Pediatric Critical Care

Longitudinal assessment of Integra in primary burn management:A randomized pediatric clinical trial*

Ludwik K. Branski, MD; David N. Herndon, MD; Clifford Pereira, MD, FRCS (Eng); Ronald P. Mlcak, PhD;Mario M. Celis, MS; Jong O. Lee, MD; Arthur P. Sanford, MD; William B. Norbury, MD;Xiao-Jun Zhang, PhD; Marc G. Jeschke, MD, PhD

During the past two decades,burn management hasevolved to the current stan-dard of care that includes

early excision of necrotic tissue andprompt coverage. This attenuates thepostburn hypermetabolic response, de-creases fluid loss, and ultimately im-

proves survival (1–3). However, earlywound closure using autograft is difficultwhen full-thickness burns exceed 40% to-tal body surface area (TBSA). Until themid-1970s, the reduced availability of na-tive skin was the limiting factor for earlyexcision of burn wounds. Allografts (ca-daver skin) served as skin substitute inseverely burned patients, and later asandwich technique was developed usingautografts with an allograft overlay (4).While this approach is still commonlyused in burn centers throughout theworld, it bears considerable risks for thepatients:

● Cadaver skin-associated factorssuch as antigenicity, cross-infec-tion, and limited availability andshelf life (5);

● Further increase of a large burnwound to near-TBSA involvement byharvesting autograft during the vul-nerable early phase of burn injury;

● Fragility of split skin grafts on full-thickness wounds that are affectedby the total loss of dermis, and sus-ceptibility to contractures and hy-pertrophic scarring (6).

The ideal repair of full-thicknessburns would therefore require replace-ment of both the dermal and the epider-mal components of skin.

Integra (Integra LifeSciences, Plains-boro, NJ), developed by Dr. Burke andcolleagues (7), consists of a 2-mm-thickinner (dermal replacement) layer made ofa porous matrix of bovine collagen andthe glycosaminoglycan chrondroitin-6-sulfate (8). This layer has a 70-�m to200-�m pore size that facilitates fibrovas-cular ingrowth, after which the originalmatrix biodegrades. The outer layer (neo-epidermis) of the membrane is a 0.23-mm-thick polysiloxane polymer with va-por transmission characteristics similarto normal epithelium. It prevents water

*See also p. 2661.From the Shriners Hospitals for Children,

Galveston, TX, and Department of Surgery, Universityof Texas Medical Branch, Galveston, TX.

Supported, in part, by Johnson & Johnson IntegraLifeSciences, the American Surgical Association Foun-dation, and the National Center for Research Re-sources, National Institutes of Health.

LKB and DNH are sharing credit as first author.The authors have not disclosed any potential con-

flicts of interest.For information regarding this article, E-mail:

majeschk@utmb.eduCopyright © 2007 by the Society of Critical Care

Medicine and Lippincott Williams & Wilkins

DOI: 10.1097/01.CCM.0000285991.36698.E2

Background: Early excision with autograft-allograft closure isstandard in severe burn management. Cadaver skin is associatedwith risks such as antigenicity, infection, and limited availabilityand shelf life. Previous studies have shown that Integra is safe touse in burns of <20% total body surface area. However, thesuitability of its use in large burns (>50% total body surfacearea), its effects on postburn hypermetabolism, and the long-termcosmetic and functional results have not yet been evaluated.

Materials and Methods: Twenty children with an average burnsize of 73 � 15% total body surface area (71 � 15% full-thickness burn) were randomized to be treated with either Integraor with autograft-allograft technique. Outcome measures such aslength of hospital stay, mortality, incidence of infection andsepsis, acute phase protein levels, and muscle fractional syn-thetic rate were compared between and within groups during theacute stay (admission to discharge). Outcome measures such asresting energy expenditure, body composition data (measured bydual-energy radiograph absorptiometry), cardiac function in-dexes, and number of reconstructive procedures were com-pared during acute hospital stay and at long-term follow-up

(up to 2 yrs postinjury). Scar evaluation was performed atlong-term follow-up.

Results: There were no significant differences between Integraand controls in burn size (70 � 5% vs. 74 � 4% total body surfacearea), mortality (40% vs. 30%), and length of stay (41 � 4 vs.39 � 4 days). In the short term, resting energy expendituresignificantly decreased (p < .01), and serum levels of constitutiveproteins significantly increased (p < .03) in the Integra groupcompared with controls. Long-term follow-up revealed a signifi-cant increase in bone mineral content and density (24 monthspostburn, p < .05), as well as improved scarring in terms ofheight, thickness, vascularity, and pigmentation (12 months and18–24 months, p < .01) in the Integra group.

