Considerations on the use of nucleic acid-based amplification for malaria parasite detection.

International audienceUNLABELLED: ABSTRACT: BACKGROUND: Nucleic acid amplification provides the most sensitive and accurate method to detect and identify pathogens. This is primarily useful for epidemiological investigations of malaria because the infections, often with two or more Plasmodium species present simultaneously, are frequently associated with microscopically sub-patent parasite levels and cryptic mixed infections. Numerous distinct equally adequate amplification-based protocols have been described, but it is unclear which to select for epidemiological surveys. Few comparative studies are available, and none that addresses the issue of inter-laboratory variability. METHODS: Blood samples were collected from patients attending malaria clinics on the Thai-Myanmar border. Frozen aliquots from 413 samples were tested independently in two laboratories by nested PCR assay. Dried blood spots on filter papers from the same patients were also tested by the nested PCR assay in one laboratory and by a multiplex PCR assay in another. The aim was to determine which protocol best detected parasites below the sensitivity level of microscopic examination. RESULTS: As expected PCR-based assays detected a substantial number of infected samples, or mixed infections, missed by microscopy (27 and 42 for the most sensitive assay, respectively). The protocol that was most effective at detecting these, in particular mixed infections, was a nested PCR assay with individual secondary reactions for each of the species initiated with a template directly purified from the blood sample. However, a lesser sensitivity in detection was observed when the same protocol was conducted in another laboratory, and this significantly altered the data obtained on the parasite species distribution. CONCLUSIONS: The sensitivity of a given PCR assay varies between laboratories. Although, the variations are relatively minor, they primarily diminish the ability to detect low-level and mixed infections and are sufficient to obviate the main rationale to use PCR assays rather than microscopy or rapid diagnostic tests. The optimal approach to standardise methodologies is to provide PCR template standards. These will help researchers in different settings to ensure that the nucleic acid amplification protocols they wish to use provide the requisite level of sensitivity, and will permit comparison between sites