novel biodegradable composite wound dressings with

Novel Biodegradable Composite Wound Dressings WithControlled Release of Antibiotics: Microstructure,Mechanical and Physical Properties

Jonathan J. Elsner, Adaya Shefy-Peleg, Meital Zilberman

Department of Biomedical Engineering, Faculty of Engineering, Tel-Aviv University, Tel-Aviv 69978, Israel

Received 10 August 2009; revised 7 October 2009; accepted 1 November 2009

Published online 2 February 2010 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.b.31599

Abstract: Wound dressings aim to restore the milieu required

for skin regeneration and protect the wound from environmen-

tal threats, including penetration of bacteria. The dressings

should be easy to apply and remove and maintain a moist

healing environment. In this study, novel biodegradable com-

posite wound dressings based on a polyglyconate mesh and a

porous PDLGA binding matrix were developed and studied.

These novel dressings were prepared by dip-coating woven

meshes in inverted emulsions, followed by freeze-drying. Their

investigation focused on the microstructure, mechanical and

physical properties, and the release profile of the antibiotic

drug ceftazidime from the binding matrix. The mechanical

properties of our wound-dressing structures were found to be

superior, combining relatively high tensile strength and ductil-

ity, which changed only slightly during 3 weeks of incubation

in an aqueous medium. The parameters of the inverted emul-

sion, the organic–aqueous phase ratio, and the type of surfac-

tant used for stabilizing the emulsion were found to affect the

microstructure of the binding matrix and the resulting proper-

ties, i.e., water absorbance, water vapor transmission rate, and

drug-release profile from the binding matrix. Appropriate selec-

tion of these parameters can yield composite structures that

have the desired physical properties and drug release behav-

ior. Thus, these unique structures are potentially very useful as

burn and ulcer dressings. VC 2010 Wiley Periodicals, Inc. J Biomed

Mater Res Part B: Appl Biomater 93B: 425–435, 2010

Key Words: water vapor transmission rate, water absorption,

ceftazidime, poly-(DL-lactic-co-glycolic acid), controlled drug

delivery

INTRODUCTION

The main goal in wound management is to achieve rapidhealing with functional and esthetic results. An ideal wounddressing can restore the milieu required for the healing pro-cess, while protecting the wound bed against penetration ofbacteria and environmental threats. The dressing shouldalso be easy to apply and remove. Most modern dressingsare designed to maintain a moist healing environment andto accelerate healing by preventing cellular dehydration andpromoting collagen synthesis and angiogenesis.1 Nonethe-less, overrestriction of water evaporation from the woundshould be avoided, as accumulation of fluid under the dress-ing may cause maceration and facilitate infection. The watervapor transmission rate (WVTR) from the skin has beenfound to vary considerably depending on the wound typeand healing stage, increasing from 204 g m�2 day�1 for nor-mal skin to 278 and as much as 5138 g m�2 day�1 for firstdegree burns and granulating wounds, respectively.2 Thephysical and chemical properties of the dressing shouldtherefore be adapted to the type of wound and to thedegree of wound exudation.

A range of dressing formats based on films, hydrophilicgels, and foams are available or have been investigated.

Thin semipermeable polyurethane films coated with a layerof acrylic adhesive, such as OptsiteV

R

(Smith & Nephew) andBioclussiveV

R

(J & J), are typically used for minor burns,postoperative wounds, and a variety of minor injuriesincluding abrasions and lacerations. Gels such as carboxy-methylcellulose-based Intrasite GelV

R

(Smith & Nephew) andalginate-based TegagelV

R

(3M) are used for many differenttypes of wounds, including leg ulcers and pressure sores.These gels promote rapid debridement by facilitating rehy-dration and autolysis of dead tissue. Foam dressings, suchas Lyofoam (Molnlycke Healthcare) and Allevyn (Smith &Nephew), are used to dress a variety of exudating wounds,including leg and decubitus ulcers, burns, and donor sites.

Films and gels have a limited absorbance capacity andare recommended for light to moderately exudating wounds,whereas foams are highly absorbent and have a high WVTRand are, therefore, considered more suitable for woundswith moderate to heavy exudation.3 The characteristics ofthe latter are controlled by the foam texture, pore size, anddressing thickness.

Bacterial contamination of a wound seriously threatensits healing. In burns, infection is the major complication af-ter the initial period of shock, and it is estimated that about

Correspondence to: M. Zilberman; e-mail: [email protected]

Contract grant sponsor: Israel Science Foundation (ISF); contract grant number: 1312/07.

Contract grant sponsor: Israel Ministry of Health; contract grant number: 3-3943

VC 2010 WILEY PERIODICALS, INC. 425

75% of the mortality following burn injuries is related toinfections rather than to osmotic shock and hypovolemia.4

Bacteria in wounds are able to produce a biofilm within�10 h. This biofilm protects them against antibiotics andimmune cells already in the early stages of the infectionprocess.5 The rapidity of biofilm growth suggests thatefforts to prevent or slow the proliferation of bacteria andbiofilms should begin immediately after creation of thewound. This has encouraged the development of improvedwound dressings that provide an antimicrobial effect byeluting germicidal compounds such as iodine (IodosorbV

R

,Smith & Nephew), chlorohexidime (BiopatchV

R

, J & J), ormost frequently silver ions (e.g., ActicoatV

R

by Smith &Nephew, ActisorbV

R

by J & J, and AquacellVR

by ConvaTec).Such dressings are designed to provide controlled releaseof the active agent through a slow but sustained releasemechanism which helps avoid toxicity yet ensures deliveryof a therapeutic dose to the wound. Some concernsregarding safety issues related to the silver ions includedin most products have been raised. Furthermore, suchdressings still require frequent change, which may bepainful to the patient and may damage the vulnerableunderlying skin, thus increasing the risk of secondarycontamination.

