A Systematic Review and Meta-Analysis of Synthetic Mesh Outcomes in Alloplastic Breast Reconstruction

Abstract

Background

Mesh implants are frequently employed in alloplastic breast reconstruction. Notably, no mesh to date has FDA approval for this indication. Several synthetic meshes have been introduced with heterogeneous properties and outcomes.

Objectives

This study aims to systematically review synthetic mesh use in alloplastic breast reconstruction, describe rates of short-term complications, and analyze these outcomes in reports comparing synthetic and biologic meshes. The authors hypothesized data from comparative and noncomparative studies would show no significant differences between synthetic and biological meshes.

Methods

The authors conducted a systematic literature review following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Thirty-one studies reporting the use of synthetic mesh and clinical outcomes were included. Eight studies directly comparing synthetic mesh and biological mesh were meta-analyzed for relative risk (RR). Nineteen noncomparative studies were analyzed for meta-rates. Outcomes, including seroma, infection, reoperation, and explant, were assessed on a per-breast basis. Resultant models were challenged for sensitivity and bias.

Results

Meta-analysis of comparative studies demonstrated no difference in the risk of infection with synthetic mesh (RR = 0.53; 95% CI [0.26-1.10]), but a reduced risk of reoperation (RR = 0.54; 95% CI [0.33-0.89]) or explant (RR = 0.43; 95% CI [0.21-0.87]). Meta-analysis of noncomparative studies demonstrated rates of seroma = 3%; 95% CI [1%-6%], infection = 4%; 95% CI [3%-6%], reoperation = 10%; 95% CI [7%-13%], and explant = 3%; 95% CI [2%-5%]).

Conclusions

Studies comparing synthetic and biologic meshes demonstrated noninferiority of synthetic in all outcomes assessed. Noncomparative studies demonstrated rates of seroma, infection, reoperation, and explant similar to literature values for biological mesh.

Level of Evidence: 2 (Risk)

Implant-based breast reconstruction after mastectomy is common, with over 100,000 cases per year in the United States alone.¹ With an evolving range of mastectomy indications, diverse patient populations, interdisciplinary collaboration, and a multitude of reconstructive techniques, reported outcomes are highly variable.^2-6 Approximately, 75% of reconstructive breast procedures are alloplastic, using tissue expanders and/or implants.^1,7

Currently, a major portion of reported alloplastic reconstructions include mesh support, with acellular dermal matrix (ADM)—a biologic mesh derived from cadaveric or animal dermis—being utilized in 52% of them between 2015 and 2022.¹ According to data available from the American Society of Plastic Surgeons’ yearly reports, its usage trended upward from 53% in 2015 to 61% in 2018, dropped to 43% in 2020, and rebounded to 50% in 2022. Emphasized by a recent paradigm shift toward prepectoral breast reconstruction, ADM holds the majority of the US market share and is the foundation of a 7 billion dollar industry.⁸ Although many publications have demonstrated effective use of various ADMs, concerns regarding costs and complications have been reported.^9-11 No mesh device has gained FDA-approved indication for breast reconstruction.

Limited previous studies, primarily conducted within Europe, have assessed outcomes in employment of various synthetic meshes, which are associated with lower costs compared with biologics. Although the literature often shows results comparable with those of ADM, data are heterogeneous and the molecular properties of both synthetics and biologics are diverse.^12,13 Despite many institutions reporting experiences with synthetic mesh in breast reconstruction, there is a paucity of data directly comparing synthetic with biologic mesh or with mesh-free reconstruction.

As a result, other investigations have attempted to systematically address the potential differences between these reconstructive approaches. In a network meta-analysis (NMA), Murphy et al found that, compared with mesh-free immediate reconstruction, ADM demonstrated significantly higher overall complication rates, whereas synthetics demonstrated insignificantly lower rates.¹⁴ In a systematic review, Makarewicz et al concluded that synthetic meshes produce equivalent or superior outcomes in terms of complications, quality of life measures, and aesthetic outcomes when compared with biologic meshes.¹⁵ In another NMA, Choi et al reported that synthetic meshes had lower rates of infection and seroma compared with ADM.¹⁶ Each of these reviews, however, was performed using outcomes solely from comparative studies, excluding critically important data from single-armed studies of synthetic and biologic meshes.

