Sensitivity and specificity of microscopic and molecular techniques for the diagnosis of taeniasis

Abstract

Background and Objectives

The diagnostic challenges associated with T. solium continue to hamper control efforts of the world's most significant foodborne parasite and leading cause of epilepsy in low and middle-income countries. This study aimed to validate two conventional PCRs for taeniasis and estimate the diagnostic performance of microscopic and molecular tools.

Methods

Formalin and ethanol-fixed samples were tested by formalin-ethyl acetate concentration technique (FECT), Malachite smear, McMaster2 method, rrnS PCR and cox1 PCR. Initial validation of PCR methods was completed on 45 microscopy positive individuals. After validation, the performance of microscopic methods and the rrnS were estimated using samples from 1,156 individuals in Laos. Bayesian latent class models (BLCMs) and a composite reference standard were used to estimate diagnostic sensitivity and specificity.

Results

On preliminary validation the rrnS was able to detect 27/45 (60.00 %) infections whereas the cox1 detected 21/45 (46.67 %). As a result, the cox1 was excluded from further performance analysis. Microscopy methods and the rrnS were highly specific with estimates above 99.02 % regardless of analytical method. The rrnS was the most sensitive test by informed BCLM (91.45 %, CrI: 73.41–99.52 %) followed by the FECT (71.20 %, CrI: 50.53–85.48 %), McMaster2 (51.31 %, CrI: 32.00–71.29 %) and Malachite smear (32.23 %, CrI: 15.40–54.47 %).

Discussion

The inability to validate the cox1 PCR suggests that it may not be suitable in its current form for routine characterisation of Taenia spp. detected by microscopy. The rrnS presents a suitable alternative to the cox1, however, requires its products to be sequenced. Given the low prevalence of taeniasis in most populations, this should be a feasible approach that may be able to be integrated with existing soil-transmitted helminth surveys that often use FECT for microscopic diagnosis.