Efficacy of Dehydrated Human Amniotic Membrane Allograft for the Treatment of Severe Extravasation Injuries in Preterm Neonates

Introduction. A peripheral intravenous (PIV) catheter is placed in 60% to 70% of neonatal intensive care unit (NICU) infants. Extravasation injuries occur in 18% to 33%, with 70% in neonates < 27 weeks of gestational age. Despite such frequent use of PIV therapy, evidence on best practice, injury prevention, management, and treatment of extravasations is lacking. Objective. This case series of 4 neonatal patients describes the experience and efficacy of using a dehydrated human amniotic membrane allograft (dHAMA) in the treatment of severe extravasation injuries. Materials and Methods. The 4 preterm, critically ill neonates, all with stage 4 extravasations, were treated with 1 to 2 applications of the dHAMA to facilitate the repair process. Prior to treatments, standard of care included either enzymatic (collagenase ointment) or autolytic debridement (active Leptospermum honey) followed by mechanical debridement prior to allograft placement. Results. The 4 full-thickness wounds exhibited recalcitrant healing. The dHAMA invigorated the wounds after standard management failed to induce repair. Application was easy and follow-up care was minimal. All wounds healed without contractures and with minimal soft scars and normal pigmentation at the 1- to 2-month follow-up visits. Conclusions. The dHAMA proved to be an effective, safe, and easy-to-apply treatment in this case series, leading to regeneration and healing of deep neonatal wounds associated with extravasations.

Every year 1 in 10 babies is born prematurely.¹ Advances in care and technology have allowed previously nonviable neonates to continue their journey, despite extreme prematurity and impairments of various organs. The most common procedure used in the neonatal intensive care unit (NICU) is peripheral intravenous (PIV) catheter placement.² Intravenous (IV) route of medication delivery revolutionized medicine but not without complications; the most common complication is the delivery of infusate into the subcutaneous tissues instead of the vein.² Incidences of neonatal infiltrations vary. A PIV is placed in 60% to 70% of infants admitted to the NICU.² Extravasation injuries occur in 18% to 33% of PIVs, with 70% in extremely preterm babies of < 27 weeks gestational age.³ Despite such a frequent use of PIV therapy, evidence on best practice, injury prevention, management, and treatment of extravasations is less than optimal. The lack of adequate visual description of the damage caused by internal mechanical and chemical forces necessitates extrapolation of the treatment area based on the original IV placement site and physical exam. Treatment options are limited; many are based on adult or pediatric studies.³

In the authors’ experience, full-thickness injuries occur in 10% to 15% of severe neonatal extravasations. Neonates have strong regenerative capabilities, but many factors can contribute to a slow healing trajectory: immaturity, critical illness, inability to mount a strong cellular response, suppression of immune system, and often lack of knowledge of neonate-compatible skin products.⁴ The risks of a protracted open wound are numerous: secondary infection, pain, prolonged hospital stay, scar formation, functional limitation, parental angst, and potential for future litigation.⁴ Wound healing typically occurs in an organized, step-wise trajectory, but certain phases may become prolonged and lead to recalcitrant wounds. The principles of wound healing should be applied to such wounds to optimize their healing potential.

The amniotic membrane has been used in medicine since the early 20th century.⁵ Gynecology, ophthalmology, burn, surgery, and neurosurgery are a few adult disciplines that use this modality; pediatric burn centers are starting to utilize unique properties of amnion. AMNIOEXCEL (dHAMA; Integra LifeSciences, Plainsboro, NJ) is a minimally manipulated dehydrated human amniotic membrane allograft (dHAMA) with intact, malleable tissues that retains important components for dynamic reciprocity of healing. As described by Schultz et al,⁶ dynamic reciprocity is an ongoing, bidirectional interaction among cells and their surrounding microenvironment. It is especially important during healing as certain chemicals and cells play pivotal roles in regulating tissue regenerative responses.⁶

Amniotic membrane provides an anchor for a developing fetus throughout pregnancy; its structural integrity and metabolic plasticity allow change, growth, and remodeling under the governing of paracrine factors.⁵ In the same way it structurally supports a developing fetus, dHAMA offers a scaffold for wound healing elements, providing anti-inflammatory effects, promotion of cellular differentiation and adhesion, infection suppression, neovascularization, pain suppression, epithelization without immunogenicity (paramount to the immature preterm immune system), and an antiscarring effect.

Application of new innovative biologics in the preterm population is in its infancy. There is limited data on neonatal wounds, treatments, and outcomes in general. The purpose of this study is to describe the authors’ experience with and the efficacy of dHAMA in the treatment of severe extravasation injuries in the neonatal population.