The present study aimed to fill this gap in data inclusion by collating current publications reporting outcomes with synthetic meshes in alloplastic breast reconstruction after mastectomy, including data from both comparative and noncomparative studies. The result is a truly comprehensive systematic review with meta-analysis of acute outcomes reported in studies comparing synthetic and biologic meshes, and meta-rates of acute outcomes reported in noncomparative studies using synthetics. The authors hypothesized that comparative studies would show no significant differences between synthetic mesh and biological mesh in rates of seroma, infection, reoperation, and explantation. Furthermore, the authors hypothesized that noncomparative studies would demonstrate rates of seroma, infection, reoperation, and explantation comparable with those reported for biologics.

METHODS

Studies through August 2022 were found and included according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines with independent review by 2 authors and conflicts solved through discussion (Supplemental Table 1). The PubMed, Medline, and Google Scholar databases were queried with multiple iterations of related search terms (Supplemental Figure 1). Citations of identified background literature and included studies were reviewed for identification of additional records.

For full-text assessment, inclusion criteria were clinical studies reporting use of a synthetic mesh in alloplastic breast reconstruction with clinical outcomes at minimum 1 month follow-up. Studies excluded were published prior to 2012, not in English language, without breast-level data, or containing data from fewer than 24 reconstructed breasts. “Breast-level data” refer to outcome data that are reported for each breast individually, rather than aggregated at the patient level. Studies from 2012 onward were included to represent contemporary literature, after the peak of ADM-focused publication.^17-19

Records obtained from the initial query were imported into Rayyan.²⁰ Prescreening exclusions were applied, and remaining records were screened by title and abstract. Additional studies were also sought out through citation auditing. Thereafter, full-text review and NIH quality assessment were conducted.²¹ For quantitative analysis, studies were divided into comparative and noncomparative, and comparative studies were further divided into those comparing synthetic with biological mesh and those comparing synthetic with mesh-free reconstruction. Only those comparing synthetic and biological mesh were assessed in meta-analysis. Characteristics of all synthetic and biological meshes discussed in this study are reported in Supplemental Table 2.

Outcomes analyzed were those deemed most likely to be affected by mesh properties: seroma, infection, reoperation, and explantation. Major complications, defined as reoperation or hospitalization, were also analyzed. True randomized clinical trials are rare and potential confounders are broad and heterogeneous. All comparative studies were included in analysis. By the nature of meta-analysis statistics, confounders could not be included in models. In lieu of this, comparative studies were examined for surgical techniques and patient factors with any differences noted. Outcomes were assessed on a per-breast basis.

Statistical Methods

Comparative meta-analyses employed the Mantel–Haenszel (MH) method without continuity correction for meta-calculation of risk ratio. The analysis of seroma rate was not included due to less than half of comparative studies including its rate in their outcomes and poorly defined parameters for its diagnosis. One study was removed from the analysis of infection secondary to incomplete reporting. The analysis of noncomparative studies for meta-proportion of outcomes used a generalized linear mixed model with logit transformation and continuity correction. The random effects model was selected for analyses demonstrating heterogeneity (P ≥ .05).

To assess for impact of small-study effects and publication bias, models were challenged with leave-one-out analysis and funnel plots were generated. To assess sensitivity of noncomparative study analyses to these potential biases, trim-fill models were generated. All statistical analyses were performed on R Version 4.2.2 (R Foundation, Austria) with tools available in CRAN packages “meta” and “metafor.”

RESULTS

Overview of Study Selection

Initially, the systematic review identified 617 records and 3 additional records from citation auditing. After removing 149 duplicates and pre-2010 records, 471 records were screened by title and abstract. Subsequently, 58 full-text reports were sought for retrieval, with 1 report not retrieved. From these, 57 reports were assessed for eligibility based on the predefined inclusion criteria, leading to the exclusion of 26 reports for reasons such as incorrect outcomes, incorrect procedures, repeat data, poor NIH quality, and small sample size. Ultimately, 31 studies met the criteria for inclusion in the qualitative review. Of these, 27 studies provided sufficient data for inclusion in the quantitative meta-analysis, with 8 of 12 comparative and all 19 noncomparative studies making up the final analyzed set. Four comparative studies were excluded as they did not contain the appropriate comparison between synthetic and biologic mesh (Figure 1).

Figure 1.

PRISMA diagram.

