Characteristics and Uses of Biologic Dressings
and Skin Substitutes
Basil A. Pruitt, Jr, MD, Norman S. Levine, MD
\s=b\ Biologic dressings used for temporary coverage of open
wounds exert both mechanical and physiologic effects by
protecting the wound, maintaining microbial control, and
hastening wound maturation. Synthetic wound dressings
modeled after the biologic dressings have been evaluated by
several investigators. Those studies have shown unilaminate
synthetic membranes to be ineffective when applied to full\x=req-\
thickness wounds and have identified a bilaminate construc-
tion as being necessary for membranes to serve as effective
skin substitutes. The desirable properties of skin substitutes
have been identified and incorporated in the design of recently
developed membranes. Recent studies in this and other labo-
ratories have resulted in the development of collagen-syn-
thetic bilaminates, a totally synthetic biologically inert bilami-
nate, and tissue culture\p=n-\derivedmembranes. The characteris-
tics and limitations of each skin substitute determine the
optimum usage of these composite membranes and define the
modifications needed to improve the effectiveness of such
dressings.
(Arch Surg 1984;119:312-322)
closure of a wound by the use of material
Temporary
other than the patient's own skin has become com¬
monplace during the past two decades. Biologic dressings
play a key role in the treatment of those patients whose
extensive burns leave so little available donor site that their
wounds cannot be covered by a single harvest of autografi
Accepted for publication Nov 7, 1983.
From the US Army Institute of Surgical Research, Fort Sam Houston,
Tex (Dr Pruitt); and the Division of Plastic Surgery, Oklahoma University
Health Science Center, Oklahoma City (Dr Levine).
The opinions or assertions contained herein are the private views of the
authors and are not to be construed as official or as reflecting the views of
the Department of the Army or the Department of Defense.
Reprint requests to Librarian, US Army Institute of Surgical Research,
Fort Sam Houston, TX 78234.
Downloaded From: http://archsurg.jamanetwork.com/ by a Georgetown University Medical Center User on 05/21/2015
skin. Cutaneous allograft is the most frequently used and
most effective biologic dressing. All other such dressings
must be compared with allograft to determine their relative
effectiveness and rank in the hierarchy of biologic dress¬
ings. In addition to what might be termed natural biologic
dressings, a variety of synthetic skin substitutes have been
fabricated since Ivalon sponge (a polyvinyl alcohol polymer
with a thin outer layer of silicon rubber; Unipoint Indus¬
tries, High Point, NC) was first introduced in 1961, and
recently developed materials have shown promise in early
clinical trials.
PROPERTIES AND USES OF BIOLOGIC DRESSINGS
A variety of beneficial effects have been attributed to
biologic dressings and, to a greater or lesser (usually lesser)
extent, to synthetic skin substitutes. In areas where a
biologic dressing adheres to the underlying wound bed, the
population density of bacteria on the wound surface de¬
creases; the dressing also aids in microbial control when
topical therapy has been discontinued following debride-
ment of the bulk of the nonviable burned tissue.1 In
where there is no adherence of a biologic dressing to the
areas
wound bed, submembrane suppuration commonly occurs,
and it appears to hasten debridement of residual nonviable
tissue, presumably by the combined action of both bacterial
and WBC enzymes. By physically covering the burn wound,
biologic dressings prevent desiccation of an open wound and
decrease evaporative water loss and heat loss. Physical
protection of the wound by a biologic dressing also prevents
further contamination and injury of the newly developing
granulation tissue. Application of allograft skin also de¬
creases wound pain, which, in the case of burns overlying
joints, facilitates motion of the burned part and the execu¬
tion of the daily physical therapy program. Finally, the
biologie dressings appear to hasten and make more orderly
the healing of second-degree burns; when applied to full-
thickness burns, they increase granulation tissue capillary
density.2
The beneficial effects of allograft skin are realized when it
is used for coverage of full-thickness burn wounds after
eschar separation and prior to autograft application, and for
immediate coverage of superficial second-degree burns to
hasten healing. Coverage with allograft skin of excised burn
wounds and other wounds such as those resulting from
mechanical trauma or operative procedures that cannot be
closed primarily is also beneficial. Also, allograft skin can
be used as a test material to determine the likelihood of
autograft take. When allograft skin shows general ad¬
herence to a burn wound and evidence of graft vasculariza-
tion within 48 to 72 hours of application, one can anticipate
an excellent take of autograft skin applied to that wound
following removal of the allograft skin. Such preparation of
a "receptive" wound bed is particularly important in the
care of those patients with extensive burns and a paucity of
donor sites, to avoid unnecessary loss of autograft skin.
Allograft skin can be obtained from a family member or
other living volunteer, but is most commonly harvested
from the bodies of patients who have died in the hospital.
Consequently, the supply of allograft does not always match
the demand for such tissue. Other limitations of fresh
allograft skin include the need for refrigerated storage
conditions and a limited "effective" shelf life of approx¬
imately seven to ten days when the tissue is stored at 4 °C.3
The viability and take of allograft skin decrease with time
after harvest and are greatest when allograft skin is used
immediately after harvest. The possibility of disease trans¬
mission requires careful donor selection. The elimination of
potential donors with cutaneous malignant neoplasms and
infections, systemic viral disease such as hepatitis, and
hématologie malignant disease reduces the size of the donor
pool available for the harvest of allograft skin.
Fresh-frozen allograft skin and lyophilized allograft skin
each overcome some of the limitations encountered with the
fresh material, but have limitations of their own.4·5 The
demanding processing procedures and requirements con¬
fine the preparation of either material to special centers
having proper facilities. The lyophilized material has an
essentially unlimited nonrefrigerated shelf life, while the
frozen material has a similarly prolonged shelf life provided
proper refrigeration is maintained. Either material can be
easily and rapidly prepared for use by rehydration or
thawing. When harvested at a thickness of greater than
0.381 mm, the lyophilized material is subject to epidermal-
dermal separation that permits desiccation of the subse¬
quently exposed dermis and rapid sloughing of the material
from the wound. Lyophilized allograft skin generally ad¬
heres less well to the wound and is less able to reduce the
bacterial count on the wound surface than fresh allograft
skin.6
Amnion, which has also been used as a biologic wound
dressing, is readily available from the delivery room and is
inexpensive to prepare.7 When applied to a wound, amnion
must be covered with an occlusive dressing to prevent
Downloaded From: http://archsurg.jamanetwork.com/ by a Georgetown University Medical Center User on 05/21/2015
desiccation and a resulting rapid sloughing of the graft.
Amnion adherence to the wound bed is also less secure than
is that of cutaneous allograft, and true biologic union with
host-to-graft vascular connection is uncertain.8 Faulk et al
have reported that application of amnion to wounds exerts
an angiogenic effect and increases the capillary density of
the underlying wound bed.9
Cutaneous xenografts from a variety of species have been
used clinically with porcine material (fresh, fresh-frozen,
and lyophilized) commercially available from several
sources. The sterility and viability of some commercially
available fresh and fresh-frozen porcine xenograft skin has
been uncertain, and contamination of such material, even by
organisms that are ordinarily considered saprophytic, rep¬
resents a potential risk for the severely burned patient
whose immunologie capacity has been markedly compro¬
mised. Porcine cutaneous xenografts are not as effective as
cutaneous allografts in reducing the bacterial population
density of the underlying wound.10 Moreover, cutaneous
xenografts adhere to wound beds less well than do cutane¬
ous allografts. The graft-to-host union appears to be ef¬
fected by fibrovascular ingrowth of granulation tissue into
the dermis of the graft, and vessel-to-vessel connection
between host and graft does not occur as it does with
cutaneous allografts." The nourishment of the outer layers
of xenograft skin appears to be dependent on a limited
plasmatic circulation, and the material undergoes progres¬
sive degenerative necrosis rather than classic rejection.
Moreover, secure union of the graft to the wound bed
develops over a period of days, during which time the grafts
and the involved body part must be immobilized.
One additional concern limits the use of all wound dress¬
ings. If dressings are applied to a heavily contaminated
wound, marked subgraft microbial proliferation and suppu¬
ration may occur and rapidly progress to frank infection. In
the case of partial-thickness burns, such infection may
convert the injury to full-thickness necrosis. In the case of
full-thickness burns, the infection may become invasive,
with involvement of local unburned tissue and even, by
systemic dissemination, remote tissues and organs. To
avoid that complication, biologic dressings should not be
applied to deep, partial-thickness burns prior to removal of
the superficial debris, to full-thickness burns prior to
eschar separation or removal, or to any wound with a
bacterial population density of over 105 organisms per gram
of tissue, as determined by biopsy monitoring of the burn
wound.
Although cutaneous graft viability appears, on the basis
of clinical experience, to be important in the control of the
subgraft microbial population, wound dressing viability
does not appear to be essential in terms of physical coverage
of excised wounds or systemic resistance to infectious
challenge following excision of extensive burn wounds. In a
murine model, excision of 60% of the skin results in essen¬
tially uniform mortality if the excised area remains uncov¬
ered, is covered with fine-mesh gauze, or is covered with an
impermeable membrane such as Vi-drape surgical drape
material (Parke-Davis & Co, Detroit). Coverage of such
excised wounds with either viable allograft or a synthetic

