Cellular and molecular regulation of fibrotic postoperative abdominal adhesions.

Anti-adhesion therapy is inherently preventative, as effective medical treatments for established adhesions are lacking and surgical adhesiolysis carries a substantial risk of adhesion reformation. Early prevention strategies focused primarily on patients undergoing laparotomy, reflecting both the predominance of open surgery at the time and its greater associated peritoneal trauma. Foundational approaches emphasized refinements in surgical technique, principles that remain widely recommended today, including gentle tissue handling, meticulous hemostasis, minimization of tissue desiccation and operative duration, thorough peritoneal lavage, and increased adoption of minimally invasive approaches ( 31 , 41 , 122 - 124 ). Early interventional efforts also sought to restore the integrity of the peritoneal barrier through routine peritoneal closure. However, this strategy failed to demonstrate consistent benefit ( 4 ). Additional technical modifications, including cooling the abdomen, supplementation of CO 2 pneumoperitoneum with nitric oxide, and combinations of multiple protective measures, have shown variable but sometimes promising reductions in adhesion formation, though reproducibility and clinical translation have been inconsistent ( 122 , 125 - 127 ). Adjunctive strategies to prevent postoperative adhesions have included pharmacologic agents, peritoneal instillates, and, most prominently, physical adhesion barriers designed to transiently separate injured peritoneal surfaces during the critical 3-to-5-day window of mesothelial repair ( 11 , 32 , 128 , 129 ). Oxidized regenerated cellulose (ORC; Interceed ® , Ethicon, Johnson & Johnson, New Brunswick, NJ USA) was the first FDA-approved adhesion barrier in the United States, followed by hyaluronate-carboxymethylcellulose (HA/CMC; Seprafilm ® , Baxter Healthcare, Deerfield, IL USA), and expanded polytetrafluoroethylene (ePTFE; Gore-Tex ® , W. L. Gore & Associates, Newark, DE USA). Among these materials, HA/CMC has the most robust clinical evidence, with randomized trials and meta-analyses demonstrating reductions in adhesion incidence and severity ( 11 , 130 ). Some studies have additionally reported decreased rates of reoperation for adhesive SBO ( 11 ) and potential improvements in fertility and chronic pain outcomes ( 131 ). However, these latter benefits are supported by limited data, and cost-effectiveness analyses have yielded mixed conclusions ( 132 , 133 ). Several practical limitations have restricted the widespread use of adhesion barriers. These include difficulty conforming solid films to complex peritoneal geometries, limited applicability during laparoscopic procedures, dependence on complete hemostasis for efficacy (particularly for ORC), the need for fixation and subsequent removal or permanent implantation of certain materials (e.g. ePTFE), and increased upfront costs in the context of uncertain long-term cost-benefit ratios ( 129 , 132 , 133 ). Although adhesion barriers are not consistently associated with major adverse effects, reports of complications such as abscess formation, fistula development, or anastomotic leakage underscore the importance of careful patient and procedural selection rather than routine use ( 134 ). Recent meta-analyses have not demonstrated a strong signal for increased patient harm; nevertheless, equivocal clinical benefit, variable cost-effectiveness, and the absence of a clearly superior or mechanism-targeted product continue to limit their widespread adoption ( 134 ). Iteration on prophylactic barrier design has continued following the FDA approval of ORC, HA/CMC, and ePTFF with a focus on preventing macrophage aggregation. With few exceptions, these studies remain mechanism-agnostic, and many are limited to in vitro findings ( 19 ). Collectively, the variable efficacy of physical adhesion barriers underscores a fundamental limitation of strategies that act solely at the tissue interface without directly modulating the cellular programs governing mesothelial repair, inflammation, and fibroblast activation. These shortcomings highlight the need for mechanism-based approaches that target the molecular determinants distinguishing adaptive peritoneal healing from pathologic fibrotic adhesion formation. Early interventions to prevent adhesion formation focused on restoring the coagulation cascade to homeostasis, a strategy that must balance adhesive clot dissolution with hemostasis. Initial attempts to lyse the fibrin clot utilizing general peptidases were superseded in the 1950s by thrombolytic agents, including streptokinase, urokinase, plasmin preparations, and recombinant human tissue plasminogen activators (r-PAs). Of these methods, r-PAs show the greatest efficacy, degrading fibrin clots and reducing adhesion formation without compromising hemostasis or wound healing in both preclinical and patient settings ( 22 ). More recent approaches have started to modulate endogenous fibrinolytic machinery instead of supplementing it. Application of potato tuber carboxypeptidase inhibitor (PTCI), an inhibitor of thrombin-activated fibrinolysis inhibitor (TAFI), resulted in significant reductions in adhesive burden, fibrin content, and inflammation in a Sprague–Dawley cecal abrasion model ( 135 ). However, concerns over safety, dosage, and treatment duration remain with profibrinolytic approaches due to the risks of hemostatic instability. Efforts targeting cell-cell communication have addressed a wide range of inflammatory and MMT mediators to reduce adhesion formation. Although limited small-molecule approaches have been adopted to target acute inflammation, Tsai et al., showed inhibition of Gr-1 and stimulation of MCP-1 limits NETosis while attracting monocytes to the site of injury, reducing adhesion formation ( 56 ). The identification and perturbation of anti-MMT targets has been substantially more diverse, and includes anti-mTOR, MEK/ERK, HIF, TGFβ, AMPK, MSLN, and IL-6 small molecules and antibodies ( 19 ). The most recent strategies for adhesion prevention have coupled mechanistic approaches with novel application methods that surpass the limitations of prophylactic barriers. As discussed throughout this review, JUN-targeted therapies demonstrate great promise in reducing adhesion fibrosis, with novel developments being made by incorporating the T-4225 inhibitor into a slow-release hydrogel for one-time application ( 25 ). Integration of stem-cell-based therapies with hydrogel delivery to modulate the adhesion microenvironment has also proved therapeutically beneficial ( 114 ). These one-time-application prophylactic methods combine operational ease-of-use with well-established biological mechanisms, resulting in robust platforms for adhesiogenic amelioration ( 19 ). Future endeavors to prevent adhesion formation must either find a way to address the multifactorial mechanisms of adhesion initiation or look downstream to target maturation of the adhesive ECM. GPCM dispersal ( 21 ), coupled with inhibition of mesothelial cell activation, may be sufficient to prevent immune cell aggregates and minimize the migration of peritoneal cells into any developing clots, keeping the provisional adhesion decellularized. These therapies would rely on a combination of therapeutics, delivered at the time of surgery to prevent initial adhesions from forming. Alternatively, inhibiting the establishment of large fibrillar ECM networks, regardless of initial tissue adhesion, may be sufficient to maintain a weak provisional matrix that becomes dislodged with stress. This therapeutic strategy could be accomplished by targeting cell fate determination, similar to the JUN inhibition trial, or by disrupting ECM deposition and crosslinking mechanisms ( 25 ). Due to our limited knowledge of the adhesive ECM, the latter strategy will require additional characterization of adhesiolysis tissue to identify distinguishing profibrotic factors. While these long-term therapies could be delivered via slow-release hydrogels at the time of surgery, they also are more amenable to other delivery strategies since the effects do not need to be acute.