Conclusion: Integra can be used for immediate wound cover-age in children with severe burns without the associated risks ofcadaver skin. (Crit Care Med 2007; 35:2615–2623)

KEY WORDS: Integra; burns; severe burns; children; reconstruc-tion; dual-energy radiograph absorptiometry; resting energy ex-penditure; autograft; allograft

2615Crit Care Med 2007 Vol. 35, No. 11

loss and protects the underlying dermisfrom mechanical trauma and bacterial in-vasion. Integra serves as a matrix for theinfiltration of fibroblasts, macrophages,lymphocytes, and capillaries derived fromthe wound bed. This promotes rapid neo-dermis formation. Once the tissue is ad-equately vascularized, the outer siliconemembrane is replaced with a thin epithe-lial autograft. In adults, Integra has beenused successfully in immediate and de-layed closure of major burns, with reduc-tion in length of hospital stay, a favorablecosmesis, and improved functional out-come (4, 7, 9, 10).

We conducted a randomized clinicaltrial with the use of Integra in the man-agement of severe full-thickness burns of�50% TBSA in a pediatric patient popu-lation, and compared it to standard au-tograft-allograft technique. We hypothe-sized that early excision of burned skinand prompt closure with Integra wouldattenuate the hypermetabolic responseand improve clinical outcome duringacute hospitalization. We further hypoth-esized that the dermal regeneration af-forded by Integra would improve long-term cosmetic and functional outcomes.

METHODS

Severely burned children admitted to theShriners Hospitals for Children in Galveston,TX, between November 2001 and March 2003were considered for this study. Permission forconducting the study was obtained from theInstitutional Review Board at the University ofTexas Medical Branch of Galveston, TX. Thefollowing inclusion criteria were used:

● Burn size �50% TBSA and �40% TBSAfull-thickness burn;

● Patients admitted within 72 hrs of in-jury;

● Patients not septic at admission.

Informed written consent was obtainedin all cases, and patients were randomizedfor grafting with either Integra or standardautograft-allograft technique. First surgerywas performed within 24 – 48 hrs of admis-sion. Unmeshed Integra sheets were care-

fully applied to the wounds after tangentialexcision of burned tissue. Care was taken toappose sheet lines and prevent tenting orwrinkling of the material. In cases whereIntegra was placed over joints (mainly el-bows and knees), patients were immobilizedin splints. Silver nitrate and sulfamylonsoaks were used on top of the Integra sheets,and Acticoat was applied to the seams, toprevent development of infections under-neath the Silastic layer. Once adequatelyvascularized, usually after 14 –21 days, theSilastic epidermis was stripped and replacedwith thin (0.05– 0.13 mm) epidermal au-tograft. In the control group, expanded au-tograft (meshed 1:4) with allograft overlaywas applied to as much area as it was possible tocover. The rest of the wound area was coveredwith unexpanded fresh allograft (meshed 1:1.5).Donor sites were recropped when healed, andallograft surgically excised and consecutively re-placed with autograft skin.

Table 1. Sepsis score based upon modified Amer-ican Academy of Chest Physicians/Society of Crit-ical Care Medicine definitions

Variable Score

1. T �38.5 or �36.5°C 12. HR �20% above NL for age 13. RR �20% above NL for age 14. WBC �12,000 or �4000 15. Bacteremia or fungemia or invasion

tissue infection (�105/g)1

T, temperature; HR, heart rate; NL, normal;RR, respiratory rate; WBC, white blood cellcount.

Sepsis defined as variable 5 and at least two ofvariables 1–4.

Table 2. Patient demographics

Treatment Groups Integra Control

Sample size 10 10Survivors 6 7Sex, male 7 9Age, mean in yrs 7.4 6.2Inhalation injury 4 6TBSA, % 70 � 5 74 � 43rd degree burns, % 65 � 6 70 � 3

TBSA, total body surface area.Percentile data presented as means � SD.

Figure 1. Serial cardiac measurements at admission, discharge, and postburn months 6, 9, 12, 18, and24. Data expressed as means � SEM. A, cardiac index (L/min/m2). B, heart rate, expressed as percentof predicted. No significant changes between the groups.

2616 Crit Care Med 2007 Vol. 35, No. 11

Demographics. Mortality rates, length ofhospital stay, cumulative hospital stay �1 yr,total number of procedures performed duringacute admission, and total operating room(OR) time until 95% healing were recorded ineach group. Weights were measured approxi-mately 5 days after admission and at dischargeusing standard clinical sling scales. Subse-quent measurements after discharge atmonths 6, 9, 12, 18, and 24 were measuredusing standing scales. The clinical scales werecalibrated monthly.

Cardiac Function and Liver Size. Cardiacfunction measurements included heart rate,stroke volume, cardiac index, and cardiac out-put. All ultrasound measurements were madewith the Sonos 100 CF echocardiogram(Hewlett Packard Imaging Systems, Andover,MA) with a 3.5-MHz transducer. Recordingswere performed with the subjects in a supineposition and breathing freely. M-mode trac-ings were obtained at the level of the tips ofthe mitral leaflets in the parasternal long-axisposition and measurements were performedaccording to the American Society of Echocar-diography recommendations (11). Left ven-tricular volumes determined at end diastoleand end systole were used to calculate strokevolume, cardiac output, and cardiac index.Three measurements were performed and av-eraged for data analysis. Liver sizes were mea-sured with standard ultrasound equipment us-ing a 3.5-MHz transducer. Data were recordedat admission, on discharge, and at postburnmonths 6, 12, 18, and 24.

