A Multicenter, Blinded, Randomized Controlled Clinical Trial Evaluating the Effect of Omega-3–Rich Fish Skin in the Treatment of Chronic, Nonresponsive Diabetic Foot Ulcers

The economic burden of diabetic foot ulcers (DFUs) remains significant in the United States. Historically, annual costs for the treatment of DFUs through Medicare alone have ranged from $6 billion to $19 billion; inclusion of the amount paid by insurance payers could place the total annual costs for DFU management in the United States at twice that level.¹ In 2010, it was estimated that the mean cost of DFUs per episode of care from a health care public payer perspective was $17 245.² A systematic review of cost-of-illness studies published in 2017 reported that the mean cost of DFUs per episode of care from a health care public payer perspective was $31 024, an amount considerably higher than the previously estimated total.³ Of course, the severity of the ulcer and whether it heals affects the cost; another 2017 study from Europe stated that average in-hospital costs were $10 827 USD (range, $702-$82 880) per DFU episode. On avergae, primary healed DFU costs were $4830, single minor amputations were $13 580, multiple minor amputations were $31 835, and major amputations were $73 813 per episode. Costs differed significantly between groups (P < .001).⁴ The most often cited reference on the efficacy of standard of care (SOC) showed approximately 70% of DFUs heal with SOC treatment; at least 30% become chronic wounds.⁵ In addition, SOC treatment is a lengthy process with low efficacy. The percentage of wounds healed with SOC therapy is 24.2% by 12 weeks and 30.9% after 20 weeks.⁵ Over the past 20 years, multiple prospective studies placed SOC closure rates (excluding the use of total contact casting) at 20% to 30%.^6,7

A chronic, nonresponsive diabetic foot wound is at high risk for infection, with the potential of leading to lower extremity amputation, even if the wound is not severe.⁸ Therefore, once a DFU forms, timely healing is necessary to stave off infection. A DFU may begin superficially and then spread to the contiguous subcutaneous tissues. Left untreated, the infection will spread to muscles, tendons, bones, and joints, and then progress to septic gangrene, which eventually leads to lower extremity amputation.^9,10 The difference in cost between infected DFUs with and without amputation is significant, depending on the amputation level.⁴ Patients who have undergone a major amputation have a 65% 4-year mortality rate, while those who have undergone a minor amputation have a 45% 4-year mortality rate.¹¹ Older studies indicated that patients who underwent a major amputation had a 68% risk of needing another amputation in the next 5 years.¹² A 2018 study indicated an incidence of contralateral major amputation of 4.8 per 100 person years.¹³ Therefore, the availability of advanced therapies is crucial to improve healing rates, reduce the risk of amputation, improve patient outcomes, and decrease treatment costs. 

Cellular and/or tissue-based products (CTPs) are increasingly used as advanced therapeutics in wound care treatment. One category of these CTPs is extracellular matrices (typically xenografts). These CTPs are made from animal tissues, such as bovine skin, porcine intestinal submucosa, and ovine stomach, and they provide structural support and biological molecules that can modulate wound healing. The microstructure of these products is highly variable, however, owing to differing tissue origins and processing methods. The device used in the present study, fish skin graft (Omega3 Wound; Kerecis), is the intact skin of Atlantic cod, which has been decellularized and sterilized. It closely resembles human skin in composition and structure. The fish skin graft is thicker, however, with a porosity and 3-dimensional microstructure that provide a foundation for efficient ingrowth of dermal and epidermal cells as well as for supporting vascularization. Moreover, the unique biomechanical properties of fish skin promote cell proliferation and differentiation, which are hallmarks for tissue regeneration.¹⁴ 

The fish skin graft is made from the skin of wild-caught Atlantic cod originating from North Atlantic Icelandic fishery. This source yields a dehydrated product that is rich in bioactive compounds, especially omega-3 polyunsaturated fatty acids. While devoid of fish scales, fish skin graft retains 3 basic layers of skin: epidermis, dermis, and hypodermis.¹⁵ In 2 double-blind, prospective, randomized clinical trials of acute wound healing use of this fish skin graft resulted in significantly faster healing outcomes compared with use of porcine intestinal submucosa and dehydrated chorion amniotic membrane.^16,17

