Autoimmune Disease and Collagen Dermal Implants

IN RESPONSE:

We vigorously disagree with Drs. Singh and Fries. Our study does not fully establish causality but is not biologically implausible. Bovine collagen injections produce delayed-type hypersensitivity reactions in 3% of patients and humoral immune responses in another 10% [1]. Immunization with type II collagen of bovine, human, chick, or rat origin induces autoimmune arthritis in rats [2], mice [3], and monkeys [4]. Other inducible autoimmune diseases include experimental allergic encephalomyelitis, thyroiditis, orchitis, and uveitis [5]. With only 322 patients participating in the 1985-86 survey by Lyon and colleagues, the probability of finding a patient with dermatomyositis or polymyositis who also had collagen exposure was less than 0.32.

The two patients who received skin tests only were included because they were exposed to the same collagen product through the same intradermal route as were treated patients. Expansion of the estimated collagen-exposed population to account for slight differences in the ever-tested and ever-treated groups does not significantly alter the analytical results. As explained in our Methods section, the study period was limited by the availability of adverse reaction reports that only covered the period from July 1980 through June 1988. Lacking a collagen patient registry, we cannot be certain that no new cases occurred in the period between July 1988 and June 1991.

Singh and Fries criticize our use of incidence rates for dermatomyositis or polymyositis adjusted to the second decade of the 20-year study by Oddis and colleagues, rather than to the last 5-year period (which had the highest rate of all periods). However, our adjusted rates considerably exceeded the generally reported incidence rates for the disorders of 2.2 to 5.5 cases per million persons and probably overestimated the expected number of cases for the collagen-exposed group by a factor of 1.5 to 3.8. Our underascertainment from passive reporting of adverse reactions far exceeded Oddis and colleagues' underascertainment of cases in their exhaustive, hospital-based, medical records review. Reported incidence rates for dermatomyositis or polymyositis of 5.0 (1947 to 1968) and 5.5 (1963 to 1982) suggest that background rates for the disorders have probably remained relatively stable over time. Our methods are standard epidemiologic techniques for disease cluster investigations.

The FDA ad hoc panel based its same-day conclusion solely on published data after listening to oral presentations and testimonials. The Centers for Disease Control and Prevention and the FDA had previously spent months conducting a much more thorough assessment of the new data. Both agencies concluded that a significant statistical association exists between bovine collagen and dermatomyositis or polymyositis. The FDA now requires disclosure of this association in the warnings on the package insert.

Dr. Lewy makes a good point. Although our collagen implant study focused on rates of the disorders, which seemed particularly elevated, evidence of other autoimmune diseases existed among the exposed population; these included scleroderma, system lupus erythematosus, and rheumatoid arthritis. As more studies are done on the collagen-exposed population, we may discover a much broader spectrum of autoimmune diseases.

Although many case series have provided anecdotal evidence of a possible association between silicone breast implants and various autoimmune diseases, no studies have yet measured incidence rates; Several major differences between autoimmune diseases thought to be associated with silicone breast implants and those associated with bovine collagen implants are readily apparent. Whereas specific humoral or cell-mediated immune responses to bovine collagen are present in our patients with dermatomyositis or polymyositis who received collagen, the presence of “silicone antibodies” in patients with breast implants has yet to be proven. Although dermatomyositis or polymyositis was diagnosed approximately 6 months after collagen implants, “human adjuvant disease” occurs about 15 years (range, 2 to 25 years) after implantation. Nevertheless, the FDA has severely restricted the use of silicone breast implants but has not yet taken any definitive steps to restrict the cosmetic uses of bovine collagen.

In conclusion, we believe that the strong temporal association combined with evidence of biological plausibility suggest that the fivefold increased risk for dermatomyositis or polymyositis among collagen recipients during the period shortly after exposure is not likely to be purely coincidental.

• Copyright ©2004 by the American College of Physicians