Scope and Criteria for the Toxicology Review
This chapter reviews studies of the toxicology of silicone compounds carried out over the past 50 years. It does not review immunological studies, except occasionally when immune system toxicology is part of a report covering other toxicology. Otherwise, immunological studies are discussed in Chapter 6. Silicone compounds include a great many chemical entities; a recent compilation lists toxicological data on 56 different siloxanes (Silicones Environmental Health and Safety Council, 1995 ). This chapter identifies silicone compounds as they are listed in individual reports, but it is organized by route of exposure not by type of compound. Silicone fluids, gels, and elastomers are covered since they are components of silicone breast implants.
Although the most relevant exposures are reviewed, that is, tissue injections and subcutaneous implants, the committee, unlike other recent reviews (Kerkvliet, 1998 ) also decided to include other (nonimplantation) exposure routes, such as dermal, oral, and inhalation, since data from such studies may provide some insights into systemic silicone toxicology. The committee included citations on the toxicology of silica in the reference list of this report, because there has been considerable mention of silica as a component of breast implant elastomers. However, the toxicology of silica is not reviewed here because the committee found no valid scientific evidence for the presence of or exposure to silica in tissues of women with breast implants. Some compounds not found in breast implants (and identified as such) are included briefly, sometimes to complete a survey of silicone species and other times because they have been mentioned in the current debate on the toxic effects of implants. It is important to note that toxicology studies often report silicone dose levels substantially in excess of any doses that could be achieved on a relative weight basis in women with silicone breast implants.
Earlier in this report, the committee emphasizes the relevance of published, peer-reviewed scientific reports and assigns secondary importance to technical reports from industry. In this chapter, however, studies done in-house by industry or by commercial testing laboratories have been analyzed. Such reports are often reviewed first in-house, then by the sponsor and panels of outside experts, and eventually by a regulatory agency, which also looks at original data. The conflict of interest inherent in experimentation by an organization with an economic interest in the outcome is recognized. Nevertheless, the committee found many of industry's technical studies informative, useful, and consistent with sound science. The studies cited here consisted of about 50 individual articles from the open scientific literature between 1948 and 1999 and about the same number of industry technical reports. Reviews available to the committee summarized data from some reports not reviewed by or not available to the committee. For example, the Silicones Environmental Health and Safety Council (1995) examined may reports on various organic silicon compounds that are not found in breast implants and reviewed some reports not accessible to the committee. This review was useful in presenting an overall picture of the generally low toxicity of silicones and identifying particular compounds that had toxicity. The report of the Independent Review Group (IRG, 1998 ) (and earlier versions of the Medical Devices Agency's work), and the report of the National Science Panel (Kerkvliet, 1998 ) which are described in Appendix C looked at essentially the same body of toxicology information as the committee. The IRG report included proprietary data not available to the committee, and as noted, the committee examined routes of exposure and listed silica references neither of which are included in the IRG or National Science Panel reports. Since the IRG, which had some proprietary data, concluded that silicones were bland substances with little toxicity, such data seem unlikely to have changed the committee's findings in any substantial way. Also, the committee believes that the inclusion of dermal, oral and inhalation toxicology studies in this report provided additional security in conclusions about the biological and toxicological behavior of relevant silicones.
Kerkvliet lists three major reasons why toxicology studies are helpful in assessing the safety of a drug or consumer product such as silicone breast implants. (1) Toxicology studies in animals may identify a hazard—that is, whether a given product can cause adverse health effects. (2) Studies may also clarify dose responses—that is, how much of an entity is necessary to produce effects. (3) Studies may provide mechanistic information—that is, how and under what circumstances an agent produces effects (Kerkvliet, 1998 ). Such studies, reviewed here, will not ''fulfill the manufacturers' responsibility to demonstrate the safety of. implants" as Kessler urged in 1992 (Angell, 1995), since unanticipated events cannot be predicted or complete safety proven. Accumulating qualitative and quantitative data on the general toxicity of silicones, however, allow a reasonable degree of confidence that silicone compounds in breast implants are not hazardous.
Brief History of Silicone Toxicology
The principles of safety evaluation have not changed much over the past 50 years. However, analytical tools, the ability to measure chemicals in the body, and the science of molecular biology, which allows association of complex changes in a few cells or molecules with various disease states, have advanced considerably. These advances affect evaluations of the toxicology of silicones over time and are reflected in more recent studies.
One of the first (if not the first) systematic evaluations of the toxicology of commercial silicones was conducted during World War II at the Dow Chemical Company. Silicone intermediates (chlorosilanes and ethoxysilanes) and selected commercial silicones were tested in rats, rabbits, and guinea pigs. The chlorosilanes and some ethoxysilanes were found to be highly corrosive; they represented significant industrial handling hazards. Methyl-and mixed methyl-and phenylpolysiloxanes, on the other hand, had very low toxicity. For practical purposes, they were divided into three groups: fluids, compounds, and resins. Five methylpolysiloxane and two methylphenylpolysiloxane fluids were tested (hexamethyldisiloxane, 0.35 centistoke [cS]; dodecamethylpentasiloxane, 2 cS;
DC 200 fluid, 50 cS; DC 550 fluid, 550 cS; DC 702 fluid, 35 cS; DC 200 fluid, 350 cS; and DC 200 fluid, 12,500 cS). None of these killed rats or guinea pigs when given orally at doses up to 30 ml/kg. Some of the fluids had laxative effects not unlike mineral oil. DC 200 fluid (50 cS) "seemed literally to flow through the animals." The fluid with the lowest viscosity (hexamethyldisiloxane, 0.65 cS) did not have a laxative effect, but produced some mild inebriation and subsequent central nervous system depression. This suggests that there might be some absorption of this compound from the gastrointestinal tract. Repeated administration of DC 200 oil (350 cS) by stomach tube, up to dose levels of 20 g/kg, did not produce gross signs of toxicity such as reduced weight gain, changes in organ weight, or organ pathology.