The peritoneum is a thin, highly specialized serous membrane that lines the abdominal cavity and envelopes intra-abdominal organs. It consists of a parietal layer lining the abdominal wall and a visceral layer covering the surfaces of organs ( 40 ), together forming a continuous, highly dynamic interface. Extensive folding and reflection of the peritoneum creates mesenteries, ligaments, and omental structures that suspend and stabilize abdominal organs while preserving mobility. The peritoneal cavity itself represents a potential space containing a small volume of lubricating fluid that permits near-frictionless organ movement under physiological conditions (i.e., in the absence of adhesions) ( 40 , 41 ). This structural organization is essential for normal abdominal function but also renders the peritoneum uniquely vulnerable to injury. Disruption of the peritoneal surface, whether from surgical manipulation, ischemia, or inflammation, initiates a reparative response that can deviate toward fibrosis and ultimately result in postoperative abdominal adhesion formation. Postoperative adhesions arise predominantly from injury to the visceral peritoneum, where extensive surface area, organ mobility, and close apposition of organs promote persistent fibrin bridges and aberrant fibrotic tissue connections ( 24 , 41 ). This risk of adhesion formation is, therefore, closely linked not only to the severity of peritoneal injury, but to its anatomic location and spatial context. The peritoneum itself is comprised of three layers ( 42 ). The mesothelial cell monolayer which lines the cavity has a central role in maintaining peritoneal homeostasis by directing membrane lubrication, immune modulation, metabolic regulation, and barrier function ( 43 ). At homeostasis, peritoneal mesothelial cells (PMCs) assume an epithelial morphology which varies from squamous to cuboidal throughout the cavity, attaching to one another via tight junctions to modulate peritoneal solute transport ( 44 ). To lubricate the peritoneal cavity, PMCs construct a hyaluronic acid (HA)-rich glycocalyx along their ciliated apical face to reduce friction ( 45 , 46 ). PMCs anchor to a conventional basal lamina comprised of collagen IV (COL4) and laminins ( 43 ), which in turn is connected to the underlying parenchyma by a stroma comprised primarily of collagen I (COL1) and fibronectin (FN) and bisected by an elastic lamina ( 42 ). This submesothelial stroma is populated by resident fibroblasts which appear largely quiescent at homeostasis, expressing transient levels of COL1 to maintain the interstitium ( 42 , 47 ). Development of the peritoneum begins when invagination of the lateral plate mesoderm (LPM) forms the coelomic cavity and becomes lined with coelomic epithelium ( 48 ). Lineage tracing in mice has shown that the coelomic epithelium will develop into mesothelin-positive (MSLN + ) cells and serve as a common lineage for the mesothelium, fibroblasts, and smooth muscle cells throughout the abdominal cavity. Remarkably, these MSLN + PMCs remain a post-natal source of smooth muscle cells (SMCs), participating in visceral muscle regeneration following irradiation, although the duration of this functionality is unclear ( 49 ). Peritoneal fluid assists in regulating the peritoneal microenvironment. Comprised of cellular exudate enriched in glycosaminoglycans and carbohydrates, peritoneal fluid undergoes regular turnover and is circulated throughout the peritoneal cavity by forces applied via respiration and gravity, a requirement for effective resident immune cell surveillance ( 21 , 42 ). Cavity-resident immune cells include large and small peritoneal cavity macrophages, dendritic cells, and plasma cells, which are involved, to varying degrees, in the formation of abdominal adhesions ( 50 ). In mice, Gata6 + large peritoneal cavity macrophages (GPCMs) are the primary responders to peritoneal injury, aggregating over damaged tissue to prevent transmission of contents between the peritoneal cavity and the surrounding vasculature ( 21 ). During development, GPCMs originate from yolk sac progenitors and maintain their population through self-renewal at homeostasis ( 51 ). However, following aggregation, the GPCM population is depleted ( 51 , 52 ). While studies have demonstrated that bone-marrow-derived monocytes can be converted into GPCMs through multiple differentiation pathways, a comprehensive mechanism for GPCM replenishment following the macrophage disappearance reaction remains elusive and may provide insight into recalcitrant adhesion formation ( 51 , 52 ). These resident cell populations mediate a delicate physiological response to the loss of peritoneal homeostasis that, when imbalanced, results in the formation of abdominal adhesions. Postoperative adhesions may result from direct surgical trauma, inflammatory processes, or, most commonly, a combination of both. Surgical injury disrupts the mesothelial monolayer and exposes the underlying ECM, initiating coagulation and inflammatory cascades. Concurrent intra-abdominal inflammation, which can arise from peritonitis, endometriosis, radiation injury, peritoneal dialysis, ischemia, or contamination with blood, microbes, gastrointestinal contents, or foreign materials, further amplifies this response ( 9 , 41 ). Although these etiologies are conceptually distinct, they converge on a shared pathophysiologic sequence characterized by mesothelial injury, inflammatory cell recruitment, fibrin deposition, and impaired fibrinolysis ( 41 ). These mechanistic insights help explain the consistently higher incidence of adhesions following open surgery compared with laparoscopy, reflecting greater visceral peritoneal trauma with open surgical approaches. Adhesions form as fibrous connections between bowel loops, solid organs, and the parietal peritoneum at sites of injury, with anatomic distribution varying by surgical context ( 29 , 41 , 53 ). In gynecologic surgery, adhesions frequently involve the adnexa of the uterus, ovaries, and fallopian tubes, particularly after myomectomy ( 54 ), whereas following gastrointestinal surgery the small intestine is most affected and accounts for the majority of small bowel obstructions ( 1 , 3 , 53 ). Adhesions also form at incision sites, with incidence influenced by both incision location and surgical approach: midline vertical incisions are associated with approximately 30% higher adhesion rates than suprapubic transverse incisions ( 55 ), while laparoscopic approaches reduce incision-associated adhesion formation by 25-50% ( 15 , 29 ). To develop effective and broadly applicable anti-adhesive therapies, common targets must be extracted from the processes that result from these various etiologies.