Fig 1.—Tissue obtained from wound bed following removal of
polyvinyl alcohol sponge (Ivalon) dressing. Note irregular retained
fragment of polyurethane foam with intense foreign body inflam¬
matory reaction in surrounding tissue.
skin substitute results in survival of over 75% of the
animals. Other animal studies have shown protection from
the lethal effects of an intraperitoneal Pseudomonas chal¬
lenge following burning by immediate wound excision and
coverage of the excised wound with either viable allograft
skin or nonviable formaldehyde-treated allograft skin.12
These and other studies, as well as the increased use of
early burn wound excision procedures that produce wounds
requiring immediate coverage, have reawakened interest in
the development of effective wound dressings—skin substi¬
tutes.
SYNTHETIC WOUND DRESSINGS
Early Wound Dressings
A variety of synthetic wound dressings and skin substi¬
tutes have been evaluated over the past three decades, with
the earliest materials being a spray-on plastic (a copolymer
of hydroxyvinylchloride-acetate and sebacic acid),13 polyox
(a polymer of ethylene oxide),1 and Ivalon sponge.14 These
materials permitted submembrane suppuration and attend¬
ant sepsis, especially when applied to full-thickness inju-

Fig 2.—Left, Full-thickness burn wounds on posterior aspect of leg were covered with viable
cutaneous allograft and gauze-backed, poly-amino acid film. Right, Five days after application,
allograft skin is well vascularized and adherent throughout. Film was adherent only at periphery of
wound through serous crust formation (note remnant of crust at lower margin of wound above
lateral malleolus). Elevation of film reveals that it was semiattached to the wound surface, floating
on layer of serosanguineous fluid visible on undersurface of membrane.

Downloaded From: http://archsurg.jamanetwork.com/ by a Georgetown University Medical Center User on 05/21/2015

 Table 1.—Parametric Analysis of the Change in Mean Log Bacterial Counts for All Wounds

Human Cadaver Porcine Cutaneous

Cutaneous Allograft Xenograft Synthetic Dressing Coarse-Mesh Gauze
Wounds on which allograft took (N = 23)
Bacterial count before treatment 4.72 4.57 4.81 4.61
Change treatment*_-1.20_+0.13_+0.02_-0.25_
in count after
SE for change_038_039_035_O40_
95% confidence limits for change_-0.44 to -1.96_-0.65 to +0.91_-0.68 to +0.72_-0.65 to +1.05
Significance of change, f test_P<.01_NS_NS_NS_
Wounds on which allograft did not take (N 18)t
=

treatment_4^23_4.77_4_24_4J2_
Bacterial count before
treatment*_+1.00_+0.57_+0.95_+0.47_
Change in count after
SE for change_0^54_+0.40_0.47_043_
+1.37_0.00
95% confidence limits for change_-0.09 to +2.09_-0.23 to to +1.90_-0.39 to +1.29
Significance change,
of f test P<.1 P<.2 P<.1 P<.3
*Minus indicates decrease in bacterial count after treatment. Plus sign indicates increase.
fFor the combined results of all four treatments for wounds on which allograft did not take, there was a change of +0.70 (P<.01, =
72).

