INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Giardiasis is by far the most common enteric parasitic infection in the United States, with an estimated 500 thousand to 1 million cases occurring annually (5). Until relatively recently, definitive diagnosis of this infection was dependent upon microscopic examination of stool specimens for the characteristic cyst and trophozoite forms of Giardia lamblia. Unfortunately, the sensitivity of conventional ovum-and-parasite (O&P) examination on a single stool specimen for G. lamblia has been shown in several studies to be less than optimal (1, 4, 9, 11) and in a recent study conducted in our laboratory was determined to be only 74% (3). The poor sensitivity of a single O&P examination for diagnosing giardiasis is primarily due to intermittent or low-level shedding of parasites by infected individuals (4) and is one of the principal reasons that parasitology textbooks and laboratory manuals recommend that multiple, independently collected stool specimens be obtained for O&P examination (6, 7).

The commercial availability of enzyme immunosorbent assays (EIAs) for detecting Giardia-specific antigens in stool specimens has provided a potentially attractive alternative to conventional O&P examination for diagnosing giardiasis. A number of clinical evaluations of Giardia EIAs have found them to be rapid and, perhaps more importantly, cost-effective tools for diagnosing infection with G. lamblia (1, 2, 8, 9, 11, 12, 14). The sensitivity of EIA in these studies has been at worst comparable (8, 14) and in most cases somewhat superior (1, 2, 9, 11, 12) to O&P examination. Indeed, for those laboratories that serve patient populations in whom the prevalence of pathogenic enteric parasites other than G. lamblia approaches zero, routine use of Giardia EIA rather than O&P examination has been suggested (2).

G. lamblia is the most commonly identified pathogenic enteric parasite in specimens submitted to the Hennepin County Medical Center (HCMC) laboratory for O&P examination, with an annual per-specimen prevalence rate of approximately 10% (3). This high prevalence of G. lamblia infection is in part attributable to the institution's role in mandated health screening of newly domiciled refugees, many of whom are asymptomatically infected with intestinal parasites. The issue of how to most cost-effectively utilize Giardia EIA testing in a high-prevalence setting was, therefore, the primary motivator for this study, since most of the studies published to date comparing Giardia EIA with O&P examination have not assessed whether EIA testing of a single specimen results in a diagnostic yield comparable to that obtained if multiple stool samples are examined microscopically. We sought to answer this question by prospectively comparing the sensitivity of a commercial Giardia EIA with routine O&P examination in the high-prevalence population served by HCMC.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Between February and June 1999, 206 stool samples collected from 103 patients attending HCMC or its outlying clinics were evaluated for the presence of G. lamblia by both conventional O&P examination and the ProSpecT Giardia Microplate assay (Alexon-Trend Inc., Ramsey, Minn.). To be included in the evaluation, patients had to have two stool samples submitted to the laboratory for O&P examination that had been independently collected within a 7-day time period. All specimens were received in the ParaPak ULTRA Stool System (Meridian Diagnostics, Cincinnati, Ohio), consisting of stool preserved in 10% formalin in one vial and stool fixed in polyvinyl alcohol in the other. Formalin-preserved material was concentrated prior to examination by use of a formalin-ethyl acetate sedimentation technique as recommended by the manufacturer of the specimen collection kit. Concentrates were read as wet mounts by examining the entire area under a 22- by 50-mm coverslip, using a 10× objective for screening and 40× objective for parasite identification. Permanently stained preparations of stool specimens were made with the material preserved in polyvinyl alcohol by use of Wheatley's (13) modified trichrome stain (Meridian Diagnostics). Trichrome-stained smears were read by examining at least 300 fields using a 100× oil immersion objective.

Giardia EIA testing was performed on unconcentrated formalin-preserved specimens using the manufacturer's recommended procedure. The EIA testing was performed in a blinded manner; specimens were assigned unique identifiers and unlinked from conventional O&P results. Upon completion of assay procedures, absorbance values were measured using a spectrophometer and interpreted using criteria provided by the manufacturer.

In the absence of a true "gold standard" for determining the diagnostic sensitivities of the two test procedures, specimens meeting the following criteria were considered to be positive for G. lamblia: (i) positive by O&P examination and positive or negative by Giardia EIA and (ii) negative by O&P examination and positive by Giardia EIA if the specimen was positive by Giardia EIA upon repeat and (a) the other specimen in the pair was positive by O&P examination or (b) the other specimen in the pair was negative by O&P examination but repeatedly positive by Giardia EIA or (c) the other specimen in the pair was negative by Giardia EIA but the patient was clinically symptomatic and was treated for giardiasis.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Of the 206 stool samples tested, 54 specimens (26.2%) collected from 30 patients (29.1%) were determined to be positive for G. lamblia. Conventional O&P examination detected 88.9% (48 of 54) of the positive specimens, while the Giardia EIA exhibited a sensitivity of 96.3% (52 of 54). Two O&P-negative specimens that were positive on initial testing with the Giardia EIA were negative upon repeat testing, and since the other sample in each pair was negative by both Giardia EIA and conventional O&P examination and neither patient was symptomatic, these were considered false-positive results. The specificity of the Giardia EIA was, therefore, determined to be 98.7% (150 of 152).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

A number of evaluations of the ProSpecT Giardia EIA have been published in the peer-reviewed literature (1, 2, 8, 9, 11), with reported sensitivities varying from 91% (1) to 100% (2) and specificities of 98% (1) to 100% (2). In most of these studies, Giardia EIA was compared on a direct per-specimen basis with microscopic examination of stool specimens, by conventional O&P examination either alone (1, 2, 11) or in combination with direct fluorescent antibody staining (8). In our evaluation, Giardia EIA detected 52 of 54 specimens determined to be positive for G. lamblia using O&P examination plus clinical criteria as the gold standard, for a diagnostic sensitivity of 96.3% (conventional O&P examination had a sensitivity of 88.9%). Two specimens were deemed to have given false-positive Giardia EIA results, and thus the specificity of the test in our hands was 98.7%. Our data substantially confirm, therefore, those of previous studies in demonstrating that the ProSpecT Giardia EIA is a sensitive and specific alternative to conventional O&P examination for diagnosing giardiasis.