The loss of peritoneal homeostasis initiates a wound response that consists of three pillars: the innate immune response, coagulation, and the peritoneal cell response, which result in abdominal adhesion formation ( 19 ). While these pillars have substantial temporal and mechanistic overlap, this framework isolates three events that are, on their own, sufficient to adhere adjacent peritoneal surfaces ( Figure 1 ) ( 21 - 23 ). As such, anti-adhesive therapeutics must account for all three pillars to effectively prevent adhesions from forming. Injuries to the peritoneum trigger an innate immune response, the extent of which determines whether healing eventually resolves or progresses towards fibrosis. Initially, GPCMs activate and aggregate at the site of injury, quickly sealing the wound. Depending on the magnitude of the damage, these macrophages may become depleted in the surrounding tissues in a process known as the macrophage disappearance reaction (MDR), and remain in a pro-fibrotic aggregate, allowing an excess of infiltrating immune cells in the absence of routine macrophage clearance. This platelet-like aggregation is mediated by macrophage scavenger receptor 1 (MSR1) and macrophage receptor with collagenous structure (MARCO), which may represent initial targets for anti-adhesion intervention ( 21 ). In contrast to GPCMs, circulating monocytes and small peritoneal cavity macrophages (SPCMs) may have a protective role against abdominal adhesion formation, supported by evidence revealing that exogenous monocyte chemoattractant protein-1 (MCP-1/CCL2) administration leads to a reduction in adhesion severity ( 56 ). The balance of resident and recruited leukocytes at the site of injury is heavily dependent on cytokine and chemokine signaling. In a rat adhesion model, Uyama et al., showed that blockade of interleukin-6 (IL-6) receptor signaling reduced leukocyte infiltration, fibrin thickness, and collagen deposition, indicating a central role for IL-6 inflammation in fibrotic progression ( 57 ). Immune cell secretion of inflammatory chemokines CXCL1, MCP-1, and CXCL2 contributes to a persistent inflammatory state, recruiting neutrophils from local circulation and leading to a sustained innate immune response in the context of MDR ( 56 , 58 ). In a mouse cecal abrasion model, Tsai et al., demonstrated that neutrophil content in the peritoneal cavity rises sharply within six hours and remains elevated for at least four days, in contrast to resolving injury, in which neutrophil numbers drop off by 72 hours ( 56 ). Neutrophils further contribute to inflammation through cytokine release and the formation of neutrophil extracellular traps (NETs), which are created through the release of DNA webs during neutrophil apoptosis. NETs have been observed at adhesion interfaces, and treatment with DNAse reduces adhesion severity, suggesting that neutrophils utilize DNA to create an adhesive surface and propagate adhesiogenesis ( 56 , 59 ). In an environment characterized by general macrophage depletion, these apoptotic bodies remain uncleared, causing release of their contents into the surrounding microenvironment, provoking tissue damage, and amplifying the immune-driven fibrotic response seen after peritoneal injury. Peritoneal injury activates coagulation, leading to deposition of a fibrin-rich matrix across damaged serosal surfaces. In murine models, this fibrin layer is evident histologically within hours of surgery ( 60 ). Under physiologic wound healing conditions, the provisional fibrin matrix is cleared by plasmin-mediated fibrinolysis within several days as the mesothelium re-epithelializes, restoring homeostasis. However, when fibrinolytic capacity is overwhelmed or suppressed, the fibrin scaffold persists and becomes a permissive substrate for fibroblast adhesion, migration, and ECM deposition ( 60 , 61 ). In the rat peritoneal injury model used by Nguyen et al., treatment with an exogenous fibrinolytic enzyme resulted in minimal fibrin accumulation by post-operative Day 7, whereas, in control animals, fibrin was still present at this timepoint and was associated with greater adhesion burden, suggesting a causal relationship between fibrin persistence and the development of subsequent fibrosis. Persistent fibrin reflects the overall inflammatory and metabolic state of the injured peritoneum and is a consequence of disruption of the fibrinolytic balance through inflammatory signaling. Postoperative adhesions are accompanied by increased oxidative stress markers, reduced antioxidant activity, and elevated IL-6 and transforming growth factor β 1 (TGF-β1) levels in peritoneal tissue, and these changes correlate with thicker fibrin and higher adhesion scores ( 60 , 62 ). Hypoxia-induced signaling and toll-like receptor 4 (TLR4) activation suppress fibrinolytic enzymes and promote antifibrinolytic mediators such as plasminogen activator inhibitor-1 (PAI-1), further stabilizing the fibrin gel ( 63 ). Consistent with this framework, enhancing fibrinolysis reduces adhesion severity and decreases fibroblast adhesion and migration ( 61 ). Together, these studies support the idea that fibrin persistence is a waypoint on the path towards abdominal adhesion formation. The mesothelium has a central role in the response to peritoneal injury. Fischer et al., used a novel in vitro adhesiogenesis assay to demonstrate that within hours of injury, a propagating calcium wave causes stressed PMCs to extend membrane protrusions to form physical bridges with neighboring cells, fusing and initiating pro-adhesion marker expression ( MYL9 , ARF-GAP1 , Rho GTPases, and AKAP12 ), even in healthy cells. Subsequently these PMCs adopt mesenchymal markers and integrate into the developing adhesion ( 23 ). Lineage-tracing experiments have shown that PMCs undergo mesothelial-mesenchymal transition (MMT) and contribute directly to adhesion-forming myofibroblast populations defined by Wilm’s-tumor-1 (WT1) and α-smooth-muscle-actin (αSMA) expression ( 23 , 64 ). Other groups have shown that NETosis promotes epithelial-to-mesenchymal transition (EMT) in the context of pulmonary and cardiac fibrosis via upregulation of αSMA, Snail, and Twist, implying that neutrophil infiltration may similarly stimulate the analogous process of MMT in peritoneal adhesiogenesis ( 65 , 66 ). PMCs also play an active immunoregulatory role during this early phase, releasing pro-inflammatory mediators into the peritoneal space. Loss of mitochondrial homeostasis within PMCs may represent one upstream mechanism linking metabolic dysregulation to inflammation and fibrin persistence. Reduced SIRT3 expression in PMCs leads to mitochondrial dysfunction, increased reactive oxygen species (ROS), and activation of the NLRP3 inflammasome, with higher levels of interlukin-1β/interleukin-18 (IL-1β/IL-18) and more severe adhesions ( 67 ). Furthermore, Zindel et al., demonstrated that intraperitoneal microbial contamination enhances post-surgical adhesion formation, largely through immune cell-derived epidermal growth factor receptor (EGFR) signaling, providing another mechanism for MMT following peritoneal injury ( 64 ). These data were consistent with elevated EGFR expression in human adhesion biopsies, suggesting that EGFR may represent an anti-fibrotic therapeutic target for amelioration of abdominal adhesions by targeting the MMT process. Together, these data highlight the central role PMCs play in the acute peritoneal and inflammatory responses to homeostatic imbalance. As the provisional fibrin matrix persists, activated fibroblasts infiltrate the injured peritoneum and proliferate. Lineage-tracing and transcriptomic studies indicate that adhesion-prone fibroblasts derive from both tissue-resident peritoneal fibroblasts and PMCs undergoing MMT ( 23 , 24 , 64 ). Fischer et al., demonstrated that stressed PMCs that initially form calcium-dependent bridges between injured surfaces, later express mesenchymal markers and integrate into the adhesion stroma.( 23 ) Building on these findings, Foster and colleagues identified a distinct fibroblast population within adhesions positive for platelet-derived growth factor receptor α (PDGFRα), c-Jun (JUN), and S100 calcium binding protein A4 (S100A4), and showed that these cells are clonally expanded from local precursors rather than recruited from the circulation, indicating that adhesion fibroblasts are derived through local activation and expansion within the injured peritoneum ( 68 ). Despite their heterogeneity in origin, adhesion-associated fibroblasts differentiate towards a myofibroblast phenotype characterized by αSMA expression, contractility, and sustained production of collagen-rich ECM ( 25 , 59 ). In this state, fibroblasts also exhibit stable transcriptional reprogramming, including JUN-dependent upregulation of profibrotic signaling and proliferation ( 24 ). Progressive matrix stiffening further promotes myofibroblast activation and maintenance of a profibrotic microenvironment, creating a self-sustaining feedback loop between fibroblasts and the ECM which is a hallmark of multiple fibrotic diseases. Mature adhesions are the endpoint of pathologic adhesiogenesis following peritoneal injury. In preclinical models, by postoperative Days 7–10, adhesions progress from soft, translucent fibrous bands into opaque, vascularized, collagen-rich structures that are difficult to separate mechanically ( 56 , 60 , 61 ). Histologically, mature adhesions show dense, aligned collagen bundles with Masson’s trichrome or Sirius Red staining, reduced fibrin content, and a relatively lower proportion of inflammatory mediators compared with earlier stages. The current data supports a model in which stable adhesions develop from the cumulative interactions between sustained inflammation, impaired fibrinolysis, and activation/expansion of peritoneal fibroblasts.

