ABSTRACT
Malaria parasite infections that are only detectable by molecular methods are highly prevalent and represent a potential transmission reservoir. The methods used to detect these infections are not standardized, and their operating characteristics are often unknown. We designed a proficiency panel of Plasmodium spp. in order to compare the accuracy of parasite detection of molecular protocols used by labs in a clinical trial consortium. Ten dried blood spots (DBSs) were assembled that contained P. falciparum, P. vivax, P. malariae, and P. ovale; DBSs contained either a single species or a species mixed with P. falciparum. DBS panels were tested in 9 participating laboratories in a masked fashion. Of 90 tests, 68 (75.6%) were correct; there were 20 false-negative results and 2 false positives. The detection rate was 77.8% (49/63) for P. falciparum, 91.7% (11/12) for P. vivax, 83.3% (10/12) for P. malariae, and 70% (7/10) for P. ovale. Most false-negative P. falciparum results were from samples with an estimated ≤ 5 parasites per µl of blood. Between labs, accuracy ranged from 100% to 50%. In one lab, the inability to detect species in mixed-species infections prompted a redesign and improvement of the assay. Most PCR-based protocols were able to detect P. falciparum and P. vivax at higher densities, but these assays may not reliably detect parasites in samples with low P. falciparum densities. Accordingly, formal quality assurance for PCR should be employed whenever this method is used for diagnosis or surveillance. Such efforts will be important if PCR is to be widely employed to assist malaria elimination efforts.