Closed Incision Management With Negative Pressure Wound Therapy (CIM)

Introduction

Underlying comorbidities such as obesity, diabetes, and poor vascular status as well as risk factors such as smoking, radiation therapy, chemotherapy, and use of steroids present potential challenges in maintaining incision closure after an open surgical procedure.^1-3 Closed incisions with a high risk of complications include those from hip and knee arthroplasty, ⁴ lower-extremity bypass, ⁵ abdominal laparotomies,^6,7 and cardiothoracic procedures. ⁸ These incisions have been traditionally closed by primary intention using sutures, staples, adhesive strips, or a combination thereof. In addition to the higher risk for surgical site complications in these procedure groups, the use of sutures and staples induce stress concentrations where they engage the tissue, and elevated stress concentrations can cause ischemia, fibrosis, or other tissue injury.

Negative pressure wound therapy (NPWT), as delivered by VAC Therapy (KCI USA, Inc, San Antonio, TX),^9-11 involves the controlled application of intermittent or continuous subatmospheric pressure to the wound bed typically via a pressure-manifolding dressing (eg, a reticulated, open-cell foam dressing). NPWT has been shown to have significant clinical success in the treatment of wounds that have a tissue deficit.^9-11 This success with open wounds has led some clinicians to use NPWT on closed surgical incisions using modifications of dressings that were designed for open wounds. In a randomized prospective clinical trial comparing NPWT to standard postoperative dressings used over closed incisions following high-energy trauma, Stannard et al^5,12 reported the incidence of dehiscence and infections to be lower in the NPWT-treated group. The authors recommend that NPWT be considered for high-risk wounds following severe skeletal trauma. A retrospective study by Atkins et al ⁸ found that when patients who were at high risk for sternal infection after sternotomy were treated postsurgically with NPWT, only 1 incision out of 213 (0.5%) became infected. This rate was lower than the authors’ prediction of a 6.1% ± 4% sternotomy infection complication rate for high-risk patients, based on risk scores developed by Fowler et al. ¹³ These results, and more recent case studies, ¹⁴ suggest that NPWT on clean closed surgical incisions may help reduce the risk of infection and other postoperative complications.

Clinical benefits reported when applying NPWT over clean, closed surgical incisions have led to the development of a closed incision management with negative pressure wound therapy (CIM) system (Prevena Incision Dressing, KCI USA, San Antonio, TX). Whereas the mechanisms of how NPWT delivered by VAC Therapy (KCI USA, Inc, San Antonio, TX) facilitates healing in chronic and traumatic wounds is the subject of many papers,^9,15-18 the mechanism for how CIM using Prevena Incision Dressing, facilitates healing in closed incisions has not yet been evaluated. The dressing under negative pressure is hypothesized to affect local stresses around closed incisions in a beneficial manner. Computer simulation using finite element analysis (FEA) has been used extensively in biomedical engineering to simulate cells, tissues, organs, vasculature, and musculoskeletal systems to model mechanical behavior.^19-21 As a complementary approach, bench models of tissue analogues provide a hands-on experimental method to validate or replicate FEA models using physical materials and measurements.^22,23 Both techniques have also been applied to previous NPWT research.^24-26

This article provides initial evidence from computer modeling and benchtop experiments on the manner in which CIM with the Prevena Incision Dressing applied to an incision alters the local biomechanical environment, which affects incisional healing.

Methods

The CIM system (Prevena Incision Management System, KCI USA, San Antonio, TX) consists of a small, single-patient, disposable therapy unit and a dressing specifically designed for use over closed surgical incisions that are at high risk for postsurgical complications caused by location and/or patient comorbidities. This system, which is cleared by the FDA and commercially available, consists of a battery-operated therapy unit, preset to −125 mm Hg and a peel-and-place reticulated open-cell foam dressing with a skin-friendly interfacial layer bonded to the foam component (Prevena Incision Dressing, KCI USA, San Antonio, TX) To provide flexibility for application to anatomical contours, the dressing has a pleated nonadherent drape surrounding the foam, which is also scored to allow bending. Computer-simulated and actual Prevena Incision Dressings were used in this study.

