The relative sensitivities of a commercially available enzyme immunoassay (EIA) (ProSpecT Giardia; Alexon-Trend Inc., Ramsey, Minn.) and conventional ovum-and-parasite (O&P) examination for the detection of Giardia lamblia in preserved stool specimens were determined. Paired stool samples collected independently within a 7-day period from 103 patients were analyzed by both methods. A total of 54 specimens from 30 patients (18 asymptomatically infected with G. lamblia and 12 with symptoms consistent with intestinal giardiasis) were determined to be positive for G. lamblia, of which 48 (88.9%) were positive by microscopy and 52 (96.3%) were positive by EIA. Both specimens submitted were positive for G. lamblia by O&P examination for 66.7% (20 of 30) of the positive patients; for 26.7% (8 of 30) a single specimen was positive by O&P examination, and for 6.7% (2 of 30) of those determined to be infected with G. lamblia, both samples were negative by microscopy. The sensitivity of conventional O&P examination was somewhat higher in symptomatically infected individuals, with 75% (9 of 12) of patients in this category having G. lamblia detected in both samples, compared with 61% (11 of 18) of asymptomatic patients. A total of 24 positive patients (80%) had G. lamblia antigen detected by EIA in both submitted samples, 4 positive patients (13.3%) had one specimen positive by EIA, and the EIA was negative in both specimens from 2 infected individuals (6.5%), the sensitivity of EIA was substantially equivalent in asymptomatic and symptomatic individuals (77 versus 83% of patients with positive results on both specimens). Although the sensitivity of EIA for the detection of G. lamblia on a single stool specimen was somewhat higher than that of conventional O&P examination in symptomatic patients (83 versus 75%), in asymptomatic patients (77 versus 61%), and overall (80 versus 67%), examination of two specimens by either EIA or microscopy was necessary to achieve a diagnostic sensitivity of greater than 90%.