Comparison of Outcomes of Synthetic Mesh vs Suture Repair of Elective Primary Ventral Herniorrhaphy A Systematic Review and Meta-analysis

Transcription

1 Research Original Investigation ASSOCIATION OF VA SURGEONS Comparison of Outcomes of Synthetic vs Repair of Elective Primary Ventral Herniorrhaphy A Systematic Review and Meta-analysis Mylan T. Nguyen, MS; Rachel L. Berger, BA; Stephanie C. Hicks, PhD; Jessica A. Davila, PhD; Linda T. Li, MD; Lillian S. Kao, MD; Mike K. Liang, MD IMPORTANCE More than ventral hernias are repaired in the United States annually, of which 75% are primary ventral hernias (eg, umbilical or epigastric hernias). Despite the volume, there is insufficient evidence to support the use of sutures vs mesh for primary ventral hernia repairs. CME Quiz at jamanetworkcme.com OBJECTIVE To compare suture vs mesh repairs for 3 outcomes: hernia recurrence, surgical site infection (SSI), and seromas. DATA SOURCES Randomized controlled trials, case-control, and cohort studies were identified from OVID, PubMed, and reference lists from January 1, 1980, through June 1, STUDY SELECTION English-language studies with adult patients were eligible for if there was mention of both suture and mesh techniques used during elective repair of a primary ventral hernia. Two study authors independently ed the 1492 articles originally identified and selected 9 for analysis. The Downs and Black 26-item checklist was used to critically assess the risk of bias. DATA EXTRACTION Year of publication, study design, inclusion and exclusion criteria, number of patients, follow-up duration, use of preoperative antibiotics, size of hernias repaired, age, body mass index (calculated as weight in kilograms divided by height in meters squared), American Society of Anesthesiologists grade, repair techniques, incidence of hernia recurrence, seroma, and SSI. DATA EXTRACTION AND SYNTHESIS Three separate univariate meta-analyses for each end point followed by a multivariate meta-analysis were performed. Across all 9 studies, there were 637 mesh repairs and 1145 suture repairs. The pooled mesh repairs demonstrated a 2.7% recurrence rate, 7.7% seroma rate, and 7.3% SSI rate The pooled suture repairs demonstrated an 8.2% recurrence rate, 3.8% seroma rate, and 6.6% SSI rate. On the basis of results from the multivariate meta-analysis, recurrences (log odds ratio, 1.04; 95% CI, 1.58 to 0.52) were more common with suture repair, whereas seromas (0.84; ) and SSIs (0.65; ) were more common with mesh repair. CONCLUSIONS AND RELEVANCE repair has a small reduction in recurrence rates compared with suture repairs for primary ventral hernias, but an increased risk of seroma and SSI was observed. Further high-quality studies are necessary to determine whether suture or mesh repair leads to improved outcomes for primary ventral hernias. JAMA Surg. 2014;149(5): doi:10.1/jamasurg Published online February 19, Author Affiliations: Department of Surgery, The University of Texas Health Sciences Center, Houston, Texas (Nguyen, Kao, Liang); Department of Surgery, Baylor College of Medicine, Houston, Texas (Berger, Li); Department of Statistics, Dana Farber Cancer Institute, Boston, Massachusetts (Hicks); Houston Veterans Affairs Health Services Research and Development Center for Excellence, Baylor College of Medicine, Houston Texas (Davila). Corresponding Author: Mike K. Liang, MD, Department of Surgery, The University of Texas Health Sciences Center, Lyndon B. Johnson Hospital, 5656 Kelley Street, Houston, TX (mkliang@yahoo.com and mike.liang@uth.tmc.edu). 415