Two FEA models as well as benchtop testing evaluated the impact of NPWT with the Prevena Incision Dressing on the biomechanical environment in and around previously closed cutaneous incisions. The measurements for the reticulated open-cell foam and underlying tissue protective layer of the Prevena Incision Dressing are as follows: 19 mm thickness, 63 mm width, and 254 mm length. A schematic cross-section of this dressing is shown in Figure 1.

OPEN IN VIEWER

Figure 1.

Schematic of the Prevena Incision Dressing in cross-section (KCI USA, Inc, San Antonio, TX). The blue shaded areas indicate skin adhesive surfaces. Green shaded areas are adhesive bonds within the dressing

Finite Element Analyses

FEA entails the use of computer tools to simulate and analyze real-world scenarios. Two distinct and independent FEA approaches were used in this study (FEA-1 and FEA-2). The models were limited to 2-dimensional cross-sections through the incision, tissue, and dressing to avoid end effects and reduce computation time. Tissue and dressing components were assigned mechanical material models obtained from biomechanics literature and test data (Table 1).^27-31 The models of tissue were flat to eliminate complicating any loading effects caused by anatomical curvature. Abaqus/Explicit versions 6.7 and 6.9 (Dassault Systemes Simulia Corp, Providence, RI) finite element software were used to perform the simulations.

OPEN IN VIEWER

Table 1.

Tissue and Dressing Mechanical Material Models: Parameter Units Are SI (mm): mm, N, 10³ kg, 10³ kg/m³, s, MPa, mJ

Tissue Type	Mechanical Model	Strain Energy Function	Material Parameters and Mass Density 1	Source
Epidermis, dermis	Hyperelastic 18	Polynomial	C10 = 0.0094; C01 = 0.082; D1 = 0.25; ρ = 1.5 × 10−8	Hendriks 27
Adipose tissue (fat)	Hyperelastic	Ogden	µ1 = 0.05; α1 = 5; D1 = 1; ρ = 1.0 × 10−8	Oomens et al 28
Muscle	Hyperelastic	Ogden	µ1 = 0.18; α1 = 30; D1 = 0.25; ρ = 1.5 × 10−8	Oomens et al 28
Prevena Component
Drape	Hyperelastic	Ogden	µ1 = 0.0147; α1 = 4.3; µ2 = 5.44; α2 = −1.97; D1 = 0.01; D2 = 0; ρ = 1.5 × 10−9	Test data model fits
Foam, fabric	Hyperfoam 19	N/A	µ1 = 0.0255; α1 = 25; µ2 = 1.22 × 10−5; α2 = −2.15; ν1 = 0; ν2 = 0; ρ = 1.5 × 10−8	Test data model fits

FEA-1: Closed incision model with subcutaneous void

This initial FEA evaluation focused on depiction of lateral forces across the incision overlying a subcutaneous void with and without NPWT. Controls were not evaluated in this earlier model, which was intended to be relatively more qualitative than FEA-2 (described in the next section). An open incision, 25 mm deep × 2 mm wide, extended to a subcutaneous void that was modeled as a rectangular space 50 mm wide × 5 mm high, below the epidermis (Figure 2A). The overall tissue dimensions were 100 mm wide × 50 mm high. The mechanical properties used in this model were those of adipose tissue because this is expected to be a predominant tissue type in the anatomical areas where NPWT is to be used (see Table 1). Suture closure of the incision was simulated by pulling the incision sides together, which resulted in appositional tension across the incision line (Figure 2B). The dressing modeled was a reticulated, open-cell foam block 20 mm thick. The dressing was modeled to adhere to the tissue (no slippage between foam and skin). The boundary conditions for the tissue block were bottom and sides fixed with the top of the tissue free to move. Negative pressure (−125 mm Hg; −16.7 kPa) was then applied to the dressing (Figure 2C).

OPEN IN VIEWER

Figure 2.

