Application of a composite faecal egg count procedure in diagnostic parasitology

Application of a composite faecal egg count procedure in diagnostic parasitology

veterinary parasitology ELSEVIER Veterinary Parasitology 52 (1994) 337-342 Short communication Application of a composite faecal egg count procedur...

342KB Sizes 0 Downloads 25 Views

veterinary parasitology ELSEVIER

Veterinary Parasitology 52 (1994) 337-342

Short communication

Application of a composite faecal egg count procedure in diagnostic parasitology Jason Nicholls, David L. Obendorf* Animal Health Laboratory, Tasmanian Department of Primary Industry, Mt. Pleasant Laboratories, P.O. Box 46, Kings Meadows, 7249 Tas., Australia

(Accepted 21 June 1993 )

Abstract Egg counts from a simple composite faecal counting procedure using equal amounts of sample from ten sheep were compared statistically against the arithmetic means of the same ten samples prepared by a conventional egg counting method. Forty separate data sets were analysed in an untransformed bivariate plot and after natural logarithmic transformation. A sign test analysis indicated a high degree of similarity between the two data sets. A confidence interval for the composite count (n = 10) was calculated to give a result between five eggs more and 15 eggs less than the arithmetic mean count of the ten samples. When multiple faecal samples are to be examined, the composite method has significant advantages in time saving and increased throughput whilst still providing an accurate resuit. This technique has been used to monitor gastrointestinal helminthosis and for faecal egg count reduction testing to assess anthelmintic efficacy. Key words: Sheep-Nematoda; Diagnosis-Nematoda; Egg counts

1. Introduction Q u a n t i t a t i v e assessment o f strongyloid o v a in the faeces o f sheep has been widely used to d e t e r m i n e the degree o f pasture c o n t a m i n a t i o n with h e l m i n t h eggs ( G o r d o n , 1967). In young sheep, o v a counts can also be used to predict w o r m b u r d e n s ( M c K e n n a , 1987; M c K e n n a a n d Simpson, 1987 ). Currently ova counts are being p e r f o r m e d to assess the effectiveness o f anthelmintics (Waller, 1989) a n d as a m e a n s o f m o n i t o r i n g parasite control programs. *Corresponding author. 0304-4017/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSD10304-4017 ( 93 ) 00560-L

338

J. Nicholls. D.L. Obendorf/ Veterinary Parasitology 52 (1994) 337-342

In recent years, the demand for parasitology testing services has increased considerably. This is primarily due to a greater awareness of the widespread occurrence of nematodes resistant to the major broad-spectrum anthelmintic groups (benzimidazoles and levamisole) and a need to use drenches more sparingly. Ova counting procedures which assess the number of nematode eggs per gram (epg) of fresh faeces are all based on a dilution counting procedure in which an aliquot of the faecal suspension from a known volume of sample is examined microscopically. Several methods and subsequent modifications have been reported; however, the McMaster method forms the basis of the most universally used technique in Australia and elsewhere (Whitlock, 1948 ). When testing many faecal samples daily, the McMaster method is both time consuming and labour intensive. Due to the variability of ova counts between individual sheep and the generally poor diagnostic value of single ova counts, an accepted flock monitoring approach has been to assess the ova count of several sheep in a flock (Gordon, 1967 ). This method has been successfully applied in the faecal egg count reduction (FECR) test to assess anthelmintic efficacy. In response to drenching trials, it is recommended that faecal samples from ten animals are tested individually. From the individual ova counts, the arithmetic or geometric mean can be calculated for the ten animals. When faecal samples are collected 10-14 days after anthelmintic treatment, the effectiveness of suitably administered drenches can be assessed by comparison of the mean ova counts from non-treated control and treated groups. From a statistical perspective, FECR tests based on group means also permit the calculation of percentage ova reduction within confidence intervals (Waller, 1989 ). With the introduction of diagnostic flock testing services for anthelmintic resistance and for monitoring sheep nematode control programs, it was necessary to consider alternative parasitological methods for obtaining estimates of mean egg counts. Combining faecal samples from several individuals before conducting an ova count has the benefit of substantially reducing the number of tests performed. This paper presents a method which uses individual faecal samples in a composite counting procedure. Analysis of the results obtained by this method are also compared with the arithmetic mean of the same samples by individual ova count tests.