Bioresorbable dressings successfully address this short-coming, as they do not need to be removed from the woundsurface once they have fulfilled their role. Biodegradablefilm dressings made of lactide-caprolactone copolymerssuch as TopkinV

R

(Biomet) and OprafolVR

(Lohmann &Rauscher) are currently available. Bioresorbable dressingsbased on biological materials such as collagen and chitosanhave been reported to perform better than conventional andsynthetic dressings in accelerating granulation tissue forma-tion and epithelialization.6,7 However, controlling the releaseof antibiotics from these materials is challenging because oftheir hydrophilic nature. In most cases, the drug reservoir isdepleted in less than 2 days, resulting in a very short anti-bacterial effect.8,9

There is currently no available synthetic dressing thatcombines the advantages of occlusive dressings with biode-gradability and intrinsic topical antibiotic treatment. Wehave recently reported the development of biodegradablecomposite core/shell fibers loaded with antibiotics whichcould potentially be used as a basic unit for such wounddressings.10 In this study, we developed a composite wounddressing based on this concept. We propose a plain-woventextile made of biodegradable polyglyconate fibers boundtogether with a continuous porous matrix that affords thedressing an occlusive nature. The reinforcing fibers’ excel-lent mechanical properties afford good mechanical strength,whereas the continuous binding matrix can be tailored toproduce the desired effect: e.g., drug release kinetics, waterabsorbance, and other physical properties that promotewound healing. The goal of this study was, therefore, to de-velop this type of dressing and characterize its main fea-tures: mechanical and physical properties and drug releaseprofile to elucidate whether it may have better wound heal-ing properties.

MATERIALS AND METHODS

MaterialsPolymers. MaxonTM, polyglyconate monofilament sutureswith a diameter of 0.20–0.25 mm, United States Surgical,USA, were woven to produce a mesh.

Bioresorbable porous structures (binding matrices) weremade of 50/50 poly(DL-lactic-co-glycolic acid) (PDLGA),inherent viscosity (i.v.) ¼ 0.56 dL/g (in CHCl3 at 30�C,MW � 100 KDa), Absorbable Polymer Technologies, USA.

Drug. Ceftazidime hydrate, 90–105%, Sigma (C-3809).

Surface active agent. Bovine serum albumin (BSA), MW ¼66,000 Da, Sigma (A-4503).

Sorbitan monooleate (Span 80), MW 428.608 g mol�1,Sigma (85548).

Reagents. 1,1,1,3,3,3-Hexafluoro-2-propanol (H1008) waspurchased from Spectrum Chemical Mfg. Corp.

Wound dressing preparationFiber surface treatment. The sutures were surface treatedto dispose of the original fiber coating and enhance adhe-sion between the core fiber and the binding porous matrix.The MaxonTM fibers were gently wrapped around flexibleTeflon frames and dipped in 1,1,1,3,3,3 hexafluoro-2-propa-nol for 40 s. The fibers were then washed with 70% ethanoland dried. The surface treated Maxon fibers were hand-wo-ven into a plain-woven fabric (the most basic fundamentaltextile weave) with 1.5 mm gaps between adjacent fibers.

Inverted emulsion formation. An organic solution was pre-pared by dissolving a known amount of PDLGA (15% w/v)in chloroform. An aqueous solution was obtained by dissolv-ing a known amount of the drug (5% w/w relative to thepolymer load) and surfactant. The aqueous solution was thenpoured into the organic solution (in a test tube) and homoge-nized to an inverted emulsion using a Kinematica PT-3100Polytron homogenizer operating at 18,000 rpm for 1.5 min.Three formulations were prepared:

a. Formulation with organic–aqueous phase ratio (O:A) ¼6:1 containing 1% w/v BSA (relative to the aqueous vol-ume) as surfactant. This formulation is termed BSA1.

b. Formulation with O:A ¼ 12:1 containing 1% w/v BSA(relative to the aqueous volume) as surfactant. This for-mulation is termed BSA2.

c. Formulation with O:A ¼ 12:1 containing 1% w/v Span80 (relative to the organic phase volume) as surfactant.This formulation is termed SPAN.

The woven structure was then dip coated (while placedon holders) in fresh inverted emulsions and then immedi-ately frozen in a liquid nitrogen bath. The samples werethen placed in a precooled (�105�C) freeze-dryer (Virtis101 equipped with a nitrogen trap) and freeze-dried to pre-serve the microstructure of the emulsion-based structures.Drying was performed in two stages: The freeze-dryerchamber pressure was reduced to 100 mTorr while the

426 ELSNER ET AL. NOVEL BIODEGRADABLE COMPOSITE WOUND DRESSINGS

temperature remained at �100�C. After 3 h a hot plate wasturned on to �45�C for an additional 12 h. The condenserwas then turned off and its plate temperature graduallyincreased to room temperature while the pressure wasmonitored between 100 and 700 mTorr. During this step,the liquid nitrogen trap condensed the excess water and sol-vent vapors. The dried samples were stored in desiccatorsuntil use.

Morphological characterizationThe morphology of the wound dressing’s structures wasobserved using a Jeol JSM-6300 scanning electron micro-scope (SEM) at an accelerating voltage of 5 kV. Cryo-genically fractured surfaces were Pd-sputtered beforeobservation.

In vitro changes in the morphology of wet wound dress-ing structures were characterized using an environmentalSEM (Quanta 200 FEG ESEM) at an accelerating voltage of10 kV and a pressure of 4.5 Torr. Specimens were fixed to aspecial base, fractured surface facing up, and maintained inPBS (37�C, pH 7.0) for 14 days. Initial fractographs weretaken and coordinates were recorded so as to enable returnto these coordinates at 2, 4, and 14 days after immersion inPBS.