Finite element analysis-1: a 2-mm wide incision cuts from the skin surface to the void (A). The incision is joined by sutures (B) and lateral stress develops (color contours, MPa). Negative pressure (−125 mm Hg) is applied to the dressing (C), and lateral stresses become more compressive

FEA-2: Closed incision model with sliding fascial separation

The FEA-2 model used multiple layers of tissue to provide a more realistic model and was intended to be more quantitative than FEA-1 (Figure 3A). The tissue model included the nonlinear mechanical behavior of epidermal, dermal, and subdermal tissue layers. Tissue thicknesses were defined to be as follows: epidermis, 0.212 mm; dermis, 2.97 mm; fat layer above fascial separation, 9.94 mm; fat layer below fascial separation, 10 mm; muscle layer, 10 mm. The dermis and epidermis thicknesses were within the range described in the literature. ³² The tissue block width was 200 mm. A vertical incision cut through the epidermis, dermis, and upper fat layer. A fascial separation between the fat layers extended laterally for 25 mm on both sides of the incision bottom. Sutured closure was modeled as locally joined element faces at the incision and may be considered a control condition (Figure 3B). Tension in the dermis and epidermis was induced by applying pressure (−150 kPa) to the dermis and epidermis exposures at the model sides. This opened a 2-mm incision width prior to suturing (Figure 3A). Tissue self-contact within the incision was modeled as frictionless. The incision was assumed to have fluid patency with the surface, permitting negative pressure distribution throughout the subcutaneous void. This is based on the assumption that any fluids entering the incision from the tissue are evacuated by the negative pressure at a rate greater than their input, so no hematoma/seroma can develop. Negative pressure (−125 mm Hg) was applied uniformly within the CIM dressing cavity, distributed along the drape and tissue surfaces and throughout the incision.

OPEN IN VIEWER

Figure 3.

Finite element analysis-2: lateral stress color contour plot of the incision (A) after suturing (B) and with application of the dressing with negative pressure to the model (C). Maximum principal tensile stress directions are indicated by the red vectors on each element

The Prevena Incision Dressing dimensions were modeled from the bottom up, as follows: lower drape inner edge, 25 mm from center; outer edge, 96 mm from center; foam width, 63.5 mm; foam thickness, 17.5 mm; fabric width, 63.5 mm; and fabric thickness, 0.5 mm. Drape was simulated to be adhered to the underlying skin, modeling the adhesive coating on the CIM-dressing. Beneath the foam portion of the dressing, the fabric–skin interaction was defined to have a friction coefficient of 0.25, which is consistent with values previously used in the peer-reviewed literature. ³³ Loading was applied stepwise. At time 0, no loads were applied to the model. Then, skin tension was applied as a smooth increase from 0 to 150 kPa (−1125 mm Hg) over 0.2 s. Negative pressure was applied as a smooth increase from 0 to 16.7 kPa (−125 mm Hg), starting at 0.4 s and attaining target negative pressure at 1 s; model responses were observed.

Benchtop Modeling

We developed an instrumented physical model of soft tissue with an incision representing a clean, closeable surgical incision with the capability to measure the closing force or resistance to opening as the tissue is pulled in tension. The tissue model consisted of cast layers composed of various mixtures of room-temperature–vulcanized liquid silicone (Transil 40-1, PlatSil Gel 10) and PlatSil Gel 10 deadener (Polytek Development, Easton PA), simulating 3 layers of tissue: 2 mm skin, 10 mm fat, and 10 mm muscle. These tissue analogues were evaluated in uniaxial compression tests and had 20% compression tangent moduli of 190, 27, and 229 kPa for muscle, fat, and skin, respectively. At the axis of the model, an incision 9 inches long was made down to the muscle layer (Figure 4). A lateral dissection between the skin and fat layers was created, extending from 1 cm outside the incision plane to beyond the tissue model edges, and thin steel plates, with holes added for interdigitation, were implanted and bonded in place by room-temperature–vulcanized silicone. The distal ends of the plates were secured to the grips of a Universal Test System (Model 100R12, Test Resources, Inc, Shakopee MN) configured for horizontal extension force testing with a 250-lb force-load cell (Figure 5). The strain rate was set at 1 mm/s for maximum extension up to 10 mm, followed by a 5-s hold and then returned to the starting point at the same strain rate. Force and displacement data were recorded continuously. Two test clinical scenarios, each with and without negative pressure, were evaluated: (1) incision closure with sutures (2-0 Ethilon black monofilament, Ethicon Inc, Somerville, NJ) in a continuous running stitch spaced 5 mm apart, (2) sutures with NPWT dressing at −125 mm Hg, (3) incision closure with staples, and (4) staples with NPWT dressing at −125 mm Hg.