Metabolism. Measurements of the hyper-metabolic response (resting energy expendi-ture) were performed by indirect calorimetryusing a SensorMedics (Yorba Linda, CA) 2900metabolic cart. All indirect calorimetric mea-surements were made at 30°C. Composition ofinspired and expired gases were sampled andanalyzed at 60-sec intervals. Values of VCO2,VO2, respiratory quotient, and resting energyexpenditure were accepted when at steadystate for 5 mins. The average resting energyexpenditure was calculated from these steady-state measurements. For statistical compari-sons, energy expenditure was expressed as aquotient of resting energy expenditure and thebasal metabolic rate, using the Harris-Benedict equation (12) to find a percentage ofpredicted resting energy expenditure(PPREE). The measurements were performedat five time points during the acute hospitalstay (admission; 7–10, 13–15, and 20–22 dayspostburn; discharge) and at postburn months6, 9, 12, 18, and 24.

Muscle Fractional Synthetic Rate. The de-gree of protein catabolism was quantified us-ing stable isotope tracers. Protein kineticstudies were performed beginning between 5am and 7 am, 5 days to 7 days (first study) and12 days to 14 days (second study) after the firstexcision and grafting procedure. All stable iso-tope studies consisted of a 5-hr infusion of2H5-phenylalanine, as previously described(13). Vastus lateralis muscle biopsies were

taken from the study leg at 2 hrs and 5 hrswhile the subject was under intravenous con-scious sedation. Samples were immediatelywashed with normal saline to remove anyblood, blotted dry, and snap-frozen in liquidnitrogen for storage at �70°C. Muscle tissuesamples were ground, and intracellular-freeamino acids and muscle proteins were ex-tracted as previously described (13). Muscleintracellular free enrichment of phenylalaninewas determined by gas chromatography-massspectrometry. Mixed muscle protein-boundphenylalanine enrichment was analyzed by gaschromatography-mass spectrometry after pro-tein hydrolysis and amino acid extraction (13).We calculated the fractional synthetic rate ofmixed muscle proteins by measuring the in-corporation rate of the phenylalanine tracerinto the proteins and using the precursor-product model to calculate the synthesis rate:

fractional synthetic rate � (�Ep/t)/(EM(1)

� EM(2)/2) 60 100 [1]

where �p is the increment in protein-boundphenylalanine enrichment between two se-

quential biopsies, t is the time between thetwo sequential biopsies, and EM(1) � EM(2) arethe phenylalanine enrichments in the free in-tracellular pool in the two sequential biopsies(13). Data are expressed as percent per hour.

Body Composition. Total weight, lean bodymass, bone mineral content, and bone mineraldensity were measured by dual-energy radio-graph absorptiometry. A Hologic (Waltham,MA) QDR-4500A absorptiometer was used forthese studies. To minimize systematic devia-tions, the system was calibrated daily against aspinal phantom in the anteroposterior, lateral,and single-beam modes. Individual pixels werecalibrated against a tissue bar phantom to deter-mine whether the pixel was reading bone, fat,lean tissue, or air. All studies were performedafter feedings and intravenous fluids were dis-continued to minimize exogenous potassiumcontamination. Measurements were performedat admission, on discharge, and at postburnmonths 6, 12, 18, and 24, and data expressedas percent change since admission.

Serum Proteins. Levels of serum constitu-tive hepatic proteins (such as transferrin, pre-

Figure 2. Percentage of predicted resting energy expenditure (PPREE, calculated using Harris-Benedict equation). Data expressed as means � SEM. A, acute hospitalization period with measure-ments at admission; postburn wks 1, 2, and 3; and discharge. B, long-term measurements at dischargeand postburn months 6, 9, 12, 18, and 24. *Levels of PPREE were significantly lower at 3 wks postburnand discharge as compared with controls.

2617Crit Care Med 2007 Vol. 35, No. 11

albumin, and retinol binding protein), type Iserum acute phase proteins (such as �1-acidglycoprotein and C-reactive protein), type IIserum acute phase proteins (such as �2-macroglobulin, haptoglobin, and triglyceride)were measured on admission and 7 days to 10days post-treatment using a nephelometer(Dade Behring, Deerfield, IL) with N-antiserato different serum proteins (Behring Diagnos-tics, Westwood, MA).

Sepsis Scores and Infection. Incidence ofsepsis in the acute phase was recorded as per themodified American Academy of Chest Physi-cians/Society of Critical Care Medicine sepsisscores for burns (14). Invasive wound infectionwas defined by quantitative cultures of �105

organisms per gram of tissue, wound biopsiesbeing taken on clinical suspicion (Table 1).

