Comparison of Sensitivity of Immunofluorescent Microscopy to That of a Combination of Immunofluorescent Microscopy and Immunomagnetic Separation for Detection of Cryptosporidium parvum Oocysts in Adult Bovine Feces

Abstract

A direct immunofluorescence assay (DFA) (Merifluor; Meridian Diagnostics, Inc., Cincinnati, Ohio) was compared to an immunomagnetic separation (IMS) assay (Dynabeads; Dynal, Inc., Lake Success, N.Y.) coupled with immunofluorescent microscopy (Waterborne, Inc., New Orleans, La.) for their ability to detect low concentrations of Cryptosporidium parvum oocysts in adult bovine fecal material. IMS-DFA resulted in a 2-log-unit increase in sensitivity (10 oocysts/g) compared to DFA alone (1,000 oocysts/g). The higher sensitivity obtained with IMS-DFA resulted from testing 2 g of fecal material instead of the 13 to 19 mg of fecal material tested in the DFA; the increased sensitivity was not attributable to a higher percent recovery.

Cryptosporidium parvum is a primary enteric pathogen which can infect a wide variety of mammalian hosts (6). Clinical cryptosporidiosis and asymptomatic oocyst shedding is typically more common in younger individuals within a host population (2–4, 10, 15, 18, 19). The primary source(s) of C. parvum oocysts for infection of newborn and young individuals remains unidentified. Given the prolonged contact between postparturient mammalian females and their young (12), it is plausible that postparturient females serve as the primary source of infection for their young. A transient rise in shedding of C. parvum has been reported for postparturient ewes (19) but has not been observed for other livestock species, such as dairy cattle (3). We have determined previously that a diagnostic test will need to detect concentrations as low as 1 to 10 oocysts/g if we are to reliably measure the percentage of postparturient cattle shedding low levels of C. parvum oocysts and determine what role such shedding plays in the epidemiology of calfhood infection (3).

Previous evaluations of direct immunofluorescent microscopy (DFA) with bovine fecal samples have shown that this procedure can detect oocyst concentrations as low as 1,000 to 6,000 oocysts/g (16, 17). Xiao and Herd (17) used calf fecal samples to determine the sensitivity of DFA. Given that calf fecal material has a higher fat content than adult bovine fecal material, the use of calf feces in DFA may not be representative of the performance of DFA with adult bovine fecal material, since the high fat content of human stools has been shown to reduce the sensitivity of immunofluorescence microscopy by several fold (13, 14).

We reevaluated the sensitivity of DFA for adult cattle fecal material and examined whether immunomagnetic separation (IMS) could improve the sensitivity of DFA for detecting low concentrations of oocysts in adult bovine feces. The ability of IMS to concentrate oocysts either failed to improve the sensitivity of PCR for bovine fecal samples (8) or improved sensitivity by 1 to 2 log units (16), which allowed the detection of 80 to 100 oocysts/g of feces. This level of sensitivity is insufficient for detection of oocyst shedding by periparturient cattle; therefore, an improved method is required. An evaluation by Rochelle et al. (11) indicated that oocyst recovery efficiencies were as high as 67% for IMS coupled with DFA for seeded bovine fecal samples, suggesting that a combined procedure may generate the requisite sensitivity for identifying low levels of oocyst shedding in adult cattle.

C. parvum oocysts were purified from feces from naturally infected dairy calves by using a previously described technique (1), and the concentration of oocysts was determined with a hemacytometer. A series of 4.5-g bovine fecal samples containing no detectable C. parvum oocysts were spiked with 0.5-ml aliquots of a C. parvum suspension to yield final concentrations of 100, 1,000, and 10,000 oocysts/g of feces. A negative control was prepared with 0.5 ml of distilled water in 4.5 g of C. parvum-negative fecal matter. After overnight incubation at 4°C, the samples were washed, sieved, and concentrated according to previous protocols (2, 4). Oocysts were detected by using the Merifluor Cryptosporidium/Giardia detection kit (Meridian Diagnostics, Inc., Cincinnati, Ohio) DFA with six replicates per concentration, according to the manufacturer’s instructions. Slides were weighed prior to and immediately after a 10-μl loopful of the 5-ml fecal suspension was placed onto the well. The Merifluor assay was chosen for this study because we have used it in previous epidemiologic studies of C. parvum infection in cattle populations (3, 4).

Percent recovery was calculated as n/(wc), where n is the number of oocysts counted in the smear, w is the weight of the smear (in grams), and c is the number of oocysts per gram of fecal suspension. Given that the amount of material applied to each slide well with a 10-μl loop was not constant, we measured the effect of smear weight on the number of oocysts recovered per smear by using Poisson regression (7).