Fibrosis can affect nearly every major organ in the human body, scarring functional tissue through the excessive deposition of fibrillar ECM ( 69 , 70 ). In the context of adhesion formation, fibrosis converts weak, cellular, fibrin-rich adhesive tracts into strong, collagen-rich tissues that are highly resistant to tensile forces ( 20 , 68 ). Although a causal relationship has not been established between adhesion fibrosis and negative clinical outcomes such as aSBO, mature adhesions are not susceptible to blunt dissection like their immature counterparts. As such, these tenacious adhesions pose a greater barrier to adhesiolysis and are considered a more significant risk to patient health ( 25 ). Given the clinical burden of fibrotic adhesions, and their near-ubiquitous formation despite heterogeneous adhesive etiologies, the mechanisms of adhesion fibrosis represent an ideal target for therapeutic intervention. In all tissues, the ECM and embedded cells are constantly receiving bi-directional feedback from one another, with immediate implications for tissue biomechanics, signaling dynamics, and gene expression ( 71 ). This dynamic reciprocity defines the local microenvironment and, when dysregulated, perpetuates fibrosis, making it essential to understand both the cellular processes that give rise to fibrotic adhesions and the architecture of the adhesive ECM.( 71 , 72 ) While recent preclinical work has provided novel insights into adhesiogenesis by identifying the predominant transcriptional regulators of adhesion fibrosis( 24 , 73 ), the complex signaling networks that regulate these mechanisms have not been interrogated. The ECM architecture represents an even larger gap in our collective knowledge, as the mechanisms of outside-in regulation by which ECM composition and mechanical state influence cell mobility, physiology, and fate in the adhesive interface have not been investigated. However, contemporary transcriptomic analyses have provided a glimpse into the adhesive matrisome, revealing a selection of ECM proteins upregulated in multiple fibrotic diseases. Here, we outline the activation and contribution of profibrotic cell populations to the developing abdominal adhesion and highlight components of the adhesive matrisome with therapeutic potential. Thus far, we have broadly referred to mesenchymal, collagen-secreting cells within the adhesion interface as adhesion-associated fibroblasts, but the origins of this population are multilineage, resulting in confusion over which cell types leverage specific adhesiogenic mechanisms ( 19 ). Both PMCs ( 64 , 74 ) and submesothelial fibroblasts ( 24 ) can activate and migrate into the adhesive interface, but the timing, proportion, mechanisms, and context of this influx differs between studies ( Figure 2 ). Further complicating interpretation of these results are vague uses of cell-state terminology. To accurately consolidate nomenclature, we refer to the review of fibroblast and myofibroblast activation from Younesi et al ( 75 ). Quiescent progenitors, responsible for maintaining interstitial homeostasis, become activated fibroblasts once they begin proliferating and migrating into provisional matrix. These fibroblasts transition into matrix fibroblasts as they start producing ECM and establishing focal complexes. They then transition into contractile myofibroblasts by first forming stress fibers and mature focal adhesions (proto-myofibroblast) and then incorporate αSMA into the stress fibers to increase contractility, resulting in a mature myofibroblast phenotype. Transient myofibroblasts undergo apoptosis to resolve the wound healing response, while the persistence of myofibroblasts causes pathological stiffening of the ECM. This progression from quiescent progenitor to contractile myofibroblast parallels the accumulation and stiffening of the local ECM ( 75 ). Although the current adhesiogenic literature is not sufficiently resolved by time or marker annotation to delineate between each fibroblast phenotype, this framework provides the structure necessary to address questions of PMC and submesothelial fibroblast fate, ECM spatial heterogeneity, and the divergence of homeostatic and pathological wound healing responses. Aggregation and analysis of over 20 mouse single cell RNA sequencing (scSeq) datasets by Kadri et al., has identified shared pathological mechanisms across peritoneal, pleural, and pericardial mesothelium. Following disruption, healthy peritoneal cells enter a metabolically active state defined by Ifi27l2a and Crip1 expression before progressing through proteolytic ( Dcn , Plac8 ) and fibrogenic ( Mgp , Sparc ) phenotypes. Similar populations were identified in aggregate human data, and ex and in vivo overexpression of mesothelial state markers in fibrotic mouse lungs confirmed that these markers are sufficient to remodel the ECM by degrading the established collagen matrix and constructing a de novo ECM.( 76 ) While these data do not derive from models of adhesion formation, the data strongly suggests that PMCs undergo multiple cell fate decisions before committing to MMT. Rectifying this model of PMC activation with the calcium wave model will help distinguish the cell-fate determinants that control PMC activation, MMT commitment, and profibrotic ECM synthesis, ultimately providing a clear delineation between physiological and pathological response ( 23 , 76 ). Hypoxia inducible factor 1 α (HIF1α) is a potent PMC fate determinant under hypoxic conditions, in which direct inhibition of HIF1α via small molecule inhibitors significantly reduces adhesion formation in preclinical models ( 73 , 74 , 77 , 78 ). Although this mechanism has been demonstrated repeatedly, its influence on adhesiogenesis is likely proportionally influenced by the inciting insult. HIF1α drives MMT, resulting in the expression of MSLN and podoplanin (PDPN) which label 80-90% of cells within hypoxic adhesive interfaces ( 24 , 47 , 74 , 79 ). However, animal models utilizing mechanical injury contain roughly 50% MSLN + cells, and even fewer PDPN + cells, suggesting the relevance of other cell fate determinants when hypoxia is not the primary etiology ( 24 ). TGF-β is also sufficient to drive MMT. Bulk RNA sequencing (bulkSeq) following TGF-β stimulation in vitro reveals a massive transcriptional shift in PMCs involving over 800 genes, and this transformation can be attenuated by insulin growth factor binding protein 4 (IGFBP4) and bone morphogenetic protein 4 (BMP4) signaling ( 79 ). In vivo , it is unclear how TGF-β signaling intersects with