Graft Adherence to De-epithelialized Surfaces: A Comparative Study MICHAEL J. TAVIS, M.D., JAMES W. THORNTON, JOHN H. HARNEY, E. AUBREY WOODROOF, PH.D.,* ROBERT H. BARTLETT, M.D.
Graft adherence may be divided into two distinct phases: Phase I, which is fibrin dependent and Phase II, which begins after 72 hours with fibro-vascular ingrowth or vascular anastomosis with the graft material. Adherence values for autograft, homograft, heterograft, silicone membrane and a modified collagen membrane were evaluated during the fibrin-dependent Phase I period at S and 72 hours on dermal, fascial and granulating surfaces on rats. Modified collagen membrane demonstrated a superior adherence at both times tested on dermal and fascial surfaces, while autograft and homograft were significantly more adherent on granulating surfaces at 72 hours. The inert silicone membrane was consistently the least adherent. Granulating surfaces produced the highest adherence values at 5 hours and fascial surfaces at 72 hours. The higher values found with collagen indicate that future research directed toward the production of a synthetic wound dressing or skin should be directed toward biologically derived materials, rather than inert materials. The data supports the concept of the role offibrin as the bonding factor in Phase I adherence and implies that collagen, rather than elastin, is primarily responsible for early graft adherence.
THE MOST IMPORTANT property of a biological or prosthetic material applied to a de-epithelialized surface is adherence to the wound base. Adherence, more than any other factor, defines the success of the graft. It is an absolute prerequisite for vascularization or "take" of autograft, and is essential for the prevention of suppuration beneath prosthetic materials. This concept of the primary importance of adherence is Submitted for publication February 7, 1976. * Shiley Laboratories, Santa Ana, California 92705. Reprint requests: Michael J. Tavis, M.D. Department of Surgery, University of California, Irvine Irvine, California 92717. Supported in part by a grant from the Donald E. Baxter Foundation. Presented to the Sixtieth Clinical Congress ofthe American College of Surgeons, Miami, Florida (preliminary report).
From the Department of Surgery, University of California, Irvine, Irvine, California 92717
supported by our own observations6-8 and the demonstration that adherence of a skin substitute to split-thickness defects such as donor sites and partial-thickness burns can reduce pain, limit infection, and consequently optimize the rate of healing. '5 It has also been shown that adherence of a dressing to a granulating or freshly excised surface can significantly reduce tissue infection3'4'6'8 with a subsequent reduction in morbidity and increased graft take. This study was designed to define the parameters involved in the adherence of graft materials, with the ultimate goal of producing a functional prosthetic skin. The specific objectives of this study were to establish baseline adherence values for biological and synthetic materials on various wounds. To accomplish this objective, a method was devised to quantitate the adherence of graft materials using a rat model on three types of surfaces: dermal, fascial, and granulation tissue. Methods and Materials
Graft materials evaluated in the study included: 1) Autograft-fresh isografts from syngeneic rats (.010 in. thick); 2) Homograft-fresh, obtained from SpragueDawley donors (.010 in. thick); 3) Heterograft-frozen, irradiated pigskint; 4) Silicone polymer membrane-
594
t Hancock Laboratories, Orange, California.
VOl. 184.* NO. S
GRAFT ADHERENCE
595
approximately 0.5 mil thick and laminated to cotton gauze backingt (See Ref. 5); 5) Modified bovine collagen membrane -approximately 6 mils thick and produced by extruding a fibrous suspension of cutaneously derived collagen onto a gauze backing and stabilized by treatment with a cross-linking agent. Amino acid composition, studied by chromatographic techniques, was identical to
bovine skin collagen.t All graft materials were cut with a standard 6 mm biopsy punch and placed immediately in longitudinal rows on the anesthetized recipient animal using a controlled rotation schedule to assure equivalent sampling. Six samples were selected to test multiple materials on the same animal in a controlled fashion, to generate a significant sample size with a minimum of 25 values per material, and to minimize shear forces on a curved dorsal surface. All five materials were tested on each animal, and a total of 85 animals were used to evaluate 840 grafts. Immediately after placement of the grafts, the animals were returned to their cages and allowed to move about freely, providing a random exposure of the grafts to trauma. Food and water were provided ad lib. Quantitative determinations of graft adherence were made at 5 and 72 hours. The 5-hour sample was used to evaluate early graft adherence. The 72-hour determinations were chosen to evaluate the innate adherence of the material in the absence of granulation tissue ingrowth (Phase I) into biological materials and the fabric backings of synthetic materials after a clot had formed between the graft and the wound surface. Once fibro-vascular ingrowth has occurred (Phase II), measurements of tensile strength, in fact, become a measurement of the testing of the underlying tissues, i.e. a measurement of wound healing. Dermal surfaces were produced by creating splitthickness defects, approximately 5 x 7 cm on the dorsal surfaces of syngeneic rats, using a small Padgett dermatome set at 0.0010 inch. Fascial surfaces were produced in one of two ways using full-thickness skin excisions. Defects produced on animals to be used for 5-hour samples were made by simple full-thickness surgical excisions to deep fascia below the panniculus carnosus of approximately 5 x 7 cm. Crusty serous exudates formed on the large excised surfaces by 72 hours and tended to entrap the grafts so that attempts to pull off one graft required excessive force and tended to dislodge adjacent grafts. To avoid this problem, individual 1 cm full-thickness defects were used to evaluate adherence at 72 hours. Granulating surfaces were produced in 8 to 10 days after full-thickness dorsal excisions. Similar serous exudates described previously were also encountered in this group. Here again, multiple 1 cm defects were used pure
t Edwards Laboratories, Inc., Santa Ana, Califomia.
