3.1 Descriptive study of animal population
The study involved 60 (75.94%) goat farms and 19 (24.1%) sheep farms; from which, a total of 277 (77.59%) goat samples and 80 (22.41%) sheep blood samples were collected and tested for brucellosis. The median ages of goat and sheep populations were 1.5 years each, while the median flock size of the goat and sheep farms was 48 and 100, respectively. The species wise herd characteristics of goat and sheep are depicted in Table 1. From the survey, 17.65% (12/68) of sheep were either purchased from nearby herds or brought from India, while 61% (105/172) of the goats were either collected from neighboring districts or brought from abroad. About ninety per cent of the goat herds included in the study were registered farms, but more than half of the sheep herds were not registered. Interestingly, an indigenous community mainly residing at the terai belt of south western Nepal maintained most sheep flocks, which was the primary means of their livelihoods.
Sero-prevalence of goat and sheep population
The flock level prevalence for sheep and goat farms was 30% (6/19; 95% CI: 12%, 54%) and 3.33% (2/60, 95%CI: 0.92%, 11.36 %) respectively. Of the total of 80 sheep samples tested, 12 (15%; 95% CI: 8.79-24.41), and among 277 goat samples tested 3 (1.1%; 95% CI: 0.37-3.14) were seropositive to Brucella (Table 1). There is significant difference between the proportion of sheep and goat populations that have seropositivity to brucellosis ( χ2 =29.78, p<0.001).
Antibodies to brucella were detected only in female goats, but in the sheep populations, a higher proportion of males 18.75% (3/16) were seropositive to brucella than females 14.1% (9/64). Only the local goat breeds, such as Khari, were seropositive to brucella by ELISA. Lampuchre is an indigenous sheep breed that had the highest burden of disease. The detailed illustrations of the sex-wise and breed wise comparison of seroprevalence of Brucella among goats and sheep by both RBT and ELISA are described in table 1.
There was a significant association (χ2= 28.29, p<0.001) between seropositivity of RBT and ELISA tests for brucellosis in sheep and goats [46] and an extremely high level of agreement
(κ = 0.95, 95% CI: 0.85-1, p<0.001 [35].
Table 1:Comparison of seroprevalence of Brucella among goats and sheep by sex and breed wise classification
Variables
|
Category
|
Total number (%)
|
RBT positive (%)
|
ELISA positive (%)
|
Overall Prevalence (95%CI)
|
Species
|
|
|
|
|
|
Goat
|
Male
|
65 (23.46)
|
0.00
|
0.00
|
1.1% (0.37-3.14)
|
|
Female
|
212 (76.53)
|
1.81% (5/277)
|
1.1 % (3/277)
|
|
|
|
|
|
|
Sheep
|
Male
|
16 (20)
|
18.75 % (3/16)
|
18.75 % (3/16)
|
15% (8.79-24.41)
|
|
Female
|
64 (80)
|
12.5% (8/64)
|
14.1% (9/64)
|
Breeds
|
|
|
|
|
|
Goat*
|
Local
|
135 (48.74)
|
2.22 % (3/135)
|
2.22 % (3/135)
|
1.1% (0.37-3.14)
|
|
Exotic
|
142 (51.26)
|
1.41 % (2/142)
|
0.00
|
|
|
|
|
|
|
Sheep
|
Lampuchre
|
75(93.75)
|
14.67% (11/75)
|
16% (12/75)
|
15% (8.79-24.41)
|
|
Baruwal
|
5 (6.25)
|
0.0
|
0.00
|
* Exotic breeds of goat included Boer and Jamunapari. Local breeds of goat included Khari and Terai.
Univariable regression analysis
The bivariate analysis of the sheep and goat data was depicted in Tables 2 and 3, respectively. Sheep greater than 1.5 years of age had significantly higher odds of brucellosis (OR= 4.29, 95%CI: 1.16, 20.63, p=0.0406) than the sheep of age ≤ 1.5 years. There was significantly higher odds of brucellosis among sheep when flock size was >100 (OR= 4.2, 95% CI: 1.19,15.91, p = 0.026) than the sheep herds of ≤100. Sheep that had parity greater than 1.5 were 4.11 more likely to be detected with brucellosis compared to sheep ≤ 1.5, but the result was statistically borderline significant (OR= 4.11, 95%CI: 0.98, 21.29, p = 0.055).
To accommodate for empty cells in the two by two contingency tables in the goat bivariate analysis a correction was made, with the addition of 0.5 to all entries and the odds ratios then calculated [37], [38-40] ](Table 3). In bivariate analysis, the only variable associated with seropositivity was that the goats that were taken for grazing had significantly higher odds (OR=14.5, 95% CI: 1.1, 283.9, p = 0.003) of detecting brucellosis compared to goats stall-fed at farms (Table 3).
Multivariable logistic regression analysis
The variables that qualified from the sheep data for multivariable analysis (p<0.20) were age category, gender, grazing system, disinfection process applied at the farm entry point. Similarly, for the goat data, the same sets of the variables were included in final firth multivariable logistic regression based on the cut off criteria of p<0.20.
In the multivariable regression analysis, sheep of older age (>1.5 years) had significantly higher odds (OR= 5.56, 95%CI: 11.39, 29.38, p = 0.02) of Brucella seropositivity compared to the younger sheep (≤1.5 years) (Table 4). The sheep farms of flock size greater than 100, had higher odds (OR = 4.74, 95% CI: 1.23, 20.32, p = 0.03) of brucella seropositivity than those of smaller farm size .