The mean pore diameter (n ¼ 100 pores) and porosityof the observed morphologies was analyzed using the SigmaScan Pro software and statistics were calculated using theSPSS 10 software. Statistical significance was determinedusing the ANOVA (Tukey-Kramer) method. The area occu-pied by the pores was calculated for each SEM fractographusing the Sigma Scan Pro software to evaluate the porosityof the samples. Porosity was determined as the area occu-pied by the pores divided by the total area.

Water vapor transmission rateThe moisture permeability of the wound dressing wasdetermined by measuring the WVTR across the dressing.A Sheen Payne permeability cup with an exposure area of10 cm2 was filled with 10 mL PBS and covered with a circu-lar wound dressing. The cup was placed in a straight posi-tion inside an oven at 37�C, containing 1 kg of freshly driedsilica gel to maintain relatively low humidity conditions. Theweight of the assembly was measured every hour for 12 hand a graph of the evaporated water versus time was plot-ted. WVTR was calculated by the formula:

WVTR ¼ slope� 24

area

g

m2 � day� �

(1)

Water uptake abilityThe fluid absorbing capacity of a wound dressing is an im-portant criterion for maintaining a moist environment overthe wound bed. Water uptake was measured over 7 days.Dry wound dressings were cut into 1 cm � 1.5 cm rectan-gles, weighed, and placed in bottles containing 2 mL PBS(pH 7.0). The bottles were closed and placed in an incuba-

tor at 37�C. The weight of the samples was measured after6 h, 12 h, 1, 2, 3, and 7 days by removing the PBS and blot-ting them gently to remove excess fluid. The water uptakewas calculated as:

Water uptake ð%Þ ¼ Wwet �Wdry

Wdry� 100 (2)

Tensile mechanical propertiesThe wound dressing’s tensile mechanical properties weremeasured at room temperature, under unidirectional ten-sion at a rate of 10 mm/min, using a 5500 Instron machine.Each wound dressing sample was cut into a dog bone shape(neck length 5 cm, width 1 cm). The tensile strength wasdefined as the maximum strength in the stress–strain curve.The maximal strain was defined as the breaking strain.Young’s modulus was defined as the slope of the stress–strain curve in the elastic (linear) region. Four sampleswere tested for each type of specimen. Additional specimenswere immersed in phosphate buffered saline (PBS), pH 7.0,at 37�C for 1, 2, and 3 weeks, after which they were driedand tested in the same manner.

The means and standard deviations were calculatedusing the SPSS 10 software. ANOVA (Tukey-Kramer) wasused for group comparison.

In vitro drug release studiesThe composite wound dressings were immersed in PBS at37�C for 28 days to determine the various drug releasekinetics from these structures. The release studies wereconducted in closed glass vessels containing 1.5 mL PBS.The medium was removed (completely) periodically, at eachsampling time (6 h, 1, 2, 3, 7, 14, 21, and 28 days), andfresh medium was introduced.

Ceftazidime assay. The medium’s ceftazidime content wasdetermined using a Jasco HPLC with a UV 2075 plus detec-tor and a reverse phase column (InterstilV

R

ODS-3V 5 lm,inner diameter d ¼ 4.6 mm, length ¼ 250 mm), kept at25�C. The mobile phase consisted of a mixture of PBS andacetonitrile (95/5, v/v) at a flow rate of 1 mL/min with aquaternary gradient pump (PU 2089 plus) without gradient.Twenty microliters of samples were injected with an auto-sampler (AS 2057 Plus). The column effluent was eluted for22 min and detected at 254 nm. The area of each elutedpeak was integrated using EZstart software version 3.1.7. Acalibration curve was prepared for concentrations rangingfrom 1.0 to 200.0 lg/mL (correlation coefficient > 0.999,slope: 0.0000318).

Residual drug recovery from composite dressingsResidual drug recovery from the composite dressings wasmeasured as follows: the fibers were placed in 1 mL meth-ylene chloride to dissolve the remaining PDLGA coating.Two mililiters of water was then added to dissolve thehydrophilic drug residues. The materials were vortexed for30 s and then left to stand until phase separation occurred.

ORIGINAL RESEARCH REPORT

JOURNAL OF BIOMEDICAL MATERIALS RESEARCH B: APPLIED BIOMATERIALS | MAY 2010 VOL 93B, ISSUE 2 427

The aqueous phase was filtered to dispose of polymer par-ticles. The drug concentration was estimated using theaforementioned assays.

RESULTS

Morphological characterization of dry wound dressingsA composite wound-dressing construction composed of aplain-woven mesh of polyglyconate fibers bound by a con-tinuous PDLGA porous matrix loaded with the antibiotic cef-tazidime was developed and studied. A SEM fractographshowing a basic unit structure of the wound dressing is pre-sented in Figure 1(a). The PDLGA matrix adhered well to

the fibers, forming a skin layer �60 lm thick, as demon-strated in Figure 1(b).

The freeze-drying of inverted emulsions technique thatwas used to create the PDLGA-binding matrix resulted in aporous microstructure [shown in Figure 1(c,d)] which alsoacts as a reservoir for the antibiotic that is incorporated init. Three different emulsion formulations were used in thisstudy, as listed in Table I. These three formulations yieldeddifferent resultant matrix microstructures. The microstruc-ture of the BSA1 sample is highly porous [Figure 2(a)] withan average porosity of 63 6 4% and pore diameter of 1.46 0.3 lm (Table I). An increase in the emulsions’ organic–aqueous phase ratio from 6:1 to 12:1 (formulation BSA2)resulted in larger polymer domains between pores, lesspore connectivity, and a lower porosity of 35% 6 2% [Fig-ure 2(b) and Table I]. However, it did not significantly affectthe pore size, which remained 1.4 6 0.3 lm. When the sur-factant BSA which was used to stabilize the emulsion wasreplaced with Span 80 (formulation SPAN) for the same O–A phase ratio (12:1), the mean pore size decreased to 1.1 60.3 lm, whereas porosity increased to 45% 6 7% [Figure2(c) and Table I].