2 Research Original Investigation vs s for Primary Ventral Hernia Repairs More than ventral hernias are repaired in the United States annually, of which 75% are primary ventral hernias (nonincisional). 1,2 To mitigate the risk of hernia recurrence that is associated with the use of suture repair techniques, mesh repair has become a popular alternative. Despite the strong trend toward the use of mesh repairs, particularly with large primary ventral hernias, there is insufficient quality data supporting this practice. In addition, outcomes of seroma formation and surgical site infection (SSI) are less commonly reported but are important outcomes associated with significant morbidity and cost. Furthermore, wound complications and seromas may be associated with increased rates of hernia recurrence. This study assesses hernia recurrence, seroma formation, and SSI after 2 treatment options for hernia repair mesh and suture. To appropriately account for the correlation between the multiple end points, 2 statistical techniques a univariate meta-analysis that examines each complication independently and a multivariate meta-analysis that examines all 3 complications concurrently are used. Methods Search Methods for the Identification of Studies A literature search following Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines and the Metaanalysis of Observational Studies in Epidemiology checklist for observational studies was performed with the MEDLINE database, using PubMed and OvidSP, with the following medical subject heading (MeSH) terms: ventral hernia and suture Figure 1. Systematic Literature Search Flowchart 1492 Records identified and deduplicated, with MeSH limits 1131 PubMed 360 OVIDSP/MEDLINE 1 Reference 1492 Titles ed 108 Abstracts ed 24 Articles ed 9 Studies included in quantitative synthesis (meta-analysis) MeSH indicates medical subject heading Titles excluded (not related to primary ventral hernia repair) 84 Abstracts excluded Study design not suitable No or not relevant comparator group Outcomes irrelevant Hernias not of interest 15 Articles excluded <10 patients per study group Included cirrhotic patients Hernias not of interest or mesh. 3,4 These searches were restricted according to the following MeSH limits: (1) January 1, 1980, through June 1, 2012, (2) human, (3) adults 19 to 44 years and 45 years or older, and (4) English-language articles. A search of the references of captured articles was also performed. Titles, abstracts, and entire articles were ed by 2 independent ers (R.L.B. and M.K.L.), and discrepancies were resolved by consensus among the authors. Inclusion and Exclusion Criteria Articles were eligible for if there was mention of both suture and mesh techniques used during elective repair of a primary ventral hernia. Exclusion criteria included the following: description of a surgical technique only, only evaluated patients with cirrhosis, assessed emergency cases only, consisted of fewer than 10 patients in either study group (mesh or suture), consisted of laparoscopic cases only, included incisional hernias (including recurrent primary ventral hernias), or did not report comparative outcomes of SSI, hernia recurrence, or seroma formation. Data Extraction The following information was extracted from each article: year of publication, study type, inclusion and exclusion criteria, number of patients, follow-up duration, use of preoperative antibiotics, size of hernias repaired, age, body mass index (calculated as weight in kilograms divided by height in meters squared), American Society of Anesthesiologists grade, repair techniques, and incidence of hernia recurrence, seroma, and SSI. All studies were graded for risk of study bias using the Downs and Black 26-item checklist for the assessment of methodologic quality that included the following: 10 variables for assessing quality of reporting, 3 for external validity, 7 for bias, and 6 for confounding. 5 Two independent ers (R.L.B. and M.K.L.) ed and scored each study on a scale of 0 (lowest) to 26 (highest) and then gave a global grade: low, moderate, or high risk of bias. Reviewers were not masked, and studies were evaluated in no particular order. Discrepancies between scores were discussed and resolved between the ers. Statistical Analysis The statistical software R was used for the analysis of data. 6 A univariate meta-analysis for each end point was performed to calculate the odds ratios (OR) with 95% CIs. A multivariate meta-analysis was performed to jointly estimate the log OR of mesh vs suture for the 3 end points: hernia recurrence, seroma, and SSI. Heterogeneity was assessed using the Higgins I 2 statistic. The fixed-effects model was used throughout. Results Systematic Review The results of the systematic are shown in Figure 1. A total of 1492 potential citations were initially identified using the MeSH terms and limits. An evaluation of titles 416 JAMA Surgery May 2014 Volume 149, Number 5 jamasurgery.com

