- Email: [email protected]
Fecal shedding of Mycobacterium avium subsp. paratuberculosis by dairy cows
Veterinary Microbiology 107 (2005) 257–263 www.elsevier.com/locate/vetmic
Fecal shedding of Mycobacterium avium subsp. paratuberculosis by dairy cows Beate M. Crossley a, Francisco J. Zagmutt-Vergara a, Terry L. Fyock b, Robert H. Whitlock b, Ian A. Gardner a,* a
Department of Medicine and Epidemiology, School of Veterinary Medicine, University of California, Davis, CA 95616, USA b Johne’s Research Laboratory, New Bolton Center, University of Pennsylvania, 382 West Street Road, Kennett Square, PA 19348, USA Received 17 August 2004; received in revised form 21 January 2005; accepted 31 January 2005
Abstract Between 1982 and 2000, fecal samples were obtained from 786 cows that were shedding Mycobacterium avium subsp. paratuberculosis (Map). These cows were resident on 93 Pennsylvania dairies (mean herd size, 64 milk cows) that had no or minimal previous testing for Map. Feces were cultured on four tubes of Herrold’s egg yolk medium and the distribution of mean Map colony forming units (CFU) was evaluated. Most cows were light (<10 CFU/tube, 51.4%) or high (>50 CFU/tube, 30.8%) fecal shedders with fewer cows in the moderate category (10–50 CFU/tube). Of the 786 cows, 192 (24.4%) had colonies in only one of four tubes. In the multivariable negative binomial model, there were significant associations between mean CFU/tube and prevalence, herd size, and season and an interaction between herd size and season. The linear mixed model of continuous tube counts with a random herd effect yielded similar findings with associations with herd size as a continuous variable, season, and an interaction between categorized prevalence and continuous herd size. Variability in CFU/tube was greatest among cows in the same herd, intermediate for replicate tubes from the same cow, and smallest among cows in different herds. Reduction in the number of replicate tubes from four would have reduced the sensitivity of fecal culture for Map by approximately 6% (for three tubes) to 12% (for two tubes). # 2005 Elsevier B.V. All rights reserved. Keywords: Mycobacterium paratuberculosis; Colony forming units; Isolation; Cattle; Feces
1. Introduction
* Corresponding author. Tel.: +1 530 752 6992; fax: +1 530 752 0414. E-mail address: [email protected] (I.A. Gardner).
Mycobacterium avium subsp. paratuberculosis infection (Johne’s disease or bovine paratuberculosis) costs U.S. cattle producers up to $ 250 million annually (Ott et al., 1999). Mycobacterium a. paratuberculosis (Map) is spread mainly by the
0378-1135/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.vetmic.2005.01.017
258
B.M. Crossley et al. / Veterinary Microbiology 107 (2005) 257–263
fecal-oral route and infected cattle are categorized as being in one of three stages depending on fecal shedding of Map and clinical signs. Cattle in stage 1 are defined as prepatent and preclinical, and no organisms are detectable by culture. Stage-2 cattle are shedding detectable numbers of Map but without clinical signs. The shedding of Map organisms can be intermittent and detection by culture is imperfect, especially when few organisms are shed in feces (Merkal et al., 1968). Cattle in stage 2 often are further categorized as light (<10 colony forming units (CFU)/tube), moderate (10–50 CFU/tube), or high fecal shedders (>50 CFU/tube) based on culture results on conventional media such as Herrold’s egg yolk (Whitlock et al., 2000). Categorizations of mean CFU/tube with more levels have been used in some studies (Wells et al., 2002). Infected cattle with stage3 disease typically lose weight, have diarrhea that is not responsive to antibiotic treatment, and nearly always are culture and antibody positive (Whitlock et al., 2000). Cattle with clinical signs represent only a small fraction of the Map-infected animals in a herd (Sherman et al., 1990; Whitlock and Buergelt, 1996). Accurate detection of Map is dependent on the sensitivity and specificity of test methods and stage of infection, which affects fecal shedding of bacteria and antibody response to infection. Even though a ‘‘passthrough’’ phenomenon has been postulated for organisms in the intestinal tract (Sweeney et al., 1992) and some investigators believe that ‘‘passthrough’’ may be responsible for more low shedders in some heavily infected farms, fecal culture generally is considered to be 100% specific. Hence, fecal culture often is preferred to serologic testing for individual and herd classification of Map status (Colgrove et al., 1989; Sweeney et al., 1995). However, the procedures (Eamens et al., 2000; Pavlik et al., 2000), the distribution of colony counts (Bech-Nielsen et al., 1992; Giese and Ahrens, 2000; Kalis et al., 1999; Pavlik et al., 2000; Reichel et al., 1999; Whitlock et al., 2000), and the classification of infected cattle by the extent of shedding of Map in feces (Reichel et al., 1999; Whitlock et al., 2000) have not been standardized. The lack of standardization has made comparison of results from different studies and determination of associations between fecal colony counts and risk factors difficult.
The objectives of the present study were to (1) describe the distribution of Map CFU in dairy cows in herds with no or minimal testing for the organism, (2) determine if the distribution of herd CFU varied with prevalence, herd size, and season, (3) calculate the relative variability in Map CFU among or within herds, and within replicate samples from the same cow, and (4) assess the relationship between fecalculture prevalence and the proportion of light shedders (<10 CFU/tube) in a herd.