None of the variables included in the multivariate analysis were identified as significant risk factors for brucellosis in goat population after firth logistic regression . Goats from the frequent grazing herds had higher odds (OR = 13.82, 95% CI: 0.70, 272.20) of brucella than goats from isolated herds (Table 5), but this was not statistically significant (p = 0.08).
Table 2: Univariable analysis results of potential risk factors associated with sero-positivity of sheep population against Brucella spp.
Determinants
|
Total no of sheep
|
Brucella positive
|
Brucella negative
|
Odds ratio (OR)
|
95% CI
|
P value
|
Animal Origin
|
|
|
|
|
|
|
Purchased
|
12
|
3
|
9
|
2.19
|
(0.85, 2.21)
|
0.32
|
Home Breed
|
68
|
9
|
59
|
Ref
|
|
|
Age (median=1.5 years)
|
|
|
|
|
|
|
>1.5
|
37
|
9
|
28
|
4.29
|
(1.16, 20.63)
|
0.041*
|
≤1.5
|
43
|
3
|
40
|
Ref
|
|
|
Herd size (median= 100)
|
|
|
|
|
|
|
>100
|
24
|
7
|
17
|
4.2
|
(1.19,15.91)
|
0.026*
|
≤100
|
56
|
5
|
51
|
Ref
|
|
|
Parity (median=1)
|
|
|
|
|
|
|
>1
|
24
|
6
|
18
|
4.11
|
(0.98,21.29)
|
0.055
|
≤1
|
40
|
3
|
37
|
Ref
|
|
|
Gender
|
|
|
|
|
|
|
Male
|
16
|
3
|
13
|
1.41
|
(0.28,5.53)
|
0.646
|
Female
|
64
|
9
|
55
|
Ref
|
|
|
Grazing system
|
|
|
|
|
|
|
Yes
|
74
|
12
|
62
|
2.6
|
(0.12, 49.16)
|
0.154
|
No
|
6
|
0
|
6
|
Ref
|
|
|
Repeat breeding
|
|
|
|
|
|
|
Yes
|
11
|
3
|
8
|
2.94
|
(0.62,2.63)
|
0.199
|
No
|
53
|
6
|
47
|
Ref
|
|
|
*P value<0.05 means statistically significant
Table 3: Univariable analysis results of potential risk factors associated with sero-positivity of goat population against Brucella spp.
Determinants
|
Total no of goats
|
Brucella Positive
|
Brucella Negative
|
Odds Ratios (OR)
|
95% CI
|
P value
|
Animal Origin
|
|
|
|
|
|
|
Purchased
|
105
|
2
|
103
|
1.2
|
0.11, 26.11
|
0.87
|
Home bred
|
172
|
1
|
171
|
Ref
|
|
|
Age (median=1.5 years)
|
|
|
|
|
|
|
≤1.5
|
194
|
3
|
191
|
3.1
|
0.16, 59.74
|
0.12
|
>1.5
|
83
|
0
|
83
|
Ref
|
|
|
Herd size (median= 48)
|
|
|
|
|
|
|
≤48
|
140
|
3
|
137
|
7
|
0.36, 136.8
|
0.06
|
>48
|
137
|
0
|
137
|
Ref
|
|
|
Parity (median=1)
|
|
|
|
|
|
|
≤1
|
113
|
3
|
110
|
6.5
|
0.33, 127.2
|
0.04*
|
>1
|
102
|
0
|
102
|
Ref
|
|
|
Gender
|
|
|
|
|
|
|
Female
|
211
|
3
|
209
|
2.19
|
0.112, 42.9
|
0.34
|
Male
|
65
|
0
|
65
|
ref
|
|
|
Grazing system
|
|
|
|
|
|
|
Yes
|
92
|
3
|
89
|
14.5
|
1.1, 283.9
|
0.01*
|
No
|
185
|
0
|
185
|
Ref
|
|
|
Repeat breeding
|
|
|
|
|
|
|
Yes
|
42
|
1
|
41
|
2.1
|
0.18, 23.28
|
0.29
|
No
|
171
|
2
|
169
|
Ref
|
|
|
*P value<0.05 means statistically significant
Table 4: Multivariable analysis results of risk factors (p<0.05) associated with sero-positivity of sheep population against Brucella spp.
Determinants
|
Category
|
Coefficient
|
Standard Error
|
Odds ratio (OR
|
95% CI
|
P value
|
Age (median=1year)
|
>1.5
|
1.72
|
0.76
|
5.56
|
(1.39, 29.38)
|
0.02*
|
|
≤1.5
|
|
|
|
|
|
Herd size (median=100)
|
>100
|
1.56
|
0.70
|
4.74
|
(1.23, 20.32)
|
0.03*
|
|
≤100
|
|
|
|
|
|
|
|
|
|
|
|
|
*P value < 0.05 means statistically significant
Table 5: Multivariable analysis results of risk factors (p<0.05) associated with sero-positivity of goat population against Brucella spp.
Determinants
|
Category
|
Coefficient
|
Standard Error
|
Odds ratio (OR)
|
95% CI
|
P value
|
Age
(median=1 year)
|
>1.5
|
0.98
|
1.54
|
2.24
|
(0.12, 45.89)
|
0.60
|
|
≤1.5
|
|
|
|
|
|
Herd size
(median = 48)
|
>48
|
1.86
|
1.53
|
6.44
|
(0.32, 128.16)
|
0.22
|
|
≤48
|
|
|
|
|
|
Grazing System
|
Yes
|
2.63
|
1.52
|
13.82
|
(0.70, 272.20)
|
0.08 a
|
|
No
|
|
|
|
|
|
a: This variable has borderline significant p-value and could be a potential risk factor