3 vs s for Primary Ventral Hernia Repairs Original Investigation Research Table 1. Characteristics of the Studies Included in the Meta-analyses Source Arroyo et al, Asolati et al, Berger et al, Bowley et al, Eryilmaz et al, Halm et al, Polat et al, Sanjay et al, Stabilini et al, Study Design randomized control trial observational observational randomized control trial observational Total Patients Preoperative Antibiotic Use Follow-up 200 NR Mean, 64 mo (range, mo) Repair Technique Simple Repair Technique a placed in the extraperitoneal plane. plug used on defects up to 3 cm diameter. sheet inserted for larger hernias. Risk of Bias Inclusion and Exclusion Criteria Low Excluded if ASA grade 4, emergency 229 NR NR NR Onlay, inlay, combined High Excluded if laparoscopic, not elective, recurrent, ASA grade 4, ascites, liver insufficiency, emergency 411 Not routinely used Median, 70 mo (range, mo) 473 NR Median, 25 mo (range, mo) 111 Yes in all cases Mean, 37 mo (range, 6-66 mo) 110 Not routinely used Median, 32 mo (range, mo) 50 Yes in all cases Mean, 22 mo (range, 6-44 mo) Not routinely used in suture cases; yes in all mesh cases 98 Not routinely used Median, 4.5 y (range, 1-8 y) Mean, 52.9 mo (range, 8-60 mo) Simple PHS and underlay Medium Included if elective primary ventral hernia repair with suture or polypropylene hernia system mesh Mayo NR High Excluded if pediatric, incidental Simple Mayo and simple Onlay High Excluded if laparoscopic, no follow-up Preperitoneal plane High NR Mayo PHS and onlay High Excluded if strangulated, recurrent, omphalocele or periumbilical fistulas, defects >4 cm, ASA grade 4, COPD, CAD, ascites, CRF, DM Mayo and simple Mayo Flat mesh placed in pre-peritoneal plane. plug used for hernia <5 cm diameter. Flat mesh placed in the preperitoneal space. High High Included if umbilical and paraumbilical, elective NR Abbreviations: ASA, American Society of Anesthesiologists; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; CRF, chronic renal failure; DM, diabetes mellitus; NR, not reported; PHS, Prolene Hernia System. a All studies, except that by Berger et al, 9 reported use of polypropylene mesh type. resulted in omission of 1384 articles with titles not related to primary hernia repair. Review of the remaining 108 article abstracts resulted in the exclusion of an additional 84 articles because of study design issues, unsuitable comparator groups, irrelevant outcomes, or included hernias not of interest. Of the 24 articles remaining, 15 were excluded because they had sample size issues, only evaluated patients with cirrhosis, or included incisional hernias. Nine articles met the inclusion and exclusion criteria and were included in the final analysis. The characteristics of the 9 included articles are summarized in Table 1. The studies included consisted of prospective randomized control trials (n = 2), retrospective medical record s (n = 4), and prospective observational studies (n = 3). The use of preoperative antibiotics varied among the studies, with 1 study reporting a difference in use between suture and mesh repairs. The global risk of bias for the articles varied. After consensus among the ers, 7 articles had a high risk of bias and 1 article each had a medium and low risk of bias. Although the risk of bias varied because each article met the minimum criteria for inclusion, the articles were included in the meta-analysis. Table 2 presents the extracted information for each article with respect to suture and mesh use. There were 1782 total cases, with 637 mesh repairs and 1145 suture repairs. Univariate