calcium wave activation of PMCs on Day 1 to drive MMT ( 23 ). Notably, senescent PMCs are resistant to the transforming effects of TGFβ, suggesting mesothelial responses may differ between young and aged cohorts ( 80 ). A thorough accounting of cell populations in the adhesive interface is necessary to resolve lineage concerns and requires a robust panel of population-level markers. In addition to the pathologic and anatomic markers recently identified by Kadri et al., several proteins have been targeted for immunolabeling and lineage tracing ( 76 ). At homeostasis, PMCs have been identified with immunolabeling of transmembrane proteins Leucine Rich Repeat Neuronal 4 (LRRN4) and MSLN ( 81 ), intermediate filaments like keratin ( 82 ) and Uroplakin 3B (UPK3B) ( 81 ), and tight junction protein zona-occludens 1 (ZO-1) ( 82 ). Most of these proteins are downregulated upon MMT (with the exception of MSLN), as the PMCs lose their epithelial phenotype and upregulate MSLN, PDPN, and WT1, which have all been used to label activated or transformed PMCs with varying degrees of success ( 49 , 74 , 79 ). Lineage tracing of WT1 has variable results, with some studies reporting anywhere from <10% to 90% occupation of the adhesive interface by WT1-lineage cells ( 24 , 81 ). Furthermore, WT1 is upregulated in a subset of PMCs during development, suggesting it is subject to MSLN-dependent fate specification ( 49 ). As mentioned previously, PDPN lineage labeling suffers from similar variability ( 24 , 74 ). MSLN has been repeatedly used for PMC identification ( 76 ), comprehensively labels the PMC population ( 49 ), and is less susceptible to variability ( 24 ), making it a strong candidate for PMC lineage tracing across the adhesiogenic timeline. Few studies have specified the role of submesothelial fibroblasts in adhesion formation, instead lumping all collagen producing cells into a single population. As such, the publications by Foster et al., have redefined the role of these cells in adhesiogenesis ( 24 , 25 ). scSeq data suggests that heterogeneous submesothelial populations converge on a collagen-positive myofibroblast phenotype due to JUN expression. In this model of adhesion fibrosis, JUN expression is upregulated by intercellular PDGF and IL-6 signaling derived from the acute inflammatory response. JUN + fibroblasts then maintain IL-6 production, leading to chronic Janus-kinase/signal transducer-and-activator-of-transcription (JAK/STAT) signaling and perpetual fibrosis. Subsequently, inhibition of JUN and IL-6 in this model are both sufficient to attenuate adhesion formation, confirming the necessity of these two proteins in the profibrotic feedback loop ( 24 , 57 ). While direct JAK/STAT inhibition has not been trialed in abdominal adhesion models, inhibition of JAK2/STAT3 with small molecule AG490 successfully reduced peritendinous adhesion formation, suggesting overlap of adhesiogenic mechanisms ( 83 ). Compared to PMCs, the markers of quiescent submesothelial fibroblasts are sparse, although staining for COL1 has been sufficient to demarcate the population at homeostasis ( 47 ). As discussed previously, the combination of JUN, S100A4, αSMA, and PDGFRα label most adhesion-associated fibroblasts following cecal abrasion, although JUN expression alone is insufficient since it is divided between the former fibroblast population and MSLN + cells. αSMA positivity suggests these cells are myofibroblasts, and analysis of these markers alongside ECM and focal adhesion protein expression can provide additional confirmation of fibroblast state ( 75 ). While PDGFRα is not definitively exclusive to submesothelial fibroblasts, PMCs do not express PDGFRα at homeostasis, making it a potential option for conditional lineage tracing ( 24 ). The ECM is a fundamental regulator of cell fate and function, with local differences in molecular composition, mechanochemical stimuli, and biomechanical force propagation having direct impact on cell adhesion, contractility, and transcriptional activity ( 75 ). Since fibrotic abdominal adhesions are a pathology of the ECM, this review adopts contemporary efforts led by the Matrisome Project to standardize ontology. The human matrisome contains approximately 300 core structural proteins which are grouped into collagens ( 44 ), proteoglycans ( 36 ), and non-collagenous glycoproteins (197), and an additional ~700 ECM-associated proteins which fall into ECM-affiliated proteins (171), ECM regulators (238), or secreted factors (344) ( 71 ). While our current assessments of the adhesive ECM depend heavily on interpretation of transcriptomic and genetic knockout experiments, advances in proteomics, high-throughput immunostaining, and atomic force microscopy technologies must be leveraged to provide a comprehensive picture of the adhesion-associated ECM. Of the different fibrotic pathologies, the initial stages of abdominal adhesion formation most closely resemble those of skin wound healing, as the establishment and remodeling of a provisional matrix is central to both processes ( 47 , 72 , 75 , 84 ). It is important to note that similarities in fibrotic processes diverge at the level of cell population, since the Engrailed-positive (EN1 + ) fibroblasts that form dermal scars are absent in abdominal adhesions ( 24 , 84 ). In the dermis, the provisional matrix is a granulation tissue comprised of platelets and fibrin that is remodeled through stages of FN, Collagen III (COL3), and COL1 deposition, gradually transforming into a fibrillar ECM network with increased tissue stiffness ( 84 ). While this linear process has not been comprehensively defined throughout adhesiogenesis, formation of the early-stage provisional matrix has been documented histologically and is required to facilitate fibroblast recruitment and fibrillar matrix assembly ( 85 , 86 ). Additionally, therapeutic inhibition of JUN changes the proportion of fibrillar ECM from COL1- to COL3-dominant, making adhesions substantially weaker and suggesting that the matrix is constrained to a more provisional state ( 25 ). This regulation of biomechanical properties through ECM remodeling is a foundational mechanism of fibrotic pathology that remains understudied in the context of adhesion formation ( 72 ). Collagens are essential to adhesion integrity, as they provide the tensile resistance necessary to prevent dissolution of the acute-stage adhesive tract ( 25 ). Twenty-eight collagen subtypes exist in humans, each comprised of three alpha chains encoded by forty-four distinct genes. Of the six categories of collagen