Cosmetic Outcome. Burn scars were as-sessed by four blinded volunteer clinicians us-ing the Hamilton burn-scar rating system (15,16). Briefly, this scoring system is based onthe evaluation of scar height/thickness, irreg-ularity, vascularity, and pigmentation/color,with a maximum of 14 points given for allcategories. The evaluators were blinded to theidentity and treatment group of the patientswhose photographs were assessed, and nonewas involved in the treatment of these pa-tients. Each observer analyzed all photographsof representative scars on the chest, back, up-per thigh, and arms, at 12 months and at18–24 months postinjury.

Functional Outcome and Cost. Time tofirst reconstructive procedure, cumulative re-constructive procedures required during 2 yrs,and the cumulative OR time required for theseprocedures were used as indicators of func-tional outcome as well as long-term cost-efficiency indexes. Long-term functional out-come was measured as a function of thenumber of reconstructive procedures requiredby the survivors during 2 yrs postinjury.

Statistics. Body composition and metabolicand cardiac function measurements were ana-lyzed using a two-way analysis of variance forrepeated measures. If significant differenceswere found between groups, posthoc multiplecomparisons (Bonferroni method and Student’st-test) were performed to compare differencesbetween groups and within groups. Differencesbetween serum protein measurements and mus-cle fractional synthetic rates were analyzed usingpaired Student’s t-tests. Differences betweenHamilton scar scores were examined with un-paired Student’s t-tests after normality was con-firmed (Kolmogorov-Smirnov test). The rangeof 0 to 14 with intervals of 0.25 was determinedas large enough for the use of a parametric test. Inall cases, significance was accepted at p � .05.Statistical analysis was carried out using SigmaStatfor Windows version 2.03 (SPSS, Chicago, IL).

RESULTS

Between November 2001 and March2003, 20 acutely burned children were en-rolled into the study. Demographics of the

two groups are listed in Table 2 and demon-strate comparable groups in terms of age,gender, extent of injury, and mortality rates.

Cardiac Function and Liver Size. Wefound no differences in cardiac functionbetween the groups and within thegroups when examining short-term andlong-term results. This indicates thatboth groups were physiologically similarbefore treatment and that Integra did notadversely affect cardiorespiratory func-tion. Cardiac index (Fig. 1A) was elevatedin both groups and remained elevated forup to 24 months postinjury; heart rate(Fig. 1B) gradually declined in bothgroups, but remained elevated comparedwith normal values. Liver size increased,but remained similar in both groups(data not shown).

Metabolism. Early excision followedby coverage with sterile Integra changed

the postburn hypermetabolic response.Metabolic rate measured as PPREE waselevated in both groups within the first 7days as compared with the value at ad-mission. While the control group did notshow a tendency to decrease throughoutthe acute stay, patients in the Integragroup showed a significantly loweredPPREE at 3 wks after admission (i.e.,around the time the silicone layer wasreplaced with autograft) and at discharge(Fig. 2A). However, in the long-term eval-uation, no significant difference was seenbetween the groups, because the PPREEnormalized in both Integra and controlpatients (Fig. 2B).

Muscle Fractional Synthetic Rate.The fractional synthetic rate of peripheralmuscle protein showed no significant dif-ferences between the Integra group andthe control group during the first study

Figure 3. Serial body composition measurements at admission, discharge, and postburn months 6, 12,18, and 24. Data expressed as means � SEM. A, bone mineral content and (B) bone mineral density aresignificantly increased in the Integra group at 24 months compared with controls (marked *). Nosignificant differences seen between the groups in lean body mass and weight.

2618 Crit Care Med 2007 Vol. 35, No. 11

(0.09% � 0.02% vs. 0.12% � 0.03%,respectively) and the second study (0.1% �0.4 vs. 0.08 � 0.01, respectively). No differ-ences were found within the groups be-tween the 2 study time points.

Body Composition. Dual-energy ra-diograph absorptiometry measurementsshowed significantly improved bone min-eral content (Fig. 3A) and bone mineraldensity (Fig. 3B) in the Integra group 2yrs postburn. At 18 months postburn,bone mineral content and bone mineraldensity showed a trend toward improve-ment; however, using the rigid two-wayrepeated measures analysis of variancefor statistical analysis, the p value did notreach significance (p � .09 for bone min-eral content and p � .07 for bone mineraldensity). No significant differences werefound between groups in terms of leanbody mass and total body weight. Bothgroups showed the typical decrease seenfor the first year postburn in severelyburned children, and a gradual improve-ment in the second year postburn.

Serum Proteins. Acute phase proteinswere measured at admission and at dis-charge (or when 95% of involved area washealed) as an indicator of the hepaticacute phase response. Complement C-3and haptoglobin were increased signifi-cantly (both, p � .01), indicating agreater acute phase response in the In-tegra group as compared with the controlgroup (Fig. 4). Constitutive proteins suchas prealbumin and transferrin were im-proved significantly in the Integra group(p � .02 and p � .01, respectively), indi-cating improved liver function in thesepatients (Fig. 5). Apolipoprotein B levelswere higher in the Integra group (p �.01), while triglyceride levels were notsignificantly different between treatmentgroups (Fig. 6).