ries with residual nonviable tissue. Ivalon fragmented on

removal, with the retained particles of sponge eliciting a
foreign-body inflammatory reaction that complicated
wound closure (Fig 1). The spray-on plastic film, when
applied to full-thickness injuries in the immediate postburn
period, was associated with the development of wound
infection, and the polyox gel, if not covered by an impermea¬
ble membrane or a dressing, rapidly desiccated and disap¬
peared.
Subsequent experience has been reported with the use of
laminated nylon velour,15 polyurethane foam laminated to
polypropylene film,16 poly-amino acid films," films of Gore-
tex (polytetrafluoroethylene; W. L. Gore & Associates,
Elkton, Md) and ultrathin (0.8 mil) silicone polymer mem¬
branes laminated to a meshed inner layer.18 The previously
noted problems of submembrane suppuration and highly
variable biologic union with irregular membrane adherence
to the wound were encountered with these materials.
Fig 3.—Unilaminate, small-pore-size polytetrafluoroethylene
Unilaminate impervious, or nearly impervious, films com¬ membrane at top was applied to underlying excised rat wound
monly adhere to the margin of a wound and a highly variable seven days prior to biopsy. Note that membrane is attached to
fraction of the wound surface by formation of a serous crust, granulation tissue that has formed on wound by dark band of
neutrophil-laden, serofibrinous exúdate.
yet are elevated from other portions of the wound surface by
a serous or seropurulent film (Fig 2). Foams and other
laminates permit suppuration within the lacunae or pores of with an average burn extent of 12% of the total body
the inner membrane above the level of the ingrowing wound surface.19 The Hydron dressing was applied to burn wounds
granulation tissue and show variable degrees of fragmenta¬ with an average extent of 5% of the body surface. The
tion on removal, with the residual fragments of the syn¬ investigators found that Hydron applied to a moist ex¬
thetic material enmeshed in the wound-bed granulation udative wound would not cure properly and easily "lifted
tissue serving as foci of inflammatory reaction. off" the wound. In more than half of the patients, the
Recently Developed Wound Dressings dressing lost its integrity, requiring either local repair or
complete reapplication. Eighty percent of the patients
Recently developed unilaminate membranes are best showed significant peeling of the dressing, especially in
considered wound dressings rather than skin substitutes. those areas where it came in contact with the bedclothes.
The Hydron (polyhydroxyethylmethacrylate; Hydron Lab¬ Submembrane serum collections often occurred, and in
oratories, Ine, New Brunswick, NJ) burn dressing is a thin eight patients these showed bacterial growth. Frank infec¬
transparent unilaminate membrane that is formed in situ on tion may occur beneath this material, and instances of septic
the burn wound. Polyhydroxyethylmethacrylate powder conversion of partial-thickness injury to full-thickness loss
and liquid polyethylene glycol are alternately sprayed on have been reported by others, sometimes in association
the burn wound to form the membrane. Warren and Snell- with systemic sepsis. Thirteen of the 15 patients studied
ing have reported on the use of this membrane in 15 patients credited the Hydron with reducing wound pain. The in-

Downloaded From: http://archsurg.jamanetwork.com/ by a Georgetown University Medical Center User on 05/21/2015

