A five-country evaluation of a point-of-care circulating cathodic antigen urine assay for the prevalence of Schistosoma mansoni

Abstract

We evaluated a commercial point-of-care circulating cathodic antigen (POC-CCA) test for assessing Schistosoma mansoni infection prevalence in areas at risk. Overall, 4,405 school-age children in Cameroon, Côte d'Ivoire, Ethiopia, Kenya, and Uganda provided urine for POC-CCA testing and stool for Kato-Katz assays. By latent class analysis, one POC-CCA test was more sensitive (86% versus 62%) but less specific (72% versus ~100%) than multiple Kato-Katz smears from one stool. However, only 1% of POC-CCA tests in a non-endemic area were false positives, suggesting the latent class analysis underestimated the POC-CCA specificity. Multivariable modeling estimated POC-CCA as significantly more sensitive than Kato-Katz at low infection intensities (< 100 eggs/gram stool). By linear regression, 72% prevalence among 9-12 year olds by POC-CCA corresponded to 50% prevalence by Kato-Katz, whereas 46% POC-CCA prevalence corresponded to 10% Kato-Katz prevalence. We conclude that one urine POC-CCA test can replace Kato-Katz testing for community-level S. mansoni prevalence mapping.

Figure 1.

Range and variation of Schistosoma mansoni prevalence of infection in 63 participant schools. Dark bars indicate prevalence as determined by standard Kato-Katz microscopy of the first daily stool specimen; superimposed light bars indicate prevalence measured using concurrent point-of-care circulating cathodic antigen (POC-CCA) detection in the first urine specimen. Bars are arranged in order of ascending prevalence based on the Kato-Katz results.

Figure 2.

Modeled estimation of the sensitivity of a single point-of-care circulating cathodic antigen (POC-CCA) in the five study countries (as shown in the variously thin or dotted lines as indicated in the Figure) or the five taken together (thick gray line), according to the “true” local Schistosoma mansoni prevalence of infection among 9- to 12-year-old children. Comparison is shown to results for a single Kato-Katz stool examination (thick black line). The model is based on logistic regression in which “true” disease status (yes or no) of each subject is treated as a binary latent variable, and the likelihood of schistosomiasis was hypothesized to be a function of age, gender, and community to which each subject belonged.

Figure 3.

Correlation of prevalence in individual schools (based on a single Kato-Katz examination) with that from a single point-of-care circulating cathodic antigen (POC-CCA) urine testing. Linear regression parameters are shown in the legend. Small black circles, Kenya; large triangles, Côte d'Ivoire; squares, Uganda; large circles with an X, Cameroon; large inverted triangles, Ethiopia. Vertical dotted lines represent school-age prevalence Kato-Katz thresholds used in the World Health Organization (WHO) guidelines for timing and coverage of mass drug administration (MDA)-based schistosomiasis control programs and horizontal dotted lines show the corresponding cutoffs for prevalence determined by the POC-CCA assay.