Several prospective and retrospective studies involving DFUs treated with fish skin grafts have demonstrated its effect on wound closure. In a study by Dorweiler et al,¹⁸ weekly application of the fish skin graft to 25 amputated and bone-exposed wounds resulted in closure of 17 wounds (68%) between 9 weeks and 41 weeks. A reduction of analgesics intake was also noted after initiation of fish skin graft treatment. In a study in which 8 patients with diabetes underwent forefoot surgery followed by applications of fish skin graft once a week for 6 weeks, Woodrow et al¹⁹ reported the wound area was reduced more than 84.9% at 6 weeks with no complications. In 2019, Michael et al²⁰ published a retrospective study of 51 patients with 58 full-thickness DFUs that were treated with fish skin graft. The study compared the initial wound surface area at first application of fish skin with the final surface area after a 16-week treatment period. At the 16-week endpoint, a mean reduction in surface area of 87.57% was noted, and 35 of 58 wounds (60.34%) were fully healed. A greater than 90% reduction in surface area was measured in 43 wounds (74.14%), and a greater than 75% reduction was seen in 49 wounds (84.48%). In the 35 wounds in which full healing was achieved, the average number of applications of fish skin was 4.9 and the median time to full healing was 10 weeks. 

Although these data are enlightening, this prospective, randomized, controlled trial is intended to further the research on this topic. Therefore, the present authors sought to evaluate the fish skin graft in the care of DFUs. Clearly, randomizing patients with DFUs to gauze and saline as a comparator is not appropriate. Thus, an advanced wound care product was sought. Collagen alginate dressing (Fibracol Plus Collagen Wound Dressing with Alginate; 3M) is an innovative wound dressing technology that has been established in clinical trials^7,21 as a competitor to biologic wound products. This wound care device is composed of collagen and calcium alginate fibers. It contains 80% more collagen than the original collagen alginate dressing. The unique combination of natural biopolymers created by a patented process incorporates the structural support of collagen and the gel-forming properties of alginates into a sterile, soft, absorbent, conformable topical wound dressing. In the presence of wound fluid, collagen alginate dressing maintains a moist microenvironment at the wound surface that is conducive to the formation of granulation tissue and to epithelialization, thereby enabling rapid healing.^22,23 

This study was designed to assess the efficacy of the fish skin graft in the treatment of resistant DFUs in comparison with collagen alginate dressings, henceforth referred to as SOC. 

This prospective, multicenter, parallel-group, randomized, controlled trial was designed to collect patient outcome data for the treatment of DFUs. The trial was single-blinded concerning wound healing assessments; the confirmation of wound healing was overseen by an independent wound care adjudicator. This study is registered with ClinicalTrials.gov (ID: NCT04133493).

Both groups (arm 1 and arm 2) underwent SOC treatment, which consisted of offloading of the DFU with a walker (A-W0800BLK Walker Brace Equalizer; Össur America-Royce Medical). The protocol allowed for the use of total contact casting if there were adherence issues or if the patient's foot was too large for a controlled ankle motion boot; however, no patient in either cohort was treated with this option. All patients in both cohorts underwent sharp debridement at the first visit of the run-in phase and the first visit of randomization. Prospective participants were excluded if they were being treated with systemic antibiotics at the time of randomization. Systemic antibiotics were permitted during the treatment phase only for managing infection in connection with debridement. 

Patients in study arm 1 received SOC treatment plus application of fish skin graft secured with surgical adhesive strips, sutures, or staples and covered with a nonadherent dressing (ADAPTIC Non-Adhering Dressing; 3M). The fish skin graft was bolstered down to the wound bed with gauze or an open-cell foam, covered with a moisture-retentive foam dressing and hydrogel as needed to retain adequate moisture balance, and padded with stretch gauze and self-adherent wrap. The fish skin was reapplied and the dressing changed by the site investigator once a week. 

Patients in study arm 2 received SOC treatment only, consisting of a wound care covering of collagen alginate dressing followed by a padded dressing comprising 4 in × 4 in gauze pads, stretch gauze, and self-adherent wrap. The wound was dressed by the site investigator once weekly and by the patients or their caregivers at home 3 times weekly. This frequency of dressing changes was established per the product instructions for use (Figure 1).

