Introduction
An incisional hernia is a frequent complication following abdominal surgery and is defined as any abdominal wall gap with or without a bulge in the area of a postoperative scar perceptible or palpable by clinical examination or imaging [
1,
2]. Bulging, discomfort or pain and cosmetic issues have a significant impact on the quality of life of patients [
3].
At a weighted mean follow-up of 23.7 months, 12.8% of the patients have developed an incisional hernia and the estimated risk to undergo an incisional hernia repair after a midline incision is 5.2% [
4]. In high-risk patients, the incidence increases to up to 30% [
5,
6]. The identified patient-related risk factors are obesity (BMI > 25 kg/m
2), the presence of an abdominal aortic aneurysm and congenital connective tissue disorders [
7‐
12]. Furthermore, postoperative wound complications are an independent risk factor for the development of an incisional hernia [
9].
With unacceptable high recurrence rates after incisional hernia repair, which ranges from 23.8% to 32% [
13‐
16], a preventive strategy has become the focus of scientific research. In high-risk patients, preventive mesh placement proves an effective tool in reducing incisional hernia rates [
6,
17,
18]. However, if the mesh is placed intra-peritoneally, delayed wound healing and increased pain after 6 week follow-up are observed [
19]. Thus, improving the technique of primary abdominal wall closure might be advantageous.
The technique for closure of the abdominal wall after a midline incision has evolved from using interrupted sutures to a continuous running monofilament suture technique [
20,
21]. Also, a small stitch suture technique, with a 4:1 suture length to wound length ratio, decreases the incidence of incisional hernia [
22,
23]. The European Hernia Society formulated a guideline for the closure of the abdominal wall incorporating the previous statements. This guideline makes a weak recommendation for abdominal wall closure with slow-absorbable sutures, which is adopted in standard clinical practice [
24]. Multiple meta-analyses report conflicting results regarding the preferable suture material, caused by substantial heterogeneity between studies and inclusion criteria [
4,
20,
21,
25].
Two meta-analyses suggested the use of slow-absorbable sutures based on a lower occurrence of suture sinuses and wound pain, but in the absence of a difference in incisional hernia occurrence [
21,
25]. The most recent meta-analysis recommended slow-absorbable sutures based on a decrease of incisional hernia compared to fast-absorbable sutures, but a separate comparison between only slow-absorbable and non-absorbable sutures was not performed [
20]. Lastly, Bosanquet et al. suggested in a meta-regression that suture material does not have influence on incisional hernia rate or on the occurrence of suture sinuses [
4].
Collagen metabolism plays a central role in the healing of the abdominal wall. Collagen maturation and collagen breakdown in particular are crucial. A decreased collagen I/III ratio is indicative of a low presence of mature collagen. Also increased MMP activity, responsible for collagen type I denaturation, is involved in abdominal hernia development [
26]. However, the healing process of the abdominal wall is still not fully understood. This unknown factor complicates the discussion regarding incisional hernia prevention. The aim of this study is to investigate the effect of different types of suture materials on the healing of the abdominal wall in a rat model. Giving insight to physiological and pathophysiological processes might lead to new starting points for preventive strategies.
Materials
A non-absorbable monofilament polypropylene suture (Prolene 4/0, Ethicon Inc; Johnson & Johnson, Somerville, NJ, USA), a slow-absorbable monofilament polydioxanone suture (PDS II 4/0, Ethicon Inc; Johnson & Johnson, Somerville, NJ, USA), and a fast-absorbable multifilament polyglactine suture (Vicryl 4/0, Ethicon Inc; Johnson & Johnson, Somerville, NJ, USA) were obtained commercially.
Study design
33 male Wistar rats between 240 and 260 g were acquired from a registered breeding facility (Envigo, Horst, the Netherlands), housed at the Maastricht University animal facilities and cared for according to local protocol. The animals were socially housed in filter-topped cages with a 12 h day-night cycle and had free access to food and water. Because of hormonal influences on wound healing by progesterone and oestrogen [
27] and a faster postoperative weight recovery of males compared to females [
28], only male rats were used in this experiment. The animals were randomly assigned to one of the three groups for abdominal wall closure of two separate midline incisions by either Prolene 4/0, PDS 4/0 or Vicryl 4/0. There were two time points for tissue evaluation, after 7 days and 21 days of follow-up.
