Prophylactic Mesh Placement for the Prevention of Incisional Hernia in High-Risk Patients After Abdominal Surgery: A Systematic Review and Meta-Analysis

Background and objectives In high-risk populations, the efficacy of mesh placement in incisional hernia (IH) prevention after elective abdominal surgeries has been supported by many published studies. This meta-analysis aimed at providing comprehensive and updated clinical implications of prophylactic mesh placement (PMP) for the prevention of IH as compared to primary suture closure (PSC). Materials and methods PubMed, Science Direct, Cochrane, and Google Scholar were systematically searched until March 3, 2020, for studies comparing the efficacy of PMP to PSC in abdominal surgeries. The main outcome of interest was the incidence of IH at different follow-up durations. All statistical analyses were carried out using Review Manager version 5.3 (The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) and Stata 11.0 (Stata Corporation LP, College Station, TX). The data were pooled using the random-effects model, and odds ratio (OR) and weighted mean differences (WMD) were calculated with the corresponding 95% confidence interval (CI). Results A total of 3,330 were identified initially and after duplicate removal and exclusion based on title and abstract, 26 studies comprising 3,000 patients, were included. The incidence of IH was significantly reduced for PMP at follow-up periods of one year (OR= 0.16 [0.05, 0.51]; p=0.002; I2=77%), two years (OR= 0.23 [0.12, 0.45]; p<0.0001; I2=68%), three years (OR= 0.30 [0.16, 0.59]; p=0.0004; I2= 52%), and five years (OR=0.15 [0.03, 0.85]; p=0.03; I2=87%). However, PMP was associated with an increased risk of seroma (OR=1.67 [1.10, 2.55]; p= 0.02; I2=19%) and chronic wound pain (OR=1.71 [1.03, 2.83]; p= 0.04; I2= 0%). No significant difference between the PMP and PSC groups was noted for postoperative hematoma (OR= 1.04 [0.43, 2.50]; p=0.92; I2=0%), surgical site infection (OR=1.09 [0.78, 1.52]; p= 0.62; I2=12%), wound dehiscence (OR=0.69 [0.30, 1.62]; p=0.40; I2= 0%), gastrointestinal complications (OR= 1.40 [0.76, 2.58]; p=0.28; I2= 0%), length of hospital stay (WMD= -0.49 [-1.45, 0.48]; p=0.32; I2=0%), and operating time (WMD=9.18 [-7.17, 25.54]; p= 0.27; I2=80%). Conclusions PMP has been effective in reducing the rate of IH in the high-risk population at all time intervals, but it is associated with an increased risk of seroma and chronic wound pain. The benefits of mesh largely outweigh the risk, and it is linked with positive outcomes in high-risk patients.


Introduction
Any procedure that requires access to the abdominal wall carries with itself the precarious complication of incisional hernia (IH). This is especially common in patients undergoing open bariatric surgery and abdominal aneurysm repair. The incidence of IH is approximately 11%-20% in patients post laparotomy, but it can be as high as 39.9% in high-risk populations such as obesity, prior abdominal operation, abdominal aortic aneurysm, or patients suffering from neoplastic diseases [1][2][3][4]. Annually, 150,000 patients are operated for IH in the United States alone, with one-third repairs occurring within nine years. IH not only creates a financial burden but also leads to poor healthrelated quality of life (QoL) in patients. It is also associated with poor body image and a lower sense of self-worth [2,[5][6].
Mesh placement has been found effective in reducing occurrences of umbilical hernia, inguinal hernia, and parastomal hernia. Previous systematic reviews have also yielded supportive findings regarding the efficacy of prophylactic mesh placement (PMP) in preventing IH [4,[7][8]. However, they did not evaluate the time-based effectiveness of PMP as compared to primary suture closure (PSC) and did not study the differences among various population subgroups, as they were limited by small sample size. Furthermore, the literature remains inconclusive on whether the mesh is efficacious in reducing chronic wound pain [8].
Due to the lack of sufficient quality evidence, there is a need for further high-quality studies to support the use of mesh for IH prevention in high-risk patients [8]. Several new studies have been published since the last meta-analysis, and therefore, we sought to conduct an updated meta-analysis of all studies to date. The larger sample size enabled us to provide a holistic, well-powered assessment of the efficacy of a prophylactic mesh in preventing IH. One of the reasons why the efficacy of a prophylactic mesh has remained unclear is maybe because of varying effectiveness in different patient subgroups. Hence, we also aimed to conduct a range of subgroup analyses to identify specific patient populations in which a prophylactic mesh might be beneficial. Additionally, we sought to evaluate seldom-evaluated aspects of mesh placement, including hematoma, seroma, chronic wound pain, surgical site infections, gastrointestinal complications, operating time, and length of hospital stay.