Patient recruitment and randomization

The complete inclusion and exclusion criteria used by site investigators to screen patients for study eligibility are listed in Table 1. 

Patients were required to have had a DFU for a minimum of 4 weeks and demonstrate adequate renal function and adequate perfusion to the affected extremity (Table 1). Prior to randomization, patients who met the inclusion criteria were first treated with SOC for a 2-week screening period, during which they were evaluated on-site weekly for ulcer assessment and/or measurements as well as sharp debridement. During the first screening visit, patients underwent a comprehensive physical examination and their medical history was documented. If multiple ulcers were present, the largest ulcer was selected as the study ulcer (referenced as index ulcer).

All wounds underwent a 14-day roll-in period in the screening phase in which comprehensive SOC was administered. This included collagen alginate dressings, gauze, a soft roll, and a compressive dressing applied to the ulcer. 

During the roll-in period prior to randomization, all patients underwent debridement, offloading, and moist wound care. If the index ulcer reduced in area by 20% or more after 14 days of SOC, the patient did not undergo randomization and was excluded from the trial as a screening failure; the patient was not considered treatment resistant. Any patient who had signed a consent form and been assigned a screening number but had not been randomized was classified as a screen failure. If the wound area was reduced by less than 20%, the patient was randomized and enrolled in the study.

Procedures

The treatment phase (≤ 12 weeks) began with a series of assessments designed to confirm the patients’ continued eligibility. As necessary, the investigators debrided the ulcers in accordance with SOC. Patients whose ulcers continued to meet eligibility criteria were then randomized to 1 of 2 groups—study arm 1: SOC with fish skin graft; and study arm 2: SOC with collagen alginate dressing. Additionally, for the most compassionate care, ulcers in either arm that had not reduced by 50% at week 7 were considered a treatment failure and those patients were permitted to exit the study and obtain additional outside treatment. See Figure 2 for the complete study flowchart. 

Validation of healing

Complete ulcer healing was based on the site investigator’s assessment, as evidenced by complete (100%) reepithelialization without drainage and need of dressing. A follow-up validation visit was conducted 1 week after ulcer closure was first observed to confirm durability of closure.

An independent panel of wound care experts who were blinded to the patient allocation process and the principal investigator’s assessment reviewed all study-related decisions made by the site investigators and confirmed healing status. The validation team included a vascular surgeon, a podiatrist, and an internal medicine specialist.

Study outcomes

The primary outcome of the study was the comparison of the proportion of index ulcers healed at 12 weeks; wounds were deemed either healed or not healed. Secondary outcome measures included time to heal (for DFUs that healed) and wound area reduction by percentage at 12 weeks.

Sample size calculations and statistical analysis

All testing for endpoints was two-sided, with alpha set at .05 as the level of significance for demographic comparisons between arms and interim analysis; .047 was set as the significant level for final analysis if statistical testing was conducted for primary and secondary endpoints. The null hypothesis of this study was that the proportion of wounds healed at 12 weeks would be equal for arm 1 and arm 2, after up to 12 weeks of SOC with fish skin graft or up to 12 weeks of SOC with collagen alginate dressing. 

The primary analysis consisted of the proportion of index wounds closed at 12 weeks (arm 1 vs arm 2) using the χ² test. Analysis was adjusted using logistic regression.

The secondary analysis consisted of 2 factors: (1) time to heal within 12 weeks using Kaplan-Meier analysis, and (2) percentage wound area reduction at 12 weeks (arm 1 vs arm 2) using the Mann-Whitney test. Analysis of tertiary endpoints did not use the comparative statistical analysis.

This study took place from June 20, 2019, to November 30, 2020. After giving their consent to participate in the study, 58 patients were screened. Of these, 9 patients exhibited greater than 20% healing during the screening period and were excluded from the study. The 49 patients who were eligible to participate were randomized into arm 1 with SOC plus fish skin graft or arm 2 with SOC. All patients received their assigned intervention and were included in the intent-to-treat analysis. 