2. Material and methods 2.1. Fecal samples Between 1982 and 2000, fecal samples were obtained from the rectum of 786 cows on 93 Holstein dairies (herd size, 4–302) in Pennsylvania with no or minimal previous testing for Map. Most dairy herds practiced indoor calving in the winter months (November–March) and outdoor calving during the rest of the year. Calving occurred year round with a slight increase in the late winter and early spring months to take advantage of lower forage costs in the spring and summer. Owners of study herds and herd veterinarians often had recognized cattle with clinical Johne’s disease, and hence had requested a whole herd fecal culture to obtain an estimate of Map prevalence. Clinically affected cows had been culled from herds prior to testing, and hence, infected cows detected in this study were in stage 2. Samples were taken from all adult cows, regardless of stage of lactation or gestation. Animals were identified by ear-tag number and/or neck chain, date of birth, and purchase history. Fecal cultures were done at the Johne’s Research Laboratory, New Bolton Center, University of Pennsylvania, Kennett Square, PA. 2.2. Sample preparation and culture Two cultivation procedures were used, as described in the following section. For both methods, 2 g of feces and 35 ml water were mixed in a plastic tube until a homogeneous suspension was created (placed on a verimix tube rocker for 20 min), and then the suspension was allowed to settle for 30 min. In the first
B.M. Crossley et al. / Veterinary Microbiology 107 (2005) 257–263
method (Whipple et al., 1992), which was used from 1982 until October 1991, decontamination was done by adding 5 ml supernatant to 30 ml of 0.75% hexadecylpyridium chloride (HPC) solution. Samples were incubated for 18–24 h at room temperature, followed by centrifugation for 30 min at 990 g. The pellet was re-suspended in 1 ml of 100 mg amphotericin B/ml sterile water and then inoculated onto each of four tubes of Herrold’s egg yolk medium (HEYM) with mycobactin J with 200 ml of suspension/tube. After October 1991, the Cornell modified decontamination and inoculation method was used (Whitlock and Rosenberger, 1994). Decontamination was done by adding 5 ml supernatant to a 0.75% HPC solution, which contained 25 ml of 0.9% HPC in a 50% concentrated brain heart infusion (BHI) broth solution. Samples were incubated for 18–24 h at 35–37 8C, followed by centrifugation for 30 min at 900 g at room temperature. The pellet was resuspended in 1 ml of an antibiotic solution containing 50% BHI with 100 mg/ml vancomycin, 100 mg/ml nalidixic acid, and 50 ml/ml amphotericin B, and incubated overnight at 35–37 8C. Four tubes of HEYM, each from a different media lot and without antibiotics were each inoculated with 200 ul of the solution. For both methods, incubated tubes were kept at 37 8C in a slanted position for 1 week, after which their caps were tightened and the tubes were placed in an upright position. Samples were checked for growth of Map colonies every other week with the final reading at 16 weeks. The number of CFU in each tube was counted. If the CFU were too numerous to count (TNTC), numbers were recorded as TNTC and given a value of 75 for purposes of data analysis. For each cow, the mean CFU was based on colony counts in noncontaminated tubes. If all four tubes from a cow were contaminated, the results were excluded from analysis. If a single colony was found in only one of four non-contaminated tubes, the resultant mean was 0.25 CFU/tube. 2.3. Statistical analysis Colony counts were categorized as light (<10 CFU/ tube), moderate (10–50 CFU/tube), or high (>50 CFU/ tube) (Whitlock et al., 2000). Associations with three potential risk factors (within-herd prevalence, herd size measured as the number of cows, and season of sample
259
collection) were assessed. Prevalence (number of cows that were fecal-culture positive/total number of cows tested) and herd size were evaluated as continuous variables and in categories. Prevalence categories were low (0–10%), medium (>10–20%), medium high (>20–30%) and high (>30%). Herd size (number of cows) was categorized as small (<50), medium (>50– 100), or large (>100). Season at the time of sample collection was designated as summer (June–September), winter (November–March), or other (April, May, and October). Culture methods either were centrifugation (before October 1991) or Cornell (after October 1991). Univariable associations between categorized or continuous mean CFU/tube and continuous risk factor data were evaluated using Kruskal–Wallis one-way ANOVA or Pearson’s correlation coefficient, respectively. Spearman rank correlation was used for categorized CFU and categorized risk factors. A P-value of 0.05 was considered significant. Analyses were done with SPSSTM Version 11.0 (SPSS Inc., Chicago, IL). The multivariable relationship between mean CFU per cow and prevalence, herd size, season, and culture method categories was assessed using a generalized linear model (GLM) with a negative binomial distribution for CFU (GLMNB). S-Plus 6.1 (Insightful Corp., Seattle, Washington) and the glm.nb function (Venables and Ripley, 1999) available at http:// www.stats.ox.ac.uk/pub/MASS4/Software.html#Windows, were used for this analysis. The GLMNB model was chosen because the distribution of the mean colony CFU/tube was highly right skewed; some cows had a 0 or very low CFU (1 CFU on a single tube) and others had CFU that were TNTC on all tubes (equivalent to 300 CFU/tube). The negative binomial has a variance larger than the mean and has been shown to fit such a data distribution well (Crofton, 1971; Elston et al., 2001; Wilson et al., 1996). The model was first fitted using all four factors (prevalence, herd size, season and method) and all possible two- and three-way interactions, and then a backward elimination procedure was performed using the Akaike information criterion (AIC). The variable with the largest P value was eliminated, and the model was refitted until the AIC value did not decrease further. If the P value was between 0.05 and 0.1, the selected model was compared with a model without the term, using a likelihood ratio test. If the goodnessof-fit between the reduced model and the model with
260
B.M. Crossley et al. / Veterinary Microbiology 107 (2005) 257–263
the additional term was not significantly reduced, the tested term was removed. The variability in CFU between and within herds, depending on the variables studied, was assessed with a linear mixed-effects model (LME) using the nlme3.3.2 library for S-Plus 6.1. Mean colony counts were log transformed to fulfill normality assumptions of the LME. Contrasts (comparisons of each variable level to a baseline category) were used to test differences between levels of a categorical variable. Procedures for selection of the final LME model were similar to those used for the GLMNB. Assumptions of the GLMNB and LME models were assessed by evaluating histograms of the residuals and plots of the deviance residuals versus predicted values. The goodness-of-fit was evaluated with a scatter plot of fitted versus observed values.
these, 192 (24.4%) had colonies in only a single tube. The mean CFU per tube and mean per individual positive tube were 12.5 (39,354 colonies/3144 tubes) and 15.2 (39,354 colonies/2529 tubes with colonies), respectively. The overall mean herd prevalence was 13.1% (median 8.6%), and ranged between 1 and 100% for tested herds. Descriptive statistics for the mean and median CFU and the percentage of animals by shedding and risk factor category are in Table 1. Probability density function for mean Map colony counts per cow demonstrated clustering at the extremes of less than one colony per cow or more than 70 colonies per cow in 90% of all cases. For the 93 study herds, the proportion of positive cows that were light shedders increased with prevalence category (P < 0.001). For prevalence categories of 0–5, >5–10, >10–20, >20–30, and >30%, the median percentages of light shedders per herd were 1.6, 2.1, 3.3, 6.5, and 9.8%, respectively.
3. Results
3.2. Factors associated with fecal Map colony counts
3.1. Descriptive statistics
In univariable analysis, significant associations or correlations (P < 0.05) were evident between mean CFU (either categorized or in continuous form) and herd size and prevalence (either categorized or in
Fecal samples from 786 dairy cows had detectable colonies in at least one of the four HEYM tubes and, of
Table 1 Distribution of fecal colony counts (CFU) of Mycobacterium avium subsp. paratuberculosis from Holstein cows on 93 dairies in Pennsylvania, U.S.A., 1982–2000, categorized by risk factors Factor
Herd size (no. of cows) Small (4–49) Medium (50–100) Large (101–302) Season Summer (June–September) Winter (November–March) Other (April, May and October) Culture method Centrifugation Cornell modified Prevalence (%) 0–5 >5–10 >10–20 >20–30 >30 Total a
No. of animals
No. of herds
Mean CFU (S.D.)
Median CFU (S.D.)
212 265 309
35 40 18
23.5 (30.0) 23.6 (30.0) 32.5 (31.2)
304 368 114
41 31 21
342 444 73 72 309 107 225 786
Shedding category (%)a Light
Moderate
High
23.8 (31.4) 22.8 (31.4) 34.4 (34.7)
55.7 57.4 43
19.3 17.4 17.5
25 25.3 39.5
22.2 (28.8) 32.8 (31.4) 21.7 (30.5)
21.9 (30.8) 34.2 (34.3) 21.1 (31.4)
56.3 43.5 63.2
20.7 17.1 13.2
23 39.4 23.7
49 44
23.9 (29.6) 29.5 (31.4)
23.2 (31.1) 30.9 (34.3)
53.5 49.5
21.9 14.9
24.6 35.6
31 18 26 9 9 93
21.9 22.5 31.5 30.1 22.6 27.1
20.9 21.7 33.3 31.2 22.2 27.6
60.3 62.5 43 45.8 58.7 51.3
19.2 13.9 18.8 18.7 17.3 17.9
20.5 23.6 38.2 35.5 24 30.8
(28.7) (30.8) (30.6) (31.4) (30.5) (30.7)
(30.1) (31.4) (34.4) (34.1) (30.9) (33.1)
Shedding categories: light (mean tube CFU < 10), moderate (mean tube CFU, 10–50) and high (mean tube CFU > 50).