FIG. 1. Vertically aligned tensiometer with fully adjustable platform allowing consistently perpendicular displacement of grafts.
for the 72-hour determinations and large single defects were used for the 5-hour measurements. Quantitation was achieved by the use of a specially designed tensiometer (Fig. 1). The tensiometer was designed to allow reproducible perpendicular displacement of the grafts. Perpendicular displacement of grafts minimizes shear forces, is easily mechanically aligned and generates the most reproducible values (Table 1). Shear forces were avoided to obviate the need to overcome the elastic potential of the material before adherence, per se, could be measured, and to avoid artifacts due to the variable curvature of the rat. At the time of measurement, the anesthetized animal was taped to a platform which allowed rotation in all directions. The 6 mm graft in place on the animal's back was attached to the tensiometer with a 5 mm metal disk using cyanoacrylate and raised at a TABLE 1. Adherence as a Function of Angular Displacement of Collagen at 5 Hours
Angle
Mean (gM/cm2)
S.D.
900 600
115.2 17.3 22.7
55.4 13.4 15.4
450
Ann. Surg. o November 1976
TAVIS AND OTHERS
596
5 HOURS MATERIAL ADHERENCE %ADHERENCE 72 Autograft 89 gm/cm2 Homograft 160 65 288 78 Collogen 74 Pigskin 85 74 93 SLSilicone
MATERIAL
72 HOURS = ADHERENCE %ADHERENCE
Autograft 484gm/cm2 578 Homograft 435 Collagen 400 Pigskin 499 Silicone
44 39 68 47 27
400
gm-0/o cm2
300-
Collogen Homogroft .... =
200
Autogrof t
FIG. 2. Split-thickness surface-cumulative index values at 5 and 72 hours. Cumulative index derived by multiplying adherence (tensile strength in gm/cm2) x% adherence (% of grafts remaining intact at time of measurement), to give values in terms of gm-%/cm.2
.-, Pigskin .>*A__ ~
Silicone
__
100-
Hours
constant velocity of
1 cm/sec. The force necessary to dislodge the graft was recorded on a strip chart and converted into gm/cm.2
72
measurement. In order to most accurately represent the total adherence of the materials tested, the mean tensile strength in gm/cm2 for all grafts in a given test situation was multiplied by the per cent of the grafts remaining Results intact on the surface to give a cumulative index in This calculation includes a zero value for gm-%/cm2. In determining the type of materials to be extensively non-adherent samples. tested, a pilot study was conducted with the following On split-thickness surfaces (Fig. 2), collagen membrane results: demonstrated the highest cumulative index. At 5 hours, 1. No difference could be demonstrated in the adhertensile was 288 gm/cm2 with 78% of the grafts strength ence of fresh frozen and frozen irradiated pigskin. remaining intact, resulting in a cumulative index of 2. No difference in adherence could be found when fresh autograft or homograft was compared to autograft approximately 173. At 72 hours, tensile strength was 435 and homograft which had been frozen in liquid nitrogen. gm/cm2 with 68% remaining adherent giving an index of 3. No difference in adherence could be demonstrated in 295. These values were followed in decreasing order by homograft, autograft, pigskin, and silicone. collagen with or without a gauze backing. On fascial surfaces (Fig. 3) modified collagen memFrom these initial studies it was concluded that the brane again had the highest cumulative index, with values backing material on the collagen membrane and the at 5 and 72 hours of 86 and 438. Following in decreasing viability of the graft materials were not significant factors order were pigskin, homograft, autograft, and silicone. in surface adherence at 5 and 72 hours. On granulating surfaces (Fig. 4), the cumulative index The reproducibility of the model was tested by for all materials decreased at 72 hours due to a consistent comparing the tensile strength of autograft on splitthickness defects at 5 and 72 hours on ten animals. decrease in the number of grafts remaining adherent on the Separate animals showed no significant differences in surface although their tensile strengths, with the exception of collagen increased. The cumulative index was highest means or standard deviations at either time tested. Two factors were considered important in evaluating for autograft, followed by homograft, collagen, pigskin, the adherence of the materials; the tensile strength in and silicone. The decreased tensile strength of collagen gm/cm2, (measured only on the grafts which remained on was felt to be secondary to collagenous breakdown on the the surface at the time of measurement), and the per cent granulating surface. Group means and standard deviaof the grafts which remained on the surface at the time of tions for each surface are listed in Table 2.
Vol. 184 . No. S
597
GRAFT ADHERENCE 5 HOURS
MATERIAL ADHERENCE %ADHERENCE 76 Autogrof1 84 gm/cm2 67 86 Homograft 81 106 Collogen 74 63 Pigskin 98 Silicone 54
72 HOURS ADHERENCE %ADHERENCE 81 Autogroft 360gm/cm2 491 78 Homogroft 521 84 Collagen 473 84 Pigskin 348 81 Silicone Collagen MATERIAL
400FIG. 3. Fascial surfacecumulative index values at 5 and 72 hours.
Homograft
gm-%
'1.
Cm2
300-
Autograft .