Meta-analysis The univariate meta-analysis for recurrence included all 9 articles. The pooled recurrence rate was 2.7% (n = 17) for mesh repair and 8.2% (n = 94) for suture repair. With respect to recurrence, there was no suggestion of heterogeneity among the studies. The I 2 test was 0% (P =.49), suggesting little variation among the included studies. The analysis of pooled data revealed that mesh repair was associated with a lower recurrence rate compared with suture repair (OR, 0.31; 95% CI, ; P <.001). Only 1 study noted the interval from operative repair to hernia recurrence in its cohort (median, 28 months; range, months) 12 (Figure 2A). Six studies were included in the analysis for seroma complication. The pooled seroma rate was 7.7% (n = 32) for mesh repair and 3.8% (n = 21) for suture repair. The I 2 test was insignificant for heterogeneity among the articles (I 2 = 0%, P =.48). The pooled OR was 2.27 (95% CI, ; P =.004) for risk of seroma formation with mesh compared with suture repair. Patients undergoing mesh repair had an increased risk of developing a seroma complication (Figure 2B). Seven studies evaluated SSI as a complication after primary ventral hernia repair using mesh or suture. The pooled SSI rate was 7.3% (n = 31) for mesh repair and 6.6% (n = 43) jamasurgery.com JAMA Surgery May 2014 Volume 149, Number 5 417

4 Research Original Investigation vs s for Primary Ventral Hernia Repairs Table 2. Characteristics of and Use Source Hernia Size Age, y a BMI a ASA Grade Arroyo et al 7 >3 cm: 30 <3 cm: NR I/II: 85 III: 15 (%) Recurrence Seroma SSI 11 (11.0) 5 (5.0) 3 (3.0) Asolati et al 8 97 NR NR NR NR 7 (7.7) NR NR Berger et al Mean (SD), 2.0 (0.2) cm (0.6) 30.5 (0.3) I: 15 II: 128 III: 127 IV: (8.2) 11 (3.9) 21 (7.5) Bowley et al NR NR NR NR 16 (4.0) NR NR Eryilmaz et al <3 cm 43.7 (12.2) NR NR 9 (14.0) 4 (6.0) 2 (3.0) Halmetal NR NR NR NR 14 (14.3) NR 9 (8.0) Polat et al NR 49.7 (14) NR I/II: 7 2 (11.1) 0 1 (5.6) III: 2 Sanjay et al NR 53 Median, 31.2 NR 7 (11.5) 1 (1.6) 7 (11.5) Stabilini et al Mean (SD), (14.7) (2.2) cm Arroyo et al 7 >3 cm: 32 <3 cm NR I/II: 88 III: 12 1 (1.0) 6 (6) 2 (2) Asolati et al NR NR NR NR 4 (3.0) NR NR Berger et al Mean (SD), 4.7 (0.3) cm (0.9) 32.5 (0.4) I: 3 II: 41 III: 83 IV: 3 7 (5.4) 18 (13.8) 25 (19.2) Bowley et al NR NR NR NR 2 (2.5) NR NR Eryilmaz et al >3 cm 49.4 (12.0) NR NR 1 (2.0) 4 (8) 3 (6) Halmetal NR NR NR NR 0 NR 0 Polat et al NR Range, NR I/II: 12 III: (3.1) 0 Sanjay et al NR 54 Median, NR Stabilini et al Mean (SD), 2.8 (1.6) cm NR 2 (3.3) 3 (4.8) 1 (1.1) Abbreviations: ASA, American Society of Anesthesiologists; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); NR, not reported; SSI, surgical site infection. a Reported as means with or without SD unless noted. for suture repair. The resultant pooled univariate OR showed that mesh repair did not differ in SSI occurrences when compared with suture repair (OR, 1.53; 95% CI, ; P =.08). However, analysis of heterogeneity revealed an I 2 of 47.1% (P =.08), indicating that the studies were dissimilar (Figure 2C). Multivariate Meta-analysis A multivariate meta-analysis was performed to jointly estimate the effect sizes. This meta-analysis incorporated the correlation of the estimates, which can be increased by borrowing strength from external studies. All 9 articles were included in the analysis. Recurrence was associated with suture repair (log OR, 1.05; 95% CI, 1.58 to 0.52; P <.001), whereas risk of seroma formation and SSI occurrence was associated with mesh repair (0.84; ; P <.004; and 0.65; ; P =.02, respectively). When assessing heterogeneity, an I 2 of 16.8% (P =.25) was found, indicating that the studies were not heterogeneous. Discussion Primary ventral hernias are distinguishable from incisional hernias and warrant independent research to guide best practice. Using a multivariate meta-analysis, this study found that the use of mesh in primary ventral hernia repairs resulted in fewer hernia recurrences but more seromas and SSIs. Only 1 other meta-analysis has assessed the