The sensitivity of microscopy and PCR-based detection methods affecting estimates of prevalence of blood parasites in birds.

Inferences about the evolution of host-parasitic relationships are often made based on the prevalence of avian malaria, which is usually estimated in a large sample of birds using either microscopic or molecular screening of blood samples. However, different techniques often have variable accuracy; thus, screening methodology can raise issues about statistical bias if method sensitivity varies systematically across parasites or hosts. To examine this possibility, published information was collected on the prevalence of species in 4 genera of avian blood parasites ( Plasmodium, Haemoproteus, Leucocytozoon, and Trypanosoma) from various sources that used different tools. The data were tested to determine if the application of different methods provided different estimates for the same hosts. In these comparisons between the main methodologies, the PCR-based molecular methods were generally found to provide higher estimates for Plasmodium spp. prevalence than microscopic tools, while there was no significant tendency for such a trend in species of Haemoproteus and Leucocytozoon. When analyzing intraspecific variance of prevalence within molecular studies, some studies provided consistently higher estimates for Haemoproteus spp. prevalence than others, indicating that differences between studies can affect detected estimates. Within microscopic studies, surveys that examined more microscopic fields were more likely to report higher prevalence for Plasmodium spp. than those relying on fewer microscopic fields. Consequently, studies making comparisons across parasite genera and/or host species from different sources need to consider several types of bias originating from variation in method sensitivity.