Materials And Methods
The current study has been carried out in accordance with the Preferred Reporting Items for Systemic Reviews and Meta-analysis (PRISMA) guidelines. Two independent reviewers carried out the literature search, quality assessment, data extraction, and statistical analyses. In case of any conflict, a third reviewer was consulted.

Search strategy
Online databases, including PubMed, Science Direct, and CENTRAL Register of Controlled Trials (Cochrane), were systematically searched from the inception of databases till March 3, 2020, without time or language restrictions. Google Scholar was also searched for gray literature. References of relevant reviews were also manually searched for additional studies. The search strategy for each database is given in Table 1

Data extraction and outcomes
Information regarding study characteristics, demographics, and reported outcomes were extracted. Four different mesh locations were considered: (1) Onlay position (above the anterior rectus sheath or below the abdominal fascia); (2) Retrorectus, also known as the sublay position (between the rectus abdominis muscle and posterior rectus sheath); (3) Preperitoneal (between the posterior rectus sheath and parietal peritoneum), and (4) Intraperitoneal (in the abdominal cavity).
The main outcome of interest was the incidence of IH at different follow-up durations. Other outcomes included seroma, chronic wound pain, hematoma, wound dehiscence, surgical site infection, respiratory and gastrointestinal complications, hospital stay, and operating time. The incidence of IH was confirmed by clinical examination or imaging modalities, such as ultrasonography or computed tomography (CT) scan, and no difference was made between IH diagnosed clinically or through imaging modalities. Gastrointestinal complications included ascites, bowel obstruction, bowel perforation, intra-abdominal abscess, and paralytic ileus. We accepted the primary study investigator's definition for seroma and all remaining outcomes.
The number of patients that presented during follow-up was considered as the denominator instead of randomization numbers for meta-analyses of outcomes. Studies were classified in each follow-up group based on follow-up time. Where specific follow-up was not mentioned, mean or median followup was used to classify the study. Study characteristics and early complications were extracted from earlier publications of a trial if publication of the latest follow-up data lacked them. The incidence of IH was recorded from each follow-up duration. Studies that did not provide means and standard deviations (SD) for hospital stay duration and operation time were not included in the respective analysis.

Statistical analysis
Review Manager v.5.3 (The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) and Stata 11.0 (Stata Corporation LP, College Station, TX) were used for all statistical analyses.
Patients' data were divided into two groups -PSC or PMP -according to the procedure. Weighted mean differences (WMD) and Mantel-Haenszel (MH) odds ratios (OR) were calculated with a 95% confidence interval (CI) and pooled using a random-effects model. Subgroup analyses were performed by stratifying studies according to study design (RCT and observational), mesh location (onlay, retrorectus, preperitoneal, and intraperitoneal), BMI (<40 and >40), and study population (bariatric, neoplastic, vascular, and mixed). The chi-squared test (p-interaction) was used to assess subgroup differences.
Statistical heterogeneity was quantified across studies using Higgin's I 2 statistics, and a value of 25%-50% was considered mild, 50%-75% as moderate, and >75% as severe. The leave one out analysis was performed to determine whether any single study had a disproportionate effect on the pooled results.

Quality assessment
The quality assessment of studies was done using the Cochrane Collaboration risk-of-bias tool and the Newcastle-Ottawa scale for RCTs and observational studies, respectively. Publication bias was assessed using a funnel plot and Egger's regression test. A p-value of ≤ 0.05 was considered significant in all cases.

Results
A total of 3,330 records were identified in the initial search, 3,319 from electronic databases and 11 through references of relevant studies (other sources). After removing duplicates and excluding articles based on title and abstract screening, the full texts of 73 articles were reviewed for eligibility. A total of 26 articles met the inclusion criteria [1-3,5, . Figure 1 shows the PRISMA flowchart summarizing the literature search. The follow-up time in the included studies ranged from six to 60 months. Out of the 26 studies, 17 were RCTs and nine were observational (seven prospective cohorts and two retrospective studies). The included studies enrolled a total of 3,349 participants, from which 3,000 were analyzed (1,397 receiving PMP and 1,603 receiving PSC). The rest were either lost to follow-up or excluded during surgery. Study characteristics and demographics are given in Table 2