Four preterm, critically ill neonates with stage 4 extravasations were treated with 1 to 2 applications of dHAMA to facilitate the repair process. Prior to treatments, standard of care included either enzymatic (collagenase ointment) or autolytic debridement (active Leptospermum honey) followed by mechanical debridement with monofilament polyester fibers mitten (Debrisoft Lolly; Lohmann & Rauscher, Milwaukee, WI). Wounds were covered by a secondary silicone-based atraumatic dressing.

The dHAMA is donated by prescreened mothers during planned Caesarean sections (C-sections). The preservation of growth factors and anti-inflammatory cytokines (ie, platelet-derived growth factors, vascular endothelial growth factor, transforming growth factor beta, interleukins 4 and 6, tissue inhibitors of metalloproteinases [TIMPs], and others) is ensured via the DRY flex (Integra LifeSciences) process of membrane preparation.

Case 1

A 28-week-old male infant was born via C-section at an outside tertiary hospital in Long Island, New York, where his course was complicated by respiratory distress syndrome, a large patent ductus arteriosus (PDA), and anemia of prematurity. At 45 days of life (DOL), he was transferred to Steven and Alexandra Cohen Children’s Medical Center (New Hyde Park, NY) for a cardiothoracic surgery consult with subsequent PDA closure surgery. On postoperative day (POD) 2, the infant was on a high-frequency oscillator from a PDA ligation at the time of the extravasation.

Due to surgery, the infant's food and fluids were withheld (nothing by mouth; NPO), and he received total parenteral nutrition (TPN; 10% dextrose, 4% amino acids) via a PIV in his right foot. The extravasation site included a large blister (6 cm x 4 cm) with surrounding areas of necrotic skin (Figure 1A). There was a pressure injury on the dorsum of the foot from the hub of the device as well as gross edema from the foot to the calf area. He received treatment with hyaluronic acid as per unit’s protocol. The large blister was decompressed twice in the first 48 hours with a butterfly needle while keeping the outer covering intact and avoiding deroofing the area. Within 3 days, it was clear the outer skin was necrotic and would likely slough off, but the depth of injury was not initially obvious.

The accepted open wounds protocol of using Leptospermum honey, to ensure moist healing, antimicrobial effects, debridement, granulation, and epithelization enhancement, covered by secondary silicone-based dressings with daily wound assessment were followed (Figure 1B). Over a 14-day period following honey and dressing application, a full-thickness wound (level of tendon) was debrided twice, once with a curette and once mechanically due to significant slough formation (Figure 1C). The wound was not infected, but the infant was on a 7-day course of systemic antibiotics postoperatively.

After 2 weeks of the aforementioned accepted treatment and failure to close (wound size: 7.5 cm x 4 cm), the dHAMA (measuring 8 cm x 4 cm) was applied following mechanical debridement. Eight days after dHAMA application, the wound showed significant improvement (5 cm x 3 cm) as well as granulation tissue growth (Figure 1D). At the same time point, a second allograft (2 cm x 3 cm) was placed. After 2 dHAMA applications (day 21), there was 90% wound closure. The honey treatment was continued and obtained complete closure 6 weeks after initial dHAMA application (Figure 1E). No functional limitations were noted. Patient was discharged 2 months post initial allograft application; the skin at the injury site was not discolored and a barely visible, soft scar was seen.

Case 2

A 31.3-week-old male infant was born via vaginal delivery at Steven and Alexandra Cohen Children’s Medical Center with a prenatal course significant for fetal imaging consistent with Tetralogy of Fallot (TOF), an absent pulmonary valve, and hydronephrosis. Initial course was complicated by respiratory distress syndrome for which the infant received nasal intermittent mandatory ventilation. Echo upon admission to the NICU confirmed the diagnosis of TOF, absent pulmonary valve syndrome, and discontinuous pulmonary arteries with anomalous origin of the left pulmonary artery (LPA) from the ascending aorta. There was a large left-to-right shunt due to the anomalous LPA causing pulmonary vascular congestion.

At DOL 4, the infant remained NPO and was receiving TPN via a PIV in the left hand that resulted in an extravasation injury. The entire dorsal hand had significant tightness and nonblanchable skin. The surrounding area appeared potentially necrotic (Figure 2A). Extensive edema was noted along the forearm. Hyaluronidase 1 mL of the 15 mL/unit concentration was injected in five 0.2 mL aliquots along with a normal saline wash-out of the punctures secondary to significant tightness of the involved area (4 punctures are made with an 18 G needle; the needle is withdrawn and the cannula is inserted subcutaneously through a puncture before normal saline is injected and allowed to flow out freely).