subtypes, fibril-forming collagens (including types I, III, and V) are of the greatest relevance to the tenacity of abdominal adhesions ( 71 ). Fibril-forming collagens are linear alpha chain trimers that assemble into large bundles (fibrils) that can extend up to centimeters in length ( 87 ). COL1 is the most abundant of these, comprising the majority of interstitial tissues due to its ability to form vast, highly organized fibril networks ( 87 , 88 ). COL1 is also the primary determinant of adhesion tenacity and mature adhesion formation ( 25 , 85 ). Tissue biomechanics are largely regulated by collagen content, with tensile resistance being further modulated by fibril assembly, fibril diameter, and collagen crosslinking ( 87 - 89 ). As such, reports of spatially variable collagen content across the adhesive interface are of particular interest, suggesting the profibrotic mechanisms of adhesion formation do not produce a uniform biomechanical microenvironment. While this spatial heterogeneity has been well documented by histological profiling across clinical adhesiolysis samples, the mechanisms that cause spatial heterogeneity are unknown, offering opportunities to identify targetable cell fate determinants by comparing microenvironmental factors between collagen sparse and dense adhesiolysis tissue ( 90 , 91 ). The transcriptomic analyses of adhesiolysis tissue and stimulated PMC cultures has identified multiple upregulated ECM components whose heterogeneous distribution could cause divergences in the adhesiogenic microenvironment ( 24 , 74 , 76 , 79 ). Recently, collagen crosslinking has emerged as a potential target for adhesion amelioration ( 92 ). The strength of COL1 fibrillar networks comes from both the concentration and alignment of fibrils ( 93 - 95 ), and also their degree of crosslinking, which can be mediated either enzymatically via transglutaminases (TG) and lysyl oxidases (LOX), or non-enzymatically by age-associated glycation end-products (AGE) ( 96 , 97 ). While the predominant mediator of collagen crosslinking in the developing abdominal adhesion is unidentified, LOX has previously been targeted via siRNA knockdown and the LOX inhibitor ß-aminopropionitrile in a mouse model of peritoneal fibrosis to effectively limit peritoneal thickening, positioning enzymatic crosslinking as a viable target for reducing postoperative adhesion fibrosis ( 92 ). FN is essential for both collagen fibril assembly and cell adhesion and thus comprises the initial remodeling stage of the provisional matrix ( 75 , 98 ). Fibrin-fibronectin interactions are a prerequisite for the formation of a FN network, as cryptic binding sites within the fibrin clot keep relaxed extracellular FN dimers localized within the remodeling tissue ( 86 , 99 ). Integrin binding to the relaxed extracellular FN then causes a conformational shift which elongates the homodimer, allowing its assembly into fibrillar networks ( 71 ). In turn, the tensile forces transduced through these networks are essential for the formation of mature focal adhesions. FN binding sites enable interaction with the collagen network and signaling molecules such as latent TGF-β binding protein (LTBP), enabling the transmission force from αSMA positive contractile fibers to regulate collagen network stiffness and the availability of ECM-associated secreted growth factors ( 71 , 75 ). Of note, relaxed FN networks are a hallmark of diseased ECM, comprising various tumor stroma and pericardial fibrosis samples, and may be driven by direct myofibroblast adhesion to the collagen network ( 94 , 100 , 101 ). Labeling FN tensional states may present a method for differentiating between physiological and pathological peritoneal wound healing responses, allowing the identification of key molecular regulators ( 100 ). Matricellular proteins are a subtype of non-collagenous glycoproteins characterized by their general antagonism to cell adhesion and absence of structural function. This diverse group associates with a wide variety of proteins including the structural ECM, growth factors, cell surface receptors, and integrins and other cell adhesion molecules, and are frequently involved in profibrotic pathologies ( 102 , 103 ). Regulation of cell adhesion is primarily conducted via competition with FN for adhesion proteins like integrins and syndecans, suggesting a potential mechanism of pathological FN network collapse ( 102 , 104 ). Factors frequently upregulated in adhesiogenic processes include the thrombospondin and tenascin families, osteonectin ( SPARC ), osteopontin ( SPP1 ), and periostin ( POSTN ) ( 24 , 71 , 74 , 76 , 79 ). Knockdown of SPP1 has previously been shown to attenuate postoperative adhesion formation, although the mechanism remains undefined ( 105 ). Tenascin-C ( TNC ) is also a potential candidate for adhesion prophylaxis, as it is required for other fibrotic diseases of the peritoneum like endometriotic lesion formation ( 106 ), ovarian cancer metastasis ( 107 ), and systemic sclerosis ( 108 ).

Adhesions are fundamentally a fibrotic disease, with the confluence of multiple wound healing mechanisms resulting in a provisional matrix primed for pathological ECM deposition. Despite advances in our understanding of adhesiogenic profibrotic transcriptional regulators, the influence of the adhesion ECM remains largely uncharacterized and provides opportunities for overcoming the similarities between physiological and pathological peritoneal wound healing that have historically limited anti-adhesive therapeutics. Analysis of fibroblast lineage, phenotype, and transcriptional state throughout adhesion formation enables the identification of adhesion-specific cell fate determinants and should be coupled with comprehensive characterization of ECM composition and biomechanical properties. Comparisons between sparse and dense regions of collagen throughout the adhesive interface, and between adhesions and adjacent wound healing sites like anastomoses, provide well-controlled methods to study fibroblast fate determination. Recent transcriptomic analyses of adhesiogenic samples have provided a strong foundation to interrogate the individual cell-matrix dynamics that direct abdominal adhesion formation, enabling novel approaches for abdominal adhesion amelioration. Moreover, cross-species comparative transcriptomic atlases spanning pre-clinical and clinical data will identify critical opportunities for therapeutic intervention going forward, with the field now at a critical inflection point to move toward tangible translational strategies.

