CLINICAL OBSERVATIONS

POLYVINYL SPONGE IMPLANTS: EXPERIMENTAL AND CLINICAL OBSERVATIONS By CUTHBERT E. DUKES, M.D., M.Sc., F.R.C.S., and BERNARD C. V. MITCHLEY, M.I. Biol., A.I.S.T. Chester Beatty Research Institute, Institute of Cancer Research,! Royal Cancer Hospital, London POLYVmyL sponge is porous, resilient, and relatively inert. For these reasons it is sometimes of use to plastic surgeons because, when implanted subcutaneously, it causes only very slight inflammation and foreign-body reaction. Grindlay and Waugh (I95I), in their pioneer work on the surgical uses of plastic sponge, were struck by the apparent inertness of this plastic and commented on this unusual.characteristic in these words :-- " It is not possible to explain the mechanism by which living tissue adopts polyvinyl sponge and uses it as a framework for fibrous tissue or to explain why the body does not surround the sponge by a protective barrier... Perhaps tissue fails to recognise polyvinyl sponge as a foreign body, because tissue fluids enter it. Cells follow fluid, and what was inert becomes living." Our own experiments with polyvinyl sponge implanted subcutaneously in rats have confirmed the fact that it almost seems to be " adopted ".as if a natural,constituent of the body and that fibroblasts soon begin to grow into the interstices of the sponge, being accompanied by a stream of phagocytic cells and by sprouting blood capillaries. The phagocytes tend to attach themselves to the solid framework of the sponge forming multinucleated giant cells (Fig. I). The fibroblasts begin to form collagen which increases in density as time passes (Fig. 2), and nodules.of ossification may develop here and there (Fig. 3). A point of practical importance, to which further reference will be made later, is that tiny particles of the sponge may break off and work their way, or be transported elsewhere. Such migrating particles may be found buried in neighbouring fat or fibrous tissue (Fig. 4). Implants of polyvinyl sponges have often been used by plastic surgeons,.especially during the past ten years. Occasionally the implants have subsequently been removed, because of either trauma or infection, but hitherto there has been no evidence that they have given rise to any form of neoplasm in man. Edgerton and McClary (I958), who made special enquiries with regard to this, were unable to find any reports of cancer arising in an area adjacent to a plastic sponge, but they added" the final chapter in the use of this material cannot be told for a number of years." Similar statements to the effect that polyvinyl sponge has not been reported to cause cancer in man have been made by Hamit (I957), Clarkson (I96O), and Harris (I96I). At first experimental work on laboratory animals also seemed to support the view that implanted polyvinyl sponge would not be carcinogenic. Thus Brown and Moore (1952) found no evidence of any tumour growth in fifty mice which were watched for ten months after implantation of sheets and fragments of Ivalon polyvinyl sponge. Neuman (1956) reported similar negative results in 3 A 225

226 BRITISH JOURNAL OF PLASTIC SURGERY FIG. I Section through margin of sponge thirteen months after implantation subcutaneously in rat, to show infiltration of interstices of sponge by phagocytes and connective tissue cells. Note absence of fibrous capsule round sponge. A, Subcutaneous tissues ; B, Sponge. (H~matoxylin and Eosin.) x IOO. FIG. 2 Section through central regions of sponge eighteen months after implantation subcutaneously in rat, to show dense collagenous fibrous tissue occupying interstices of sponge. (H~cmatoxylin and Eosin.) x Ioo.

EXPERIMENTAL AND CLINICAL OBSERVATIONS IN POLYVINYL SPONGE IMPLANTS 227" FIG. 3 Section through sponge removed twenty-four months after implantation subcutaneously in rat to show ossification in the fibrous tissue which has~ invaded the sponge. (Hmmatoxylin and Van Gieson.) Ioo. Fro. 4 Section through connective and adipose tissue adjacent to a sponge removed twenty months after implantation subcutaneously in rat, to show tiny particles o f sponge broken off and transported into neighbouring tissues. Particles of sponge marked by arrows. (H~ematoxylin and Eosin.) Ioo.