2. Materials and methods

Forty sets of ten faecal samples were tested in this study. Individual ova counts were conducted on a 2 g (_+0.1 g) sample according to a modified McMaster method (Whitlock, 1948). For the composite sampling, 2 g ( + 0.1 g) of faeces, irrespective of consistency, from each of ten samples was placed into a Stomacher ® plastic bag (170 m m × 3 0 0 mm, Labco, Victoria) and 200 ml of water were added. The composite sample was thoroughly homogenised in a Stomacher ® laboratory blender (BA 6021 Steward Laboratories, UAC House, Black-

J. Nicholls, D.L. Obendorf / Veterinary Parasitology 52 (1994) 337-342

339

friars Rd., London SE1 9UG, UK) until all faecal solids were well dispersed. Using digital scales, a 44 g aliquot of the mixed composite suspension was removed and 76 ml of saturated NaC1 solution (specific gravity 1.19 ) were added. The subsample was mixed and four chambers (0.5 ml each) of a Whitlock Universal counting slide (Whitlock Universal Counting Chamber, 4 × 0 . 5 ml, J.A. Whitlock and Co., N.S.W., Australia) were filled using a sieve-ended tube. All the strongyloid eggs in the chambers were counted, summed and multiplied by 15 to express the reading as the epg of faeces.

3. Results A graph of the arithmetic means plotted against the composite egg count for each of the 40 faecal sample sets is shown in Fig. 1. Thirty-seven of the 40 data points lie on or near the equivalence line ( x = y ) . Approaching zero, the variation is greater but this is to be expected for a test which obeys Poisson distribution principles. A sign test analysis (Hettmansperger, 1984), a non-parametric procedure used to test whether the two data sets can be considered to be just random subsamples of the same parent population, yielded P = 0.268, indicating a high degree of similarity. The expression of the numerical data obtained by the two methods, namely an arithmetic mean of ten individual egg counts expressed in epg and a single composite egg count multiplied by 15, means that identical results for each set are unlikely (due to differing sensitivities). Nevertheless, the correlation is high, given the inherent inaccuracies of the quantitative counting

8OOO 6000 1

"

/ /

C C]

¢~ 2000

"---0 c- 1000

/"'

• ,gll •

ip

/

i

1000

2000

//~ 6000

8000

C o m p o s i t e court: Fig. 1. Regression line for the arithmetic means of ten individual egg counts plotted against the respective composite egg counts obtained from 40 sets of ten faecal samples.

340

J. Nicholls, D.L. Obendorf / Veterinary Parasitology 52 (1994) 337-342

10 111



03 © c-

I~e •

c-





"V @ ~

2 i @ 0

i

i

i

i

2

z.

6

8

Loge(som

Desire

coui'qt)

10

Fig. 2. Natural logarithm transformed regression for the arithmetic means of ten individual egg counts plotted against the respective composite egg counts obtained from 40 sets of ten faecal samples.

method. By extending this result to a larger population than the sample on which it is based, a confidence interval about the median difference can be calculated. It is predicted that a composite ova count by this method is likely to give a result between five eggs more and 15 eggs less than the arithmetic mean of the ten samples. Using a natural logarithmic transformation of the data set, the two distributions are remarkably alike (Fig. 2 ). There is a slight tendency for most points to lie just below the y = x line, i.e. arithmetic mean counts are evidently marginally higher than composite counts. Data points are most tightly aggregated about the line y = x at the higher values (more than 90 epg) and least so at the lower values (less than 90 epg). The three outlying data points relate to zero composite counts with corresponding arithmetic means of 3, 5 and 18. Variation is greatest between mean counts and composite estimates where the log~ (composite plus I ) falls below 4 (i.e. less than 60 epg).

4. Discussion

The above-mentioned method of composite faecal sampling for quantitative ova counting has been used in Tasmania for diagnostic flock monitoring using Wormtest kits (Regional Veterinary Laboratory, Benalla, Vic., Australia). These faecal sampling kits were designed for use on farm by sheep graziers. Faecal samples can be submitted to a laboratory for analysis in these kits. Results are reported directly to the grazier to allow for planning of drenching and assessment

