The articles by Cukier and colleagues [1] and Bridges and colleagues [2] and the superb editorial by Hochberg [3] raise two further questions: Can an adequate control group for such toxicity and environmental studies possibly exist, and have we completely defined the proposed toxic syndrome yet? If the definition is too narrow, so will be the control group chosen. According to our preliminary data on 177 consecutive women with breast implants (silicone, "saline," or polyurethane, and those with or without collagen and silicone injections), controlling for these factors would make the comparison group too small for the most aggressive epidemiologist.

Total implant years of the group were 1764, with a mean use of 10.95 years (±5.33 years). In the study by Bridges and colleagues [2], the average age of implant recipients was 44.14 years (±8.75 years). Even here, 8% had been additionally exposed to polyurethane, whereas 8% had had breast cancer (some with subsequent postoperative radiotherapy and chemotherapy), 7% had previous fibrocystic disease, and still others had had collagen or silicone injections. The most common symptoms identified were fatigue (84%) and arthralgias (83%), but extremity numbness (82%), neuropathic burning (88%), and memory loss (66%), which may have been neurologic manifestations, were as common as rheumatologic symptoms.

We conclude that these patients may be difficult to control for statistically because they are heterogeneous in their toxin exposures, basic health before implants, numbers and types of implants, diversity of medications, and surgical history. As an aside, the "natural" entity with which to compare implant recipients may be the demyelinating diseases rather than the collagen vascular group. We need to define this syndrome further to better select our controls (and to ensure that our patients are not caught between the diagnostic criteria of neurologists and those of rheumatologists).