 vestigators concluded that the optimum use of Hydron was
in the treatment of small, noncircumferential, superficial,
partial-thickness burns with intact blister epithelium that
did not cross a joint surface.
Studies by Dressier et al of the effect of Hydron on the
healing of full-thickness scald burns in a murine model and
full-thickness excisional wounds in rabbit ears have shown
that the Hydron burn dressing exerted no deleterious
effects on wound healing.20 The material has also been
reported to be useful as a donor-site dressing, but the cost of
such a material compared with fine-mesh gauze militates for
use of the latter.
Another unilaminate membrane in the form of a transpar¬
ent liquid gel has been developed by Wokalek et al for use as
a dressing for fresh operative wounds.21 The gel is composed
of agar copolymerized with acrylamide. The gel is consid¬
ered chemically inert and nonallergenic. The developers
have reported that the gel reduces wound discomfort and
that changes of the dressing are painless. The gel's trans¬
parency allows early identification of submembrane suppu¬
ration and its porosity allows drainage of wound secretions.
The low tensile strength of the gel makes it susceptible to
disruption, and the lack of firm adherence makes this
material a poor candidate for use as a skin substitute and
speaks for its use solely as a temporary wound dressing.
Fibrin film has also been used as a unilaminate wound
dressing. Nishi has reported on the use of chrome-hardened
fibrin films in 333 burned patients.22 The films are credited
with having an oligodynamic effect and with reducing the
healing time of partial-thickness burns. The material has
hemostatic properties and is considered nonantigenic and of
low tissue reactivity. Application of the fibrin film reduced
wound pain and maintained wound bacterial counts at low
levels in the study patients. The investigator recommended
that the fibrin film be covered with gauze dressings to
maintain close adherence to the wound bed. The absence of
a control group and of specific bactériologie data in the
study make evaluation difficult, but the fibrin film appears
to function like a serous crust.
Collagen in various forms has been evaluated as a biologic
dressing by several investigators. Collagen is an attractive
skin substitute since it can be isolated from tissues of other
species and purified in large amounts, is of low antigenicity,
exerts a hemostatic effect, can be produced in a variety of
physical forms, ie, gels, films, fibrils, extruded strips,
porous mats, sponges, and powder, and can be bonded to
other materials for use as a composite. Generally, the
tensile strength of collagen is enhanced by chemical or
physical means to prevent fragmentation, but such treat¬
ment decreases elasticity and drapability when the collagen
is used as a film, sponge, or strip. The initial adherence of
membranes to granulation tissue appears to be a fibrin-
dependent phenomenon.23 Although fibrin binds preferen¬
tially to native collagen, it is uncertain whether bonding
occurs between fibrin and the exogenous collagen of mem¬
branes.24 Nevertheless, Tavis et al, using a "modified"
collagen membrane, found good initial adherence to wounds
on which bacterial population density was low.25 Increased
adherence to such wounds was noted 72 hours after applica-
Downloaded From: http://archsurg.jamanetwork.com/ by a Georgetown University Medical Center User on 05/21/2015
tion, except with those collagen membranes applied to
granulation tissue, in which case only approximately one
third of the membrane surface was adherent to the wounds
by that time. Water vapor transmission through collagen
membranes is related to membrane thickness and, in modi¬
fied collagen sheets 6 mils thick, approached that of burn
wound eschar. Burn wounds fully excised seven days after
burn and covered with collagen membranes showed slightly
higher bacterial counts than did autografted wounds, but
showed lower bacterial densities than did wounds to which
allograft or xenograft skin had been applied. Autograft
skin subsequently applied to wounds initially dressed with
collagen membranes following excision showed only a 42%
take.
A collagen sponge membrane, when applied to an excised
animal wound, showed good initial adherence, but gradu¬
ally dried and developed a hardened outer surface as
adherence decreased across time.26 The unilaminate col¬
lagen membranes appear to function like a serous crust.
The decrease in adherence with time indicates little, if any,
fibrovascular ingrowth to form an effective biologic union
between wound and graft, and reflects the action of both
microbial and host-tissue collagenase.
NECESSARY PROPERTIES OF SKIN SUBSTITUTES
Many studies have permitted the identification of those
properties needed for a clinically effective skin substitute,
which are as follows.
Absence of antigenicity
Tissue compatibility
Absence of local and systemic toxicity
Water vapor transmission similar to normal skin
Impermeability to exogenous microorganisms
Rapid and sustained adherence to wound surface
Inner surface structure that permits ingrowth of fibro¬
vascular tissue
Flexibility and pliability to permit conformation to irreg¬
ular wound surface
Elasticity to permit motion of underlying body tissue
Resistance to linear and shear stresses
Prevention of proliferation of wound surface flora and
reduction of bacterial density of wound
Tensile strength to resist fragmentation and retention of
membrane fragments when removed
Biodegradability (important for "permanently" im¬
planted membranes)
Low cost
Indefinite shelf life
Minimal storage requirements
Studies of an ultrathin silicone polymer membrane identi¬
fied the importance of water vapor permeability of the outer
layer of bilaminate membranes in the prevention of subgraft
fluid collections.18 The early studies of skin substitutes also
indicated that the pore size of the membrane must be such
that it will act as a barrier to microorganisms. Such a small
pore size will not permit biologic union by ingrowth of
fibrovascular tissues from the wound bed, a process requir¬
ing a pore diameter of at least 80 µ (Fig 3).26 It appears
that optimum function of a skin substitute requires that it
 be bilaminate, with the pore sizes of the inner and outer
layers as described. Theoretically, a unilaminate membrane
having different pore sizes in the inner and outer portions
should function satisfactorily, but such has not been devel¬
oped.
Other properties necessary for a membrane to function as
an effective skin substitute include the following: freedom
from local or systemic toxicity, absence of antigenicity, and
resistance to linear and shear stress. It is also important
that the membrane not promote bacterial growth and,
ideally, it should reduce the microbial density at the wound-
membrane interface by providing an environment in which
host defenses can function. The membrane should also be
sufficiently thin, pliable, and elastic to tolerate the stresses
and motion associated with limb and underlying muscle
activity, and permit the membrane to conform and inti¬
mately adhere to the many irregularities of a wound bed.
Desirable properties from a logistical standpoint include
low cost, an indefinite shelf life, and minimal storage
requirements.
RECENTLY DEVELOPED SKIN SUBSTITUTES
Recent work by several investigative groups has led to
the development of three types of membranes for use as skin
substitutes, ie, bilaminate membranes that are a composite
of heterologous biodegradable tissue and synthetic mate¬
rial, bilaminate membranes that are totally synthetic and
biologically inert, and membranes composed of autologous
tissue that either persists or is gradually replaced by host
tissue.
Collagen-Synthetic Composite Membranes
Biobrane (Woodroof Laboratories Ine, Santa Ana, Calif)
is a bilaminate membrane that has been evaluated for use as
both a wound and skin graft donor-site dressing. The outer
layer of the membrane is composed of an ultrathin layer of
silicone rubber that is mechanically bonded to a fine-knit,
flexible nylon fabric. The silicone membrane is semiper¬
meable and allows passage of water vapor from the wound
surface to the environment while preventing passage of
bacteria from the environment to the wound surface. Type I
porcine collagen is covalently bonded to the fabric to
provide an inert hydrophilic mesh inner layer into which
granulation tissue can grow. The water vapor transport
characteristics of this composite are directly related to the
thickness of the membrane, which can be controlled so the
water vapor transmission approximates that of skin. The
elasticity of the composite (100% elongation in any direc¬
tion) provides drapability and permits full range of motion
of a grafted body part. Bactériologie studies have shown an
inverse relationship between adhesion and bacterial col¬
onization,27 indicating that this material is best suited for
application to "clean wounds" such as donor sites, or per¬
haps burn wounds excised before heavy microbial coloniza¬
tion has occurred.
Clinical studies of this material used for immediate cover¬
age of split-thickness skin graft donor sites have shown the