Within 1 week following the aforementioned treatment, a full-thickness (to subcutaneous tissue) wound was appreciated (8 cm x 3 cm). Initial treatment with honey was not successful; hard, thick eschar developed (7 cm x 3 cm). The eschar was softened with collagenase ointment (once daily over the next 5 days) and cross-hatching technique (Figure 2B). Eventually, eschar and surrounding slough were debrided; a final mechanical debridement (Figure 2C) converted a stalled wound to a fresh, bleeding wound. On day 11 from initial injury, the dHAMA was placed (Figure 2D) and covered by surgical tape strips (Steri-Strips; 3M, St Paul, MN) to ensure immobility; a nonadhesive silicone layer, a secondary dressing, and a gauze wrapping also were applied at this time.

The wound was checked at days 4 and 8 post dHAMA application. Remarkable closure was induced after a single allograft application (Figure 2E) shows the wound 16 days after dHAMA application), which was then followed by 1 more week of honey application to ensure complete closure (Figure 2E). The parents were very pleased with the outcome.

At the 1-month follow-up visit, the wound remained fully closed without functional limitations or hypertrophic scarring.

Case 3

A 25-week-old female infant was delivered via C-section with a history of evolving chronic lung disease requiring persistent ventilator support. She was transferred for a cardiothoracic consult with a presumed need for PDA ligation; on DOL 46, the PDA ligation took place.

On DOL 50, she was restarted on trophic feedings in addition to TPN infusing via a PIV in the right hand that resulted in an extravasation injury. There was an extensive intact blister (5 cm x 4 cm) with gross edema extending from the fingers to the upper arm. The blister was allowed to deroof naturally, which is a common approach in burn victims and provides natural cytokines and other anti-inflammatory mediators, therefore minimizing the area of necrosis and eventual wound size.⁷ After 2 weeks of expectant management and slow healing, the wound was mechanically debrided and the dHAMA was applied.

The wound achieved 90% closure after 10 days. By 5 weeks, complete closure was obtained with appropriate wound contraction. The size of the scar was 10% of the original wound size (6 mm x 2 mm) and barely visible at the 1-month follow-up post discharge after a single application of dHAMA.

Case 4

A 24-week-old male infant was born via C-section at a tertiary hospital in Long Island, New York, whose course was complicated by respiratory distress syndrome requiring high-frequency ventilation and 2 courses of dexamethasone. Serial echocardiograms revealed a large PDA with a left-to-right shunt. The infant completed an antibiotic course for presumed sepsis secondary to maternal chorioamnionitis. He had a grade 4 intraventricular hemorrhage, progressing to seizures as a complication. The infant was transferred to the authors’ facility for a PDA ligation.

On DOL 34, the PDA ligation was performed. On DOL 36, 2 days postop from ligation, the infant remained NPO with TPN infusing via a right hand PIV that resulted in an extravasation injury. The initial assessment of the hand included epidermal stripping after removal of the dressing as well as grossly edematous fingertips, dorsum of the hand, and forearm (Figure 3A); the authors were concerned the surrounding tissues would progress to necrosis given initial presentation. It was difficult to decide the depth of injury initially as swelling predominated. Hyaluronic acid was administered and a few punctures were performed with a 22 G angiocath secondary to significant swelling. Leptospermum honey was started immediately, but 5 days later a clear crescentic area of skin necrosis and separation became visible (3 cm x 7 cm) (Figure 3B).

Honey was continued; 10 days from initiation of honey, necrotic eschar detached completely, leaving a deep wound (4 cm x 5.5 cm) (Figure 3C). On day 11, the dHAMA (4 cm x 8 cm) was placed in the appropriately prepared wound bed and covered by surgical tape strips, a nonadhesive layer, and a secondary dressing. The first dHAMA application led to extensive extracellular matrix deposition, filling in the defect nicely. Superficial slough was noted and the wound was mechanically debrided. On day 19, a second allograft (2 cm x 2 cm) was placed (Figure 3D). At 5 weeks, the wound achieved complete closure (Figure 3E). The initial, mild hypopigmentation was barely visible 2 months later.

The purpose of this study is to describe the use of dHAMA in extremely preterm, critically ill newborns who sustained significant wounds from extravasation injuries. To the best of the authors’ knowledge, this is the first study describing the use of this dHAMA product in the preterm population with IV injuries.

