Background: Current methods for detecting malaria parasites are invasive and associated with poor compliance when repeated sampling is required. New methods to detect and quantify parasites in a less-invasive manner would greatly enhance the potential for longitudinal surveillance in clinical trials.
Methods: Saliva, urine, and blood samples from 386 Gambian outpatients with suspected malaria infections were analyzed by nested polymerase chain reaction (nPCR) to detect infection and to evaluate diagnostic accuracy in comparison to expert microscopy. The amount of parasite DNA in malaria-positive samples was estimated using real-time quantitative PCR (qPCR).
Results: Blood parasite density as estimated by qPCR correlated well with parasite counts established by microscopy (p = 0.94; P < .001). qPCR results for saliva had a significant correlation with microscopy counts (p = 0.58; P < .001), whereas qPCR results for urine had a positive but poor correlation with microscopy counts (p = 0.20; P = .117). The mean amounts of parasite DNA quantified in blood were greater than the mean amounts quantified in saliva and urine samples obtained concurrently from the same individual, by approximately 600-fold and approximately 2500-fold, respectively. When nPCR results were compared with microscopy results, nPCR of saliva had a sensitivity of 73% and a specificity of 97%; its sensitivity increased to 82% in samples with a parasite density of > or = 1000 parasites/microL. nPCR of urine had a sensitivity of 32% and a specificity of 98%.
Conclusion: Saliva sampling is a promising less-invasive approach for detecting malaria infection.