Treatment of full skin thickness burn injury using cultured epithelial grafts

Bums (1991)

17,

Printed in Great Britain

(6), 495-498

495

Treatment of full skin thickness burn injury using cultured epithelial grafts A. Blight, E. M. Mountford, I. M. Cheshire, J. M. P. Clancy and P. L. Levick Skin Culture Laboratory,

Birmingham

Accident Hospital, Birmingham,

This report presents the experience gained from 26 patients freated with autogenic cultured epifhelial grafts [auto-CEG). All auto-CEG were applied to wounds clinically a&red as full skin thickness inju y. In total 89 separate sites were grafted. The overall estimate of ‘take’ranged from 0 to 98 per cent with a mean value of 15 per cent. 7’he highest level of ‘take’(43 per cent) was observed when auto-CEG were applied to wounds which had been previously covered with allogenic split-thickness skin grafts. An increased incidence of wound colonization with pathogenic species of bacteria corresponded with a decreased gruff take’. Ps. aeruginosa and Staph. aureus werefound to be present on 32.6 per cent and 60.5 per cent ofwound swabs respectively, where 10 per cent or less ‘fake’of auto-CEG was seen, indicating that bacterial infection is in part responsible for gruff failure. However, in 20.9 per cent of such instances, no growth of bacteria was detecfed, perhaps suggesting that certain wound beds may not present the correct physical environment necessary 10 supporf proliferating epithelial cells isolated from their underlying dermal component.

Introduction Following the reports by Professor Howard Green and co-workers (Gallico et al., 1984) of the successful coverage of large area bum wounds with epithelial skin grafts grown in vitro, this hospital and many bum centres established tissue culture facilities. The aim was to produce autogenic cultured epithelial grafts (auto-CEG), primarily for the treatment of bum patients. At Birmingham Accident Hospital a programme of grafting with CEG was subsequently initiated for all patients with severe bum injury. The majority of patients were treated within the West Midlands Regional Unit. Patients were also treated with CEG at six other bum centres within Britain. Encouraging reports have been published, frequently on small series of patients, giving clinical estimations of auto-CEG ‘take’ as high as 80 per cent (Gallico et al., 1984; O’Connor et al., 1984; Munster et al., 1990). However, the Birmingham group and others have reported lower levels of ‘take’ with CEG (Eldad et al., 1987; Levick et al., 1987). Reports that allogenic CEG (allo-CEG) could be successfully grafted across histocompatibility barriers (Thivolet et al., 1986a,b) led to the widespread use of donor material. The efficacy of allo-CEG in closing full skin thickness defects mirrors that reported for auto-CEG. Up to 85 per cent ‘epithelial regeneration’ was recorded in some cases (DeLuca et al., 1989), whilst others found allo-CEG unsuccessful (Madden et al., 1986). It has now been established that c, 1991 Buttenvorth-Heinemann 0305-4 179/91/060495-04

Ltd

UK

allogenic material does not persist at the wound site for more than 1 week (Brain et al., 1989; Burt et al., 1989). In view of the confused and conflicting data, it is difficult to establish the efficacy of CEG under given circumstances and how best to employ this new technology for the most effective treatment of burn injury. To address and clarify this issue, this report presents data collated over a series of patients treated since March 1985. All patients were treated with auto-CEG, applied to wounds which were clinically defined as full skin thickness injury, since they were insensitive to pin prick, and excision to subdermal fat was required to reach viable tissue.

Methods Autologous cultures were initiated for 59 bum patients. Patient age ranged from 10 months to 80 years. Patients selected for treatment with auto-CEG normally had bums covering greater than 30 per cent body surface area (BSA), although exceptionally one patient had as little as 2 per cent BSA bum injury. The maximum BSA bum was 95 per cent and the mean value was 55 per cent. Skin biopsies were usually taken within 1 week of admission (49/59 biopsies) but have been taken in excess of 10 weeks after injury. Initially, full skin thickness biopsies were taken under local anaesthesia in order to establish primary keratinocyte cultures, but latterly, partial skin thickness biopsies have been taken at the time of the first grafting procedure. Cells were extracted from full skin thickness biopsies by repeated 30-min treatments (five to six treatments) with 0.2 per cent tryspin:0.2 per cent ethylenediaminetetra acetic acid in calcium and magnesium-free phosphate buffered saline, at 3 7°C. Partial skin thickness biopsies were also processed by this method or by a single overnight trypsinization at 4°C. This alternative processing method was assessed in six paired samples and was not found to be detrimental to the subsequent colony-forming ability of the keratinocytes harvested. Cells derived from biopsies were inoculated into 75 cm’ growth area tissue culture flasks containing lethally irradiated Swiss 3T3.J2 feeder layers and keratinocyte medium as described by Green et al. (1979) and O’Connor et al. (1984). Primary or secondary cultures were prepared for grafting, ideally within 1 week of confluence, tertiary cultures were not used. Cultures for grafting were changed on to a medium from which cholera toxin had been omitted, at least