OPEN IN VIEWER

Figure 4.

Schematic of bench test setup: overhead view (A) and cross-section (B)

OPEN IN VIEWER

Figure 5.

Incision bench model configured for horizontal extension-force testing with the negative pressure wound therapy dressing applied

Differences between negative-pressure–treated and nondressed conditions were evaluated separately for the sutured and stapled conditions. The data were evaluated for suitability of parametric analyses by checking for normality using the Shapiro-Wilk test and for equal variances using the Levene test and then compared between groups using the nonparametric Wilcoxon Signed Rank test.

Results

FEA-2: Closed Incision Model With Sliding Fascial Separation

Contour color plots of the lateral stress are depicted in Figures 3B and 3C. This figure also shows the maximum principal stress magnitude and direction as arrows within each element. When the open incision (Figure 3A) is closed with superficial and deep sutures (Figure 3B), there is increased lateral tension in both sutures. Furthermore, the lines of stress in the tissue are bent toward the suture line, concentrating the tension at the sutures and resulting in a nonhomogeneous stress distribution and suboptimal contiguity of apposing tissue surfaces, resulting in gaps.

With only sutures in place but prior to application of negative pressure, there was substantial lateral tensile stress at the superficial (27.8 kPa) and deep (8.4 kPa) sutures. The average lateral stress at the incision site was 28.05 ± 1.98 (SE) kPa in the epidermis, 14.50 ± 0.08 kPa in the dermis, and 3.31 ± 1.42 kPa in fat. When negative pressure is applied, the dressing collapses vertically and contracts laterally. The gap in the incision closes, and the vertical compression in the sides of the incision is eliminated (Figures 3B and 3C). As indicated by the red vectors, the tensile stresses at the incision become horizontal and uniform. The lateral tensile stress at the superficial sutures decreased to 15.4 kPa (a decrease of 45%) and at the deep suture to 4.2 kPa (a decrease of 50%). The average lateral stress at the incision site decreased to 14.82 ± 1 kPa in the epidermis (P = .0235; a decrease of 47%) and increased to 15.3 ± 0.2 kPa in the dermis (P = .0407; an increase of 6%); the average lateral tensile stress in fat increased to 4.2 ± 0.1 kPa, not significantly different from prenegative pressure values (P = .4529). Table 2 provides detailed data.

OPEN IN VIEWER

Table 2.

FEA-2 Statistical Comparison of Tissue Stress Metrics Between Sutures Alone and Sutures + NPWT

	Treatment
		Sutures Only		Sutures + NPWT
Biomechanical Metric	Tissue (n)	Mean	SE	Mean	SE	Percentage Change	P Value
Lateral stress (kPa)	Epidermis (2)	28.05	1.98	14.82	0.58	−47.16	.0235 a
	Dermis (2)	14.50	0.08	15.34	0.15	5.79	.0407 a
	Fat (12)	3.31	1.42	4.22	0.04	27.28	.4529 b
Shear stress (kPa)	Epidermis (2)	3.67	0.14	0.12	0.11	−96.76	.0029 a
	Dermis (2)	2.16	0.20	0.38	0.26	−82.44	.0327 a
	Fat (12)	1.08	0.25	0.04	0.01	−96.56	<.0001 b
Orientation (deviation from horizontal, degrees)	Epidermis (2)	7.39	0.48	0.42	0.55	−94.28	.0054 a
	Dermis (2)	7.52	1.32	1.37	1.26	−81.82	.0413 a
	Fat (12)	14.05	7.81	0.55	0.47	−96.10	<.0001 c

Abbreviations: FEA, finite element analysis; NPWT, Negative pressure wound therapy; SE, standard error.

a

t Test.

b

Wilcoxon signed-rank test.

c

Welch ANOVA.