B.M. Crossley et al. / Veterinary Microbiology 107 (2005) 257–263
continuous form), season, and culture method. In general, higher mean CFU was associated with herd sizes >100 cows, moderate or moderate-high (10– 30%) prevalence, sample collection during winter and use of the Cornell method for fecal culture (Table 1). These relationships were further explored in the multivariable models. In the negative binomial model for cow-level data (n = 775 cows with complete data), there was a significant association between continuous mean CFU/tube and the categorized prevalence (P = 0.030), herd size (P = 0.046), and season (P = 0.0001). Two-way interactions were found between prevalence and herd size categories (P = 0.0002), and a borderline significant interaction between prevalence and season (P = 0.064). There was no significant association between culture method and the mean CFU. In the LME model for herdlevel data, significant positive associations were evident between the log transformed continuous tube counts and continuous herd size (P = 0.006), season (P = 0.019), and there was a borderline significant association with categorized prevalence (P = 0.087), and an interaction between prevalence and herd size (P = 0.077). These last two terms where kept in the model because of their biological plausibility and because the fit of the model was significantly improved compared with reduced model without the terms (AIC 10,035 versus 10,038: likelihood ratio test = 15.3, P = 0.017). Table 2 shows the estimated variability (S.D.) in the mean CFU among herds, within herds (between cows within a particular herd), and within individual cows after controlling for the effects of season, herd size, prevalence, and the two-way interaction between herd size and prevalence. Estimated S.D. was greatest for cows within the same herd, intermediate for tube replicate within cow, and smallest among cows from different herds. Table 2 Estimation of the variability (S.D.) at the herd and cow level of log transformed mean fecal colony counts (CFU) of Map (cultured in four HEYM tubes) and obtained from Holstein cows on 93 dairies in Pennsylvania, U.S.A., 1982–2000 Level
S.D.
95% CI
Among herds Among cows within the same herd Within cows
0.57 1.42 0.89
0.40, 0.80 1.34, 1.51 0.87, 0.92
Estimates expressed in actual tube CFU (exponentiation of the S.D. values – 0.1).
261
4. Discussion In the present study mean Map CFU values at the cow and herd level were associated with prevalence category, larger herd size and sampling during winter. Mean CFU values were a conservative estimate of the true CFU values for risk factor groups because cows that had Map CFU on fecal culture that were too numerous to count were recorded as having 75 CFU. It is likely that the TNTC grouping included a wide range of Map CFU values. Prevalence estimates for each herd were culturebased and therefore are likely to substantially underestimate the true within-herd prevalence. Higher prevalence suggests increased environmental contamination with Map due to increased number of shedding cows. In these herds, replacement heifers would have a higher chance of infection by the fecal-oral route and control of environmental contamination would be an appropriate control strategy though adult infections are believed to occur much less frequently than in the first 6 months of life (Whitlock and Buergelt, 1996). Significant interaction between categorized season and prevalence using the GLMNB model was not evident whereas we found such an interaction using the LME model with a random herd effect, suggesting that fluctuation in CFU values at different times of the year and in different prevalence categories are influenced by the herd management, body condition and for other biological reasons. In the current study, larger herds had greater mean CFU per cow or per herd. The reasons for the herd size relationship are speculative but larger herds often have higher stocking densities and management systems that promote exposure to Map at a young age. Stressors might trigger shedding of Map, shorten the incubation period (Whittington and Sergeant, 2001), and influence the numbers of organisms shed in feces. Also, the probability of having a subclinical cow in a large herd is greater than in smaller herds. Larger herds are also more difficult to monitor particular when stocking rates are low, pasture is spelled, or when animals are culled at relatively young ages (Whittington and Sergeant, 2001). In the NAHMS Dairy 1996 study, about 40% of larger herds were considered to be infected compared with about 20% of smaller herds based on ELISA testing of 30–40 animals per herd (Wells and Wagner, 2000). However,
262
B.M. Crossley et al. / Veterinary Microbiology 107 (2005) 257–263
the within-herd ELISA prevalence did not differ across herd size categories similar to those we used (1–49 cows = 2.6%; 50–99 cows = 2.3%, and 100–299 cows = 2.5%; Bruce Wagner, Unpublished data). Collection of fecal samples during winter was associated with higher mean Map CFU. Adverse weather conditions (Jorgensen, 1977) and an increased frequency of calvings in late winter, which is a stress factor (Jakobsen et al., 2000), are possible explanations for this association. However, the explanation might be political rather than biological. For many years of this study, farmers were more reluctant to cull high Map shedders during winter months because their base annual milk production was factored strongly during the winter months, when production costs were greatest. Therefore, the more milk sold during the winter stabling period, the higher the price for milk produced during the following summer months. A more critical evaluation of seasonal effects will require repeated sampling of animals over time in the same herds. No significant association was found between CFU and culture method, although the culture method changed approximately midway through the study. The primary difference was the incorporation of double incubation which has minimal effect on Map recovery, compared with other factors such as antibiotics in the media and using fewer tubes or smaller inoculum size. Hence, it was appropriate to not make time periodspecific comparisons. At a herd-level, an unknown impact on results may have occurred with improved knowledge leading to new disinfection methods, pasteurization of hospital milk (Grant et al., 2002; Lund et al., 2000; Obasanjo et al., 1997; Stabel, 2001), improved rations, vaccination (Kormendy, 1994) and different housing systems preventing calves from contact with feces from dairy cows (Obasanjo et al., 1997) over the 18 years of the study. In a prior study of 10 Pennsylvania dairy herds (Whitlock et al., 2000), the proportions of light (<10 CFU/tube), moderate (10–50 CFU/tube) and high (>50 CFU/tube) fecal shedders were estimated to be approximately 70, 10, and 20%, respectively. In the present study, most sampled cows were either light (51.4%) or high (30.8%) fecal shedders, with fewer (17.8%) in the intermediate category. The distribution of shedding was similar to findings of a study in Dutch dairy herds (Kalis et al., 1999) in which feces of 150 Map positive cows were cultured by the Jorgensen
method. In a recent study of 376 culture-positive diary cattle 59% of cows were light shedders and 26% were heavy shedders (unpublished data, RHW). Reasons why so few cattle are in the intermediate shedding category are speculative but it might in part be an artifact of the inability to estimate the most probable number of CFU in the TNTC category without dilution of sample inocula. An interesting finding in the herdlevel analysis was the higher proportion of light shedders with increasing Map prevalence. Although our calculations in low prevalence herds were based on few shedding cows, this finding provides support for the hypothesis that higher prevalence herds have more low Map shedders which might be attributable to a ‘‘pass through’’ phenomenon, as previously suggested (Sweeney et al., 1992). In 192 cases, we isolated Map from only a single tube per cow. Difficulties in the isolation of Map from rare fecal shedders are well known and include intermittent shedding of the bacterium (Merkal et al., 1968) and a substantial decrease in Map numbers attributable to sample decontamination (Whittington and Sergeant, 2001). One consequence of the finding of colonies in only one of four HEYM tubes is that approximately 48 cows would not have been detected as Map infected if only three tubes without mycobactin had been used, or 96 cows using two tubes. The net loss in the relative sensitivity of fecal culture would have been approximately 6% (three tubes) and 12% (two tubes) compared with culture on four tubes of HEYM. The present study highlights the importance of multiple CFU estimates (replicate tubes) per cow and the influence of prevalence, herd size, and season of sampling on animal and herd-level CFU. Future studies of risk factors for CFU at the cow and herd level should consider possible confounding effects of these variables. In addition, use of four tubes of HEYM without antibiotics as the standard fecal culture procedure is recommended, to ensure adequate sensitivity of culture.