Bums (1991) Vol. 17/No. 6

496

48 h prior to application to a patient. Sterility of the cultures was verified by ‘streaking’ aliquots of culture conditioned medium onto blood agar plates at least 24 h prior to application of the auto-CEG. Grafts were detached using the neutral protease, Dispase. Following detachment the cell sheet was secured with Liga clips (Ethicon, Edinburgh, UK) to a supporting dressing of petroleum jelly gauze (Chesebrough Ponds, Greenwich, USA) or N-A Dressing (Johnson and Johnson, Slough, UK). The prepared CEG were stored in medium without serum and maintained in modular incubators flushed through with a 10 per cent CO, in air gas mixture. This procedure was coordinated with the operating schedule in order to have the CEG ready with minimal delay before use. Wound swabs were taken from areas of full skin thickness injury prior to application of auto-CEG. Grafts were placed edge to edge to cover the surface of the wound bed completely. Dressings of petroleum jelly gauze, surgical gauze, absorbent cotton wool and a securing crepe bandage were applied over the CEG. These dressings were changed as and when required, in accordance with standard practice, usually every 2 days. The CEG supporting dressing was left undisturbed for 7-10 days unless signs of significant wound infection indicated its removal at an earlier date. Upon removal of the dressing an estimation of the success of the applied auto-CEG was made by subjective visual assessment of the area covered with new epithelium as a percentage of the site grafted, described as ‘% take’. Photographs were taken to assist the visual assessment of graft ‘take’, when considered appropriate.

Results Cultures were established from 59 biopsies, 10 of these were infected and four biopsies failed to give rise to confluent monolayers. Auto-CEG were applied and the ‘take’ assessed in 26 patients out of the 59 selected for treatment: 24 patients died and two healed prior to grafting, in seven surviving patients cultures suitable for grafting were not established for the reasons given above. The mean estimated area of bum for these 26 patients was 50 per cent (range 2-75 per cent), (TableI). Nine patients received serial applications, and in total 89 separate sites were grafted. The overall estimate of ‘take’ of auto-CEG for these patients is shown in Tablel, ‘take’ ranged from 0 to 98 per cent with a mean percentage value of 15. Auto-CEG were applied to a variety of wound beds. None of the bums treated with auto-CEG were excised to Table I. The severity of injury of patients who were selected for treatment with cultured epithelial grafts and the subsequent graft success Estimated of burn f%) O-l 0 1 l-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-l 00 Not known

Auto-CEG

area Biopsied 1 1 2 8 12 5 7 10 3 1 5

Grafted 1 0 1 5 9 2 4 4 0 0 0

(%)

‘Take’

Range

0

0 -

41.2 9.2 2.1 45.9 15.9 25.3

O-60 o-33 O-25 O-70 O-98 O-70

-

100 0

0

75

000 000

00

:: 0 B Ii 5

Ez

00

0

50

0

$ F 8 000 0

25

0

0

000 000

0 0 :

,.

-

0

0 -

”

zzz% kzm% 0000

0

000000

Ext.

