Evaluating the diagnostic test accuracy of molecular xenomonitoring methods for characterising community burden of lymphatic filariasis

Background
Molecular xenomonitoring (MX), the detection of pathogen DNA in mosquitoes, is a recommended approach to support lymphatic filariasis (LF) elimination efforts. Potential roles of MX include detecting presence of LF in communities and quantifying progress towards elimination of the disease. However, the relationship between MX results and human prevalence is poorly understood.
Methods
We conducted a systematic review and meta-analysis from all previously conducted studies that reported the prevalence of filarial DNA in wild-caught mosquitoes (MX rate) and the corresponding prevalence of microfilaria (mf) in humans. We calculated a pooled estimate of MX sensitivity for detecting positive communities at a range of mf prevalence values and mosquito sample sizes. We conducted a linear regression to evaluate the relationship between mf prevalence and MX rate.
Results
We identified 24 studies comprising 144 study communities. MX had an overall sensitivity of 98.3% (95% CI 41.5, 99.9%) and identified 28 positive communities that were negative in the mf survey. Low sensitivity in some studies was attributed to small mosquito sample sizes (<1,000) and very low mf prevalence (<0.25%). Human mf prevalence and mass drug administration status accounted for approximately half of the variation in MX rate (R2 = 0.49, p<0.001). Data from longitudinal studies showed that, within a given study area, there is a strong linear relationship between MX rate and mf prevalence (R2 = 0.78, p < 0.001).
Conclusion
MX shows clear potential as a tool for detecting communities where LF is present and as a predictor of human mf prevalence.