How many stool examinations are necessary to detect pathogenic intestinal protozoa?

Abstract

Recent reports have suggested that examination of one stool specimen is sufficient to detect protozoa in symptomatic patients. To verify this hypothesis, we examined specimens submitted to the regional laboratory of a large, prepaid health plan that provides primary care to a membership representative of the general population in northern California. We examined the number and results of stool specimen examinations in the two-month period before detection of Entamoeba histolytica in 125 infected persons, of Giardia lamblia in 154 infected persons, and of Dientamoeba fragilis in 60 persons infected with this parasite. About 13% of the entire group had a single stool specimen examined, 19% had two examined, 65% had three examined, and 6% had more than three examined. Most infections were detected in the first specimen or specimens submitted, but many were not detected until later examination. Thirteen patients with E. histolytica (10.4%) required 4-9 examinations. The sensitivity of one examination was compared with that of three examinations; with the additional examinations, the yield increased 22.7% (95% confidence interval [CI] = 11.8-33.5%) for E. histolytica, 11.3% (95% CI = 6.9-81.8%) for G. lamblia, and 31.1% (95% CI = 17.7-54.4%) for D. fragilis. Our results suggest that a single stool specimen examination will miss many pathogenic protozoan infections in symptomatic persons.