Qualitative Review

Comparison With Biologic Mesh

Eight studies comparing synthetic and biologic mesh were identified with follow-up ranging from 3 to 23 months and cohort size ranging from 48 to 2655 breasts (Table 1). Most described subpectoral plane (7/8), 5 included only immediate reconstructions, and there was an approximately even distribution of direct-implant and tissue-expander implantation (Supplemental Table 3). The largest study, Potter et al,²² included prospectively collected 3-month cross-sectional follow-up of 417 cases employing synthetic mesh and 1999 cases employing biologic mesh in various techniques from 81 institutions across the United Kingdom. The authors noted BMI, smoking, and previous radiation as predictors of complication. No association was found between complication and mesh employment or mesh type. All additional studies included mean follow-up of at least 6 months with notably smaller cohorts.

Table 1.

Study and Patient Characteristics for Included Comparative Studies

First author	Year	Journal	Country	Study type	Patients	Breasts	Institutions	Mean/median age	Mean/median BMI	Mean/median follow-up	Synthetic mesh	Biologic mesh	Seroma	Infection	Major complication	Explant
					n	n	n	yr	kg/m2	mo	brand; n breasts		n synthetic/n biologic
Gao	2022	Breast Cancer	China	Retrospective	116	176	1	41	21	17	TiLoop; 55	Surgisis; 121	0/6	1/4	0/5	0/5
Hansson	2021	J Surg Oncol.	Sweden	Prospective	24	48	1	43	23	16	TIGR; 24	Veritas; 24	9/2	4/0	6/2	4/1
Schuler	2021	In Vivo	Germany	Retrospective	157	188	1	51	26	12	Seragyn; 54 TiLoop; 94	Strattice; 40	11/11	9/2	20/11	8/11
Levy	2020	Ann Plast Surg.	USA	Retrospective	169	304	1	49	24	15	Phasix; 112	Allomax; 192	1/6	12/34	14/31	12/19
Hallberg	2019	Eur J Surg Oncol.	Sweden	Prospective	71	116	1	46	23	23	TIGR; 65	Surgisis; 116	NA	NA	3/12	2/12
Quah	2019	Plast Reconstr Surg GO	Australia	Retrospective	150	216	1	45	NA	12	TiLoop; 179	Veritas; 37	3/19	5/7	14/13	3/3
Potter	2019	Lancet Oncol.	UK	Prospective	2108	2655	81	49	25	3	Various; 417	Various; 1999	NA	61/251	49/193	24/90
Gschwantler-Kaulich	2016	Eur J Surg Oncol.	Austria	Prospective	48	65	4	48	23	6	TiLoop; 35	Protexa; 30	1/3	1/3	4/6	2/7

mo, month; NS, not applicable; yr, year.

Four studies described experience with TiLOOP (PFM Medical, Cologne, Germany), a permanent titanium-coated polypropylene implant with a 3-dimensional (3D) “bra” shape. Gschwantler-Kaulich et al²³ prospectively compared with Protexa (Tecnoss, Giaveno, Italy) ADM finding no significant difference in outcomes. Quah et al²⁴ retrospectively compared with Veritas (Baxter Inc., Deerfield, IL) ADM finding significantly lower complication rate in the synthetic mesh cohort. Schuler et al²⁵ retrospectively compared with Seragyn (SERAG-WEISSNER, Naila, Germany), an absorbable polydioxanone mesh, and Strattice (Allergan, Dublin, Ireland) ADM finding no difference in complications between synthetic meshes and a significantly higher complication rate when employing the biologic mesh. Gao et al²⁶ retrospectively compared with Surgisis (Cook Biotech, West Lafayette, IN) ADM demonstrating no significant difference in complications between cohorts.

Two studies, from 1 institution, described experience with TIGR (Novus, Uppsala, Sweden), an absorbable mesh woven from 2 co-polymers of different properties. Hallberg et al²⁷ prospectively compared with Surgisis ADM finding a lower explantation rate with the synthetic mesh. In a follow-up study, Hansson et al²⁸ compared with Veritas ADM in bilateral reconstructions using the alternative mesh in each breast of each patient. This study demonstrated significantly lower rate of seroma in synthetic mesh reconstructions with other outcomes not reaching significance.

Although the slowly absorbed synthetic polymer poly-4-hydroxybutyrate meshes like GalaFlex (Galatea Surgical, Lexington, MA) and Phasix (Becton Dickinson, Franklin Lakes, NJ) have gained some popularity for breast surgery in the United States, just 1 identified study examined its use in breast reconstruction. In a 169-patient series, Levy et al²⁹ retrospectively compared Phasix with AlloMax (Bard/Davol, Warwick, RI) human-derived ADM in immediate subpectoral reconstructions, describing a more favorable complication profile with the synthetic.