Sepsis Scores and Infection. Averagesepsis scores taken for duration of acutephase were not significantly different be-tween the two groups (p � .82; Fig. 7).Four patients in each group had invasiveburn wound infections, and organismscausing these were both bacterial andfungal (Staphylococcus aureus, Entero-coccus, Aspergillus, and Candida beingthe most common organisms in decreas-ing frequency). In the Integra group,these four patients had 7 � 3% of thetotal covered surface removed because ofthe superficial infection. No differencewas seen in wound infection rates orcausative organism in either group.

Cosmetic Outcome. The Hamiltonburn-scar scoring system revealed an aes-

thetically improved scar in the Integragroup at both 12 months and 18 –24months postinjury. There was also a sig-nificant improvement in scarring over

time (12–24 months) within the Integragroup (Fig. 8 and Table 3).

Functional Outcome and Cost. Short-term cost-efficiency indicators such as

Figure 4. Acute phase proteins in both groups at admission and discharge. Data expressed as means � SEM.A, complement C-3 levels increased significantly at discharge in the Integra group as compared withcontrols. *p � .01 Integra vs. control. B, haptoglobin levels increased significantly at discharge in theIntegra group as compared with controls. *p � .01 Integra vs. control.

Figure 5. Constitutive protein levels in both groups at admission and discharge. Data expressed asmeans � SEM. A, prealbumin levels increased significantly at discharge in the Integra group ascompared with controls. *p � .02 Integra vs. control. B, transferrin levels increased significantly atdischarge in the Integra group as compared with controls. *p � .01 Integra vs. control.

2619Crit Care Med 2007 Vol. 35, No. 11

length of hospital stay, cumulative hospitalstay during 1 yr, total number of proce-dures performed during acute admission,and OR time were used to determine out-comes in both groups. Time to first recon-structive procedure, cumulative recon-structive procedures required during 2 yrs,and the cumulative OR time required forthese procedures were used as indicators offunctional outcomes and long-term cost-efficiency indexes. No differences werenoted between treatment groups in anyof these parameters at 2 yrs postburn(Table 4).

DISCUSSIONEarly and complete excision of burned

areas along with prompt coverage is an

integral part of burn management (1–3).Immediate wound coverage with nativeskin is only possible when sufficient do-nor area is available to cover the excisedareas. Integra is proven to be a safe skinsubstitute for immediate closure of majorburns in adults (4, 7, 9, 10). Its use inseverely burned children has not beenstudied, except for a case report pub-lished by Dr. King (17), where Integrawas shown to reduce nutritional require-ments in an 11-yr-old boy with 60%TBSA burns. We assessed the use of In-tegra in comparison to standard au-tograft-allograft technique for differencesin physiologic and metabolic responses totreatment by measuring various indexesof cardiorespiratory function, liver func-

tion, and liver response (in terms of acutephase proteins), as well as metabolic rateand the translation of it (in terms ofchanges in muscle protein synthesis,preservation of lean body mass, and dif-ferences in bone composition).

Demographics of both groups (Table1) indicate adequate randomization andcomparability between groups pretreat-ment. Of the ten patients in each group,three patients died in the control groupand four in the Integra group. One of thesurvivors in the Integra group subse-quently required cultured epithelial au-tograft because of paucity of donor sites,owing to 97% TBSA involvement, andwas excluded from final analysis. Themortality rate seen in these patients (30%to 40%) was much higher than the ex-pected mortality rate for burns of similarseverity at our unit. This is due to theunfortunate fact that our unit was expe-riencing a spate of multiresistant infec-tions during this period. Interestingly, al-though earlier reports (7, 18) didhighlight problems with an increased riskof infection and sepsis with the use ofIntegra, we did not experience any differ-ence in wound infection rates betweenthe Integra and control patients. Nor wasany difference noted in causative organ-isms cultured from wound biopsies in thecases that did get infected. Sepsis scores,averaged for the entire acute hospitaliza-tion period, did not reveal any differ-ences, either. This is consistent with therecent multicenter postapproval clinicaltrial wherein incidence of invasive infec-tion at Integra-treated sites was only3.1% with superficial infection rates be-ing only 13.2% (19). This reduction ininfection rate, we believe, is due to im-proved technique and experience with theuse of Integra. Also, because septic pa-tients have an additional 40% increase inmetabolic rate and protein catabolism,relative to nonseptic burn patients ofcomparable size, the prevention of sero-mas, hematomas, and infections is an im-portant clinical consideration while usingIntegra (20).