228 BRITISH JOURNAL OF PLASTIC SURGERY a series of experiments on rats. In fact, polyvinyl sponge implants seemed so apparently harmless that they have been recommended as a way of collecting normal connective tissue for biochemical and enzyme analysis (Boucek and Noble, 1955) and for the quantitative analysis of deoxyribonucleic acid and.collagen (Woessner and Boucek, 1961). Thus there is ample evidence that polyvinyl sponge is remarkably well tolerated at first, but the object of the present paper is to discuss the long-term effects when FIG. 5 Structure of Ivalon sponge as revealed by section and microscopic examination. (Hmmatoxylin and Van Gieson.) 13. this plastic is implanted subcutaneously in rats. The research we are reporting was initiated by the late Professor Eric Horning as a natural continuation of his previous work on the subcutaneous implantation of plastic films (Alexander and Homing, 1959). In this he had noticed that a correlation existed between carcinogenicity of implanted plastic films and the actual size of their surface area. Therefore in planning these new experiments with plastic sponge he decided to compare the long-term effects of "thin" sponges (depth 2 ram.) with "thick " sponges (depth 5 ram.). Professor Homing died in November 1959, soon after this experiment had been planned. We have undertaken and completed it and are reporting our findings because our results may be of importance in relation to the current use of polyviny! sponge by plastic surgeons. ExperimentaL--Commercially produced plastic sponge consisting of cross-linked polyvinyl alcohol (sold under the name of " Ivalon ") was cut into squares 2 by 2 cm., of ~wo thicknesses, namely 2 and 5 ram. Each piece of sponge was thoroughly extrae~d.with water and alcohol and sterilised by autoclaving at 15 lb. for twenty minutes, The structure as revealed by section and microscopic

EXPERIMENTAL AND CLINICAL OBSERVATIONS IN POLYVINYL SPONGE IMPLANTS 229 examination is illustrated in Fig. 5. The sponges were implanted subcutaneously into the right flank of 7o-day-old Chester Beatty stock rats. Twenty rats were used for implantation of" thin" and twenty for "thick " sponges. RESULTS I. " Thin "Sponges (2 by 2 cm. by 2 mm.).--two of the twenty rats implanted with "thin " sponge died within twelve months. The remaining eighteen animals were kept under observation until they became sick or developed a palpable tumour FIG. 6 Section through portion of" thick " sponge ten months after implantation subcutaneously in rat. A sarcoma arising outside the sponge has spread inwards, replacing the pale connective tissue occupying the interstices of the sponge by denser and darker malignant growth. A, Sponge fiued with " connective tissue" ; B, Spread of sarcoma into sponge replacing "connective tissue "; C, Primary focus of sarcoma outside sponge. (Hmmatoxylin and Van Gieson.) 5"5. at the site of implantation of sponge. They were then killed and material was removed for microscopic examination both from the implanted sponge and surrounding tissues. Only one malignant turnout was found--a spindle-cell sarcoma in a rat implanted twenty-two months previously. 2. " Thick "Sponges (2 by 2 cm. by 5 mm.).--six of the twenty rats implanted subcutaneously with thick sponge died within ten months of implantation. The remaining fourteen were killed when they developed a palpable tumour at: the site of implantation. In each case the palpable turnout proved to be a sarcoma. In each case also it was clear from the dissection and microscopic examination that the sarcoma had arisen from the tissues adjacent to the sponge, and in many the malignant tumour had subsequently spread by direct continuity into the sponge (Figs. 6, 7, and 8). This experiment is summarised in the table (p. 231).

~230 BRITISH JOURNAL OF PLASTIC SURGERY FIG. 7 S a m e case as Figs. 6 a n d 8 to show histology of spindle-cell sarcoma arising outside t h e sponge. ( H m m a t o x y l i n a n d Eosin.) 17o. FIG. 8 Same case as Figs. 6 a n d 7 to show spread of s a r c o m a into sponge, replacing t h e connective tissue w h i c h h a d previously filled its interstices. (Hmmatoxylin a n d Eosin.) x IOO.