Comparison With Mesh-Free Reconstruction

The previously discussed landmark 2019 study by Potter et al²² compared synthetic, biologic, and mesh-free reconstructions >2000 patients across 81 UK institutions. No significant difference in complications was identified between techniques at 3-month cross-sectional follow-up. Four additional studies comparing synthetic mesh with mesh-free reconstructions were identified.

Dieterich et al³⁰ prospectively compared TiLOOP with implant-alone reconstructions in 98 breasts finding no significant difference in complications between cohorts. Ganz et al³¹ retrospectively compared 115 partial subpectoral reconstructions employing Vicryl (Ethicon, Raritan, NJ), an absorbable polyglactin mesh, with 46 complete subpectoral reconstructions also finding no significant differences.

Baldelli et al³² retrospectively compared outcomes of immediate subpectoral reconstruction in 70 cases using Surgimesh PET (Apside, Boulogne-Billancourt, France), a nonabsorbable polyester monofilament, vs 136 cases with standard fascial coverage. They demonstrated no difference in early postoperative complication rate between cohorts, with radiation therapy being the only significant predictor. Most recently, Patzelt et al³³ prospectively compared Seragyn with autologous dermal flap in 128 prophylactic mastectomies finding no difference in complications between cohorts and noting that autologous flaps may be preferrable in large, ptotic breasts.

Noncomparative Studies

Nineteen studies were identified reporting experience employing synthetic mesh without a comparator cohort. Eighteen of the 19 demonstrated results of at least 6 months mean follow-up. Study and patient characteristics are reported (Table 2). Eleven used TiLOOP, 4 TIGR, 2 Vicryl, 1 Seragyn, and 1 ULTRAPRO (Ethicon), a partially absorbable poliglecaprone mesh.

Table 2.

Study and Patient Characteristics for Included Noncomparative Studies

First author	Year	Journal	Country	Study type	Patients	Breasts	Institutions	Mean/median age	Mean/median BMI	Mean/median follow-up	Mesh	Seroma	Infection	Major complication	Explant
					n	n	n	yr	kg/m2	mo	Brand	n	n	n	N
Michno	2022	Surg Oncol.	Germany	Prospective	258	345	8	49	23	2a	TiLoop	22	15	88	0
Nguyen-Strauli	2022	J Plast Reconstr Aesthet Surg.	Switzerland	Retrospective	43	63	1	44	22	6	TiLoop	8	3	8	2
Wow	2022	Cancer (Basel)	Poland	Retrospective	170	232	3	50	22	20	TIGR	63	14	24	19
Ng	2021	ANZ J Surg.	Australia	Retrospective	80	109	1	49	23	23	TiLoop	11	8	9	9
Gentile	2021	Bioact Mater.	Italy	Retrospective	276	328	1	54	27	44	TiLoop	NA	24	43	15
Faulkner	2020	Plast Reconstr Surg.	US	Retrospective	227	376	1	52	NA	12	Vicryl	4	8	83	17
Bonomi	2019	Ann Plast Surg.	Italy	Retrospective	56	62	1	47	24	20	Vicryl	1	2	4	1
Casella	2019	Plast Reconstr Surg.	Italy	Retrospective	179	250	1	56	23	39	TiLoop	0	2	6	3
Casella	2019	J Plast Reconstr Aesthet Surg.	Italy	Prospective	187	237	1	56	25	37	TiLoop	3	7	16	9
Machleidt	2018	Arch Gynecol Obstet.	Germany	Retrospective	119	148	4	49	24	7	Seragyn	38	21	24	17
Casella	2018	Breast J.	Italy	Prospective	46	92	1	43	28	24	TiLoop	0	0	2	9
Hallberg	2018	J Plast Surg Hand Surg.	Sweden	Prospective	49	65	1	46	23	17	TIGR	2	1	7	2
Pompei	2018	Clin Plast Surg.	Italy	Retrospective	49	60	1	51	NA	12	TIGR	2	1	7	2
Pukancsik	2017	Eur J Surg Oncol.	Hungary	Retrospective	102	174	1	43	23	23	ULTRAPRO	9	5	20	4
Casella	2015	Plast Reconstr Surg GO	Italy	Prospective	25	25	1	60	22	12	TiLoop	0	4	5	0
Casella	2014	Eur J Plast Surg.	Italy	Prospective	63	73	1	36	23	12	TiLoop	0	2	1	1
Dietrich	2013	Plast Reconstr Surg.	Germany	Retrospective	187	231	8	47	23	14	TiLoop	11	14	31	18
Becker	2013	Aesthet Plast Surg.	US	Retrospective	52	70	1	54	NA	17	TIGR	2	4	8	0
Dietrich	2012	Eur J Surg Oncol.	Germany	Retrospective	42	45	2	NA	22	20	TiLoop	2	0	2	2

NA, not applicable; US, United States. ^aCross-sectional follow-up.