Water vapor transmission rateWVTR was measured for the three types of dressings, basedon the formulations in Table I. Evaporative water lossthrough the various dressings [Figure 3(a)] was linearlydependant on time (R2 > 0.99 in all cases), resulting in aconstant WVTR. A WVTR of 3452 6 116 g m�2 day�1 wasmeasured for a dressing based on the BSA1 formulation.When the O:A phase ratio was increased to 12:1 (BSA2), theWVTR decreased significantly to 480 6 69 g m�2 day�1.A WVTR of 2641 6 42 g m�2 day�1 was measured whenthe surfactant BSA was replaced with Span 80 (SPAN). TheWVTR was determined experimentally also for a dense non-porous PDLGA film, prepared using the solution castingtechnique as described elsewhere,11 to serve an analoguefor biodegradable films currently in use for wound care, aswell as to elucidate the effect of porosity on the WVTR. AWVTR of 356 6 106 g m�2 day�1 was measured in thiscase. The WVTR of an exposed aqueous surface was alsodetermined experimentally, in order to simulate a conditionin which no dressing is applied to the wound surface. Inthis case, the WVTR was 6329 6 725g m�2 day�1.

FIGURE 1. SEM fractographs of a ceftazidime-loaded composite

wound dressing: (a) A representative plain-weave basic unit structure,

(b) a cross section view of the reinforcing fibers and binding drug-

loaded porous matrix, (c) and (d) detailed views of the continuous po-

rous matrix structure.

TABLE I. Structural Characteristics of the Ceftazidime-Loaded Porous Matrix

Specimen

Organic–Aqueous

Phase RatioDrug Loadinga

(w/w)

Polymer Contentin the OrganicPhaseb (w/w) Surfactantb

Freeze-Dried Emulsion

Porosity Pore Diameter

BSA1 6:1 5% 15% BSA (1% w/v in theaqueous phase)

63 6 4% 1.4 6 0.3 lm

BSA2 12:1 BSA (1% w/v in theaqueous phase)

35 6 2% 1.4 6 0.3 lm

SPAN 12:1 Span80 (1% w/v in theorganic phase)

45 6 7% 1.1 6 0.3 lm

a Relative to the polymer weight.b Relative to a liquid phase volume (organic or aqueous).


Water uptakeBSA1 (O:A¼6:1) and BSA2 (O:A¼12:1) formulations wereused to measure the fluid absorption capacity of our newwound dressings. Samples were placed in PBS (pH. 7.0) tosimulate the water absorption behavior in the presence ofwound fluids. Water absorption ability was calculatedaccording to the formula described in Experimental Section.Both types of dressings displayed similar temporal water

absorption patterns, consisting of the following stages [Fig-ure 4(a)]:

1. A rapid initial water uptake within the first 24 h. A con-siderable share of the initial water uptake actuallyoccurred during the first 6 h [Figure 4(a), inset]. A 65%(w/w) increase in the weight of the BSA1 type dressingswas observed at this stage, whereas dressings of typeBSA2 increased by 56%.

2. A slight decrease in water content during the following 3days, reaching �45% w/w for both types of dressings.

3. A steady increase over the next 2 weeks, reaching a125% increase in weight after 3 weeks, for both types ofdressings.

The changes in the microstructure of the wet matrixsamples because of water uptake showed initial reorganiza-tion of the porous structure with a small decrease in poresize, followed by gradual thickening of the polymeric walls,as presented in Figures 4(b–e). These changes led tocoarser porous structures.

Tensile mechanical propertiesThe effect of polymer degradation on the wound dressing’smechanical properties was determined using the BSA1 for-mulation. The results are presented in Figure 5. Dressingsincubated for 1, 2, and 3 weeks displayed similar tensilestrengths (21–27 MPa) and strains at break (55–63%).When failure occurred, it was initiated because of break ofthe reinforcing fibers and not by the matrix. The initialYoung’s modulus of the dressing (126 6 27 MPa) was pre-served for 1 week of incubation (117 6 19 MPa) and thendecreased to �70 MPa after 2 weeks.

In vitro drug-release studiesThe release profile of antibiotics from wound dressings var-ied considerably with the changes in formulation (Figure 6).Ceftazidime release from the dressings based on the BSA1formulation was relatively short, reaching almost completerelease of the encapsulated drug within 24 h. An increase inthe emulsion’s O:A phase ratio from 6:1 to 12:1 reduced theburst release. Specifically, burst release values of 97 and57% were recorded after 6 h for formulations BSA1 andBSA2, respectively, after which the release of the antibioticsfrom BSA2 dressings continued for 5 days at a decreasingrate. The ceftazidime release profile from the SPAN formula-tion was totally different. It exhibited a low burst release of6% during the first 6 h of incubation and then a releasepattern of a nearly constant rate for 10 days.

DISCUSSION

A composite wound dressing based on a fibrous polyglyco-nate mesh embedded within a porous drug-loaded PDLGAporous matrix was developed and studied. Composites aremade up of individual materials, matrix, and reinforcement.The matrix component supports the reinforcement materialby maintaining its relative positions and the reinforcementmaterial imparts its special mechanical properties toenhance the matrix properties. Taken together, both

FIGURE 2. SEM fractographs showing the effect of a change in the

emulsion’s formulation parameters on the microstructure of the bind-

ing matrix containing 5% w/w ceftazidime and 15% w/v polymer (50/

50 PDLGA, MW 100 KDa): (a) O:A phase ratio of 6:1 and 1% w/v BSA

(BSA1), (b) O:A phase ratio of 12:1 and 1% w/v BSA (BSA2), (c) O:A

phase ratio of 12:1 and 1% w/v Span 80 (SPAN).



materials synergistically produce properties unavailable inthe individual constituent materials, allowing the designerto choose an optimum combination. In our application, thereinforcing polyglyconate mesh affords the necessary me-

chanical strength to the dressing, whereas the porousPDLGA binding matrix is aimed to provide adequate mois-ture control and release of antibiotics to protect the woundbed from infection and promote healing. Both structural

FIGURE 4. (a) Water uptake by two dressing formulations containing: 5% w/w ceftazidime, 15% w/v polymer (50/50 PDLGA, MW 100 KDa), stabi-

lized with 1% w/v: (l) BSA1 (O:A ¼ 6:1) and (n) BSA2 (O:A ¼ 12:1). (b–e) E-SEM fractographs of BSA1, recorded during the water uptake experi-

ment at 0 (b), 2 (c), 4 (d) and 14 days (e). In addition, larger domains of (b) and (e) are presented. [Color figure can be viewed in the online

issue, which is available at www.interscience.wiley.com.]