Silicone
200-
100-
72
Hours
analysis of variance was used to study measurements of tensile strength and cumulative index as a function of the The data were analyzed by accepting the criteria for the type of material tested and the time of sampling (Table 3). analysis and homogeneity of variance. A two-way The results of the analysis show that on all three surfaces, Statistical Analysis
5 HOURS
|
72 HOURS
MATERIAL ADHERENCE -%ADHERENCE 95 Autograft 240 gm/cm2
MATERIAL |ADHERENCE %ADHERENCE 30 Autograft 406gm/cm2
Homograft Collogen Pigskin Silicone
Homograft Collagen Pigskin Si I icone
213 250 177 119
96 96 97 98
343 184 309 350
30 34 18 15
400FIG. 4. Granulating surface-cumulative index values at 5 and 72 hours.
gm
-%
cm2
300'
200
A.Autoggraf I
Homograft Sil logen
100
""Silicone Hou rs
72
Ann.
TAVIS AND OTHERS
598
Surg. a November 1976
TABLE 2. Cumulative Index -Mean and S.D.
On granulating surfaces at 5 hours no significant differences were found among the groups. At 72 hours, Material Mean-5 hrs-S.D. Mean-72 hrs-S.D. two distinctly significantly different groups are found with Split-thickness autograft and homograft contained in the more adherent Autograft 64 ± 29 213 + 233 and collagen, heterograft, and silicone contained in group; Homograft 65 ± 20 225 ± 63 the less adherent group. There was, however, no Collagen 181 + 77 296 80 Heterograft 63 ± 25 188 ± 83 significant difference among any of the materials within Silicone 69 ± 25 135 ± 67 either the higher or the lower groups. Fascial This analysis demonstrates that collagen membrane has Autograft 65 ± 24 292 ± 122 Homograft 57 ± 22 383 ± 121 significantly better adherence than the other materials on Collagen 114 ± 60 438 ± 131 both split-thickness and fascial defects, while autograft Heterograft 47 ± 23 397 ± 134 and homograft are significantly more adherent after 72 Silicone 53 + 29 282 ± 59 Granulating hours on a granulating surface. Silicone was consistently Autograft 228 + 121 122 ± 54 the least adherent. Homograft 205 ± 101 103 ± 34 When surfaces were compared (Table 5) using a similar Collagen 240 ± 99 61 + 12 Heterograft 172 + 84 56 ± 19 one-way analysis of cumulative index, granulating surSilicone 117 39 52 ± 27 faces were consistently higher at 5 hours and fascial surfaces were consistently superior in terms of adherence at 72 hours. An analysis of tensile strength produced both the type of material tested and the time of sampling showed significant differences at the 95% level of identical results. confidence. Discussion Graft materials were compared using a one-way analysis of variance conducted on cumulative index Graft adherence may be divided into two distinct numbers to determine significant differences between phases. Phase I occurs during the initial placement of the groups at 5 and 72 hours at a 95% level of confidence. The graft when adherence is due to fibrin bonding2'9 and lasts results of this analysis are summarized below and listed in for approximately 72 hours. The second phase (Phase II) Table 4. begins with the onset of fibro-vascular ingrowth or On split-thickness surfaces at 5 hours, collagen vascular anastomosis with the graft material. This study demonstrated a significantly greater adherence than the was designed to examine and quantitate Phase I remaining groups, among which there was no significant adherence. difference. At 72 hours, collagen was significantly more There was no consistent pattern in the type of biological adherent than the other materials and silicone was material that adhered best to all surfaces, at both times significantly less adherent. tested (Table 4). This finding implies that there is little or On fascial surfaces at 5 hours, collagen was significantly no effect on graft adherence associated with either foreign more adherent than the other four materials whose values or native dermal components during Phase I. The only were not significantly different. At 72 hours, a sig- consistent finding was that the inert silicone grafts were nificantly different bimodal grouping was found with consistently less adherent than the biological materials. collagen, heterograft, and homograft comprising the This finding explains the general failure of inert laminates higher group; and autograft and silicone comprising the as synthetic wound coverings in most circumstances. lower group. There was, however, no significant differ- Adherence of these materials is largely dependent upon ence among materials within either the higher or lower fibro-vascular ingrowth into the interstices of the backing groups. materials during Phase II. Cotton gauze backing failed to TABLE 3. Two-way Analysis of Variance
df
Split-thickness Fascial
Granulating
Time Material Time Material Time Material
F 1, F 6, F 1, F 6,
237 237 237 237
F 1, 139 F 6, 139
Critical F 3.84 2.21 3.84 2.21 3.92 2.29
Calculated F Tensile Strength
Calculated F Cum. Index
690 3.86 647 6.52
616 41.0 470 5.43
22.8 2.78
58.2 3.44
VOl. 184.o NO. 5
599
GRAFT ADHERENCE
provide an adequate surface for fibrin bonding during Phase I, which led to loss of early adherence with eventual suppuration before ingrowth could occur. The superior adherence of biological materials, regardless of their origin, appears to be an innate property of the material and implies that future work directed toward the production of synthetic skins should involve the evaluation of biologically derived materials as opposed to inert materials with fabric backings. This concept is clearly demonstrated by the excellent potential for modified collagen membrane as a dressing for splitthickness and fascial surfaces found in this study. Although the inadequacy of collagen as a covering for granulating surfaces was initially discouraging, it is felt that the poor results were due to breakdown by collagenase. Ongoing biochemical and physical modifications are being conducted to improve collagen adaptability to this surface. Granulation tissue consistently produced the highest values of tensile strength and cumulative index at 5 hours. By 72 hours, the cumulative index for granulating surfaces fell below those of split-thickness and fascial surfaces. Fascial surfaces were consistently the highest at 72 hours, supporting the concept of early excision to fascia in the treatment of full-thickness skin losses and burns and coverage with either a temporary covering or immediate autografting. Although the cumulative index values for granulating surfaces was greater at 5 hours, the tensile strength for all materials and surfaces at 72 hours was not significantly different. This observation implies some differences between granulation tissue and the fresh surfaces which gradually dissipate over time. The greatly increased vascularity and bleeding of granulation tissue (which was scraped prior to the graft placement) provides an explanation for the superior adherence at 5 hours (more fibrin) and the decreased cumulative index at 72 hours (grafts lost due to bleeding). Fascial and split-thickness surfaces were characteristically smooth and relatively dry. The continued oozing and rough contour characteristic of the granulating surface predisposed grafts to loosening and eventual loss from the surface at 72 hours in a random trauma situation. This loss of graft material and the resultant decrease in cumulative index represents an experimental artifact on granulating surfaces at the 72-hour time period. At 5 hours the freshly scraped granulations provided a fibrin-rich surface on which the grafts demonstrated superior adherence in both tensile strength and cumulative index. This observation is consistent with Burleson' s fibrin bond hypothesis2 and our own observations9 concerning the role of fibrin as the bonding agent in Phase I adherence. Burleson concluded from studies with pigskin on granulating surfaces that fibrin linked elastin in the wound
TABLE 4. One-way Analysis of Materials
Surface
Time Autograft Homograft Heterograft Collagen Silicone
Split-
thickness 5 0 0 0 + 0
72 0 0 0 + -
Fullthickness
Granulating
5 0 0 0 + 0
5 0 0 0 0 0
72 0/0/+ 0/+ 0/+ 0/-
72
0/+ 0/+ 0/0/0/-
+, significantly greater; -, significantly lower; 0, no significant difference.