Dieterich et al³⁴ are the first to report experience with TiLOOP, in 45 breasts, and note benefits in implant stabilization but indicate ADM may be preferrable with poor soft-tissue coverage. A 2013 multi-institutional follow-up³⁵ of 231 breasts concludes “acceptable complication profile,” and suggests indication in only primary cases.

Five studies utilizing TiLOOP were included from Casella et al. In 2014³⁶ and 2015,³⁷ they respectively demonstrated favorable results at 1 year with direct-to-implant in 73 breasts and tissue expander in 25 breasts. Following this, in 2018³⁸ and 2019,³⁹ they described success in 92 prophylactic cases and 237 tissue expanders, respectively, both with favorable complication rates and aesthetic outcomes at long-term follow-up. An additional study in 2019⁴⁰ noted low complication rates and high patient satisfaction in 250 direct-to-implant prepectoral reconstructions.

Gentile et al⁴¹ reported experience and histological data with TiLOOP in 328 prepectoral direct-to-implant and tissue-expander reconstructions demonstrating complete mesh integration and favorable complication rates with both techniques. Ng et al⁴² described a similar study of 109 breasts and noted improved patient satisfaction with prepectoral reconstructions. In a smaller study of 63 prepectoral reconstructions, Nguyen-Sträuli et al⁴³ noted convenience in the 3D shape of TiLOOP, but warned of morbidity and complication associated with inappropriate patient selection. Most recently, Michno et al⁴⁴ mirrored this in 345 reconstructions, describing “acceptable complication rates” and demonstrating BMI and smoking as risk factors.

TIGR mesh was the second most common. Becker and Lind⁴⁵ described experience in 77 primary and secondary reconstructions, demonstrating low complication rates and concluding it a viable alternative to ADM. Pompei et al⁴⁶ and Hallberg et al⁴⁷ provided experience with long-term follow-up in 60 and 65 breasts, respectively, both demonstrating acceptable complication rates with similar conclusions. Most recently, Wow et al reported results in 232 prepectoral and subpectoral reconstructions, with low complication rates in both, concluding it be safe and allow good cosmetic results.⁴⁸

Two noncomparative studies reported experiences with Vicryl. Bonomi et al⁴⁹ described reconstruction after 62 wise-pattern skin reducing mastectomies with a moderate complication profile and favorable aesthetic outcomes. Faulkner et al⁵⁰ demonstrated results from 376 reconstructions over a 7-year period with very low complication rates and noted that 1.2 million US dollars were saved using Vicryl over ADM.

Two studies employed other meshes. Pukancsik et al⁵¹ reported low complications with ULTRAPRO in 174 direct-to-implant reconstructions concluding it to be a safe cheaper alternative and Machleidt et al⁵² noted complication rates with Seragyn comparable with ADM in 148 reconstructions of various techniques.

Meta-Analysis

Comparative Studies

Studies reporting infection, major complication (reoperation or hospitalization), and explantation rate showed significant heterogeneity of outcomes (I² = 69%-74%), thus the random effects model was utilized. The risk of infection was not significantly different between synthetic and biologic mesh cohorts with relative risk (RR) = 0.53, 95% CI (0.26-1.10). The risk of major complication was found to be significantly higher in biologic mesh with RR = 0.54, 95% CI (0.33-0.89) in favor of synthetic mesh. The risk of explantation was also found to be significantly higher in biologic mesh with RR = 0.43, 95% CI (0.21-0.87) in favor of synthetic mesh (Figure 2).

Figure 2.

Meta-analysis of comparative studies. Random effects models were chosen for all secondary to heterogeneity. One study was excluded from model for infection due to no reporting.

Of note, the largest study, Potter et al,²² demonstrated contradictory results with no significant difference between synthetic and biologic meshes for any outcomes. Leave-one-out model challenges redacting this study reduced heterogeneity in analyses resulting in shortened confidence intervals and demonstrating RRs in favor of synthetic mesh for all 3 outcomes. Further descriptions of comparative model challenges are available as supplementary content (Supplemental Figure 2).