EXPERIMENTAL AND CLINICAL OBSERVATIONS IN POLYVINYL SPONGE IMPLANTS TABLE Incidence of Sarcomata in Rats at Site of Subcutaneous Implantation of Ivalon Sponges of Two Different Sizes 231 Number of Rats. Size of Sponge. Survivors at Time of First Tumour (ten months). Average Survival Time of Rats living for Ten Months or More. Number of Rats which developed Sarcomata. 20 20 20 5 mm. 14 t 14 months 14 20 20 x2o 2 mm. I8 2o months I DISCUSSION Two important conclusions may be drawn from this experiment: (I) that polyvinyl sponge may eventually be carcinogenic for rats if left long enough in the tissues, and (2) that the degree of carcinogenicity is related to the thickness of the implanted sponge. We do not propose to discuss the possible explanations of this carcinogenicky. It has been attributed either to simple repeated trauma or to a specific (unrecognised) " carcinogen." These alternatives have been examined by Hueper (I96I), who concluded that "the physical state of polymers is carcinogenically important inasmuch as k favours an intimate contact of the polymers with body fluid and cells, thereby permkting a prolonged and direct chemical interaction between implants and tissues, which accounts for the formation of cancers from polymers." Putting on one side, then, the still obscure problem of why polyvinyl sponge causes sarcoma in rats, we shall confine ourselves only to the very practical question as to whether or not the proved carcinogenicity for laboratory animals implies that there may be some danger of human cancer following the use of pieces of polyvinyl sponge as implants for such purposes as to remedy amastia or hypomastia. Against such a possible danger of cancer following plastic implants, it might be argued that a variety of different chemical and physical agents are carcinogenic for rats and other laboratory animals, though hitherto there is no evidence that they have caused cancer in man. Secondly, the pieces of plastic sponge used in these experiments differed in size and shape from implants such as are used by plastic surgeons. They were pieces of sponge measuring 2 cm. square and either 2 or 5 ram. thick, with sharp edges and corners, and would naturally cause much more local trauma than carefully fashioned and moulded implants. These are two arguments against the likelihood of human cancer occurring as the result of the employment of polyvinyl plastic implants, but, on the other hand, we must not overlook the fact that the initial tissue response to implanted plastic in man seems to be of a strikingly similar character to that observed in the rats used in our experiments. We are able to say this because we have had the opportunity of examining eight polyvinyl sponge implants after their removal from patients. These implants had remained in poskion for periods ranging from four to eighteen months. The eight implants were derived from four patients (aged 22, 28, 48, and 54 years).

:232 BRITISH JOURNAL OF PLASTIC SURGERY In each cas e we found that the plastic sponge had been infiltrated by connective tissue to a depth of about half an inch : in the deeper regions the sponge contained tissue fluid and cellular debris only. In the portion of sponge infiltrated by connective tissue the histological picture closely resembled that already described in our experimental rats. Thus Fig. 9 is a low'power view of a section through the margin of an Ivalon sponge implanted between the mammary gland and pectoral muscle of a woman aged 22. It was removed ten months later because the patient complained that it had shrunk and become hard FIG. 9 Ivalon sponge removed ten months after implantation between mammary gland and pectoral muscles of woman aged 23, to show infiltration of interstices of sponge by living tissue. (Framework of sponge appears jet black.) (Hmmatoxylin and Van Gieson.) x 17. and uncomfortable. The interstices of this portion of the sponge were filled with collagen-producing fibroblasts, phagocytes, and multinucleated giant cells, nourished by new capillaries (Fig. io). A thin encircling rim of apparently normal tissue was attached to some parts of the implant, and sections through this showed collections of phagocytes and foreign-body giant cells which appeared to be attempting to dislodge particles of the sponge (Fig. II). We found some of these particles of sponge in a portion of the mammary gland overlying the implant (Fig. 12). (In our experiments with the implantation of plastic sponge subcutaneously in rats, we had noticed that particles might break off and migrate into surrounding tissues and have illustrated this phenomenon in Fig. 4.) We did not find any evidence of neoplasm in the eight human implants we have examined, and as far as we have been able to ascertain no case of human cancer attributable to a plastic sponge implant has hitherto been reported. None the less, we think a careful watch should be kept for this, and whenever an implant

EXPERIMENTAL AND CLINICAL OBSERVATIONS IN POLYVINYL SPONGE IMPLANTS 233; FIG. IO High-power view of same case as Fig. 9 to show collagen-producing fibroblasts, phagocytes, and capillaries in interstices of sponge. (Framework of sponge jet black.) (H~matoxylin and Van Gieson.) IOO. FIG, I I Section through margin of Ivalon implant in same case as Figs. 9 and IO. Note that there is no "fibrous capsule" around implant and that collections of phagocytes and foreign-body giant cells appear to be attempting to displace particles at margin of sponge. A, Collection of phagocytes at margin of sponge. (H~ematoxylin and leosin.) x 85.

