Living Cells or Collagen Matrix: Which Is More Beneficial in the Treatment of Diabetic Foot Ulcers?

Ulceration occurs in 4%–10% of people with diabetes in the United States and leads to amputation in approximately 30% of patients with diabetes who are 40 years and older.¹ Each foot ulcer is associated with direct costs that can frequently exceed $45,000.² The American Diabetes Association estimates that more than one-half of all amputations can be prevented with proper patient therapy. Diabetic neuropathy leads to functional impairment of microcirculation and can result in foot tissue hypoxia and reduced healing capacity, even in the presence of adequate blood flow. The standard treatment recommended by the American College of Foot and Ankle Surgeons consists of extensive debridement of necrotic tissue, treatment of infection, non-weightbearing or offloading techniques that decrease the pressure applied on the affected extremity, and arterial revascularization, if indicated.³ Topical treatments for diabetic foot ulcers include enzymatic debridement, total contact casting, hyperbaric oxygen, antibiotics, growth factors, and dressings that maintain a moist wound environment while also protecting granulating tissue from mechanical injury. Recently, several tissue/cell-based (human skin equivalents) wound dressings have been developed and are undergoing clinical evaluation as well.
The present study compared OASIS^® Wound Matrix [ECM] (Healthpoint, Fort Worth, Tex), which is an acellular, bioactive, collagen-based wound dressing made from porcine intestinal submucosa to Dermagraft^® [LSE] (Advanced BioHealing, La Jolla, Calif), a living skin equivalent consisting of human fibroblast cells derived from newborn foreskin tissue that have been seeded onto a bioabsorbable polyglactin mesh scaffold.
In a randomized, non-blinded study, the two materials were used for treating full-thickness diabetic ulcers as an adjunct to standard therapy that consisted of wound debridement and use of saline moistened gauze dressings. The focus of the study was to compare patient outcomes following diabetic ulcer treatment with these two materials. The relative rate of wound closure when an acellular collagen-based matrix was compared to a living cellular product became a point of particular interest in the course of this study. The living cellular product (LSE) was essentially devoid of collagen, but contained living cells. The authors hypothesized that there would be no difference in the rate of wound closure or in the percentage of wounds that achieved closure when comparing treatment with an acellular collagen-based matrix versus a living cellular material.

Materials and Methods
This study involved 4 study sites (Weil Foot and Ankle Institute, Des Plaines, Ill; Coastal Podiatry, Inc., Virginia Beach, Va; Ocean County Foot and Ankle Surgical Associates, Toms River, NJ; The Foot and Ankle Institute of South Florida, South Miami, Fla) in which patients were randomized to treatment with either an acellular, porcine-derived, bioactive, collagen matrix material ([ECM], OASIS), or a living skin equivalent ([LSE], Dermagraft) consisting of human fibroblast cells derived from newborn foreskin tissue that have been seeded onto a bioabsorbable polyglactin mesh scaffold.

Patients were screened to determine if they met the inclusion and exclusion criteria (Table 1). Following a 1-week phase-in period, subjects were randomized to treatment with either living fibroblast cells or with collagen matrix. Patients were examined at least once weekly for the first 8 weeks, and subsequently, every other week until closure was achieved, for up to 12 weeks. An extended observation period of 8 additional weeks followed. In cases where wound closure was confirmed, the wounds were re-evaluated 1 week later to reconfirm closure.
The critical endpoint of this study was wound closure, and was defined as full epithelization without any evidence of drainage or bleeding. A central monitor (Clinical Science Corp., Skokie, Ill) logged all adverse events that occurred.
Randomization was accomplished when the investigative site identified a qualified candidate and contacted an independent site (MED Institute, Inc., West Lafayette, Ind) that randomly assigned patients to one of the two study arms. Randomization was performed across all study sites—the result was a balanced pool of subjects between study groups.

All wounds were debrided to remove all callous and any necrotic tissue where possible. The wounds were thoroughly cleansed, and either the ECM or the LSE was applied to the wound. Each wound was then backed by saline moistened gauze, which was left in place for 1 week. Both dressing materials were prepared and applied according to the each manufacturer’s specifications.
Offloading consisted of a well-padded fixed ankle removable boot to be worn at all times while weightbearing. In 2 cases where the boot could not be tolerated, the patients were permitted to wear a diabetic shoe, which consisted of a stiff, extra-depth shoe that was lined with the same plastizote material. The liner of the shoe gear or fixed ankle boot was examined at each office visit to confirm that offloading was performed properly. Additional accommodations were made as necessary.
Patients were given a maximum of 3 LSE grafts or 8 ECM dressings. Less LSE grafts were used because of proposed Medicare treatment standards in effect at the time of the study, and the relatively high cost of the LSE compared to the ECM product.
Forty potential subjects were screened, 31 of whom met the study criteria. Of this group, 26 study subjects completed the study. Study subjects who met all inclusion and exclusion criteria were randomized to treatment with either ECM or LSE. Since patients were gathered from 4 different sites, training of the study investigators was provided in a uniform manner at each location to ensure consistent treatment and application techniques. Data collection was monitored by the Clinical Science Corporation (Skokie, Ill).
The size of the wound was documented before initiating treatment and at follow-up visits. Each target ulcer was photographed before and after cleansing and debridement. The ulcer was measured after debridement. In addition, ulcer location, duration, and a description of the wound base was also recorded.
The wounds were irrigated with normal saline and then the ECM or LSE was placed over the debrided wound. Each product was applied in accordance with the manufacturer’s recommendations.
Wounds were evaluated at 1, 2, 3, 4, 6, 8, 10, and 12 weeks. If closure did not occur until the 12th week, then an additional wound check was done at the 13th week. Follow-up examinations were performed at week 16 and week 20.
The ECM product was reapplied to the wounds in the ECM group if more than one-half of the graft was not adhering to the wound. If a caramel colored gel was observed on the wound, this was rinsed using gentle irrigation with sterile saline, leaving the remaining material intact. Before reapplying the ECM product, non-adherent portions of the ECM were trimmed before a new piece was reapplied. Up to 8 ECM applications were permitted.
The LSE could be reapplied 2 more times (total of 3 applications) at weeks 2 and 4 only if closure had not been achieved in wounds in the LSE study arm.
At subsequent weekly evaluations, the wound was cleaned, evaluated, photographed, and measured at each time point. The wound closure time was considered when full epithelialization and no wound drainage were observed.
The time required for complete healing of each wound was recorded. The average time for healing was calculated within each study arm. The time required for complete wound healing was compared between the ECM and LSE arms with a confidence level of 95% using a 2-sided Student’s t-test.