membrane to reduce evaporative water loss from the wound
and to decrease wound pain markedly. In such cases, the
Downloaded From: http://archsurg.jamanetwork.com/ by a Georgetown University Medical Center User on 05/21/2015
membrane has remained soft and pliable during the healing
period and lifts off the wound as reepithelialization occurs.
Laboratory studies in a murine model, using the membrane
as a dressing for full-thickness wounds, have shown that
connective tissue ingrowth occurs by 72 hours after applica¬
tion with adherence said to be equal to or better than that of
allograft skin.27 Such dressings remained intact for up to
one month after application and apparently inhibited wound
contraction. In all study animals, "normalize! adherence
'
of this composite was not as good as that of allograft skin,
but was better than that of xenograft skin. In a recent
clinical study, Biobrane was applied over mesh grafts that
were expanded to a 6:1 or greater ratio to prevent graft
slippage, the development of infection in the graft in¬
terstices, and mechanical disruption of the thin mesh of the
graft. The membrane could be easily removed on the fifth
postgraft day, by which time the mesh autografts were
revascularized; maintenance of the prosthetic covering be¬
yond that time completely blocked epithelialization of the
interstices by the autograft.28
Burke et al have developed a bilaminate membrane
consisting of a bovine collagen-based dermal analogue and a
Silastic epidermis.29 The biocompatible dermis, which is
slowly biodegraded, serves as a template for growth of host
connective tissue to form what has been termed a neo-
dermis. The dermal layer of the composite is a porous,
collagen-chondroitin 6-sulfate fibrillar matrix that is read¬
ily populated by fibroblasts and capillaries from the wound
bed. The physical and biochemical properties (such as pore
size, cross-link density, helical structure of collagen, and
collagenase resistance) of the dermis are controlled by the
glycosaminoglycan content of the material as well as the
sterilization procedures (heat and exposure to glutaralde-
hyde) used in preparation. The epidermis of the composite
is a 0.1-mm-thick layer of Silastic applied in liquid form to
the dermal component, to which it firmly binds as it cures.
This epidermal analogue permits water flux at levels similar
to normal skin and protects the wound from mechanical
trauma and microbial invasion. The composite can be stored
in 70% isopropyl alcohol or in freeze-dried form.
The developers have reported on the use of this material
for immediate closure of excised burn wounds in ten pa¬
tients with extensive burns. The composite was applied to
the wounds, which ranged in extent from 15% to 60% of the
total body surface (average, 27%). Despite careful place¬
ment and suturing of the membranes, some wrinkling
occurred, necessitating debridement of the wrinkles to
drain underlying serous collections. Although collagen
sponge laminated to a polyurethane film showed, in studies
by Oluwasanmi and Chvapil, less adherence to tangentially
excised rabbit wounds than collagen sponge per se,26 this
composite was firmly adherent to the wound bed within
minutes of application, and early vascularization of the
dermal component was evident within three to five days.
When donor sites were subsequently available, the Silastic
epidermis was removed and the neodermis was covered
with thin autograft skin 0.1 mm thick. Adherence of this
membrane to the excised wounds ranged from 95% to 100%,
and the membranes were left intact for 14 to 64 days, with
Downloaded From: http://archsurg.jamanetwork.com/ by a Georgetown University Medical Center User on 05/21/2015
 Fig 4.—Lack of available donor sites required repeated application
of totally synthetic composite membrane to burn wounds on legs of
this patient prior to autografting. Top left, Sheets of totally synthetic
composite membrane have just been applied over granulation
tissue on burns of legs. Membrane is held in place by elastic net
gauze to prevent dislodgment by contact with bedclothes. Note that
membrane elasticity permits flexion of knees. Top right, "Strike-
through" staining of composite membrane due to exúdate and
debris within membrane interstices indicates that it should be
changed. Center left, Removal of third set of composite membranes
reveals clean granulation tissue. Note clean appearance of under¬
side of membrane, integrity of membrane, intimate attachment to
wound bed, and lack of fragmentation of inner layer at point of
separation from wound surface, indicated by the reflected edge of
membrane. Center right, Appearance of full-thickness burn wounds
ready for autografting following three applications of totally syn¬
thetic composite membrane. Note uniformity of granulation tissue
and absence of inflammatory changes In wound. Bottom left, Take
of subsequently applied mesh autografts expanded in 3:1 ratio was
excellent. Bottom right, Appearance of burns after healing of mesh
grafts. Note paucity of hypertrophie scarring. Lack of wound granu¬
loma formation confirms absence of retained membrane fragments.
take of the subsequently applied autografts ranging from
85% to 95%. Submembrane hematoma formation caused
loss of a portion of the skin substitute in one patient and
seroma formation occurred beneath areas of membrane
wrinkling in several patients. In areas of wear or motion,
epidermal-dermal separation of the composite occurred and
desiccation of the dermal analogue in those areas reduced
the take of subsequently applied autograft skin.
Histologie studies showed progressive host tissue inva¬
sion and replacement of the dermal analogue with no bovine
collagen evident seven weeks after application. Moderate
fibrosis of the neodermis was noted and some scarring was
also identified below the neodermis, which was considered
to represent fibrotic change of unexcised heat-injured tis¬
sue. Pore size appears to be critical to the successful
function of the dermal analogue, since a pore size signifi¬
cantly smaller than that of normal dermis retarded cellular
invasion and led to development of thick, fibrous tissue. No
inflammation or immunologie reactions were noted.
Long-term follow-up is lacking, but there is said to be no
evidence to date of hypertrophie scar formation or clinical
contracture following use of this composite. Areas treated
with the skin substitute were found to be softer and more
elastic and pliable than areas in which the excised wounds
were closed by immediate application of meshed autograft
skin. Although no graft or subgraft infections were noted in
the study patients, the grafts were applied only to excised
wounds, and the material has not been applied to wounds
covered with granulation tissue and contaminated with
microbial organisms.
Yannas et al have reported further studies of the skin
substitute composite using a guinea pig model of a full-
thickness wound. In vitro seeding of the dermal analogue
with dissociated autologous basal cells was followed by
rapid basal cell proliferation and formation of sheets of
keratinized epidermis at the laminar interface of the com¬
posite within 14 days of application to a wound.30 In that
model, what was considered to be new, functional skin had
been generated in less than four weeks.
Downloaded From: http://archsurg.jamanetwork.com/ by a Georgetown University Medical Center User on 05/21/2015
Totally Synthetic Bilaminate Membrane
Another direction has been taken by Levine et al, who
have developed a totally synthetic, biologically inert bilami¬
nate membrane. A number of materials were evaluated for
potential use as the inner layer of a composite skin substi¬
tute in terms of initial membrane adherence to the wound
surface, conformation to the wound surface, adherence of
the membrane to the wound surface five days after applica¬
tion, occurrence of submembrane suppuration, and growth
of fibroblastic tissue into the membrane. Loose-knit nylon
stocking fabric, coarse-mesh surgical gauze, and fine-mesh
surgical gauze appeared similar, and those materials were
superior in terms of initial adherence, adherence at five
days, absence of subgraft suppuration, and extent of fibro¬
blastic ingrowth to Silastic-coated gauze, collagen mem¬
branes, and three polytetrafluoroethylene membranes.31
Subsequent testing of the three best materials showed the
nylon stocking fabric to be superior in terms of tissue
compatibility, lack of fragmentation and retention of foreign
bodies at the time of membrane removal, flexibility and
elasticity, and conformity to irregular wound surfaces.
Studies of the extent of fibroblastic ingrowth into the
nylon stocking fabric were used to determine the optimal
thickness of the dermal analogue. Clot formation occurred
beneath the serous crust, which rapidly formed on the outer
surface of wounds covered with single and double layers of
the fabric. Fibroblastic ingrowth always occurred when
four and eight layers of the mesh fabric were used, with only
rare elevation of the dressing from the wound surface
occurring beneath four layers of fabric. These studies also
showed that a minimal thickness of 0.635 mm of the dermal