Early total excision of large burnwounds (�50% TBSA) and coverage withautograft and/or cadaver skin within 2–3days of injury reduces the metabolic rateby 40%, compared with a burn that is notcovered until 1 wk postinjury (20). Con-sidering that covering the excised areaswith Integra did not require simulta-neous autografting, thus allowing the pa-tient to tide over the initial injury with-out the additional insult of donor site

Figure 6. Apolipoprotein B and triglyceride levels in both groups at admission and discharge. Dataexpressed as means � SEM. A, apolipoprotein B levels increased significantly at discharge in the Integragroup as compared with controls. *p � .01 Integra vs. control. B, triglyceride levels did not differsignificantly at admission or discharge in either group.

Figure 7. Sepsis scores averaged during acute hospitalization period. Scores based upon the modifiedAmerican Academy of Chest Physicians/Society of Critical Care Medicine definitions. None of the timepoints showed statistically significant differences between the groups.

2620 Crit Care Med 2007 Vol. 35, No. 11

wounds, we assumed that patients woulddo better clinically. We also envisagedthat Integra would attenuate the post-burn hypermetabolic response in severeburns. According to our expectations, wesaw an improvement in PPREE in theshort term. At approximately 3 wks post-burn, PPREE—normalized for age, gen-der, height, and weight—dropped signif-icantly in the Integra group, reachingnormal levels and staying at this leveluntil discharge. This finding suggests In-tegra provides a great benefit in severelyburned patients. The results of the con-trol group, where PPREE at dischargeexceeded a mean of 150% of normal val-

ues, are consistent with previous findingsfrom our institute, where, in a similarburned population, the PPREE values re-mained elevated at discharge (21). How-ever, this positive finding was not sup-ported by improved cardiorespiratoryfunction indexes, reduced liver size,weight gain, or increase in muscle mass(as measured by dual-energy radiographabsorptiometry and muscle protein ki-netic results) during the acute hospital-ization. This is still positive, because itproves that Integra can be used in �95%TBSA burns, where donor site paucity is areal problem and expeditious woundhealing a requirement. The fact that In-tegra did not adversely affect cardiorespi-ratory physiology adds further evidenceto its safety in large burns in children.

In the long term, Integra significantlyimproved bone mineral content and den-sity compared with controls at 24 monthspostburn, showing a clear trend towardthis by 18 months postburn. These re-sults suggest that the attenuation of hy-permetabolism in the short term mightprove beneficial for long-term improve-

ments in body composition. The positivefindings of bone mineral content anddensity increase, however, cannot bematched by significant differences in leanbody mass or weight between the twotreatment groups. Although the Integrapatients showed a trend toward higherlean body mass and weight 18 monthsand 24 months postburn, differences be-tween groups did not reach significance.

Interestingly, the use of Integra signif-icantly increased prealbumin (a constitu-tive protein), C-3 complement and hap-toglobin (both type I and II acute phaseproteins), and apolipoprotein B (a mem-ber of the free fatty acid transport mech-anism). Fatty acid infiltration is a com-mon finding postburn. It is importantowing to its hepatic accumulation, fur-ther compromising liver function. In-creased peripheral lipolysis, due to a lackof transporter proteins such as low-density and high-density lipoproteins andtheir apolipoproteins A and B, has beendiscussed as a possible mechanism (22).We have recently shown that the hyper-metabolic response of the liver in severelyburned pediatric patients, despite ade-quate nutritional support, is prolonged toup to 80 days postinjury (23). Levels ofconstitutive hepatic proteins (such as se-rum prealbumin, transferrin, and retinolbinding proteins) were shown to declinebelow normal immediately postburn andremained low up to 80 days after theinjury. This decline was most likely dueto decreased hepatic synthesis and ex-pression. Wound coverage with Integraseems to attenuate postburn inducedliver dysfunction with improved prealbu-min and transferrin levels, while main-taining retinol binding protein levels tolevels seen in patients treated with thestandard autograft-allograft technique.

Increased hepatic acute phase proteinsin response to severe injury are assumedto serve important functions in restoringhomeostasis. Overproduction of these, es-pecially after a severe burn, may sustaininflammatory reactions, including hyper-metabolism, catabolism, and multiple or-gan failure (24). The effect of the increasein C-3 complement and haptoglobinwithin the Integra group is unclear, con-sidering that it did not adversely affectthe PPREE, lean body mass, or cardiac orrespiratory function.

Past studies have shown that the his-tologic structure and physical propertiesof the skin at sites of Integra-supportedautograft are very similar to those of nor-mal skin. Dermal remodeling occurred

Figure 8. Long-term results with Integra vs. autograft-allograft technique. (A) and (C) show pathologicscarring with keloids and hypertrophy in control patients. (B) and (D) show good clinical results inIntegra patients’ corresponding areas. All photographs taken 18–24 months after injury.

Table 3. Hamilton burn scar scores at 12 monthsand 18–24 months after injury

Integra Control p Value

12 months 5.4 � 1.7 7.7 � 2.6 .00318–24 months 4.3 � 2.2 6.6 � 3.1 .02p value .035 .403

Data presented as means � SD; p values sig-nificant at �.05.