Mesh, suture, and surgery details
In all except four studies, patients underwent midline laparotomy. In three studies, laparoscopic surgery was done, and in one study, patients underwent bilateral subcostal incisions [17,[21][22]25]. A variety of meshes were used in the studies, with polypropylene (PP) being the most common one (n=11 studies). Three studies used different biologic meshes, i.e., Alloderm, Surgisis Gold, and Bovine pericardium [14,[18][19]. One study used an unspecified biosynthetic mesh [6]. Other meshes used included, but were not limited to, polyglactin, propylene polyglycolic acid, and polypropylenepolyvinylidene fluoride. Meshes were placed in four different locations. Two studies planted mesh in two separate locations [5,13]. Nine studies placed the mesh in the onlay position. The most commonly used technique in studies for aponeurosis closure was continuous. The diagnostic modalities for IH, mesh, and suture details for included studies are summarized in the Appendices section.

Quality assessment and publication bias
The majority of the RCTs and observational studies were of robust methodological quality. Half of the RCTs either had a high or unclear risk of bias in the blinding of participants and personnel ( Figure 2). Details of bias assessment in observational studies are present in Table 3. The funnel plot showed significant publication bias (Figure 3), and it was confirmed by Egger's regression test (p=0.031).

Results of meta-analyses
The summarized results of all outcomes discussed below are given in Figure 4. Individual outcomes with their forest plots are given under their respective subheadings. A table summarizing the effects of the leave one out analysis for each outcome is given in the appendices section.  ). All forest plots are given in Figure 6.     ) showed significant reduction in IH incidence at the three-years follow-up after PMP. Individual forest plots for all analyses are given in Figure 9.

Incidence of IH at two years
The different subgroups analyses carried out for IH at the two-year and three-year follow-up are given in Table 5.  IH, incisional hernia; No Sig. Diff., no significant difference; Sig. Diff, Significant difference † -Significant difference was found only between onlay and intraperitoneal mesh (p=0.02). † † -Neoplastic group has a significantly lower incidence of IH than the bariatric group (p=0.006).     IH, incisional hernia; OR, odds ratio; No Sig. Diff., no significant difference; Sig. Diff, significant difference †-Significant difference was found between the mixed group with the bariatric (p=0.01), neoplastic (p=0.02), and vascular (p=0.008) groups.

Sensitivity analysis by excluding non-midline incisions and laparoscopic surgeries
Additional sensitivity analyses were done by excluding studies that employed non-midline incisions or laparoscopic procedures for outcomes, namely risk of IH at one year and two years, and seroma ( Figure 13). The three-year and five-year follow-up data for IH and chronic wound pain did not include non-midline incision studies, so they were exempted from this sensitivity analysis. Results did not differ significantly after sensitivity analysis and PMP was still found to significantly reduce the