Demographics

No significant difference was noted between the study groups in terms of demographics, renal function, or blood glucose. Table 2 reports the patient demographics. 

Wound closure and area reduction rates

As noted, 49 patients were included in the final analysis. At 6 weeks, the percentage area reduction was 72.8% in arm 1 and 41.2% in arm 2 (P = .044). At 12 weeks, 67% of the index wounds in arm 1 had fully closed, compared with 32% in arm 2 (P = .0152; N = 49) (Figure 3, Figure 4). The time to closure for the healed wounds was about equal in each arm (6 weeks). The median number of applications to achieve closure was 5 in arm 1.

Wound area reduction was calculated by comparing the wound area after 6 weeks and 12 weeks with the starting area at week 1. If the wound area reached 0 before 6 weeks or 12 weeks, the wound was considered 100% reduced in area. For patients who dropped out of the study before the proposed timepoint, the data were kept in the study record and reason for discontinuation was documented, but the data were not included in the results. At week 6, the number of participants was 24 in arm 1 and 25 in arm 2.

Per protocol, patients exited from study if their wound had healed less than 50% at 6 weeks and was not considered healed at 12 weeks. These patients were not included when wound area reduction at 12 weeks was calculated. At 12 weeks, the number of participants was 21 in arm 1 and 13 in arm 2. At 12 weeks, the area reduction was 97.3% ± 6%  and 76.8% ± 35.3%, respectively. A test for variances between the 2 arms was performed and determined the 2 treatment arms had unequal variance. A two-tail student t-test was performed for unequal variance groups, and a P value of less than 0.05 was established as statistically significant. The P value was 0.06 at 12 weeks.

Given the previously noted costs of treating DFUs,^2,3 determining the most cost-effective means of improving wound closure should be a top priority. When evaluating CTPs, their primary use in DFUs is to enhance the percentage and speed of closure. The number of applications of the CTP multiplied by its cost will equal the cost of the therapy.  In this trial in which treatment using SOC plus fish skin graft was compared with SOC using collagen alginate dressing, both methods demonstrated an improvement in the wound-healing trajectory. However, there was a significant two-fold enhancement in healing rate for DFUs compared with SOC treatment alone. The xenograft is cost effective to produce and has been shown to have the potential to reduce the cost of DFU treatments compared with SOC.²⁴

The premarket approval process is a rigorous prospective, randomized trial system, which when successfully navigated means that approved products are indicated specifically for the treatment of diabetic foot. The 4 currently approved products are a bi-layered bioengineered skin substitute (Apligraf; Organogenesis, Inc), cryopreserved human fibroblast derived dermal substitute (Dermagraft; Organogenesis, Inc), dermal regeneration matrix (Omnigraft; Integra LifeSciences), and platelet-derived growth factor (PDGF; Regranex; Smith+Nephew). The bi-layered bioengineered skin substitute and cryopreserved human fibroblast derived dermal substitute are cell-containing therapies that are expensive to produce. When compared with conventional therapy in randomized controlled trials, they resulted in 56% and 30% complete wound closure at 12 weeks, respectively.^25,26 Both products required multiple applications to achieve closure. The dermal regeneration matrix is a bioengineered product that demonstrated 51% closure rate compared with saline gel at 16 weeks.²⁷ In this trial, the median number of applications to achieve closure was 1, which enhanced the cost effectiveness. The PDGF data for DFUs showed a wound closure rate of 50% at 16 weeks vs 35% for the placebo gel.²⁸ This product is applied daily and has best results when coupled with daily debridement. By comparison, the present similarly designed trial using a xenograft made of acellular fish skin showed similar results, with 67% of patients' DFUs being healed over 12 weeks vs 32% with SOC. Therefore, the authors believe acellular fish skin is appropriate to include in the care plan of patients with DFUs that fail to respond to offloading after 4 weeks. 