Noncomparative Studies

Meta-proportions of seroma, infection, major complication, and explantation were calculated from the included noncomparative studies. These studies also showed evidence of heterogeneity (I² = 52%-89%), thus random effects model was again employed. Reported rates of seroma ranged from 0% to 26% yielding a meta-rate of 3% and 95% CI (1%-6%). The rates of infection ranged from 0% to 16% yielding a meta-rate of 4% and 95% CI (3%-6%). Major complication ranged from 1% to 26% yielding a meta-rate of 10% and 95% CI (7%-13%). Explantation rate ranged from 0% to 11% yielding a meta-rate of 3% and 95% CI (2%-5%; Figure 3).

Figure 3.

Meta-analysis of noncomparative studies. Random effects models were chosen for all secondary to heterogeneity. One study was excluded from model for seroma due to no reporting.

Constructed funnel plots showed leftward skews for all noncomparative analyses. This demonstrates that studies with lower standard error (larger cohorts) generally reported higher rates of complication. This may be an indicator of small-study bias and reflect inappropriately low complication rates described by resulting analysis (Figure 4).

Figure 4.

Funnel plots for noncomparative studies. A leftward skew indicates that smaller studies (with higher standard errors) tend to report lower rates of the outcome assessed compared with larger studies.

Model sensitivity was assessed by trim-fill, a statistical method in which duplicate mirror studies are generated for reports with high standard error or with results outside of the confidence interval.⁵³ The rates of all analyzed outcomes increased in these models. The reported infection rates were relatively insensitive to bias (trim-fill meta-rate = 6%; 95% CI [4%-9%]), whereas other outcomes were notably sensitive. Trim-fill meta-rate for seroma was 12%; 95% CI (6%-24%), that for major complication was 17%; 95% CI (12%-24%), and for explantation was 6%; 95% CI (4%-9%; Supplemental Figure 3).

DISCUSSION

Various meshes of differing properties have emerged as potential cost-effective, nonbiologic alternatives to ADM in breast reconstruction. This study is the first truly comprehensive description of various synthetic meshes utilized in breast reconstructions and meta-analysis of their available data. We describe a heterogenous collection of studies cataloging 7 employed synthetics: TiLOOP, TIGR, Vicryl, Seragyn, ULTRAPRO, Phasix, and SurgiMesh. Importantly, this analysis was done using data from both comparative and single-armed studies of synthetic and biologic meshes.

This is particularly relevant given the 2023 meta-analysis by Murphy et al, which claims to be the first NMA to compare human ADM, xenograft ADM, and synthetic mesh with no ADM or mesh for implant-based reconstruction. In their study, they concluded that ADM is associated with a significantly higher complication risk and no significant difference in capsular contracture compared with mesh-free reconstructions.¹⁴ Although that study found no differences between outcomes with synthetic mesh vs mesh-free, the paucity of reports reduces power of the comparison. An additional recent systematic review by Makarewicz et al claims to be the first to comprehensively evaluate all publications comparing biologic and synthetic meshes in implant-based reconstruction, and concludes synthetic meshes appear to produce equivalent or superior outcomes when compared with biologic meshes.¹⁵ In another recent study, Choi et al provide an NMA in this space, this time comparing between 4 categories of mesh using 2-armed studies: ADM, absorbable synthetic mesh, nonabsorbable synthetic mesh, and no mesh.¹⁶ According to their analysis, synthetic meshes had significantly lower rates of infection and seroma compared with ADM.

Although each of these studies make great contributions to the body of literature on this topic, each of their analyses are limited to comparative studies only, without consideration for single-armed publications. In addition, a few of them are lacking in a complete description of available meshes and individual study outcomes.^14,16 The inclusion of a large body of additional data from noncomparative studies enhances the generalizability and power of the analysis, providing a more complete statistical comparison.