The shear stress in the epidermis decreased significantly with negative pressure from 3.7 ± 0.2 kPa to 0.1 ± 0.2 kPa (P = .0029; a decrease of 97%), while in the dermis at the incision site, shear stress decreased significantly with negative pressure from 2.2 ± 0.3 kPa to 0.4 ± 0.4 kPa (P = .0327; a decrease of 82%). Shear stress in the fat layer decreased significantly with negative pressure from 1.08 ± 0.87 kPa to 0.04 ± 0.04 kPa (P < .0001; a decrease of 97%). The average orientation of the principal stress from the horizontal in the epidermis decreased from 7.4° ± 0.5° to 0.4° ± 0.5° (P = .0054; a decrease of 94%). The average orientation of the principal stress from the horizontal in the dermis decreased from 7.5° ± 1.3° to 1.4° ± 1.3° (P = .0413; a decrease of 82%). The average orientation of the principal stress from the horizontal in the fat tended to decrease from 14.1° ± 7.8° to 0.5° ± 0.5° (P < .0001; a decrease of 96%). Table 2 provides detailed data.

Discussion

There was a remarkable convergence of the 3 independent models used in this study. The change in superficial suture stress on application of negative pressure was about 50% for both FEA models (55% for lateral tension in FEA-1, 45% for the superficial suture, and 47% for the epidermis in FEA-2). The reduction in lateral tension is key to maintaining the integrity of a closed incision. Both FEA models corresponded very well with the 43% to 51% increase in the force needed to stretch the bench model by 10 mm when negative pressure was applied. The FEAs showed that not only was there a decrease in the lateral tension around the incision with negative pressure, but also, the lines of tension were normalized and similar to preincision levels. The decrease in the magnitude and normalization of the stress distribution may allow for stronger contact of apposing incisional surfaces, which was corroborated by the bench model results.

In FEA-2, prior to application of negative pressure, the lines of stress in the tissue at the incision face are directed toward the sutures, resulting in a bowing out of the incision faces and thereby having less than optimal contact between the 2 faces of the incision (Figure 3B). This suboptimal contact can result in voids that lead to hematomas/seromas, which can further compromise the integrity of the overall incision site. On application of negative pressure, there was significant (>80%) normalization of stresses and their orientation along the entire interface as a result of reduction of the shear stress concentrations at the sutures. This resulted in the straightening of the previously bowed/curved interface and more uniform apposition of the interfacial surfaces. More specifically, in addition to the decrease in lateral stress at the superficial suture, there was a significant reduction (50%) in the lateral stress at the deeper suture, which is in the fat layer, along with a significant decrease in shear stress (97%) and angular deviation from normal (96%) in the fat layer. This suggests that even though there was no change in average lateral stress in the fat layers, application of negative pressure may keep the apposing faces of the fat layer in close contact. This implication is supported by the orientation data, which show a decrease from 14.0° to 0.5° in the deviation from the horizontal when negative pressure is applied. Lateral stress in the dermis was also affected similarly; there was minimal clinically relevant change with the application of negative pressure (14.5 kPa to 15.3 kPa), but there were significant decreases in the shear stress (82%) and angular deviation from normal (82%). Thus, reduction of stresses at the suture is critical in maintaining interfacial tissue apposition and, hence, the integrity of the incisional closure.

Mechanical cues are known to modulate biological responses at the macrolevel through the nanolevel. ³⁴ Reduction in incisional stresses has significant implications. It is known that excessive lateral tensions in and around the incision following surgery increase the likelihood of dehiscence, scarring, poor cosmesis, or a combination thereof. ³⁵ It is also generally accepted that collagen synthesis and organization are affected by mechanical stimuli. ³⁶ Mechanical forces are known to be involved in the fibroblast to myofibroblast transition ³⁷ and may contribute to the increased numbers of myofibroblasts and their activity in hypertrophic scars. ³⁸ Animal studies have shown that scarring may be partly a result of inhibition of cellular apoptosis as a result of excessive mechanical load. ³⁹ Therefore, therapies thought to decrease myofibroblast numbers in the healing tissue may potentially have an effect on cosmesis and functionality. Furthermore, tension can modulate directionality of vascular capillary sprouts,^40,41 and the reduction in shear stress in the epidermis, dermis, and fat may provide added clinical benefit—for example, reducing the risk of blistering. ⁴² Thus, the decreases in tension and normalization of stress distribution around the incision, as indicated by our models, may contribute to decreased risk of dehiscence and possible improvement in cosmetic outcome.

