Drs. Margolick and Vogt raise questions about the comprehensiveness of the immunologic battery and the accuracy of specific laboratory techniques. We agree that more extensive and more technically sophisticated evaluation might be preferable. Our choice of immunologic evaluation was pragmatic: We used the exact "panel" of studies typically done by environmental physicians for assessment of chemically sensitive patients and on which a diagnosis of multiple chemical sensitivity is often based. Although these data may have some limitations in the study of immunologic mechanisms, they clearly argue against the utility of immunologic assessment as currently practiced.

We thank Dr. Stricker for his suggestion regarding IgG subclass, IgA, and complement testing, and may pursue it.

Drs. Krohn and Jacobson and Ms. Ryan raise questions about selection of cases and controls. We can address the three specific issues raised. As stated, controls were screened for symptoms of chemical sensitivity, and two patients were excluded. Controls were not, however, screened out because of previous chemical exposure. To do so would have been inappropriate in a study designed to assess the relation between chemical sensitivity and chemical exposure. Although research psychiatric interviews have important limitations, we do not agree that research on chemical sensitivity should ignore the domain of psychological distress, given previous research. Their other concerns regarding confounding, bias, and failure to match on "important factors" are too vague. Extensive data were collected on exposure history, occupational history, and severity of illness, and we would be happy to respond privately to more specific questions.

We appreciate Ms. Coxe's point that for a significant minority of patients with multiple chemical sensitivity, psychological symptoms are clearly not "central".

Both Dr. Byers and Dr. McCampbell question how funding from the Boeing Company may have influenced our research. Partial support for laboratory expenses was provided by the International Association of Machinists/Boeing Health and Safety Institute. This institute is operated jointly by Boeing and its largest union to promote research and education in occupational health. Grants for independent research are awarded after review by a committee of labor and management representatives. Given the controversy surrounding chemical sensitivity, we argue that representation of both workers and management minimizes the potential for bias.

Dr. McCampbell questions whether immunologic findings might have differed if analyzed according to severity of chemical sensitivity or according to the proportion of patients with abnormal results. Neither of these approaches produces a different picture. More severe chemical sensitivity was not associated with greater immunologic abnormality and the two groups (cases and controls) did not differ significantly in the proportion of patients with abnormal findings (using laboratory reference standards). As for the dating of previous symptoms, we agree that recall of remote symptoms is imprecise, but recall is the only available measure. Our analyses were based on patients' own reports of the onset of chemical sensitivity and of individual symptoms.

Finally, we stand by our comments regarding avoidance. Our intent was to highlight the lack of evidence supporting avoidance of low-level chemical exposures. To date, no controlled studies show any physiologic injury in chemically sensitive patients that might be exacerbated by low-level chemical exposure. In our opinion, the disability and isolation caused by avoidance regimens argue strongly for a rehabilitative approach.