Seroepidemiology of schistosomiasis japonica by ELISA in the Philippines. II. Unreliability of stool examination in the measurement of incidence.

Abstract

It was shown in a previous study that a single quantitative stool examination (by technics currently used in the Philippines) is so sensitive that it underestimated by 50% the prevalence of schistosomiasis japonica in children detected by a single serologic assay (ELISA). This paper proves that measurement of incidence of the infection among these same children by stool examination is unreliable as well. Three important sources of error in the measurement of incidence by stool examination are: (a) the inclusion of already infected children in the group of presumably uninfected children in which conversion (change from negative to positive) is being measured; (b) the insensitivity of stool examination in identifying those who do acquire infection among the initially uninfected; and (c) spontaneous, possibly temporary, stool reversion. An overestimation of incidence results from error (a) because already infected children have 5-7 times the tendency to convert than do uninfected children; error (b) tends to produce an underestimation of incidence since stool examination will detect only about one-half of the children who do become infected; while error (c) also tends to cause an underestimation of incidence and is dependent on the frequency of stool examinations--34% of the conversions observed by multiple examinations may not be detected in a single year-end examination. The equivalent sources of error are insignificant when ELISA is employed to measure incidence. In the 3 years of observation, the annual incidence of schistosomiasis japonica measured by ELISA showed no significant change (13.5%, 16.7%, and 15.6%); on the other hand, the annual incidence measured by stool examination showed a significant reduction from 37.5% down to 16.0% and 15.1%. Significant sources of error invalidate measurements by stool examination; it is therefore concluded that incidence measurements have failed to produce acceptable evidence of reduction in transmission in the area of study in the 3 years of control operations. We believe that insistence on relying almost exclusively on the parasitologic technic can result in an erroneous evaluation of the control program currently being implemented.