to elastin in the graft to provide the initial Phase I bond. Our studies, however, demonstrate that bovine collagen which contains no elastin frequently has a greater adherence than those materials that do contain elastin. Although this finding does not eliminate elastin as a bonding factor, it does imply that it is probably not of primary importance; rather, that collagen is. Clarification of the relative contributions of collagen and elastin during Phase I adherence is dependent upon adherence determinations of elastin grafts which will be available to us in the near future. The model is currently being used in additional experiments to define the role of fibrin and blood particles in graft adherence. Synthetic materials with demonstrated Phase I adherence are being evaluated in Phase II studies which involve the characterization and quantitation of
fibro-vascular ingrowth. Conclusions A reliable method for quantitating the adherence of a variety of materials to wound surfaces during the initial period (Phase I) of adherence has been developed. Quantitative adherence values for autograft, homograft, heterograft, modified bovine collagen membrane, and silicone polymer on split-thickness, fascial, and granulating surfaces have been determined. Viability of graft TABLE 5. One-way Analysis of Surfaces
Material
Time
Splitthickness
Fullthickness
6.38 29.2* 5.77 38.3*
Autograft
5 72
Homograft
5 72
6.41 21.3 10.4 22.5
Heterograft
5 72
6.29 18.8
4.66 39.7*
Collagen
5 72
8.59 43.8*
Silicone
5 72
17.3 29.6 6.88 13.5
* Significantly higher.
5.29 28.2*
Granulating 22.8* 12.2 20.5* 10.3 17.2* 5.56 24.0* 6.26 11.7* 5.25
600
TAVIS AND OTHERS
materials before the onset of vascular ingrowth is not a significant factor in adherence at periods up to 72 hours. Biological materials were significantly more adherent than inert materials, but no single biological material was consistently the most adherent. Granulating surfaces produced the highest cumulative index values at five hours, and fascial surfaces at 72 hours. All surfaces tended toward a similar mean value for tensile strength at 72 hours. Modified collagen membrane has excellent potential as a dressing for split-thickness and fascial surfaces. The role of fibrin in graft adherence is supported. Collagen, rather than elastin, is primarily responsible for graft adherence. Future work directed toward the production of synthetic wound dressings should involve biologically derived materials rather than or as a substrate for inert materials. References 1. Burleson, R. and Eiseman, B.: Effect of Skin Dressings and Topical Antibiotics on Healing of Partial-Thickness Skin Wounds in Rats, Surg. Gynecol. Obstet., 136:958, 1973.
Ann. Surg.
*
November 1976
2. Burleson, R. and Eiseman, B.: Mechanisms of Antibacterial Effect of Biologic Dressings, Surgery, 72:315, 1972. 3. Burleson, R. and Eiseman, B.: Nature of the Bond Between Partial-Thickness Skin and Wound Granulations, Ann. Surg., 177:181, 1973. 4. Eade, G. G.: The Relationship Between Granulation Tissue, Bacteria and Grafts in Burned Patients, Plast. Reconstruct. Surg., 22:42, 1958. 5. Kornberg, J., Burns, N. E., Kafesjian, R. and Bartlett, R. H.: Ultra Thin Silicone Polymer Membrane: A New Synthetic Skin Substitute, Trans. Am. Soc. Artif. Int. Organs, 18:39, 1972. 6. Tavis, M. J., Harney, J. H., Thornton, J. W., et al.: Modified Collagen Membrane as a Skin Substitute: Preliminary Studies, J. Biomed. Mat. Res., 9:285, 1975. 7. Tavis, M. J., Thornton, J. W., Harney, J. H., et al.: Adherence to De-epithelialized Surfaces: A Comparative Study, Surg. Forum, 25:39, 1974. 8. Tavis, M. J., Harney, J. H., Thornton, J. W. and Bartlett, R. H.: A Comparative Study of Burn Wound Management in Rats. Submitted to Surgery, 1976. 9. Thornton, J. W., Tavis, M. J., Harney, J. H., et al.: The Adherence of Collagen Grafts to Wound Surfaces: The Role of Fibrin. Submitted to Science, 1976.