Acknowledgements The study was supported in part by the Pennsylvania Department of Agriculture, Harrisburg, PA, 3-year research grant on Johne’s disease and the Center for
B.M. Crossley et al. / Veterinary Microbiology 107 (2005) 257–263
Food Animal Health, School of Veterinary Medicine, University of California, Davis. References Bech-Nielsen, S., Shulaw, W.P., Frandsen, P.L., Jorgensen, J.B., Feld, N.C., Ahrens, P., 1992. Comparison of subjective and objective test evaluations for use of Mycobacterium phleiadsorbed serum in a dot-enzyme-linked immunosorbent assay for diagnosis of paratuberculosis in cattle. Am. J. Vet. Res. 53, 1386–1391. Colgrove, G.S., Thoen, C.O., Blackburn, B.O., Murphy, C.D., 1989. Paratuberculosis in cattle: a comparison of three serologic tests with results of fecal culture. Vet. Microbiol. 19, 183–187. Crofton, H.D., 1971. A quantitative approach to parasitism. Parasitology 63, 343–364. Eamens, G.J., Whittington, R.J., Marsh, I.B., Turner, M.J., Saunders, V., Kemsley, P.D., Rayward, D., 2000. Comparative sensitivity of various faecal culture methods and ELISA in dairy cattle herds with endemic Johne’s disease. Vet. Microbiol. 77, 357–367. Elston, D.A., Moss, R., Boulinier, T., Arrowsmith, C., Lambin, X., 2001. Analysis of aggregation, a worked example: numbers of ticks on red grouse chicks. Parasitology 122, 563–569. Giese, S.B., Ahrens, P., 2000. Detection of Mycobacterium avium subsp. paratuberculosis in milk from clinically affected cows by PCR and culture. Vet. Microbiol. 77, 291–297. Grant, I.R., Hitchings, E.I., McCartney, A., Ferguson, F., Rowe, M.T., 2002. Effect of commercial-scale high-temperature, shorttime pasteurization on the viability of Mycobacterium paratuberculosis in naturally infected cows’ milk. Appl. Environ. Microbiol. 68, 602–607. Jakobsen, M.B., Alban, L., Nielsen, S.S., 2000. A cross-sectional study of paratuberculosis in 1155 Danish dairy cows. Prev. Vet. Med. 46, 15–27. Jorgensen, J.B., 1977. Survival of Mycobacterium paratuberculosis in slurry. Nord. Vet. Med. 29, 267–270. Kalis, C.H., Hesselink, J.W., Russchen, E.W., Barkema, H.W., Collins, M.T., Visser, I.J., 1999. Factors influencing the isolation of Mycobacterium avium subsp. paratuberculosis from bovine fecal samples. J. Vet. Diagn. Invest. 11, 345–351. Kormendy, B., 1994. The effect of vaccination on the prevalence of paratuberculosis in large dairy herds. Vet. Microbiol. 41, 117– 125. Lund, B.M., Donnelly, C.W., Rampling, A., 2000. Heat resistance of Mycobacterium paratuberculosis. Lett. Appl. Microbiol. 31, 184–185. Merkal, R.S., Larsen, A.B., Kopecky, K.E., 1968. Comparison of examination and test methods for early detection of paratuberculous cattle. Am. J. Vet. Res. 29, 1533–1538. Obasanjo, I.O., Grohn, Y.T., Mohammed, H.O., 1997. Farm factors associated with the presence of Mycobacterium paratuberculosis infection in dairy herds on the New York State Paratuberculosis Control Program. Prev. Vet. Med. 32, 243–251. Ott, S.L., Wells, S.J., Wagner, B.A., 1999. Herd-level economic losses associated with Johne’s disease on US dairy operations. Prev. Vet. Med. 40, 179–192.
263
Pavlik, I., Matlova, L., Bartl, J., Svastova, P., Dvorska, L., Whitlock, R., 2000. Parallel faecal and organ Mycobacterium avium subsp. paratuberculosis culture of different productivity types of cattle. Vet. Microbiol. 77, 309–324. Reichel, M.P., Kittelberger, R., Penrose, M.E., Meynell, R.M., Cousins, D., Ellis, T., Mutharia, L.M., Sugden, E.A., Johns, A.H., de Lisle, G.W., 1999. Comparison of serological tests and faecal culture for the detection of Mycobacterium avium subsp. paratuberculosis infection in cattle and analysis of the antigens involved. Vet. Microbiol. 66, 135–150. Sherman, D.M., Gay, J.M., Bouley, D.S., Nelson, G.H., 1990. Comparison of the complement-fixation and agar gel immunodiffusion tests for diagnosis of subclinical bovine paratuberculosis. Am. J. Vet. Res. 51, 461–465. Stabel, J.R., 2001. On-farm batch pasteurization destroys Mycobacterium paratuberculosis in waste milk. J. Dairy Sci. 84, 524–527. Sweeney, R.W., Whitlock, R.H., Hamir, A.N., Rosenberger, A.E., Herr, S.A., 1992. Isolation of Mycobacterium paratuberculosis after oral inoculation in uninfected cattle. Am. J. Vet. Res. 53, 1312–1314. Sweeney, R.W., Whitlock, R.H., Buckley, C.L., Spencer, P.A., 1995. Evaluation of a commercial enzyme-linked immunosorbent assay for the diagnosis of paratuberculosis in dairy cattle. J. Vet. Diagn. Invest. 7, 488–493. Venables, W.N., Ripley, B.D., 1999. Modern Applied Statistics with S-Plus. Springer-Verlag, New York. Wells, S.J., Wagner, B.A., 2000. Herd-level risk factors for infection with Mycobacterium paratuberculosis in US dairies and association between familiarity of the herd manager with the disease or prior diagnosis of the disease in that herd and use of preventive measures. J. Am. Vet. Med. Assoc. 216, 1450–1457. Wells, S.J., Whitlock, R.H., Lindeman, C.J., Fyock, T., 2002. Evaluation of bacteriologic culture of pooled fecal samples for detection of Mycobacterium paratuberculosis. Am. J. Vet. Res. 63, 1207–1211. Whipple, D.L., Kapke, P.A., Andersen, P.R., 1992. Comparison of a commercial DNA probe test and three cultivation procedures for detection of Mycobacterium paratuberculosis in bovine feces. J. Vet. Diagn. Invest. 4, 23–27. Whitlock, R.H., Buergelt, C., 1996. Preclinical and clinical manifestations of paratuberculosis (including pathology). Vet. Clin. North. Am: Food Anim. Pract. 12, 345–356. Whitlock, R.H., Rosenberger, A.E., 1994. Fecal culture protocol for Mycobacterium paratuberculosis: a recommended procedure. In: Proceedings of the 94th U.S. Animal Health Association, Denver, CO, pp. 280–285. Whitlock, R.H., Wells, S.J., Sweeney, R.W., Van Tiem, J., 2000. ELISA and fecal culture for paratuberculosis (Johne’s disease): sensitivity and specificity of each method. Vet. Microbiol. 77, 387–398. Whittington, R.J., Sergeant, E.S., 2001. Progress towards understanding the spread, detection and control of Mycobacterium avium subsp. paratuberculosis in animal populations. Aust. Vet. J. 79, 267–278. Wilson, K., Grenfell, B.T., Shaw, D.J., 1996. Analysis of aggregated parasite distributions: a comparison of methods. Funct. Ecol. 10, 592–601.