Three different fecal processing procedures were evaluated prior to IMS. We also evaluated the accuracy of IMS for detecting C. parvum oocysts in two types of bovine feces, one with no magnetic particles and another which contained 20 to 50 mg of magnetic particulate matter after 2 g of the sample had been sieved. Although the exact percentage of cattle feces containing magnetic material is unknown, we have observed magnetic material in bovine feces on several occasions (unpublished data); therefore, we undertook to determine whether the presence of magnetic material reduced oocyst recovery and whether removal of this particulate matter prior to IMS would improve recovery (5). Oocyst enumeration was performed by using the Dynabeads anti-Cryptosporidium assay (Dynal, Inc., Lake Success, N.Y.) and a fluorescein isothiocyanate-labeled anti-Cryptosporidium monoclonal antibody assay (Waterborne, Inc., New Orleans, La.) according to the manufacturer’s instructions. We chose Dynal as the source of immunomagnetic beads based on the higher recovery efficiencies of Dynal compared to Crypto Scan (Clearwater Diagnostics, Portland, Maine) for oocysts spiked into water concentrates (5, 11).

Bovine fecal material from three different farms and testing negative for C. parvum was spiked with C. parvum oocysts to yield final concentrations ranging from 10 to 1,000 oocysts/g of feces. The negative control had only sterile, distilled water added. Each sample was resuspended in 40 ml of phosphate-buffered saline, strained through a 4- by 4-inch piece of gauze for sieved protocols, and centrifuged for 10 min at 1,000 × g. Supernatants were discarded, and pellets were resuspended in 10 ml of sterile, distilled water and transferred to Leighton tubes for IMS.

To determine if removal of magnetic particles prior to IMS would improve the percent recovery, fecal samples were suspended in 30 ml of phosphate-buffered saline, and the heavier particulate matter was allowed to settle for 5 min. Supernatants were collected, strained through cotton gauze, and centrifuged for 10 min at 1,000 × g. Supernatants were discarded, and pellets were resuspended in 10 ml of sterile, distilled water and transferred to Leighton tubes for IMS according to the manufacturer’s directions (Dynal, Inc.). Given that an entire 2-g sample of fecal material was tested per assay, percent recovery was calculated as n/(2c), where n is the number of oocysts counted in the smear, 2 represents the 2-g sample of tested fecal material, and c is the number of oocysts per gram of the spiked fecal aliquot.

Using Poisson regression (7), we analyzed whether sieving the sample increased the number of recovered oocysts compared to not sieving, whether the presence of magnetic material reduced the number of recovered oocysts, and whether removing magnetic material prior to IMS increased the number of recovered oocysts. The likelihood ratio statistic was used to determine if the indicated effect (i.e., the procedure or fecal constituent) significantly altered the expected number of oocysts recovered per assay. We used a P value of ≤0.05 for significance.

With a single smear weighing from 13 to 19 mg, the DFA procedure recovered an average of 5 and 54 oocysts/smear at concentrations of 1,000 and 10,000 oocysts/g, respectively (Table 1). This 1-log-unit difference in the number of detected oocysts was significantly different (P, <0.001). The mean percent recoveries for DFA at 1,000 and 10,000 oocysts/g were not significantly different, at 28 and 34%, respectively. No oocysts were detected in samples containing either 0 or 100 oocysts/g of feces. Assuming that oocysts were randomly distributed in the fecal pellet and that oocyst counts in fecal smears are Poisson distributed, and based on our observation that the mean oocyst count was 5 per smear for a concentration of 1,000 oocysts/g, the probability of detecting oocysts at a concentration of ≥1,000 oocysts/g was ≥99.3% (1 − e⁻⁵ = 0.993) (9). The probability of detecting oocysts with a 16-mg smear remained above 90% for concentrations as low as 600 oocysts/g (9). This level of sensitivity is much higher than those previously reported for analyses of either human or bovine fecal samples. Previous studies using the same DFA as our study reported detectable concentrations of 5,000 to 10,000 oocysts/g for human fecal samples and 4,000 to 6,000 oocysts/g for bovine fecal samples (13, 14, 16). This lower sensitivity of DFA for human stools likely resulted from their higher fat content and the subsequent fat-removal procedures used, both of which likely reduce the probability of recovering oocysts from the fecal matrix compared to recovering oocysts from the low-fat, watery fecal matrix characteristic of adult cattle manure (13, 14).

TABLE 1.

Percent recovery of C. parvum from adult bovine feces by DFA

No. of oocysts added/g of stool	Mean weight (range) of smear (mg)	Mean no. (range) of detected oocysts	Mean % recoverya (range)
0	16.8 (12.5–22.2)	0	0
100	NDb	0	0
1,000	15.8 (4.5–19.5)	5 (3–7)	28 (18–49)
10,000	15.5 (12.6–18.6)	54 (34–63)	34 (19–50)

^aPercent recovery was calculated as n/(wc), where n is the number of oocysts counted in the smear, w is the weight (in grams) of the fecal smear, and c is the number of oocysts per gram in the spiked fecal sample. 