Amnion is composed of 5 distinct layers: epithelium, basement membrane, compact layer, fibroblast, and a spongy layer. Epithelial layer, normally closest to the developing fetus, consists of a single layer of cells on the basement membrane, which is composed mostly of collagens 3and 4, glycoproteins, laminin, nidogen, and fibronectin.⁸ The compact layer, devoid of cells, is the main fibrous structure of the amnion. Interstitial collagens 1 and 3 form bundles important for structural integrity, and collagens 5 and 6 form filamentous connections to the basement membrane. The fibroblast layer is the thickest layer, with fibroblast embedded in a loose collagen and noncollagenous proteins. Lastly, the spongy layer is the interface between the amnion and chorion, composed of nonfibrillar meshwork of collagen 3 and an abundant content of proteoglycans and glycoproteins. This composition is partially preserved during the dressing preparation in order to contribute to important scaffolding properties. Clinically, the authors have noticed significantly faster extracellular matrix generation compared with other treatments, possibly in part due to allograft components and collagen contribution.

The “D/TIME” acronym — debridement/tissue, inflammation, moisture, and edge — is important for wound bed preparation, which means. Inflammation is a common deterrent of wound healing and efforts toward controlled inflammation are important.⁹ Multiple regenerative biomolecules are preserved during amniotic allograft preparation. Past studies^10-14 have demonstrated the presence of various growth factors, such as epidermal, platelet-derived, transforming, and fibroblast. These, along with anti-inflammatory cytokines, TIMPs (which help regulate matrix metalloproteinase activity), lack of 2-microglobulin and human leukocyte antigens A and C, contribute to an immunologically privileged environment, which is important to neonates who are immunologically weak; it will lead to full acceptance of an allograft without rejection or negative reactions.^13,14

In the present case series, culture swabs were sent sporadically throughout the treatment; the authors found culture-negative surfaces in all 4 wounds after initial allograft administration. Preventing infection is paramount to successful wound closure. Recent reports^14-16 described the unique molecules within amnion (eg, defensins, which have antimicrobial properties that may play a role in keeping a wound infection free). To the best of the authors’ knowledge, no neonatal studies are available, but pediatric studies describing burn patients support adult literature on the efficacy, antimicrobial properties, and scar efficacy of amniotic products.¹⁶

The adult literature describes successful use of amniotic allografts in the treatment of diabetic foot ulcers, lower extremity wounds, and burns.^11-16 Fetterolf and Snyder¹⁰ published a thorough review of clinical support of biochemical properties of dehydrated allografts, including staining for various growth factors, cytokines, and cells. Koob et al⁸ studied the concentrations of cytokines, chemokines, and growth factors in the graft after processing at various temperatures and over time. Clear retention of the above-mentioned molecules has been demonstrated.⁸ Upregulation of growth factors by fibroblasts have been shown after allograft placement as well as slow release of growth factors, chemokines, and cytokines from the tissue itself.^10,15 Tenenhaus¹⁶ reported successful repair in partial-thickness and deep burn wounds.

With neonatal wound care still in its infancy, product and dressing usage is extrapolated from adult use, which is often not applicable secondary to safety profile, size, or simply unknown interactions with developing immature skin.⁴ Preparation for amniotic membrane application rests on the same principles as any wound preparation. The wound bed must be clean, ie, debrided if necessary, which is often a challenging task in a tiny neonate where sharp or surgical debridement may not be tolerated.

Herein, the authors found preparing a wound with enzymatic or autolytic debridement followed by mechanical debridement immediately prior to allograft placement prepared the wound perfectly for allograft acceptance. The soft, transparent dHAMA used in this case series comes in a variety of sizes, incorporates into the wound bed quickly, and folds in easily, avoiding product waste. In their experience, the allograft stayed in place without any issues, even with challenging anatomical locations (ie, side of the ankle). The authors followed the manufacturer’s recommendations and reinforced the graft with the surgical tape strips over the wound, followed by a nonadhesive dressing and a secondary atraumatic silicone-based dressing to maintain a moist wound bed. The dHAMA can be applied dry or a few drops of normal saline can be placed following application, depending on the wound bed. If left undisturbed, the dHAMA applications incorporate into the wound by day 7 or 8. All wounds herein decreased by a minimum of 60% in size after 1 application. No babies developed contractures related to hypertrophic, jagged scars that can be seen in these injuries. All scars were small, soft, and close to original skin pigmentation; when followed-up at the clinic after discharge, the wound often decreased in size and thickness even more.

Although the authors’ observations are limited to a small number of preterm neonates with wounds developed due to IV extravasations, the results can be extrapolated to other wound etiologies and patient populations. These data build upon already growing evidence of the efficacy of amniotic products in general use. Future research should compare different combinations of amniotic membrane constituents and their efficacy in various neonatal wounds.

Affiliation: Steven and Alexandra Cohen Children’s Medical Center, New Hyde Park, NY

Correspondence: Vita Boyar, MD, 269-01 76th Avenue, Suite 344, New Hyde Park, NY 11040; VBoyar@northwell.edu

Disclosure: The authors disclose no financial or other conflicts of interest.