Fecal shedding of Mycobacterium avium subsp. paratuberculosis by dairy cows Beate M. Crossley a, Francisco J. Zagmutt-Vergara a, Terry L. Fyock b, Robert H. Whitlock b, Ian A. Gardner a,* a
Department of Medicine and Epidemiology, School of Veterinary Medicine, University of California, Davis, CA 95616, USA b Johne’s Research Laboratory, New Bolton Center, University of Pennsylvania, 382 West Street Road, Kennett Square, PA 19348, USA Received 17 August 2004; received in revised form 21 January 2005; accepted 31 January 2005
Abstract Between 1982 and 2000, fecal samples were obtained from 786 cows that were shedding Mycobacterium avium subsp. paratuberculosis (Map). These cows were resident on 93 Pennsylvania dairies (mean herd size, 64 milk cows) that had no or minimal previous testing for Map. Feces were cultured on four tubes of Herrold’s egg yolk medium and the distribution of mean Map colony forming units (CFU) was evaluated. Most cows were light (<10 CFU/tube, 51.4%) or high (>50 CFU/tube, 30.8%) fecal shedders with fewer cows in the moderate category (10–50 CFU/tube). Of the 786 cows, 192 (24.4%) had colonies in only one of four tubes. In the multivariable negative binomial model, there were significant associations between mean CFU/tube and prevalence, herd size, and season and an interaction between herd size and season. The linear mixed model of continuous tube counts with a random herd effect yielded similar findings with associations with herd size as a continuous variable, season, and an interaction between categorized prevalence and continuous herd size. Variability in CFU/tube was greatest among cows in the same herd, intermediate for replicate tubes from the same cow, and smallest among cows in different herds. Reduction in the number of replicate tubes from four would have reduced the sensitivity of fecal culture for Map by approximately 6% (for three tubes) to 12% (for two tubes). # 2005 Elsevier B.V. All rights reserved. Keywords: Mycobacterium paratuberculosis; Colony forming units; Isolation; Cattle; Feces
1. Introduction
* Corresponding author. Tel.: +1 530 752 6992; fax: +1 530 752 0414. E-mail address: [email protected] (I.A. Gardner).
Mycobacterium avium subsp. paratuberculosis infection (Johne’s disease or bovine paratuberculosis) costs U.S. cattle producers up to $ 250 million annually (Ott et al., 1999). Mycobacterium a. paratuberculosis (Map) is spread mainly by the
0378-1135/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.vetmic.2005.01.017
258
B.M. Crossley et al. / Veterinary Microbiology 107 (2005) 257–263
fecal-oral route and infected cattle are categorized as being in one of three stages depending on fecal shedding of Map and clinical signs. Cattle in stage 1 are defined as prepatent and preclinical, and no organisms are detectable by culture. Stage-2 cattle are shedding detectable numbers of Map but without clinical signs. The shedding of Map organisms can be intermittent and detection by culture is imperfect, especially when few organisms are shed in feces (Merkal et al., 1968). Cattle in stage 2 often are further categorized as light (<10 colony forming units (CFU)/tube), moderate (10–50 CFU/tube), or high fecal shedders (>50 CFU/tube) based on culture results on conventional media such as Herrold’s egg yolk (Whitlock et al., 2000). Categorizations of mean CFU/tube with more levels have been used in some studies (Wells et al., 2002). Infected cattle with stage3 disease typically lose weight, have diarrhea that is not responsive to antibiotic treatment, and nearly always are culture and antibody positive (Whitlock et al., 2000). Cattle with clinical signs represent only a small fraction of the Map-infected animals in a herd (Sherman et al., 1990; Whitlock and Buergelt, 1996). Accurate detection of Map is dependent on the sensitivity and specificity of test methods and stage of infection, which affects fecal shedding of bacteria and antibody response to infection. Even though a ‘‘passthrough’’ phenomenon has been postulated for organisms in the intestinal tract (Sweeney et al., 1992) and some investigators believe that ‘‘passthrough’’ may be responsible for more low shedders in some heavily infected farms, fecal culture generally is considered to be 100% specific. Hence, fecal culture often is preferred to serologic testing for individual and herd classification of Map status (Colgrove et al., 1989; Sweeney et al., 1995). However, the procedures (Eamens et al., 2000; Pavlik et al., 2000), the distribution of colony counts (Bech-Nielsen et al., 1992; Giese and Ahrens, 2000; Kalis et al., 1999; Pavlik et al., 2000; Reichel et al., 1999; Whitlock et al., 2000), and the classification of infected cattle by the extent of shedding of Map in feces (Reichel et al., 1999; Whitlock et al., 2000) have not been standardized. The lack of standardization has made comparison of results from different studies and determination of associations between fecal colony counts and risk factors difficult.
The objectives of the present study were to (1) describe the distribution of Map CFU in dairy cows in herds with no or minimal testing for the organism, (2) determine if the distribution of herd CFU varied with prevalence, herd size, and season, (3) calculate the relative variability in Map CFU among or within herds, and within replicate samples from the same cow, and (4) assess the relationship between fecalculture prevalence and the proportion of light shedders (<10 CFU/tube) in a herd.
2. Material and methods 2.1. Fecal samples Between 1982 and 2000, fecal samples were obtained from the rectum of 786 cows on 93 Holstein dairies (herd size, 4–302) in Pennsylvania with no or minimal previous testing for Map. Most dairy herds practiced indoor calving in the winter months (November–March) and outdoor calving during the rest of the year. Calving occurred year round with a slight increase in the late winter and early spring months to take advantage of lower forage costs in the spring and summer. Owners of study herds and herd veterinarians often had recognized cattle with clinical Johne’s disease, and hence had requested a whole herd fecal culture to obtain an estimate of Map prevalence. Clinically affected cows had been culled from herds prior to testing, and hence, infected cows detected in this study were in stage 2. Samples were taken from all adult cows, regardless of stage of lactation or gestation. Animals were identified by ear-tag number and/or neck chain, date of birth, and purchase history. Fecal cultures were done at the Johne’s Research Laboratory, New Bolton Center, University of Pennsylvania, Kennett Square, PA. 2.2. Sample preparation and culture Two cultivation procedures were used, as described in the following section. For both methods, 2 g of feces and 35 ml water were mixed in a plastic tube until a homogeneous suspension was created (placed on a verimix tube rocker for 20 min), and then the suspension was allowed to settle for 30 min. In the first
B.M. Crossley et al. / Veterinary Microbiology 107 (2005) 257–263