^bND, not determined. 

Smear weight was negatively correlated with percent recovery (P, <0.001), likely the result of oocysts becoming obscured by the feed particles that accumulate at higher smear weights. This impact on the sensitivity of DFA for fecal smears ranging from 13 to 19 mg may help explain the lower sensitivity observed by Xiao and Herd (17) with their assay, which used 20 μl of fecal suspension per smear. In our study, the application of 20 mg of fecal suspension per smear would have resulted in percent recoveries as low as 15%, consistent with the 14.8% reported by Xiao and Herd (17).

IMS-DFA permitted the detection of oocysts in concentrations as low as 10 oocysts/g of feces (Table 2). By using parameters from the Poisson regression model to assess the results obtained for fecal samples sieved prior to IMS, the probability of detecting oocysts at a concentration of ≥10 oocysts/g was determined to be ≥93% (9). The mean percent recovery after sieving a sample which contained no magnetic particles was 35%, compared to a mean percent recovery of 25% for samples that had not been sieved. The mean percent recovery after sieving a sample which contained magnetic particles was 23%, compared to 20% for samples whose magnetic particles had been removed prior to IMS. By using the total number of oocysts recovered from a sieved sample with no magnetic particles as the reference value for Poisson regression, we determined that not sieving a sample resulted in only 57% (95% confidence interval, 53 to 62%; P, <0.001) of oocysts being recovered compared to sieving and only 58% (95% confidence interval, 50 to 70%; P, <0.001) of oocysts being recovered compared to sieving feces that contained magnetic material. Attempting to remove magnetic material prior to IMS did not increase the percent recovery (P, 0.93), similar to previous work on water concentrates in which the use of preclearing IMS beads for removal of magnetizable material not only failed to improve recovery but appeared to reduce recovery of oocysts (5). Regardless of the procedure used or the presence of magnetic particles, the percent recovery was negatively correlated with oocyst concentration, declining 7 to 12% as the concentration increased from 10 to 1,000 oocysts/g.

TABLE 2.

Recovery of C. parvum oocysts by Dynal IMS concentration of 2 g of adult bovine fecal material

Feces type	Treatment method	No. of oocysts added/g of feces	Mean no. of oocysts in 2 g of fecesa	Range of oocysts in 2 g of feces	Mean % recovery
No magnetic particles	Sieve	0	0
		100	96	88–104	48
		200	139	132–153	35
		600	336	263–465	28
		1,000	543	278–923	27
	Nonsieve	100	75	61–84	38
		1,000	255	124–348	13
Magnetic particles	Sieve and no magnetic treatment prior to IMS	0	0
		10	8	5–11	40
		50	24	15–30	24
		100	38	24–53	19
		1,000	209	123–311	10
	Sieve and magnetic treatment prior to IMS	10	6	2–14	32
		100	39	17–78	20
		1,000	147	43–231	7

^aMean of three fecal samples, each collected from a separate farm. 

The presence of large feed particles (due to not sieving) or magnetic particles in a fecal sample interfered with the ability of Dynal IMS beads to recover oocysts. Previous studies on the performance of Dynal IMS in water concentrates did not observe a strong association between turbidity and percent recovery (5, 11). Attempting to remove magnetic particles prior to IMS did not improve the percent recovery, similar to the observation that attempts to remove magnetizable material from water concentrates prior to IMS inadvertently removed oocysts from the sample (5).

In conclusion, our IMS-DFA protocol was approximately 1 log unit more sensitive than IMS coupled with PCR (8, 16) and resulted in a 2-log-unit improvement over the sensitivity of DFA alone. This higher sensitivity of IMS-DFA compared to DFA alone was not the result of a higher percent recovery for the combined technique, but resulted from concentrating 2 g of fecal material during IMS. Only 13 to 19 mg of fecal material was tested per smear for immunofluorescent microscopy (mean ≈ 16 mg). Hence, the 2-log-unit increase in sensitivity observed in IMS-DFA can be explained by the fact that IMS-DFA tests 125 times (2,000 versus 16 mg per smear) more fecal material than DFA. IMS-DFA has the requisite sensitivity to detect concentrations of C. parvum oocysts as low as 10 oocysts/g of adult bovine feces. Although the cost of reagents for IMS-DFA is quite high, the procedure is cost-effective in comparison to DFA alone when one considers that 2 g of sample is tested simultaneously in each smear. Moreover, IMS-DFA provides for relatively accurate quantification of the load of oocysts in adult cattle feces. This is essential if we are to develop reliable estimates of the rate of environmental loading of C. parvum attributable to cattle and if we are to determine the risk that cattle grazing poses to microbial water quality at the watershed scale.