cally possible in less developed settings (Nickerson et al., 1990). The technique offers a hypothetical sensitivity limit of a single malaria gene per specimen, but in practice it requires 10 to 100 gene copies.
Problems with contamination from parasite DNA that was not originally in the sample and difficulties in specimen processing need to be resolved before PCR can be used routinely as a diagnostic tool. In addition, personnel who perform the test must be highly trained, and the costs of PCR thermal cyclers and materials remain high. Improvements in PCR technology, such as rapid sample preparation, packaging of reagents to prevent cross-contamination, and the development of nonradioactive signal systems, are bringing the technique closer to the stage where it will be of practical use. The PCR technique has been used to diagnose malaria infection with DNA primers specific for P. falciparum and P. vivax (W. J. Martin, University of Southern California Medical Center, unpublished data, 1991). By sequencing the dihydrofolate reductase gene associated with pyrimethamine resistance, specific primer sequences have been identified that react with the mutant gene but not the wild-type gene. These sequences potentially could be used for clinical diagnosis, identification of the parasite to the species level, and determination of pyrimethamine resistance.
Unless results can be accurately quantified and correlated with parasitemia, PCR may not be the test of choice in areas where malaria is endemic and where the majority of inhabitants harbor low levels of parasites without disease. The optimal use of PCR may be in sophisticated clinical laboratories in developed countries that have expertise in using complex technology but are unable to maintain competency in malaria microscopy. PCR may also be useful as an epidemiologic tool. For example, the technique could be used to determine the extent of falciparum malaria in an area of mixed infection and to identify drug-resistant strains in patient specimens or in mosquitoes.
Serological assays can detect antimalarial antibodies but cannot determine whether the antibodies result from current or past infection. Therefore, such assays are not appropriate for diagnosis but can be used for certain epidemiologic applications (Voller and Draper, 1982).
Several assay methods, including indirect immunofluorescence, hemagglutination, ELISA, and avidin-biotin-peroxidase complex ELISA, have been tested for their ability to diagnose malaria. Assays have been developed to measure antibodies to crude parasite antigens (Demedts et al., 1987; Sato et al., 1990); antigamete antibodies, which may influence parasite infectivity (Mendis et al., 1987); antibodies to merozoite surface anti-