method (Whipple et al., 1992), which was used from 1982 until October 1991, decontamination was done by adding 5 ml supernatant to 30 ml of 0.75% hexadecylpyridium chloride (HPC) solution. Samples were incubated for 18–24 h at room temperature, followed by centrifugation for 30 min at 990 g. The pellet was re-suspended in 1 ml of 100 mg amphotericin B/ml sterile water and then inoculated onto each of four tubes of Herrold’s egg yolk medium (HEYM) with mycobactin J with 200 ml of suspension/tube. After October 1991, the Cornell modified decontamination and inoculation method was used (Whitlock and Rosenberger, 1994). Decontamination was done by adding 5 ml supernatant to a 0.75% HPC solution, which contained 25 ml of 0.9% HPC in a 50% concentrated brain heart infusion (BHI) broth solution. Samples were incubated for 18–24 h at 35–37 8C, followed by centrifugation for 30 min at 900 g at room temperature. The pellet was resuspended in 1 ml of an antibiotic solution containing 50% BHI with 100 mg/ml vancomycin, 100 mg/ml nalidixic acid, and 50 ml/ml amphotericin B, and incubated overnight at 35–37 8C. Four tubes of HEYM, each from a different media lot and without antibiotics were each inoculated with 200 ul of the solution. For both methods, incubated tubes were kept at 37 8C in a slanted position for 1 week, after which their caps were tightened and the tubes were placed in an upright position. Samples were checked for growth of Map colonies every other week with the final reading at 16 weeks. The number of CFU in each tube was counted. If the CFU were too numerous to count (TNTC), numbers were recorded as TNTC and given a value of 75 for purposes of data analysis. For each cow, the mean CFU was based on colony counts in noncontaminated tubes. If all four tubes from a cow were contaminated, the results were excluded from analysis. If a single colony was found in only one of four non-contaminated tubes, the resultant mean was 0.25 CFU/tube. 2.3. Statistical analysis Colony counts were categorized as light (<10 CFU/ tube), moderate (10–50 CFU/tube), or high (>50 CFU/ tube) (Whitlock et al., 2000). Associations with three potential risk factors (within-herd prevalence, herd size measured as the number of cows, and season of sample
259
collection) were assessed. Prevalence (number of cows that were fecal-culture positive/total number of cows tested) and herd size were evaluated as continuous variables and in categories. Prevalence categories were low (0–10%), medium (>10–20%), medium high (>20–30%) and high (>30%). Herd size (number of cows) was categorized as small (<50), medium (>50– 100), or large (>100). Season at the time of sample collection was designated as summer (June–September), winter (November–March), or other (April, May, and October). Culture methods either were centrifugation (before October 1991) or Cornell (after October 1991). Univariable associations between categorized or continuous mean CFU/tube and continuous risk factor data were evaluated using Kruskal–Wallis one-way ANOVA or Pearson’s correlation coefficient, respectively. Spearman rank correlation was used for categorized CFU and categorized risk factors. A P-value of 0.05 was considered significant. Analyses were done with SPSSTM Version 11.0 (SPSS Inc., Chicago, IL). The multivariable relationship between mean CFU per cow and prevalence, herd size, season, and culture method categories was assessed using a generalized linear model (GLM) with a negative binomial distribution for CFU (GLMNB). S-Plus 6.1 (Insightful Corp., Seattle, Washington) and the glm.nb function (Venables and Ripley, 1999) available at http:// www.stats.ox.ac.uk/pub/MASS4/Software.html#Windows, were used for this analysis. The GLMNB model was chosen because the distribution of the mean colony CFU/tube was highly right skewed; some cows had a 0 or very low CFU (1 CFU on a single tube) and others had CFU that were TNTC on all tubes (equivalent to 300 CFU/tube). The negative binomial has a variance larger than the mean and has been shown to fit such a data distribution well (Crofton, 1971; Elston et al., 2001; Wilson et al., 1996). The model was first fitted using all four factors (prevalence, herd size, season and method) and all possible two- and three-way interactions, and then a backward elimination procedure was performed using the Akaike information criterion (AIC). The variable with the largest P value was eliminated, and the model was refitted until the AIC value did not decrease further. If the P value was between 0.05 and 0.1, the selected model was compared with a model without the term, using a likelihood ratio test. If the goodnessof-fit between the reduced model and the model with
260
B.M. Crossley et al. / Veterinary Microbiology 107 (2005) 257–263
the additional term was not significantly reduced, the tested term was removed. The variability in CFU between and within herds, depending on the variables studied, was assessed with a linear mixed-effects model (LME) using the nlme3.3.2 library for S-Plus 6.1. Mean colony counts were log transformed to fulfill normality assumptions of the LME. Contrasts (comparisons of each variable level to a baseline category) were used to test differences between levels of a categorical variable. Procedures for selection of the final LME model were similar to those used for the GLMNB. Assumptions of the GLMNB and LME models were assessed by evaluating histograms of the residuals and plots of the deviance residuals versus predicted values. The goodness-of-fit was evaluated with a scatter plot of fitted versus observed values.
these, 192 (24.4%) had colonies in only a single tube. The mean CFU per tube and mean per individual positive tube were 12.5 (39,354 colonies/3144 tubes) and 15.2 (39,354 colonies/2529 tubes with colonies), respectively. The overall mean herd prevalence was 13.1% (median 8.6%), and ranged between 1 and 100% for tested herds. Descriptive statistics for the mean and median CFU and the percentage of animals by shedding and risk factor category are in Table 1. Probability density function for mean Map colony counts per cow demonstrated clustering at the extremes of less than one colony per cow or more than 70 colonies per cow in 90% of all cases. For the 93 study herds, the proportion of positive cows that were light shedders increased with prevalence category (P < 0.001). For prevalence categories of 0–5, >5–10, >10–20, >20–30, and >30%, the median percentages of light shedders per herd were 1.6, 2.1, 3.3, 6.5, and 9.8%, respectively.
3. Results
3.2. Factors associated with fecal Map colony counts
3.1. Descriptive statistics
In univariable analysis, significant associations or correlations (P < 0.05) were evident between mean CFU (either categorized or in continuous form) and herd size and prevalence (either categorized or in
Fecal samples from 786 dairy cows had detectable colonies in at least one of the four HEYM tubes and, of
Table 1 Distribution of fecal colony counts (CFU) of Mycobacterium avium subsp. paratuberculosis from Holstein cows on 93 dairies in Pennsylvania, U.S.A., 1982–2000, categorized by risk factors Factor
Herd size (no. of cows) Small (4–49) Medium (50–100) Large (101–302) Season Summer (June–September) Winter (November–March) Other (April, May and October) Culture method Centrifugation Cornell modified Prevalence (%) 0–5 >5–10 >10–20 >20–30 >30 Total a
No. of animals
No. of herds
Mean CFU (S.D.)
Median CFU (S.D.)
212 265 309
35 40 18
23.5 (30.0) 23.6 (30.0) 32.5 (31.2)
304 368 114
41 31 21
342 444 73 72 309 107 225 786
Shedding category (%)a Light
Moderate
High
23.8 (31.4) 22.8 (31.4) 34.4 (34.7)
55.7 57.4 43
19.3 17.4 17.5
25 25.3 39.5
22.2 (28.8) 32.8 (31.4) 21.7 (30.5)
21.9 (30.8) 34.2 (34.3) 21.1 (31.4)
56.3 43.5 63.2
20.7 17.1 13.2
23 39.4 23.7
49 44
23.9 (29.6) 29.5 (31.4)
23.2 (31.1) 30.9 (34.3)
53.5 49.5
21.9 14.9
24.6 35.6
31 18 26 9 9 93
21.9 22.5 31.5 30.1 22.6 27.1
20.9 21.7 33.3 31.2 22.2 27.6
60.3 62.5 43 45.8 58.7 51.3
19.2 13.9 18.8 18.7 17.3 17.9
20.5 23.6 38.2 35.5 24 30.8
(28.7) (30.8) (30.6) (31.4) (30.5) (30.7)
(30.1) (31.4) (34.4) (34.1) (30.9) (33.1)
Shedding categories: light (mean tube CFU < 10), moderate (mean tube CFU, 10–50) and high (mean tube CFU > 50).