Graft adherence may be divided into two distinct phases: Phase I, which is fibrin dependent and Phase II, which begins after 72 hours with fibro-vascular ingrowth or vascular anastomosis with the graft material. Adherence values for autograft, homograft, heterograft, silicone membrane and a modified collagen membrane were evaluated during the fibrin-dependent Phase I period at S and 72 hours on dermal, fascial and granulating surfaces on rats. Modified collagen membrane demonstrated a superior adherence at both times tested on dermal and fascial surfaces, while autograft and homograft were significantly more adherent on granulating surfaces at 72 hours. The inert silicone membrane was consistently the least adherent. Granulating surfaces produced the highest adherence values at 5 hours and fascial surfaces at 72 hours. The higher values found with collagen indicate that future research directed toward the production of a synthetic wound dressing or skin should be directed toward biologically derived materials, rather than inert materials. The data supports the concept of the role offibrin as the bonding factor in Phase I adherence and implies that collagen, rather than elastin, is primarily responsible for early graft adherence.
THE MOST IMPORTANT property of a biological or prosthetic material applied to a de-epithelialized surface is adherence to the wound base. Adherence, more than any other factor, defines the success of the graft. It is an absolute prerequisite for vascularization or "take" of autograft, and is essential for the prevention of suppuration beneath prosthetic materials. This concept of the primary importance of adherence is Submitted for publication February 7, 1976. * Shiley Laboratories, Santa Ana, California 92705. Reprint requests: Michael J. Tavis, M.D. Department of Surgery, University of California, Irvine Irvine, California 92717. Supported in part by a grant from the Donald E. Baxter Foundation. Presented to the Sixtieth Clinical Congress ofthe American College of Surgeons, Miami, Florida (preliminary report).
From the Department of Surgery, University of California, Irvine, Irvine, California 92717
supported by our own observations6-8 and the demonstration that adherence of a skin substitute to split-thickness defects such as donor sites and partial-thickness burns can reduce pain, limit infection, and consequently optimize the rate of healing. '5 It has also been shown that adherence of a dressing to a granulating or freshly excised surface can significantly reduce tissue infection3'4'6'8 with a subsequent reduction in morbidity and increased graft take. This study was designed to define the parameters involved in the adherence of graft materials, with the ultimate goal of producing a functional prosthetic skin. The specific objectives of this study were to establish baseline adherence values for biological and synthetic materials on various wounds. To accomplish this objective, a method was devised to quantitate the adherence of graft materials using a rat model on three types of surfaces: dermal, fascial, and granulation tissue. Methods and Materials
Graft materials evaluated in the study included: 1) Autograft-fresh isografts from syngeneic rats (.010 in. thick); 2) Homograft-fresh, obtained from SpragueDawley donors (.010 in. thick); 3) Heterograft-frozen, irradiated pigskint; 4) Silicone polymer membrane-
594
t Hancock Laboratories, Orange, California.
VOl. 184.* NO. S
GRAFT ADHERENCE
595
approximately 0.5 mil thick and laminated to cotton gauze backingt (See Ref. 5); 5) Modified bovine collagen membrane -approximately 6 mils thick and produced by extruding a fibrous suspension of cutaneously derived collagen onto a gauze backing and stabilized by treatment with a cross-linking agent. Amino acid composition, studied by chromatographic techniques, was identical to
bovine skin collagen.t All graft materials were cut with a standard 6 mm biopsy punch and placed immediately in longitudinal rows on the anesthetized recipient animal using a controlled rotation schedule to assure equivalent sampling. Six samples were selected to test multiple materials on the same animal in a controlled fashion, to generate a significant sample size with a minimum of 25 values per material, and to minimize shear forces on a curved dorsal surface. All five materials were tested on each animal, and a total of 85 animals were used to evaluate 840 grafts. Immediately after placement of the grafts, the animals were returned to their cages and allowed to move about freely, providing a random exposure of the grafts to trauma. Food and water were provided ad lib. Quantitative determinations of graft adherence were made at 5 and 72 hours. The 5-hour sample was used to evaluate early graft adherence. The 72-hour determinations were chosen to evaluate the innate adherence of the material in the absence of granulation tissue ingrowth (Phase I) into biological materials and the fabric backings of synthetic materials after a clot had formed between the graft and the wound surface. Once fibro-vascular ingrowth has occurred (Phase II), measurements of tensile strength, in fact, become a measurement of the testing of the underlying tissues, i.e. a measurement of wound healing. Dermal surfaces were produced by creating splitthickness defects, approximately 5 x 7 cm on the dorsal surfaces of syngeneic rats, using a small Padgett dermatome set at 0.0010 inch. Fascial surfaces were produced in one of two ways using full-thickness skin excisions. Defects produced on animals to be used for 5-hour samples were made by simple full-thickness surgical excisions to deep fascia below the panniculus carnosus of approximately 5 x 7 cm. Crusty serous exudates formed on the large excised surfaces by 72 hours and tended to entrap the grafts so that attempts to pull off one graft required excessive force and tended to dislodge adjacent grafts. To avoid this problem, individual 1 cm full-thickness defects were used to evaluate adherence at 72 hours. Granulating surfaces were produced in 8 to 10 days after full-thickness dorsal excisions. Similar serous exudates described previously were also encountered in this group. Here again, multiple 1 cm defects were used pure
t Edwards Laboratories, Inc., Santa Ana, Califomia.