Procedure
After a 1-week acclimatisation period, preoperative pain medication (buprenorphine 0.05 mg/kg) was administered. Anaesthesia was induced using 3–4% isoflurane and maintained with 2.5% isoflurane through inhalation of an air mixture. The animal was placed on a sterile field after the abdomen was shaved and the skin was disinfected with 2% iodine solution.
Via a midline incision of 6 cm in the skin, the abdominal wall was exposed. Two smaller full-thickness incisions of approximately 1 cm were made in the midline of the exposed abdominal wall, with a minimal distance of 2 cm between the two incisions. The two incisions were closed with either Prolene 4/0, PDS 4/0 or Vicryl 4/0 using a continuous suture technique; the skin was closed with Monocryl 4/0 (Ethicon Inc; Johnson & Johnson, Somerville, NJ, USA). Postoperatively, the animals were administrated fluid resuscitation and recovered under a heat lamp.
After 7 days of follow-up, one of the closed midline incisions was explanted and after 21 days of follow-up the second previously closed midline incision was explanted after killing, using the operative procedure as described above. Using the previous midline incision, randomly either the upper or lower abdominal wall incision was explanted equally distributed over the experimental groups. The explanted specimens were cut in half and one-half was preserved in liquid nitrogen and the other half in formaldehyde 4% for further analysis. The animals were killed by carbon dioxide overdose at the completion of the follow-up at 21 days.
Histology
Half of the explanted incision was fixed in 4% formaldehyde, dehydrated and embedded in paraffin. Subsequently, 4 µm thick tissue sections were cut and a haematoxylin–eosin staining was performed. An experienced animal pathologist, who was blinded to the group allocation, evaluated the stained sections microscopically. To compare inflammation and collagen deposition, granulocytes, macrophages, foreign body giant cells and collagen fibres were scored using a four-point semi-quantitative scoring system (not present, slightly present, moderately present or abundantly present) [
29‐
32].
RNA isolation and quantitative real time PCR
Total RNA was isolated from the snap-frozen abdominal wall specimen using TRI reagent (Sigma, NL). 750 ng DNAse-treated RNA was used to synthesise cDNA (SensiFAST™, cDNA synthesis kit, Bioline, London, UK). For qPCRs, a volume of 10 µl consisting of the cDNA equivalent of 2.5 ng total RNA, 1 × Absolute qPCR SYBR Green Fluorescein Mix (SensiFAST™ SYBR
® Hi-ROX Kit, Bioline, London, UK) and 0.15 μM of gene-specific primers (Sigma, NL) was used (Supplementary Table 1). The LightCycler
® 480 Instrument II (Roche Molecular Systems, Inc., was used to perform the qPCR. LinRegPCR software was used to establish gene expression levels. The geometric mean of two internal control genes, Rplp0 and beta-actin (Actb), was calculated and used as normalisation factor. In one sample in the Vicryl group, insufficient cDNA was available for analysis with Actb, resulting in unreliable data. Rplp0 showed no expression in one sample from the Prolene group. For both samples, one reliable housekeeping gene was available, which was used as normalisation factor. Relevant primers were identified from the literature and build using a primer designing tool (Primer-blast) [
33]; the sequences are reported in Table
1.