FIGURE 3. (a) Water loss versus time plots: n-Uncovered surface, l-BSA1 formulation, n-BSA2 formulation, ~ SPAN formulation, and h-dense

(non-porous) PDLGA (50/50, MW 100 KDa) film, (b) Water vapor transmission rates for the various wound dressings. [Color figure can be viewed

in the online issue, which is available at www. interscience.wiley.com.]


constituents are biodegradable, thus enabling easy removalof the wound dressing from the wound surface once it hasfulfilled its role.

The freeze-drying of inverted emulsions technique whichwas used to create the porous binding matrix is unique inits ability to preserve the liquid structure in the solid state.The viscous emulsion, consisting of a continuous PDLGA/chloroform solution phase and a dispersed aqueous drug so-lution, formed good contact with the mesh during the dip-coating process. Consequently, an unbroken solid porousmatrix was deposited by the emulsion following freeze-dry-ing, as demonstrated in Figure 1(a).

The freeze-drying of inverted emulsions technique hasseveral advantages. First, it enables attaining a thin uninter-rupted barrier, which unlike mesh or gauze alone can betterprotect the wound bed against environmental threats anddehydration. Second, it entails very mild processing condi-

tions which enable the incorporation of sensitive bioactiveagents such as antibiotics10 and even growth factors12 tohelp reduce the bio-burden in the wound bed and acceler-ate wound healing. Third, the microstructure of the freeze-dried matrix can be customized through modifications ofthe emulsion’s formulation to exhibit different attributes,namely different porosities or drug release profiles. Threeformulations were used in this study and their effects onthe microstructure, the resulting physical and mechanicalproperties, and the drug release profile were studied.

Microstructural and physical propertiesIn this study, we focused on three matrix formulations(listed in Table I), which display distinctly different micro-structural features (Figure 2 and Table I). The effect of theO:A phase ratio was examined on formulations containingBSA as surfactant. As expected, a higher O:A phase ratio, i.e.lower aqueous phase quantity, resulted in a smaller porosityof the solid structure. However, both microstructures werehomogenous and characterized by a similar average poresize. The stabilization effect of Span 80 was even higherthan that obtained using BSA and therefore resulted in asmaller pore size (Table I).

Successful wound healing requires a moist environment.Two parameters must therefore be determined: the wateruptake ability of the dressing and the WVTR through thedressing. An excessive WVTR may lead to wound dehydra-tion and adherence of the dressing to the wound bed,whereas a low WVTR might lead to maceration of healthysurrounding tissue and buildup of a back pressure and painto the patient. A low WVTR may also lead to leakage fromthe edges of the dressing which may result in dehydration

FIGURE 5. (a) Tensile stress–strain curves for wound dressings immersed in PBS for 0 (–0–), 1 (–1–), 2 (–2–), and 3 (–3–) weeks. (b) Young’s mod-

ulus, (c) tensile strength, and (d) maximal tensile strain as a function of immersion time. Comparison was made using ANOVA and significant

differences are indicated (*). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

FIGURE 6. Cumulative release of ceftazidime from wound dressings

based on the following formulations: l-BSA1, n-BSA2, ~-SPAN.

[Color figure can be viewed in the online issue, which is available at

www.interscience.wiley.com.]



and bacterial penetration.3,13 It has been claimed that awound dressing should ideally possess a WVTR in the rangeof 2000–2500 g m�2 day�1, half of that of a granulatingwound.13 In practice, however, commercial dressings do notnecessarily conform to this range and have been shown tocover a larger spectrum of WVTR, ranging from 90(DermiflexV

R

, J & J14) to 3350 g m�2 day�1 (BeschitinVR

, Uni-tika15). Clearly, the WVTR is related to the structural prop-erties (thickness, porosity) of the dressing as well as to thechemical properties of the material from which it is made.The results for our dressings, shown in Figure 3, demonstratehow the WVTR can be customized based on modifications ofthe porous matrix’s microstructure. In this study, we exam-ined what we consider the upper (12:1) and lower (6:1) O:Aphase ratio boundaries which would be suitable for this typeof application, in terms of emulsion stability and in terms ofthe desired drug release profiles. We were able to cover alarge range of WVTRs, between 480 and 3452 g m�2 day�1.The lowest value is similar to that reported for film-typedressings (e.g., Tegaderm, 491 6 44 g m�2 day�1),16 whereasthe highest value is similar to that of foam-type dressings(e.g., Lyofoam, 3052 6 684 g m�2 day�1).16 Based on ourprevious results, where it was shown that the porosity offreeze-dried emulsions can be altered progressively by modifi-cation of the emulsions’ O:A ratio,10 there is good reason tobelieve that a WVTR in this range can be fine tuned by alter-ing the O:A phase ratio of formulations which include BSA assurfactant anywhere between 6:1 and 12:1, and specificallyto the 2000–2500 g m�2 day�1 range. A WVTR of 2641 642 g m�2 day�1 that was achieved with the SPAN formulationis close to this range and seems the most appropriate.