analogue was needed to ensure consistent fibroblastic in¬
growth. It was found that a 1-mil-thick poly¬
tetrafluoroethylene membrane with O.l-µ pores permit¬
ted passage of water vapor, was impermeable to liquid
water, and impermeable to bacteria inoculated on the
membrane with a wetting agent. On the basis of these
studies, a composite wound dressing (skin substitute) has
been developed, consisting of a dermal analogue of six
layers of 14 denier nylon stocking fabric and the 1-mil-thick,
O.l-µ pore size polytetrafluoroethylene membrane as the
epidermal analogue. This dressing can be sterilized prior to
use in an autoclave or with ethylene oxide.
Animal studies have shown this membrane to be closely
attached to the bed of the excised wound five days after
application. Fibrovascular tissue of the host extended into
the lower two thirds of the thickness of the dermal analogue;
by ten days, organized collagen was identified in the nylon
matrix. When applied to animals with experimentally con¬
taminated wounds, the membrane limited microbial pro¬
liferation and prevented the development of fatal invasive
infection. The wound protective effect of the bilaminate was
evidenced by its ability to promote survival in rats in which
excisional wounds of 60% of the body surface area were
dressed with the membrane—a result not achieved by
covering such wounds with coarse-mesh gauze.32
The membrane has also been evaluated clinically and
compared with cutaneous allograft, porcine xenograft, and
coarse-mesh gauze when applied to wounds covered with
 granulation tissue on 21 patients with burns of from 30% to
66% of the total body surface. Each wound was divided into
four areas and the four dressings were applied to randomly
assigned areas. Wound appearance was assessed qualita¬
tively by an experienced burn surgeon when the dressings
were removed 48 hours after application. Quantitative
aerobic wound bactériologie studies were performed by the
standard surface-swab technique, both before dressing
application and at the time of removal, 48 hours after
application.
Of those wounds in which allograft skin adhered to the
granulation tissue, no significant difference in wound ap¬
pearance existed between those covered with autograft
skin, coarse-mesh gauze, and the bilaminate. Those areas
covered with porcine xenograft were judged to be signifi¬
cantly worse in appearance. In wounds on which allograft
skin did not take, xenograft-treated wounds had the worst
clinical appearance, while autograft-treated wounds were
significantly worse than those treated with coarse-mesh
gauze or the synthetic dressing. Wounds treated with
coarse-mesh gauze and synthetic dressing were indistin¬
guishable from each other.
These findings confirm clinical observations indicating
that if a wound accepts allograft skin, the take of autograft
skin is assured, and that the wound should be closed with
autografts, if available. If autografting cannot be per¬
formed for some reason, eg, limited donor sites or intercur¬
rent illness, the application of allograft skin is the treat¬
ment of choice until autografting can be accomplished. For
wounds on which allograft skin will not adhere, the superior
performance of the bilaminate skin substitute and the
coarse-mesh gauze is a reflection of the greater debriding
action of those materials compared with the allograft and
xenograft tissue. The greater inflammatory response, fiber
fragmentation, and retention of cotton fibers at the time of
removal of cotton gauze make the synthetic bilaminate the
preferred dressing for such wounds.
The wounds of the study patients were colonized by a
variety of organisms, most commonly Pseudomonas aeru-
ginosa and Staphylococcus aureus. The organism popula¬
tion density significantly decreased in those areas of the
wounds on which allograft skin took, but showed no change
in areas treated with the other materials (Table 1). In
wounds on which allografts did not take, there was no
significant change in the bacterial population density be¬
neath any of the materials. However, in over half of the
treatment areas of the wounds, the bacterial counts in¬
creased after 48 hours of dressing application, indicating
that for wounds to which biologic dressings or skin substi¬
tutes do not adhere, those membranes should be changed
more frequently than every two days.
The bilaminate dressing has also been applied to areas of
burn wound granulation tissue representing from 1% to 30%
of the total body surface in 15 additional patients with burns
of from 30% to 77% of the body surface. In those patients
whose wounds appeared to be ready for closure, the bilami¬
nate dressing was generally left in place for 48 to 72 hours,
following which autografting was carried out. In other
cases, where the wounds had foci of residual necrotic tissue
Downloaded From: http://archsurg.jamanetwork.com/ by a Georgetown University Medical Center User on 05/21/2015
and were heavily contaminated, the dressing was usually
changed daily to effect wound debridement. Since the
previously noted bactériologie studies showed no reduction
in microbial density in wounds covered with this dressing,
the laminate was soaked in an antimicrobial solution (5%
mafenide acetate) immediately prior to application to pre¬
vent bacterial proliferation and reduce subgraft suppura¬
tion. The laminate was held in place with elastic mesh
dressings to ensure apposition to the wound bed and
prevent shearing of the graft from the wound surface by
contact with the bedclothes (Figure 4, top left). The dress¬
ing conformed well to irregular wound surfaces and its
flexibility and modest elasticity in two directions were
sufficient to permit active motion of extremities to which it
was applied. There was no fragmentation or retention of the
fabric forming the dermal analogue at the time of removal,
even when the dressing had been in place for three or more
days.
Although the semitransparency of the composite pre¬
vents direct visualmonitoring of the wound surface, wound
exúdate was readily evident by staining of the outer surface
of the dressing. Such "strike through" staining was usually
apparent in wounds considered unready for grafting within
24 hours of application of the dressing, and reflected the
presence of debris and exúdate within the dressing (Fig 4,
top right). Debridement of such material was achieved by
daily dressing changes of the bilaminate membrane. In
wounds considered ready for grafting, staining was not
evident for up to three or four days. The staining did not
appear to be associated with significant wound infections,
since take of autografts immediately applied to wounds
from which stained dressings were removed was usually
excellent (Fig 4, center left to bottom right).
Tissue Culture-Derived Skin Substitutes
The most recent line of development has been the tissue
culture growth of epidermal cells obtained from the pro¬
spective recipient who will require grafting. Studies by
Eisinger et al have shown that single-cell suspensions of
human epidermal cells can be grown on collagen film into
confluent multilayered sheets, with the lower layer of cells
morphologically similar to the basal cells of normal epider¬
mis.33 When these sheets were implanted into nude mice,
either under the kidney capsule or into subcutaneous
tissue, the cells grew into a multilayer structure with
keratohyaline granules in the upper layers. The tumori-
genie potential of the cells appears to be low, since no
abnormal growth was observed in these studies. Canine
epidermal cell sheets, prepared in the same fashion, were
applied to either fresh or granulating wounds and held in
place by an occlusive dressing. When the occlusive dressing
remained intact, the sheet of cells expanded to cover the
wound in four of six animals by the fifth day. In two of the
animals, the grafts became infected thereafter and the cells
did not survive. In the other two animals, the wounds
appeared, by clinical criteria, to be healed after one week,
but it was uncertain whether the tissue-cultured cells
persisted or were replaced by cells from the recipient.
Wound contraction seemed to occur in the test animals and
 Table 2.—Uses and Limitations of Wound Dressing Membranes and Skin Substitutes*
Submembrane
Susceptibility Fluid Loss of
Dressing Optimum Use to Infection Collection Integrity Adherence
Hydron Early coverage of small, Low on
noncircumferential, superficial, exudative wounds
partial-thickness burns not involving
joint
Polymerized agar Temporary dressing of clean wounds + + + + Low
Fibrin film Temporary dressing of partial- + + + Requires dressing
thickness burns
Collagen membranes Temporary dressing of excised wounds + Decreases with time
Biobrane Dressing for donor site or excised Less than
wound allograft skin
Silastic-collagen Excised burn wounds + Excellent
composite29
Nylon-microporous Wounds with granulation tissue Excellent on
membrane composite32 (when
treated with clean wound
5% mafenide acetate)
Culture-grown Freshly excised wounds or those with Good under
epithelial sheets granulation tissue occlusive dressing
Culture-grown Excised wounds Good with sutures and
composite an occlusive dressing