B.M. Crossley et al. / Veterinary Microbiology 107 (2005) 257–263
continuous form), season, and culture method. In general, higher mean CFU was associated with herd sizes >100 cows, moderate or moderate-high (10– 30%) prevalence, sample collection during winter and use of the Cornell method for fecal culture (Table 1). These relationships were further explored in the multivariable models. In the negative binomial model for cow-level data (n = 775 cows with complete data), there was a significant association between continuous mean CFU/tube and the categorized prevalence (P = 0.030), herd size (P = 0.046), and season (P = 0.0001). Two-way interactions were found between prevalence and herd size categories (P = 0.0002), and a borderline significant interaction between prevalence and season (P = 0.064). There was no significant association between culture method and the mean CFU. In the LME model for herdlevel data, significant positive associations were evident between the log transformed continuous tube counts and continuous herd size (P = 0.006), season (P = 0.019), and there was a borderline significant association with categorized prevalence (P = 0.087), and an interaction between prevalence and herd size (P = 0.077). These last two terms where kept in the model because of their biological plausibility and because the fit of the model was significantly improved compared with reduced model without the terms (AIC 10,035 versus 10,038: likelihood ratio test = 15.3, P = 0.017). Table 2 shows the estimated variability (S.D.) in the mean CFU among herds, within herds (between cows within a particular herd), and within individual cows after controlling for the effects of season, herd size, prevalence, and the two-way interaction between herd size and prevalence. Estimated S.D. was greatest for cows within the same herd, intermediate for tube replicate within cow, and smallest among cows from different herds. Table 2 Estimation of the variability (S.D.) at the herd and cow level of log transformed mean fecal colony counts (CFU) of Map (cultured in four HEYM tubes) and obtained from Holstein cows on 93 dairies in Pennsylvania, U.S.A., 1982–2000 Level
S.D.
95% CI
Among herds Among cows within the same herd Within cows
0.57 1.42 0.89
0.40, 0.80 1.34, 1.51 0.87, 0.92
Estimates expressed in actual tube CFU (exponentiation of the S.D. values – 0.1).
261
4. Discussion In the present study mean Map CFU values at the cow and herd level were associated with prevalence category, larger herd size and sampling during winter. Mean CFU values were a conservative estimate of the true CFU values for risk factor groups because cows that had Map CFU on fecal culture that were too numerous to count were recorded as having 75 CFU. It is likely that the TNTC grouping included a wide range of Map CFU values. Prevalence estimates for each herd were culturebased and therefore are likely to substantially underestimate the true within-herd prevalence. Higher prevalence suggests increased environmental contamination with Map due to increased number of shedding cows. In these herds, replacement heifers would have a higher chance of infection by the fecal-oral route and control of environmental contamination would be an appropriate control strategy though adult infections are believed to occur much less frequently than in the first 6 months of life (Whitlock and Buergelt, 1996). Significant interaction between categorized season and prevalence using the GLMNB model was not evident whereas we found such an interaction using the LME model with a random herd effect, suggesting that fluctuation in CFU values at different times of the year and in different prevalence categories are influenced by the herd management, body condition and for other biological reasons. In the current study, larger herds had greater mean CFU per cow or per herd. The reasons for the herd size relationship are speculative but larger herds often have higher stocking densities and management systems that promote exposure to Map at a young age. Stressors might trigger shedding of Map, shorten the incubation period (Whittington and Sergeant, 2001), and influence the numbers of organisms shed in feces. Also, the probability of having a subclinical cow in a large herd is greater than in smaller herds. Larger herds are also more difficult to monitor particular when stocking rates are low, pasture is spelled, or when animals are culled at relatively young ages (Whittington and Sergeant, 2001). In the NAHMS Dairy 1996 study, about 40% of larger herds were considered to be infected compared with about 20% of smaller herds based on ELISA testing of 30–40 animals per herd (Wells and Wagner, 2000). However,
262
B.M. Crossley et al. / Veterinary Microbiology 107 (2005) 257–263
the within-herd ELISA prevalence did not differ across herd size categories similar to those we used (1–49 cows = 2.6%; 50–99 cows = 2.3%, and 100–299 cows = 2.5%; Bruce Wagner, Unpublished data). Collection of fecal samples during winter was associated with higher mean Map CFU. Adverse weather conditions (Jorgensen, 1977) and an increased frequency of calvings in late winter, which is a stress factor (Jakobsen et al., 2000), are possible explanations for this association. However, the explanation might be political rather than biological. For many years of this study, farmers were more reluctant to cull high Map shedders during winter months because their base annual milk production was factored strongly during the winter months, when production costs were greatest. Therefore, the more milk sold during the winter stabling period, the higher the price for milk produced during the following summer months. A more critical evaluation of seasonal effects will require repeated sampling of animals over time in the same herds. No significant association was found between CFU and culture method, although the culture method changed approximately midway through the study. The primary difference was the incorporation of double incubation which has minimal effect on Map recovery, compared with other factors such as antibiotics in the media and using fewer tubes or smaller inoculum size. Hence, it was appropriate to not make time periodspecific comparisons. At a herd-level, an unknown impact on results may have occurred with improved knowledge leading to new disinfection methods, pasteurization of hospital milk (Grant et al., 2002; Lund et al., 2000; Obasanjo et al., 1997; Stabel, 2001), improved rations, vaccination (Kormendy, 1994) and different housing systems preventing calves from contact with feces from dairy cows (Obasanjo et al., 1997) over the 18 years of the study. In a prior study of 10 Pennsylvania dairy herds (Whitlock et al., 2000), the proportions of light (<10 CFU/tube), moderate (10–50 CFU/tube) and high (>50 CFU/tube) fecal shedders were estimated to be approximately 70, 10, and 20%, respectively. In the present study, most sampled cows were either light (51.4%) or high (30.8%) fecal shedders, with fewer (17.8%) in the intermediate category. The distribution of shedding was similar to findings of a study in Dutch dairy herds (Kalis et al., 1999) in which feces of 150 Map positive cows were cultured by the Jorgensen
method. In a recent study of 376 culture-positive diary cattle 59% of cows were light shedders and 26% were heavy shedders (unpublished data, RHW). Reasons why so few cattle are in the intermediate shedding category are speculative but it might in part be an artifact of the inability to estimate the most probable number of CFU in the TNTC category without dilution of sample inocula. An interesting finding in the herdlevel analysis was the higher proportion of light shedders with increasing Map prevalence. Although our calculations in low prevalence herds were based on few shedding cows, this finding provides support for the hypothesis that higher prevalence herds have more low Map shedders which might be attributable to a ‘‘pass through’’ phenomenon, as previously suggested (Sweeney et al., 1992). In 192 cases, we isolated Map from only a single tube per cow. Difficulties in the isolation of Map from rare fecal shedders are well known and include intermittent shedding of the bacterium (Merkal et al., 1968) and a substantial decrease in Map numbers attributable to sample decontamination (Whittington and Sergeant, 2001). One consequence of the finding of colonies in only one of four HEYM tubes is that approximately 48 cows would not have been detected as Map infected if only three tubes without mycobactin had been used, or 96 cows using two tubes. The net loss in the relative sensitivity of fecal culture would have been approximately 6% (three tubes) and 12% (two tubes) compared with culture on four tubes of HEYM. The present study highlights the importance of multiple CFU estimates (replicate tubes) per cow and the influence of prevalence, herd size, and season of sampling on animal and herd-level CFU. Future studies of risk factors for CFU at the cow and herd level should consider possible confounding effects of these variables. In addition, use of four tubes of HEYM without antibiotics as the standard fecal culture procedure is recommended, to ensure adequate sensitivity of culture.