By examining paired, independently collected stool samples, we determined the diagnostic sensitivity of a single O&P examination to be 66.7% (20 of 30 cases detected). This is similar to the sensitivity of 74% reported for a previous study conducted in our laboratory (3) and comparable to sensitivities reported by other investigators (1, 11, 12). This relative lack of sensitivity of O&P examination, which presumably is reflective of low parasite numbers or intermittent shedding of organisms, means that at least two independently collected stool specimens need to be submitted for O&P examination to obtain a diagnostic sensitivity of greater than 90%. A sensitivity of 93.3% was achieved in the present study by microscopically examining two stool samples for parasites. Given that the sensitivity of the Giardia EIA, at least on a per-specimen basis, is higher than that of O&P examination, it seemed conceivable that the sensitivity of a single Giardia EIA might approach the sensitivity of paired-stool-sample O&P examination. In the present study, however, the sensitivity of Giardia EIA on a single sample was 80% (24 of 30 cases detected), which is a considerable improvement over conventional O&P examination of a single specimen (14%; four additional cases), but not equivalent in sensitivity to O&P examination of two stool specimens. The improvement in sensitivity effected by using Giardia EIA rather than conventional O&P examination on a single specimen was somewhat greater in asymptomatically infected individuals undergoing routine health screening (17%; three additional cases) than in individuals with symptoms of intestinal giardiasis (8%; one additional case); however, in neither population was a single EIA comparable in sensitivity to O&P examination of paired specimens.

Only one of the previously published evaluations of Giardia EIAs used analysis of paired, independently collected stool samples to evaluate the comparative sensitivity of conventional O&P examination and EIA (9). Mank et al. (9) determined the sensitivity of a single O&P examination to be 80%, with microscopic examination of two stool samples increasing the diagnostic yield to 96.4%. Use of a Giardia EIA improved the sensitivity of a single-stool-sample analysis to 92.7%, with a negative predictive value substantially equivalent to that achieved by microscopically examining two specimens (98.7 versus 99.4%). Based on these findings, those authors concluded that a single Giardia EIA could replace multiple O&P examinations for patients at minimal risk for infection with parasites other than G. lamblia. Interestingly, the 12.7% increased sensitivity (92.7 versus 80%) of single-sample Giardia EIA testing over single-sample microscopic examination reported by Mank and colleagues (9) was essentially matched by the 13.3% increase (80 versus 66.7%) observed in our study. The primary difference in findings between the two studies was in the sensitivity of a single O&P examination. In our study, in only 66.7% of documented cases were both stool samples positive for G. lamblia by O&P examination, whereas in the study by Mank et al. (9), G. lamblia was detected microscopically in both samples on 80% of occasions. A possible explanation for this considerable difference in sensitivity of G. lamblia detection by O&P examination is that while all the specimens examined in the previously published study were obtained from patients with symptoms of intestinal giardiasis, in the present study only 40% (41 of 103) of patients from whom specimens were obtained were symptomatic. Indeed, in our study, the sensitivity of a single O&P examination for patients exhibiting symptoms of giardiasis was 75%, compared with only 61% in individuals asymptomatically infected with G. lamblia (Table 1). Nevertheless, unlike in the previous study (9), even in symptomatically infected patients the sensitivity of a single Giardia EIA was not comparable to that of O&P examination of two stool samples (83.3 versus 100%).

In conclusion, the results of our study confirm previous findings, demonstrating the superior sensitivity of Giardia EIA testing to conventional O&P examination in both symptomatic and asymptomatic individuals. The improvement in sensitivity achieved by detecting G. lamblia antigen was, however, insufficient to achieve parity between single-specimen testing with Giardia EIA and multiple-specimen evaluation by conventional O&P examination. In settings where the prevalence of enteric parasitic infection is low and where the primary pathogen identified is G. lamblia, routine testing of a single stool specimen by conventional O&P examination has been recommended (10). Based on our findings and on those of Mank and colleagues (9), utilization of Giardia EIA as the routine means of enteric parasite detection appears to be a preferable approach in such environments, since it would improve diagnostic yield, shorten turnaround time, and decrease labor cost. In higher-prevalence settings, however, or for patients who are at increased risk for giardiasis, the negative predictive value of the ProSpecT Giardia EIA on a single stool sample is not sufficiently high to exclude the possibility of G. lamblia infection. To evaluate patients for whom the level of clinical suspicion for G. lamblia infection is moderate or high and infection with other intestinal parasites is low, therefore, we recommend that two stool samples be submitted to the laboratory for Giardia EIA testing, with analysis of the second sample being performed if the first assay yields a negative result.