Results
Subjects in each group were examined for demographic characteristics. Each group contained 13 subjects, and contained no statistically significant differences (Table 2). Average wound size at study initiation was 1.85 cm² ± 1.83 cm² in the ECM group (n = 13), compared to 1.88 cm² ± 1.39 cm² in the LSE group (n = 13). There was no statistically significant difference in initial wound size (P = 0.94).
In this study, 10 wounds closed in the ECM group (76.9%), with an average time to closure of 35.67 ± 41.47 days, while 11 wounds closed in the LSE group (84.6%), with an average closure time of 40.90 ± 32.32 days. There was no statistically significant difference in the time to closure between these groups (P = 0.73).
Average number of LSE dressings applied was 2.54 ± 0.78, while the average number of ECM applied was 6.46 ± 1.39. This translates to an estimated cost of $3505 for LSE dressings ($1380 each) and an estimated cost of $807 for ECM dressings ($125 each; Figure 1).
Figure 2 shows a Kaplan-Meier curve, which compares probability of closure between ECM (Group 1) and LSE (Group 2). The curves for both groups are essentially indistinguishable: P = 0.37.

Discussion
In this study, no significant difference was found in the time to closure or in the percentage of wound closure between the two study groups. Additionally, there was no significant difference in the demographics of the subjects enrolled. However, the power of the pilot study presented here is low (6.5%), indicating that the sample size would need to be larger to determine if the two treatments presented here are truly equivalent.
Nonetheless, there are several important differences to note between these groups. The application of ECM (OASIS) was significantly easier then the application of LSE (Dermagraft), primarily due to variations in the process required for preparing each material. The ECM, which is stored on the shelf as a dry material, simply required hydration, while LSE had to be ordered in advance, and required a series of steps to defrost and apply the material. The LSE material would expire and have to be returned if not used immediately. Consequently, the costs associated with shipping and handling of the LSE made this product substantially more expensive than ECM. In this study, material costs averaged $807 ($125 per piece) for ECM, and averaged $3505 ($1380 per piece) for LSE. (Please note: the costs cited here are based on the actual cost of materials today. When the study was originally conducted, the cost of the ECM material was $28 per piece). However, in actual clinical practice, this cost difference would probably be much more substantial since there was a limit of 3 LSE grafts applied in this study. Also in actual clinical practice, Medicare and other providers would approve up to 8 applications at an estimated cost of $11,040 for the LSE material alone.
Although each material was applied to the surface of diabetic foot ulcers, the products are dramatically different. The LSE provides living fibroblasts to the wound surface, which can act to generate and release growth factors to the wound bed, thereby stimulating angiogenesis, mitogenesis, and chemotaxis.^4,5 Conversely, the ECM acts as a collagen bioscaffold, providing a substrate for vascular infiltration.⁶ In every case, there are wounds that are in need of cellular and vascular proliferation, while others have already developed a granular base and need a substrate to stabilize these components in order to gain structural integrity. In essence, the bed of a wound is analogous to bricks and mortar. Sometimes a wound needs “bricks” (collagen scaffold) and sometimes the wound needs “mortar” (angiogenesis and cellular components). Based on the results of the present study, one can be certain that wound healing is enhanced with either component. The average closure time of approximately 7 weeks found in the present study represents a significant improvement over standard care with saline-moistened gauze, where wound closure times have been reported in the 9- to 12-week range.⁷
It is also likely that both products overlap in their actions. Fibroblast-generated cellular components ultimately trigger a cascade of events that will lead to collagen assembly and accumulation. Similarly, collagen is capable of triggering a biochemical cascade that leads to angiogenesis, mitogenesis, and chemotaxis.⁸ This is particularly true with type III collagen, which has been compared to stem cells for their ability to trigger the accumulation of various tissue components.⁹ Type III collagen represents approximately 23% of the collagen found in the human body at the fetal stage, and gradually decreases as birth approaches. By adulthood, type III collagen is replaced with type I collagen, leaving behind approximately 1% of type III collagen.^10,11 The ability of this small percentage of type III collagen to stimulate wound healing illustrates the value of this component in wound healing.

Conclusion
Wound closure is enhanced by the addition of either collagen “bricks” or living cell “mortar.” The authors believe that all wounds have deficiencies that must be addressed to facilitate wound healing. A highly granular wound bed would likely benefit from the addition of collagen to stabilize this vascular tissue, and would provide integrity for the continued advancement of the wound. Similarly, wounds lacking cellular proliferation and active development of granular tissue would more likely benefit from the addition of living cellular components. Ultimately, the ability to achieve wound closure appears to be highly dependent on the clinician’s ability to assess the needs of the wound at hand.