2621Crit Care Med 2007 Vol. 35, No. 11

within a month’s time with the appear-ance of a structure resembling papillaryand reticular dermis. Also, most of theoriginal layer of collagen and glycosami-noglycan undergoes biodegradation in3–4 wks (25). In a patient series spanning10 yrs, no significant hypertrophic scarwas noted in patients treated with Integraand 93% of patients had no or minimalhypertrophic scars. Areas grafted with In-tegra in children grew with the childrenand were cosmetically superior to areastreated with autograft alone (26). Integrahas the advantage of requiring little ad-ditional resurfacing or reconstruction.Consistent with these findings, we dem-onstrated improved photographic scarscores at both 12 months and 18–24months post-treatment in the Integra-treated patients. Moreover, Integra-treated areas improved significantly overtime between 12 and 24 months.

Because required donor grafts in theIntegra group were thinner than in thecontrols, we envisaged a quicker healingof donor sites translating to decreased ORvisits and decreased length of stay (27).This was not the case in our study. Nodifferences were demonstrable in lengthof acute stay, cumulative hospital stayduring 12 months, and OR time. Wethink this is because Integra-treated pa-tients require additional clinical surveil-lance, wound checks, and a more de-manding postoperative nursing care andrehabilitation than autografted patients.Integra patients also still required fullcoverage with autograft after the first 3wks.

Improved scarring did not translate todecreased reconstructive procedures orOR time spent on reconstruction for thefirst 2 yrs postinjury. In our patient pop-

ulation, most of the reconstructive pro-cedures were for contracture releasesover joints of the hand, elbows, and axil-lae; the mode of primary skin coveragedid not influence the need for and fre-quency of these procedures. However,multicenter trials have proven the use ofIntegra in secondary reconstruction andcontracture releases especially over joints(28, 29). This success is dependent uponcomplete fixation and prevention of sheerforces at the wound-matrix interface.

Limitations of the Study. The data ofthis study need to be interpreted withcaution, considering the small samplesize. There are disadvantages to the use ofIntegra compared with cadaver skin. Cov-erage of large wound areas with Integra isvery expensive and therefore not afford-able by many healthcare providers in de-veloping countries. The handling of In-tegra needs to be practiced, even byexperienced surgeons, and the learningcurve to maximize biointegration is re-portedly steep, with high failure rates ini-tially. The silicone layer does not drapeand bend very well and is prone to del-amination due to shearing over joints.However, in experienced hands, the tech-nique for grafting Integra is relativelyquick and relatively atraumatic whencompared with allograft and xenograftthat have to be surgically excised fromthe wound bed.

CONCLUSIONS

In summary, Integra can be used forimmediate wound coverage in childrenwith severe burns without the associatedrisks of cadaver skin. Its use was notassociated with an increased incidence ofinfection or sepsis in our patient cohort.

Integra was associated with an attenua-tion of postburn hepatic dysfunction andimproved resting energy expenditure inthe short term, and an improved aes-thetic outcome in the long term.

ACKNOWLEDGMENTS

We thank David Chinkes for assistancewith statistical analysis and EileenFigueroa and Steve Schuenke for help inthe preparation of this manuscript.

REFERENCES

1. Herndon DN, Parks DH: Comparison of serialdebridement and autografting and early mas-sive excision with cadaver skin overlay in thetreatment of large burns in children.J Trauma 1986; 26:149–152

2. Janzekovic Z: A new concept in the earlyexcision and immediate grafting of burns.J Trauma 1970; 10:1103–1108

3. Thompson P, Herndon DN, Abston S, et al:Effect of early excision on patients with ma-jor thermal injury. J Trauma 1987; 27:205–207

4. Muller MJ, Herndon DN: Operative woundmanagement. In: Herndon DN (Ed). TotalBurn Care. Second Edition. London, UK,W.B. Saunders, 2001, pp 221–231

5. Blome-Eberwein S, Jester A, Kuentscher M,et al: Clinical practice of glycerol preservedallograft skin coverage. Burns 2002;28(Suppl 1):S10–S12

6. Nanchahal J, Ward CM: New grafts for old? Areview of alternatives to autologous skin.Br J Plast Surg 1992; 45:354–363

7. Burke JF, Yannas IV, Quinby WC Jr, et al:Successful use of a physiologically acceptableartificial skin in the treatment of extensiveburn injury. Ann Surg 1981; 194:413–428

8. Tompkins RG, Burke JF: Progress in burntreatment and the use of artificial skin.World J Surg 1990; 14:819–824

9. Yannas IV, Burke JF, Orgill DP, et al: Woundtissue can utilize a polymeric template tosynthesize a functional extension of skin.Science 1982; 215:174–176

10. Yannas IV, Burke JF, Warpehoski M, et al:Prompt, long-term functional replacementof skin. Trans Am Soc Artif Intern Organs1981; 27:19–23