Animal models are indispensable to abdominal adhesion research, as the multifactorial etiology of adhesion formation cannot be fully captured by in vitro culture systems or post hoc analyses of clinical adhesiolysis samples. Our knowledge of adhesiogenesis, as discussed in the previous sections, is almost entirely derived from studies conducted in animal models, making their utility and importance abundantly clear. However, several different species and surgical techniques have been used to study adhesion formation, each with unique strengths and drawbacks. This has resulted in the lack of a single standardized model, complicating the synthesis of experimental results across studies of adhesion fibrosis ( 19 ). Therefore, we outline the major species and models used to study the pathogenesis of adhesion formation, using the ‘fidelity, homology, and discrimination’ framework defined by Carmichael et al., to highlight the key challenges and opportunities associated with each ( 109 ). When comparing between organisms, fidelity is their structural likeness at the anatomical, cellular, and molecular scales, homology the similarity between their disease processes, including etiology and progression, and discrimination is the resemblance of their mechanical responses to therapeutic intervention ( 109 ). Taken together, this framework describes the external validity of a model, or its ability to accurately predict the mechanisms of pathophysiology and therapeutic intervention in a human population ( 109 ). Assessment of a model’s fidelity, homology, and discrimination is essential when designing or interpreting adhesiogenic studies and enables a transparent synthesis of the literature. While organisms as fundamental as the sea urchin have been used to study the dynamics of coelomic cavity immune surveillance and wound healing ( 51 ), laboratory mice ( Mus musculus ) and rats ( Rattus norvegicus domestica ) are the primary models of adhesiogenesis due to their relative external validity, high reproduction rate, ease of maintenance, and genetic consistency and manipulability ( 51 , 109 - 111 ). C57Bl/6J, BALB/c, and SWR/J mice, and Sprague-Dawley and Wistar rats have all been used to model adhesion formation and share anatomical features that may bias the translation of adhesiogenic studies ( 109 , 112 - 114 ). For example, appendectomies only generate a small percentage of adhesions formed in the human clinical setting, but the rodent cecum, somewhat analogous to the human appendix, is substantially larger relative to the rest of the gastrointestinal (GI) tract and serves as a common location for the induction of adhesion formation in rodents ( 3 , 14 , 115 ). Similarly, the greater omentum participates in up to 90% of human adhesion events, but in rodents this connective structure does not extend past the pancreas, preventing its direct involvement in models localized to the parietal peritoneum and lower gastrointestinal tract ( 3 , 14 , 27 , 115 , 116 ). Recent scSeq analyses have identified unique PMC markers across GI, omental, and adipose tissues, suggesting that differences in rodent omental and GI anatomy may shift the visceral mesothelial response towards clinically underrepresented PMC subpopulations ( 76 ). Determining whether these markers have a functional impact on PMC activation and MMT would clarify initial assessments of rodent model fidelity and homology. scSeq has also advanced our understanding of the rodent and human immune systems with similar implications for model fidelity. While GPCMs comprise roughly 95% of the resident immune population in mice and are the primary mediators of the innate immune response to peritoneal damage, less than 5% of human peritoneal macrophages are GATA6 positive ( 21 , 117 ). Instead, humans maintain a population of intermediate-stage macrophages whose contribution to adhesion formation remains unknown, making it difficult to assess the homology of the innate immune responses ( 117 ). Complicating these efforts further are the differences in immune response between rodent strains, as C57Bl/6J mice produce a monocyte-heavy acute inflammatory response to peritoneal damage which is more protective against adhesion formation than that of BALB/c mice ( 118 ). This suggests that fidelity and homology across rodent strains is not equivalent, but direct comparisons between a greater number of strains, and between the inflammatory responses of human and rodent peritoneal cavity macrophages, are still required to inform model organism selection. Surgical technique contributes substantially to the homology and discrimination of a model, and a variety of methods have been employed to generate adhesions including excision, abrasion, ligation, electrocautery, clamping, and desiccation of the peritoneum, anastomosis of the small bowel, and exposure of the peritoneal cavity to foreign bodies or autologous blood ( 25 , 109 , 119 ). Cecal abrasion and peritoneal button models of adhesion formation have become favored because they excel at producing severe adhesive burdens, reducing the sample size and variability of surgical cohorts and enabling the study of molecular mechanisms that would otherwise be impenetrable ( Figure 3 ) ( 85 , 109 , 111 , 120 ). However, these models are limited in their ability to discriminate between anti-adhesive therapies since the surgical mechanisms of adhesion generation are divorced from clinically relevant procedures. Differences between these models may also result in conflicting reports of the peritoneal cell contribution during adhesiogenesis ( 19 ). Though individual abrasion or peritoneal button protocols vary substantially between studies, a potential driver of divergent outcomes is the induction of punctate bleeding via abrasion. Since the capillary bed is sublaminar, observation of punctate bleeding confirms defacement of the basal lamina and the direct exposure of submesothelial fibroblasts to the acute inflammatory response, potentially promoting their involvement in adhesiogenesis. Studies utilizing cecal abrasion as the primary defect detect more fibroblasts throughout the adhesion interface ( 24 , 47 , 48 ). In contrast, studies utilizing ischemic buttons are biased towards PMC involvement, and the ischemia and subsequent hypoxic response induced in PMCs by peritoneal buttons adds an additional source of disparity between the two models ( 73 , 74 ). Despite these biases, cecal abrasion and ischemic button models remain the most reliable models available, generating well powered data without the need for large sample sizes ( 19 , 85 ). Rigorous studies of adhesiogenic mechanisms will address the limitations of external validity by testing hypotheses with multiple surgical models and trialing therapeutic interventions in organisms with robust fidelity and discrimination ( 24 , 25 ). The species and surgical models presented thus far are best suited for identifying and interrogating the cellular and molecular mechanisms of adhesion formation, where the sacrifices in external validity are justified by the reproducibility of the model. However, alternatives optimized for external validity are better suited to trial anti-adhesive therapies since the model’s response to therapy (discrimination) must be as predictive as possible ( 109 ). Recently, small bowel anastomoses were performed in both mice and fetal pigs ( Sus scrofa domesticus ) to test the safety and efficacy of an anti-adhesive hydrogel ( 25 ). The procedure is more complicated than cecal abrasion or peritoneal buttoning, limiting cohort size and causing variation in the adhesiogenic response. However, fidelity and homology are improved since the technique directly reproduces a clinical procedure, and discrimination is improved as the disease burden accurately reflects the variability observed in the patient population ( 3 ). Additionally, burst testing of the anastomosis provides a metric to assess the safety of the therapeutic and any possible impact on surgical wound healing ( 25 ). Postmortem scoring of the abdominal cavity is the current standard for assessing adhesive burden, but this process is not standardized and stems from a disconnect between the scoring process and tangible negative clinical outcomes (i.e. aSBO) ( 109 ). Guidelines for blinded clinical scoring of adhesion formation frequently consider the number and location of adhesions formed, the visible size, opacity, and vascularity of these adhesions, and occasionally, their physical quality, such as the difficulty of dissection ( 24 , 64 , 74 , 121 ). As such, adhesive burden is a general term which encapsulates the area occupied by an adhesiogenic response and its resistance to external dissolution. However, no quantitative correlations have ever been established between any of these qualitative factors and the risk that a negative clinical outcome, such as aSBO, actually occurs in the subject ( 18 ). As such, current clinical scoring guides do not reflect a risk posed to the subject’s health but are instead a gross readout of the strength of the adhesiogenic response. Despite these shortcomings, current scoring guides that quantify adhesive burden remain a valuable tool for assessing adhesiogenic mechanisms, but it is critical to remember that successful reductions in adhesive burden via preclinical therapeutic trials indicate a blunting of the adhesiogenic response and do not necessarily reflect a decreased risk to the subject. As such, future work to establish a standardized, comprehensive scoring system that accounts for both adhesive burden and complication risk will be extremely valuable. In the interim, the use of guidelines that rely, at least in part, on the number and size of the adhesions present, such as Nair’s clinical scoring guide, provide the most comprehensive criteria at this time as the presence or absence of adhesions is the greatest predictor of negative clinical outcomes ( 121 ). Additional factors such as adhesion opacity, vascularity, and resistance to dissection, can be helpful qualifiers, but interpretation of these is highly subjective and should not take precedent over size and quantity. Standardized photographic documentation of the post-operative peritoneal cavity, especially of the primary outcome site, e.g. parietal peritoneum to cecum for cecal abrasion models, should accompany this clinical scoring and be published for transparency.