Water uptake by the wound dressing may occur eitheras the result of water entry into accessible voids in the po-rous matrix structure (hydration effect) or as the polymermatrix material gradually uptakes water and swells (swel-ling effect). Our water uptake patterns for wound dressingsbased on the BSA1 and BSA2 formulations (Figure 4) dem-onstrated both these effects. As mentioned in the resultssection, both types of wound dressing demonstrated athree-stage water uptake pattern. Examination of the wateruptake process through temporal microstructural changes inthe polymeric matrix sheds light on these stages, as follows:

Stage 1 (governed by hydration): during this stage ofwater uptake, a quick flux of water associated with hydra-tion of the porous structures was measured within 6 h afterimmersion in PBS for both types of dressings, as demon-strated in Figure 4(a). Water content then plateaued at val-ues of 65% for the BSA1 formulation with the higher poros-ity (63%) and 55% for the BSA2 formulation with thesmaller porosity (35%).

Stage 2: A small decrease in water content during days2–4, probably due to gradual shrinkage of pore walls andreorganization of the porous matrix [Figures 4(b–d) and Ta-ble II]. Such changes could be provoked by a combinedeffect of the polymer’s glass transition temperature that isvery close to the incubation temperature (37 �C) in combi-nation with a softening effect of water on the polymer. Ithas been shown, for instance, that amorphous electrospun

PDLGA fibrous mats undergo drastic shrinkage after 1 dayof in vitro incubation because of the relaxation of extendedamorphous chains.17 A 26% decrease in the void fraction ofthe polymer matrix [Figure 4(e) and Table II) evidentlycaused a reduction in the water volume contained withinthe matrix (�20%) during this phase.

Stage 3 (governed by swelling): After the 4th day ofimmersion in the aqueous medium, water uptake increasedgradually and similarly for both types of dressing over theduration of 3 weeks, ultimately reaching a twofold increase.The process of swelling is dependent on the polymer’swater affinity. As PDLGA is not as hydrophilic as hydrogelsor natural polymers used in this application, swellingoccurred slowly. It is believed that the swelling effect wasenhanced over time as hydrophilic end groups became moreabundant because of polymer degradation by hydrolysis.This stage is also characterized by gradual thickening of thepolymer walls because of the increased water uptake andcreation of larger voids in the matrix because of polymerdegradation [Figure 4(e)]. The combination of these twochanges, which occurred in parallel, resulted is a coarsermicrostructure.

We appreciate that the gradual increase of porositybecause of degradation would consequently result in anincrease over time in the WVTR values presented thus far.Such changes in WVTR are not a matter for concern as it isexpected that the barrier properties of the regenerating skinwould be restored at a corresponding rate. Furthermore, agradual increase in WVTR over time could encouragewound closure as shown by Schunck et al. (2005) in arecent study comparing in vivo response to occlusive versusnonocclusive dressings.18 They discovered that the use ofocclusive dressings during the entire or 2 weeks of the heal-ing period promotes cell migration but delays wound clo-sure and so a gradual reduction in the dressing’s barrierproperties over 1–2 weeks could have a beneficial effect onwound closure.

It should also be mentioned that delamination of thefiber-matrix interface may affect the integrity of the wounddressing as well as barrier properties. We have, therefore,examined the effect of degradation on the interface andfound that the dressing material maintained its integrity forthe duration of at least 2 weeks exposure to aqueous me-dium, despite a gradual ongoing process of degradation ofthe matrix component. SEM observations [Figure 7] showedhigh quality fiber-matrix interface, i.e. contact between thetwo components still exists as degradation proceeds.

TABLE II. Structural Characteristics of the BSA1 Wound

Dressing at Various Times of Incubation in Aqueous Medium

Incubation Time (Days) Porosity (%) Pore Diameter (lm)

0 63 6 4 1.4 6 0.32 53 6 6 1.2 6 0.34 37 6 2 1.2 6 0.3

14 57 6 3 N.A.


Tensile mechanical propertiesThe mechanical properties of a wound dressing are an im-portant factor in its performance, whether it is to be usedtopically to protect cutaneous wounds or as an internalwound support, e.g. for surgical tissue defects or herniarepair. Furthermore, in the clinical setting, appropriate me-chanical properties of dressing materials are needed toensure that the dressing will not be damaged by handling.Porous structures typically possess inferior mechanicalproperties compared to dense structures, yet in woundhealing applications porosity is an essential requirement for

diffusion of gasses, nutrients, cell migration, and tissuegrowth. Most wound dressings are, therefore, designedaccording to the bilayer composite structure concept andconsist of an upper dense ‘‘skin’’ layer to protect the woundmechanically and prevent bacterial penetration and a lowerspongy layer designed to adsorb wound exudates andaccommodate newly formed tissue. Our new dressing designintegrates both structural/mechanical and functional com-ponents (e.g. drug release and moisture management) in asingle composite layer. It combines relatively high tensilestrength and modulus together with good flexibility (elonga-tion at break). It actually demonstrated better mechanicalproperties than most other dressings currently used orstudied, as demonstrated in Table III.

The initial mechanical properties of natural polymerssuch as collagen or gelatin can be satisfactory. However,considerable degradation of these properties is expected tooccur rapidly due to hydration20 and enzymatic activity.23

The results of the 3 weeks degradation study of our wounddressings show a significant decrease only in Young’s modu-lus (Figure 5). The maximal stress and strain of our com-posite wound dressing (24 MPa and 55%, respectively) aredictated mainly by the mechanical properties of the rein-forcing fibers which fail first during breakage. At these timeperiods, they are not subjected to considerable degradation,which explains the constancy in these properties. In contra-distinction, the Young’s modulus of the dressings is consid-erably affected by the properties of the binding matrix thatmakes up the largest part of the cross-sectional area. Thedegradation of the matrix material which is clearly in

FIGURE 7. SEM fractographs demonstrating the interface between the

reinforcing fiber and porous matrix for specimens based on formula-

tion BSA1, immersed in aqueous medium for: (a) start point, (b) 1

week, and (c) 2 weeks.