Intraperitoneal injection was well tolerated, except for hexamethyldisiloxane, which produced extensive adhesions within the peritoneal cavity. This compound also produced inflammation and necrosis at the sites of subcutaneous and intradermal injections and proved lethal on repeated intraperitoneal injections. Other silicone fluids in the peritoneal cavity elicited only reactions "typical. of an irritating foreign body" with nodules containing the fluid in the omentum and visceral peritoneum. Eye irritation was transitory and no skin irritation was observed with these fluids (Rowe et al. 1948 ).
Shortly after the report by Rowe et al. Kern et al. (1949) reported their results from feeding rats 0.05%-0.2% silicone-containing diets (a poly-dimethylsiloxane [PDMS], G.E. Dri-Film, No. 9977) and injecting silicone suspensions at unknown (but probably low) doses, intraperitoneally and intravenously in mice, and intra-and subcutaneously and in the muscles of rabbits. Hematological and gross and microscopic pathology examinations after 13 weeks were all normal, and the animals had no loss in body weight or other signs of toxicity (Kern et al. 1949 ).
Two silicone compounds (DC 4 Ignition sealing compound and DC Antifoam A) were examined. Both agents caused transient conjunctival irritation, but no corneal damage when introduced directly into the eyes. No skin irritation was seen. Feeding of Antifoam A at concentrations up to 1% to rats did not produce any untoward effects. In a six-month feeding study in dogs, Antifoam A also exhibited no toxicity (Child et al. 1951 ). Three types of silicone resins (DC 2102, a methylpolysiloxane, DC 993, a methylphenylpolysiloxane; and DC Pan Glaze, which was similar to DC 993) were evaluated. Acute oral administration of up to 3 g/kg in guinea pigs was not toxic (higher doses could not be administered), and intraperitoneal injection in rats or dermal application in rabbits produced no signs of irritation. Rats fed Pan Glaze at concentrations up to 3% for 50 days gained weight normally, and on microscopic examination, their organs did not show any signs of toxicity (Rowe et al. 1948 Rowe et al. 1950 ).
The studies described by Rowe et al. (1948) reflect state-of-the-art toxicity testing at that time. They were done in a respected laboratory by competent toxicologists. The untoward effects observed with some compounds did not alarm toxicologists. These effects were found only after exposure to high doses of the test agent. According to an old classification, substances with a probable human lethal dose in excess of 15 g/kg were considered practically nontoxic (Casarett, 1975 ). These investigators commented that "for the past few years, an attempt has been made to keep pace with the rapid development of these products so that toxicological information would be available upon which the health hazards of these materials could be evaluated." Only a few selected samples from each class of compounds were studied, but the experimental toxicology of silicone compounds did not yield data that suggested a need for fundamental, mechanistically oriented experimentation.
When these and some other early studies were reviewed in 1950, silicone fluids with a viscosity of 350 cS were described as having exceedingly low toxicity. Some animal toxicity tests, such as oral and subcutaneous administration and eye irritation, were even performed on one of the authors of this study (Barondes et al. 1950 ). By then-current standards of toxicology, silicone fluids had to be considered harmless, devoid of any obvious acute toxic potential, and thus presumably safe.
The Current Database
A recent review of silicone toxicology summarized a substantial database (Silicones Environmental Health and Safety Council, 1995 ). This document does not list any references which makes it impossible to determine whether the data were published or to discover when the studies were done. It is not possible, therefore, to evaluate adherence to modern good laboratory practice regulations, protocols, and procedural requirements. Carcinogenesis studies done before the mid-1970s had different protocols and procedural requirements than later studies and, by today's standards, must be considered less reliable. This may apply to other test systems as well. The Silicones Council review analyzed a total of 629 studies (see Table 4-1 ), more than half of them done with PDMS linears (Chemical Abstracts Service [CAS] No. 63148-62-9). Compounds that are of concern because a large number of people are exposed to them and because they are found in breast implants, that is, D4 and D5 (where D4 and D5 represent cyclic tetramer and pentamer, respectively), comprise 17% of studies. There are few chronic lifetime or carcinogenesis studies (less than 3%) and immunological studies (less than 5%). Acute and sub-acute toxicity and irritation studies are in the majority (57%). Some of the Silicones Council studies summarized briefly in this current database may also be reviewed subsequently in other parts of this chapter. As noted, this material presents an overall picture of silicone toxicity based on a general review of many data sources covering a wide variety of compounds. Specific studies on breast implant compounds are relied on by the committee for conclusions relevant to the safety of silicone breast implants, however.