Gram -M (n = 20)

000000

000

Gram

PDS

(nTy6)

(nly2)

Wound

888880

PCEG in:7

8)

bed

Figure 1. The effect of wound bed on the success of autologous cultured epithelial grafts. The type of wound bed was classified in

89 applications of auto-CEG and the subsequent ‘take’assessed. Categories of wound bed were: EXC, excised; Gran., granulation tissue; PDS, pretreated with allogenic split skin grafts; PCEG, pretreated with CEG; +M, with meshed graft; -M without meshed graft. Horizontal rules = mean.

fascia. In general bums were excised to vital tissue and CEG were applied to this wound bed, usually fat or granulation tissue. The resulting ‘take’ by clinical assessment is presented in Figure 1. Due to the time required to produce auto-CEG, cultures were most frequently applied to granulation tissue, which was found to be a poor recipient bed. ‘Take’ was considerably improved, when auto-CEG were applied to close the interstices of meshed autologous split-thickness skin grafts (SSG), the figures given in Figure I are corrected to exclude the healing attributable to the applied meshed SSG, e.g. if a 1: 6 meshed SSG was applied beneath auto-CEG then 17 per cent of the wound would be covered by the mesh and only the remaining 83 per cent could potentially be healed by the applied auto-CEG. In general, it was observed that when the auto-CEG failed the meshed SSG also failed. The highest level of ‘take’ (43 per cent) was achieved when auto-CEG were applied to wounds which had been excised and covered with allogenic SSG. The allogenic SSG was usually obtained from multiple organ donors, frozen in a viable state, and stored for up to 6 months at - 70°C. Wounds were covered with allogenic SSG, which was replaced as rejection occurred, from the time of excision until application of the auto-CEG, usually from 2 to 3 weeks. A single swab was taken from each site immediately prior to grafting, to determine the bacterial flora of the wound. An increased incidence of wound colonization with potentially pathogenic species of bacteria correlated with decreased graft ‘take’ (Table II). The incidence of bacterial species found

Blight et al.: Cultured epithelial grafts

497

Table II. Percentage of sites colonized with pathogenic bacteria prior to grafting and the percentage graft ‘take’with autologous cultured epithelial grafts % ‘Take’ 50

% Sites colonized 86 44 30

Table III. Presence of bacterial species on wound swabs in bum patients prior to grafting with autogenic cultured epithelial grafts Bacterial speices

% Incidence *

Streptococcus pyogenes Staphylococcus aureus Acinetobacter anitratus Pseudomonas aeruginosa Streptococcus faecalis Klebsiella spp. ‘Coliform bacilli’ Proteus spp. Candida spp.

11 47 37 26 4 11 5 2 4

Micrococcus spp. Diphtheroides spp.

23 5

No growth

19

*Swabs were obtained from 57 sites in 18 patients prior to wound preparation with antiseptic agents and before application of autoCEG.

swabs is given in Table Ill. Where ‘take’ of CEG was less than 10 per cent the prevalent bacteria Ps. aeruginosa and Sfaph. aureus, were found to be present on 32.6 per cent and 60.9 per cent of wound swabs respectively, where the number of swabs was M= 43. In sites where greater than 50 per cent ‘take’ was observed, wound colonization by these species was never recorded (n = 5). However, it should also be appreciated that in 20.9 per cent of the sites where less than 10 per cent ‘take’ occurred, no growth of bacteria was detected. On sites where over 50 per cent ‘take’ was observed, wound beds were as frequently colonized with Acinefobacfer anifrafw as sites where less than 10 per cent ‘take’ was seen (39.5 per cent and 3 7.2 per cent respectively). Patients known to be colonized with beta haemolytic streptococcus (Lancefield group A) were not grafted and no instance of colonization with this strain at time of grafting was recorded. on wound

Discussion Application of auto-CEG to full skin thickness bum wounds has resulted in limited success. When auto-CEG were applied to freshly excised areas the average graft success was disappointing, as low as 8 per cent. This contrasts with conventional SSG, where freshly excised areas would constitute a good recipient wound bed. Lack of adherence of keratinocytes to fibrin-coated collagen has been reported in vitro (Shakespeare and Shakespeare, 1987), and such conditions could be created in a freshly bleeding wound bed. DeLuca and co-workers attributed loss of auto-CEG to haematoma in two patients who had poor grafting results (DeLuca et al., 1989). Granulation tissue was also found to be a poor wound bed for reception of auto-CEG in this and other studies.