11. Sahn DJ, DeMaria A, Kisslo J, et al: Recom-mendations regarding quantitation in M-mode echocardiography: Results of a surveyof echocardiographic measurements. Circu-lation 1978; 58:1072–1083

12. Hart DW, Wolf SE, Chinkes DL, et al: Effectsof early excision and aggressive enteral feed-ing on hypermetabolism, catabolism, andsepsis after severe burn. J Trauma 2003; 54:755–764

13. Wolfe RR, Chinkes DL: Isotope Tracers inMetabolic Research: Principles and Practiceof Kinetic Analysis. Second Edition. Hobo-ken, NJ, Wiley-Liss, 2005

Table 4. Overview of short-term and long-term cost-efficiency indicators in treatment groups

Treatment Groups Integra Control p Value

Short-termLength of stay, daysa 32.6 � 15.0 38.2 � 11.1 .49Total operationsb 5.5 � 0.6 5.7 � 2.3 .89First operation time, hrs 3.1 � 0.6 2.8 � 0.9 .54Blood loss, litersc 2.3 � 1.6 2.8 � 0.9 .61Cumulative OR time, hrsd 11.2 � 3.3 13.8 � 4.8 .33

Long-termCumulative stay, dayse 21.8 � 16.4 16.0 � 10.1 .51Reconstructive proceduresf 4.8 � 1.7 4.2 � 3.5 .77Reconstructive OR time, hrs 11.9 � 2.8 11.4 � 6.5 .89

OR, operating room.aDuring acute admission; bto achieve 95% healing of burned area; cestimated, during first

operation; dfor all operations during acute admission to achieve 95% healing of burned area; efromthe time of discharge up to 2 yrs after injury; ftotal up to 2 years after injury. Data presented asmeans � SD.

2622 Crit Care Med 2007 Vol. 35, No. 11

14. Bone RC, Balk RA, Cerra FB, et al: Defini-tions for sepsis and organ failure and guide-lines for the use of innovative therapies insepsis. The ACCP/SCCM Consensus Confer-ence Committee. American College of ChestPhysicians/Society of Critical Care Medicine.Chest 1992; 101:1644–1655

15. Smith GM, Tompkins DM, Bigelow ME, et al:Burn-induced cosmetic disfigurement: Canit be measured reliably? J Burn Care Rehabil1988; 9:371–375

16. Crowe JM, Simpson K, Johnson W, et al:Reliability of photographic analysis in deter-mining change in scar appearance. J BurnCare Rehabil 1998; 19:183–186

17. King P: Artificial skin reduces nutritionalrequirements in a severely burned child.Burns 2000; 26:501–503

18. Clayton MC, Bishop JF: Perioperative andpostoperative dressing techniques for IntegraArtificial Skin: Views from two medical cen-ters. J Burn Care Rehabil 1998; 19:358–363

19. Heimbach DM, Warden GD, Luterman A, et

al: Multicenter postapproval clinical trial ofIntegra dermal regeneration template forburn treatment. J Burn Care Rehabil 2003;24:42–48

20. Hart DW, Wolf SE, Chinkes DL, et al: Deter-minants of skeletal muscle catabolism aftersevere burn. Ann Surg 2000; 232:455–465

21. Przkora R, Barrow RE, Jeschke MG, et al:Body composition changes with time in pe-diatric burn patients. J Trauma 2006; 60:968–971

22. Aarsland A, Chinkes D, Wolfe RR: Contribu-tions of de novo synthesis of fatty acids tototal VLDL-triglyceride secretion during pro-longed hyperglycemia/hyperinsulinemia innormal man. J Clin Invest 1996; 98:2008–2017

23. Jeschke MG, Barrow RE, Herndon DN: Ex-tended hypermetabolic response of the liverin severely burned pediatric patients. ArchSurg 2004; 139:641–647

24. Bone RC: Toward a theory regarding thepathogenesis of the systemic inflammatory

response syndrome: What we do and do notknow about cytokine regulation. Crit CareMed 1996; 24:163–172

25. Stern R, McPherson M, Longaker MT: Histo-logic study of artificial skin used in the treat-ment of full-thickness thermal injury. J BurnCare Rehabil 1990; 11:7–13

26. Sheridan RL, Hegarty M, Tompkins RG, et al:Artificial skin in massive burns—results toten years. Eur J Plast Surg 1994; 17:91–93

27. Heimbach D, Luterman A, Burke J, et al:Artificial dermis for major burns. A multi-center randomized clinical trial. Ann Surg1988; 208:313–320

28. Frame JD, Still J, Lakhel-LeCoadou A, et al:Use of dermal regeneration template in con-tracture release procedures: A multicenterevaluation. Plast Reconstr Surg 2004; 113:1330–1338

29. Young RC, Burd A: Paediatric upper limbcontracture release following burn injury.Burns 2004; 30:723–728

2623Crit Care Med 2007 Vol. 35, No. 11