*0 indicates none; +, minimal; , moderate; + + +, marked; + + + +, severe.

the cultured cells may have simply been sloughed as the
wound narrowed.
O'Connor et al cultured recipient epidermal cells until
confluent sheets filled a 50-mm tissue culture disk (14 to 21
days), and applied the sheets to full-thickness burn wounds
of two patients.34 The epithelial sheets did not have a
stratum corneum, but developed one "about" eight days
after application to either granulation tissue or freshly
excised wounds. Histologie examination performed approx¬
imately six months after application showed that all epider¬
mal layers were present, with no adenexal structures. The
epidermis was mildly hyperkeratotic, lacked rete pegs, was
thinner than normal, and had flattened basal cells. Mild
chronic inflammatory changes were evident in the connec¬
tive tissue underlying the culture-grown epidermal grafts.
Similar changes were seen beneath scar epidermis formed
by migration of epithelium from the wound margins. The
absence of a stratum corneum made these epidermal sheets
susceptible to desiccation and required that petroleum
jelly-gauze dressings be applied over the grafts for three to
four weeks. The main cause of failure of the culture-grown
material was infection, and the authors considered that it
was most successful when applied to freshly excised burn
wounds. Takes were noted not only with primary cultures
but also with second and tertiary subcultures, but the take
of the subcultured tissue was less certain. Noninfected
epidermis taken from patients critically ill with sepsis had a
low colony-forming capacity, and the authors recommended
that epidermis for such culture growth be harvested as soon
after admission as possible.
Recently, Bell et al have developed what they term "a
living skin equivalent," a composite consisting of a fibro-
blast-seeded collagen fibrillar lattice upon which dissoci¬
ated epidermal cells are cultured and proliferate.35 Those
investigators found that the addition of fibroblasts, which
had grown out from fragments of a full-thickness skin
biopsy, to a solution of Type I collagen caused the collagen to
form a fibrillar lattice that subsequently decreased in
volume and increased in consistency. This lattice prevented
shrinkage when implanted and provided tensile strength to
the composite. When grafted to full-thickness animal
wounds, rapid vascularization of the dermal component of
the composite was found to prevent graft necrosis. The
graft is said to inhibit wound contraction, with contraction
being "blocked by at least 75% or more in 80% of the grafts"
in a series of 52 grafts.
These composite grafts are said to persist for as long as
ten months, but it is uncertain whether the implanted
tissue survived or served as a scaffold for replacement by
host tissue. Considerable remodeling of the dermal compo¬
nent occurs, as indexed by a progressive increase in bire¬
fringence and a decrease in the ratio of labeled to unlabeled
fibroblasts across time, although radioactively labeled fi-
broblasts do persist for up to five weeks. Ten weeks after
implantation, the graft dermis had thinned to approx¬
imately half the thickness of adjacent host dermis, and
birefringence was greatest at the area of transition between
graft and host tissue.36 Additionally, epidermal hypertro¬
phy has been observed and the epidermal scaling described
by the investigators may represent sloughing of the graft
epidermis. Histologie sections showed tongues of host
epidermis invading the periphery of such composite grafts
along the epidermal-dermal interface of the grafts, sug¬
gesting replacement of the graft epidermis by scar epi¬
thelium of the host. The time required to form a
100-sq-cm composite graft from a 1-sq-cm biopsy specimen
is 20 to 26 days, and represents a major limitation to the
clinical use of this material, since grafting of burn wounds
must often be performed earlier in the postburn course,
particularly if excision is employed.