Graft failure in these granulating lesions has been attributed to the presence of bacteria, since such wounds are usually heavily colonized (Teepe et al., 1987; DeLuca et al., 1989). Our study showed an association between graft failure and the presence of bacterial colonization of the wound bed. However, the sample sizes were too small to allow statistical analysis and correlation between graft ‘take’, bacterial colonization and different wound beds. Ps. aeruginosa was observed in 32.6 per cent of all cases of less than 10 per cent ‘take’ of auto-CEG, thus there is weak evidence to suggest that the presence of Ps. aeruginosa may contribute to graft failure. More significantly, Staph. aurew was found in 60.5 per cent of sites where less than 10 per cent ‘take’ occurred, suggesting that this organism is partly responsible for graft failure. In this study, A. anifratus was not considered to cause graft failure since the same frequency of colonization was observed in both high and low ‘take’ groups, although the number of observations in the ‘high take’ group is small (n= 5). Recent in vitro studies using fibroblasts and a transformed keratinocyte cell line indicated that extracts of media in which A. unitrufus had been grown were as toxic as extracts derived from cultures of Ps. aeruginosa or Staph. aweus (Taylor et al., 1990). However, differences were observed between strains of the same species of bacteria and this may explain discrepancies between the results obtained in the in vitro study and the clinical data presented here. Whilst the presence of bacteria is clearly important, more detailed study of the conditions which facilitate graft success are needed, as other factors must also be influencing the fate of auto-CEG since less than 10 per cent ‘take’ was observed in some sites where no pathogenic bacteria were detected. It is possible that the granulation tissue may not provide the physical environment necessary to support proliferating keratinocytes isolated from an underlying dermis (Appleby et al., 1991). Auto-CEG appear to be effective only on wound beds which have been appropriately prepared by using allogenic SSG as a biological dressing. Although slightly lower, our results for this type of wound bed (mean 43 per cent) correlate well with those of Green and co-workers, who reported the treatment of a series of 21 paediatric patients where ‘take’ ranged from 0 to 80 per cent with a mean value of 55 per cent (Compton et al., 1989). Application of allogenic SSG significantly modifies the wound bed, most notably by increasing the density of capillaries in the granulation tissue as well as reducing bacterial colonization (Pruitt and Levine, 1984). These changes may influence both attachment and nutrition of the auto-CEG subsequently applied. Future investigations into these changes may indicate why auto-CEG have failed to ‘take’ on wounds where bacterial colonization was not detected and therefore not considered to be the cause of failure of the auto-CEG. Furthermore, blistering, graft instability and an absence of rete pegs and ridges have been observed by a number of groups for a period after graft ‘take’ (Gallico and O’Connor, 1985; Woodley et al., 1988), and anchoring fibrils took up to 1 year to develop (Compton et al., 1989). Thus the nature of the wound bed may explain the poor grafting results seen clinically in freshly excised and ‘clean’ granulating wounds. The application of cultured skin grafts has been regarded as very expensive, however, the cost of materials required to produce an average sized auto-CEG from a 75 cm2 flask is less than f10 (c.US$lT), and a basic laboratory can be equipped for under f 15 000 (c. $25 000), plus staffing costs _ this was not considered prohibitive. However, the circumstances in which auto-CEG can be utilized suc-

Burns (1991) Vol. 17/No.

498

cessfully are still very restricted, at present its use is confined to those patients where the use of allogenic-SSG (or a substitute biological dressing) can be justified. Further research is required to improve the success of auto-CEG on other wound beds before the technique can be effectively used as a generalized treatment for bums of all sizes. The combination of auto-CEG with a dermal component prior to grafting may improve ‘take’ of auto-CEG on other wound beds, since these composite grafts would more closely resemble conventional SSG which would be expected to perform well under these conditions. The permanent survival of the carrier dermis may not be crucial since a ‘new dermis’ appears to re-form below the auto-CEG (Compton et al., 1989). What is crucial is that the keratinocytes are presented in a state where they can proliferate and remain, and that epithelial cover of the wound remains stable.

Acknowledgements We would like to acknowledge the following for their contributions to this project: Mr J. P. Gowar and Miss R. L. Lester, Bums Consultants, for allowing us to report their patients and Dr J. C. Lawrence and staff of the Bums Research Group for the bacteriological studies. We thank Professor H. Green for kindly donating the 3T3.J2 cell line, Dr C. George-Nascimento of Chiron Corporation for donations of human EGF and MS K. E. Young for technical assistance.