Table 1
All primers were tested for transcription of the intended gene
rplp0 (ribosomal protein lateral stalk subunit P0) | 190 | f | 55.00 | CCTCACCGAGATTAGGGACA |
| r | 45.00 | ATCGCTCAGGATTTCAATGG |
actb (actin, beta) | 297 | f | 55.00 | CCGCGAGTACAACCTTCTTG |
| r | 55.00 | CAGTTGGTGACAATGCCGTG |
il6 (Interleukin 6) | 246 | f | 57.14 | CTCTCCGCAAGAGACTTCCAG |
| r | 47.62 | TTCTGACAGTGCATCATCGCT |
nos2 (iNOS) | 234 | f | 52.38 | TAGTCAACTACAAGCCCCACG |
| r | 60 | GTGAGGAACTGGGGGAAACC |
cd86 (CD86) | 164 | f | 45.45 | AGACATGTGTAACCTGCACCAT |
| r | 55 | TACGAGCTCACTCGGGCTTA |
il10 (Interleukin 10) | 186 | f | 52.38 | CGACGCTGTCATCGATTTCTC |
| r | 60.00 | CAGTAGATGCCGGGTGGTTC |
clec10a (C-type lectin domain containing 10a) | 164 | f | 60.00 | GAGGCTTGAGCCAGAAGGTG |
| r | 52.38 | TGCTGAGCCGTTGTTCTTGAG |
mrc1 (mannose receptor C typ 1) | 212 | f | 60.00 | CCCGCTCCTCAAGACAATCC |
| r | 55.00 | AAATACGGTGACTGCCCACC |
cd163 (CD163) | 131 | f | 60 | CTCTGAAGCGACGACAGACC |
| r | 50 | ATGCCAACCCGAGGATTTCA |
tgfb1 (transforming growth factor-β) | 115 | f | 60.00 | GGCTGAACCAAGGAGACGGA |
| r | 55.00 | CCTCGACGTTTGGGACTGAT |
vegfa (vascular endothelial growth factor a) | 235 | f | 60 | AGAAGGGGAGCAGAAAGCCC |
| r | 47.83 | GATCCGCATGATCTGCATAGTGA |
angpt2 (angiopoietin 2) | 168 | f | 55 | CATGATGTCATCGCCCGACT |
| r | 52.38 | TCCATGTCACAGTAGGCCTTG |
nos3 (eNOS) | 139 | f | 52.38 | GAATGGAGAGAGCTTTGCAGC |
| r | 60 | CCGCCAAGAGGATACCAGTG |
col1a1 (collagen type 1 alpha 1 chain) | 237 | f | 60 | CTGACTGGAAGAGCGGAGAG |
| r | 55.00 | CAGGATCGGAACCTTCGCTT |
mmp1 (matrix metallopeptidase 1) | 144 | f | 55.00 | AAGGCCACTGGTGATCTTGC |
| r | 43.48 | GGTATTTCCAGACTGTTTCCACA |
fn1 (Fibronectin 1) | 165 | f | 63.16 | TCCCCTCCCAGAGAAGTGG |
| r | 43.48 | TTGGGGAAGCTCATCTGTCTTTT |
Statistical analysis
A sample-size calculation was performed in preparation of the experiment. A difference of 20% in inflammation on a histological level was considered relevant, with a variance of ± 16%. Alpha was chosen at 0.05 and with a power of 0.80, resulting in a needed group size of 11 animals per group.
All data were expressed as a median with range or mean with 95% confidence interval. Nonparametric tests were performed using a Kruskal–Wallis test. In case of significance, a Mann–Whitney U post hoc test was performed to identify specific differences between the groups. A Bonferroni correction was used to correct for multiple testing, so a p value of 0.017 (0.05/3) was considered significant. SPSS 23.0 for Mac (SPSS Inc., Chicago, IL, USA) was used for the statistical analysis.
Discussion
The technique for closure of the abdominal wall is evolving to prevent incisional hernia development. Both the use of a small bite technique shows a reduction in the incidence of incisional hernia [
22,
23], as well as mesh placement after a laparotomy prevents incisional hernia in high-risk patients [
6,
18]. Reports in the literature regarding the optimal suture material for the closure of the abdominal wall are conflicting and focus on clinical outcomes, such as incisional hernia, suture sinuses and wound pain [
4,
21,
25,
44].
The meta-analyses comparing non-absorbable sutures versus slow-absorbable sutures are confounded by factors in the study population; multiple types of incisions, emergency versus elective surgery and different suture materials or techniques are included. In the comparison between non-absorbable and slow-absorbable sutures, no differences are detected regarding the incisional hernia rate [
21,
25]. However, slow-absorbable sutures do result in less wound pain and suture sinuses. This experiment intends to provide a pathophysiological foundation for the choice in suture material, hypothesising that improved healing of the abdominal wall leads to a reduction of the incidence of incisional hernia.
A fast-absorbable, slow-absorbable and non-absorbable suture material was compared in a rat model. In general, the foreign body reaction to biomaterials is subdivided into four phases: protein absorption, cell recruitment and adhesion, foreign body giant cell formation, and finally extracellular matrix formation and fibrotic encapsulation [
45,
46]. In this experiment, the focus lies on the last three phases. Tissues were microscopically examined regarding the presence of different types of cells and collagen fibres. Consequently, the gene expression was determined using qPCR to explore vascularisation, fibroblast activity and macrophage polarisation.