Acknowledgements The study was supported in part by the Pennsylvania Department of Agriculture, Harrisburg, PA, 3-year research grant on Johne’s disease and the Center for
B.M. Crossley et al. / Veterinary Microbiology 107 (2005) 257–263
Food Animal Health, School of Veterinary Medicine, University of California, Davis. References Bech-Nielsen, S., Shulaw, W.P., Frandsen, P.L., Jorgensen, J.B., Feld, N.C., Ahrens, P., 1992. Comparison of subjective and objective test evaluations for use of Mycobacterium phleiadsorbed serum in a dot-enzyme-linked immunosorbent assay for diagnosis of paratuberculosis in cattle. Am. J. Vet. Res. 53, 1386–1391. Colgrove, G.S., Thoen, C.O., Blackburn, B.O., Murphy, C.D., 1989. Paratuberculosis in cattle: a comparison of three serologic tests with results of fecal culture. Vet. Microbiol. 19, 183–187. Crofton, H.D., 1971. A quantitative approach to parasitism. Parasitology 63, 343–364. Eamens, G.J., Whittington, R.J., Marsh, I.B., Turner, M.J., Saunders, V., Kemsley, P.D., Rayward, D., 2000. Comparative sensitivity of various faecal culture methods and ELISA in dairy cattle herds with endemic Johne’s disease. Vet. Microbiol. 77, 357–367. Elston, D.A., Moss, R., Boulinier, T., Arrowsmith, C., Lambin, X., 2001. Analysis of aggregation, a worked example: numbers of ticks on red grouse chicks. Parasitology 122, 563–569. Giese, S.B., Ahrens, P., 2000. Detection of Mycobacterium avium subsp. paratuberculosis in milk from clinically affected cows by PCR and culture. Vet. Microbiol. 77, 291–297. Grant, I.R., Hitchings, E.I., McCartney, A., Ferguson, F., Rowe, M.T., 2002. Effect of commercial-scale high-temperature, shorttime pasteurization on the viability of Mycobacterium paratuberculosis in naturally infected cows’ milk. Appl. Environ. Microbiol. 68, 602–607. Jakobsen, M.B., Alban, L., Nielsen, S.S., 2000. A cross-sectional study of paratuberculosis in 1155 Danish dairy cows. Prev. Vet. Med. 46, 15–27. Jorgensen, J.B., 1977. Survival of Mycobacterium paratuberculosis in slurry. Nord. Vet. Med. 29, 267–270. Kalis, C.H., Hesselink, J.W., Russchen, E.W., Barkema, H.W., Collins, M.T., Visser, I.J., 1999. Factors influencing the isolation of Mycobacterium avium subsp. paratuberculosis from bovine fecal samples. J. Vet. Diagn. Invest. 11, 345–351. Kormendy, B., 1994. The effect of vaccination on the prevalence of paratuberculosis in large dairy herds. Vet. Microbiol. 41, 117– 125. Lund, B.M., Donnelly, C.W., Rampling, A., 2000. Heat resistance of Mycobacterium paratuberculosis. Lett. Appl. Microbiol. 31, 184–185. Merkal, R.S., Larsen, A.B., Kopecky, K.E., 1968. Comparison of examination and test methods for early detection of paratuberculous cattle. Am. J. Vet. Res. 29, 1533–1538. Obasanjo, I.O., Grohn, Y.T., Mohammed, H.O., 1997. Farm factors associated with the presence of Mycobacterium paratuberculosis infection in dairy herds on the New York State Paratuberculosis Control Program. Prev. Vet. Med. 32, 243–251. Ott, S.L., Wells, S.J., Wagner, B.A., 1999. Herd-level economic losses associated with Johne’s disease on US dairy operations. Prev. Vet. Med. 40, 179–192.
263
Pavlik, I., Matlova, L., Bartl, J., Svastova, P., Dvorska, L., Whitlock, R., 2000. Parallel faecal and organ Mycobacterium avium subsp. paratuberculosis culture of different productivity types of cattle. Vet. Microbiol. 77, 309–324. Reichel, M.P., Kittelberger, R., Penrose, M.E., Meynell, R.M., Cousins, D., Ellis, T., Mutharia, L.M., Sugden, E.A., Johns, A.H., de Lisle, G.W., 1999. Comparison of serological tests and faecal culture for the detection of Mycobacterium avium subsp. paratuberculosis infection in cattle and analysis of the antigens involved. Vet. Microbiol. 66, 135–150. Sherman, D.M., Gay, J.M., Bouley, D.S., Nelson, G.H., 1990. Comparison of the complement-fixation and agar gel immunodiffusion tests for diagnosis of subclinical bovine paratuberculosis. Am. J. Vet. Res. 51, 461–465. Stabel, J.R., 2001. On-farm batch pasteurization destroys Mycobacterium paratuberculosis in waste milk. J. Dairy Sci. 84, 524–527. Sweeney, R.W., Whitlock, R.H., Hamir, A.N., Rosenberger, A.E., Herr, S.A., 1992. Isolation of Mycobacterium paratuberculosis after oral inoculation in uninfected cattle. Am. J. Vet. Res. 53, 1312–1314. Sweeney, R.W., Whitlock, R.H., Buckley, C.L., Spencer, P.A., 1995. Evaluation of a commercial enzyme-linked immunosorbent assay for the diagnosis of paratuberculosis in dairy cattle. J. Vet. Diagn. Invest. 7, 488–493. Venables, W.N., Ripley, B.D., 1999. Modern Applied Statistics with S-Plus. Springer-Verlag, New York. Wells, S.J., Wagner, B.A., 2000. Herd-level risk factors for infection with Mycobacterium paratuberculosis in US dairies and association between familiarity of the herd manager with the disease or prior diagnosis of the disease in that herd and use of preventive measures. J. Am. Vet. Med. Assoc. 216, 1450–1457. Wells, S.J., Whitlock, R.H., Lindeman, C.J., Fyock, T., 2002. Evaluation of bacteriologic culture of pooled fecal samples for detection of Mycobacterium paratuberculosis. Am. J. Vet. Res. 63, 1207–1211. Whipple, D.L., Kapke, P.A., Andersen, P.R., 1992. Comparison of a commercial DNA probe test and three cultivation procedures for detection of Mycobacterium paratuberculosis in bovine feces. J. Vet. Diagn. Invest. 4, 23–27. Whitlock, R.H., Buergelt, C., 1996. Preclinical and clinical manifestations of paratuberculosis (including pathology). Vet. Clin. North. Am: Food Anim. Pract. 12, 345–356. Whitlock, R.H., Rosenberger, A.E., 1994. Fecal culture protocol for Mycobacterium paratuberculosis: a recommended procedure. In: Proceedings of the 94th U.S. Animal Health Association, Denver, CO, pp. 280–285. Whitlock, R.H., Wells, S.J., Sweeney, R.W., Van Tiem, J., 2000. ELISA and fecal culture for paratuberculosis (Johne’s disease): sensitivity and specificity of each method. Vet. Microbiol. 77, 387–398. Whittington, R.J., Sergeant, E.S., 2001. Progress towards understanding the spread, detection and control of Mycobacterium avium subsp. paratuberculosis in animal populations. Aust. Vet. J. 79, 267–278. Wilson, K., Grenfell, B.T., Shaw, D.J., 1996. Analysis of aggregated parasite distributions: a comparison of methods. Funct. Ecol. 10, 592–601.