The potential mechanisms of the enhanced healing properties of fish skin graft have been studied in vitro.²⁹ The fish skin graft is gently processed and provided as lyophilized acellular full-thickness fish skin that has undergone terminal sterilization. It retains a distinct porous and homologous scaffold for skin graft that is non-allogeneic and gentle to humans. The natural structure and chemical complexity of the fish skin provides optimal conditions for cell ingrowth. In addition, this general processing allows for the inclusion of fatty acids in the product, which is unique and differs from other grafts.²⁹ Due to the processing necessary for mammalian xenografts, there is no fat left in those products. Animal research studies have shown that the omega-3 fatty acids that remain in fish skin graft have an inhibitory effect on bacterial growth and improve epithelial cell migration.³⁰ Theoretically, the ongoing application of a product that has a naturally occurring bacterial static component and modulates the inflammatory response is ideal.³¹ Inflamed or infected wounds are associated with pain. Previous studies found a decrease in postprocedural pain with the acellular fish skin graft,³² which could improve adherence with therapy.

The authors are aware that clinical trials attract adherent patients who fit a narrow spectrum of care. In DFU trials, such patients often have well-controlled hemoglobin A1c and are adherent with offloading. This prospective, randomized trial is no different in that regard, but the model used is the best option for evaluating a product such as the one studied herein. It is not possible to blind either the person applying the product or the patient to the material being applied. Thus, this trial included a third investigator who was blinded to evaluation of closure. Intrinsic bias remains a concern owing to the patient and caregiver knowing the therapy being applied. 

Both study arms received a once-weekly visit that included debridement, reapplication, and dressing change in the clinic; patients in the SOC arm were additionally allowed dressing changes at home, whether by themselves or a caregiver. Such dressing changes were done in accordance with the frequency and care level presented by the manufacturer in the instructions for use for the specific product. Additional dressing changes at home can expose a wound to unknown factors, potentially resulting in stalled healing. To minimize the risk of delayed healing, patients and/or their caregivers were assessed for adherence, and both verbal and written instructions were given. In most cases the caregiver was a licensed home care nurse. 

This trial is appropriately powered to show that in DFUs defined as superficial ulcers not involving tendon capsule or bone, the addition of the fish skin graft to a treatment protocol consisting of appropriate off-loading and debridement can result in a significantly improved closure rate compared with SOC with collagen alginate dressing. This finding, in combination with the findings of a cost analysis by Winters et al,²⁴ which showed the fish skin graft is not only more effective at healing DFUs than SOC but also results in lower overall costs due to reduced hospitalization, fewer amputations, and a lower chance of ulcer recurrence, indicates that fish skin graft is an attractive therapy for DFUs.²⁰ The supporting retrospective studies as well as the findings from this prospective, randomized, controlled trial indicate the fish skin graft should be included in the list of CTPs for patients with DFU for whom 4 weeks of standard off-loading has not resulted in appropriate wound area reduction.

Authors: Eric J. Lullove, DPM¹; Brock Liden, DPM²; Christopher Winters, DPM³; Patrick McEneaney, DPM⁴; Allen Raphael, DPM⁵; and John C. Lantis II, MD⁶

Affiliations: ¹West Boca Center for Wound Healing, Coconut Creek, FL; ²Surgical Services, Berger Health System, Circleville, OH; ³Department of Surgery, St Vincent Hospital, Indianapolis, IN; ⁴Northern Illinois Foot and Ankle Specialists, Crystal Lake, IL; ⁵Village Podiatry Centers, Smyrna, GA; ⁶Department of Surgery, Icahn School of Medicine, Mount Sinai Morningside and West Hospitals, New York, NY

Contributions: Principal investigators in this study were Eric J. Lullove, DPM; Brock Liden, DPM; Christopher Winters, DPM; Allen Raphael, DPM; Bert Altmanshofer, DPM; Patrick McEneaney, DPM; Belinda Marcus, MD; Igor Zilberman, DPM; and Madelin Ramil, DPM.

Correspondence: John C. Lantis II, MD, Chief, Division of Vascular and Endovascular Surgery, St Luke’s - Roosevelt Hospital, Vascular/Endovascular Surgery, 1111 Amsterdam Avenue, MU 208, New York, NY 10025; JLantis@chpnet.org or John.Lantis@mountsinai.org 

Disclosure: Statistical analysis was conducted by Michelle LaPradd (Pi-Squared, LLC). Dr John C. Lantis II provides consulting services to Kerecis LLC, including significant input into the design, conduct, and interpretation of this study.