TABLE III. Mechanical Properties of Various Wound

Dressings

Material/FormatElastic

Modulus (MPa)

TensileStrength(MPa)

Elongationat Break (%)

BSA1 (compositepolyglyconatemesh, coatedwith PDLGAporous matrix)

126 6 27 24.2 6 4.5 55 6 5

Electrospunpoly-(L-lactide-co-e-caprolactone)(50:50) mat19

8.4 6 0.9 4.7 6 2.1 960 6 220

Electrospungelatin mat19

490 6 52 1.6 6 0.6 17.0 6 4.4

Electrospuncollagen mat20

11.4 6 1.2

ResolutVR

LTregenerativemembrane(Gore). Glycolidefiber mesh coatedwith an occlusivePDLGA membrane21

11.7 20

KaltostatVR

(ConvaTec)Calcium/SodiumAlginate fleece22

1.3 6 0.2 0.9 6 0.1 10.8 6 0.4



progress after 2 weeks of exposure to PBS [Figure 4(e)]thus leads to a decrease in Young’s modulus. The mechani-cal properties of our wound dressings are superior to thosereported before and remain good even after 3 weeks of deg-radation (Young’s modulus of 69 MPA, maximal stress 24MPa and maximal strain 61%), as demonstrated in Figure 5.

In vitro drug-release studiesThe incorporation of broad-spectrum antibiotics such as cef-tazidime in wound dressings can help reduce the bio-bur-den in the wound bed and thus prevent infection andaccelerate wound healing. The controlled release of thisantibiotic may help prevent the occurrence of complica-tions associated with conventional topical treatments. Thelocal antibiotic release profile should potentially exhibit aconsiderable initial burst release rate to respond to theelevated risk of infection from bacteria introduced duringthe initial trauma, followed by a release of antibiotics atan effective level long enough to inhibit latent infection.With characteristic healing reported to take 3–7 weeks(depending on the location, size and degree of injury, aswell as the rate of tissue regeneration), our aim would beto maintain effective antibiotic levels for at least 2–3weeks.

The main challenge in designing a device for the releaseof low molecular weight hydrophilic antibiotics is to over-come the rapid discharge of the drug from the device. Com-mon strategies that have been described in an attempt toovercome the problem of rapid drug release include entrap-ment of the hydrophilic drug within a hydrophobic sub-stance as a means for delaying water penetration24 and out-ward drug diffusion or enhancement of drug bonding to thecarrying matrix.25 These strategies are usually contrary tobasic requirements from wound dressings such as porosityand hydrophilicity, desired to provide adequate gaseousexchange and absorption of wound exudates. A delicate bal-ance between the two must clearly be achieved.

The advantage of the freeze-drying of inverted emul-sions technique used in our study is that the drug is incor-porated within a porous structure during the fabricationprocess, to obtain its release in a controlled desired manner.Our results show that the highly porous (63%) matrixbased on the BSA1 formulation (with an O–A phase ratio of6:1) exhibited a very high burst release of antibiotics andtotal release of the encapsulated drug within 24 h. Lowerporosities, achieved by employing emulsions with a higher12:1 O:A phase ratio, reduced the burst release and pro-longed the antibiotic’s release (Figure 6). Dressings basedon the BSA2 and SPAN formulations displayed a burstrelease of 57 and 6% and overall release spans of 5 and 10days, respectively. Replacing the hydrophilic surfactant albu-min with the hydrophobic Span 80 resulted not only in aconsiderable 10-fold decrease in the antibiotic’s burstrelease but also in a twofold increase in total elution span.A finer microstructure with thicker polymer walls betweenpores thus enables slower diffusion of the hydrophilic anti-biotic molecules to the surrounding. The incorporation of asurfactant may contribute to more than just a stabilizing

(microstructural) effect or serve as a chemical moderator.By binding to various drugs through specific interactions,we have demonstrated that albumin, a predominant drug-binding protein in the body may reduce the burst releaseof antibiotics and contribute to a more moderate releaserate overall, making it an attractive surfactant with a dualrole.10

SUMMARY AND CONCLUSIONS

Novel biodegradable occlusive wound dressings based on apolyglyconate mesh and a porous PDLGA binding matrixwere developed and studied. These composite dressingswere prepared by dip-coating woven meshes in invertedemulsions, followed by their freeze-drying. Their investiga-tion focused on the microstructure, mechanical and physicalproperties, and the release profile of the antibiotic drug cef-tazidime from the binding matrix.

The physical properties of the wound dressing (waterabsorbance and WVTR) as well as drug release profile canbe controlled by changes in the inverted emulsion’s formu-lation. An increase in the emulsion’s O:A phase ratioresulted in decreased porosity of the matrix, and thus in alower WVTR and some reduction in the initial water uptake.It also reduced the burst release of antibiotics from the ma-trix. The surfactants used for stabilizing the inverted emul-sion significantly affected the microstructure and the result-ing physical properties. Specimens containing albumin assurfactant exhibited a relatively high burst release and lowWVTR at a 12:1 O:A phase ratio, whereas specimens with asimilar O:A ratio containing Span 80 produced an optimalWVTR of 2641 6 42 g m�2 day�1 and a very low burstrelease followed by a moderate and constant release ratefor 7 days.

The initial mechanical properties of our wound-dressingstructures were found to be superior in combining resist-ance to tear (tensile strength of �24 MPa) and ductility(elongation at break of �55%). Three weeks of in vitroincubation in PBS did not affect these properties, and we,therefore, conclude that adequate strength retention ispromised during the anticipated duration of usage as awound dressing.