Discussion
Our study shows that PMP offers a greater benefit in the prevention of IH than PSC. Mesh placement showed a reduction in IH risk at intervals of one-year, 18-months, two-years, three-years, and fiveyears postoperatively. On subgroup analysis, only study population and mesh location were found to influence hernia development.  [7,30]. The exact reason for the difference in findings is unclear, but it was noted that most of the included studies in the previous meta-analyses had a shorter time interval of six months to one year as compared to ours, which evaluated for a longer follow-up of two to three years [7,30]. For shorterterm periods of six months, PMP was inefficacious at IH prevention. Only two studies have inspected the outcomes at a five-year interval [2][3]. However, to determine the accurate recurrence following IH repair, we recommend that patients should be followed for a longer period of time (10-15-year followup).
Not all mesh locations were effective at IH prevention, and the best effect was observed with onlay and intraperitoneal mesh placement. The onlay position was superior to intraperitoneal in efficacy but with higher complication rates. Midline and lateral incisions are best reinforced when onlay mesh placement is adopted. Though not extensively discussed, some studies hinted at the superior efficacy of the onlay position [4,8]. The safety and efficacy of mesh type were not extensively studied in our meta-analysis, as they were beyond the scope of our discussion.
Among the secondary outcomes, only the risk of seroma and chronic wound pain were found to be significantly increased in the PMP group. Seroma was significantly increased (about two times) in those with PMP. This concurred with the analysis by Borab  reported that onlay and preperitoneal PMP were linked with a higher risk of seroma development, which was further aggravated when the PP mesh was placed in the onlay position [4]. This may be well-explained by the extensive dissection in onlay position, thus increasing the likelihood of postoperative complications.
Most seroma cases were less morbid and were treated conservatively with antibiotics and percutaneous drainage. However, some mentioned the removal of mesh due to infection [13,28]. To decrease seroma incidence, subcutaneous drainage and appropriate tissue management were advised [6].
Chronic wound pain significantly impacts QoL in patients after any surgical procedure. The degree of pain is closely associated with the type and extent of surgery, nerve damage, intensity of radio and chemotherapy, and psychosocial factors. The risk of chronic wound pain with mesh placement is of much conjecture, as few studies reported a lower incidence of chronic wound pain with mesh use, but the meta-analysis by Wang et al. found that mesh failed to provide any significant reduction in chronic wound pain [8]. The results of Wang et al. were limited by a small sample size (3 studies; 229 participants) [8]. Our analysis (6 studies; 741 participants) showed that mesh was associated with a significantly increased incidence of chronic wound pain compared to suture closure. To quantify the debilitating burden of chronic pain, few studies included standardized scoring such as visual analog scale (VAS) for pain and the EQ-5D (EuroQol-5 Dimension) and SF-36 (36-Item Short Form Survey) questionnaires for QoL [5,20]. No difference in QoL was found between the mesh and suture groups [5,20]. Patients in both the PSC and PMP groups suffered from chronic pain, which, however, was well-tolerated and rarely interfered with routine activities, hence resulting in higher patient satisfaction [20].
The difference in other secondary outcomes was non-significant. SSI, an infrequent complication in mesh hernioplasties, is influenced by certain risk factors such as mesh type, obesity, smoking history, mean operative time, and degree of emergency [5,16,20,25]. The lack of a significant difference between PMP and PSC for SSI could be attributed to a few postulates. Firstly, only a few studies adopted complication assessment protocols and standardized assessment scales to gauge the effect of PMP on different outcomes and QoL. Secondly, there may be an underestimation of the additional complications due to underreporting.
No significant difference was observed for the length of hospital stay and mean operative time in our study, whereas Wang et al. found mesh use to be associated with increased operative time [8]. Though insignificant, the results showed a trend of reduced length of hospital stay with mesh use but with increased operative time. Additionally, studies failed to analyze the effect of strenuous activities and early resumption of work. We excluded 'reoperation' and 're-hospitalization' outcomes due to discrepancies in the defining criteria.

Strengths, limitations, and future suggestions
To the best of our knowledge, this is the first study to run a follow-up duration-based analysis of IH and included clinical outcomes in various patient populations.
This updated analysis has adopted a more integrated, extensive (including both observational studies and RCTs), and comparative approach to gain better insight into the outcomes. Subgroup analysis for IH outcomes at different follow-up intervals and other significant outcomes (seroma and chronic wound pain) may help predict the postoperative outcomes better.
However, some aspects may have been missed owing to insufficient studies, ambiguous reporting, or the redundancy of the results. There is a lack of universal agreement on the definition of hernia recurrence and the indications for surgical repair. To establish the long-term viability of mesh, the effect of materials and techniques on the outcomes needs to be addressed. Furthermore, the efficacy of mesh type and its location needs to be extensively evaluated.
Since this only catered to elective cases, a comparison analysis between emergency and elective procedures and among other hernia types is warranted. Additionally, the influence of the surgeon's technique and expertise on postoperative outcomes is less frequently addressed and studies should be carried out to explore this aspect as well. These measures may assist in bridging the major gaps in clinical practice.

Clinical implications
Mesh placement has proven to be, repeatedly, effective in decreasing the incidence of IH after elective midline laparotomy and laparoscopy. This stands true for some cases of parastomal hernia and emergency laparotomy. Even with existing infection, mesh use is associated with better results in both hernia prevention and in lowering wound morbidity [2,10].
Additionally, the adoption of a benefit vs. risk approach in vulnerable (high-risk) patient populations may assist in reducing the complications. In one of the included studies by N Argudo et al. (2018), the selection of patients for mesh placement utilizing an algorithm decreased the recurrence of the hernia, lowered the number of complications, and saved a considerable cost burden [24]. Therefore, a standardized approach for mesh placement can assist in lowering the cost burden and in decreasing the mortality rates.

Conclusions
PMP has been effective in decreasing the recurrence rates of IH for both shorter and longer time periods. It is, however, associated with an increased incidence of seroma and chronic wound pain. No significant difference was found between the PMP and PSC groups for hematoma, surgical site infection, wound dehiscence, gastrointestinal complications, length of hospital stay, and operating time. The benefits of PMP largely outweigh the risk of complications and is beneficial for high-risk patient populations. There is a need for trials with extensive follow-up durations of 10-15 years to study the long-term benefits of mesh, and more studies with uniform reporting criteria are needed for accurately analyzing chronic wound pain outcomes. Furthermore, studies evaluating the efficacy of one mesh type over another are warranted.