As a result, the present analysis represents the first comprehensive analysis comparing synthetic and biologic meshes in implant-based reconstruction for both comparative and noncomparative studies. Another key point of differentiation is this study's meta-analysis approach for comparative studies, namely the MH method vs NMA. Although NMA is capable of comparing multiple interventions using direct and indirect evidence across a network of studies, MH is particularly well-suited at generating in depth analysis that is focused on 2 interventions with respect to binary outcomes, especially in the context of sparse data. This makes it the appropriate choice for robust, pairwise comparisons between biological and synthetic meshes alone.^54,55

Qualitative assessment of publications in this present study showed overall support of synthetic meshes employed, with no study concluding employment of a specific synthetic to be unviable and no comparative study demonstrating inferiority of synthetic to biologic mesh. Further, meta-analysis of comparative studies demonstrated no significant difference in infection rate and superiority over biologic mesh in rates of major complication or explantation, whereas noncomparative studies demonstrated generally favorable rates of acute complications. The results of this analysis lend conclusions of safety and efficacy to synthetics currently described.

As study inclusion was broad, the heterogeneity of outcomes is not unexpected. Alloplastic breast reconstruction describes a collection of techniques performed after variable mastectomies for a range of indications in a diverse population with often suboptimal comorbidity profiles. Fortunately, technique and timing were relatively homogenous between comparative studies. Additionally, reports represented a range of follow-up (3-44 months), and reporting of complications such as seroma and infection can vary subjectively. As 95% of noncomparative studies described mean follow-up of at least 6 months and 84% of at least 12 months, outcomes should include radiation effects and both first and second stages when applicable. This analysis does not aim to describe selection factors for mesh-based alloplastic reconstruction and cannot include such controls. Although this limits generalizability of the results herein to specific practices, meta-analysis of 3594 comparative cases and 2985 noncomparative cases should comprehensively describe ideal results in a broadly inclusive patient population.

The analysis of comparative studies showed superiority of synthetic to biologic mesh in risk of major complication and explantation. However, in a rigorous study conducted by Potter et al²² to mirror a true randomized clinical trial, and found no significant difference in 3-month complication rates between synthetic and biologic mesh. One may thus not necessarily conclude that synthetic mesh is superior, but that included studies unequivocally describe noninferiority. One limitation worth noting alongside these findings, however, is that some complications associated with mesh and implants may manifest after longer periods of time, even after 2 years. As an additional caveat for surgeons in the United States, synthetic meshes described herein may not be available at many institutions and biologics described in comparative trials are not those most frequently used. However, no mesh has received FDA approval for the indication of breast reconstruction.

In noncomparative studies, ranges and meta-descriptions of seroma, infection, major complication, and explantation were low overall and comparable with or below those reported for ADM.^56-58 Of note, TiLOOP was the most-employed mesh device in included reports. In addition, as a nonabsorbable titanium-coated mesh, its molecular and biologic properties are inherently different from ADM and resorbable synthetic meshes. As a result, a potential limitation in utilizing these TiLOOP studies is the possibility of double counting of patients, as the studies were conducted by the same author during overlapping time periods.^36-40 Simultaneously, the higher volume of TiLOOP studies means that its outcomes could have a disproportionate influence on the findings regarding synthetic meshes as a whole. Further research is needed to evaluate other synthetic mesh products more thoroughly.

Additionally, with funnel plots demonstrating strong leftward skews, there is risk of small-study and publication bias in this analysis.^59,60 Constructed trim-fill models demonstrated notable sensitivity to these effects for rates of seroma, major complication, and explantation. Resulting meta-rates demonstrated confidence intervals of more than double those of the overall analysis. Although conclusions cannot be made as to if this represents a “truer” estimation, these pseudo-adjusted rates are more comparable with the broader literature utilizing biological meshes, the landmark Potter et al, and with our institutional experience.²² Meta-rates described may represent idealistic results, and ranges of reported complications may yield a more reliable assessment.

CONCLUSIONS

This reports a contemporary systematic review and meta-analysis of short-term complications associated with synthetic mesh use in alloplastic breast reconstruction. The results indicate comparative outcomes demonstrating, at minimum, noninferiority of synthetic mesh when compared with biologic and noncomparative outcomes demonstrating generally low and favorable complication rates. With demonstrated potential for bias in outcomes reported, the models presented may underestimate true complication rates. Further research should examine patient-reported, aesthetic, and long-term outcomes. Additional studies comparing synthetic and biologic meshes, particularly those widely used in the United States, are necessary for advancement of the field.

Supplemental Material

This article contains supplemental material located online at https://doi.org/10.1093/asjof/ojae066.

Disclosures

The authors declared no potential conflicts of interest with respect to the research, authorship, and publication of this article.

Funding

The authors received no financial support for the research, authorship, and publication of this article, including payment of the article processing charge.