FIG. 1. Vertically aligned tensiometer with fully adjustable platform allowing consistently perpendicular displacement of grafts.
for the 72-hour determinations and large single defects were used for the 5-hour measurements. Quantitation was achieved by the use of a specially designed tensiometer (Fig. 1). The tensiometer was designed to allow reproducible perpendicular displacement of the grafts. Perpendicular displacement of grafts minimizes shear forces, is easily mechanically aligned and generates the most reproducible values (Table 1). Shear forces were avoided to obviate the need to overcome the elastic potential of the material before adherence, per se, could be measured, and to avoid artifacts due to the variable curvature of the rat. At the time of measurement, the anesthetized animal was taped to a platform which allowed rotation in all directions. The 6 mm graft in place on the animal's back was attached to the tensiometer with a 5 mm metal disk using cyanoacrylate and raised at a TABLE 1. Adherence as a Function of Angular Displacement of Collagen at 5 Hours
Angle
Mean (gM/cm2)
S.D.
900 600
115.2 17.3 22.7
55.4 13.4 15.4
450
Ann. Surg. o November 1976
TAVIS AND OTHERS
596
5 HOURS MATERIAL ADHERENCE %ADHERENCE 72 Autograft 89 gm/cm2 Homograft 160 65 288 78 Collogen 74 Pigskin 85 74 93 SLSilicone
MATERIAL
72 HOURS = ADHERENCE %ADHERENCE
Autograft 484gm/cm2 578 Homograft 435 Collagen 400 Pigskin 499 Silicone
44 39 68 47 27
400
gm-0/o cm2
300-
Collogen Homogroft .... =
200
Autogrof t
FIG. 2. Split-thickness surface-cumulative index values at 5 and 72 hours. Cumulative index derived by multiplying adherence (tensile strength in gm/cm2) x% adherence (% of grafts remaining intact at time of measurement), to give values in terms of gm-%/cm.2
.-, Pigskin .>*A__ ~
Silicone
__
100-
Hours
constant velocity of
1 cm/sec. The force necessary to dislodge the graft was recorded on a strip chart and converted into gm/cm.2
72
measurement. In order to most accurately represent the total adherence of the materials tested, the mean tensile strength in gm/cm2 for all grafts in a given test situation was multiplied by the per cent of the grafts remaining Results intact on the surface to give a cumulative index in This calculation includes a zero value for gm-%/cm2. In determining the type of materials to be extensively non-adherent samples. tested, a pilot study was conducted with the following On split-thickness surfaces (Fig. 2), collagen membrane results: demonstrated the highest cumulative index. At 5 hours, 1. No difference could be demonstrated in the adhertensile was 288 gm/cm2 with 78% of the grafts strength ence of fresh frozen and frozen irradiated pigskin. remaining intact, resulting in a cumulative index of 2. No difference in adherence could be found when fresh autograft or homograft was compared to autograft approximately 173. At 72 hours, tensile strength was 435 and homograft which had been frozen in liquid nitrogen. gm/cm2 with 68% remaining adherent giving an index of 3. No difference in adherence could be demonstrated in 295. These values were followed in decreasing order by homograft, autograft, pigskin, and silicone. collagen with or without a gauze backing. On fascial surfaces (Fig. 3) modified collagen memFrom these initial studies it was concluded that the brane again had the highest cumulative index, with values backing material on the collagen membrane and the at 5 and 72 hours of 86 and 438. Following in decreasing viability of the graft materials were not significant factors order were pigskin, homograft, autograft, and silicone. in surface adherence at 5 and 72 hours. On granulating surfaces (Fig. 4), the cumulative index The reproducibility of the model was tested by for all materials decreased at 72 hours due to a consistent comparing the tensile strength of autograft on splitthickness defects at 5 and 72 hours on ten animals. decrease in the number of grafts remaining adherent on the Separate animals showed no significant differences in surface although their tensile strengths, with the exception of collagen increased. The cumulative index was highest means or standard deviations at either time tested. Two factors were considered important in evaluating for autograft, followed by homograft, collagen, pigskin, the adherence of the materials; the tensile strength in and silicone. The decreased tensile strength of collagen gm/cm2, (measured only on the grafts which remained on was felt to be secondary to collagenous breakdown on the the surface at the time of measurement), and the per cent granulating surface. Group means and standard deviaof the grafts which remained on the surface at the time of tions for each surface are listed in Table 2.
Vol. 184 . No. S
597
GRAFT ADHERENCE 5 HOURS
MATERIAL ADHERENCE %ADHERENCE 76 Autogrof1 84 gm/cm2 67 86 Homograft 81 106 Collogen 74 63 Pigskin 98 Silicone 54
72 HOURS ADHERENCE %ADHERENCE 81 Autogroft 360gm/cm2 491 78 Homogroft 521 84 Collagen 473 84 Pigskin 348 81 Silicone Collagen MATERIAL
400FIG. 3. Fascial surfacecumulative index values at 5 and 72 hours.
Homograft
gm-%
'1.
Cm2
300-
Autograft .
Silicone
200-
100-
72
Hours
analysis of variance was used to study measurements of tensile strength and cumulative index as a function of the The data were analyzed by accepting the criteria for the type of material tested and the time of sampling (Table 3). analysis and homogeneity of variance. A two-way The results of the analysis show that on all three surfaces, Statistical Analysis
5 HOURS
|
72 HOURS
MATERIAL ADHERENCE -%ADHERENCE 95 Autograft 240 gm/cm2
MATERIAL |ADHERENCE %ADHERENCE 30 Autograft 406gm/cm2
Homograft Collogen Pigskin Silicone
Homograft Collagen Pigskin Si I icone
213 250 177 119
96 96 97 98
343 184 309 350
30 34 18 15
400FIG. 4. Granulating surface-cumulative index values at 5 and 72 hours.
gm
-%
cm2
300'
200
A.Autoggraf I
Homograft Sil logen
100
""Silicone Hou rs
72
Ann.