morbidity associated with mesh vs suture repair of primary ventral hernias. The meta-analysis by Aslani and Brown 16 found that mesh offered a reduction in recurrence rates but no difference in wound complications in the repair of primary umbilical hernias. However, this meta-analysis failed to exclude studies with incisional hernias, including recurrent hernias. Incisional hernias result from failure of a previous incision, indicating a difference in cause when compared with primary ventral hernias. This meta-analysis also included studies that involved cases in patients requiring emergency surgery and incarcerated patients, which are vastly 418 JAMA Surgery May 2014 Volume 149, Number 5 jamasurgery.com

5 vs s for Primary Ventral Hernia Repairs Original Investigation Research Figure 2. Associations Between Outcome and Repair Type A Experimental Control Study, OR (95% CI) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Fixed effects model ( ) Fixed, % Random, % Random effects model 0.35 ( ) Heterogeneity: I 2 = 0%, τ 2 = 0, P = OR (95% CI) B Experimental Control Study, OR (95% CI) ( ) ( ) ( ) ( ) ( ) ( ) Fixed effects model ( ) Fixed, % Random, % Random effects model 2.32 ( ) Heterogeneity: I 2 = 0%, τ 2 = 0, P = OR (95% CI) C Experimental Control Study, OR (95% CI) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Fixed effects model ( ) Fixed, % Random, % Random effects model 0.94 ( ) Heterogeneity: I 2 = 47.1%, τ 2 = 0.8, P = OR (95% CI) A, Hernia recurrence; B, seroma; C, surgical site infection. OR indicates odds ratio. different from elective repair. In addition, wound infections, seroma formation, and hematoma formation were combined into 1 category: wound complications. The test for heterogeneity when assessing the association between wound complications and repair type in cohort studies was I 2 = 47% for the metaanalysis by Aslani and Brown, indicating that the studies were heterogeneous. 16 Two Cochrane s that assessed mesh vs suture repair found mesh to be associated with lower hernia recurrence in both inguinal and incisional hernia repairs. In the setting of inguinal hernia repairs, one of the s did not find a difference in the occurrences of seromas and wound infections, whereas the other found that mesh use resulted in an increase in wound infections in incisional hernias. 17,18 Aside from these studies, analysis of the effects of mesh and suture use in the setting of primary ventral hernia repairs has not been analyzed adequately. The studies included in our analysis used several different techniques for repair, particularly a simple suture or the Mayo technique for suture repair and polypropylene mesh placed in different locations. Although there are many kinds of mesh, most studies included in this analysis used polypropylene. The reason for this use was not clearly stated in the studies, but polypropylene is relatively economical and easy to handle, making it a common choice for hernia repairs. jamasurgery.com JAMA Surgery May 2014 Volume 149, Number 5 419

6 Research Original Investigation vs s for Primary Ventral Hernia Repairs In some studies, a mesh plug was also used. When comparing the different techniques for mesh placement, the onlay (prefascial) technique has resulted in fewer recurrences, but the results of the analysis were not significant. 18 In a study by Lau and Patil, 19 simple suture repair resulted in a higher recurrence rate when compared with the Mayo technique. In this meta-analysis, the different techniques for suture and mesh repair were too varied to allow for adequate analysis. The results of this meta-analysis have important implications for clinical practice. The use of mesh may result in an increase in seroma formation and SSI risk, but use of mesh may be more beneficial, especially in recurrence. A cost-analysis of treatments for inguinal hernias found that mesh proved to be more cost-effective in the long run. In addition, this study found that the recurrence rate was highly correlated with the cost-effectiveness ratio of the differing techniques. 