;75:1–3.

21. Israelsson LA, Smedberg S, Montgomery A, et al. Incisional hernia re-

pair in Sweden 2002. Hernia 2006;10:258–61.

22. Gleysteen JJ. Mesh-reinforced ventral hernia repair: Preference for 2

techniques. Arch Surg 2009;144:740–5.

23. Forte A, Zullino A, Manfredelli S, et al. Incisional hernia surgery: re-

port on 283 cases. Eur Rev Med Pharmacol Sci 2011;15:644–8.

24. Abdollahi A, Maddah GH, Mehrabi BM, et al. Prosthetic incisional her-

nioplasty: clinical experience with 354 cases. Hernia 2010;14:569–73.

25. Kingsnorth AN, Sivarajasingham N, Wong S, et al. Open mesh repair

of incisional hernias with significant loss of domain. Ann R Coll Surg

Engl 2004;86:363–6.

26. de Vries Reilingh TS, van Geldere D, Langenhorst B, et al. Repair of

large midline incisional hernias with polypropylene mesh: comparison

of three operative techniques. Hernia 2004;8:56–9.

27. Coskun M, Peker Y, Tatar F, et al. Median incisional hernia and the

factors affecting the recurrence of median incisional hernia repair. Er-

ciyes Tip Derg 2009;31:244–9.

28. Weber G, Baracs J, Horvath OP. ["Onlay" mesh provides significantly

better results than "sublay" reconstruction. Prospective randomized

multicenter study of abdominal wall reconstruction with sutures

only, or with surgical mesh–results of a five-years follow-up]. Magy

Seb 2010;63:302–11.

29. Binnebosel M, Rosch R, Junge K, et al. Biomechanical analyses of

overlap and mesh dislocation in an incisional hernia model in vitro.

Surgery 2007;142:365–71.

30. Garcia-Urena MA, Vega Ruiz V, Diaz Godoy A, et al. Differences in

polypropylene shrinkage depending on mesh position in an experimen-

tal study. Am J Surg 2007;193:538–42.

31. White TJ, Santos MC, Thompson JS. Factors affecting wound compli-

cations in repair of ventral hernias. Am Surg 1998;64:276–80.

32. Ward RA, Brier ME. Retrospective analyses of large medical da-

tabases: what do they tell us? J Am Soc Nephrol 1999;10:

429–32.

988 The American Journal of Surgery, Vol 207, No 6, June 2014

Appendix 1

Search Strategy

Embase 144(hernia/exp OR scar/exp OR ‘postoperative complica-

tion’/exp OR (herni* OR scar* OR cicatr* OR postoperat*OR (post NEXT/1 operat*)):ab,ti) AND (onlay OR chevrelOR prefascial OR ‘pre fascial’):ab,ti AND (sublay OR(rives NEAR/1 stoppa) OR underlay OR ‘under lay’ ORsubfascial OR ‘sub fascial’ OR preperitoneal OR ‘preperitoneal’ OR retrorect* OR (retro NEXT/1 rect*)):ab,ti

Medline OvidSP 106(exp hernia/ OR exp Cicatrix/ OR exp ‘‘Postoperative

Complications’’/ OR (herni* OR scar* OR postoperat* ORpost-operat*).ab,ti.) AND (onlay OR chevrel OR prefas-cial* OR pre fascial*).ab,ti. AND (sublay OR (rives ADJ1stoppa) OR underlay* OR under lay* OR subfascial* ORsub-fascial* OR preperitoneal* OR pre peritoneal* ORretrorect* OR (retro ADJ rect*)).ab,ti.

Cochrane Central 3((herni* OR scar* OR cicatr* OR postoperat* OR

(post NEXT/1 operat*)):ab,ti) AND (onlay OR chevrelOR prefascial OR ‘pre fascial’):ab,ti AND (sublayOR (rives NEAR/1 stoppa) OR underlay OR ‘under lay’

OR subfascial OR ‘sub fascial’ OR preperitoneal OR‘pre peritoneal’ OR retrorect* OR (retro NEXT/1 rect*)):ab,ti

Web of Science 104TS5(((herni* OR scar* OR cicatr* OR postoperat* OR

(post NEXT/1 operat*))) AND (onlay OR chevrel ORprefascial OR ‘pre fascial’) AND (sublay OR (rives NEAR/1 stoppa) OR underlay OR ‘under lay’ OR subfascial OR‘sub fascial’ OR preperitoneal OR ‘pre peritoneal’ ORretrorect* OR (retro NEXT/1 rect*)))

Scopus 143TITLE-ABS-KEY((herni* OR scar* OR cicatr* OR post-

operat*) AND (onlay OR chevrel OR prefascial OR ‘‘prefascial’’) AND (sublay OR (rives W/1 stoppa) OR underlayOR ‘‘under lay’’ OR subfascial OR ‘‘sub fascial’’ OR preper-itoneal OR ‘‘pre peritoneal’’ OR retrorect* OR (retro-rect*)))

PubMed Publisher 6((herni*[tiab] OR scar*[tiab] OR postoperat*[tiab] OR

post-operat*[tiab])) AND (onlay[tiab] OR chevrel[tiab] ORprefascial*[tiab] OR pre fascial*[tiab]) AND (sublay[tiab]OR rives stoppa[tiab] OR underlay*[tiab] OR under lay*[tiab] OR subfascial*[tiab] OR sub-fascial*[tiab] OR pre-peritoneal*[tiab] OR pre peritoneal*[tiab] OR retrorect*[tiab] OR retro rect*[tiab]) AND publisher[sb]