Our novel composite structures combine good mechani-cal properties with desired physical properties and con-trolled release of the antibiotic drug ceftazidime from thebinding matrix and are therefore potentially very useful asburn and ulcer dressings. Changing the emulsion formula-tion enables adapting the desired properties to the woundcharacteristics and can thus enhance wound healing.

ACKNOWLEDGMENTS

The authors would like to thank Dr. Zahava Barkai (Center ofMaterials, Tel-Aviv University’s Faculty of Exact Sciences)for her generous assistance with Environmental SEMobservations.

REFERENCES1. Field FK, Kerstein MD. Overview of wound healing in a moist

environment. Am J Surg 1994;167:2S–6S.


2. Lamke LO, Nilsson GE, Reithner HL. The evaporative water loss

from burns and the water-vapour permeability of grafts and artifi-

cial membranes used in the treatment of burns. Burns 1977;3:

159–165.

3. Boateng JS, Matthews KH, Stevens HN, Eccleston GM. Wound

healing dressings and drug delivery systems: a review. J Pharm

Sci 2008;97:2892–2923.

4. RevathiG,Puri J, JainBK.Bacteriologyof burns.Burns1998;24:347–349.

5. Harrison-Balestra C, Cazzaniga AL, Davis SC, Mertz PM. A wound-

isolated Pseudomonas aeruginosa grows a biofilm in vitro within

10 hours and is visualized by light microscopy. Dermatol Surg

2003;29:631–635.

6. Pruitt BA Jr, Levine NS. Characteristics and uses of biologic

dressings and skin substitutes. Arch Surg 1984;119:312–322.

7. Chung LY, Schmidt RJ, Hamlyn PF, Sagar BF, Andrews AM,

Turner TD. Biocompatibility of potential wound management

products: Fungal mycelia as a source of chitin/chitosan and their

effect on the proliferation of human F1000 fibroblasts in culture.

J Biomed Mater Res 1994;28:463–469.

8. Ruszczak Z, Friess W. Collagen as a carrier for on-site delivery of

antibacterial drugs. Adv Drug Deliv Rev 2003;55:1679–1698.

9. Galdbart JO, Branger C, Andreassian B, Lambert-Zechovsky N,

Kitzis M. Elution of six antibiotics bonded to polyethylene vascu-

lar grafts sealed with three proteins. J Surg Res 1996;66:174–178.

10. Elsner JJ, Zilberman M. Antibiotic-eluting bioresorbable composite

fibers for wound healing applications: Microstructure, drug delivery

andmechanical properties. Acta Biomater 2009;5:2872–83.

11. Aviv M, Berdicevsky I, Zilberman M. Gentamicin-loaded biore-

sorbable films for prevention of bacterial infections associated

with orthopedic implants. J Biomed Mater Res A 2007;83:10–19.

12. Zilberman M. Novel composite fiber structures to provide drug/

protein delivery for medical implants and tissue regeneration.

Acta Biomater 2007;3:51–57.

13. Queen D, Gaylor JD, Evans JH, Courtney JM, Reid WH. The pre-

clinical evaluation of the water vapour transmission rate through

burn wound dressings. Biomaterials 1987;8:367–371.

14. Wu P, Fisher AC, Foo PP, Queen D, Gaylor JD. In vitro assessment

of water vapour transmission of synthetic wound dressings. Bio-

materials 1995;16:171–175.

15. Mi FL, Wu YB, Shyu SS, Schoung JY, Huang YB, Tsai YH, Hao

JY. Control of wound infections using a bilayer chitosan wound

dressing with sustainable antibiotic delivery. J Biomed Mater Res

2002;59:438–449.

16. Lamke LO. The influence of different ‘‘skin grafts’’ on the evapora-

tive water loss from burns. Scand J Plast Reconstr Surg 1971;5:

82–86.

17. Zong X, Ran S, Kim KS, Fang D, Hsiao BS, Chu B. Structure and

morphology changes during in vitro degradation of electrospun

poly(glycolide-co-lactide) nanofiber membrane. Biomacromole-

cules 2003;4:416–423.

18. Schunck M, Neumann C, Proksch E. Artificial barrier repair in

wounds by semi-occlusive foils reduced wound contraction and

enhanced cell migration and reepithelization in mouse skin.

J Invest Dermatol 2005;125:1063–1071.

19. Lee J, Tae G, Kim YH, Park IS, Kim SH, Kim SH. The effect of gel-

atin incorporation into electrospun poly(L-lactide-co-epsilon-cap-

rolactone) fibers on mechanical properties and cytocompatibility.

Biomaterials 2008;29:1872–1879.

20. Rho KS, Jeong L, Lee G, Seo BM, Park YJ, Hong SD, Roh S, Cho

JJ, Park WH, Min BM. Electrospinning of collagen nanofibers:

effects on the behavior of normal human keratinocytes and early-

stage wound healing. Biomaterials 2006;27:1452–1461.

21. Milella E, Ramires PA, Brescia E, La Sala G, Di Paola L, Bruno V.

Physicochemical, mechanical, and biological properties of com-

mercial membranes for GTR. J Biomed Mater Res 2001;58:

427–435.

22. Chiu CT, Lee JS, Chu CS, Chang YP, Wang YJ. Development of

two alginate-based wound dressings. J Mater Sci Mater Med

2008;19:2503–2513.

23. Lee SB, Kim YH, Chong MS, Hong SH, Lee YM. Study of gelatin-

containing artificial skin V: Fabrication of gelatin scaffolds using a

salt-leaching method. Biomaterials 2005;26:1961–1968.

24. Krasko MY, Golenser J, Nyska A, Nyska M, Brin YS, Domb AJ.

Gentamicin extended release from an injectable polymeric

implant. J Control Release 2007;117:90–96.

25. Sripriya R, Kumar MS, Sehgal PK. Improved collagen bilayer

dressing for the controlled release of drugs. J Biomed Mater Res

B Appl Biomater 2004;70:389–396.



novel biodegradable composite wound dressings with

Documents