234 BRITISH JOURNAL OF PLASTIC SURGERY is removed it should be sectioned and examined microscopically. It can be fixed in formalin, dehydrated, and sectioned as if it were an ordinary pathological specimen. One reason for carrying out such an examination is because, if a malignant tumour were to develop from tissues at the margin of the sponge, it is probable that it would soon spread by direct continuity into the interstices of the sponge. We found that in rats which developed sarcomas there was evidence of rapid spread into the sponge as though this were the path of least resistance (see Fig. 6). It even looked as if the connective tissue in the interstices of the sponge was exceptionally fertile soil for malignant growth (see Fig. 8). FIG. I2 Same case as Figs. 9, IO, and Ii. Section through mammary gland overlying implant to show particles of sponge, broken off and embedded in mammary gland. A, Mammary gland. Arrows indicate particles of sponge. (Hmmatoxylin and Eosin.) 85. Our experimental work with the subcutaneous implantation of plastic sponge has proved that this substance may be carcinogenic for the rat, and that the degree of this carcinogenicity is dependent on the size of the implant and its duration of residence in the host. The immediate reaction to an implanted sponge seems to be similar both in man and the rat, and it is natural to ask, therefore, if the remote effect is likely to be different. This is a question which cannot be answered at the present time. Further observations are necessary, both experimental and clinical. In the meantime the best advice which could be given to a plastic surgeon is to use sheets of sponge as thin as possible and as porous as possible, and to remove them (if this also is possible) when their purpose has been served. We realise that plastics, such as polyvinyl sponge, have many uses in surgery and may be used more extensively in the future, and we hope that this paper may stimulate further study of the reaction of host tissues to plastic sponge implants.

EXPERIMENTAL AND CLINICAL OBSERVATIONS IN POLYVINYL SPONGE IMPLANTS 235 SUMMARY Experiments are reported which prove that polyvinyl sponge when implanted subcutaneously in rats may eventually give rise to a local sarcoma. The degree of carcinogenicity of this plastic depends on the size and thickness of the implanted.sponge. Eight polyvinyl sponge implants have also been examined microscopically after their removal from human patients and the histology compared with that of plastic sponge implanted subcutaneously in rats. The question is discussed as to whether or not this proved carcinogenicity for laboratory animals implies that there may be some danger of human cancer following the implantation of pieces of polyvinyl sponge. We are much indebted to Mr Patrick Clarkson, F.R.C.S., and other surgeons who have supplied us with implants for examination from their patients, and to Mr K. Moreman for help with photography. The work was supported by grants from the Medical Research Council, British Empire ~Cancer Campaign, United States Public Health Service, and the Anna Fuller Fund. REFERENCES ALEXANDER, P., and HORNING, E. S. (1959). " Carcinogenesis: of Actionfl' p. I~. London : Ciba Foundation Symposium. Mechanism BOUCEK, R. J., and NOBLE, N. L. (1955). Arch. Path., Chicago, 59, 553. BROWN, J. B., and MOORE, A. M. (1952). Plast. reconstr. Surg., io, 451. CLARKSON, P. (196o). Proc. R. Soc. Med., 53, 880. EDGERTON, M. T., and MCCLARY, A. R. (1958). Plast. reconstr. Surg., 2i, 279. GRINDLAY, J. H., and WAUGtt, J. M. (1951). Arch. Surg., Chicago, 63, 288. HAMIT, H. F. (1957). Arch. Surg., Chicago, 75, 224. HARRIS, H. I. (I96I). ft. int. Coll. Surg., 35, 630. HUEPER, W. C. (1961). Path. Microbiol., 24, 77. NEUMAN, Z. (1956). Brit. ft. plast. Surg., 9, 195. WOESSNER, J. F., and BOUCEK, R. J. (1961). Arch. Biochem., 93, 85.