Human skin is adhered tightly to the body’s underlying structures and is under endogenous tension. ³⁴ As a result, incisional edges retract when an incision or laceration is made in skin, as also shown in the finite-element models used in this study. This native skin tension opposes the closure of the incision. In certain circumstances, such as in the lower extremities, in obese patients, or in locations where there is movement, this fact may result in excessive tension on sutures. ¹ These incisions are then at high risk for dehiscence. Therefore, therapies to help prevent complications in incisions would increase dermal to dermal contact (apposition) and decrease the excessive lateral tension around the incision line.

In this study, 3 distinct models were used to analyze the impact of CIM on the biomechanical environment around a closed incision. Although all models have inherent assumptions, constraints, and limitations, the results from the models used in this study are in concordance with reports of clinical findings.^4,5,12,43 The finite-element models used in this study were 2-dimensional and were restricted to evaluating a lateral slice through the center of the incision line. The models also assumed that the tissue substrates were flat; future analyses may consider the subtle implication of curved subjacent tissue surfaces. The models developed for this study used standard assumptions, such as material homogeneity and isotropy to make the problem tractable. However, the materials properties assigned to the various tissues were consistent with those observed in bulk tissues. ²⁷

Both FEA-1 and FEA-2 models assumed a clinically relevant 2-mm wide incision prior to application of sutures. The skin tension induced by sutured closure in FEA-1 (1.85 kPa) and FEA-2 (11.2 kPa) are in general agreement with the ranges used in Flynn’s parametric studies (2.07 to 13.14 kPa).^44,45 Frictional coefficients, dimensions, and materials properties were not considered as parametric variables (where a range of constants would be evaluated) in the FEA models in this study. Values used, such as the friction parameters, only affect the time taken for the model to reach equilibrium, or the magnitudes of the model outputs, but not the overall conclusion. Dynamic loading resulting from patient movement was not considered. Individual sutures were not modeled; instead, the suture lines (both superficial and deep) were modeled as single entities.

The silicone benchtop model provided experimental corroboration of the 2 FEA models. Additionally, the silicone model showed that incisions closed with staples had slightly greater strength than those closed with sutures—this is consistent with other reports in the literature. ⁴⁶ The use of a synthetic material like silicone provides tight control of the material properties, allowing for accurate force measurements at controlled strain rates without degradation of properties. Silicone is widely used as a soft-tissue proxy in biomechanical studies. ²³ The various silicones used to create the layers, as described in our Methods section, had compression tangent moduli similar to that reported in the literature. ²³ It is recognized that, unlike the FEA models, single-material (eg, silicone only) models do not capture the complex nonlinear hyperelastic properties of tissue at higher strains. This seems to have a limited effect on the results, based on the convergence of the results from the 3 independent models used in this study.

The results from this study support and explain the findings of a number of previously published studies. In their article describing several case reports, Gomoll et al ⁴ commented on the possible benefits of a semirigid foam (under negative pressure) bolstering the incision line—benefits that are supported and explained by the findings in this study. According to these authors, the reduced frequency of NPWT dressing changes (compared with other therapies) is particularly advantageous with larger, obese patients or with difficult-to-access incision locations and may also diminish the risk of incision contamination. Stannard et al ⁵ showed in a prospective randomized controlled trial that closed surgical incisions with hematomas drained in a significantly shorter time when treated with NPWT compared with the control pressure dressing (average 1.6 days vs 3.1 days, respectively; P = .03). Infection rates decreased from 16% to 8% with NPWT, consistent with the assertions by Gomoll et al. ⁴ In addition to the decreased frequency of dressing changes, the reduction in contamination may also be a result of decreased risk of incision-line dehiscence because of increased appositional strength of the incision. A third possible reason for the decrease in infection rates may be the NPWT-induced reduction in stress concentrations in tissue surrounding the insertion sites of sutures and staples, as suggested by the results of the current study. High stress concentrations can lead to ischemia, which reduces the ability to fight infection and can lead to necrosis. ⁴⁷

In conclusion, all 3 models used in this study—the 2 finite element models and the bench model—showed mutually corroborative results. Application of CIM mitigated and normalized the stress distributions around the closed incision in skin to homeostatic levels and directions. This modulation of the biomechanical environment around the incision site may facilitate enhanced apposition of the incision line and may be particularly beneficial for obese patients and those with prolonged swelling and incisions at sites over joints where there are increased stresses as a result of motion/movement. Future preclinical and human clinical studies are expected to corroborate these assertions.