References Appleby M., Leary T., Blight A. et al. (1991) Epidennal keratinocyte self renewal is dependant upon dermal integrity 1. Invest. Demafol. (submitted) Brain A., Purkis P., Coates P. et al. (1989) Survival of cultured allogenic keratinocytes transplanted to deep dermal bed assessed with probe specific for Y chromosome. Br. Med. 1.298, 917. Burt A. M., Pallett C. D., Sloane J. P. et al. (1989) Survival of cultured allografts in patients with bums assessed with probe specific for Y chromosome. Br. Med. J 298,915. Compton C. C., Gill J. M., Bradford D. A. et al. (1989) Skin regenerated from cultured epithelial autografts on full-thickness bum wounds from 6 days to 5 years after grafting: a light, electron microscopic and immunohistochemical study. Lab. Invest. 60, 600. DeLuca M., Albanese E., Bondanza S. et al. (1989) Multicentre experience in the treatment of bums with autologous and allogenic cultured epithelium, fresh or preserved in a frozen state. Burns 15, 303. Eldad A., Burt A., Clarke J. A. et al. (1987) Cultured epithelium as a skin substitute. Burns 13, 173.

6

Gallico G. G., O’Connor N. E., Compton C. C. et al. (1984) Permanent coverage of large bum wounds with autologous cultured human epithelium. New Engl J Med. 311, 448. Gallico G. G. and O’Connor N. E. (1985) Cultured epithelium as a skin substitute. C/in. Phsf. Surg. 12, 149. Green H., Kehinde 0. and Thomas J. (1979) Growth of cultured human epidermal cells into multiple epithelia suitable for grafting. Pmt. Nat/. Acad. Sci. 76, 5665. Levick P. L., Blight A., Clancy J. M. P. et al. (1987) Generation of autograft; current techniques and results. In: Teepe R. G. C. (ed.), Clinical tcse of Cultured Epifhelium in Surgery and Demafology. Wheathampstead, Herts: Medical and Scientific Conferences, pp. 15-18. Madden M. R., Finkelstein T. L., Stiano-Coico et al. (1986) Grafting of cultured allogenic epidermis on second and third degree bum wounds in 26 patients. 1. Trauma 26, 955. Munster A. M., Weiner S. H. and Spence R. J. (1990) Cultured epidermis for the coverage of massive bum wounds; a single centre experience. Ann. Surg. 211, 676. O’Connor N. E., Gallico G. G., Compton C. et al. (1984) Grafting of bums with cultured epithelium prepared from autologous epidermal cells. II. intermediate term results on three paediatric patients. In: Hunt T. K., Heppenstall K. B., Pines E. et al. (eds), soft and Hard Tissue Repair: Biological and Clinical Aspects, vol. 2, New York: Praeger, chap. 14, pp. 283-292. Pruitt B. A. and Levine N. S. (1984) Characteristics and uses of biological dressings and skin substitutes. Arch. Surg. 119,312, Shakespeare V. A. and Shakespeare P. G. (1987) Growth of cultured human keratinocytes on fibrous dermal collagen: a scanning electron microscope study. Burns 13,343. Taylor D., Whatling C., Keamey J. N. et al. (1990) The effect of bacterial products on human fibroblast and keratinocyte detachment and viability. Br. 1. Demafol. 112, 23. Teepe R. G. C., Ponec M., Kempenaar B. et al. (1987) Clinical histological and ultrastructural aspects of cultured epithelium. In: Teepe R. G. C. (ed.), Clinical Use of Cultured Epifhelium in Surgery and Dermatology. Wheathampstead, Herts: Medical and Scientific Conferences, pp. 37-45. Thivolet T., Faure M., Demidem A. et al. (1986a) Cultured human epidermal allografts are not rejected for a long period. Arch. Demafol. Res. 278, 252. Thivolet T., Faure M., Demidem A. et al. (1986b) Long term survival and immunological tolerance of human epidermal allografts produced in culture. Tramphnfafion 42, 274. Woodley D. T., Peterson H. D., Herzog S. R. et al. (1988) Bum wounds resurfaced by cultured epidermal autografts show abnormal reconstitution of anchoring fibrils. JAA4A 259,2566.

Paper accepted

19 June 1991.

Correspondence should be addressed to: Mrs A. Blight, Skin Culture Laboratory Accident Hospital, Bath Row, Birmingham Bl5 INA, UK.