20 With the increased risk of SSIs and seromas, the cost-effectiveness of mesh repair with marginal recurrence rates may be mitigated compared with other repair types. The use of more advanced statistical techniques to combine information across several studies is being sought to establish consensus quantitative evidence that may contribute to hypothesis generation. Because of its ability to synthesize information to assist in decision making, the metaanalytic technique has been used in several different disciplines. A meta-analysis has the ability to detect associations that are typically not seen in individual studies because of low statistical power. 21 In addition, the metaanalytic technique has the advantage of making results more generalizable, thereby improving the external validity of an association. 22 Typically, studies use a univariate metaanalysis to assess the effect size across studies for a single outcome of interest. However, in clinical practice, there are numerous outcomes that may be the result of a technique, and these outcomes may correlate. As such, to more appropriately understand the relationship between an intervention and its outcomes, a multivariate meta-analysis is used. A multivariate meta-analysis allows for a pooled statistic for each end point to be calculated while accounting for any correlation between end points and discrepancies that may exist within a study and among different studies. 23 This study uses 2 statistical approaches to analyze the morbidity of mesh and suture repairs of primary ventral hernias. In the first approach, we analyzed the results using a standard univariate meta-analysis for each morbidity independently. From this, we found mesh to be associated with less recurrence but more seroma formations. The univariate analysis did not find a difference in SSIs between mesh and suture repairs. In the second approach, the multivariate metaanalysis found that SSI was associated with mesh use. Using a multivariate meta-analysis allowed us to corroborate the findings from other studies by ruling out possible interaction among the 3 most common morbidities assessed in the literature after mesh and suture repair of primary ventral hernias. This meta-analysis included only 1 study that rated low on the Downs and Black checklist. Most of the studies had a high degree of risk of bias, lending to the possibility that the metaanalysis may not appropriately capture valuable individual patient data. Although this may be the case, our results are aligned with the current literature. Hernia size is likely to be a confounder for the choice of mesh or suture repair, but the studies included did not uniformly report on hernia size. Therefore, a subgroup analysis was not possible. Conclusions The results of our multivariate meta-analysis support previous reports that found that the use of mesh for primary ventral hernias is associated with a lower recurrence risk. In addition, mesh use resulted in an increased risk of seroma formation and SSIs compared with suture repair. Despite being one of the most common surgical procedures, the level of evidence available on repair of primary ventral hernias remains poor with largely level II to IV data. Additional prospective randomized control trials are warranted to corroborate the findings of this meta-analysis. ARTICLE INFORMATION Accepted for Publication: July2,2013. Published Online: February 19, doi:10.1/jamasurg Author Contributions: Dr Liang had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Nguyen, Berger, Li, Liang. Acquisition of data: Berger, Li, Liang. Analysis of data: Nguyen, Berger, Hicks, Davila, Li, Kao, Liang. Drafting of the manuscript: Nguyen, Berger, Hicks, Liang. Critical revision of the manuscript for important intellectual content: Nguyen, Berger, Davila, Li, Kao, Liang. Statistical Analysis: Nguyen, Berger, Hicks, Davila, Liang. Administrative, technical, or material support: Nguyen, Li, Liang. Study supervision: Liang. Conflict of Interest Disclosures: Dr Liang is currently receiving a grant from Lifecell and Acell. No other disclosures were reported. Previous Presentation: Presented at the Annual Meeting of the Association of Veterans Affairs Surgeons; April 22, 2013; Milwaukee, Wisconsin. REFERENCES 1. Poulose BK, Shelton J, Phillips S, et al. Epidemiology and cost of ventral hernia repair: making the case for hernia research. Hernia. 