Downloaded From: http://archsurg.jamanetwork.com/ by a Georgetown University Medical Center User on 05/21/2015

 USES AND LIMITATIONS OF
MEMBRANE DRESSINGS
During the past decade, the desirable properties of an
effective burn-wound membrane dressing (skin substitute)
have been defined and a variety of candidate membranes
have been developed and evaluated. Certain membranes,
especially the bilaminate composites, appear, on the basis of
early studies, to be promising, but all have functioned
imperfectly and have specific limitations even when used
optimally (Table 2). The unilaminate membranes and some
bilaminates composed of either degradable or nondegrad-
able synthetic materials function like a serous crust, de¬
creasing wound pain but remaining susceptible to subgraft
suppuration and infection. The biodegradable and culture-
grown membranes appear to be susceptible to infection and
References
1. Pruitt BA Jr, Silverstein P: Methods of resurfacing denuded skin
areas. Transplant Proc 1971;3:1537-1545.
2. Miller TA, Switzer WE, Foley FD, et al: Early homografting of second
degree burns. Plast Reconstr Surg 1967;40:117-125.
3. Donaldson RC, Payne J, Hershey FB: Effects of storage on enzyme
activity and viability of skin. Surg Gynecol Obstet 1960;110:1-4.
4. Graham WB III, Hamilton RW, Lehr HB: Versatility of skin allografts:
Desirability of a viable frozen tissue bank. J Trauma 1971;11:494-501.
5. Young JM, Hyatt GW: Stored skin homografts in extensively burned
patients. Arch Surg 1960;80:208-213.
6. Pruitt BA Jr, Sell K, O'Neill JA Jr, et al: Clinical Evaluation of Freeze\x=req-\
Dried and Fresh Frozen Homograft on Burned Patients, Annual Research
Progress Report. Fort Sam Houston, Tex, US Army Surgical Research
Unit, Brooke Army Medical Center, section 10, June 30, 1967.
7. Robson MC, Krizek TJ, Koss N, et al: Amniotic membranes as a
temporary wound dressing. Surg Gynecol Obstet 1973;136:904-906.
8. Colocho G, Graham WB III, Greene AE, et al: Human amniotic
membrane as a physiologic wound dressing. Arch Surg 1974;109:370-373.
9. Faulk WP, Matthews R, Stevens PJ, et al: Human amnion as an adjunct
in wound healing. Lancet 1980;1:1156-1158.
10. Levine NS, Lindberg RA, Salsbury RE, et al: Comparison of coarse
mesh gauze with biologic dressings on granulating wounds. Am J Surg
1976;131:727-729.
11. Silverstein P, Curreri PW, Munster AM: Evaluation of Fresh Viable
Porcine Cutaneous Xenografts as a Temporary Burn Wound Cover, Annual
Research Progress Report. Fort Sam Houston, Tex, US Army Surgical
Research Unit, Brooke Army Medical Center, section 51, June 30, 1971.
12. Levine NS, Salisbury RE, Mason AD Jr: The effect of early surgical
excision and homografting on survival of burned rats and of interperi-
toneally-infected burned rats. Plast Reconstr Surg 1975;56:423-429.
13. Choy DSJ: Clinical trials of a new plastic dressing for burns and
surgical wounds. Arch Surg 1954;68:33-43.
14. Chardack WM, Brueske DA, Santomauro AP, et al: Experimental
studies on synthetic substitutes for skin and their use in the treatment of
burns. Ann Surg 1962;155:127.
15. Hall CW, Liotta D, DeBakey ME: Artificial skin. Trans Am Soc Artif
Intern Organs 1966;12:340-343.
16. Alexander JW, Wheeler LM, Rooney RC, et al: Clinical evaluation of
Epigard, a new synthetic substitute for homograft and heterograft skin.
J Trauma 1973;13:374-383.
17. Walder AI, May PD, Bingham CP, et al: Evaluation of synthetic films
as wound covers. Trans Am Soc Artif Intern Organs 1969;15:29-32.
18. Kornberg J, Burns NE, Kafesjian R, et al: Ultra thin silicon polymer
membrane: A new synthetic substitute: A preliminary study. Trans Am Soc
Artif Intern Organs 1972;18:39-43.
19. Warren RJ, Snelling CFT: Clinical evaluation of the Hydron burn
dressing. Plast Reconstr Surg 1980;66:361-368.
20. Dressler DP, Barbee WK, Sprenger R: The effect of Hydron burn
Downloaded From: http://archsurg.jamanetwork.com/ by a Georgetown University Medical Center User on 05/21/2015
are either replaced (to a variable extent) by scar tissue
produced by the recipient, or, if persistent, resemble trans¬
planted scars. The totally synthetic, biologically inert
composite membrane appears to have no intrinsic antibac¬
terial activity and functions best when pretreated with an
antimicrobial solution and applied to clean granulation
tissue. Allograft skin remains the optimum biologic dress¬
ing for temporary wound coverage. Generally, all the com¬
posite skin substitutes appear to function best when applied
to excised wounds or wounds with clean granulation tissue.
Further modifications of such membranes are necessary to
increase their range of application by enhancing their
resistance to infection and accelerating submembrane heal¬
ing to reduce scar formation and subsequent contracture
formation.
wound dressing on burned rat and rabbit ear wound healing. J Trauma
1980;20:1024-1028.
21. Wokalek H, Sch\l=o"\pfE, Vaubel E, et al: Erste erfahrungen mit einem
Transparent Flussigkeits-Gel bei der Behandlung frischer operations\x=req-\
wunden und chronischer Epitheldefekte der Haut. Aktuel Dermatol 1979;
5:255-265.
22. Nishi S: Clinical application of fibrin film for burns. Med Postgradu-
ates 1974;12:419-442.
23. Burleson RL, Ennulat N: Fibrin adherence to biologic tissues. J Surg
Res 1978;25:523-529.
24. Norton L, Chvapil M: Comparison of newer synthetic and biological
wound dressings. J Trauma 1981;21:463-468.
25. Tavis MJ, Harney JH, Thornton JW, et al: Modified collagen mem-
brane as a skin substitute. J Biomed Mater Res 1975;9:285-301.
26. Oluwasanmi J, Chvapil M: A comparative study of four materials in
local burn care in rabbit model. J Trauma 1976;16:348-353.
27. Tavis MJ, Woodroof EA, Roth J, et al: New types of skin prostheses:
Preliminary studies. Read before the 11th Annual Meeting of the American
Burn Association, New Orleans, March 15, 1979.
28. Lin SD, Robb EC, Nathan P: A comparison of IP-758 and Biobrane in
rats as temporary protective dressings on widely expanded meshed auto-
grafts. J Biomed Mater Res 1982;3:220-222.
29. Burke JF, Yannas IV, Quinby WC Jr, et al: Successful use of a
physiologically acceptable artificial skin in the treatment of extensive burn
injury. Ann Surg 1981;194:413-428.
30. Yannas IV, Burke JF, Orgill DP, et al: Wound tissue can utilize a
polymeric template to synthesize a functional extension of skin. Science
1982;215:174-176.
31. Levine NS, Salisbury RE, Peterson HD: Continued Evaluation of
Split-Thickness Cutaneous Xenograft and Synthetic Materials as Tempo-
rary Biologic Wound Covers for Burned Soldiers, Annual Research Prog-
ress Report. Fort Sam Houston, Tex, US Army Institute of Surgical
Research, Brooke Army Medical Center, section 40, June 1974.
32. Levine NS, Peterson HD, Mason AD Jr: Use of a Synthetic Dressing
on Denuded Wounds in Burned Patients, Annual Research Progress
Report. Fort Sam Houston, Tex, US Army Institute of Surgical Research,
Brooke Army Medical Center, June 1975, pp 343-363.
33. Eisinger M, Morden M, Raaf JH, et al: Wound coverage by a sheet of
epidermal cells grown in vitro from dispersed single cell preparations.
Surgery 1980;88:287-293.
34. O'Connor NE, Mulliken JB, Banks-Schlegel S, et al: Grafting of
burns with cultured epithelium prepared from autologous epidermal cells.
Lancet 1981;1:75-78.
35. Bell E, Ehrlich HP, Sher S, et al: Development and use of a living skin
equivalent. Plast Reconstr Surg 1981;67:386-392.
36. Bell E, Ehrlich HP, Buttle DJ, et al: Living tissue formed in vitro and
accepted as skin-equivalent tissue of full thickness. Science 1981;211:1052\x=req-\
1054.