TAVIS AND OTHERS
598
Surg. a November 1976
TABLE 2. Cumulative Index -Mean and S.D.
On granulating surfaces at 5 hours no significant differences were found among the groups. At 72 hours, Material Mean-5 hrs-S.D. Mean-72 hrs-S.D. two distinctly significantly different groups are found with Split-thickness autograft and homograft contained in the more adherent Autograft 64 ± 29 213 + 233 and collagen, heterograft, and silicone contained in group; Homograft 65 ± 20 225 ± 63 the less adherent group. There was, however, no Collagen 181 + 77 296 80 Heterograft 63 ± 25 188 ± 83 significant difference among any of the materials within Silicone 69 ± 25 135 ± 67 either the higher or the lower groups. Fascial This analysis demonstrates that collagen membrane has Autograft 65 ± 24 292 ± 122 Homograft 57 ± 22 383 ± 121 significantly better adherence than the other materials on Collagen 114 ± 60 438 ± 131 both split-thickness and fascial defects, while autograft Heterograft 47 ± 23 397 ± 134 and homograft are significantly more adherent after 72 Silicone 53 + 29 282 ± 59 Granulating hours on a granulating surface. Silicone was consistently Autograft 228 + 121 122 ± 54 the least adherent. Homograft 205 ± 101 103 ± 34 When surfaces were compared (Table 5) using a similar Collagen 240 ± 99 61 + 12 Heterograft 172 + 84 56 ± 19 one-way analysis of cumulative index, granulating surSilicone 117 39 52 ± 27 faces were consistently higher at 5 hours and fascial surfaces were consistently superior in terms of adherence at 72 hours. An analysis of tensile strength produced both the type of material tested and the time of sampling showed significant differences at the 95% level of identical results. confidence. Discussion Graft materials were compared using a one-way analysis of variance conducted on cumulative index Graft adherence may be divided into two distinct numbers to determine significant differences between phases. Phase I occurs during the initial placement of the groups at 5 and 72 hours at a 95% level of confidence. The graft when adherence is due to fibrin bonding2'9 and lasts results of this analysis are summarized below and listed in for approximately 72 hours. The second phase (Phase II) Table 4. begins with the onset of fibro-vascular ingrowth or On split-thickness surfaces at 5 hours, collagen vascular anastomosis with the graft material. This study demonstrated a significantly greater adherence than the was designed to examine and quantitate Phase I remaining groups, among which there was no significant adherence. difference. At 72 hours, collagen was significantly more There was no consistent pattern in the type of biological adherent than the other materials and silicone was material that adhered best to all surfaces, at both times significantly less adherent. tested (Table 4). This finding implies that there is little or On fascial surfaces at 5 hours, collagen was significantly no effect on graft adherence associated with either foreign more adherent than the other four materials whose values or native dermal components during Phase I. The only were not significantly different. At 72 hours, a sig- consistent finding was that the inert silicone grafts were nificantly different bimodal grouping was found with consistently less adherent than the biological materials. collagen, heterograft, and homograft comprising the This finding explains the general failure of inert laminates higher group; and autograft and silicone comprising the as synthetic wound coverings in most circumstances. lower group. There was, however, no significant differ- Adherence of these materials is largely dependent upon ence among materials within either the higher or lower fibro-vascular ingrowth into the interstices of the backing groups. materials during Phase II. Cotton gauze backing failed to TABLE 3. Two-way Analysis of Variance
df
Split-thickness Fascial
Granulating
Time Material Time Material Time Material
F 1, F 6, F 1, F 6,
237 237 237 237
F 1, 139 F 6, 139
Critical F 3.84 2.21 3.84 2.21 3.92 2.29
Calculated F Tensile Strength
Calculated F Cum. Index
690 3.86 647 6.52
616 41.0 470 5.43
22.8 2.78
58.2 3.44
VOl. 184.o NO. 5
599
GRAFT ADHERENCE
provide an adequate surface for fibrin bonding during Phase I, which led to loss of early adherence with eventual suppuration before ingrowth could occur. The superior adherence of biological materials, regardless of their origin, appears to be an innate property of the material and implies that future work directed toward the production of synthetic skins should involve the evaluation of biologically derived materials as opposed to inert materials with fabric backings. This concept is clearly demonstrated by the excellent potential for modified collagen membrane as a dressing for splitthickness and fascial surfaces found in this study. Although the inadequacy of collagen as a covering for granulating surfaces was initially discouraging, it is felt that the poor results were due to breakdown by collagenase. Ongoing biochemical and physical modifications are being conducted to improve collagen adaptability to this surface. Granulation tissue consistently produced the highest values of tensile strength and cumulative index at 5 hours. By 72 hours, the cumulative index for granulating surfaces fell below those of split-thickness and fascial surfaces. Fascial surfaces were consistently the highest at 72 hours, supporting the concept of early excision to fascia in the treatment of full-thickness skin losses and burns and coverage with either a temporary covering or immediate autografting. Although the cumulative index values for granulating surfaces was greater at 5 hours, the tensile strength for all materials and surfaces at 72 hours was not significantly different. This observation implies some differences between granulation tissue and the fresh surfaces which gradually dissipate over time. The greatly increased vascularity and bleeding of granulation tissue (which was scraped prior to the graft placement) provides an explanation for the superior adherence at 5 hours (more fibrin) and the decreased cumulative index at 72 hours (grafts lost due to bleeding). Fascial and split-thickness surfaces were characteristically smooth and relatively dry. The continued oozing and rough contour characteristic of the granulating surface predisposed grafts to loosening and eventual loss from the surface at 72 hours in a random trauma situation. This loss of graft material and the resultant decrease in cumulative index represents an experimental artifact on granulating surfaces at the 72-hour time period. At 5 hours the freshly scraped granulations provided a fibrin-rich surface on which the grafts demonstrated superior adherence in both tensile strength and cumulative index. This observation is consistent with Burleson' s fibrin bond hypothesis2 and our own observations9 concerning the role of fibrin as the bonding agent in Phase I adherence. Burleson concluded from studies with pigskin on granulating surfaces that fibrin linked elastin in the wound
TABLE 4. One-way Analysis of Materials
Surface
Time Autograft Homograft Heterograft Collagen Silicone
Split-
thickness 5 0 0 0 + 0
72 0 0 0 + -
Fullthickness
Granulating
5 0 0 0 + 0
5 0 0 0 0 0
72 0/0/+ 0/+ 0/+ 0/-
72
0/+ 0/+ 0/0/0/-
+, significantly greater; -, significantly lower; 0, no significant difference.