J. Nicholls, D.L. Obendorf / Veterinary Parasitology 52 (1994) 337-342

341

of worm control on the property. Rapid and cost-effective information is available using this procedure. The kits can also be used to test for anthelmintic resistance by following the Drenchtest instructions (Regional Veterinary Laboratory, Benalla, Vic., Australia) and submitting another set of ten faecal samples, 10-14 days after drenching a mob of sheep. Using pre-treatment composite egg counts as a control (recommended count greater than 300 epg), a determination of drench efficacy can be given, based on the FECR of the post-drench sample. Adoption of this composite method has permitted this popular parasitology service to be offered through a diagnostic veterinary laboratory as a valuable feefor-service test. After 16 months of operation, the Wormtest/Drenchtest service is averaging 80 kits per month and all the sample preparation, counting and reporting is handled by one technician. This study gave confidence that the increased efficiency in processing of faecal samples did not jeopardise the accuracy of the results. The confidence limits for the composite of less than 15 and greater than five epg than the arithmetic mean of the ten samples are acceptable for the purpose for which this composite test was designed, namely FECR tests and monitoring of nematode control on sheep properties. The logarithmic transformation of this data set confirms that, over most of the range, the composite estimate is essentially equivalent to the mean count. However, the variation between the composite and the group mean estimates was greatest when the count fell below 60 epg. Another composite egg counting procedure was similarly evaluated by Baldock et al. (1990). This study was conducted in a summer rainfall zone where Haemonchus contortus and Trichostrongylus colubriformis were the predominant species. The composite method differed in that smaller masses of faeces (0.5 g) were used from three to five animal samples and only two chambers of a Whitlock Universal counting chamber were counted. In addition, the multiplication factors used to express counts as epg were higher (20-33). The egg counts reported by Baldock et al. (1990) are significantly higher than those reported here. The predominant species in the Tasmanian study were Trichostrongylus spp. and Ostertagia (Teladorsagia) circumcincta. The composite method described in this study used a standard number of samples (n = 10) and aimed at ensuring adequate amounts of sample (2 g) and minimising the epg multiplication factor ( 15 ) for each egg counted by counting four chambers of the slide (2 ml). Concern has been expressed as to the usefulness of composite sampling techniques for drench resistance FECR tests (Baldock et al., 1990). This method only provides one value and consequently does not allow for the calculation of confidence intervals when assessing the FECR percentage for a given test anthelmintic. This may be a disadvantage when assessing low grade resistance (i.e. the percentage FECR is close to the accepted cut-off of 95%). In these situations, the low sensitivity of this test may cause some FECR tests to be incorrectly rated. As with all responses to drenching trials using FECR tests, a high degree of care with field work is necessary to ensure that there are no missed dosings of sheep and, where several anthelmintics are being tested concurrently, that faecal samples are properly allocated to the correct groups for testing. In the case of a highly effective anthelmintic (i.e. greater than 95% FECR), the presence of an undosed

342

J. N icholls, D.L. Obendorf / Veterinary Parasitology 52 (1994) 337-342

individual in the group will not be detected by the composite sampling method. Depending on the effect this individual faecal sample has on the composite count, the apparent efficacy of the anthelmintic may be reduced. Given these precautions and reservations, this method has been employed in a recent state survey of anthelmintic resistance (N. Murray, J. Nicholls and D.L. Obendorf, personal communication, 1989). At the post-drench sampling, errors in drenching or allocation of faecal samples into test groups can lead to incorrect results and interpretation. In order to minimise this a further modification to the technique can be used. This involves making two composites, each containing five faecal samples. A 22 g subsample is taken from each and 38 ml of saturated NaC1 added. Two chambers of a counting slide ( 1 ml) are filled with aliquots from the two five-sample composites. If the number of eggs counted for each composite are comparable, the result can be accepted. If the counts are dissimilar, then individual counts are performed. In our experience, with tests where a drench was expected to be highly effective, then four individual counts return zero or very low egg counts whilst the one remaining sample has a high count.

Acknowledgement The authors would like to thank Bob Gittins for his assistance with the statistical calculations.

References Baldock, F.C., Lyndal-Murphy, M. and Pearse, B., 1990. An assessment of a composite sampling method for counting strongyle eggs in sheep faeces. Aust. Vet. J., 67:165-167. Gordon, H.McL., 1967. The diagnosis ofhelminthosis in sheep. Vet. Med. Rev., 67: 140-168. Hettmansperger, T.P., 1984. Statisical Inference Based on Ranks. Wiley, New York, pp. 3-5. McKenna, P.B., 1987. The estimation of gastrointestinal strongyle worm burdens in young sheep flocks: a new approach in the interpretation of faecal egg counts. I. Development. NZ Vet. J., 35: 94-97. McKenna, P.B. and Simpson, B.H., 1987. The estimation of gastrointestinal strongyle worm burdens in young sheep flocks: a new approach in the interpretation of faecal egg counts. II. Evaluation. NZ Vet. J., 35: 98-100. Waller, P.J., 1989. Anthelmintic Resistance. Report of the Working Party for the Animal Health Committee of the Standing Committee on Agriculture. SCA Technical Report Series - - No. 28. CSIRO Australia, East Melbourne, Victoria, 26 pp. Whitlock, H.V., 1948. Some modifications of the McMaster helminth egg counting technique and apparatus. J. Counc. Sci. Ind. Res. (Australia), 21:177-180.