Despite substantial advances in surgical technique and increasing insight into the biology of peritoneal repair, postoperative adhesions remain among the most pervasive and clinically consequential sequelae of abdominopelvic surgery. Adhesions are exceedingly common, developing in the vast majority of patients after open abdominal or pelvic procedures (67–100%) ( 26 - 28 ) and persist in more than half of patients overall (54%), even in the laparoscopic era ( 29 ). These high incidence rates underscore the resilience of adhesion-forming pathways despite procedural refinements and development of a few adhesion-prevention products. Adhesion formation occurs across a broad spectrum of surgical populations and is shaped by patient-specific, disease-related, and operative factors. Nonmodifiable characteristics such as age, as well as clinical conditions reflecting the magnitude of peritoneal injury (including inflammation, infection, and ischemia) are associated with increased adhesion risk ( 30 ). However, the extent of surgical trauma remains the dominant determinant. The degree of peritoneal injury and procedural complexity strongly predict adhesion formation and anatomic distribution, reflecting focal disruption of the peritoneal surface at the site of surgical manipulation ( 4 , 29 , 31 ). Accordingly, open surgical approaches, repeat operations, prolonged operative times, and procedures requiring extensive dissection are consistently associated with higher adhesion density and recurrence ( 28 ). Prior abdominal surgery further amplifies risk and often predicts sites of adhesion reformation, contributing to marked variation in adhesion prevalence across surgical disciplines, including gastrointestinal (66%), obstetric and gynecologic (51%), and urologic (22%) procedures ( 29 ). Although laparoscopy reduces adhesion formation by approximately 25-50% compared with open surgery, these reductions translate into only modest improvements in long-term clinical outcomes ( 15 , 29 , 31 ). The clinical consequences of post-operative adhesions extend well beyond the index operation and represent a substantial source of chronic morbidity and healthcare utilization. While many adhesions remain asymptomatic, clinically significant adhesions are associated with chronic abdominal or pelvic pain, infertility, and ischemic complications ( 4 , 32 ). aSBO is the most severe and frequent manifestation, accounting for approximately 56% of all cases of bowel obstruction and occurring in approximately 2% of patients following abdominal or pelvic surgery ( 4 ). Large-scale epidemiologic studies have firmly established the long-term burden of adhesion-related disease. The landmark Surgical and Clinical Adhesions (SCAR) study demonstrated that within 10 years of open abdominopelvic surgery, one in three patients (34.6%) were readmitted for conditions potentially related to adhesions, and 5.7% required hospitalization for directly attributable complications. Notably, over one-fifth (22.1%) of all adhesion-related readmissions occurred within the first postoperative year, with risk persisting throughout the decade-long follow-up period ( 33 ). Subsequent analyses of the SCAR cohort and related populations have confirmed that adhesion-related readmissions and complications remain common despite widespread adoption of minimally invasive surgical techniques ( 3 , 34 ). Beyond their direct clinical manifestations, adhesions substantially complicate subsequent surgical interventions by creating dense fibrotic planes that obscure anatomy and increase technical difficulty. These changes prolong operative time and elevate the risk of intraoperative complications, including bleeding and inadvertent enterotomy. Adhesion-related operative complexity is associated with increased postoperative morbidity, including greater need for intensive care, parenteral nutrition, and prolonged hospitalization ( 35 - 39 ). Reoperation in the setting of dense adhesions is particularly hazardous. Adhesiolysis itself carries considerable risk, with inadvertent bowel injury occurring in approximately 6% of cases and associated reported mortality rates as high as 13% ( 4 , 39 ). Collectively, these persistent and costly sequelae highlight the urgent need for improved mechanistic understanding of adhesion formation to inform the development of effective preventive and therapeutic strategies.

Abdominal adhesions are networks of scar tissue that form between peritoneal surfaces following the majority of abdominopelvic operations, causing small bowel obstructions ( 1 ), infertility ( 2 ), chronic pain ( 3 ), and complications during repeat surgical operations ( 4 ). Adhesive small bowel obstructions (aSBO) alone represent a significant healthcare burden that places excessive strain on emergency surgery departments ( 5 ), bears a 7.2% risk of death ( 6 ), and exceeds $1.7 billion in annual clinical expenditures ( 5 ). These adhesion-related complications have presented a recurrent challenge to surgeons and scientists throughout history, but recent advances in our understanding of adhesion pathophysiology set a promising framework for the development of anti-adhesive therapies. While documentation of small bowel obstructions occurred as early as the 16 th century BC( 7 ), the first descriptions of abdominal adhesions were reported in 1835 by Richard Bright following autopsies of surgical patients ( 8 ). Then, in 1889, T. Dembrowski initiated over a century of scientific investigation into the etiology and pathophysiology of abdominal adhesion formation when he generated the first animal model of adhesiogenesis, creating parietal peritoneal defects in dogs ( 9 , 10 ). Since then, scientific inquiry has made substantial advances in our knowledge of adhesion pathophysiology, but the resulting anti-adhesive therapies (prophylactic barriers) have not been widely adopted due to concerns over cost-effectiveness, safety, and efficacy ( 11 , 12 ). In parallel, surgeons have attempted to optimize their practices to reduce and treat adhesion-related complications ( 4 ). The first laparoscopic surgery was performed in 1933 to lyse adhesive tissue ( 13 ), and although the technique would not be adopted to relieve aSBO until 1991 ( 7 ), its implementation alongside measures to protect the peritoneum from desiccation and exposure to foreign bodies have significantly reduced the rates of postsurgical adhesion formation ( 3 , 4 , 14 , 15 ). Despite this, careful surgical technique and adhesiolysis (resection) remain the only treatments for abdominal adhesions, and adhesiolysis alone is insufficient due to its significant risk of inadvertent enterotomy and failure to prevent subsequent adhesion formation ( 3 , 4 , 15 - 17 ). As such, novel therapeutic approaches are still required to ameliorate adhesiogenesis and improve the care of abdominopelvic surgery patients ( 18 , 19 ). To advance therapeutic options for adhesion prevention, the Surgical Adhesions Improvement Project, initiated by the American College of Surgeons (ACS), recently published a landmark review reflecting the growing consensus that anti-adhesive therapies must address three pillars of adhesion formation: innate immunity, coagulation, and the peritoneal cell response ( 19 ). Together, these mechanisms culminate in the formation of ectopic fibrotic networks ( 19 , 20 ). Although each pillar, in isolation, is sufficient to adhere adjacent peritoneal surfaces ( 21 - 23 ), the strengthening of these adhesions through the construction of a robust extracellular matrix (ECM) facilitates the development of intractable adhesive burdens ( 24 , 25 ). Here, we review the fibrotic mechanisms underpinning abdominal adhesion formation, expanding upon previous discussions of the peritoneal cell response to identify potential mechanisms that remain largely uncharacterized, and highlighting key areas for future research. Careful interrogation of the cellular and molecular mechanisms that drive fibrosis of the abdominal adhesion will provide novel therapeutic targets for their prevention and amelioration.