to elastin in the graft to provide the initial Phase I bond. Our studies, however, demonstrate that bovine collagen which contains no elastin frequently has a greater adherence than those materials that do contain elastin. Although this finding does not eliminate elastin as a bonding factor, it does imply that it is probably not of primary importance; rather, that collagen is. Clarification of the relative contributions of collagen and elastin during Phase I adherence is dependent upon adherence determinations of elastin grafts which will be available to us in the near future. The model is currently being used in additional experiments to define the role of fibrin and blood particles in graft adherence. Synthetic materials with demonstrated Phase I adherence are being evaluated in Phase II studies which involve the characterization and quantitation of
fibro-vascular ingrowth. Conclusions A reliable method for quantitating the adherence of a variety of materials to wound surfaces during the initial period (Phase I) of adherence has been developed. Quantitative adherence values for autograft, homograft, heterograft, modified bovine collagen membrane, and silicone polymer on split-thickness, fascial, and granulating surfaces have been determined. Viability of graft TABLE 5. One-way Analysis of Surfaces
Material
Time
Splitthickness
Fullthickness
6.38 29.2* 5.77 38.3*
Autograft
5 72
Homograft
5 72
6.41 21.3 10.4 22.5
Heterograft
5 72
6.29 18.8
4.66 39.7*
Collagen
5 72
8.59 43.8*
Silicone
5 72
17.3 29.6 6.88 13.5
* Significantly higher.
5.29 28.2*
Granulating 22.8* 12.2 20.5* 10.3 17.2* 5.56 24.0* 6.26 11.7* 5.25
600
TAVIS AND OTHERS
materials before the onset of vascular ingrowth is not a significant factor in adherence at periods up to 72 hours. Biological materials were significantly more adherent than inert materials, but no single biological material was consistently the most adherent. Granulating surfaces produced the highest cumulative index values at five hours, and fascial surfaces at 72 hours. All surfaces tended toward a similar mean value for tensile strength at 72 hours. Modified collagen membrane has excellent potential as a dressing for split-thickness and fascial surfaces. The role of fibrin in graft adherence is supported. Collagen, rather than elastin, is primarily responsible for graft adherence. Future work directed toward the production of synthetic wound dressings should involve biologically derived materials rather than or as a substrate for inert materials. References 1. Burleson, R. and Eiseman, B.: Effect of Skin Dressings and Topical Antibiotics on Healing of Partial-Thickness Skin Wounds in Rats, Surg. Gynecol. Obstet., 136:958, 1973.
Ann. Surg.
*
November 1976
2. Burleson, R. and Eiseman, B.: Mechanisms of Antibacterial Effect of Biologic Dressings, Surgery, 72:315, 1972. 3. Burleson, R. and Eiseman, B.: Nature of the Bond Between Partial-Thickness Skin and Wound Granulations, Ann. Surg., 177:181, 1973. 4. Eade, G. G.: The Relationship Between Granulation Tissue, Bacteria and Grafts in Burned Patients, Plast. Reconstruct. Surg., 22:42, 1958. 5. Kornberg, J., Burns, N. E., Kafesjian, R. and Bartlett, R. H.: Ultra Thin Silicone Polymer Membrane: A New Synthetic Skin Substitute, Trans. Am. Soc. Artif. Int. Organs, 18:39, 1972. 6. Tavis, M. J., Harney, J. H., Thornton, J. W., et al.: Modified Collagen Membrane as a Skin Substitute: Preliminary Studies, J. Biomed. Mat. Res., 9:285, 1975. 7. Tavis, M. J., Thornton, J. W., Harney, J. H., et al.: Adherence to De-epithelialized Surfaces: A Comparative Study, Surg. Forum, 25:39, 1974. 8. Tavis, M. J., Harney, J. H., Thornton, J. W. and Bartlett, R. H.: A Comparative Study of Burn Wound Management in Rats. Submitted to Surgery, 1976. 9. Thornton, J. W., Tavis, M. J., Harney, J. H., et al.: The Adherence of Collagen Grafts to Wound Surfaces: The Role of Fibrin. Submitted to Science, 1976.
