Dann et al. (1) tested immune responses of volunteers exposed to Cryptosporidium oocysts using an enzyme-linked immunosorbent assay (ELISA). They found fecal immunoglobulin A (IgA) but no serum IgG2 response to infection. They concluded that in the naive individual, the parasite elicits a strong secretory antibody response but fails to provide adequate immune stimulation for serum antibody detectable by ELISA.

Since prior research indicates the ELISA, using unpurified antigen, may lack the sensitivity to detect a serological response to infection (2-4), we believe this conclusion is misleading. The implied conclusion that these infected volunteers did not respond contradicts findings of another study of the same volunteer group (5). Moss et al., using a Western blot assay, found a characteristic antibody response to either 15/17- or 27-kDa markers following infection in 10 of the 11 volunteers who experienced symptomatic illness (5). The Moss study also found serological response to Cryptosporidium antigens at day 0 in 21 of 29 volunteers (5), contradicting the assumptions of the Dann study that volunteers were immunologically naive at baseline. Volunteers immunologically naive by Western blotting excreted a higher number of oocysts following infection (5).

We tracked serological responses to infection, using a previously described Western blot assay (3), for one laboratory worker accidentally infected with C. parvum. During the 1.5 years prior to the accidental oocyst exposure, the worker had no IgA or IgM serological response to either the 15/17- or 27-kDa marker and no IgG response to the 15/17-kDa marker. He had a very weak IgG response to the 27-kDa marker that was only periodically detectable. He would, therefore, be classified as immunologically naive by Western blotting (5). Classical cryptosporidiosis developed 10 days following the exposure and 4 weeks later he developed an intense IgA, IgM, and IgG response to both the 15/17- and 27-kDa markers. The IgG and IgM responses peaked at 4 weeks postinfection and declined rapidly. IgM reached low levels by 22 weeks postinfection while IgA remained significantly elevated at 42 weeks postinfection. The IgG response peaked at 14 weeks postinfection. Response to the 15/17-kDa marker declined to near background by 42 weeks postinfection whereas response to the 27-kDa marker remained elevated at 71 weeks postinfection. If infection in immunologically naive individuals fails to provide adequate immune stimulation for serum antibody to develop, as suggested by Dann et al. (1), we should not have observed these classical serum antibody responses to infection.

Because of the large number of antigens contained in an oocyst, any assay based on aggregate serological responses to all of these antigens may be less sensitive than an assay, such as Western blotting, that examines responses to Cryptosporidium-specific antigens (2-4). Given published concerns with the ELISA, the failure of the authors to cite Western blot results for the same volunteers is puzzling. Their ELISA was clearly better able to detect secretory than serum responses to infection. This may have occurred either because the response was stronger or because of less cross-reacting secretory antibody.

This issue should not detract from the important and other supported findings of the study. Improved methods to detect infection may help better understand the epidemiology of Cryptosporidium infection.

• Copyright © 2001 American Society for Microbiology