2012;16(2): Muysoms FE, Miserez M, Berrevoet F, et al. Classification of primary and incisional abdominal wall hernias.hernia. 2009;13(4): Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting: Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283(15): Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic s and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration.bmj. 2009;339:b2700. doi: /bmj.b Downs SH, Black N. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. J Epidemiol Community Health. 1998;52(6): Gasparrini A, Armstrong B, Kenward MG. Multivariate meta-analysis for non-linear and other multi-parameter associations. Stat Med. 2012;31(29): Arroyo A, García P, Pérez F, Andreu J, Candela F, Calpena R. Randomized clinical trial comparing 420 JAMA Surgery May 2014 Volume 149, Number 5 jamasurgery.com

7 vs s for Primary Ventral Hernia Repairs Original Investigation Research suture and mesh repair of umbilical hernia in adults. Br J Surg. 2001;88(10): Asolati M, Huerta S, Sarosi G, Harmon R, Bell C, Anthony T. Predictors of recurrence in veteran patients with umbilical hernia: single center experience. Am J Surg.2006;192(5): Berger RL. Clapp ML, Chu C, Hicks SC, Awad SS, Liang MK. repair vs prolene hernia system for elective primary ventral hernia repairs. In Submission. 10. Bowley DM, Boffard KD. Penetrating lung hernia with pulmonary evisceration: case report. J Trauma. 2001;50(3): Eryilmaz R, Sahin M, Tekelioglu MH. Which repair in umbilical hernia of adults: primary or mesh? Int Surg. 2006;91(5): Halm JA, Heisterkamp J, Veen HF, Weidema WF. Long-term follow-up after umbilical hernia repair: are there risk factors for recurrence after simple and mesh repair. Hernia. 2005;9(4): Polat C, Dervisoglu A, Senyurek G, Bilgin M, Erzurumlu K, Ozkan K. Umbilical hernia repair with the Prolene hernia system. Am J Surg. 2005;190(1): Sanjay P, Reid TD, Davies EL, Arumugam PJ, Woodward A. comparison of mesh and sutured repair for adult umbilical hernias. Hernia. 2005;9(3): Stabilini C, Stella M, Frascio M, et al. versus direct suture for the repair of umbilical and epigastric hernias: ten-year experience. Ann Ital Chir. 2009;80(3): Aslani N, Brown CJ. Does mesh offer an advantage over tissue in the open repair of umbilical hernias? a systematic and meta-analysis. Hernia. 2010;14(5): Scott N, Go PM, Graham P, McCormack K, Ross SJ, Grant AM. Open mesh vs non-mesh for groin hernia repair. Cochrane Database Syst Rev. 2011;(10):CD doi:10.2 / CD den Hartog D, Dur AH, Tuinebreijer WE, Kreis RW. Open surgical procedures for incisional hernias. Cochrane Database Syst Rev. 2008;(1):CD doi:10.2/ cd pub Lau H, Patil NG. Umbilical hernia in adults. Surg Endosc. 2003;17(12): Stylopoulos N, Gazelle GS, Rattner DW. A cost-utility analysis of treatment options for inguinal hernia in 1,513,008 adult patients. Surg Endosc. 2003;17(2): Cohn LD, Becker BJ. How meta-analysis increases statistical power. Psychol Methods. 2003;8(3): Shadish WR, Cook TD, Campbell DT. Experimental and Quasi-Experimental Designs for Generalized Causal Inference. Boston, MA: Houghton Mifflin; Riley RD. Multivariate meta-analysis: the effect of ignoring within-study correlation. J R Stat Soc. 2009;172(4):1-23. CORRECTION Error in Byline and Author Contributions: In the Observation entitled Reappraisal of Central Pancreatectomy: A 12-Year Single-Center Experience, publishedintheapril2014issueof JAMA Surgery (2014; 149 [4]: doi:10.1 /jamasurg ), the third author s name in the byline should have been given as Safi Dokmak, and in the Author Contributions as Dokmak. This article was corrected online. jamasurgery.com JAMA Surgery May 2014 Volume 149, Number 5 421