The presence of macrophages was significantly higher in the Vicryl group after 21 days and simultaneously more foreign body giant cells were encountered compared to Prolene at 7 days and compared to PDS at 21 days. PDS and Vicryl are both absorbable sutures which rely on hydrolysis for degradation [
47,
48]. Polylactic and polyglycolid acids, polymers both present in Vicryl, depend on phagocytosis by macrophages and especially foreign body giant cells for complete absorption [
45,
49]. Vicryl elicits a stronger macrophage response in comparison to Prolene and PDS, which also results in more foreign body giant cells. The interaction between host and suture might play a role in its degradation and absorption. Although, in that case a similar finding would be expected in PDS. PDS differs from Vicryl in the polymer used and the fact that it is a monofilament suture in contrast to Vicryl which is braided [
48]. These factors contribute to the difference in absorption rate between PDS and Vicryl, which, in the case of the faster absorption rate of Vicryl, can negatively affect the occurrence of incisional hernia. This concurs with clinically based results from a previous report, in which slow-absorbable sutures were considered superior regarding the incidence of incisional hernia compared to fast-absorbable sutures [
20].
After determining the quantity of macrophages present, the polarisation between macrophage subtype 1 and 2 was evaluated using qPCR. Macrophage subtype 2 and its signalling play an essential role in liver regeneration, skeletal muscle healing and scar formation after injury of the skin [
50‐
52]. Therefore, it was hypothesised that the macrophage subtype 2, which modulates extracellular matrix and activates fibroblasts [
34‐
36], would have a positive effect on abdominal wall healing.
C
lec10a expression, typical for macrophage subtype 2, was significantly higher in the PDS group compared to the Prolene group. This suggests a more dominant presence of macrophage subtype 2 in the healing process. A striking number of genes (
il10,
mrc1,
cd163,
tgf-
b) typical for subtype 2 macrophages showed higher expression, although not significantly, when the abdominal wall was closed with PDS instead of Prolene after 7 days follow-up. The expression levels of the genes
mrc1 and
cd163 were higher in the presence of PDS compared to Vicryl and Prolene after 7 days, without reaching significance.
Nos2, typically expressed by the macrophage subtype 1, and
mmp1 showed a similar pattern. Although no significant differences could be detected, multiple genes typically expressed by subtype 2 macrophages were high in expression in the PDS group versus one subtype 1 specific gene. This could be interpreted as a distinct pattern, suggesting a dominant macrophage subtype 2 presence after abdominal wall closure with PDS. The effect disappears after 21 days of follow-up, which coincides with the progression of the healing process. The findings regarding macrophage polarisation concur with reports on scaffolds of different biomaterials, concluding a mainly anti-inflammatory macrophage polarisation in reaction to PDS (polydioxanone). Regarding polypropylene and polylactic acid, both the pro-inflammatory and anti-inflammatory macrophage responses were reported [
53].
The presence of different cell types in the analysed samples might result in various expression patterns, which could influence qPCR results. In addition, the macrophage polarisation is described as a spectrum rather than a black and white differentiation [
54‐
56]. This may cause heterogeneity in the qPCR data and might limit the ability to reach a statistical difference. This might also be an argument for a larger sample size than calculated for this experiment to reach adequate power.
Recording the mere presence of macrophages no longer suffices in evaluation of the healing process of the abdominal wall and novel insights in macrophage polarisation need to be taken into account. Subtyping and activation of macrophages provide additional information on the regenerative process taking place in the abdominal wall. The anti-inflammatory subtype 2 macrophage is associated with tissue regeneration and is, therefore, assumed to have a positive effect on the healing process [
34‐
36,
50‐
52]. The results of this experiment suggest a favourable macrophage response to PDS in comparison to Prolene and Vicryl, which in turn might benefit the regenerative capacity of the abdominal wall. This provides a new perspective on the dilemma of appropriate suture material in abdominal wall closure, which is of added value to the existing literature, which is mostly based on clinical outcomes.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.